Skip to contentSkip to conversation

target/ppc: PPC TCG Improvements (decodetree migrations + ISA 2.07 flag updates)

PATCH Superseded Chinmay Rath 6 participants 83 messages Aug 26 – Aug 27 Page 1 of 2

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover target/ppc: PPC TCG Improvements (decodetree migrations + ISA 2.07 flag updates)

This series primarily contains 3 changes :

1. PPC instructions movement to decodetree specification (Only patch 19 needs Rb)
Link to the original v1 : https://lore.kernel.org/qemu-devel/[email protected]/
Changes from v1 :
- Patch 19 commit log updated as suggested by Glenn.
- Retained Rbs received.

2. PPC ISA 2.07 flag cleanup and updates (Already reviewed)
Link to the original v2 : https://lore.kernel.org/qemu-devel/[email protected]/
This series is based on top the above decodetree series.
- No changes. Only Rbs added.

3. Stepping up as an active maintainer for PowerPC TCG CPUs module.

Aboorva Devarajan (3):
target/ppc: Move vector merge instructions to decodetree target/ppc: Move vector pack instructions to decodetree target/ppc: Move GPR atomic load/store instructions to decodetree

Amit Machhiwal (2):
target/ppc: Move fixed-point Shift insns to decodetree target/ppc: Move fixed-point byte-reversal store insns to decodetree

Chinmay Rath (9):
target/ppc: make do_ea_calc_ra available for 32 bit builds target/ppc: Replace PPC2_VSX207 flag with PPC2_ISA207 target/ppc: Use PPC2_ISA207 instead of PPC2_BCTAR_ISA207 target/ppc: Use PPC2_ISA207 instead of PPC2_LSQ_ISA207 target/ppc: Use PPC2_ISA207 instead of PPC2_ALTIVEC_207 target/ppc: Use PPC2_ISA207 instead of PPC2_ISA207S target/ppc: Reorder PPC2 flags
target/ppc: Add ICBT support for ISA version 2.07 target/ppc: Add self as maintainer for PowerPC TCG CPUs

Mukesh Kumar Chaurasiya (IBM) (2):
target/ppc: convert slw, srw instruction via decode spec target/ppc: convert sraw[i] instruction via decode spec

Nikhil Kumar Singh (3):
target/ppc: Migrate extswsli to decodetree target/ppc: Migrate atomic loads to decodetree target/ppc: Convert cache instructions to decodetree

Ojaswin Mujoo (7):
target/ppc: Move isync instruction to decodetree. target/ppc: Convert b{a, l, la} to decode tree target/ppc: move various conditional branch insns to decodetree target/ppc: Fix TRANS* macro variadic arguments handling target/ppc: Move wait instruction to decodetree target/ppc: Move sleep & friends to decodetree target/ppc: Refactor sleep and its variants to use a common helper

Shivang Upadhyay (4):
target/ppc: Move Condition Register access instructions to decodetree. target/ppc: Move Condition Register logical instructions to
decodetree.
target/ppc: Move Fixed-Point Load/Store String instructions to
decodetree.
target/ppc: Move VMX integer arithmetic and BCD instructions to
decodetree.

Shivani Nittor (1):
target/ppc: Convert mcrf to decode tree

Tanushree Shah (2):
target/ppc: Move rlwimi, rlwinm instructions to decodetree target/ppc: Move lmw, stmw instructions to decodetree

Utkarsh Verma (1):
target/ppc: Move st{b, h, w, d, q}cx instructions to decodetree

Vishal Chourasia (3):
target/ppc: Move mfmsr, mtmsr[d] instructions to decodetree target/ppc: Move byte-reverse instructions to decodetree target/ppc: Move system call and rfi instructions to decodetree

 MAINTAINERS                                   |    2 +-
 hw/ppc/spapr_caps.c                           |    2 +-
 linux-user/ppc/elfload.c                      |    6 +-
 target/ppc/cpu.h                              |   38 +-
 target/ppc/cpu_init.c                         |    6 +-
 target/ppc/cpu_init.h                         |    7 +-
 target/ppc/helper.h                           |  100 +-
 target/ppc/insn32.decode                      |  254 ++
 target/ppc/int_helper.c                       |   76 +-
 target/ppc/internal.h                         |    7 -
 target/ppc/mem_helper.c                       |   18 +-
 target/ppc/tcg-excp_helper.c                  |   18 +-
 target/ppc/translate.c                        | 2297 +++--------------
 target/ppc/translate/bhrb-impl.c.inc          |    4 +-
 target/ppc/translate/branch-impl.c.inc        |  183 +-
 target/ppc/translate/fixedpoint-impl.c.inc    |  938 ++++++-
 target/ppc/translate/fp-impl.c.inc            |    4 +-
 target/ppc/translate/misc-impl.c.inc          |  170 ++
 .../ppc/translate/processor-ctrl-impl.c.inc   |   38 +-
 target/ppc/translate/storage-ctrl-impl.c.inc  |  117 +
 target/ppc/translate/vmx-impl.c.inc           |  457 ++--
 target/ppc/translate/vmx-ops.c.inc            |   59 +-
 target/ppc/translate/vsx-impl.c.inc           |   56 +-
 target/ppc/translate/vsx-ops.c.inc            |   24 +-
 24 files changed, 2494 insertions(+), 2387 deletions(-)
Signature · 1 line
2.55.0
#1 lore raw

PATCH v2 1/37 target/ppc: Migrate extswsli to decodetree

From: Nikhil Kumar Singh <[email protected]>

Move EXTSWSLI to decodetree using the XS-form layout with a custom %xs_sh field to represent the fractured shift encoding.

This replaces legacy GEN_HANDLER-based implementations with a single trans_EXTSWSLI() handler.

The implementation intentionally operates directly on the destination GPR to preserve exact TCG output and avoid introducing temporary variables, matching legacy behavior.

Testing:
- Verified TCG equivalence with legacy implementation.

---
 target/ppc/insn32.decode |  8 ++++++++
 target/ppc/translate.c   | 43 ++++++++++++++--------------------------
 2 files changed, 23 insertions(+), 28 deletions(-)
target/ppc/insn32.decode +8 -0
index 0e9c68f2fb..342e65dadf 100644
@@ -117,6 +117,11 @@
 &X_sa_rc        rs ra rc
 @X_sa_rc        ...... rs:5 ra:5 ..... .......... rc:1          &X_sa_rc
 
+# XS-form with fractured shift amount
+%xs_sh           1:1 11:5
+&XS              rs ra sh rc
+@XS              ...... rs:5 ra:5 ..... ......... . rc:1        &XS sh=%xs_sh
+
 %x_frtp         22:4 !function=times_2
 %x_frap         17:4 !function=times_2
 %x_frbp         12:4 !function=times_2
@@ -1301,6 +1306,9 @@ XVF64GERPN      111011 ... -- .... 0 ..... 10111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNP      111011 ... -- .... 0 ..... 01111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 
+##Extend Sign Word and Shift Left Immediate XS-form
+EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
+
 ## Vector Division Instructions
 
 VDIVSW          000100 ..... ..... ..... 00110001011    @VX
target/ppc/translate.c +15 -28

61 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 1/37
#2 lore raw

Re: [PATCH v2 01/37] target/ppc: Migrate extswsli to decodetree

Quoted 85 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Move EXTSWSLI to decodetree using the XS-form layout with a custom %xs_sh field to represent the fractured shift encoding.

This replaces legacy GEN_HANDLER-based implementations with a single trans_EXTSWSLI() handler.

The implementation intentionally operates directly on the destination GPR to preserve exact TCG output and avoid introducing temporary variables, matching legacy behavior.

Testing:
- Verified TCG equivalence with legacy implementation.

---
 target/ppc/insn32.decode |  8 ++++++++
 target/ppc/translate.c   | 43 ++++++++++++++--------------------------
 2 files changed, 23 insertions(+), 28 deletions(-)
target/ppc/insn32.decode +8 -0
index 0e9c68f2fb..342e65dadf 100644
@@ -117,6 +117,11 @@
 &X_sa_rc        rs ra rc
 @X_sa_rc        ...... rs:5 ra:5 ..... .......... rc:1          &X_sa_rc
 
+# XS-form with fractured shift amount
+%xs_sh           1:1 11:5
+&XS              rs ra sh rc
+@XS              ...... rs:5 ra:5 ..... ......... . rc:1        &XS sh=%xs_sh
+
 %x_frtp         22:4 !function=times_2
 %x_frap         17:4 !function=times_2
 %x_frbp         12:4 !function=times_2
@@ -1301,6 +1306,9 @@ XVF64GERPN      111011 ... -- .... 0 ..... 10111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNP      111011 ... -- .... 0 ..... 01111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 
+##Extend Sign Word and Shift Left Immediate XS-form
+EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
+
 ## Vector Division Instructions
 
 VDIVSW          000100 ..... ..... ..... 00110001011    @VX
target/ppc/translate.c +0 -24
index 06ed2adf10..e627f48f9e 100644
@@ -2470,30 +2470,6 @@ static void gen_sradi1(DisasContext *ctx)
     gen_sradi(ctx, 1);
 }
 
-/* extswsli & extswsli. */
-static inline void gen_extswsli(DisasContext *ctx, int n)
-{
-    int sh = SH(ctx->opcode) + (n << 5);
-    TCGv dst = cpu_gpr[rA(ctx->opcode)];
-    TCGv src = cpu_gpr[rS(ctx->opcode)];
-
-    tcg_gen_ext32s_tl(dst, src);
-    tcg_gen_shli_tl(dst, dst, sh);
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, dst);
-    }
-}
-
-static void gen_extswsli0(DisasContext *ctx)
-{
-    gen_extswsli(ctx, 0);
-}
-
-static void gen_extswsli1(DisasContext *ctx)
-{
-    gen_extswsli(ctx, 1);
-}
-
 /* srd & srd. */
 static void gen_srd(DisasContext *ctx)
 {

target/ppc/translate.c static bool resolve_PLS_D(DisasContext *ctx, arg_D *d, arg_PLS_D *a) · line 5732

Quoted 7 lines
target/ppc/translate.c +4 -0
@@ -5750,6 +5726,21 @@ static bool resolve_PLS_D(DisasContext *ctx, arg_D *d, arg_PLS_D *a)
     return true;
 }
 
+static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a)
+{
+    REQUIRE_64BIT(ctx);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA300);

TRANS64_FLAGS2 exists and could technically be used here, but it would require splitting into a do_extswsli helper + a macro expansion, adding an extra indirection with no benefit since the logic is unique to EXTSWSLI and not shared with any other instruction. The hand-written trans_EXTSWSLI looks the right choice.

Thanks,
Amit

#3 lore raw

PATCH v2 2/37 target/ppc: Migrate atomic loads to decodetree

In reply to Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Migrate load-and-reserve instructions (lbarx, lharx, lwarx, ldarx, lqarx) to decodetree using the X-form layout.

A shared helper (do_load_locked) is introduced for standard-width variants, while LQARX is handled separately due to its 128-bit semantics and register pairing constraints.

The implementation preserves legacy behavior, including:
- Reservation granularity and overwrite semantics
- Alignment requirements
- Invalid instruction cases for LQARX

Testing:
- Verified TCG equivalence with legacy implementation

---
 target/ppc/insn32.decode |   7 +++
 target/ppc/translate.c   | 124 +++++++++++++++++++--------------------
 2 files changed, 69 insertions(+), 62 deletions(-)
target/ppc/insn32.decode +7 -0
index 342e65dadf..26948e08a7 100644
@@ -1309,6 +1309,13 @@ XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 ##Extend Sign Word and Shift Left Immediate XS-form
 EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
 
+## Load and Reserve Instructions
+LBARX           011111 ..... ..... ..... 0000110100 .   @X_rc
+LHARX           011111 ..... ..... ..... 0001110100 .   @X_rc
+LWARX           011111 ..... ..... ..... 0000010100 .   @X_rc
+LDARX           011111 ..... ..... ..... 0001010100 .   @X_rc
+LQARX           011111 ..... ..... ..... 0100010100 .   @X_rc
+
 ## Vector Division Instructions
 
 VDIVSW          000100 ..... ..... ..... 00110001011    @VX
target/ppc/translate.c +62 -62

157 lines not on this page · raw message

Signature · 1 line
2.55.0
#4 lore raw

Re: [PATCH v2 02/37] target/ppc: Migrate atomic loads to decodetree

Quoted 144 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Migrate load-and-reserve instructions (lbarx, lharx, lwarx, ldarx, lqarx) to decodetree using the X-form layout.

A shared helper (do_load_locked) is introduced for standard-width variants, while LQARX is handled separately due to its 128-bit semantics and register pairing constraints.

The implementation preserves legacy behavior, including:
- Reservation granularity and overwrite semantics
- Alignment requirements
- Invalid instruction cases for LQARX

Testing:
- Verified TCG equivalence with legacy implementation

---
 target/ppc/insn32.decode |   7 +++
 target/ppc/translate.c   | 124 +++++++++++++++++++--------------------
 2 files changed, 69 insertions(+), 62 deletions(-)
target/ppc/insn32.decode +7 -0
index 342e65dadf..26948e08a7 100644
@@ -1309,6 +1309,13 @@ XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 ##Extend Sign Word and Shift Left Immediate XS-form
 EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
 
+## Load and Reserve Instructions
+LBARX           011111 ..... ..... ..... 0000110100 .   @X_rc
+LHARX           011111 ..... ..... ..... 0001110100 .   @X_rc
+LWARX           011111 ..... ..... ..... 0000010100 .   @X_rc
+LDARX           011111 ..... ..... ..... 0001010100 .   @X_rc
+LQARX           011111 ..... ..... ..... 0100010100 .   @X_rc
+
 ## Vector Division Instructions
 
 VDIVSW          000100 ..... ..... ..... 00110001011    @VX
target/ppc/translate.c +19 -57
index e627f48f9e..cc9287dcc5 100644
@@ -2947,30 +2947,6 @@ static void gen_isync(DisasContext *ctx)
     ctx->base.is_jmp = DISAS_EXIT_UPDATE;
 }
 
-static void gen_load_locked(DisasContext *ctx, MemOp memop)
-{
-    TCGv gpr = cpu_gpr[rD(ctx->opcode)];
-    TCGv t0 = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_RES);
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(gpr, t0, ctx->mem_idx, DEF_MEMOP(memop) | MO_ALIGN);
-    tcg_gen_mov_tl(cpu_reserve, t0);
-    tcg_gen_movi_tl(cpu_reserve_length, memop_size(memop));
-    tcg_gen_mov_tl(cpu_reserve_val, gpr);
-}
-
-#define LARX(name, memop)                  \
-static void gen_##name(DisasContext *ctx)  \
-{                                          \
-    gen_load_locked(ctx, memop);           \
-}
-
-/* lwarx */
-LARX(lbarx, MO_UB)
-LARX(lharx, MO_UW)
-LARX(lwarx, MO_UL)
-
 static void gen_fetch_inc_conditional(DisasContext *ctx, MemOp memop,
                                       TCGv EA, TCGCond cond, int addend)
 {
@@ -3219,42 +3195,9 @@ STCX(sthcx_, MO_UW)
 STCX(stwcx_, MO_UL)
 
 #if defined(TARGET_PPC64)
-/* ldarx */
-LARX(ldarx, MO_UQ)
 /* stdcx. */
 STCX(stdcx_, MO_UQ)
 
-/* lqarx */
-static void gen_lqarx(DisasContext *ctx)
-{
-    int rd = rD(ctx->opcode);
-    TCGv EA, hi, lo;
-    TCGv_i128 t16;
-
-    if (unlikely((rd & 1) || (rd == rA(ctx->opcode)) ||
-                 (rd == rB(ctx->opcode)))) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-        return;
-    }
-
-    gen_set_access_type(ctx, ACCESS_RES);
-    EA = tcg_temp_new();
-    gen_addr_reg_index(ctx, EA);
-
-    /* Note that the low part is always in RD+1, even in LE mode.  */
-    lo = cpu_gpr[rd + 1];
-    hi = cpu_gpr[rd];
-
-    t16 = tcg_temp_new_i128();
-    tcg_gen_qemu_ld_i128(t16, EA, ctx->mem_idx, DEF_MEMOP(MO_128 | MO_ALIGN));
-    tcg_gen_extr_i128_i64(lo, hi, t16);
-
-    tcg_gen_mov_tl(cpu_reserve, EA);
-    tcg_gen_movi_tl(cpu_reserve_length, 16);
-    tcg_gen_st_tl(hi, tcg_env, offsetof(CPUPPCState, reserve_val));
-    tcg_gen_st_tl(lo, tcg_env, offsetof(CPUPPCState, reserve_val2));
-}
-
 /* stqcx. */
 static void gen_stqcx_(DisasContext *ctx)
 {
@@ -5741,6 +5684,68 @@ static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a)
     return true;
 }
 
+/*
+ * Load-and-reserve core
+ */
+static bool do_load_locked(DisasContext *ctx, arg_X_rc *a, MemOp memop)
+{
+    TCGv EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    TCGv gpr = cpu_gpr[a->rt];
+
+    gen_set_access_type(ctx, ACCESS_RES);
+
+    tcg_gen_qemu_ld_tl(gpr, EA, ctx->mem_idx, memop | MO_ALIGN);
+
+    tcg_gen_mov_tl(cpu_reserve, EA);
+    tcg_gen_movi_tl(cpu_reserve_length, memop_size(memop));
+
+    tcg_gen_mov_tl(cpu_reserve_val, gpr);
+
+    return true;
+}

target/ppc/translate.c static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a) · line 5715

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+TRANS_FLAGS2(ATOMIC_ISA206, LBARX, do_load_locked, DEF_MEMOP(MO_UB))
+TRANS_FLAGS2(ATOMIC_ISA206, LHARX, do_load_locked, DEF_MEMOP(MO_UW))
+TRANS(LWARX, do_load_locked, DEF_MEMOP(MO_UL))
+TRANS64(LDARX, do_load_locked, DEF_MEMOP(MO_UQ))
+
+static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a)
+{
+    REQUIRE_64BIT(ctx);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207S);

The old GEN_HANDLER_E used PPC2_LSQ_ISA207, but the new code uses PPC2_ISA207S . In the current QEMU CPU model set this makes no practical difference — every CPU that sets PPC2_ISA207S also sets PPC2_LSQ_ISA207. However, PPC2_LSQ_ISA207 is the precise flag for this feature and keeps lqarx symmetric with its companion stqcx_ (line 5870) which correctly continues to use PPC2_LSQ_ISA207. Please use REQUIRE_INSNS_FLAGS2(ctx, LSQ_ISA207) for consistency and to guard against future CPU models that may set these flags independently

Quoted 15 lines
target/ppc/translate.c +15 -0
+#if defined(TARGET_PPC64)
+    TCGv EA;
+    TCGv_i128 t16;
+    /* Must use even register and avoid overlap */
+    if (unlikely((a->rt & 1) || (a->rt == a->ra) || (a->rt == a->rb))) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_RES);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    t16 = tcg_temp_new_i128();
+
+    tcg_gen_qemu_ld_i128(t16, EA, ctx->mem_idx,
+                         DEF_MEMOP(MO_128 | MO_ALIGN));

target/ppc/translate.c static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a) · line 5740

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+    tcg_gen_extr_i128_i64(cpu_gpr[a->rt + 1],
+                         cpu_gpr[a->rt],
+                         t16);
+
+    tcg_gen_mov_tl(cpu_reserve, EA);
+    tcg_gen_movi_tl(cpu_reserve_length, 16);
+
+    tcg_gen_st_i64(cpu_gpr[a->rt],
+                   tcg_env, offsetof(CPUPPCState, reserve_val));

All these multi-line calls can be collapsed to single lines — they are within 80 characters respectively and fit within the 80-column limit. The only exception is tcg_gen_st_i64(..., reserve_val2) at 85 chars, which legitimately needs wrapping. The unnecessary line splits add visual noise without any readability benefit.

Quoted 33 lines
target/ppc/translate.c +8 -5
+    tcg_gen_st_i64(cpu_gpr[a->rt + 1],
+                   tcg_env, offsetof(CPUPPCState, reserve_val2));
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+
 #include "translate/fixedpoint-impl.c.inc"
 
 #include "translate/fp-impl.c.inc"
@@ -5853,9 +5858,6 @@ GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 GEN_HANDLER(isync, 0x13, 0x16, 0x04, 0x03FFF801, PPC_MEM),
-GEN_HANDLER_E(lbarx, 0x1F, 0x14, 0x01, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
-GEN_HANDLER_E(lharx, 0x1F, 0x14, 0x03, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
-GEN_HANDLER(lwarx, 0x1F, 0x14, 0x00, 0x00000000, PPC_RES),
 GEN_HANDLER_E(lwat, 0x1F, 0x06, 0x12, 0x00000001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(stwat, 0x1F, 0x06, 0x16, 0x00000001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(stbcx_, 0x1F, 0x16, 0x15, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
@@ -5864,8 +5866,6 @@ GEN_HANDLER2(stwcx_, "stwcx.", 0x1F, 0x16, 0x04, 0x00000000, PPC_RES),
 #if defined(TARGET_PPC64)
 GEN_HANDLER_E(ldat, 0x1F, 0x06, 0x13, 0x00000001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(stdat, 0x1F, 0x06, 0x17, 0x00000001, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(ldarx, 0x1F, 0x14, 0x02, 0x00000000, PPC_64B),
-GEN_HANDLER_E(lqarx, 0x1F, 0x14, 0x08, 0, PPC_NONE, PPC2_LSQ_ISA207),
 GEN_HANDLER2(stdcx_, "stdcx.", 0x1F, 0x16, 0x06, 0x00000000, PPC_64B),
 GEN_HANDLER_E(stqcx_, 0x1F, 0x16, 0x05, 0, PPC_NONE, PPC2_LSQ_ISA207),
 #endif
Signature · 1 line
2.55.0
#5 lore raw

Re: [PATCH v2 02/37] target/ppc: Migrate atomic loads to decodetree

Quoted 155 lines · Amit Machhiwal
Quoted 144 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Migrate load-and-reserve instructions (lbarx, lharx, lwarx, ldarx, lqarx) to decodetree using the X-form layout.

A shared helper (do_load_locked) is introduced for standard-width variants, while LQARX is handled separately due to its 128-bit semantics and register pairing constraints.

The implementation preserves legacy behavior, including:
- Reservation granularity and overwrite semantics
- Alignment requirements
- Invalid instruction cases for LQARX

Testing:
- Verified TCG equivalence with legacy implementation

---
  target/ppc/insn32.decode |   7 +++
  target/ppc/translate.c   | 124 +++++++++++++++++++--------------------
  2 files changed, 69 insertions(+), 62 deletions(-)
target/ppc/insn32.decode +7 -0
index 342e65dadf..26948e08a7 100644
@@ -1309,6 +1309,13 @@ XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
  ##Extend Sign Word and Shift Left Immediate XS-form
  EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
 
+## Load and Reserve Instructions
+LBARX           011111 ..... ..... ..... 0000110100 .   @X_rc
+LHARX           011111 ..... ..... ..... 0001110100 .   @X_rc
+LWARX           011111 ..... ..... ..... 0000010100 .   @X_rc
+LDARX           011111 ..... ..... ..... 0001010100 .   @X_rc
+LQARX           011111 ..... ..... ..... 0100010100 .   @X_rc
+
  ## Vector Division Instructions
 
  VDIVSW          000100 ..... ..... ..... 00110001011    @VX
target/ppc/translate.c +19 -57
index e627f48f9e..cc9287dcc5 100644
@@ -2947,30 +2947,6 @@ static void gen_isync(DisasContext *ctx)
      ctx->base.is_jmp = DISAS_EXIT_UPDATE;
  }
 
-static void gen_load_locked(DisasContext *ctx, MemOp memop)
-{
-    TCGv gpr = cpu_gpr[rD(ctx->opcode)];
-    TCGv t0 = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_RES);
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(gpr, t0, ctx->mem_idx, DEF_MEMOP(memop) | MO_ALIGN);
-    tcg_gen_mov_tl(cpu_reserve, t0);
-    tcg_gen_movi_tl(cpu_reserve_length, memop_size(memop));
-    tcg_gen_mov_tl(cpu_reserve_val, gpr);
-}
-
-#define LARX(name, memop)                  \
-static void gen_##name(DisasContext *ctx)  \
-{                                          \
-    gen_load_locked(ctx, memop);           \
-}
-
-/* lwarx */
-LARX(lbarx, MO_UB)
-LARX(lharx, MO_UW)
-LARX(lwarx, MO_UL)
-
  static void gen_fetch_inc_conditional(DisasContext *ctx, MemOp memop,
                                        TCGv EA, TCGCond cond, int addend)
  {
@@ -3219,42 +3195,9 @@ STCX(sthcx_, MO_UW)
  STCX(stwcx_, MO_UL)
 
  #if defined(TARGET_PPC64)
-/* ldarx */
-LARX(ldarx, MO_UQ)
  /* stdcx. */
  STCX(stdcx_, MO_UQ)
 
-/* lqarx */
-static void gen_lqarx(DisasContext *ctx)
-{
-    int rd = rD(ctx->opcode);
-    TCGv EA, hi, lo;
-    TCGv_i128 t16;
-
-    if (unlikely((rd & 1) || (rd == rA(ctx->opcode)) ||
-                 (rd == rB(ctx->opcode)))) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-        return;
-    }
-
-    gen_set_access_type(ctx, ACCESS_RES);
-    EA = tcg_temp_new();
-    gen_addr_reg_index(ctx, EA);
-
-    /* Note that the low part is always in RD+1, even in LE mode.  */
-    lo = cpu_gpr[rd + 1];
-    hi = cpu_gpr[rd];
-
-    t16 = tcg_temp_new_i128();
-    tcg_gen_qemu_ld_i128(t16, EA, ctx->mem_idx, DEF_MEMOP(MO_128 | MO_ALIGN));
-    tcg_gen_extr_i128_i64(lo, hi, t16);
-
-    tcg_gen_mov_tl(cpu_reserve, EA);
-    tcg_gen_movi_tl(cpu_reserve_length, 16);
-    tcg_gen_st_tl(hi, tcg_env, offsetof(CPUPPCState, reserve_val));
-    tcg_gen_st_tl(lo, tcg_env, offsetof(CPUPPCState, reserve_val2));
-}
-
  /* stqcx. */
  static void gen_stqcx_(DisasContext *ctx)
  {
@@ -5741,6 +5684,68 @@ static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a)
      return true;
  }
 
+/*
+ * Load-and-reserve core
+ */
+static bool do_load_locked(DisasContext *ctx, arg_X_rc *a, MemOp memop)
+{
+    TCGv EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    TCGv gpr = cpu_gpr[a->rt];
+
+    gen_set_access_type(ctx, ACCESS_RES);
+
+    tcg_gen_qemu_ld_tl(gpr, EA, ctx->mem_idx, memop | MO_ALIGN);
+
+    tcg_gen_mov_tl(cpu_reserve, EA);
+    tcg_gen_movi_tl(cpu_reserve_length, memop_size(memop));
+
+    tcg_gen_mov_tl(cpu_reserve_val, gpr);
+
+    return true;
+}

target/ppc/translate.c static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a) · line 5715

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+TRANS_FLAGS2(ATOMIC_ISA206, LBARX, do_load_locked, DEF_MEMOP(MO_UB))
+TRANS_FLAGS2(ATOMIC_ISA206, LHARX, do_load_locked, DEF_MEMOP(MO_UW))
+TRANS(LWARX, do_load_locked, DEF_MEMOP(MO_UL))
+TRANS64(LDARX, do_load_locked, DEF_MEMOP(MO_UQ))
+
+static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a)
+{
+    REQUIRE_64BIT(ctx);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207S);
Quoted 8 lines

The old GEN_HANDLER_E used PPC2_LSQ_ISA207, but the new code uses PPC2_ISA207S . In the current QEMU CPU model set this makes no practical difference — every CPU that sets PPC2_ISA207S also sets PPC2_LSQ_ISA207. However, PPC2_LSQ_ISA207 is the precise flag for this feature and keeps lqarx symmetric with its companion stqcx_ (line 5870) which correctly continues to use PPC2_LSQ_ISA207. Please use REQUIRE_INSNS_FLAGS2(ctx, LSQ_ISA207) for consistency and to guard against future CPU models that may set these flags independently

Hi Amit,

Nice catch ! Thanks !
In the later patches 32 and 34 in this series, I have consolidated all ISA 2.07 flags into a single flag. Hence both PPC2_LSQ_ISA207 and PPC2_ISA207S are replaced by PPC2_ISA207, hence this is automatically taken care of.

Regards,
Chinmay

Quoted 65 lines
Quoted 16 lines
target/ppc/translate.c +15 -0
+#if defined(TARGET_PPC64)
+    TCGv EA;
+    TCGv_i128 t16;
+    /* Must use even register and avoid overlap */
+    if (unlikely((a->rt & 1) || (a->rt == a->ra) || (a->rt == a->rb))) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_RES);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    t16 = tcg_temp_new_i128();
+
+    tcg_gen_qemu_ld_i128(t16, EA, ctx->mem_idx,
+                         DEF_MEMOP(MO_128 | MO_ALIGN));

target/ppc/translate.c static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a) · line 5740

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+    tcg_gen_extr_i128_i64(cpu_gpr[a->rt + 1],
+                         cpu_gpr[a->rt],
+                         t16);
+
+    tcg_gen_mov_tl(cpu_reserve, EA);
+    tcg_gen_movi_tl(cpu_reserve_length, 16);
+
+    tcg_gen_st_i64(cpu_gpr[a->rt],
+                   tcg_env, offsetof(CPUPPCState, reserve_val));

All these multi-line calls can be collapsed to single lines — they are within 80 characters respectively and fit within the 80-column limit. The only exception is tcg_gen_st_i64(..., reserve_val2) at 85 chars, which legitimately needs wrapping. The unnecessary line splits add visual noise without any readability benefit.

Quoted 33 lines
target/ppc/translate.c +8 -5
+    tcg_gen_st_i64(cpu_gpr[a->rt + 1],
+                   tcg_env, offsetof(CPUPPCState, reserve_val2));
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+
  #include "translate/fixedpoint-impl.c.inc"
 
  #include "translate/fp-impl.c.inc"
@@ -5853,9 +5858,6 @@ GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 0x00000001, PPC_STRING),
  GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
  GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
  GEN_HANDLER(isync, 0x13, 0x16, 0x04, 0x03FFF801, PPC_MEM),
-GEN_HANDLER_E(lbarx, 0x1F, 0x14, 0x01, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
-GEN_HANDLER_E(lharx, 0x1F, 0x14, 0x03, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
-GEN_HANDLER(lwarx, 0x1F, 0x14, 0x00, 0x00000000, PPC_RES),
  GEN_HANDLER_E(lwat, 0x1F, 0x06, 0x12, 0x00000001, PPC_NONE, PPC2_ISA300),
  GEN_HANDLER_E(stwat, 0x1F, 0x06, 0x16, 0x00000001, PPC_NONE, PPC2_ISA300),
  GEN_HANDLER_E(stbcx_, 0x1F, 0x16, 0x15, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
@@ -5864,8 +5866,6 @@ GEN_HANDLER2(stwcx_, "stwcx.", 0x1F, 0x16, 0x04, 0x00000000, PPC_RES),
  #if defined(TARGET_PPC64)
  GEN_HANDLER_E(ldat, 0x1F, 0x06, 0x13, 0x00000001, PPC_NONE, PPC2_ISA300),
  GEN_HANDLER_E(stdat, 0x1F, 0x06, 0x17, 0x00000001, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(ldarx, 0x1F, 0x14, 0x02, 0x00000000, PPC_64B),
-GEN_HANDLER_E(lqarx, 0x1F, 0x14, 0x08, 0, PPC_NONE, PPC2_LSQ_ISA207),
  GEN_HANDLER2(stdcx_, "stdcx.", 0x1F, 0x16, 0x06, 0x00000000, PPC_64B),
  GEN_HANDLER_E(stqcx_, 0x1F, 0x16, 0x05, 0, PPC_NONE, PPC2_LSQ_ISA207),
  #endif
Signature · 1 line
2.55.0
#6 lore raw

Re: [PATCH v2 02/37] target/ppc: Migrate atomic loads to decodetree

Quoted 166 lines · Chinmay Rath
Quoted 155 lines · Amit Machhiwal
Quoted 144 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Migrate load-and-reserve instructions (lbarx, lharx, lwarx, ldarx, lqarx) to decodetree using the X-form layout.

A shared helper (do_load_locked) is introduced for standard-width variants, while LQARX is handled separately due to its 128-bit semantics and register pairing constraints.

The implementation preserves legacy behavior, including:
   - Reservation granularity and overwrite semantics
   - Alignment requirements
   - Invalid instruction cases for LQARX

Testing:
   - Verified TCG equivalence with legacy implementation

---
  target/ppc/insn32.decode |   7 +++
  target/ppc/translate.c   | 124
+++++++++++++++++++--------------------

  2 files changed, 69 insertions(+), 62 deletions(-)
target/ppc/insn32.decode +0 -0
index 342e65dadf..26948e08a7 100644
@@ -1309,6 +1309,13 @@ XVF64GERNN      111011 ... -- .... 0 ..... 

11111010 ..-  @XX3_at xa=%xx_xa_pair
  ##Extend Sign Word and Shift Left Immediate XS-form EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS +## Load and Reserve Instructions

target/ppc/insn32.decode +6 -0
+LBARX           011111 ..... ..... ..... 0000110100 .   @X_rc
+LHARX           011111 ..... ..... ..... 0001110100 .   @X_rc
+LWARX           011111 ..... ..... ..... 0000010100 .   @X_rc
+LDARX           011111 ..... ..... ..... 0001010100 .   @X_rc
+LQARX           011111 ..... ..... ..... 0100010100 .   @X_rc
+

  ## Vector Division Instructions
    VDIVSW          000100 ..... ..... ..... 00110001011 @VX

target/ppc/translate.c +0 -23
index e627f48f9e..cc9287dcc5 100644
@@ -2947,30 +2947,6 @@ static void gen_isync(DisasContext *ctx)
       ctx->base.is_jmp = DISAS_EXIT_UPDATE;
   }
   -static void gen_load_locked(DisasContext *ctx, MemOp memop)
-{
-    TCGv gpr = cpu_gpr[rD(ctx->opcode)];
-    TCGv t0 = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_RES);
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(gpr, t0, ctx->mem_idx, DEF_MEMOP(memop) | 
 MO_ALIGN);
-    tcg_gen_mov_tl(cpu_reserve, t0);
-    tcg_gen_movi_tl(cpu_reserve_length, memop_size(memop));
-    tcg_gen_mov_tl(cpu_reserve_val, gpr);
-}
-
-#define LARX(name, memop)                  \
-static void gen_##name(DisasContext *ctx)  \
-{                                          \
-    gen_load_locked(ctx, memop);           \
-}
-
-/* lwarx */
-LARX(lbarx, MO_UB)
-LARX(lharx, MO_UW)
-LARX(lwarx, MO_UL)
-
   static void gen_fetch_inc_conditional(DisasContext *ctx, MemOp memop,
                                         TCGv EA, TCGCond cond, int addend)

  {

target/ppc/translate.c +0 -2
@@ -3219,42 +3195,9 @@ STCX(sthcx_, MO_UW)
   STCX(stwcx_, MO_UL)
     #if defined(TARGET_PPC64)
-/* ldarx */
-LARX(ldarx, MO_UQ)
   /* stdcx. */
   STCX(stdcx_, MO_UQ)

  -/* lqarx */

target/ppc/translate.c +0 -22
-static void gen_lqarx(DisasContext *ctx)
-{
-    int rd = rD(ctx->opcode);
-    TCGv EA, hi, lo;
-    TCGv_i128 t16;
-
-    if (unlikely((rd & 1) || (rd == rA(ctx->opcode)) ||
-                 (rd == rB(ctx->opcode)))) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-        return;
-    }
-
-    gen_set_access_type(ctx, ACCESS_RES);
-    EA = tcg_temp_new();
-    gen_addr_reg_index(ctx, EA);
-
-    /* Note that the low part is always in RD+1, even in LE mode.  */
-    lo = cpu_gpr[rd + 1];
-    hi = cpu_gpr[rd];
-
-    t16 = tcg_temp_new_i128();
-    tcg_gen_qemu_ld_i128(t16, EA, ctx->mem_idx, DEF_MEMOP(MO_128 | 

MO_ALIGN));
-    tcg_gen_extr_i128_i64(lo, hi, t16);

target/ppc/translate.c +0 -7
-
-    tcg_gen_mov_tl(cpu_reserve, EA);
-    tcg_gen_movi_tl(cpu_reserve_length, 16);
-    tcg_gen_st_tl(hi, tcg_env, offsetof(CPUPPCState, reserve_val));
-    tcg_gen_st_tl(lo, tcg_env, offsetof(CPUPPCState, reserve_val2));
-}
-

  /* stqcx. */
  static void gen_stqcx_(DisasContext *ctx)
  {

target/ppc/translate.c +0 -0
@@ -5741,6 +5684,68 @@ static bool trans_EXTSWSLI(DisasContext *ctx, 
 arg_XS *a)
       return true;
   }

  +/*
+ * Load-and-reserve core
+ */
+static bool do_load_locked(DisasContext *ctx, arg_X_rc *a, MemOp memop)

target/ppc/translate.c +15 -0
+{
+    TCGv EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    TCGv gpr = cpu_gpr[a->rt];
+
+    gen_set_access_type(ctx, ACCESS_RES);
+
+    tcg_gen_qemu_ld_tl(gpr, EA, ctx->mem_idx, memop | MO_ALIGN);
+
+    tcg_gen_mov_tl(cpu_reserve, EA);
+    tcg_gen_movi_tl(cpu_reserve_length, memop_size(memop));
+
+    tcg_gen_mov_tl(cpu_reserve_val, gpr);
+
+    return true;
+}

target/ppc/translate.c static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a) · line 5715

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+TRANS_FLAGS2(ATOMIC_ISA206, LBARX, do_load_locked, DEF_MEMOP(MO_UB))
+TRANS_FLAGS2(ATOMIC_ISA206, LHARX, do_load_locked, DEF_MEMOP(MO_UW))
+TRANS(LWARX, do_load_locked, DEF_MEMOP(MO_UL))
+TRANS64(LDARX, do_load_locked, DEF_MEMOP(MO_UQ))
+
+static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a)
+{
+    REQUIRE_64BIT(ctx);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207S);
Quoted 8 lines

The old GEN_HANDLER_E used PPC2_LSQ_ISA207, but the new code uses PPC2_ISA207S . In the current QEMU CPU model set this makes no practical difference — every CPU that sets PPC2_ISA207S also sets PPC2_LSQ_ISA207. However, PPC2_LSQ_ISA207 is the precise flag for this feature and keeps lqarx symmetric with its companion stqcx_ (line 5870) which correctly continues to use PPC2_LSQ_ISA207. Please use REQUIRE_INSNS_FLAGS2(ctx, LSQ_ISA207) for consistency and to guard against future CPU models that may set these flags independently

Quoted 7 lines

Hi Amit,

Nice catch ! Thanks !
In the later patches 32 and 34 in this series, I have consolidated all ISA 2.07 flags into a single flag. Hence both PPC2_LSQ_ISA207 and PPC2_ISA207S are replaced by PPC2_ISA207, hence this is automatically taken care of.

Regards,
Chinmay

++ Amit, IDK how Thunderbird omitted you out of the sent list.

Quoted 70 lines
Quoted 70 lines
Quoted 18 lines

+#if defined(TARGET_PPC64)
+    TCGv EA;
+    TCGv_i128 t16;
+    /* Must use even register and avoid overlap */ +    if (unlikely((a->rt & 1) || (a->rt == a->ra) || (a->rt == a->rb))) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL); +        return true;
+    }

target/ppc/translate.c +7 -0
+
+    gen_set_access_type(ctx, ACCESS_RES);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    t16 = tcg_temp_new_i128();
+
+    tcg_gen_qemu_ld_i128(t16, EA, ctx->mem_idx,
+                         DEF_MEMOP(MO_128 | MO_ALIGN));

target/ppc/translate.c static bool trans_EXTSWSLI(DisasContext *ctx, arg_XS *a) · line 5740

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+    tcg_gen_extr_i128_i64(cpu_gpr[a->rt + 1],
+                         cpu_gpr[a->rt],
+                         t16);
+
+    tcg_gen_mov_tl(cpu_reserve, EA);
+    tcg_gen_movi_tl(cpu_reserve_length, 16);
+
+    tcg_gen_st_i64(cpu_gpr[a->rt],
+                   tcg_env, offsetof(CPUPPCState, reserve_val));

All these multi-line calls can be collapsed to single lines — they are within 80 characters respectively and fit within the 80-column limit. The only exception is tcg_gen_st_i64(..., reserve_val2) at 85 chars, which legitimately needs wrapping. The unnecessary line splits add visual noise without any readability benefit.

Quoted 36 lines

+    tcg_gen_st_i64(cpu_gpr[a->rt + 1],
+                   tcg_env, offsetof(CPUPPCState, reserve_val2));

target/ppc/translate.c +6 -0
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+

  #include "translate/fixedpoint-impl.c.inc"
    #include "translate/fp-impl.c.inc"

target/ppc/translate.c +0 -3
@@ -5853,9 +5858,6 @@ GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 
 0x00000001, PPC_STRING),
   GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
   GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
   GEN_HANDLER(isync, 0x13, 0x16, 0x04, 0x03FFF801, PPC_MEM),
-GEN_HANDLER_E(lbarx, 0x1F, 0x14, 0x01, 0, PPC_NONE, 
 PPC2_ATOMIC_ISA206),
-GEN_HANDLER_E(lharx, 0x1F, 0x14, 0x03, 0, PPC_NONE, 
 PPC2_ATOMIC_ISA206),
-GEN_HANDLER(lwarx, 0x1F, 0x14, 0x00, 0x00000000, PPC_RES),

  GEN_HANDLER_E(lwat, 0x1F, 0x06, 0x12, 0x00000001, PPC_NONE, PPC2_ISA300), GEN_HANDLER_E(stwat, 0x1F, 0x06, 0x16, 0x00000001, PPC_NONE, PPC2_ISA300), GEN_HANDLER_E(stbcx_, 0x1F, 0x16, 0x15, 0, PPC_NONE, PPC2_ATOMIC_ISA206),

target/ppc/translate.c +0 -2
@@ -5864,8 +5866,6 @@ GEN_HANDLER2(stwcx_, "stwcx.", 0x1F, 0x16, 
 0x04, 0x00000000, PPC_RES),
   #if defined(TARGET_PPC64)
   GEN_HANDLER_E(ldat, 0x1F, 0x06, 0x13, 0x00000001, PPC_NONE, PPC2_ISA300),
   GEN_HANDLER_E(stdat, 0x1F, 0x06, 0x17, 0x00000001, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(ldarx, 0x1F, 0x14, 0x02, 0x00000000, PPC_64B),
-GEN_HANDLER_E(lqarx, 0x1F, 0x14, 0x08, 0, PPC_NONE, PPC2_LSQ_ISA207),
   GEN_HANDLER2(stdcx_, "stdcx.", 0x1F, 0x16, 0x06, 0x00000000, PPC_64B),
   GEN_HANDLER_E(stqcx_, 0x1F, 0x16, 0x05, 0, PPC_NONE, PPC2_LSQ_ISA207),

  #endif

Signature · 1 line
2.55.0
#7 lore raw

PATCH v2 3/37 target/ppc: Convert cache instructions to decodetree

In reply to Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Convert cache management instructions to decodetree using custom X-form layouts (@X_ea, @X_l, @X_th) and dedicated translation functions. This ensures reserved bits are accurately represented and ignored by the decoder (using '-' in the patterns), rather than being erroneously parsed into instruction fields.

Legacy GEN_HANDLER-based implementations are removed.

The implementation preserves legacy semantics, including:
- MMU-visible accesses for instructions treated as loads (e.g. dcbt, dcbtst, dcbtep, dcbtstep)
- Supervisor and BookE-specific constraints
- Effective address computation via do_ea_calc()

Instructions that were defined as no-ops in the legacy implementation remain no-ops here.

Note on intentional behavior changes:
- dcbf now validates the L field against the Power ISA v3.1 set of

    legal values (0, 1, 3, 4, 6).

- Reserved bits that were previously encoded as part of the opcode mask (and trapped if set) are now properly decoded as "don't care" ('-' in the pattern).

Testing:
- Verified TCG equivalence for all cache operations

---
 target/ppc/helper.h      |   8 +-
 target/ppc/insn32.decode |  35 +++
 target/ppc/mem_helper.c  |   8 +-
 target/ppc/translate.c   | 478 ++++++++++++++++++++-------------------
 4 files changed, 292 insertions(+), 237 deletions(-)
target/ppc/helper.h +4 -4
index 6b2d19a3ad..46ad78d1d4 100644
@@ -48,12 +48,12 @@ DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
 DEF_HELPER_4(lsw, void, env, tl, i32, i32)
 DEF_HELPER_5(lswx, void, env, tl, i32, i32, i32)
 DEF_HELPER_FLAGS_4(stsw, TCG_CALL_NO_WG, void, env, tl, i32, i32)
-DEF_HELPER_FLAGS_3(dcbz, TCG_CALL_NO_WG, void, env, tl, int)
+DEF_HELPER_FLAGS_3(DCBZ, TCG_CALL_NO_WG, void, env, tl, int)
 #ifdef TARGET_PPC64
-DEF_HELPER_FLAGS_2(dcbzl, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(DCBZL, TCG_CALL_NO_WG, void, env, tl)
 #endif
-DEF_HELPER_FLAGS_2(icbi, TCG_CALL_NO_WG, void, env, tl)
-DEF_HELPER_FLAGS_2(icbiep, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBI, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBIEP, TCG_CALL_NO_WG, void, env, tl)
 
 #if defined(TARGET_PPC64)
 DEF_HELPER_4(DIVDEU, i64, env, i64, i64, i32)
target/ppc/insn32.decode +35 -0
index 26948e08a7..6b1cf9a0a9 100644
@@ -1306,6 +1306,41 @@ XVF64GERPN      111011 ... -- .... 0 ..... 10111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNP      111011 ... -- .... 0 ..... 01111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 
+## Cache Management Instructions
+
+&X_ea           ra rb
+@X_ea           ...... ..... ra:5 rb:5 .......... .     &X_ea
+
+&X_l            l ra rb
+@X_l            ...... .. l:3 ra:5 rb:5 .......... .    &X_l
+
+&X_th           th ra rb
+@X_th           ...... th:5 ra:5 rb:5 .......... .      &X_th
+
+DCBF            011111 -- ... ..... ..... 0001010110 -   @X_l
+DCBFEP          011111 ----- ..... ..... 0001111111 -   @X_ea
+DCBI            011111 ----- ..... ..... 0111010110 -   @X_ea
+DCBST           011111 ----- ..... ..... 0000110110 -   @X_ea
+DCBSTEP         011111 ----- ..... ..... 0000111111 -   @X_ea
+DCBA            011111 ----- ..... ..... 1011110110 -   @X_ea
+
+DCBT            011111 ..... ..... ..... 0100010110 -   @X_th
+DCBTEP          011111 ----- ..... ..... 0100111111 -   @X_ea
+DCBTST          011111 ..... ..... ..... 0011110110 -   @X_th
+DCBTSTEP        011111 ----- ..... ..... 0011111111 -   @X_ea
+
+DCBTLS          011111 ..... ..... ..... 0010100110 -   @X_th
+DCBLC           011111 ..... ..... ..... 0110000110 -   @X_th
+
+DCBZ            011111 ----- ..... ..... 1111110110 -   @X_ea
+DCBZEP          011111 ----- ..... ..... 1111111111 -   @X_ea
+ICBI            011111 ----- ..... ..... 1111010110 -   @X_ea
+ICBIEP          011111 ----- ..... ..... 1111011111 -   @X_ea
+
+DST             011111 ----- ..... ..... 0101010110 -   @X_ea
+DSTST           011111 ----- ..... ..... 0101110110 -   @X_ea
+DSS             011111 ----- ..... ..... 1100110110 -   @X_ea
+
 ##Extend Sign Word and Shift Left Immediate XS-form
 EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
 
target/ppc/mem_helper.c +4 -4
index 787ba2cdec..95a224e74c 100644
@@ -319,13 +319,13 @@ static void dcbz_common(CPUPPCState *env, target_ulong addr,
     clear_helper_retaddr();
 }
 
-void helper_dcbz(CPUPPCState *env, target_ulong addr, int mmu_idx)
+void helper_DCBZ(CPUPPCState *env, target_ulong addr, int mmu_idx)
 {
     dcbz_common(env, addr, mmu_idx, env->dcache_line_size, GETPC());
 }
 
 #ifdef TARGET_PPC64
-void helper_dcbzl(CPUPPCState *env, target_ulong addr)
+void helper_DCBZL(CPUPPCState *env, target_ulong addr)
 {
     int dcbz_size = env->dcache_line_size;
 
@@ -341,7 +341,7 @@ void helper_dcbzl(CPUPPCState *env, target_ulong addr)
 }
 #endif
 
-void helper_icbi(CPUPPCState *env, target_ulong addr)
+void helper_ICBI(CPUPPCState *env, target_ulong addr)
 {
     unsigned mmu_idx = cpu_mmu_index(env_cpu(env), false);
     MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, mmu_idx);
@@ -357,7 +357,7 @@ void helper_icbi(CPUPPCState *env, target_ulong addr)
     cpu_ldl_mmu(env, addr, oi, GETPC());
 }
 
-void helper_icbiep(CPUPPCState *env, target_ulong addr)
+void helper_ICBIEP(CPUPPCState *env, target_ulong addr)
 {
 #if !defined(CONFIG_USER_ONLY)
     MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, PPC_TLB_EPID_LOAD);
target/ppc/translate.c +249 -229

496 lines not on this page · raw message

Signature · 1 line
2.55.0
#8 lore raw

Re: [PATCH v2 03/37] target/ppc: Convert cache instructions to decodetree

Quoted 92 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Convert cache management instructions to decodetree using custom X-form layouts (@X_ea, @X_l, @X_th) and dedicated translation functions. This ensures reserved bits are accurately represented and ignored by the decoder (using '-' in the patterns), rather than being erroneously parsed into instruction fields.

Legacy GEN_HANDLER-based implementations are removed.

The implementation preserves legacy semantics, including:
- MMU-visible accesses for instructions treated as loads (e.g. dcbt, dcbtst, dcbtep, dcbtstep)
- Supervisor and BookE-specific constraints
- Effective address computation via do_ea_calc()

Instructions that were defined as no-ops in the legacy implementation remain no-ops here.

Note on intentional behavior changes:
- dcbf now validates the L field against the Power ISA v3.1 set of

    legal values (0, 1, 3, 4, 6).

- Reserved bits that were previously encoded as part of the opcode mask (and trapped if set) are now properly decoded as "don't care" ('-' in the pattern).

Testing:
- Verified TCG equivalence for all cache operations

---
 target/ppc/helper.h      |   8 +-
 target/ppc/insn32.decode |  35 +++
 target/ppc/mem_helper.c  |   8 +-
 target/ppc/translate.c   | 478 ++++++++++++++++++++-------------------
 4 files changed, 292 insertions(+), 237 deletions(-)
target/ppc/helper.h +4 -4
index 6b2d19a3ad..46ad78d1d4 100644
@@ -48,12 +48,12 @@ DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
 DEF_HELPER_4(lsw, void, env, tl, i32, i32)
 DEF_HELPER_5(lswx, void, env, tl, i32, i32, i32)
 DEF_HELPER_FLAGS_4(stsw, TCG_CALL_NO_WG, void, env, tl, i32, i32)
-DEF_HELPER_FLAGS_3(dcbz, TCG_CALL_NO_WG, void, env, tl, int)
+DEF_HELPER_FLAGS_3(DCBZ, TCG_CALL_NO_WG, void, env, tl, int)
 #ifdef TARGET_PPC64
-DEF_HELPER_FLAGS_2(dcbzl, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(DCBZL, TCG_CALL_NO_WG, void, env, tl)
 #endif
-DEF_HELPER_FLAGS_2(icbi, TCG_CALL_NO_WG, void, env, tl)
-DEF_HELPER_FLAGS_2(icbiep, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBI, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBIEP, TCG_CALL_NO_WG, void, env, tl)
 
 #if defined(TARGET_PPC64)
 DEF_HELPER_4(DIVDEU, i64, env, i64, i64, i32)
target/ppc/insn32.decode +23 -0
index 26948e08a7..6b1cf9a0a9 100644
@@ -1306,6 +1306,41 @@ XVF64GERPN      111011 ... -- .... 0 ..... 10111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNP      111011 ... -- .... 0 ..... 01111010 ..-  @XX3_at xa=%xx_xa_pair
 XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 
+## Cache Management Instructions
+
+&X_ea           ra rb
+@X_ea           ...... ..... ra:5 rb:5 .......... .     &X_ea
+
+&X_l            l ra rb
+@X_l            ...... .. l:3 ra:5 rb:5 .......... .    &X_l
+
+&X_th           th ra rb
+@X_th           ...... th:5 ra:5 rb:5 .......... .      &X_th
+
+DCBF            011111 -- ... ..... ..... 0001010110 -   @X_l
+DCBFEP          011111 ----- ..... ..... 0001111111 -   @X_ea
+DCBI            011111 ----- ..... ..... 0111010110 -   @X_ea
+DCBST           011111 ----- ..... ..... 0000110110 -   @X_ea
+DCBSTEP         011111 ----- ..... ..... 0000111111 -   @X_ea
+DCBA            011111 ----- ..... ..... 1011110110 -   @X_ea
+
+DCBT            011111 ..... ..... ..... 0100010110 -   @X_th
+DCBTEP          011111 ----- ..... ..... 0100111111 -   @X_ea
+DCBTST          011111 ..... ..... ..... 0011110110 -   @X_th
+DCBTSTEP        011111 ----- ..... ..... 0011111111 -   @X_ea
+

target/ppc/insn32.decode XVF64GERPN 111011 ... -- .... 0 ..... 10111010 ..- @XX3_at xa=%xx_xa_pair · line 1341

Quoted 10 lines
target/ppc/insn32.decode +10 -0
+DCBTLS          011111 ..... ..... ..... 0010100110 -   @X_th
+DCBLC           011111 ..... ..... ..... 0110000110 -   @X_th
+
+DCBZ            011111 ----- ..... ..... 1111110110 -   @X_ea
+DCBZEP          011111 ----- ..... ..... 1111111111 -   @X_ea
+ICBI            011111 ----- ..... ..... 1111010110 -   @X_ea
+ICBIEP          011111 ----- ..... ..... 1111011111 -   @X_ea
+
+DST             011111 ----- ..... ..... 0101010110 -   @X_ea
+DSTST           011111 ----- ..... ..... 0101110110 -   @X_ea

Per **PowerISA v3.1 §4.3.2.1**, dst, dstst, and dss are obsolete and must be treated as no-ops independent of whether Vector instructions are available. The new REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in all three handlers is incorrect — it causes these opcodes to be treated as illegal instructions on non-Altivec CPUs, contradicting the ISA mandate. These handlers should unconditionally return true with no flag check, matching the ISA's "always a no-op" requirement. The ra == 0 INVAL check in trans_DST and trans_DSTST should also be dropped for the same reason — if the instruction is a no-op, the operand constraint has no effect.

Quoted 426 lines
target/ppc/insn32.decode +2 -0
+DSS             011111 ----- ..... ..... 1100110110 -   @X_ea
+
 ##Extend Sign Word and Shift Left Immediate XS-form
 EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
 
target/ppc/mem_helper.c +4 -4
index 787ba2cdec..95a224e74c 100644
@@ -319,13 +319,13 @@ static void dcbz_common(CPUPPCState *env, target_ulong addr,
     clear_helper_retaddr();
 }
 
-void helper_dcbz(CPUPPCState *env, target_ulong addr, int mmu_idx)
+void helper_DCBZ(CPUPPCState *env, target_ulong addr, int mmu_idx)
 {
     dcbz_common(env, addr, mmu_idx, env->dcache_line_size, GETPC());
 }
 
 #ifdef TARGET_PPC64
-void helper_dcbzl(CPUPPCState *env, target_ulong addr)
+void helper_DCBZL(CPUPPCState *env, target_ulong addr)
 {
     int dcbz_size = env->dcache_line_size;
 
@@ -341,7 +341,7 @@ void helper_dcbzl(CPUPPCState *env, target_ulong addr)
 }
 #endif
 
-void helper_icbi(CPUPPCState *env, target_ulong addr)
+void helper_ICBI(CPUPPCState *env, target_ulong addr)
 {
     unsigned mmu_idx = cpu_mmu_index(env_cpu(env), false);
     MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, mmu_idx);
@@ -357,7 +357,7 @@ void helper_icbi(CPUPPCState *env, target_ulong addr)
     cpu_ldl_mmu(env, addr, oi, GETPC());
 }
 
-void helper_icbiep(CPUPPCState *env, target_ulong addr)
+void helper_ICBIEP(CPUPPCState *env, target_ulong addr)
 {
 #if !defined(CONFIG_USER_ONLY)
     MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, PPC_TLB_EPID_LOAD);
target/ppc/translate.c +158 -210
index cc9287dcc5..60c88a6153 100644
@@ -4296,216 +4296,6 @@ static void gen_setb(DisasContext *ctx)
 }
 #endif
 
-/***                         Cache management                              ***/
-
-/* dcbf */
-static void gen_dcbf(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbfep (external PID dcbf) */
-static void gen_dcbfep(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    CHK_SV(ctx);
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbi (Supervisor only) */
-static void gen_dcbi(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv EA, val;
-
-    CHK_SV(ctx);
-    EA = tcg_temp_new();
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, EA);
-    val = tcg_temp_new();
-    /* XXX: specification says this should be treated as a store by the MMU */
-    gen_qemu_ld8u(ctx, val, EA);
-    gen_qemu_st8(ctx, val, EA);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-/* dcdst */
-static void gen_dcbst(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbstep (dcbstep External PID version) */
-static void gen_dcbstep(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbt */
-static void gen_dcbt(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtep */
-static void gen_dcbtep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtst */
-static void gen_dcbtst(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtstep */
-static void gen_dcbtstep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtls */
-static void gen_dcbtls(DisasContext *ctx)
-{
-    /* Always fails locking the cache */
-    TCGv t0 = tcg_temp_new();
-    gen_load_spr(t0, SPR_Exxx_L1CSR0);
-    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
-    gen_store_spr(SPR_Exxx_L1CSR0, t0);
-}
-
-/* dcblc */
-static void gen_dcblc(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     */
-}
-
-/* dcbz */
-static void gen_dcbz(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-
-#ifdef TARGET_PPC64
-    if (ctx->excp_model == POWERPC_EXCP_970 && !(ctx->opcode & 0x00200000)) {
-        gen_helper_dcbzl(tcg_env, tcgv_addr);
-        return;
-    }
-#endif
-
-    gen_helper_dcbz(tcg_env, tcgv_addr, tcg_constant_i32(ctx->mem_idx));
-}
-
-/* dcbzep */
-static void gen_dcbzep(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-    gen_helper_dcbz(tcg_env, tcgv_addr, tcg_constant_i32(PPC_TLB_EPID_STORE));
-}
-
-/* dst / dstt */
-static void gen_dst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-}
-
-/* dstst /dststt */
-static void gen_dstst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-
-}
-
-/* dss / dssall */
-static void gen_dss(DisasContext *ctx)
-{
-    /* interpreted as no-op */
-}
-
-/* icbi */
-static void gen_icbi(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbi(tcg_env, t0);
-}
-
-/* icbiep */
-static void gen_icbiep(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbiep(tcg_env, t0);
-}
-
-/* Optional: */
-/* dcba */
-static void gen_dcba(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a store by the MMU
-     *      but does not generate any exception
-     */
-}
-
 /***                    Segment register manipulation                      ***/
 /* Supervisor only: */
 
@@ -5746,6 +5536,255 @@ static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a)
     return true;
 }
 
+/*
+ * Cache Management Instructions (decodetree)
+ */
+
+static bool trans_DCBA(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBA);
+    return true;
+}
+
+static bool trans_DCBT(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBTST(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBLC(DisasContext *ctx, arg_X_th *a)
+{
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+    return true;
+}
+
+static bool trans_DSS(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+    return true;
+}
+
+static bool trans_DCBTLS(DisasContext *ctx, arg_X_th *a)
+{
+    TCGv t0 = tcg_temp_new();
+
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+
+    gen_load_spr(t0, SPR_Exxx_L1CSR0);
+    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
+    gen_store_spr(SPR_Exxx_L1CSR0, t0);
+
+    return true;
+}
+
+static bool trans_DST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DSTST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DCBF(DisasContext *ctx, arg_X_l *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    /*
+     * As per PowerISA v3.1, the L field can have values 0, 1, 3, 4, or 6.
+     * Other values are Undefined Behavior (UB).
+     */
+    switch (a->l) {
+    case 0: /* dcbf */
+    case 1: /* dcbfl */
+    case 3: /* dcbflp */
+    case 4: /* dcbfps */
+    case 6: /* dcbstps */
+        break;
+    default:
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBST(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBFEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+
+    return true;
+}
+
+static bool trans_DCBSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5706

Quoted 10 lines
target/ppc/translate.c +10 -0
+    return true;
+}
+
+static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+    return true;

The #if defined(CONFIG_USER_ONLY) block looks redundant to me. REQUIRE_SV is already defined for both build configurations — in user-only builds it unconditionally calls gen_priv_opc() and returns, and in system builds it checks ctx->pr. The manual #if split just duplicates what REQUIRE_SV already does. The function can be simplified to:

  static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a)
  {
      REQUIRE_INSNS_FLAGS(ctx, CACHE);
      REQUIRE_SV(ctx);
      gen_set_access_type(ctx, ACCESS_CACHE);
      TCGv EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
      TCGv val = tcg_temp_new();
  
      gen_qemu_ld8u(ctx, val, EA);
      gen_qemu_st8(ctx, val, EA);
  
      return true;

}

Quoted 47 lines
target/ppc/translate.c +47 -0
+#else
+    TCGv EA, val;
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    val = tcg_temp_new();
+
+    gen_qemu_ld8u(ctx, val, EA);
+    gen_qemu_st8(ctx, val, EA);
+
+    return true;
+#endif
+}
+
+static bool trans_ICBI(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_ICBI);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_helper_ICBI(tcg_env, EA);
+
+    return true;
+}
+
+static bool trans_ICBIEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_ICBIEP(tcg_env, EA);
+    return true;
+}
+
+static bool trans_DCBZ(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5763

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBZ);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+#if defined(TARGET_PPC64)
+    if (ctx->excp_model == POWERPC_EXCP_970 &&
+        !(ctx->opcode & 0x00200000)) {

trans_DCBZ reads ctx->opcode directly to distinguish dcbz from dcbzl on the 970.

In a decodetree translator ctx->opcode is not reliable — the instruction fields are decoded into arg_* before the translator is called. Bit 21 (0x00200000) is the LSB of the RT field: dcbz encodes RT=0 and the 970-specific dcbzl encodes RT=1. Since the current decode pattern

011111----- ..... ..... 1111110110

treats all of RT as don't-care, both variants land in trans_DCBZ with the distinguishing bit lost.

I think a better way would to split the decode entry so the table itself separates the two:

DCBZ   011111 00000 ..... ..... 1111110110 -   @X_ea
DCBZL  011111 00001 ..... ..... 1111110110 -   @X_ea

Then trans_DCBZ needs no ctx->opcode access, and trans_DCBZL handles the 970 path cleanly with its own REQUIRE_64BIT + excp_model check.

Quoted 56 lines
target/ppc/translate.c +24 -19
+        gen_helper_DCBZL(tcg_env, EA);
+        return true;
+    }
+#endif
+
+    gen_helper_DCBZ(tcg_env, EA, tcg_constant_i32(ctx->mem_idx));
+    return true;
+}
+
+static bool trans_DCBZEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_DCBZ(tcg_env, EA, tcg_constant_i32(PPC_TLB_EPID_STORE));
+    return true;
+}
+
 #include "translate/fixedpoint-impl.c.inc"
 
 #include "translate/fp-impl.c.inc"
@@ -5910,25 +5949,6 @@ GEN_HANDLER_E(mcrxrx, 0x1F, 0x00, 0x12, 0x007FF801, PPC_NONE, PPC2_ISA300),
 #endif
 GEN_HANDLER(mtmsr, 0x1F, 0x12, 0x04, 0x001EF801, PPC_MISC),
 GEN_HANDLER(mtspr, 0x1F, 0x13, 0x0E, 0x00000000, PPC_MISC),
-GEN_HANDLER(dcbf, 0x1F, 0x16, 0x02, 0x03C00001, PPC_CACHE),
-GEN_HANDLER_E(dcbfep, 0x1F, 0x1F, 0x03, 0x03C00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbi, 0x1F, 0x16, 0x0E, 0x03E00001, PPC_CACHE),
-GEN_HANDLER(dcbst, 0x1F, 0x16, 0x01, 0x03E00001, PPC_CACHE),
-GEN_HANDLER_E(dcbstep, 0x1F, 0x1F, 0x01, 0x03E00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbt, 0x1F, 0x16, 0x08, 0x00000001, PPC_CACHE),
-GEN_HANDLER_E(dcbtep, 0x1F, 0x1F, 0x09, 0x00000001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbtst, 0x1F, 0x16, 0x07, 0x00000001, PPC_CACHE),
-GEN_HANDLER_E(dcbtstep, 0x1F, 0x1F, 0x07, 0x00000001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER_E(dcbtls, 0x1F, 0x06, 0x05, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),
-GEN_HANDLER_E(dcblc, 0x1F, 0x06, 0x0c, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),
-GEN_HANDLER(dcbz, 0x1F, 0x16, 0x1F, 0x03C00001, PPC_CACHE_DCBZ),
-GEN_HANDLER_E(dcbzep, 0x1F, 0x1F, 0x1F, 0x03C00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dst, 0x1F, 0x16, 0x0A, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dstst, 0x1F, 0x16, 0x0B, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dss, 0x1F, 0x16, 0x19, 0x019FF801, PPC_ALTIVEC),
-GEN_HANDLER(icbi, 0x1F, 0x16, 0x1E, 0x03E00001, PPC_CACHE_ICBI),
-GEN_HANDLER_E(icbiep, 0x1F, 0x1F, 0x1E, 0x03E00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcba, 0x1F, 0x16, 0x17, 0x03E00001, PPC_CACHE_DCBA),
 GEN_HANDLER(mfsr, 0x1F, 0x13, 0x12, 0x0010F801, PPC_SEGMENT),
 GEN_HANDLER(mfsrin, 0x1F, 0x13, 0x14, 0x001F0001, PPC_SEGMENT),
 GEN_HANDLER(mtsr, 0x1F, 0x12, 0x06, 0x0010F801, PPC_SEGMENT),
Signature · 1 line
2.55.0
#9 lore raw

Re: [PATCH v2 03/37] target/ppc: Convert cache instructions to decodetree

Quoted 104 lines · Amit Machhiwal
Quoted 92 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Convert cache management instructions to decodetree using custom X-form layouts (@X_ea, @X_l, @X_th) and dedicated translation functions. This ensures reserved bits are accurately represented and ignored by the decoder (using '-' in the patterns), rather than being erroneously parsed into instruction fields.

Legacy GEN_HANDLER-based implementations are removed.

The implementation preserves legacy semantics, including:
- MMU-visible accesses for instructions treated as loads (e.g. dcbt, dcbtst, dcbtep, dcbtstep)
- Supervisor and BookE-specific constraints
- Effective address computation via do_ea_calc()

Instructions that were defined as no-ops in the legacy implementation remain no-ops here.

Note on intentional behavior changes:
- dcbf now validates the L field against the Power ISA v3.1 set of

     legal values (0, 1, 3, 4, 6).

- Reserved bits that were previously encoded as part of the opcode mask (and trapped if set) are now properly decoded as "don't care" ('-' in the pattern).

Testing:
- Verified TCG equivalence for all cache operations

---
  target/ppc/helper.h      |   8 +-
  target/ppc/insn32.decode |  35 +++
  target/ppc/mem_helper.c  |   8 +-
  target/ppc/translate.c   | 478 ++++++++++++++++++++-------------------
  4 files changed, 292 insertions(+), 237 deletions(-)
target/ppc/helper.h +4 -4
index 6b2d19a3ad..46ad78d1d4 100644
@@ -48,12 +48,12 @@ DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
  DEF_HELPER_4(lsw, void, env, tl, i32, i32)
  DEF_HELPER_5(lswx, void, env, tl, i32, i32, i32)
  DEF_HELPER_FLAGS_4(stsw, TCG_CALL_NO_WG, void, env, tl, i32, i32)
-DEF_HELPER_FLAGS_3(dcbz, TCG_CALL_NO_WG, void, env, tl, int)
+DEF_HELPER_FLAGS_3(DCBZ, TCG_CALL_NO_WG, void, env, tl, int)
  #ifdef TARGET_PPC64
-DEF_HELPER_FLAGS_2(dcbzl, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(DCBZL, TCG_CALL_NO_WG, void, env, tl)
  #endif
-DEF_HELPER_FLAGS_2(icbi, TCG_CALL_NO_WG, void, env, tl)
-DEF_HELPER_FLAGS_2(icbiep, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBI, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBIEP, TCG_CALL_NO_WG, void, env, tl)
 
  #if defined(TARGET_PPC64)
  DEF_HELPER_4(DIVDEU, i64, env, i64, i64, i32)
target/ppc/insn32.decode +23 -0
index 26948e08a7..6b1cf9a0a9 100644
@@ -1306,6 +1306,41 @@ XVF64GERPN      111011 ... -- .... 0 ..... 10111010 ..-  @XX3_at xa=%xx_xa_pair
  XVF64GERNP      111011 ... -- .... 0 ..... 01111010 ..-  @XX3_at xa=%xx_xa_pair
  XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 
+## Cache Management Instructions
+
+&X_ea           ra rb
+@X_ea           ...... ..... ra:5 rb:5 .......... .     &X_ea
+
+&X_l            l ra rb
+@X_l            ...... .. l:3 ra:5 rb:5 .......... .    &X_l
+
+&X_th           th ra rb
+@X_th           ...... th:5 ra:5 rb:5 .......... .      &X_th
+
+DCBF            011111 -- ... ..... ..... 0001010110 -   @X_l
+DCBFEP          011111 ----- ..... ..... 0001111111 -   @X_ea
+DCBI            011111 ----- ..... ..... 0111010110 -   @X_ea
+DCBST           011111 ----- ..... ..... 0000110110 -   @X_ea
+DCBSTEP         011111 ----- ..... ..... 0000111111 -   @X_ea
+DCBA            011111 ----- ..... ..... 1011110110 -   @X_ea
+
+DCBT            011111 ..... ..... ..... 0100010110 -   @X_th
+DCBTEP          011111 ----- ..... ..... 0100111111 -   @X_ea
+DCBTST          011111 ..... ..... ..... 0011110110 -   @X_th
+DCBTSTEP        011111 ----- ..... ..... 0011111111 -   @X_ea
+

target/ppc/insn32.decode XVF64GERPN 111011 ... -- .... 0 ..... 10111010 ..- @XX3_at xa=%xx_xa_pair · line 1341

Quoted 10 lines
target/ppc/insn32.decode +10 -0
+DCBTLS          011111 ..... ..... ..... 0010100110 -   @X_th
+DCBLC           011111 ..... ..... ..... 0110000110 -   @X_th
+
+DCBZ            011111 ----- ..... ..... 1111110110 -   @X_ea
+DCBZEP          011111 ----- ..... ..... 1111111111 -   @X_ea
+ICBI            011111 ----- ..... ..... 1111010110 -   @X_ea
+ICBIEP          011111 ----- ..... ..... 1111011111 -   @X_ea
+
+DST             011111 ----- ..... ..... 0101010110 -   @X_ea
+DSTST           011111 ----- ..... ..... 0101110110 -   @X_ea

Per **PowerISA v3.1 §4.3.2.1**, dst, dstst, and dss are obsolete and

Quoted 8 lines

must be treated as no-ops independent of whether Vector instructions are available. The new REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in all three handlers is incorrect — it causes these opcodes to be treated as illegal instructions on non-Altivec CPUs, contradicting the ISA mandate. These handlers should unconditionally return true with no flag check, matching the ISA's "always a no-op" requirement. The ra == 0 INVAL check in trans_DST and trans_DSTST should also be dropped for the same reason — if the instruction is a no-op, the operand constraint has no effect.

Thanks Amit !
This is very informative. There are many such instances in QEMU code where insns haven't been updated as per the newer ISA.
Though, I'd like to point out that the REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in those insns is not new. This was already present as in the code, I will mark it with [A] below for reference.
This patch simply intends to move the insn to decodetree keeping the existing logic intact. The translation logic and flag checks remain same for these 3 pointed out insns. I'd like to take up your suggestion here and post a folllow-up patch to update those insns as per the new ISA. Would that be okay ? Or would you prefer the insn's correction as per the ISA in this same decodetree movement patch ?

Quoted 457 lines
Quoted 427 lines
target/ppc/insn32.decode +2 -0
+DSS             011111 ----- ..... ..... 1100110110 -   @X_ea
+
  ##Extend Sign Word and Shift Left Immediate XS-form
  EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
 
target/ppc/mem_helper.c +4 -4
index 787ba2cdec..95a224e74c 100644
@@ -319,13 +319,13 @@ static void dcbz_common(CPUPPCState *env, target_ulong addr,
      clear_helper_retaddr();
  }
 
-void helper_dcbz(CPUPPCState *env, target_ulong addr, int mmu_idx)
+void helper_DCBZ(CPUPPCState *env, target_ulong addr, int mmu_idx)
  {
      dcbz_common(env, addr, mmu_idx, env->dcache_line_size, GETPC());
  }
 
  #ifdef TARGET_PPC64
-void helper_dcbzl(CPUPPCState *env, target_ulong addr)
+void helper_DCBZL(CPUPPCState *env, target_ulong addr)
  {
      int dcbz_size = env->dcache_line_size;
 
@@ -341,7 +341,7 @@ void helper_dcbzl(CPUPPCState *env, target_ulong addr)
  }
  #endif
 
-void helper_icbi(CPUPPCState *env, target_ulong addr)
+void helper_ICBI(CPUPPCState *env, target_ulong addr)
  {
      unsigned mmu_idx = cpu_mmu_index(env_cpu(env), false);
      MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, mmu_idx);
@@ -357,7 +357,7 @@ void helper_icbi(CPUPPCState *env, target_ulong addr)
      cpu_ldl_mmu(env, addr, oi, GETPC());
  }
 
-void helper_icbiep(CPUPPCState *env, target_ulong addr)
+void helper_ICBIEP(CPUPPCState *env, target_ulong addr)
  {
  #if !defined(CONFIG_USER_ONLY)
      MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, PPC_TLB_EPID_LOAD);
target/ppc/translate.c +158 -210
index cc9287dcc5..60c88a6153 100644
@@ -4296,216 +4296,6 @@ static void gen_setb(DisasContext *ctx)
  }
  #endif
 
-/***                         Cache management                              ***/
-
-/* dcbf */
-static void gen_dcbf(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbfep (external PID dcbf) */
-static void gen_dcbfep(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    CHK_SV(ctx);
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbi (Supervisor only) */
-static void gen_dcbi(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv EA, val;
-
-    CHK_SV(ctx);
-    EA = tcg_temp_new();
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, EA);
-    val = tcg_temp_new();
-    /* XXX: specification says this should be treated as a store by the MMU */
-    gen_qemu_ld8u(ctx, val, EA);
-    gen_qemu_st8(ctx, val, EA);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-/* dcdst */
-static void gen_dcbst(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbstep (dcbstep External PID version) */
-static void gen_dcbstep(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbt */
-static void gen_dcbt(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtep */
-static void gen_dcbtep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtst */
-static void gen_dcbtst(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtstep */
-static void gen_dcbtstep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtls */
-static void gen_dcbtls(DisasContext *ctx)
-{
-    /* Always fails locking the cache */
-    TCGv t0 = tcg_temp_new();
-    gen_load_spr(t0, SPR_Exxx_L1CSR0);
-    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
-    gen_store_spr(SPR_Exxx_L1CSR0, t0);
-}
-
-/* dcblc */
-static void gen_dcblc(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     */
-}
-
-/* dcbz */
-static void gen_dcbz(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-
-#ifdef TARGET_PPC64
-    if (ctx->excp_model == POWERPC_EXCP_970 && !(ctx->opcode & 0x00200000)) {
-        gen_helper_dcbzl(tcg_env, tcgv_addr);
-        return;
-    }
-#endif
-
-    gen_helper_dcbz(tcg_env, tcgv_addr, tcg_constant_i32(ctx->mem_idx));
-}
-
-/* dcbzep */
-static void gen_dcbzep(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-    gen_helper_dcbz(tcg_env, tcgv_addr, tcg_constant_i32(PPC_TLB_EPID_STORE));
-}
-
-/* dst / dstt */
-static void gen_dst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-}
-
-/* dstst /dststt */
-static void gen_dstst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-
-}
-
-/* dss / dssall */
-static void gen_dss(DisasContext *ctx)
-{
-    /* interpreted as no-op */
-}
-
-/* icbi */
-static void gen_icbi(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbi(tcg_env, t0);
-}
-
-/* icbiep */
-static void gen_icbiep(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbiep(tcg_env, t0);
-}
-
-/* Optional: */
-/* dcba */
-static void gen_dcba(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a store by the MMU
-     *      but does not generate any exception
-     */
-}
-
  /***                    Segment register manipulation                      ***/
  /* Supervisor only: */
 
@@ -5746,6 +5536,255 @@ static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a)
      return true;
  }
 
+/*
+ * Cache Management Instructions (decodetree)
+ */
+
+static bool trans_DCBA(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBA);
+    return true;
+}
+
+static bool trans_DCBT(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBTST(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBLC(DisasContext *ctx, arg_X_th *a)
+{
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+    return true;
+}
+
+static bool trans_DSS(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+    return true;
+}
+
+static bool trans_DCBTLS(DisasContext *ctx, arg_X_th *a)
+{
+    TCGv t0 = tcg_temp_new();
+
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+
+    gen_load_spr(t0, SPR_Exxx_L1CSR0);
+    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
+    gen_store_spr(SPR_Exxx_L1CSR0, t0);
+
+    return true;
+}
+
+static bool trans_DST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DSTST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DCBF(DisasContext *ctx, arg_X_l *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    /*
+     * As per PowerISA v3.1, the L field can have values 0, 1, 3, 4, or 6.
+     * Other values are Undefined Behavior (UB).
+     */
+    switch (a->l) {
+    case 0: /* dcbf */
+    case 1: /* dcbfl */
+    case 3: /* dcbflp */
+    case 4: /* dcbfps */
+    case 6: /* dcbstps */
+        break;
+    default:
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBST(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBFEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+
+    return true;
+}
+
+static bool trans_DCBSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5706

Quoted 10 lines
target/ppc/translate.c +10 -0
+    return true;
+}
+
+static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+    return true;

The #if defined(CONFIG_USER_ONLY) block looks redundant to me. REQUIRE_SV is already defined for both build configurations — in user-only builds it unconditionally calls gen_priv_opc() and returns, and in system builds it checks ctx->pr. The manual #if split just duplicates what REQUIRE_SV already does. The function can be simplified to:

   static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a)
   {
       REQUIRE_INSNS_FLAGS(ctx, CACHE);
       REQUIRE_SV(ctx);
       gen_set_access_type(ctx, ACCESS_CACHE);
       TCGv EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
       TCGv val = tcg_temp_new();
   
       gen_qemu_ld8u(ctx, val, EA);
       gen_qemu_st8(ctx, val, EA);
   
       return true;
Quoted 1 line

}

Agreed ! Thanks !

Chinmay

Quoted 119 lines
Quoted 48 lines
target/ppc/translate.c +47 -0
+#else
+    TCGv EA, val;
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    val = tcg_temp_new();
+
+    gen_qemu_ld8u(ctx, val, EA);
+    gen_qemu_st8(ctx, val, EA);
+
+    return true;
+#endif
+}
+
+static bool trans_ICBI(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_ICBI);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_helper_ICBI(tcg_env, EA);
+
+    return true;
+}
+
+static bool trans_ICBIEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_ICBIEP(tcg_env, EA);
+    return true;
+}
+
+static bool trans_DCBZ(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5763

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBZ);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+#if defined(TARGET_PPC64)
+    if (ctx->excp_model == POWERPC_EXCP_970 &&
+        !(ctx->opcode & 0x00200000)) {

trans_DCBZ reads ctx->opcode directly to distinguish dcbz from dcbzl on the 970.

In a decodetree translator ctx->opcode is not reliable — the instruction fields are decoded into arg_* before the translator is called. Bit 21 (0x00200000) is the LSB of the RT field: dcbz encodes RT=0 and the 970-specific dcbzl encodes RT=1. Since the current decode pattern

   011111----- ..... ..... 1111110110

treats all of RT as don't-care, both variants land in trans_DCBZ with the distinguishing bit lost.

I think a better way would to split the decode entry so the table itself separates the two:

DCBZ   011111 00000 ..... ..... 1111110110 -   @X_ea
DCBZL  011111 00001 ..... ..... 1111110110 -   @X_ea

Then trans_DCBZ needs no ctx->opcode access, and trans_DCBZL handles the 970 path cleanly with its own REQUIRE_64BIT + excp_model check.

Quoted 39 lines
target/ppc/translate.c +24 -8
+        gen_helper_DCBZL(tcg_env, EA);
+        return true;
+    }
+#endif
+
+    gen_helper_DCBZ(tcg_env, EA, tcg_constant_i32(ctx->mem_idx));
+    return true;
+}
+
+static bool trans_DCBZEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_DCBZ(tcg_env, EA, tcg_constant_i32(PPC_TLB_EPID_STORE));
+    return true;
+}
+
  #include "translate/fixedpoint-impl.c.inc"
 
  #include "translate/fp-impl.c.inc"
@@ -5910,25 +5949,6 @@ GEN_HANDLER_E(mcrxrx, 0x1F, 0x00, 0x12, 0x007FF801, PPC_NONE, PPC2_ISA300),
  #endif
  GEN_HANDLER(mtmsr, 0x1F, 0x12, 0x04, 0x001EF801, PPC_MISC),
  GEN_HANDLER(mtspr, 0x1F, 0x13, 0x0E, 0x00000000, PPC_MISC),
-GEN_HANDLER(dcbf, 0x1F, 0x16, 0x02, 0x03C00001, PPC_CACHE),
-GEN_HANDLER_E(dcbfep, 0x1F, 0x1F, 0x03, 0x03C00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbi, 0x1F, 0x16, 0x0E, 0x03E00001, PPC_CACHE),
-GEN_HANDLER(dcbst, 0x1F, 0x16, 0x01, 0x03E00001, PPC_CACHE),
-GEN_HANDLER_E(dcbstep, 0x1F, 0x1F, 0x01, 0x03E00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbt, 0x1F, 0x16, 0x08, 0x00000001, PPC_CACHE),
-GEN_HANDLER_E(dcbtep, 0x1F, 0x1F, 0x09, 0x00000001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbtst, 0x1F, 0x16, 0x07, 0x00000001, PPC_CACHE),
Quoted 8 lines

target/ppc/translate.c GEN_HANDLER_E(mcrxrx, 0x1F, 0x00, 0x12, 0x007FF801, PPC_NONE, PPC2_ISA300), · line 5928

Quoted 8 lines
target/ppc/translate.c +0 -8
-GEN_HANDLER_E(dcbtstep, 0x1F, 0x1F, 0x07, 0x00000001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER_E(dcbtls, 0x1F, 0x06, 0x05, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),
-GEN_HANDLER_E(dcblc, 0x1F, 0x06, 0x0c, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),
-GEN_HANDLER(dcbz, 0x1F, 0x16, 0x1F, 0x03C00001, PPC_CACHE_DCBZ),
-GEN_HANDLER_E(dcbzep, 0x1F, 0x1F, 0x1F, 0x03C00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dst, 0x1F, 0x16, 0x0A, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dstst, 0x1F, 0x16, 0x0B, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dss, 0x1F, 0x16, 0x19, 0x019FF801, PPC_ALTIVEC),

[A] ^

Quoted 9 lines
Quoted 9 lines
target/ppc/translate.c +0 -3
-GEN_HANDLER(icbi, 0x1F, 0x16, 0x1E, 0x03E00001, PPC_CACHE_ICBI),
-GEN_HANDLER_E(icbiep, 0x1F, 0x1F, 0x1E, 0x03E00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcba, 0x1F, 0x16, 0x17, 0x03E00001, PPC_CACHE_DCBA),
  GEN_HANDLER(mfsr, 0x1F, 0x13, 0x12, 0x0010F801, PPC_SEGMENT),
  GEN_HANDLER(mfsrin, 0x1F, 0x13, 0x14, 0x001F0001, PPC_SEGMENT),
  GEN_HANDLER(mtsr, 0x1F, 0x12, 0x06, 0x0010F801, PPC_SEGMENT),
Signature · 1 line
2.55.0
#10 lore raw

Re: [PATCH v2 03/37] target/ppc: Convert cache instructions to decodetree

Quoted 5 lines · Chinmay Rath
Quoted 3 lines · Amit Machhiwal
Quoted 2 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

<snip>

Quoted 12 lines
Quoted 4 lines

target/ppc/insn32.decode XVF64GERPN 111011 ... -- .... 0 ..... 10111010 ..- @XX3_at xa=%xx_xa_pair · line 1341

Quoted 3 lines
target/ppc/insn32.decode +3 -0
+
+DST             011111 ----- ..... ..... 0101010110 -   @X_ea
+DSTST           011111 ----- ..... ..... 0101110110 -   @X_ea

Per **PowerISA v3.1 §4.3.2.1**, dst, dstst, and dss are obsolete and

Quoted 8 lines

must be treated as no-ops independent of whether Vector instructions are available. The new REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in all three handlers is incorrect — it causes these opcodes to be treated as illegal instructions on non-Altivec CPUs, contradicting the ISA mandate. These handlers should unconditionally return true with no flag check, matching the ISA's "always a no-op" requirement. The ra == 0 INVAL check in trans_DST and trans_DSTST should also be dropped for the same reason — if the instruction is a no-op, the operand constraint has no effect.

Quoted 12 lines

Thanks Amit !
This is very informative. There are many such instances in QEMU code where insns haven't been updated as per the newer ISA.
Though, I'd like to point out that the REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in those insns is not new. This was already present as in the code, I will mark it with [A] below for reference.
This patch simply intends to move the insn to decodetree keeping the existing logic intact. The translation logic and flag checks remain same for these 3 pointed out insns. I'd like to take up your suggestion here and post a folllow-up patch to update those insns as per the new ISA. Would that be okay ? Or would you prefer the insn's correction as per the ISA in this same decodetree movement patch ?

I did notice that this check already existed before this movement and I don't see it as a blocker for this series. It would make sense to fix such places in follow up patches.

~Amit

#11 lore raw

Re: [PATCH v2 03/37] target/ppc: Convert cache instructions to decodetree

In reply to Chinmay Rath
Quoted 572 lines · Chinmay Rath
Quoted 105 lines · Amit Machhiwal
Quoted 93 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Convert cache management instructions to decodetree using custom X-form layouts (@X_ea, @X_l, @X_th) and dedicated translation functions. This ensures reserved bits are accurately represented and ignored by the decoder (using '-' in the patterns), rather than being erroneously parsed into instruction fields.

Legacy GEN_HANDLER-based implementations are removed.

The implementation preserves legacy semantics, including:
   - MMU-visible accesses for instructions treated as loads (e.g. dcbt, dcbtst, dcbtep, dcbtstep)
   - Supervisor and BookE-specific constraints
   - Effective address computation via do_ea_calc()

Instructions that were defined as no-ops in the legacy implementation remain no-ops here.

Note on intentional behavior changes:
   - dcbf now validates the L field against the Power ISA v3.1 set of legal values (0, 1, 3, 4, 6).
   - Reserved bits that were previously encoded as part of the opcode mask (and trapped if set) are now properly decoded as "don't care" ('-' in the pattern).

Testing:
   - Verified TCG equivalence for all cache operations

---
  target/ppc/helper.h      |   8 +-
  target/ppc/insn32.decode |  35 +++
  target/ppc/mem_helper.c  |   8 +-
  target/ppc/translate.c   | 478
++++++++++++++++++++-------------------

  4 files changed, 292 insertions(+), 237 deletions(-)
target/ppc/helper.h +4 -4
index 6b2d19a3ad..46ad78d1d4 100644
@@ -48,12 +48,12 @@ DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, 
 env, tl, i32)
   DEF_HELPER_4(lsw, void, env, tl, i32, i32)
   DEF_HELPER_5(lswx, void, env, tl, i32, i32, i32)
   DEF_HELPER_FLAGS_4(stsw, TCG_CALL_NO_WG, void, env, tl, i32, i32)
-DEF_HELPER_FLAGS_3(dcbz, TCG_CALL_NO_WG, void, env, tl, int)
+DEF_HELPER_FLAGS_3(DCBZ, TCG_CALL_NO_WG, void, env, tl, int)
   #ifdef TARGET_PPC64
-DEF_HELPER_FLAGS_2(dcbzl, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(DCBZL, TCG_CALL_NO_WG, void, env, tl)
   #endif
-DEF_HELPER_FLAGS_2(icbi, TCG_CALL_NO_WG, void, env, tl)
-DEF_HELPER_FLAGS_2(icbiep, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBI, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBIEP, TCG_CALL_NO_WG, void, env, tl)
     #if defined(TARGET_PPC64)
   DEF_HELPER_4(DIVDEU, i64, env, i64, i64, i32)
target/ppc/insn32.decode +0 -0
index 26948e08a7..6b1cf9a0a9 100644
@@ -1306,6 +1306,41 @@ XVF64GERPN      111011 ... -- .... 0 ..... 

10111010 ..-  @XX3_at xa=%xx_xa_pair
  XVF64GERNP      111011 ... -- .... 0 ..... 01111010 ..- @XX3_at xa=%xx_xa_pair XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..- @XX3_at xa=%xx_xa_pair +## Cache Management Instructions

target/ppc/insn32.decode +22 -0
+
+&X_ea           ra rb
+@X_ea           ...... ..... ra:5 rb:5 .......... . &X_ea
+
+&X_l            l ra rb
+@X_l            ...... .. l:3 ra:5 rb:5 .......... . &X_l
+
+&X_th           th ra rb
+@X_th           ...... th:5 ra:5 rb:5 .......... . &X_th
+
+DCBF            011111 -- ... ..... ..... 0001010110 -   @X_l
+DCBFEP          011111 ----- ..... ..... 0001111111 -   @X_ea
+DCBI            011111 ----- ..... ..... 0111010110 -   @X_ea
+DCBST           011111 ----- ..... ..... 0000110110 -   @X_ea
+DCBSTEP         011111 ----- ..... ..... 0000111111 -   @X_ea
+DCBA            011111 ----- ..... ..... 1011110110 -   @X_ea
+
+DCBT            011111 ..... ..... ..... 0100010110 -   @X_th
+DCBTEP          011111 ----- ..... ..... 0100111111 -   @X_ea
+DCBTST          011111 ..... ..... ..... 0011110110 -   @X_th
+DCBTSTEP        011111 ----- ..... ..... 0011111111 -   @X_ea
+

target/ppc/insn32.decode XVF64GERPN 111011 ... -- .... 0 ..... 10111010 ..- @XX3_at xa=%xx_xa_pair · line 1341

Quoted 10 lines
target/ppc/insn32.decode +10 -0
+DCBTLS          011111 ..... ..... ..... 0010100110 -   @X_th
+DCBLC           011111 ..... ..... ..... 0110000110 -   @X_th
+
+DCBZ            011111 ----- ..... ..... 1111110110 -   @X_ea
+DCBZEP          011111 ----- ..... ..... 1111111111 -   @X_ea
+ICBI            011111 ----- ..... ..... 1111010110 -   @X_ea
+ICBIEP          011111 ----- ..... ..... 1111011111 -   @X_ea
+
+DST             011111 ----- ..... ..... 0101010110 -   @X_ea
+DSTST           011111 ----- ..... ..... 0101110110 -   @X_ea

Per **PowerISA v3.1 §4.3.2.1**, dst, dstst, and dss are obsolete and

Quoted 8 lines

must be treated as no-ops independent of whether Vector instructions are available. The new REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in all three handlers is incorrect — it causes these opcodes to be treated as illegal instructions on non-Altivec CPUs, contradicting the ISA mandate. These handlers should unconditionally return true with no flag check, matching the ISA's "always a no-op" requirement. The ra == 0 INVAL check in trans_DST and trans_DSTST should also be dropped for the same reason — if the instruction is a no-op, the operand constraint has no effect.

Thanks Amit !
This is very informative. There are many such instances in QEMU code where insns haven't been updated as per the newer ISA.
Though, I'd like to point out that the REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in those insns is not new. This was already present as in the code, I will mark it with [A] below for reference.
This patch simply intends to move the insn to decodetree keeping the existing logic intact. The translation logic and flag checks remain same for these 3 pointed out insns. I'd like to take up your suggestion here and post a folllow-up patch to update those insns as per the new ISA. Would that be okay ? Or would you prefer the insn's correction as per the ISA in this same decodetree movement patch ?

Quoted 442 lines
Quoted 425 lines
target/ppc/insn32.decode +2 -0
+DSS             011111 ----- ..... ..... 1100110110 -   @X_ea
+

  ##Extend Sign Word and Shift Left Immediate XS-form EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS diff --git a/target/ppc/mem_helper.c b/target/ppc/mem_helper.c index 787ba2cdec..95a224e74c 100644

target/ppc/mem_helper.c +2 -1
@@ -319,13 +319,13 @@ static void dcbz_common(CPUPPCState *env, 
 target_ulong addr,
       clear_helper_retaddr();
   }
   -void helper_dcbz(CPUPPCState *env, target_ulong addr, int mmu_idx)
+void helper_DCBZ(CPUPPCState *env, target_ulong addr, int mmu_idx)
   {
       dcbz_common(env, addr, mmu_idx, env->dcache_line_size, GETPC());
   }
     #ifdef TARGET_PPC64
-void helper_dcbzl(CPUPPCState *env, target_ulong addr)
+void helper_DCBZL(CPUPPCState *env, target_ulong addr)
   {
       int dcbz_size = env->dcache_line_size;
   @@ -341,7 +341,7 @@ void helper_dcbzl(CPUPPCState *env, target_ulong addr)

  }
  #endif
  -void helper_icbi(CPUPPCState *env, target_ulong addr) +void helper_ICBI(CPUPPCState *env, target_ulong addr) {
      unsigned mmu_idx = cpu_mmu_index(env_cpu(env), false); MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, mmu_idx);

target/ppc/mem_helper.c +1 -0
@@ -357,7 +357,7 @@ void helper_icbi(CPUPPCState *env, target_ulong 
 addr)
       cpu_ldl_mmu(env, addr, oi, GETPC());
   }
   -void helper_icbiep(CPUPPCState *env, target_ulong addr)
+void helper_ICBIEP(CPUPPCState *env, target_ulong addr)
   {
   #if !defined(CONFIG_USER_ONLY)

      MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, PPC_TLB_EPID_LOAD);

target/ppc/translate.c +0 -0
index cc9287dcc5..60c88a6153 100644
@@ -4296,216 +4296,6 @@ static void gen_setb(DisasContext *ctx)
   }
   #endif

  -/***                         Cache management                              ***/

target/ppc/translate.c +0 -38
-
-/* dcbf */
-static void gen_dcbf(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbfep (external PID dcbf) */
-static void gen_dcbfep(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    CHK_SV(ctx);
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbi (Supervisor only) */
-static void gen_dcbi(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv EA, val;
-
-    CHK_SV(ctx);
-    EA = tcg_temp_new();
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, EA);
-    val = tcg_temp_new();
-    /* XXX: specification says this should be treated as a store by 

the MMU */
-    gen_qemu_ld8u(ctx, val, EA);
-    gen_qemu_st8(ctx, val, EA);

target/ppc/translate.c +0 -93
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-/* dcdst */
-static void gen_dcbst(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbstep (dcbstep External PID version) */
-static void gen_dcbstep(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbt */
-static void gen_dcbt(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtep */
-static void gen_dcbtep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtst */
-static void gen_dcbtst(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtstep */
-static void gen_dcbtstep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtls */
-static void gen_dcbtls(DisasContext *ctx)
-{
-    /* Always fails locking the cache */
-    TCGv t0 = tcg_temp_new();
-    gen_load_spr(t0, SPR_Exxx_L1CSR0);
-    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
-    gen_store_spr(SPR_Exxx_L1CSR0, t0);
-}
-
-/* dcblc */
-static void gen_dcblc(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     */
-}
-
-/* dcbz */
-static void gen_dcbz(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-
-#ifdef TARGET_PPC64
-    if (ctx->excp_model == POWERPC_EXCP_970 && !(ctx->opcode & 

0x00200000)) {
-        gen_helper_dcbzl(tcg_env, tcgv_addr);
-        return;
-    }

target/ppc/translate.c +0 -3
-#endif
-
-    gen_helper_dcbz(tcg_env, tcgv_addr, 

tcg_constant_i32(ctx->mem_idx));

target/ppc/translate.c +0 -10
-}
-
-/* dcbzep */
-static void gen_dcbzep(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-    gen_helper_dcbz(tcg_env, tcgv_addr, 

tcg_constant_i32(PPC_TLB_EPID_STORE));

target/ppc/translate.c +0 -60
-}
-
-/* dst / dstt */
-static void gen_dst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-}
-
-/* dstst /dststt */
-static void gen_dstst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-
-}
-
-/* dss / dssall */
-static void gen_dss(DisasContext *ctx)
-{
-    /* interpreted as no-op */
-}
-
-/* icbi */
-static void gen_icbi(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbi(tcg_env, t0);
-}
-
-/* icbiep */
-static void gen_icbiep(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbiep(tcg_env, t0);
-}
-
-/* Optional: */
-/* dcba */
-static void gen_dcba(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a store by the MMU
-     *      but does not generate any exception
-     */
-}
-

  /***                    Segment register manipulation                      ***/ /* Supervisor only: */
  @@ -5746,6 +5536,255 @@ static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) return true;
  }
  +/*
+ * Cache Management Instructions (decodetree)
+ */

target/ppc/translate.c +94 -0
+
+static bool trans_DCBA(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBA);
+    return true;
+}
+
+static bool trans_DCBT(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBTST(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBLC(DisasContext *ctx, arg_X_th *a)
+{
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+    return true;
+}
+
+static bool trans_DSS(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+    return true;
+}
+
+static bool trans_DCBTLS(DisasContext *ctx, arg_X_th *a)
+{
+    TCGv t0 = tcg_temp_new();
+
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+
+    gen_load_spr(t0, SPR_Exxx_L1CSR0);
+    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
+    gen_store_spr(SPR_Exxx_L1CSR0, t0);
+
+    return true;
+}
+
+static bool trans_DST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DSTST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DCBF(DisasContext *ctx, arg_X_l *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    /*
+     * As per PowerISA v3.1, the L field can have values 0, 1, 3, 

4, or 6.
+     * Other values are Undefined Behavior (UB). +     */
+    switch (a->l) {
+    case 0: /* dcbf */
+    case 1: /* dcbfl */
+    case 3: /* dcbflp */
+    case 4: /* dcbfps */
+    case 6: /* dcbstps */
+        break;
+    default:
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL); +        return true;
+    }

target/ppc/translate.c +48 -0
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBST(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBFEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+
+    return true;
+}
+
+static bool trans_DCBSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5706

Quoted 10 lines
target/ppc/translate.c +10 -0
+    return true;
+}
+
+static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+    return true;

The #if defined(CONFIG_USER_ONLY) block looks redundant to me. REQUIRE_SV is already defined for both build configurations — in user-only builds it unconditionally calls gen_priv_opc() and returns, and in system builds it checks ctx->pr. The manual #if split just duplicates what REQUIRE_SV already does. The function can be simplified to:

Quoted 11 lines

   static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a) {
       REQUIRE_INSNS_FLAGS(ctx, CACHE); REQUIRE_SV(ctx);
          gen_set_access_type(ctx, ACCESS_CACHE); TCGv EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]); TCGv val = tcg_temp_new(); gen_qemu_ld8u(ctx, val, EA); gen_qemu_st8(ctx, val, EA);
          return true;
   }

Quoted 2 lines

Agreed ! Thanks !

Unfortunately, gen_qemu_st8 is not defined for !CONFIG_USER_ONLY. So will have to stick to the #if defined block. :)

Quoted 149 lines

Chinmay

Quoted 127 lines
Quoted 48 lines
target/ppc/translate.c +47 -0
+#else
+    TCGv EA, val;
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    val = tcg_temp_new();
+
+    gen_qemu_ld8u(ctx, val, EA);
+    gen_qemu_st8(ctx, val, EA);
+
+    return true;
+#endif
+}
+
+static bool trans_ICBI(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_ICBI);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_helper_ICBI(tcg_env, EA);
+
+    return true;
+}
+
+static bool trans_ICBIEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_ICBIEP(tcg_env, EA);
+    return true;
+}
+
+static bool trans_DCBZ(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5763

Quoted 10 lines
target/ppc/translate.c +10 -0
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBZ);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+#if defined(TARGET_PPC64)
+    if (ctx->excp_model == POWERPC_EXCP_970 &&
+        !(ctx->opcode & 0x00200000)) {

trans_DCBZ reads ctx->opcode directly to distinguish dcbz from dcbzl on the 970.

In a decodetree translator ctx->opcode is not reliable — the instruction fields are decoded into arg_* before the translator is called. Bit 21 (0x00200000) is the LSB of the RT field: dcbz encodes RT=0 and the 970-specific dcbzl encodes RT=1. Since the current decode pattern

   011111----- ..... ..... 1111110110

treats all of RT as don't-care, both variants land in trans_DCBZ with the distinguishing bit lost.

I think a better way would to split the decode entry so the table itself separates the two:

DCBZ   011111 00000 ..... ..... 1111110110 -   @X_ea DCBZL  011111 00001 ..... ..... 1111110110 -   @X_ea

Then trans_DCBZ needs no ctx->opcode access, and trans_DCBZL handles the 970 path cleanly with its own REQUIRE_64BIT + excp_model check.

Quoted 47 lines

+        gen_helper_DCBZL(tcg_env, EA);
+        return true;
+    }

target/ppc/translate.c +18 -0
+#endif
+
+    gen_helper_DCBZ(tcg_env, EA, tcg_constant_i32(ctx->mem_idx));
+    return true;
+}
+
+static bool trans_DCBZEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_DCBZ(tcg_env, EA, 

tcg_constant_i32(PPC_TLB_EPID_STORE));
+    return true;

target/ppc/translate.c +2 -0
+}
+

  #include "translate/fixedpoint-impl.c.inc"
    #include "translate/fp-impl.c.inc"

target/ppc/translate.c +0 -7
@@ -5910,25 +5949,6 @@ GEN_HANDLER_E(mcrxrx, 0x1F, 0x00, 0x12, 
 0x007FF801, PPC_NONE, PPC2_ISA300),
   #endif
   GEN_HANDLER(mtmsr, 0x1F, 0x12, 0x04, 0x001EF801, PPC_MISC),
   GEN_HANDLER(mtspr, 0x1F, 0x13, 0x0E, 0x00000000, PPC_MISC),
-GEN_HANDLER(dcbf, 0x1F, 0x16, 0x02, 0x03C00001, PPC_CACHE),
-GEN_HANDLER_E(dcbfep, 0x1F, 0x1F, 0x03, 0x03C00001, PPC_NONE, 
 PPC2_BOOKE206),
-GEN_HANDLER(dcbi, 0x1F, 0x16, 0x0E, 0x03E00001, PPC_CACHE),
-GEN_HANDLER(dcbst, 0x1F, 0x16, 0x01, 0x03E00001, PPC_CACHE),
-GEN_HANDLER_E(dcbstep, 0x1F, 0x1F, 0x01, 0x03E00001, PPC_NONE, 
 PPC2_BOOKE206),
-GEN_HANDLER(dcbt, 0x1F, 0x16, 0x08, 0x00000001, PPC_CACHE),
-GEN_HANDLER_E(dcbtep, 0x1F, 0x1F, 0x09, 0x00000001, PPC_NONE, 

PPC2_BOOKE206),

target/ppc/translate.c +0 -2
-GEN_HANDLER(dcbtst, 0x1F, 0x16, 0x07, 0x00000001, PPC_CACHE),
-GEN_HANDLER_E(dcbtstep, 0x1F, 0x1F, 0x07, 0x00000001, PPC_NONE, 

PPC2_BOOKE206),
-GEN_HANDLER_E(dcbtls, 0x1F, 0x06, 0x05, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),

Quoted 8 lines

target/ppc/translate.c GEN_HANDLER_E(mcrxrx, 0x1F, 0x00, 0x12, 0x007FF801, PPC_NONE, PPC2_ISA300), · line 5928

Quoted 8 lines

-GEN_HANDLER_E(dcblc, 0x1F, 0x06, 0x0c, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),

target/ppc/translate.c +0 -2
-GEN_HANDLER(dcbz, 0x1F, 0x16, 0x1F, 0x03C00001, PPC_CACHE_DCBZ),
-GEN_HANDLER_E(dcbzep, 0x1F, 0x1F, 0x1F, 0x03C00001, PPC_NONE, 

PPC2_BOOKE206),

target/ppc/translate.c +0 -3
-GEN_HANDLER(dst, 0x1F, 0x16, 0x0A, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dstst, 0x1F, 0x16, 0x0B, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dss, 0x1F, 0x16, 0x19, 0x019FF801, PPC_ALTIVEC),

[A] ^

Quoted 10 lines
Quoted 10 lines
target/ppc/translate.c +0 -2
-GEN_HANDLER(icbi, 0x1F, 0x16, 0x1E, 0x03E00001, PPC_CACHE_ICBI),
-GEN_HANDLER_E(icbiep, 0x1F, 0x1F, 0x1E, 0x03E00001, PPC_NONE, 

PPC2_BOOKE206),
-GEN_HANDLER(dcba, 0x1F, 0x16, 0x17, 0x03E00001, PPC_CACHE_DCBA), GEN_HANDLER(mfsr, 0x1F, 0x13, 0x12, 0x0010F801, PPC_SEGMENT), GEN_HANDLER(mfsrin, 0x1F, 0x13, 0x14, 0x001F0001, PPC_SEGMENT), GEN_HANDLER(mtsr, 0x1F, 0x12, 0x06, 0x0010F801, PPC_SEGMENT),

Signature · 1 line
2.55.0
#12 lore raw

Re: [PATCH v2 03/37] target/ppc: Convert cache instructions to decodetree

In reply to Amit Machhiwal
Quoted 620 lines · Amit Machhiwal
Quoted 92 lines · Chinmay Rath

From: Nikhil Kumar Singh <[email protected]>

Convert cache management instructions to decodetree using custom X-form layouts (@X_ea, @X_l, @X_th) and dedicated translation functions. This ensures reserved bits are accurately represented and ignored by the decoder (using '-' in the patterns), rather than being erroneously parsed into instruction fields.

Legacy GEN_HANDLER-based implementations are removed.

The implementation preserves legacy semantics, including:
- MMU-visible accesses for instructions treated as loads (e.g. dcbt, dcbtst, dcbtep, dcbtstep)
- Supervisor and BookE-specific constraints
- Effective address computation via do_ea_calc()

Instructions that were defined as no-ops in the legacy implementation remain no-ops here.

Note on intentional behavior changes:
- dcbf now validates the L field against the Power ISA v3.1 set of

     legal values (0, 1, 3, 4, 6).

- Reserved bits that were previously encoded as part of the opcode mask (and trapped if set) are now properly decoded as "don't care" ('-' in the pattern).

Testing:
- Verified TCG equivalence for all cache operations

---
  target/ppc/helper.h      |   8 +-
  target/ppc/insn32.decode |  35 +++
  target/ppc/mem_helper.c  |   8 +-
  target/ppc/translate.c   | 478 ++++++++++++++++++++-------------------
  4 files changed, 292 insertions(+), 237 deletions(-)
target/ppc/helper.h +4 -4
index 6b2d19a3ad..46ad78d1d4 100644
@@ -48,12 +48,12 @@ DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
  DEF_HELPER_4(lsw, void, env, tl, i32, i32)
  DEF_HELPER_5(lswx, void, env, tl, i32, i32, i32)
  DEF_HELPER_FLAGS_4(stsw, TCG_CALL_NO_WG, void, env, tl, i32, i32)
-DEF_HELPER_FLAGS_3(dcbz, TCG_CALL_NO_WG, void, env, tl, int)
+DEF_HELPER_FLAGS_3(DCBZ, TCG_CALL_NO_WG, void, env, tl, int)
  #ifdef TARGET_PPC64
-DEF_HELPER_FLAGS_2(dcbzl, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(DCBZL, TCG_CALL_NO_WG, void, env, tl)
  #endif
-DEF_HELPER_FLAGS_2(icbi, TCG_CALL_NO_WG, void, env, tl)
-DEF_HELPER_FLAGS_2(icbiep, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBI, TCG_CALL_NO_WG, void, env, tl)
+DEF_HELPER_FLAGS_2(ICBIEP, TCG_CALL_NO_WG, void, env, tl)
 
  #if defined(TARGET_PPC64)
  DEF_HELPER_4(DIVDEU, i64, env, i64, i64, i32)
target/ppc/insn32.decode +23 -0
index 26948e08a7..6b1cf9a0a9 100644
@@ -1306,6 +1306,41 @@ XVF64GERPN      111011 ... -- .... 0 ..... 10111010 ..-  @XX3_at xa=%xx_xa_pair
  XVF64GERNP      111011 ... -- .... 0 ..... 01111010 ..-  @XX3_at xa=%xx_xa_pair
  XVF64GERNN      111011 ... -- .... 0 ..... 11111010 ..-  @XX3_at xa=%xx_xa_pair
 
+## Cache Management Instructions
+
+&X_ea           ra rb
+@X_ea           ...... ..... ra:5 rb:5 .......... .     &X_ea
+
+&X_l            l ra rb
+@X_l            ...... .. l:3 ra:5 rb:5 .......... .    &X_l
+
+&X_th           th ra rb
+@X_th           ...... th:5 ra:5 rb:5 .......... .      &X_th
+
+DCBF            011111 -- ... ..... ..... 0001010110 -   @X_l
+DCBFEP          011111 ----- ..... ..... 0001111111 -   @X_ea
+DCBI            011111 ----- ..... ..... 0111010110 -   @X_ea
+DCBST           011111 ----- ..... ..... 0000110110 -   @X_ea
+DCBSTEP         011111 ----- ..... ..... 0000111111 -   @X_ea
+DCBA            011111 ----- ..... ..... 1011110110 -   @X_ea
+
+DCBT            011111 ..... ..... ..... 0100010110 -   @X_th
+DCBTEP          011111 ----- ..... ..... 0100111111 -   @X_ea
+DCBTST          011111 ..... ..... ..... 0011110110 -   @X_th
+DCBTSTEP        011111 ----- ..... ..... 0011111111 -   @X_ea
+

target/ppc/insn32.decode XVF64GERPN 111011 ... -- .... 0 ..... 10111010 ..- @XX3_at xa=%xx_xa_pair · line 1341

Quoted 10 lines
target/ppc/insn32.decode +10 -0
+DCBTLS          011111 ..... ..... ..... 0010100110 -   @X_th
+DCBLC           011111 ..... ..... ..... 0110000110 -   @X_th
+
+DCBZ            011111 ----- ..... ..... 1111110110 -   @X_ea
+DCBZEP          011111 ----- ..... ..... 1111111111 -   @X_ea
+ICBI            011111 ----- ..... ..... 1111010110 -   @X_ea
+ICBIEP          011111 ----- ..... ..... 1111011111 -   @X_ea
+
+DST             011111 ----- ..... ..... 0101010110 -   @X_ea
+DSTST           011111 ----- ..... ..... 0101110110 -   @X_ea

Per **PowerISA v3.1 §4.3.2.1**, dst, dstst, and dss are obsolete and must be treated as no-ops independent of whether Vector instructions are available. The new REQUIRE_INSNS_FLAGS(ctx, ALTIVEC) guard in all three handlers is incorrect — it causes these opcodes to be treated as illegal instructions on non-Altivec CPUs, contradicting the ISA mandate. These handlers should unconditionally return true with no flag check, matching the ISA's "always a no-op" requirement. The ra == 0 INVAL check in trans_DST and trans_DSTST should also be dropped for the same reason — if the instruction is a no-op, the operand constraint has no effect.

Quoted 426 lines
target/ppc/insn32.decode +2 -0
+DSS             011111 ----- ..... ..... 1100110110 -   @X_ea
+
  ##Extend Sign Word and Shift Left Immediate XS-form
  EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
 
target/ppc/mem_helper.c +4 -4
index 787ba2cdec..95a224e74c 100644
@@ -319,13 +319,13 @@ static void dcbz_common(CPUPPCState *env, target_ulong addr,
      clear_helper_retaddr();
  }
 
-void helper_dcbz(CPUPPCState *env, target_ulong addr, int mmu_idx)
+void helper_DCBZ(CPUPPCState *env, target_ulong addr, int mmu_idx)
  {
      dcbz_common(env, addr, mmu_idx, env->dcache_line_size, GETPC());
  }
 
  #ifdef TARGET_PPC64
-void helper_dcbzl(CPUPPCState *env, target_ulong addr)
+void helper_DCBZL(CPUPPCState *env, target_ulong addr)
  {
      int dcbz_size = env->dcache_line_size;
 
@@ -341,7 +341,7 @@ void helper_dcbzl(CPUPPCState *env, target_ulong addr)
  }
  #endif
 
-void helper_icbi(CPUPPCState *env, target_ulong addr)
+void helper_ICBI(CPUPPCState *env, target_ulong addr)
  {
      unsigned mmu_idx = cpu_mmu_index(env_cpu(env), false);
      MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, mmu_idx);
@@ -357,7 +357,7 @@ void helper_icbi(CPUPPCState *env, target_ulong addr)
      cpu_ldl_mmu(env, addr, oi, GETPC());
  }
 
-void helper_icbiep(CPUPPCState *env, target_ulong addr)
+void helper_ICBIEP(CPUPPCState *env, target_ulong addr)
  {
  #if !defined(CONFIG_USER_ONLY)
      MemOpIdx oi = make_memop_idx(MO_UL | MO_UNALN, PPC_TLB_EPID_LOAD);
target/ppc/translate.c +158 -210
index cc9287dcc5..60c88a6153 100644
@@ -4296,216 +4296,6 @@ static void gen_setb(DisasContext *ctx)
  }
  #endif
 
-/***                         Cache management                              ***/
-
-/* dcbf */
-static void gen_dcbf(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbfep (external PID dcbf) */
-static void gen_dcbfep(DisasContext *ctx)
-{
-    /* XXX: specification says this is treated as a load by the MMU */
-    TCGv t0;
-    CHK_SV(ctx);
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbi (Supervisor only) */
-static void gen_dcbi(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv EA, val;
-
-    CHK_SV(ctx);
-    EA = tcg_temp_new();
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, EA);
-    val = tcg_temp_new();
-    /* XXX: specification says this should be treated as a store by the MMU */
-    gen_qemu_ld8u(ctx, val, EA);
-    gen_qemu_st8(ctx, val, EA);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-/* dcdst */
-static void gen_dcbst(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_qemu_ld8u(ctx, t0, t0);
-}
-
-/* dcbstep (dcbstep External PID version) */
-static void gen_dcbstep(DisasContext *ctx)
-{
-    /* XXX: specification say this is treated as a load by the MMU */
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    tcg_gen_qemu_ld_tl(t0, t0, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
-}
-
-/* dcbt */
-static void gen_dcbt(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtep */
-static void gen_dcbtep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtst */
-static void gen_dcbtst(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtstep */
-static void gen_dcbtstep(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a load by the MMU but
-     *      does not generate any exception
-     */
-}
-
-/* dcbtls */
-static void gen_dcbtls(DisasContext *ctx)
-{
-    /* Always fails locking the cache */
-    TCGv t0 = tcg_temp_new();
-    gen_load_spr(t0, SPR_Exxx_L1CSR0);
-    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
-    gen_store_spr(SPR_Exxx_L1CSR0, t0);
-}
-
-/* dcblc */
-static void gen_dcblc(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     */
-}
-
-/* dcbz */
-static void gen_dcbz(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-
-#ifdef TARGET_PPC64
-    if (ctx->excp_model == POWERPC_EXCP_970 && !(ctx->opcode & 0x00200000)) {
-        gen_helper_dcbzl(tcg_env, tcgv_addr);
-        return;
-    }
-#endif
-
-    gen_helper_dcbz(tcg_env, tcgv_addr, tcg_constant_i32(ctx->mem_idx));
-}
-
-/* dcbzep */
-static void gen_dcbzep(DisasContext *ctx)
-{
-    TCGv tcgv_addr = tcg_temp_new();
-
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    gen_addr_reg_index(ctx, tcgv_addr);
-    gen_helper_dcbz(tcg_env, tcgv_addr, tcg_constant_i32(PPC_TLB_EPID_STORE));
-}
-
-/* dst / dstt */
-static void gen_dst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-}
-
-/* dstst /dststt */
-static void gen_dstst(DisasContext *ctx)
-{
-    if (rA(ctx->opcode) == 0) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-    } else {
-        /* interpreted as no-op */
-    }
-
-}
-
-/* dss / dssall */
-static void gen_dss(DisasContext *ctx)
-{
-    /* interpreted as no-op */
-}
-
-/* icbi */
-static void gen_icbi(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbi(tcg_env, t0);
-}
-
-/* icbiep */
-static void gen_icbiep(DisasContext *ctx)
-{
-    TCGv t0;
-    gen_set_access_type(ctx, ACCESS_CACHE);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    gen_helper_icbiep(tcg_env, t0);
-}
-
-/* Optional: */
-/* dcba */
-static void gen_dcba(DisasContext *ctx)
-{
-    /*
-     * interpreted as no-op
-     * XXX: specification say this is treated as a store by the MMU
-     *      but does not generate any exception
-     */
-}
-
  /***                    Segment register manipulation                      ***/
  /* Supervisor only: */
 
@@ -5746,6 +5536,255 @@ static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a)
      return true;
  }
 
+/*
+ * Cache Management Instructions (decodetree)
+ */
+
+static bool trans_DCBA(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBA);
+    return true;
+}
+
+static bool trans_DCBT(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBTST(DisasContext *ctx, arg_X_th *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+    return true;
+}
+
+static bool trans_DCBTSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    return true;
+}
+
+static bool trans_DCBLC(DisasContext *ctx, arg_X_th *a)
+{
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+    return true;
+}
+
+static bool trans_DSS(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+    return true;
+}
+
+static bool trans_DCBTLS(DisasContext *ctx, arg_X_th *a)
+{
+    TCGv t0 = tcg_temp_new();
+
+    /* Requires either PPC_BOOKE or PPC2_BOOKE206 */
+    if (!(ctx->insns_flags & PPC_BOOKE) &&
+        !(ctx->insns_flags2 & PPC2_BOOKE206)) {
+        return false;
+    }
+
+    gen_load_spr(t0, SPR_Exxx_L1CSR0);
+    tcg_gen_ori_tl(t0, t0, L1CSR0_CUL);
+    gen_store_spr(SPR_Exxx_L1CSR0, t0);
+
+    return true;
+}
+
+static bool trans_DST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DSTST(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, ALTIVEC);
+
+    if (a->ra == 0) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+    }
+
+    return true;
+}
+
+static bool trans_DCBF(DisasContext *ctx, arg_X_l *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    /*
+     * As per PowerISA v3.1, the L field can have values 0, 1, 3, 4, or 6.
+     * Other values are Undefined Behavior (UB).
+     */
+    switch (a->l) {
+    case 0: /* dcbf */
+    case 1: /* dcbfl */
+    case 3: /* dcbflp */
+    case 4: /* dcbfps */
+    case 6: /* dcbstps */
+        break;
+    default:
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBST(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_qemu_ld8u(ctx, EA, EA);
+
+    return true;
+}
+
+static bool trans_DCBFEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+
+    return true;
+}
+
+static bool trans_DCBSTEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    tcg_gen_qemu_ld_tl(EA, EA, PPC_TLB_EPID_LOAD, DEF_MEMOP(MO_UB));
+

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5706

Quoted 10 lines
target/ppc/translate.c +10 -0
+    return true;
+}
+
+static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, CACHE);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+    return true;

The #if defined(CONFIG_USER_ONLY) block looks redundant to me. REQUIRE_SV is already defined for both build configurations — in user-only builds it unconditionally calls gen_priv_opc() and returns, and in system builds it checks ctx->pr. The manual #if split just duplicates what REQUIRE_SV already does. The function can be simplified to:

   static bool trans_DCBI(DisasContext *ctx, arg_X_ea *a)
   {
       REQUIRE_INSNS_FLAGS(ctx, CACHE);
       REQUIRE_SV(ctx);
       gen_set_access_type(ctx, ACCESS_CACHE);
       TCGv EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
       TCGv val = tcg_temp_new();
   
       gen_qemu_ld8u(ctx, val, EA);
       gen_qemu_st8(ctx, val, EA);
   
       return true;

}

Quoted 49 lines
target/ppc/translate.c +49 -0
+#else
+    TCGv EA, val;
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    val = tcg_temp_new();
+
+    gen_qemu_ld8u(ctx, val, EA);
+    gen_qemu_st8(ctx, val, EA);
+
+    return true;
+#endif
+}
+
+static bool trans_ICBI(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_ICBI);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    gen_helper_ICBI(tcg_env, EA);
+
+    return true;
+}
+
+static bool trans_ICBIEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_ICBIEP(tcg_env, EA);
+    return true;
+}
+
+static bool trans_DCBZ(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS(ctx, CACHE_DCBZ);
Quoted 8 lines

target/ppc/translate.c static bool trans_LQARX(DisasContext *ctx, arg_LQARX *a) · line 5763

Quoted 8 lines
target/ppc/translate.c +8 -0
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+#if defined(TARGET_PPC64)
+    if (ctx->excp_model == POWERPC_EXCP_970 &&
+        !(ctx->opcode & 0x00200000)) {

Sorry I missed this comment.

Quoted 4 lines

trans_DCBZ reads ctx->opcode directly to distinguish dcbz from dcbzl on the 970.

In a decodetree translator ctx->opcode is not reliable — the instruction

I agree, in decodetree translation, it is not reliable to use ctx->opcode, though it is technically usable and populated.

Quoted 6 lines

fields are decoded into arg_* before the translator is called. Bit 21 (0x00200000) is the LSB of the RT field: dcbz encodes RT=0 and the 970-specific dcbzl encodes RT=1. Since the current decode pattern

   011111----- ..... ..... 1111110110
Quoted 8 lines

treats all of RT as don't-care, both variants land in trans_DCBZ with the distinguishing bit lost.

I think a better way would to split the decode entry so the table itself separates the two:

DCBZ   011111 00000 ..... ..... 1111110110 -   @X_ea
DCBZL  011111 00001 ..... ..... 1111110110 -   @X_ea

I agree with the idea of splitting the two insns. However, DCBZ should have '-' for reserved bits staying true to the ISA and DCBZL could use it's own format. However I am unable to find any official doc/ISA for DCBZL yet. Hence will drop this insn from the patch in the interest of time. Will post a followup patch for this.

Thanks,
Chinmay

Quoted 60 lines

Then trans_DCBZ needs no ctx->opcode access, and trans_DCBZL handles the 970 path cleanly with its own REQUIRE_64BIT + excp_model check.

Quoted 56 lines
target/ppc/translate.c +24 -19
+        gen_helper_DCBZL(tcg_env, EA);
+        return true;
+    }
+#endif
+
+    gen_helper_DCBZ(tcg_env, EA, tcg_constant_i32(ctx->mem_idx));
+    return true;
+}
+
+static bool trans_DCBZEP(DisasContext *ctx, arg_X_ea *a)
+{
+    TCGv EA;
+
+    REQUIRE_INSNS_FLAGS2(ctx, BOOKE206);
+
+    REQUIRE_SV(ctx);
+
+    gen_set_access_type(ctx, ACCESS_CACHE);
+    EA = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+
+    gen_helper_DCBZ(tcg_env, EA, tcg_constant_i32(PPC_TLB_EPID_STORE));
+    return true;
+}
+
  #include "translate/fixedpoint-impl.c.inc"
 
  #include "translate/fp-impl.c.inc"
@@ -5910,25 +5949,6 @@ GEN_HANDLER_E(mcrxrx, 0x1F, 0x00, 0x12, 0x007FF801, PPC_NONE, PPC2_ISA300),
  #endif
  GEN_HANDLER(mtmsr, 0x1F, 0x12, 0x04, 0x001EF801, PPC_MISC),
  GEN_HANDLER(mtspr, 0x1F, 0x13, 0x0E, 0x00000000, PPC_MISC),
-GEN_HANDLER(dcbf, 0x1F, 0x16, 0x02, 0x03C00001, PPC_CACHE),
-GEN_HANDLER_E(dcbfep, 0x1F, 0x1F, 0x03, 0x03C00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbi, 0x1F, 0x16, 0x0E, 0x03E00001, PPC_CACHE),
-GEN_HANDLER(dcbst, 0x1F, 0x16, 0x01, 0x03E00001, PPC_CACHE),
-GEN_HANDLER_E(dcbstep, 0x1F, 0x1F, 0x01, 0x03E00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbt, 0x1F, 0x16, 0x08, 0x00000001, PPC_CACHE),
-GEN_HANDLER_E(dcbtep, 0x1F, 0x1F, 0x09, 0x00000001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcbtst, 0x1F, 0x16, 0x07, 0x00000001, PPC_CACHE),
-GEN_HANDLER_E(dcbtstep, 0x1F, 0x1F, 0x07, 0x00000001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER_E(dcbtls, 0x1F, 0x06, 0x05, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),
-GEN_HANDLER_E(dcblc, 0x1F, 0x06, 0x0c, 0x02000001, PPC_BOOKE, PPC2_BOOKE206),
-GEN_HANDLER(dcbz, 0x1F, 0x16, 0x1F, 0x03C00001, PPC_CACHE_DCBZ),
-GEN_HANDLER_E(dcbzep, 0x1F, 0x1F, 0x1F, 0x03C00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dst, 0x1F, 0x16, 0x0A, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dstst, 0x1F, 0x16, 0x0B, 0x01800001, PPC_ALTIVEC),
-GEN_HANDLER(dss, 0x1F, 0x16, 0x19, 0x019FF801, PPC_ALTIVEC),
-GEN_HANDLER(icbi, 0x1F, 0x16, 0x1E, 0x03E00001, PPC_CACHE_ICBI),
-GEN_HANDLER_E(icbiep, 0x1F, 0x1F, 0x1E, 0x03E00001, PPC_NONE, PPC2_BOOKE206),
-GEN_HANDLER(dcba, 0x1F, 0x16, 0x17, 0x03E00001, PPC_CACHE_DCBA),
  GEN_HANDLER(mfsr, 0x1F, 0x13, 0x12, 0x0010F801, PPC_SEGMENT),
  GEN_HANDLER(mfsrin, 0x1F, 0x13, 0x14, 0x001F0001, PPC_SEGMENT),
  GEN_HANDLER(mtsr, 0x1F, 0x12, 0x06, 0x0010F801, PPC_SEGMENT),
Signature · 1 line
2.55.0
#13 lore raw

PATCH v2 4/37 target/ppc: Move vector merge instructions to decodetree

In reply to Chinmay Rath

From: Aboorva Devarajan <[email protected]>

Move below instructions to decodetree specification:

vmrg{hb, hh, hw, lb, lh, lw} : VX-form

The individual gen_vmrg* functions generated by GEN_VXFORM are replaced by a shared do_vmrg() helper dispatching to the existing gen_helper_VMRG* TCG helpers via the TRANS_FLAGS() macro, which also handles the ALTIVEC facility check.

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/helper.h                 | 12 ++++++------
 target/ppc/insn32.decode            |  9 +++++++++
 target/ppc/int_helper.c             | 12 ++++++------
 target/ppc/translate/vmx-impl.c.inc | 25 +++++++++++++++++++------
 target/ppc/translate/vmx-ops.c.inc  |  6 ------
 5 files changed, 40 insertions(+), 24 deletions(-)
target/ppc/helper.h +6 -6
index 46ad78d1d4..6db917935c 100644
@@ -173,12 +173,12 @@ DEF_HELPER_4(vcmpeqfp_dot, void, env, avr, avr, avr)
 DEF_HELPER_4(vcmpgefp_dot, void, env, avr, avr, avr)
 DEF_HELPER_4(vcmpgtfp_dot, void, env, avr, avr, avr)
 DEF_HELPER_4(vcmpbfp_dot, void, env, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrglb, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrglh, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrglw, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrghb, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrghh, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrghw, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGLB, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGLH, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGLW, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGHB, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGHH, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGHW, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_FLAGS_3(VMULESB, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_FLAGS_3(VMULESH, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_FLAGS_3(VMULESW, TCG_CALL_NO_RWG, void, avr, avr, avr)
target/ppc/insn32.decode +9 -0
index 6b1cf9a0a9..8aca8a00c0 100644
@@ -795,6 +795,15 @@ VNOR            000100 ..... ..... ..... 10100000100    @VX
 VXOR            000100 ..... ..... ..... 10011000100    @VX
 VEQV            000100 ..... ..... ..... 11010000100    @VX
 
+## Vector Merge Instructions
+
+VMRGHB          000100 ..... ..... ..... 00000001100    @VX
+VMRGHH          000100 ..... ..... ..... 00001001100    @VX
+VMRGHW          000100 ..... ..... ..... 00010001100    @VX
+VMRGLB          000100 ..... ..... ..... 00100001100    @VX
+VMRGLH          000100 ..... ..... ..... 00101001100    @VX
+VMRGLW          000100 ..... ..... ..... 00110001100    @VX
+
 ## Vector Integer Average Instructions
 
 VAVGSB          000100 ..... ..... ..... 10100000010    @VX
target/ppc/int_helper.c +6 -6
index ef4b2e75d6..a3adf746bd 100644
@@ -950,7 +950,7 @@ void helper_VMLADDUHM(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, ppc_avr_t *c,
 }
 
 #define VMRG_DO(name, element, access, ofs)                                  \
-    void helper_v##name(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)            \
+    void helper_V##name(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)            \
     {                                                                        \
         ppc_avr_t result;                                                    \
         int i, half = ARRAY_SIZE(r->element) / 2;                            \
@@ -963,11 +963,11 @@ void helper_VMLADDUHM(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, ppc_avr_t *c,
     }
 
 #define VMRG(suffix, element, access)          \
-    VMRG_DO(mrgl##suffix, element, access, half)   \
-    VMRG_DO(mrgh##suffix, element, access, 0)
-VMRG(b, u8, VsrB)
-VMRG(h, u16, VsrH)
-VMRG(w, u32, VsrW)
+    VMRG_DO(MRGL##suffix, element, access, half)   \
+    VMRG_DO(MRGH##suffix, element, access, 0)
+VMRG(B, u8, VsrB)
+VMRG(H, u16, VsrH)
+VMRG(W, u32, VsrW)
 #undef VMRG_DO
 #undef VMRG
 
target/ppc/translate/vmx-impl.c.inc +19 -6
index 92d6e8c603..00b359f031 100644
@@ -350,12 +350,25 @@ GEN_VXFORM_V(vsububm, MO_8, tcg_gen_gvec_sub, 0, 16);
 GEN_VXFORM_V(vsubuhm, MO_16, tcg_gen_gvec_sub, 0, 17);
 GEN_VXFORM_V(vsubuwm, MO_32, tcg_gen_gvec_sub, 0, 18);
 GEN_VXFORM_V(vsubudm, MO_64, tcg_gen_gvec_sub, 0, 19);
-GEN_VXFORM(vmrghb, 6, 0);
-GEN_VXFORM(vmrghh, 6, 1);
-GEN_VXFORM(vmrghw, 6, 2);
-GEN_VXFORM(vmrglb, 6, 4);
-GEN_VXFORM(vmrglh, 6, 5);
-GEN_VXFORM(vmrglw, 6, 6);
+
+static bool do_vmrg(DisasContext *ctx, arg_VX *a,
+                    void (*helper)(TCGv_ptr, TCGv_ptr, TCGv_ptr))
+{
+    TCGv_ptr ra, rb, rd;
+    REQUIRE_VECTOR(ctx);
+    ra = gen_avr_ptr(a->vra);
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+    helper(rd, ra, rb);
+    return true;
+}
+
+TRANS_FLAGS(ALTIVEC, VMRGHB, do_vmrg, gen_helper_VMRGHB);
+TRANS_FLAGS(ALTIVEC, VMRGHH, do_vmrg, gen_helper_VMRGHH);
+TRANS_FLAGS(ALTIVEC, VMRGHW, do_vmrg, gen_helper_VMRGHW);
+TRANS_FLAGS(ALTIVEC, VMRGLB, do_vmrg, gen_helper_VMRGLB);
+TRANS_FLAGS(ALTIVEC, VMRGLH, do_vmrg, gen_helper_VMRGLH);
+TRANS_FLAGS(ALTIVEC, VMRGLW, do_vmrg, gen_helper_VMRGLW);
 
 static void trans_vmrgew(DisasContext *ctx)
 {
target/ppc/translate/vmx-ops.c.inc +0 -6
index e28958a126..b8f77d3d2c 100644
@@ -33,12 +33,6 @@ GEN_VXFORM_DUAL(vsubuhm, bcdsub, 0, 17, PPC_ALTIVEC, PPC_NONE),
 GEN_VXFORM_DUAL(vsubuwm, bcdus, 0, 18, PPC_ALTIVEC, PPC2_ISA300),
 GEN_VXFORM_DUAL(vsubudm, bcds, 0, 19, PPC2_ALTIVEC_207, PPC2_ISA300),
 GEN_VXFORM_300(bcds, 0, 27),
-GEN_VXFORM(vmrghb, 6, 0),
-GEN_VXFORM(vmrghh, 6, 1),
-GEN_VXFORM(vmrghw, 6, 2),
-GEN_VXFORM(vmrglb, 6, 4),
-GEN_VXFORM(vmrglh, 6, 5),
-GEN_VXFORM(vmrglw, 6, 6),
 GEN_VXFORM_300(vextublx, 6, 24),
 GEN_VXFORM_300(vextuhlx, 6, 25),
 GEN_VXFORM_DUAL(vmrgow, vextuwlx, 6, 26, PPC_NONE, PPC2_ALTIVEC_207),
Signature · 1 line
2.55.0
Reviewed-by on v2 4/37
#14 lore raw

Re: [PATCH v2 04/37] target/ppc: Move vector merge instructions to decodetree

Quoted 19 lines · Chinmay Rath

From: Aboorva Devarajan <[email protected]>

Move below instructions to decodetree specification:

vmrg{hb, hh, hw, lb, lh, lw} : VX-form

The individual gen_vmrg* functions generated by GEN_VXFORM are replaced by a shared do_vmrg() helper dispatching to the existing gen_helper_VMRG* TCG helpers via the TRANS_FLAGS() macro, which also handles the ALTIVEC facility check.

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

Quoted 7 lines
---
 target/ppc/helper.h                 | 12 ++++++------
 target/ppc/insn32.decode            |  9 +++++++++
 target/ppc/int_helper.c             | 12 ++++++------
 target/ppc/translate/vmx-impl.c.inc | 25 +++++++++++++++++++------
 target/ppc/translate/vmx-ops.c.inc  |  6 ------
 5 files changed, 40 insertions(+), 24 deletions(-)
Quoted 130 lines
target/ppc/helper.h +6 -6
index 46ad78d1d4..6db917935c 100644
@@ -173,12 +173,12 @@ DEF_HELPER_4(vcmpeqfp_dot, void, env, avr, avr, avr)
 DEF_HELPER_4(vcmpgefp_dot, void, env, avr, avr, avr)
 DEF_HELPER_4(vcmpgtfp_dot, void, env, avr, avr, avr)
 DEF_HELPER_4(vcmpbfp_dot, void, env, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrglb, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrglh, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrglw, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrghb, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrghh, TCG_CALL_NO_RWG, void, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vmrghw, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGLB, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGLH, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGLW, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGHB, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGHH, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VMRGHW, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_FLAGS_3(VMULESB, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_FLAGS_3(VMULESH, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_FLAGS_3(VMULESW, TCG_CALL_NO_RWG, void, avr, avr, avr)
target/ppc/insn32.decode +9 -0
index 6b1cf9a0a9..8aca8a00c0 100644
@@ -795,6 +795,15 @@ VNOR            000100 ..... ..... ..... 10100000100    @VX
 VXOR            000100 ..... ..... ..... 10011000100    @VX
 VEQV            000100 ..... ..... ..... 11010000100    @VX
 
+## Vector Merge Instructions
+
+VMRGHB          000100 ..... ..... ..... 00000001100    @VX
+VMRGHH          000100 ..... ..... ..... 00001001100    @VX
+VMRGHW          000100 ..... ..... ..... 00010001100    @VX
+VMRGLB          000100 ..... ..... ..... 00100001100    @VX
+VMRGLH          000100 ..... ..... ..... 00101001100    @VX
+VMRGLW          000100 ..... ..... ..... 00110001100    @VX
+
 ## Vector Integer Average Instructions
 
 VAVGSB          000100 ..... ..... ..... 10100000010    @VX
target/ppc/int_helper.c +6 -6
index ef4b2e75d6..a3adf746bd 100644
@@ -950,7 +950,7 @@ void helper_VMLADDUHM(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, ppc_avr_t *c,
 }
 
 #define VMRG_DO(name, element, access, ofs)                                  \
-    void helper_v##name(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)            \
+    void helper_V##name(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)            \
     {                                                                        \
         ppc_avr_t result;                                                    \
         int i, half = ARRAY_SIZE(r->element) / 2;                            \
@@ -963,11 +963,11 @@ void helper_VMLADDUHM(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, ppc_avr_t *c,
     }
 
 #define VMRG(suffix, element, access)          \
-    VMRG_DO(mrgl##suffix, element, access, half)   \
-    VMRG_DO(mrgh##suffix, element, access, 0)
-VMRG(b, u8, VsrB)
-VMRG(h, u16, VsrH)
-VMRG(w, u32, VsrW)
+    VMRG_DO(MRGL##suffix, element, access, half)   \
+    VMRG_DO(MRGH##suffix, element, access, 0)
+VMRG(B, u8, VsrB)
+VMRG(H, u16, VsrH)
+VMRG(W, u32, VsrW)
 #undef VMRG_DO
 #undef VMRG
 
target/ppc/translate/vmx-impl.c.inc +19 -6
index 92d6e8c603..00b359f031 100644
@@ -350,12 +350,25 @@ GEN_VXFORM_V(vsububm, MO_8, tcg_gen_gvec_sub, 0, 16);
 GEN_VXFORM_V(vsubuhm, MO_16, tcg_gen_gvec_sub, 0, 17);
 GEN_VXFORM_V(vsubuwm, MO_32, tcg_gen_gvec_sub, 0, 18);
 GEN_VXFORM_V(vsubudm, MO_64, tcg_gen_gvec_sub, 0, 19);
-GEN_VXFORM(vmrghb, 6, 0);
-GEN_VXFORM(vmrghh, 6, 1);
-GEN_VXFORM(vmrghw, 6, 2);
-GEN_VXFORM(vmrglb, 6, 4);
-GEN_VXFORM(vmrglh, 6, 5);
-GEN_VXFORM(vmrglw, 6, 6);
+
+static bool do_vmrg(DisasContext *ctx, arg_VX *a,
+                    void (*helper)(TCGv_ptr, TCGv_ptr, TCGv_ptr))
+{
+    TCGv_ptr ra, rb, rd;
+    REQUIRE_VECTOR(ctx);
+    ra = gen_avr_ptr(a->vra);
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+    helper(rd, ra, rb);
+    return true;
+}
+
+TRANS_FLAGS(ALTIVEC, VMRGHB, do_vmrg, gen_helper_VMRGHB);
+TRANS_FLAGS(ALTIVEC, VMRGHH, do_vmrg, gen_helper_VMRGHH);
+TRANS_FLAGS(ALTIVEC, VMRGHW, do_vmrg, gen_helper_VMRGHW);
+TRANS_FLAGS(ALTIVEC, VMRGLB, do_vmrg, gen_helper_VMRGLB);
+TRANS_FLAGS(ALTIVEC, VMRGLH, do_vmrg, gen_helper_VMRGLH);
+TRANS_FLAGS(ALTIVEC, VMRGLW, do_vmrg, gen_helper_VMRGLW);
 
 static void trans_vmrgew(DisasContext *ctx)
 {
target/ppc/translate/vmx-ops.c.inc +0 -6
index e28958a126..b8f77d3d2c 100644
@@ -33,12 +33,6 @@ GEN_VXFORM_DUAL(vsubuhm, bcdsub, 0, 17, PPC_ALTIVEC, PPC_NONE),
 GEN_VXFORM_DUAL(vsubuwm, bcdus, 0, 18, PPC_ALTIVEC, PPC2_ISA300),
 GEN_VXFORM_DUAL(vsubudm, bcds, 0, 19, PPC2_ALTIVEC_207, PPC2_ISA300),
 GEN_VXFORM_300(bcds, 0, 27),
-GEN_VXFORM(vmrghb, 6, 0),
-GEN_VXFORM(vmrghh, 6, 1),
-GEN_VXFORM(vmrghw, 6, 2),
-GEN_VXFORM(vmrglb, 6, 4),
-GEN_VXFORM(vmrglh, 6, 5),
-GEN_VXFORM(vmrglw, 6, 6),
 GEN_VXFORM_300(vextublx, 6, 24),
 GEN_VXFORM_300(vextuhlx, 6, 25),
 GEN_VXFORM_DUAL(vmrgow, vextuwlx, 6, 26, PPC_NONE, PPC2_ALTIVEC_207),
Signature · 1 line
2.55.0
#15 lore raw

PATCH v2 5/37 target/ppc: Move vector pack instructions to decodetree

In reply to Chinmay Rath

From: Aboorva Devarajan <[email protected]>

Move below instructions to decodetree specification:

vpkuhum, vpkuwum, vpkuhus, vpkuwus : VX-form (PPC_ALTIVEC) vpkshus, vpkswus, vpkshss, vpkswss : VX-form (PPC_ALTIVEC)
vpkpx : VX-form (PPC_ALTIVEC)
vpkudum, vpkudus, vpksdus, vpksdss : VX-form (PPC2_ALTIVEC_207)

The legacy GEN_VXFORM_ENV/GEN_VXFORM macros that generated gen_vpk*() functions reading register fields from ctx->opcode are replaced by new do_vpk_env()/do_vpk() helpers in vmx-impl.c.inc that use the pre-extracted decodetree argument fields (a->vrt, a->vra, a->vrb) directly. The existing gen_helper_vpk*() implementations in int_helper.c are reused without changes.

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' option before and after the patch.

---
 target/ppc/helper.h                 | 26 +++++++--------
 target/ppc/insn32.decode            | 16 +++++++++
 target/ppc/int_helper.c             | 28 ++++++++--------
 target/ppc/translate/vmx-impl.c.inc | 51 +++++++++++++++++++++--------
 target/ppc/translate/vmx-ops.c.inc  | 13 --------
 5 files changed, 81 insertions(+), 53 deletions(-)
target/ppc/helper.h +13 -13
index 6db917935c..6df3de6516 100644
@@ -249,19 +249,19 @@ DEF_HELPER_FLAGS_4(VMSUMUBM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VMSUMMBM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VPERM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VPERMR, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
-DEF_HELPER_4(vpkshss, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkshus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkswss, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkswus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpksdss, void, env, avr, avr, avr)
-DEF_HELPER_4(vpksdus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuhus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuwus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkudus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuhum, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuwum, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkudum, void, env, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vpkpx, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_4(VPKSHSS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSHUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSWSS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSWUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSDSS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSDUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUHUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUWUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUDUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUHUM, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUWUM, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUDUM, void, env, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VPKPX, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_5(VMHADDSHS, void, env, avr, avr, avr, avr)
 DEF_HELPER_5(VMHRADDSHS, void, env, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VMSUMUHM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
target/ppc/insn32.decode +16 -0
index 8aca8a00c0..c51efbfd33 100644
@@ -1054,6 +1054,22 @@ VSTRIHR         000100 ..... 00011 ..... . 0000001101   @VX_tb_rc
 VCLRLB          000100 ..... ..... ..... 00110001101    @VX
 VCLRRB          000100 ..... ..... ..... 00111001101    @VX
 
+## Vector Pack Instructions
+
+VPKUHUM         000100 ..... ..... ..... 0000000111 0    @VX
+VPKUWUM         000100 ..... ..... ..... 0000100111 0    @VX
+VPKUDUM         000100 ..... ..... ..... 1000100111 0    @VX
+VPKUHUS         000100 ..... ..... ..... 0001000111 0    @VX
+VPKUWUS         000100 ..... ..... ..... 0001100111 0    @VX
+VPKUDUS         000100 ..... ..... ..... 1001100111 0    @VX
+VPKSHUS         000100 ..... ..... ..... 0010000111 0    @VX
+VPKSWUS         000100 ..... ..... ..... 0010100111 0    @VX
+VPKSDUS         000100 ..... ..... ..... 1010100111 0    @VX
+VPKSHSS         000100 ..... ..... ..... 0011000111 0    @VX
+VPKSWSS         000100 ..... ..... ..... 0011100111 0    @VX
+VPKSDSS         000100 ..... ..... ..... 1011100111 0    @VX
+VPKPX           000100 ..... ..... ..... 0110000111 0    @VX
+
 # VSX Load/Store Instructions
 
 LXSD            111001 ..... ..... .............. 10    @DS
target/ppc/int_helper.c +14 -14
index a3adf746bd..dc55ccadf7 100644
@@ -1465,7 +1465,7 @@ void helper_VPMSUMD(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
 #else
 #define PKBIG 0
 #endif
-void helper_vpkpx(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
+void helper_VPKPX(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
 {
     int i, j;
     ppc_avr_t result;
@@ -1488,7 +1488,7 @@ void helper_vpkpx(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
 }
 
 #define VPK(suffix, from, to, cvt, dosat)                               \
-    void helper_vpk##suffix(CPUPPCState *env, ppc_avr_t *r,             \
+    void helper_VPK##suffix(CPUPPCState *env, ppc_avr_t *r,             \
                             ppc_avr_t *a, ppc_avr_t *b)                 \
     {                                                                   \
         int i;                                                          \
@@ -1507,18 +1507,18 @@ void helper_vpkpx(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
         }                                                               \
     }
 #define I(x, y) (x)
-VPK(shss, s16, s8, cvtshsb, 1)
-VPK(shus, s16, u8, cvtshub, 1)
-VPK(swss, s32, s16, cvtswsh, 1)
-VPK(swus, s32, u16, cvtswuh, 1)
-VPK(sdss, s64, s32, cvtsdsw, 1)
-VPK(sdus, s64, u32, cvtsduw, 1)
-VPK(uhus, u16, u8, cvtuhub, 1)
-VPK(uwus, u32, u16, cvtuwuh, 1)
-VPK(udus, u64, u32, cvtuduw, 1)
-VPK(uhum, u16, u8, I, 0)
-VPK(uwum, u32, u16, I, 0)
-VPK(udum, u64, u32, I, 0)
+VPK(SHSS, s16, s8, cvtshsb, 1)
+VPK(SHUS, s16, u8, cvtshub, 1)
+VPK(SWSS, s32, s16, cvtswsh, 1)
+VPK(SWUS, s32, u16, cvtswuh, 1)
+VPK(SDSS, s64, s32, cvtsdsw, 1)
+VPK(SDUS, s64, u32, cvtsduw, 1)
+VPK(UHUS, u16, u8, cvtuhub, 1)
+VPK(UWUS, u32, u16, cvtuwuh, 1)
+VPK(UDUS, u64, u32, cvtuduw, 1)
+VPK(UHUM, u16, u8, I, 0)
+VPK(UWUM, u32, u16, I, 0)
+VPK(UDUM, u64, u32, I, 0)
 #undef I
 #undef VPK
 #undef PKBIG
target/ppc/translate/vmx-impl.c.inc +38 -13

57 lines not on this page · raw message

target/ppc/translate/vmx-ops.c.inc +0 -13
index b8f77d3d2c..2820976c0f 100644
@@ -59,19 +59,6 @@ GEN_VXFORM_300(bcdtrunc, 0, 20),
 GEN_VXFORM_300(bcdutrunc, 0, 21),
 GEN_VXFORM(vsl, 2, 7),
 GEN_VXFORM(vsr, 2, 11),
-GEN_VXFORM(vpkuhum, 7, 0),
-GEN_VXFORM(vpkuwum, 7, 1),
-GEN_VXFORM_207(vpkudum, 7, 17),
-GEN_VXFORM(vpkuhus, 7, 2),
-GEN_VXFORM(vpkuwus, 7, 3),
-GEN_VXFORM_207(vpkudus, 7, 19),
-GEN_VXFORM(vpkshus, 7, 4),
-GEN_VXFORM(vpkswus, 7, 5),
-GEN_VXFORM_207(vpksdus, 7, 21),
-GEN_VXFORM(vpkshss, 7, 6),
-GEN_VXFORM(vpkswss, 7, 7),
-GEN_VXFORM_207(vpksdss, 7, 23),
-GEN_VXFORM(vpkpx, 7, 12),
 GEN_VXFORM(vsum4ubs, 4, 24),
 GEN_VXFORM(vsum4sbs, 4, 28),
 GEN_VXFORM(vsum4shs, 4, 25),
Signature · 1 line
2.55.0
Reviewed-by on v2 5/37
#16 lore raw

Re: [PATCH v2 05/37] target/ppc: Move vector pack instructions to decodetree

Quoted 24 lines · Chinmay Rath

From: Aboorva Devarajan <[email protected]>

Move below instructions to decodetree specification:

vpkuhum, vpkuwum, vpkuhus, vpkuwus : VX-form (PPC_ALTIVEC) vpkshus, vpkswus, vpkshss, vpkswss : VX-form (PPC_ALTIVEC)
vpkpx : VX-form (PPC_ALTIVEC)
vpkudum, vpkudus, vpksdus, vpksdss : VX-form (PPC2_ALTIVEC_207)

The legacy GEN_VXFORM_ENV/GEN_VXFORM macros that generated gen_vpk*() functions reading register fields from ctx->opcode are replaced by new do_vpk_env()/do_vpk() helpers in vmx-impl.c.inc that use the pre-extracted decodetree argument fields (a->vrt, a->vra, a->vrb) directly. The existing gen_helper_vpk*() implementations in int_helper.c are reused without changes.

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' option before and after the patch.

Quoted 7 lines
---
 target/ppc/helper.h                 | 26 +++++++--------
 target/ppc/insn32.decode            | 16 +++++++++
 target/ppc/int_helper.c             | 28 ++++++++--------
 target/ppc/translate/vmx-impl.c.inc | 51 +++++++++++++++++++++--------
 target/ppc/translate/vmx-ops.c.inc  | 13 --------
 5 files changed, 81 insertions(+), 53 deletions(-)
Quoted 207 lines
target/ppc/helper.h +13 -13
index 6db917935c..6df3de6516 100644
@@ -249,19 +249,19 @@ DEF_HELPER_FLAGS_4(VMSUMUBM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VMSUMMBM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VPERM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VPERMR, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
-DEF_HELPER_4(vpkshss, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkshus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkswss, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkswus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpksdss, void, env, avr, avr, avr)
-DEF_HELPER_4(vpksdus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuhus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuwus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkudus, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuhum, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkuwum, void, env, avr, avr, avr)
-DEF_HELPER_4(vpkudum, void, env, avr, avr, avr)
-DEF_HELPER_FLAGS_3(vpkpx, TCG_CALL_NO_RWG, void, avr, avr, avr)
+DEF_HELPER_4(VPKSHSS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSHUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSWSS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSWUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSDSS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKSDUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUHUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUWUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUDUS, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUHUM, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUWUM, void, env, avr, avr, avr)
+DEF_HELPER_4(VPKUDUM, void, env, avr, avr, avr)
+DEF_HELPER_FLAGS_3(VPKPX, TCG_CALL_NO_RWG, void, avr, avr, avr)
 DEF_HELPER_5(VMHADDSHS, void, env, avr, avr, avr, avr)
 DEF_HELPER_5(VMHRADDSHS, void, env, avr, avr, avr, avr)
 DEF_HELPER_FLAGS_4(VMSUMUHM, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
target/ppc/insn32.decode +16 -0
index 8aca8a00c0..c51efbfd33 100644
@@ -1054,6 +1054,22 @@ VSTRIHR         000100 ..... 00011 ..... . 0000001101   @VX_tb_rc
 VCLRLB          000100 ..... ..... ..... 00110001101    @VX
 VCLRRB          000100 ..... ..... ..... 00111001101    @VX
 
+## Vector Pack Instructions
+
+VPKUHUM         000100 ..... ..... ..... 0000000111 0    @VX
+VPKUWUM         000100 ..... ..... ..... 0000100111 0    @VX
+VPKUDUM         000100 ..... ..... ..... 1000100111 0    @VX
+VPKUHUS         000100 ..... ..... ..... 0001000111 0    @VX
+VPKUWUS         000100 ..... ..... ..... 0001100111 0    @VX
+VPKUDUS         000100 ..... ..... ..... 1001100111 0    @VX
+VPKSHUS         000100 ..... ..... ..... 0010000111 0    @VX
+VPKSWUS         000100 ..... ..... ..... 0010100111 0    @VX
+VPKSDUS         000100 ..... ..... ..... 1010100111 0    @VX
+VPKSHSS         000100 ..... ..... ..... 0011000111 0    @VX
+VPKSWSS         000100 ..... ..... ..... 0011100111 0    @VX
+VPKSDSS         000100 ..... ..... ..... 1011100111 0    @VX
+VPKPX           000100 ..... ..... ..... 0110000111 0    @VX
+
 # VSX Load/Store Instructions
 
 LXSD            111001 ..... ..... .............. 10    @DS
target/ppc/int_helper.c +14 -14
index a3adf746bd..dc55ccadf7 100644
@@ -1465,7 +1465,7 @@ void helper_VPMSUMD(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
 #else
 #define PKBIG 0
 #endif
-void helper_vpkpx(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
+void helper_VPKPX(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
 {
     int i, j;
     ppc_avr_t result;
@@ -1488,7 +1488,7 @@ void helper_vpkpx(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
 }
 
 #define VPK(suffix, from, to, cvt, dosat)                               \
-    void helper_vpk##suffix(CPUPPCState *env, ppc_avr_t *r,             \
+    void helper_VPK##suffix(CPUPPCState *env, ppc_avr_t *r,             \
                             ppc_avr_t *a, ppc_avr_t *b)                 \
     {                                                                   \
         int i;                                                          \
@@ -1507,18 +1507,18 @@ void helper_vpkpx(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b)
         }                                                               \
     }
 #define I(x, y) (x)
-VPK(shss, s16, s8, cvtshsb, 1)
-VPK(shus, s16, u8, cvtshub, 1)
-VPK(swss, s32, s16, cvtswsh, 1)
-VPK(swus, s32, u16, cvtswuh, 1)
-VPK(sdss, s64, s32, cvtsdsw, 1)
-VPK(sdus, s64, u32, cvtsduw, 1)
-VPK(uhus, u16, u8, cvtuhub, 1)
-VPK(uwus, u32, u16, cvtuwuh, 1)
-VPK(udus, u64, u32, cvtuduw, 1)
-VPK(uhum, u16, u8, I, 0)
-VPK(uwum, u32, u16, I, 0)
-VPK(udum, u64, u32, I, 0)
+VPK(SHSS, s16, s8, cvtshsb, 1)
+VPK(SHUS, s16, u8, cvtshub, 1)
+VPK(SWSS, s32, s16, cvtswsh, 1)
+VPK(SWUS, s32, u16, cvtswuh, 1)
+VPK(SDSS, s64, s32, cvtsdsw, 1)
+VPK(SDUS, s64, u32, cvtsduw, 1)
+VPK(UHUS, u16, u8, cvtuhub, 1)
+VPK(UWUS, u32, u16, cvtuwuh, 1)
+VPK(UDUS, u64, u32, cvtuduw, 1)
+VPK(UHUM, u16, u8, I, 0)
+VPK(UWUM, u32, u16, I, 0)
+VPK(UDUM, u64, u32, I, 0)
 #undef I
 #undef VPK
 #undef PKBIG
target/ppc/translate/vmx-impl.c.inc +38 -13
index 00b359f031..65230f3461 100644
@@ -1070,19 +1070,44 @@ TRANS(VRLQMI, do_vector_rotl_quad, false, true)
 
 GEN_VXFORM_TRANS(vsl, 2, 7);
 GEN_VXFORM_TRANS(vsr, 2, 11);
-GEN_VXFORM_ENV(vpkuhum, 7, 0);
-GEN_VXFORM_ENV(vpkuwum, 7, 1);
-GEN_VXFORM_ENV(vpkudum, 7, 17);
-GEN_VXFORM_ENV(vpkuhus, 7, 2);
-GEN_VXFORM_ENV(vpkuwus, 7, 3);
-GEN_VXFORM_ENV(vpkudus, 7, 19);
-GEN_VXFORM_ENV(vpkshus, 7, 4);
-GEN_VXFORM_ENV(vpkswus, 7, 5);
-GEN_VXFORM_ENV(vpksdus, 7, 21);
-GEN_VXFORM_ENV(vpkshss, 7, 6);
-GEN_VXFORM_ENV(vpkswss, 7, 7);
-GEN_VXFORM_ENV(vpksdss, 7, 23);
-GEN_VXFORM(vpkpx, 7, 12);
+static bool do_vpk_env(DisasContext *ctx, arg_VX *a,
+                       void (*gen_helper)(TCGv_ptr, TCGv_ptr,
+                                          TCGv_ptr, TCGv_ptr))
+{
+    TCGv_ptr ra, rb, rd;
+    REQUIRE_VECTOR(ctx);
+    ra = gen_avr_ptr(a->vra);
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+    gen_helper(tcg_env, rd, ra, rb);
+    return true;
+}
+
+static bool do_vpk(DisasContext *ctx, arg_VX *a,
+                   void (*gen_helper)(TCGv_ptr, TCGv_ptr, TCGv_ptr))
+{
+    TCGv_ptr ra, rb, rd;
+    REQUIRE_VECTOR(ctx);
+    ra = gen_avr_ptr(a->vra);
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+    gen_helper(rd, ra, rb);
+    return true;
+}
+
+TRANS_FLAGS(ALTIVEC, VPKUHUM, do_vpk_env, gen_helper_VPKUHUM)
+TRANS_FLAGS(ALTIVEC, VPKUWUM, do_vpk_env, gen_helper_VPKUWUM)
+TRANS_FLAGS2(ALTIVEC_207, VPKUDUM, do_vpk_env, gen_helper_VPKUDUM)
+TRANS_FLAGS(ALTIVEC, VPKUHUS, do_vpk_env, gen_helper_VPKUHUS)
+TRANS_FLAGS(ALTIVEC, VPKUWUS, do_vpk_env, gen_helper_VPKUWUS)
+TRANS_FLAGS2(ALTIVEC_207, VPKUDUS, do_vpk_env, gen_helper_VPKUDUS)
+TRANS_FLAGS(ALTIVEC, VPKSHUS, do_vpk_env, gen_helper_VPKSHUS)
+TRANS_FLAGS(ALTIVEC, VPKSWUS, do_vpk_env, gen_helper_VPKSWUS)
+TRANS_FLAGS2(ALTIVEC_207, VPKSDUS, do_vpk_env, gen_helper_VPKSDUS)
+TRANS_FLAGS(ALTIVEC, VPKSHSS, do_vpk_env, gen_helper_VPKSHSS)
+TRANS_FLAGS(ALTIVEC, VPKSWSS, do_vpk_env, gen_helper_VPKSWSS)
+TRANS_FLAGS2(ALTIVEC_207, VPKSDSS, do_vpk_env, gen_helper_VPKSDSS)
+TRANS_FLAGS(ALTIVEC, VPKPX, do_vpk, gen_helper_VPKPX)
 GEN_VXFORM_ENV(vsum4ubs, 4, 24);
 GEN_VXFORM_ENV(vsum4sbs, 4, 28);
 GEN_VXFORM_ENV(vsum4shs, 4, 25);
target/ppc/translate/vmx-ops.c.inc +0 -13
index b8f77d3d2c..2820976c0f 100644
@@ -59,19 +59,6 @@ GEN_VXFORM_300(bcdtrunc, 0, 20),
 GEN_VXFORM_300(bcdutrunc, 0, 21),
 GEN_VXFORM(vsl, 2, 7),
 GEN_VXFORM(vsr, 2, 11),
-GEN_VXFORM(vpkuhum, 7, 0),
-GEN_VXFORM(vpkuwum, 7, 1),
-GEN_VXFORM_207(vpkudum, 7, 17),
-GEN_VXFORM(vpkuhus, 7, 2),
-GEN_VXFORM(vpkuwus, 7, 3),
-GEN_VXFORM_207(vpkudus, 7, 19),
-GEN_VXFORM(vpkshus, 7, 4),
-GEN_VXFORM(vpkswus, 7, 5),
-GEN_VXFORM_207(vpksdus, 7, 21),
-GEN_VXFORM(vpkshss, 7, 6),
-GEN_VXFORM(vpkswss, 7, 7),
-GEN_VXFORM_207(vpksdss, 7, 23),
-GEN_VXFORM(vpkpx, 7, 12),
 GEN_VXFORM(vsum4ubs, 4, 24),
 GEN_VXFORM(vsum4sbs, 4, 28),
 GEN_VXFORM(vsum4shs, 4, 25),
Signature · 1 line
2.55.0
#17 lore raw

PATCH v2 6/37 target/ppc: Move st{b, h, w, d, q}cx instructions to decodetree

In reply to Chinmay Rath

From: Utkarsh Verma <[email protected]>

Move the following store conditional instructions to decodetree specification:

stbcx. : X-form
sthcx. : X-form
stwcx. : X-form
stdcx. : X-form
stqcx. : X-form

The changes were verified by checking that the generated TCG ops remain semantically identical to the legacy translation, using logs captured with the -d in_asm,op flag.

---
 target/ppc/insn32.decode                   |   8 ++
 target/ppc/translate.c                     | 103 ---------------------
 target/ppc/translate/fixedpoint-impl.c.inc |  84 +++++++++++++++++
 3 files changed, 92 insertions(+), 103 deletions(-)
target/ppc/insn32.decode +8 -0
index c51efbfd33..701c66b347 100644
@@ -732,6 +732,14 @@ DSCLIQ          111111 ..... ..... ...... 001000010 .   @Z22_tap_sh_rc
 DSCRI           111011 ..... ..... ...... 001100010 .   @Z22_ta_sh_rc
 DSCRIQ          111111 ..... ..... ...... 001100010 .   @Z22_tap_sh_rc
 
+### Store Conditional Instructions
+
+STBCX           011111 ..... ..... ..... 1010110110 1   @X
+STHCX           011111 ..... ..... ..... 1011010110 1   @X
+STWCX           011111 ..... ..... ..... 0010010110 1   @X
+STDCX           011111 ..... ..... ..... 0011010110 1   @X
+STQCX           011111 ..... ..... ..... 0010110110 1   @X
+
 ## Vector Exclusive-OR-based Instructions
 
 VPMSUMD         000100 ..... ..... ..... 10011001000    @VX
target/ppc/translate.c +0 -103

118 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +84 -0

90 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 6/37
#18 lore raw

Re: [PATCH v2 06/37] target/ppc: Move st{b, h, w, d, q}cx instructions to decodetree

Quoted 19 lines · Chinmay Rath

From: Utkarsh Verma <[email protected]>

Move the following store conditional instructions to decodetree specification:

stbcx. : X-form
sthcx. : X-form
stwcx. : X-form
stdcx. : X-form
stqcx. : X-form

The changes were verified by checking that the generated TCG ops remain semantically identical to the legacy translation, using logs captured with the -d in_asm,op flag.

Quoted 5 lines
---
 target/ppc/insn32.decode                   |   8 ++
 target/ppc/translate.c                     | 103 ---------------------
 target/ppc/translate/fixedpoint-impl.c.inc |  84 +++++++++++++++++
 3 files changed, 92 insertions(+), 103 deletions(-)
Quoted 242 lines
target/ppc/insn32.decode +8 -0
index c51efbfd33..701c66b347 100644
@@ -732,6 +732,14 @@ DSCLIQ          111111 ..... ..... ...... 001000010 .   @Z22_tap_sh_rc
 DSCRI           111011 ..... ..... ...... 001100010 .   @Z22_ta_sh_rc
 DSCRIQ          111111 ..... ..... ...... 001100010 .   @Z22_tap_sh_rc
 
+### Store Conditional Instructions
+
+STBCX           011111 ..... ..... ..... 1010110110 1   @X
+STHCX           011111 ..... ..... ..... 1011010110 1   @X
+STWCX           011111 ..... ..... ..... 0010010110 1   @X
+STDCX           011111 ..... ..... ..... 0011010110 1   @X
+STQCX           011111 ..... ..... ..... 0010110110 1   @X
+
 ## Vector Exclusive-OR-based Instructions
 
 VPMSUMD         000100 ..... ..... ..... 10011001000    @VX
target/ppc/translate.c +0 -103
index 60c88a6153..ee14fbfd96 100644
@@ -3153,104 +3153,6 @@ static void gen_stdat(DisasContext *ctx)
 }
 #endif
 
-static void gen_conditional_store(DisasContext *ctx, MemOp memop)
-{
-    TCGLabel *lfail;
-    TCGv EA;
-    TCGv cr0;
-    TCGv t0;
-    int rs = rS(ctx->opcode);
-
-    lfail = gen_new_label();
-    EA = tcg_temp_new();
-    cr0 = tcg_temp_new();
-    t0 = tcg_temp_new();
-
-    tcg_gen_mov_tl(cr0, cpu_so);
-    gen_set_access_type(ctx, ACCESS_RES);
-    gen_addr_reg_index(ctx, EA);
-    tcg_gen_brcond_tl(TCG_COND_NE, EA, cpu_reserve, lfail);
-    tcg_gen_brcondi_tl(TCG_COND_NE, cpu_reserve_length, memop_size(memop), lfail);
-
-    tcg_gen_atomic_cmpxchg_tl(t0, cpu_reserve, cpu_reserve_val,
-                              cpu_gpr[rs], ctx->mem_idx,
-                              DEF_MEMOP(memop) | MO_ALIGN);
-    tcg_gen_setcond_tl(TCG_COND_EQ, t0, t0, cpu_reserve_val);
-    tcg_gen_shli_tl(t0, t0, CRF_EQ_BIT);
-    tcg_gen_or_tl(cr0, cr0, t0);
-
-    gen_set_label(lfail);
-    tcg_gen_trunc_tl_i32(cpu_crf[0], cr0);
-    tcg_gen_movi_tl(cpu_reserve, -1);
-}
-
-#define STCX(name, memop)                  \
-static void gen_##name(DisasContext *ctx)  \
-{                                          \
-    gen_conditional_store(ctx, memop);     \
-}
-
-STCX(stbcx_, MO_UB)
-STCX(sthcx_, MO_UW)
-STCX(stwcx_, MO_UL)
-
-#if defined(TARGET_PPC64)
-/* stdcx. */
-STCX(stdcx_, MO_UQ)
-
-/* stqcx. */
-static void gen_stqcx_(DisasContext *ctx)
-{
-    TCGLabel *lfail;
-    TCGv EA, t0, t1;
-    TCGv cr0;
-    TCGv_i128 cmp, val;
-    int rs = rS(ctx->opcode);
-
-    if (unlikely(rs & 1)) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-        return;
-    }
-
-    lfail = gen_new_label();
-    EA = tcg_temp_new();
-    cr0 = tcg_temp_new();
-
-    tcg_gen_mov_tl(cr0, cpu_so);
-    gen_set_access_type(ctx, ACCESS_RES);
-    gen_addr_reg_index(ctx, EA);
-    tcg_gen_brcond_tl(TCG_COND_NE, EA, cpu_reserve, lfail);
-    tcg_gen_brcondi_tl(TCG_COND_NE, cpu_reserve_length, 16, lfail);
-
-    cmp = tcg_temp_new_i128();
-    val = tcg_temp_new_i128();
-
-    tcg_gen_concat_i64_i128(cmp, cpu_reserve_val2, cpu_reserve_val);
-
-    /* Note that the low part is always in RS+1, even in LE mode.  */
-    tcg_gen_concat_i64_i128(val, cpu_gpr[rs + 1], cpu_gpr[rs]);
-
-    tcg_gen_atomic_cmpxchg_i128(val, cpu_reserve, cmp, val, ctx->mem_idx,
-                                DEF_MEMOP(MO_128 | MO_ALIGN));
-
-    t0 = tcg_temp_new();
-    t1 = tcg_temp_new();
-    tcg_gen_extr_i128_i64(t1, t0, val);
-
-    tcg_gen_xor_tl(t1, t1, cpu_reserve_val2);
-    tcg_gen_xor_tl(t0, t0, cpu_reserve_val);
-    tcg_gen_or_tl(t0, t0, t1);
-
-    tcg_gen_setcondi_tl(TCG_COND_EQ, t0, t0, 0);
-    tcg_gen_shli_tl(t0, t0, CRF_EQ_BIT);
-    tcg_gen_or_tl(cr0, cr0, t0);
-
-    gen_set_label(lfail);
-    tcg_gen_trunc_tl_i32(cpu_crf[0], cr0);
-    tcg_gen_movi_tl(cpu_reserve, -1);
-}
-#endif /* defined(TARGET_PPC64) */
-
 /* wait */
 static void gen_wait(DisasContext *ctx)
 {
@@ -5899,14 +5801,9 @@ GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 GEN_HANDLER(isync, 0x13, 0x16, 0x04, 0x03FFF801, PPC_MEM),
 GEN_HANDLER_E(lwat, 0x1F, 0x06, 0x12, 0x00000001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(stwat, 0x1F, 0x06, 0x16, 0x00000001, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER_E(stbcx_, 0x1F, 0x16, 0x15, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
-GEN_HANDLER_E(sthcx_, 0x1F, 0x16, 0x16, 0, PPC_NONE, PPC2_ATOMIC_ISA206),
-GEN_HANDLER2(stwcx_, "stwcx.", 0x1F, 0x16, 0x04, 0x00000000, PPC_RES),
 #if defined(TARGET_PPC64)
 GEN_HANDLER_E(ldat, 0x1F, 0x06, 0x13, 0x00000001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(stdat, 0x1F, 0x06, 0x17, 0x00000001, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER2(stdcx_, "stdcx.", 0x1F, 0x16, 0x06, 0x00000000, PPC_64B),
-GEN_HANDLER_E(stqcx_, 0x1F, 0x16, 0x05, 0, PPC_NONE, PPC2_LSQ_ISA207),
 #endif
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
target/ppc/translate/fixedpoint-impl.c.inc +84 -0
index fa0191e866..91f9c6c391 100644
@@ -204,6 +204,90 @@ TRANS64(PSTD, do_ldst_PLS_D, false, true, MO_UQ)
 TRANS64(STQ, do_ldst_quad, true, false);
 TRANS64(PSTQ, do_ldst_quad_PLS_D, true);
 
+/* Store Conditional Instructions */
+
+static bool do_store_cond(DisasContext *ctx, arg_X *a, MemOp mop)
+{
+    TCGLabel *lfail = gen_new_label();
+    TCGv ea = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    TCGv cr0 = tcg_temp_new();
+    TCGv t0 = tcg_temp_new();
+
+    tcg_gen_mov_tl(cr0, cpu_so);
+    gen_set_access_type(ctx, ACCESS_RES);
+
+    tcg_gen_brcond_tl(TCG_COND_NE, ea, cpu_reserve, lfail);
+    tcg_gen_brcondi_tl(TCG_COND_NE, cpu_reserve_length, memop_size(mop), lfail);
+
+    tcg_gen_atomic_cmpxchg_tl(t0, cpu_reserve, cpu_reserve_val,
+                              cpu_gpr[a->rt], ctx->mem_idx,
+                              DEF_MEMOP(mop) | MO_ALIGN);
+    tcg_gen_setcond_tl(TCG_COND_EQ, t0, t0, cpu_reserve_val);
+    tcg_gen_shli_tl(t0, t0, CRF_EQ_BIT);
+    tcg_gen_or_tl(cr0, cr0, t0);
+
+    gen_set_label(lfail);
+    tcg_gen_trunc_tl_i32(cpu_crf[0], cr0);
+    tcg_gen_movi_tl(cpu_reserve, -1);
+    return true;
+}
+
+TRANS_FLAGS2(ATOMIC_ISA206, STBCX, do_store_cond, MO_UB);
+TRANS_FLAGS2(ATOMIC_ISA206, STHCX, do_store_cond, MO_UW);
+TRANS(STWCX, do_store_cond, MO_UL);
+TRANS64(STDCX, do_store_cond, MO_UQ);
+
+static bool trans_STQCX(DisasContext *ctx, arg_STQCX *a)
+{
+    REQUIRE_64BIT(ctx);
+    REQUIRE_INSNS_FLAGS2(ctx, LSQ_ISA207);
+#if defined(TARGET_PPC64)
+    TCGLabel *lfail = gen_new_label();
+    TCGv ea = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    TCGv t0 = tcg_temp_new();
+    TCGv t1 = tcg_temp_new();
+    TCGv cr0 = tcg_temp_new();
+    TCGv_i128 cmp = tcg_temp_new_i128();
+    TCGv_i128 val = tcg_temp_new_i128();
+
+    if (unlikely(a->rt & 1)) {
+        gen_invalid(ctx);
+        return true;
+    }
+
+    tcg_gen_mov_tl(cr0, cpu_so);
+    gen_set_access_type(ctx, ACCESS_RES);
+
+    tcg_gen_brcond_tl(TCG_COND_NE, ea, cpu_reserve, lfail);
+    tcg_gen_brcondi_tl(TCG_COND_NE, cpu_reserve_length, 16, lfail);
+
+    tcg_gen_concat_i64_i128(cmp, cpu_reserve_val2, cpu_reserve_val);
+
+    /* Note that the low part is always in RS+1, even in LE mode.  */
+    tcg_gen_concat_i64_i128(val, cpu_gpr[a->rt + 1], cpu_gpr[a->rt]);
+
+    tcg_gen_atomic_cmpxchg_i128(val, cpu_reserve, cmp, val, ctx->mem_idx,
+                                DEF_MEMOP(MO_128 | MO_ALIGN));
+
+    tcg_gen_extr_i128_i64(t1, t0, val);
+
+    tcg_gen_xor_tl(t1, t1, cpu_reserve_val2);
+    tcg_gen_xor_tl(t0, t0, cpu_reserve_val);
+    tcg_gen_or_tl(t0, t0, t1);
+
+    tcg_gen_setcondi_tl(TCG_COND_EQ, t0, t0, 0);
+    tcg_gen_shli_tl(t0, t0, CRF_EQ_BIT);
+    tcg_gen_or_tl(cr0, cr0, t0);
+
+    gen_set_label(lfail);
+    tcg_gen_trunc_tl_i32(cpu_crf[0], cr0);
+    tcg_gen_movi_tl(cpu_reserve, -1);
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+
 /*
  * Fixed-Point Compare Instructions
  */
Signature · 1 line
2.55.0
#19 lore raw

PATCH v2 7/37 target/ppc: convert slw, srw instruction via decode spec

In reply to Chinmay Rath

From: "Mukesh Kumar Chaurasiya (IBM)" <[email protected]>

Instructions converted:

- slw
- srw

---
 target/ppc/insn32.decode                   |  6 +++
 target/ppc/translate.c                     | 49 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 48 +++++++++++++++++++++
 3 files changed, 54 insertions(+), 49 deletions(-)
target/ppc/insn32.decode +6 -0
index 701c66b347..66d2f455a9 100644
@@ -505,6 +505,12 @@ HASHCHK         011111 ..... ..... ..... 1011110010 .   @X_DW
 HASHSTP         011111 ..... ..... ..... 1010010010 .   @X_DW
 HASHCHKP        011111 ..... ..... ..... 1010110010 .   @X_DW
 
+# Fixed-Point Shift Instructions
+
+SLW		011111 ..... ..... ..... 0000011000 .	@X_rc
+SRW		011111 ..... ..... ..... 1000011000 .	@X_rc
+
+
 ## BCD Assist
 
 ADDG6S          011111 ..... ..... ..... - 001001010 -  @X
target/ppc/translate.c +0 -49

69 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +48 -0

51 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 7/37
#20 lore raw

Re: [PATCH v2 07/37] target/ppc: convert slw, srw instruction via decode spec

Quoted 21 lines · Chinmay Rath

From: "Mukesh Kumar Chaurasiya (IBM)" <[email protected]>

Instructions converted:

- slw
- srw

---
 target/ppc/insn32.decode                   |  6 +++
 target/ppc/translate.c                     | 49 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 48 +++++++++++++++++++++
 3 files changed, 54 insertions(+), 49 deletions(-)
target/ppc/insn32.decode +0 -0
index 701c66b347..66d2f455a9 100644

target/ppc/insn32.decode HASHCHK 011111 ..... ..... ..... 1011110010 . @X_DW · line 513

Quoted 9 lines
target/ppc/insn32.decode +6 -0
@@ -505,6 +505,12 @@ HASHCHK         011111 ..... ..... ..... 1011110010 .   @X_DW
 HASHSTP         011111 ..... ..... ..... 1010010010 .   @X_DW
 HASHCHKP        011111 ..... ..... ..... 1010110010 .   @X_DW
 
+# Fixed-Point Shift Instructions
+
+SLW		011111 ..... ..... ..... 0000011000 .	@X_rc
+SRW		011111 ..... ..... ..... 1000011000 .	@X_rc
+
+

Nit: this looks like an extra blank line - can be removed. Otherwise,

Quoted 138 lines
target/ppc/insn32.decode +0 -0
 ## BCD Assist
 
 ADDG6S          011111 ..... ..... ..... - 001001010 -  @X
target/ppc/translate.c +0 -49
index ee14fbfd96..f2b7505634 100644
@@ -2315,29 +2315,6 @@ GEN_PPC64_R4(rldimi, 0x1E, 0x06);
 
 /***                             Integer shift                             ***/
 
-/* slw & slw. */
-static void gen_slw(DisasContext *ctx)
-{
-    TCGv t0, t1;
-
-    t0 = tcg_temp_new();
-    /* AND rS with a mask that is 0 when rB >= 0x20 */
-#if defined(TARGET_PPC64)
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x3a);
-    tcg_gen_sari_tl(t0, t0, 0x3f);
-#else
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x1a);
-    tcg_gen_sari_tl(t0, t0, 0x1f);
-#endif
-    tcg_gen_andc_tl(t0, cpu_gpr[rS(ctx->opcode)], t0);
-    t1 = tcg_temp_new();
-    tcg_gen_andi_tl(t1, cpu_gpr[rB(ctx->opcode)], 0x1f);
-    tcg_gen_shl_tl(cpu_gpr[rA(ctx->opcode)], t0, t1);
-    tcg_gen_ext32u_tl(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rA(ctx->opcode)]);
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, cpu_gpr[rA(ctx->opcode)]);
-    }
-}
 
 /* sraw & sraw. */
 static void gen_sraw(DisasContext *ctx)
@@ -2379,30 +2356,6 @@ static void gen_srawi(DisasContext *ctx)
     }
 }
 
-/* srw & srw. */
-static void gen_srw(DisasContext *ctx)
-{
-    TCGv t0, t1;
-
-    t0 = tcg_temp_new();
-    /* AND rS with a mask that is 0 when rB >= 0x20 */
-#if defined(TARGET_PPC64)
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x3a);
-    tcg_gen_sari_tl(t0, t0, 0x3f);
-#else
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x1a);
-    tcg_gen_sari_tl(t0, t0, 0x1f);
-#endif
-    tcg_gen_andc_tl(t0, cpu_gpr[rS(ctx->opcode)], t0);
-    tcg_gen_ext32u_tl(t0, t0);
-    t1 = tcg_temp_new();
-    tcg_gen_andi_tl(t1, cpu_gpr[rB(ctx->opcode)], 0x1f);
-    tcg_gen_shr_tl(cpu_gpr[rA(ctx->opcode)], t0, t1);
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, cpu_gpr[rA(ctx->opcode)]);
-    }
-}
-
 #if defined(TARGET_PPC64)
 /* sld & sld. */
 static void gen_sld(DisasContext *ctx)
@@ -5777,10 +5730,8 @@ GEN_HANDLER_E(paste, 0x1F, 0x06, 0x1C, 0x03C00000, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER(rlwimi, 0x14, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(rlwinm, 0x15, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(rlwnm, 0x17, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(slw, 0x1F, 0x18, 0x00, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(sraw, 0x1F, 0x18, 0x18, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(srawi, 0x1F, 0x18, 0x19, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(srw, 0x1F, 0x18, 0x10, 0x00000000, PPC_INTEGER),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(sld, 0x1F, 0x1B, 0x00, 0x00000000, PPC_64B),
 GEN_HANDLER(srad, 0x1F, 0x1A, 0x18, 0x00000000, PPC_64B),
target/ppc/translate/fixedpoint-impl.c.inc +48 -0
index 91f9c6c391..9d04b96316 100644
@@ -1406,3 +1406,51 @@ TRANS(HASHST, do_hash, false, gen_helper_HASHST)
 TRANS(HASHCHK, do_hash, false, gen_helper_HASHCHK)
 TRANS(HASHSTP, do_hash, true, gen_helper_HASHSTP)
 TRANS(HASHCHKP, do_hash, true, gen_helper_HASHCHKP)
+
+static bool trans_SLW(DisasContext *ctx, arg_SLW *a)
+{
+    TCGv t0, t1;
+
+    t0 = tcg_temp_new();
+    /* AND rt with a mask that is 0 when rB >= 0x20 */
+#if defined(TARGET_PPC64)
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x3a);
+    tcg_gen_sari_tl(t0, t0, 0x3f);
+#else
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x1a);
+    tcg_gen_sari_tl(t0, t0, 0x1f);
+#endif
+    tcg_gen_andc_tl(t0, cpu_gpr[a->rt], t0);
+    t1 = tcg_temp_new();
+    tcg_gen_andi_tl(t1, cpu_gpr[a->rb], 0x1f);
+    tcg_gen_shl_tl(cpu_gpr[a->ra], t0, t1);
+    tcg_gen_ext32u_tl(cpu_gpr[a->ra], cpu_gpr[a->ra]);
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, cpu_gpr[a->ra]);
+    }
+    return true;
+}
+
+static bool trans_SRW(DisasContext *ctx, arg_SRW *a)
+{
+    TCGv t0, t1;
+
+    t0 = tcg_temp_new();
+    /* AND rt with a mask that is 0 when rB >= 0x20 */
+#if defined(TARGET_PPC64)
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x3a);
+    tcg_gen_sari_tl(t0, t0, 0x3f);
+#else
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x1a);
+    tcg_gen_sari_tl(t0, t0, 0x1f);
+#endif
+    tcg_gen_andc_tl(t0, cpu_gpr[a->rt], t0);
+    tcg_gen_ext32u_tl(t0, t0);
+    t1 = tcg_temp_new();
+    tcg_gen_andi_tl(t1, cpu_gpr[a->rb], 0x1f);
+    tcg_gen_shr_tl(cpu_gpr[a->ra], t0, t1);
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, cpu_gpr[a->ra]);
+    }
+    return true;
+}
Signature · 1 line
2.55.0
#21 lore raw

Re: [PATCH v2 07/37] target/ppc: convert slw, srw instruction via decode spec

Quoted 7 lines · Amit Machhiwal
Quoted 6 lines · Chinmay Rath

From: "Mukesh Kumar Chaurasiya (IBM)" <[email protected]>

Instructions converted:

- slw
- srw

Also, the commit log could be improved:

Move slw and srw to decodetree specification using the @X_rc format. The existing shift-masking logic from gen_slw/gen_srw is preserved verbatim in the new trans_SLW/trans_SRW handlers.

Quoted 167 lines
Quoted 15 lines
---
 target/ppc/insn32.decode                   |  6 +++
 target/ppc/translate.c                     | 49 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 48 +++++++++++++++++++++
 3 files changed, 54 insertions(+), 49 deletions(-)
target/ppc/insn32.decode +0 -0
index 701c66b347..66d2f455a9 100644

target/ppc/insn32.decode HASHCHK 011111 ..... ..... ..... 1011110010 . @X_DW · line 513

Quoted 9 lines
target/ppc/insn32.decode +6 -0
@@ -505,6 +505,12 @@ HASHCHK         011111 ..... ..... ..... 1011110010 .   @X_DW
 HASHSTP         011111 ..... ..... ..... 1010010010 .   @X_DW
 HASHCHKP        011111 ..... ..... ..... 1010110010 .   @X_DW
 
+# Fixed-Point Shift Instructions
+
+SLW		011111 ..... ..... ..... 0000011000 .	@X_rc
+SRW		011111 ..... ..... ..... 1000011000 .	@X_rc
+
+

Nit: this looks like an extra blank line - can be removed. Otherwise,

Quoted 138 lines
target/ppc/insn32.decode +0 -0
 ## BCD Assist
 
 ADDG6S          011111 ..... ..... ..... - 001001010 -  @X
target/ppc/translate.c +0 -49
index ee14fbfd96..f2b7505634 100644
@@ -2315,29 +2315,6 @@ GEN_PPC64_R4(rldimi, 0x1E, 0x06);
 
 /***                             Integer shift                             ***/
 
-/* slw & slw. */
-static void gen_slw(DisasContext *ctx)
-{
-    TCGv t0, t1;
-
-    t0 = tcg_temp_new();
-    /* AND rS with a mask that is 0 when rB >= 0x20 */
-#if defined(TARGET_PPC64)
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x3a);
-    tcg_gen_sari_tl(t0, t0, 0x3f);
-#else
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x1a);
-    tcg_gen_sari_tl(t0, t0, 0x1f);
-#endif
-    tcg_gen_andc_tl(t0, cpu_gpr[rS(ctx->opcode)], t0);
-    t1 = tcg_temp_new();
-    tcg_gen_andi_tl(t1, cpu_gpr[rB(ctx->opcode)], 0x1f);
-    tcg_gen_shl_tl(cpu_gpr[rA(ctx->opcode)], t0, t1);
-    tcg_gen_ext32u_tl(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rA(ctx->opcode)]);
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, cpu_gpr[rA(ctx->opcode)]);
-    }
-}
 
 /* sraw & sraw. */
 static void gen_sraw(DisasContext *ctx)
@@ -2379,30 +2356,6 @@ static void gen_srawi(DisasContext *ctx)
     }
 }
 
-/* srw & srw. */
-static void gen_srw(DisasContext *ctx)
-{
-    TCGv t0, t1;
-
-    t0 = tcg_temp_new();
-    /* AND rS with a mask that is 0 when rB >= 0x20 */
-#if defined(TARGET_PPC64)
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x3a);
-    tcg_gen_sari_tl(t0, t0, 0x3f);
-#else
-    tcg_gen_shli_tl(t0, cpu_gpr[rB(ctx->opcode)], 0x1a);
-    tcg_gen_sari_tl(t0, t0, 0x1f);
-#endif
-    tcg_gen_andc_tl(t0, cpu_gpr[rS(ctx->opcode)], t0);
-    tcg_gen_ext32u_tl(t0, t0);
-    t1 = tcg_temp_new();
-    tcg_gen_andi_tl(t1, cpu_gpr[rB(ctx->opcode)], 0x1f);
-    tcg_gen_shr_tl(cpu_gpr[rA(ctx->opcode)], t0, t1);
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, cpu_gpr[rA(ctx->opcode)]);
-    }
-}
-
 #if defined(TARGET_PPC64)
 /* sld & sld. */
 static void gen_sld(DisasContext *ctx)
@@ -5777,10 +5730,8 @@ GEN_HANDLER_E(paste, 0x1F, 0x06, 0x1C, 0x03C00000, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER(rlwimi, 0x14, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(rlwinm, 0x15, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(rlwnm, 0x17, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(slw, 0x1F, 0x18, 0x00, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(sraw, 0x1F, 0x18, 0x18, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(srawi, 0x1F, 0x18, 0x19, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(srw, 0x1F, 0x18, 0x10, 0x00000000, PPC_INTEGER),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(sld, 0x1F, 0x1B, 0x00, 0x00000000, PPC_64B),
 GEN_HANDLER(srad, 0x1F, 0x1A, 0x18, 0x00000000, PPC_64B),
target/ppc/translate/fixedpoint-impl.c.inc +48 -0
index 91f9c6c391..9d04b96316 100644
@@ -1406,3 +1406,51 @@ TRANS(HASHST, do_hash, false, gen_helper_HASHST)
 TRANS(HASHCHK, do_hash, false, gen_helper_HASHCHK)
 TRANS(HASHSTP, do_hash, true, gen_helper_HASHSTP)
 TRANS(HASHCHKP, do_hash, true, gen_helper_HASHCHKP)
+
+static bool trans_SLW(DisasContext *ctx, arg_SLW *a)
+{
+    TCGv t0, t1;
+
+    t0 = tcg_temp_new();
+    /* AND rt with a mask that is 0 when rB >= 0x20 */
+#if defined(TARGET_PPC64)
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x3a);
+    tcg_gen_sari_tl(t0, t0, 0x3f);
+#else
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x1a);
+    tcg_gen_sari_tl(t0, t0, 0x1f);
+#endif
+    tcg_gen_andc_tl(t0, cpu_gpr[a->rt], t0);
+    t1 = tcg_temp_new();
+    tcg_gen_andi_tl(t1, cpu_gpr[a->rb], 0x1f);
+    tcg_gen_shl_tl(cpu_gpr[a->ra], t0, t1);
+    tcg_gen_ext32u_tl(cpu_gpr[a->ra], cpu_gpr[a->ra]);
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, cpu_gpr[a->ra]);
+    }
+    return true;
+}
+
+static bool trans_SRW(DisasContext *ctx, arg_SRW *a)
+{
+    TCGv t0, t1;
+
+    t0 = tcg_temp_new();
+    /* AND rt with a mask that is 0 when rB >= 0x20 */
+#if defined(TARGET_PPC64)
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x3a);
+    tcg_gen_sari_tl(t0, t0, 0x3f);
+#else
+    tcg_gen_shli_tl(t0, cpu_gpr[a->rb], 0x1a);
+    tcg_gen_sari_tl(t0, t0, 0x1f);
+#endif
+    tcg_gen_andc_tl(t0, cpu_gpr[a->rt], t0);
+    tcg_gen_ext32u_tl(t0, t0);
+    t1 = tcg_temp_new();
+    tcg_gen_andi_tl(t1, cpu_gpr[a->rb], 0x1f);
+    tcg_gen_shr_tl(cpu_gpr[a->ra], t0, t1);
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, cpu_gpr[a->ra]);
+    }
+    return true;
+}
Signature · 1 line
2.55.0
#22 lore raw

PATCH v2 8/37 target/ppc: convert sraw[i] instruction via decode spec

In reply to Chinmay Rath

From: "Mukesh Kumar Chaurasiya (IBM)" <[email protected]>

Instruction converted:

- sraw
- srawi

---
 target/ppc/helper.h                        |  2 +-
 target/ppc/insn32.decode                   |  6 ++--
 target/ppc/int_helper.c                    |  2 +-
 target/ppc/translate.c                     | 42 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 41 +++++++++++++++++++++
 5 files changed, 47 insertions(+), 46 deletions(-)
target/ppc/helper.h +1 -1
index 6df3de6516..3aff559504 100644
@@ -64,7 +64,7 @@ DEF_HELPER_4(DIVWE, tl, env, tl, tl, i32)
 
 DEF_HELPER_FLAGS_1(POPCNTB, TCG_CALL_NO_RWG_SE, tl, tl)
 DEF_HELPER_FLAGS_2(CMPB, TCG_CALL_NO_RWG_SE, tl, tl, tl)
-DEF_HELPER_3(sraw, tl, env, tl, tl)
+DEF_HELPER_3(SRAW, tl, env, tl, tl)
 DEF_HELPER_FLAGS_2(CFUGED, TCG_CALL_NO_RWG_SE, i64, i64, i64)
 DEF_HELPER_FLAGS_2(PDEPD, TCG_CALL_NO_RWG_SE, i64, i64, i64)
 DEF_HELPER_FLAGS_2(PEXTD, TCG_CALL_NO_RWG_SE, i64, i64, i64)
target/ppc/insn32.decode +4 -2
index 66d2f455a9..af35a44223 100644
@@ -507,8 +507,10 @@ HASHCHKP        011111 ..... ..... ..... 1010110010 .   @X_DW
 
 # Fixed-Point Shift Instructions
 
-SLW		011111 ..... ..... ..... 0000011000 .	@X_rc
-SRW		011111 ..... ..... ..... 1000011000 .	@X_rc
+SLW             011111 ..... ..... ..... 0000011000 .   @X_rc
+SRAW            011111 ..... ..... ..... 1100011000 .   @X_rc
+SRAWI           011111 ..... ..... ..... 1100111000 .   @X_rc
+SRW             011111 ..... ..... ..... 1000011000 .   @X_rc
 
 
 ## BCD Assist
target/ppc/int_helper.c +1 -1
index dc55ccadf7..b161f930ec 100644
@@ -235,7 +235,7 @@ target_ulong helper_CMPB(target_ulong rs, target_ulong rb)
 }
 
 /* shift right arithmetic helper */
-target_ulong helper_sraw(CPUPPCState *env, target_ulong value,
+target_ulong helper_SRAW(CPUPPCState *env, target_ulong value,
                          target_ulong shift)
 {
     int32_t ret;
target/ppc/translate.c +0 -42

54 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +41 -0

44 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 8/37
#23 lore raw

Re: [PATCH v2 08/37] target/ppc: convert sraw[i] instruction via decode spec

Quoted 6 lines · Chinmay Rath

From: "Mukesh Kumar Chaurasiya (IBM)" <[email protected]>

Instruction converted:

- sraw
- srawi

This doesn't give the idea on what is being done - could have been more explicit that the insns are being moved to decode tree spec. But its upto you.

Quoted 31 lines
---
 target/ppc/helper.h                        |  2 +-
 target/ppc/insn32.decode                   |  6 ++--
 target/ppc/int_helper.c                    |  2 +-
 target/ppc/translate.c                     | 42 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 41 +++++++++++++++++++++
 5 files changed, 47 insertions(+), 46 deletions(-)
target/ppc/helper.h +1 -1
index 6df3de6516..3aff559504 100644
@@ -64,7 +64,7 @@ DEF_HELPER_4(DIVWE, tl, env, tl, tl, i32)
 
 DEF_HELPER_FLAGS_1(POPCNTB, TCG_CALL_NO_RWG_SE, tl, tl)
 DEF_HELPER_FLAGS_2(CMPB, TCG_CALL_NO_RWG_SE, tl, tl, tl)
-DEF_HELPER_3(sraw, tl, env, tl, tl)
+DEF_HELPER_3(SRAW, tl, env, tl, tl)
 DEF_HELPER_FLAGS_2(CFUGED, TCG_CALL_NO_RWG_SE, i64, i64, i64)
 DEF_HELPER_FLAGS_2(PDEPD, TCG_CALL_NO_RWG_SE, i64, i64, i64)
 DEF_HELPER_FLAGS_2(PEXTD, TCG_CALL_NO_RWG_SE, i64, i64, i64)
target/ppc/insn32.decode +0 -0
index 66d2f455a9..af35a44223 100644
@@ -507,8 +507,10 @@ HASHCHKP        011111 ..... ..... ..... 1010110010 .   @X_DW
 

target/ppc/insn32.decode HASHCHKP 011111 ..... ..... ..... 1010110010 . @X_DW · line 515

Quoted 10 lines
target/ppc/insn32.decode +4 -2
 # Fixed-Point Shift Instructions
 
-SLW		011111 ..... ..... ..... 0000011000 .	@X_rc
-SRW		011111 ..... ..... ..... 1000011000 .	@X_rc
+SLW             011111 ..... ..... ..... 0000011000 .   @X_rc
+SRAW            011111 ..... ..... ..... 1100011000 .   @X_rc
+SRAWI           011111 ..... ..... ..... 1100111000 .   @X_rc
+SRW             011111 ..... ..... ..... 1000011000 .   @X_rc
 
 

Nit: extra blank line can be removed. Otherwise, the changes look good to me.

Quoted 126 lines
target/ppc/insn32.decode +0 -0
 ## BCD Assist
target/ppc/int_helper.c +1 -1
index dc55ccadf7..b161f930ec 100644
@@ -235,7 +235,7 @@ target_ulong helper_CMPB(target_ulong rs, target_ulong rb)
 }
 
 /* shift right arithmetic helper */
-target_ulong helper_sraw(CPUPPCState *env, target_ulong value,
+target_ulong helper_SRAW(CPUPPCState *env, target_ulong value,
                          target_ulong shift)
 {
     int32_t ret;
target/ppc/translate.c +0 -42
index f2b7505634..f00071aa91 100644
@@ -2316,46 +2316,6 @@ GEN_PPC64_R4(rldimi, 0x1E, 0x06);
 /***                             Integer shift                             ***/
 
 
-/* sraw & sraw. */
-static void gen_sraw(DisasContext *ctx)
-{
-    gen_helper_sraw(cpu_gpr[rA(ctx->opcode)], tcg_env,
-                    cpu_gpr[rS(ctx->opcode)], cpu_gpr[rB(ctx->opcode)]);
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, cpu_gpr[rA(ctx->opcode)]);
-    }
-}
-
-/* srawi & srawi. */
-static void gen_srawi(DisasContext *ctx)
-{
-    int sh = SH(ctx->opcode);
-    TCGv dst = cpu_gpr[rA(ctx->opcode)];
-    TCGv src = cpu_gpr[rS(ctx->opcode)];
-    if (sh == 0) {
-        tcg_gen_ext32s_tl(dst, src);
-        tcg_gen_movi_tl(cpu_ca, 0);
-        if (is_isa300(ctx)) {
-            tcg_gen_movi_tl(cpu_ca32, 0);
-        }
-    } else {
-        TCGv t0;
-        tcg_gen_ext32s_tl(dst, src);
-        tcg_gen_andi_tl(cpu_ca, dst, (1ULL << sh) - 1);
-        t0 = tcg_temp_new();
-        tcg_gen_sari_tl(t0, dst, TARGET_LONG_BITS - 1);
-        tcg_gen_and_tl(cpu_ca, cpu_ca, t0);
-        tcg_gen_setcondi_tl(TCG_COND_NE, cpu_ca, cpu_ca, 0);
-        if (is_isa300(ctx)) {
-            tcg_gen_mov_tl(cpu_ca32, cpu_ca);
-        }
-        tcg_gen_sari_tl(dst, dst, sh);
-    }
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, dst);
-    }
-}
-
 #if defined(TARGET_PPC64)
 /* sld & sld. */
 static void gen_sld(DisasContext *ctx)
@@ -5730,8 +5690,6 @@ GEN_HANDLER_E(paste, 0x1F, 0x06, 0x1C, 0x03C00000, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER(rlwimi, 0x14, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(rlwinm, 0x15, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(rlwnm, 0x17, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(sraw, 0x1F, 0x18, 0x18, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(srawi, 0x1F, 0x18, 0x19, 0x00000000, PPC_INTEGER),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(sld, 0x1F, 0x1B, 0x00, 0x00000000, PPC_64B),
 GEN_HANDLER(srad, 0x1F, 0x1A, 0x18, 0x00000000, PPC_64B),
target/ppc/translate/fixedpoint-impl.c.inc +41 -0
index 9d04b96316..234404e459 100644
@@ -1454,3 +1454,44 @@ static bool trans_SRW(DisasContext *ctx, arg_SRW *a)
     }
     return true;
 }
+
+/* sraw & sraw. */
+static bool trans_SRAW(DisasContext *ctx, arg_SRAW *a)
+{
+    gen_helper_SRAW(cpu_gpr[a->ra], tcg_env,
+                    cpu_gpr[a->rt], cpu_gpr[a->rb]);
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, cpu_gpr[a->ra]);
+    }
+    return true;
+}
+
+/* srawi & srawi. */
+static bool trans_SRAWI(DisasContext *ctx, arg_SRAWI *a)
+{
+    TCGv dst = cpu_gpr[a->ra];
+    TCGv src = cpu_gpr[a->rt];
+    if (a->rb == 0) {
+        tcg_gen_ext32s_tl(dst, src);
+        tcg_gen_movi_tl(cpu_ca, 0);
+        if (is_isa300(ctx)) {
+            tcg_gen_movi_tl(cpu_ca32, 0);
+        }
+    } else {
+        TCGv t0;
+        tcg_gen_ext32s_tl(dst, src);
+        tcg_gen_andi_tl(cpu_ca, dst, (1ULL << a->rb) - 1);
+        t0 = tcg_temp_new();
+        tcg_gen_sari_tl(t0, dst, TARGET_LONG_BITS - 1);
+        tcg_gen_and_tl(cpu_ca, cpu_ca, t0);
+        tcg_gen_setcondi_tl(TCG_COND_NE, cpu_ca, cpu_ca, 0);
+        if (is_isa300(ctx)) {
+            tcg_gen_mov_tl(cpu_ca32, cpu_ca);
+        }
+        tcg_gen_sari_tl(dst, dst, a->rb);
+    }
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, dst);
+    }
+    return true;
+}
Signature · 1 line
2.55.0
#24 lore raw

PATCH v2 9/37 target/ppc: Convert mcrf to decode tree

In reply to Chinmay Rath

From: Shivani Nittor <[email protected]>

Convert mcrf to decode tree specification.The functionality was tested by comparing the qemu -D log -d op,in_asm output as well as single stepping gdb.

---
 target/ppc/insn32.decode             | 6 ++++++
 target/ppc/translate.c               | 7 -------
 target/ppc/translate/misc-impl.c.inc | 6 ++++++
 3 files changed, 12 insertions(+), 7 deletions(-)
target/ppc/insn32.decode +6 -0
index af35a44223..7d792d2318 100644
@@ -1455,6 +1455,12 @@ MSGSYNC         011111 ----- ----- ----- 1101110110 -
 SYNC            011111 -- ... --- ..   ----- 1001010110 -           @X_sync
 EIEIO           011111 ----- ----- ----- 1101010110 -
 
+# Condition Register Field Instruction
+
+&XL_bfa         bf bfa
+@XL_bfa         ...... bf:3 .. bfa:3 .. ..... ..... ..... .
+MCRF            010011 ... -- ... -- ----- 00000 00000 -       @XL_bfa
+
 # Branch History Rolling Buffer (BHRB) Instructions
 
 &XFX_bhrbe      rt bhrbe
target/ppc/translate.c +0 -7
index f00071aa91..55b515cec7 100644
@@ -3642,12 +3642,6 @@ GEN_CRLOGIC(crorc, tcg_gen_orc_i32, 0x0D);
 /* crxor */
 GEN_CRLOGIC(crxor, tcg_gen_xor_i32, 0x06);
 
-/* mcrf */
-static void gen_mcrf(DisasContext *ctx)
-{
-    tcg_gen_mov_i32(cpu_crf[crfD(ctx->opcode)], cpu_crf[crfS(ctx->opcode)]);
-}
-
 /***                           System linkage                              ***/
 
 /* rfi (supervisor only) */
@@ -5722,7 +5716,6 @@ GEN_HANDLER(bc, 0x10, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bcctr, 0x13, 0x10, 0x10, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bclr, 0x13, 0x10, 0x00, 0x00000000, PPC_FLOW),
 GEN_HANDLER_E(bctar, 0x13, 0x10, 0x11, 0x0000E000, PPC_NONE, PPC2_BCTAR_ISA207),
-GEN_HANDLER(mcrf, 0x13, 0x00, 0xFF, 0x00000001, PPC_INTEGER),
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(rfid, 0x13, 0x12, 0x00, 0x03FF8001, PPC_64B),
target/ppc/translate/misc-impl.c.inc +6 -0
index cbf82b1ea0..890fbb209e 100644
@@ -155,3 +155,9 @@ static bool trans_ATTN(DisasContext *ctx, arg_ATTN *a)
     return false;
 #endif
 }
+
+static bool trans_MCRF(DisasContext *ctx, arg_MCRF *a)
+{
+    tcg_gen_mov_i32(cpu_crf[a->bf], cpu_crf[a->bfa]);
+    return true;
+}
Signature · 1 line
2.55.0
Reviewed-by on v2 9/37
#25 lore raw

Re: [PATCH v2 09/37] target/ppc: Convert mcrf to decode tree

Quoted 3 lines · Chinmay Rath

From: Shivani Nittor <[email protected]>

Convert mcrf to decode tree specification.The functionality

minor: s/The/ The/

Quoted 11 lines

was tested by comparing the qemu -D log -d op,in_asm output as well as single stepping gdb.

---
 target/ppc/insn32.decode             | 6 ++++++
 target/ppc/translate.c               | 7 -------
 target/ppc/translate/misc-impl.c.inc | 6 ++++++
 3 files changed, 12 insertions(+), 7 deletions(-)
Quoted 60 lines
target/ppc/insn32.decode +6 -0
index af35a44223..7d792d2318 100644
@@ -1455,6 +1455,12 @@ MSGSYNC         011111 ----- ----- ----- 1101110110 -
 SYNC            011111 -- ... --- ..   ----- 1001010110 -           @X_sync
 EIEIO           011111 ----- ----- ----- 1101010110 -
 
+# Condition Register Field Instruction
+
+&XL_bfa         bf bfa
+@XL_bfa         ...... bf:3 .. bfa:3 .. ..... ..... ..... .
+MCRF            010011 ... -- ... -- ----- 00000 00000 -       @XL_bfa
+
 # Branch History Rolling Buffer (BHRB) Instructions
 
 &XFX_bhrbe      rt bhrbe
target/ppc/translate.c +0 -7
index f00071aa91..55b515cec7 100644
@@ -3642,12 +3642,6 @@ GEN_CRLOGIC(crorc, tcg_gen_orc_i32, 0x0D);
 /* crxor */
 GEN_CRLOGIC(crxor, tcg_gen_xor_i32, 0x06);
 
-/* mcrf */
-static void gen_mcrf(DisasContext *ctx)
-{
-    tcg_gen_mov_i32(cpu_crf[crfD(ctx->opcode)], cpu_crf[crfS(ctx->opcode)]);
-}
-
 /***                           System linkage                              ***/
 
 /* rfi (supervisor only) */
@@ -5722,7 +5716,6 @@ GEN_HANDLER(bc, 0x10, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bcctr, 0x13, 0x10, 0x10, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bclr, 0x13, 0x10, 0x00, 0x00000000, PPC_FLOW),
 GEN_HANDLER_E(bctar, 0x13, 0x10, 0x11, 0x0000E000, PPC_NONE, PPC2_BCTAR_ISA207),
-GEN_HANDLER(mcrf, 0x13, 0x00, 0xFF, 0x00000001, PPC_INTEGER),
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(rfid, 0x13, 0x12, 0x00, 0x03FF8001, PPC_64B),
target/ppc/translate/misc-impl.c.inc +6 -0
index cbf82b1ea0..890fbb209e 100644
@@ -155,3 +155,9 @@ static bool trans_ATTN(DisasContext *ctx, arg_ATTN *a)
     return false;
 #endif
 }
+
+static bool trans_MCRF(DisasContext *ctx, arg_MCRF *a)
+{
+    tcg_gen_mov_i32(cpu_crf[a->bf], cpu_crf[a->bfa]);
+    return true;
+}
Signature · 1 line
2.55.0
#26 lore raw

PATCH v2 10/37 target/ppc: Move fixed-point Shift insns to decodetree

In reply to Chinmay Rath

From: Amit Machhiwal <[email protected]>

Move the below instructions to decodetree specification:

s{r,l}d, and srad[i] : X-form

The changes were verified by validating that the TCG ops generated by these instructions remain the same after these changes. The TCG micro-ops were captured with the '-d op,in_asm' QEMU flag.

Move all shift insntructions to one place in insn32.decode

---
 target/ppc/helper.h                        |  2 +-
 target/ppc/insn32.decode                   | 11 ++-
 target/ppc/int_helper.c                    |  2 +-
 target/ppc/translate.c                     | 96 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 72 ++++++++++++++++
 5 files changed, 81 insertions(+), 102 deletions(-)
target/ppc/helper.h +1 -1
index 3aff559504..cd2a0948a3 100644
@@ -74,7 +74,7 @@ DEF_HELPER_FLAGS_1(CBCDTD, TCG_CALL_NO_RWG_SE, tl, tl)
 DEF_HELPER_FLAGS_2(CMPEQB, TCG_CALL_NO_RWG_SE, i32, tl, tl)
 DEF_HELPER_FLAGS_1(POPCNTW, TCG_CALL_NO_RWG_SE, tl, tl)
 DEF_HELPER_FLAGS_2(BPERMD, TCG_CALL_NO_RWG_SE, i64, i64, i64)
-DEF_HELPER_3(srad, tl, env, tl, tl)
+DEF_HELPER_3(SRAD, tl, env, tl, tl)
 DEF_HELPER_FLAGS_0(DARN32, TCG_CALL_NO_RWG, tl)
 DEF_HELPER_FLAGS_0(DARN64, TCG_CALL_NO_RWG, tl)
 #endif
target/ppc/insn32.decode +7 -4
index 7d792d2318..d34094d717 100644
@@ -119,7 +119,7 @@
 
 # XS-form with fractured shift amount
 %xs_sh           1:1 11:5
-&XS              rs ra sh rc
+&XS              rs ra sh rc:bool
 @XS              ...... rs:5 ra:5 ..... ......... . rc:1        &XS sh=%xs_sh
 
 %x_frtp         22:4 !function=times_2
@@ -512,6 +512,12 @@ SRAW            011111 ..... ..... ..... 1100011000 .   @X_rc
 SRAWI           011111 ..... ..... ..... 1100111000 .   @X_rc
 SRW             011111 ..... ..... ..... 1000011000 .   @X_rc
 
+SLD             011111 ..... ..... ..... 0000011011 .   @X_rc
+SRD             011111 ..... ..... ..... 1000011011 .   @X_rc
+SRAD            011111 ..... ..... ..... 1100011010 .   @X_rc
+SRADI           011111 ..... ..... ..... 110011101 . .  @XS
+
+EXTSWSLI        011111 ..... ..... ..... 110111101 . .  @XS
 
 ## BCD Assist
 
@@ -1382,9 +1388,6 @@ DST             011111 ----- ..... ..... 0101010110 -   @X_ea
 DSTST           011111 ----- ..... ..... 0101110110 -   @X_ea
 DSS             011111 ----- ..... ..... 1100110110 -   @X_ea
 
-##Extend Sign Word and Shift Left Immediate XS-form
-EXTSWSLI         011111 ..... ..... ..... 110111101 . .         @XS
-
 ## Load and Reserve Instructions
 LBARX           011111 ..... ..... ..... 0000110100 .   @X_rc
 LHARX           011111 ..... ..... ..... 0001110100 .   @X_rc
target/ppc/int_helper.c +1 -1
index b161f930ec..8e8076a92c 100644
@@ -261,7 +261,7 @@ target_ulong helper_SRAW(CPUPPCState *env, target_ulong value,
 }
 
 #if defined(TARGET_PPC64)
-target_ulong helper_srad(CPUPPCState *env, target_ulong value,
+target_ulong helper_SRAD(CPUPPCState *env, target_ulong value,
                          target_ulong shift)
 {
     int64_t ret;
target/ppc/translate.c +0 -96

108 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +72 -0

84 lines not on this page · raw message

Signature · 1 line
2.55.0
#27 lore raw

PATCH v2 11/37 target/ppc: Move fixed-point byte-reversal store insns to decodetree

In reply to Chinmay Rath

From: Amit Machhiwal <[email protected]>

Move the below PPC instructions to decodetree specification:

st{d,h,w}brx, st{d,w,h,b}cix : X-form

The changes were verified by validating that the TCG micro-ops generated by those insns remain the same. These TCG micro-ops were captured with the -d op,in_asm QEMU flag.

Additionally, remove GEN_STX and GEN_STX_E macros as they are no longer used anywhere else.

---
 target/ppc/insn32.decode                   |  8 +++
 target/ppc/translate.c                     | 52 +++-----------
 target/ppc/translate/fixedpoint-impl.c.inc | 79 ++++++++++++++++++++++
 3 files changed, 95 insertions(+), 44 deletions(-)
target/ppc/insn32.decode +8 -0
index d34094d717..8425c70b9d 100644
@@ -372,6 +372,14 @@ STVDX           011111 ..... ..... ..... 0010010001 -   @X
 STSKU           111110 ..... ..... ............. 0 11   @DD
 STCXU           111110 ..... ..... ............. 1 11   @DD
 
+STDBRX          011111 ..... ..... ..... 1010010100 -   @X
+STDCIX          011111 ..... ..... ..... 1111110101 -   @X
+STWCIX          011111 ..... ..... ..... 1110010101 -   @X
+STHCIX          011111 ..... ..... ..... 1110110101 -   @X
+STBCIX          011111 ..... ..... ..... 1111010101 -   @X
+STHBRX          011111 ..... ..... ..... 1110010110 -   @X
+STWBRX          011111 ..... ..... ..... 1010010110 -   @X
+
 ### Fixed-Point Compare Instructions
 
 CMP             011111 ... - . ..... ..... 0000000000 - @X_bfl
target/ppc/translate.c +8 -44

87 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +79 -0

85 lines not on this page · raw message

Signature · 1 line
2.55.0
#28 lore raw

PATCH v2 12/37 target/ppc: Move GPR atomic load/store instructions to decodetree

In reply to Chinmay Rath

From: Aboorva Devarajan <[email protected]>

Move below instructions to decodetree specification:

lwat, stwat : X-form (word atomic, ISA 3.0)
ldat, stdat : X-form (doubleword atomic, ISA 3.0)

The legacy gen_ld_atomic()/gen_st_atomic() functions that read register fields and the function code from ctx->opcode are replaced by new do_ld_atomic()/do_st_atomic() helpers in fixedpoint-impl.c.inc that use the pre-extracted decodetree argument fields (a->rt, a->ra, a->rb) directly. The RB field carries the Function Code (FC) selecting the atomic operation.

gen_fetch_inc_conditional() is similarly moved and refactored to accept the destination register number as an explicit parameter instead of extracting it from ctx->opcode.

Checked with '-d in_asm,op'. The printed TCG ops may differ from the pre-refactor log because EA now uses do_ea_calc(ctx, ra, 0), which lowers as GPR[ra] + 0 (then narrow in 32-bit mode) instead of mov/ext32u from ra alone.

---
 target/ppc/insn32.decode                   |   7 +
 target/ppc/translate.c                     | 212 --------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 215 +++++++++++++++++++++
 3 files changed, 222 insertions(+), 212 deletions(-)
target/ppc/insn32.decode +7 -0
index 8425c70b9d..a1c2f549bd 100644
@@ -343,6 +343,13 @@ LSKU            111010 ..... ..... ............. 0 11   @DD
 LCXU            111010 ..... ..... ............. 1 11   @DD
 
 
+### Fixed-Point Atomic Load/Store Instructions
+
+LWAT            011111 ..... ..... ..... 1001000110 -   @X
+STWAT           011111 ..... ..... ..... 1011000110 -   @X
+LDAT            011111 ..... ..... ..... 1001100110 -   @X
+STDAT           011111 ..... ..... ..... 1011100110 -   @X
+
 ### Fixed-Point Store Instructions
 
 STB             100110 ..... ..... ................     @D
target/ppc/translate.c +0 -212

224 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +215 -0

218 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 12/37
#29 lore raw

Re: [PATCH v2 12/37] target/ppc: Move GPR atomic load/store instructions to decodetree

Quoted 396 lines · Chinmay Rath

From: Aboorva Devarajan <[email protected]>

Move below instructions to decodetree specification:

lwat, stwat : X-form (word atomic, ISA 3.0)
ldat, stdat : X-form (doubleword atomic, ISA 3.0)

The legacy gen_ld_atomic()/gen_st_atomic() functions that read register fields and the function code from ctx->opcode are replaced by new do_ld_atomic()/do_st_atomic() helpers in fixedpoint-impl.c.inc that use the pre-extracted decodetree argument fields (a->rt, a->ra, a->rb) directly. The RB field carries the Function Code (FC) selecting the atomic operation.

gen_fetch_inc_conditional() is similarly moved and refactored to accept the destination register number as an explicit parameter instead of extracting it from ctx->opcode.

Checked with '-d in_asm,op'. The printed TCG ops may differ from the pre-refactor log because EA now uses do_ea_calc(ctx, ra, 0), which lowers as GPR[ra] + 0 (then narrow in 32-bit mode) instead of mov/ext32u from ra alone.

---
 target/ppc/insn32.decode                   |   7 +
 target/ppc/translate.c                     | 212 --------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 215 +++++++++++++++++++++
 3 files changed, 222 insertions(+), 212 deletions(-)
target/ppc/insn32.decode +7 -0
index 8425c70b9d..a1c2f549bd 100644
@@ -343,6 +343,13 @@ LSKU            111010 ..... ..... ............. 0 11   @DD
 LCXU            111010 ..... ..... ............. 1 11   @DD
 
 
+### Fixed-Point Atomic Load/Store Instructions
+
+LWAT            011111 ..... ..... ..... 1001000110 -   @X
+STWAT           011111 ..... ..... ..... 1011000110 -   @X
+LDAT            011111 ..... ..... ..... 1001100110 -   @X
+STDAT           011111 ..... ..... ..... 1011100110 -   @X
+
 ### Fixed-Point Store Instructions
 
 STB             100110 ..... ..... ................     @D
target/ppc/translate.c +0 -212
index 548a7c0437..179cb69dc5 100644
@@ -2741,212 +2741,6 @@ static void gen_isync(DisasContext *ctx)
     ctx->base.is_jmp = DISAS_EXIT_UPDATE;
 }
 
-static void gen_fetch_inc_conditional(DisasContext *ctx, MemOp memop,
-                                      TCGv EA, TCGCond cond, int addend)
-{
-    TCGv t = tcg_temp_new();
-    TCGv t2 = tcg_temp_new();
-    TCGv u = tcg_temp_new();
-
-    tcg_gen_qemu_ld_tl(t, EA, ctx->mem_idx, memop);
-    tcg_gen_addi_tl(t2, EA, memop_size(memop));
-    tcg_gen_qemu_ld_tl(t2, t2, ctx->mem_idx, memop);
-    tcg_gen_addi_tl(u, t, addend);
-
-    /* E.g. for fetch and increment bounded... */
-    /* mem(EA,s) = (t != t2 ? u = t + 1 : t) */
-    tcg_gen_movcond_tl(cond, u, t, t2, u, t);
-    tcg_gen_qemu_st_tl(u, EA, ctx->mem_idx, memop);
-
-    /* RT = (t != t2 ? t : u = 1<<(s*8-1)) */
-    tcg_gen_movcond_tl(cond, cpu_gpr[rD(ctx->opcode)], t, t2, t,
-                       tcg_constant_tl(1 << (memop_size(memop) * 8 - 1)));
-}
-
-static void gen_ld_atomic(DisasContext *ctx, MemOp memop)
-{
-    uint32_t gpr_FC = FC(ctx->opcode);
-    TCGv EA = tcg_temp_new();
-    int rt = rD(ctx->opcode);
-    bool need_serial;
-    TCGv src, dst;
-
-    gen_addr_register(ctx, EA);
-    dst = cpu_gpr[rt];
-    src = cpu_gpr[(rt + 1) & 31];
-
-    need_serial = false;
-    memop |= MO_ALIGN;
-    switch (gpr_FC) {
-    case 0: /* Fetch and add */
-        tcg_gen_atomic_fetch_add_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 1: /* Fetch and xor */
-        tcg_gen_atomic_fetch_xor_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 2: /* Fetch and or */
-        tcg_gen_atomic_fetch_or_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 3: /* Fetch and 'and' */
-        tcg_gen_atomic_fetch_and_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 4:  /* Fetch and max unsigned */
-        tcg_gen_atomic_fetch_umax_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 5:  /* Fetch and max signed */
-        tcg_gen_atomic_fetch_smax_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 6:  /* Fetch and min unsigned */
-        tcg_gen_atomic_fetch_umin_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 7:  /* Fetch and min signed */
-        tcg_gen_atomic_fetch_smin_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-    case 8: /* Swap */
-        tcg_gen_atomic_xchg_tl(dst, EA, src, ctx->mem_idx, memop);
-        break;
-
-    case 16: /* Compare and swap not equal */
-        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
-            need_serial = true;
-        } else {
-            TCGv t0 = tcg_temp_new();
-            TCGv t1 = tcg_temp_new();
-
-            tcg_gen_qemu_ld_tl(t0, EA, ctx->mem_idx, memop);
-            if ((memop & MO_SIZE) == MO_64 || TARGET_LONG_BITS == 32) {
-                tcg_gen_mov_tl(t1, src);
-            } else {
-                tcg_gen_ext32u_tl(t1, src);
-            }
-            tcg_gen_movcond_tl(TCG_COND_NE, t1, t0, t1,
-                               cpu_gpr[(rt + 2) & 31], t0);
-            tcg_gen_qemu_st_tl(t1, EA, ctx->mem_idx, memop);
-            tcg_gen_mov_tl(dst, t0);
-        }
-        break;
-
-    case 24: /* Fetch and increment bounded */
-        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
-            need_serial = true;
-        } else {
-            gen_fetch_inc_conditional(ctx, memop, EA, TCG_COND_NE, 1);
-        }
-        break;
-    case 25: /* Fetch and increment equal */
-        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
-            need_serial = true;
-        } else {
-            gen_fetch_inc_conditional(ctx, memop, EA, TCG_COND_EQ, 1);
-        }
-        break;
-    case 28: /* Fetch and decrement bounded */
-        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
-            need_serial = true;
-        } else {
-            gen_fetch_inc_conditional(ctx, memop, EA, TCG_COND_NE, -1);
-        }
-        break;
-
-    default:
-        /* invoke data storage error handler */
-        gen_exception_err(ctx, POWERPC_EXCP_DSI, POWERPC_EXCP_INVAL);
-    }
-
-    if (need_serial) {
-        /* Restart with exclusive lock.  */
-        gen_helper_exit_atomic(tcg_env);
-        ctx->base.is_jmp = DISAS_NORETURN;
-    }
-}
-
-static void gen_lwat(DisasContext *ctx)
-{
-    gen_ld_atomic(ctx, DEF_MEMOP(MO_UL));
-}
-
-#ifdef TARGET_PPC64
-static void gen_ldat(DisasContext *ctx)
-{
-    gen_ld_atomic(ctx, DEF_MEMOP(MO_UQ));
-}
-#endif
-
-static void gen_st_atomic(DisasContext *ctx, MemOp memop)
-{
-    uint32_t gpr_FC = FC(ctx->opcode);
-    TCGv EA = tcg_temp_new();
-    TCGv src, discard;
-
-    gen_addr_register(ctx, EA);
-    src = cpu_gpr[rD(ctx->opcode)];
-    discard = tcg_temp_new();
-
-    memop |= MO_ALIGN;
-    switch (gpr_FC) {
-    case 0: /* add and Store */
-        tcg_gen_atomic_add_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 1: /* xor and Store */
-        tcg_gen_atomic_xor_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 2: /* Or and Store */
-        tcg_gen_atomic_or_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 3: /* 'and' and Store */
-        tcg_gen_atomic_and_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 4:  /* Store max unsigned */
-        tcg_gen_atomic_umax_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 5:  /* Store max signed */
-        tcg_gen_atomic_smax_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 6:  /* Store min unsigned */
-        tcg_gen_atomic_umin_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 7:  /* Store min signed */
-        tcg_gen_atomic_smin_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
-        break;
-    case 24: /* Store twin  */
-        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
-            /* Restart with exclusive lock.  */
-            gen_helper_exit_atomic(tcg_env);
-            ctx->base.is_jmp = DISAS_NORETURN;
-        } else {
-            TCGv t = tcg_temp_new();
-            TCGv t2 = tcg_temp_new();
-            TCGv s = tcg_temp_new();
-            TCGv s2 = tcg_temp_new();
-            TCGv ea_plus_s = tcg_temp_new();
-
-            tcg_gen_qemu_ld_tl(t, EA, ctx->mem_idx, memop);
-            tcg_gen_addi_tl(ea_plus_s, EA, memop_size(memop));
-            tcg_gen_qemu_ld_tl(t2, ea_plus_s, ctx->mem_idx, memop);
-            tcg_gen_movcond_tl(TCG_COND_EQ, s, t, t2, src, t);
-            tcg_gen_movcond_tl(TCG_COND_EQ, s2, t, t2, src, t2);
-            tcg_gen_qemu_st_tl(s, EA, ctx->mem_idx, memop);
-            tcg_gen_qemu_st_tl(s2, ea_plus_s, ctx->mem_idx, memop);
-        }
-        break;
-    default:
-        /* invoke data storage error handler */
-        gen_exception_err(ctx, POWERPC_EXCP_DSI, POWERPC_EXCP_INVAL);
-    }
-}
-
-static void gen_stwat(DisasContext *ctx)
-{
-    gen_st_atomic(ctx, DEF_MEMOP(MO_UL));
-}
-
-#ifdef TARGET_PPC64
-static void gen_stdat(DisasContext *ctx)
-{
-    gen_st_atomic(ctx, DEF_MEMOP(MO_UQ));
-}
-#endif
-
 /* wait */
 static void gen_wait(DisasContext *ctx)
 {
@@ -5584,12 +5378,6 @@ GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 GEN_HANDLER(isync, 0x13, 0x16, 0x04, 0x03FFF801, PPC_MEM),
-GEN_HANDLER_E(lwat, 0x1F, 0x06, 0x12, 0x00000001, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER_E(stwat, 0x1F, 0x06, 0x16, 0x00000001, PPC_NONE, PPC2_ISA300),
-#if defined(TARGET_PPC64)
-GEN_HANDLER_E(ldat, 0x1F, 0x06, 0x13, 0x00000001, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER_E(stdat, 0x1F, 0x06, 0x17, 0x00000001, PPC_NONE, PPC2_ISA300),
-#endif
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
 GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
target/ppc/translate/fixedpoint-impl.c.inc +107 -0
index 25a60d3d3a..4d35133adc 100644
@@ -1646,3 +1646,218 @@ static bool trans_SRAWI(DisasContext *ctx, arg_SRAWI *a)
     }
     return true;
 }
+
+static void do_fetch_inc_conditional(DisasContext *ctx, MemOp memop,
+                                     TCGv EA, int rt,
+                                     TCGCond cond, int addend)
+{
+    TCGv t = tcg_temp_new();
+    TCGv t2 = tcg_temp_new();
+    TCGv u = tcg_temp_new();
+
+    tcg_gen_qemu_ld_tl(t, EA, ctx->mem_idx, memop);
+    tcg_gen_addi_tl(t2, EA, memop_size(memop));
+    tcg_gen_qemu_ld_tl(t2, t2, ctx->mem_idx, memop);
+    tcg_gen_addi_tl(u, t, addend);
+
+    /* mem(EA,s) = (t cond t2 ? u = t + addend : t) */
+    tcg_gen_movcond_tl(cond, u, t, t2, u, t);
+    tcg_gen_qemu_st_tl(u, EA, ctx->mem_idx, memop);
+
+    /* RT = (t cond t2 ? t : 1<<(s*8-1)) */
+    tcg_gen_movcond_tl(cond, cpu_gpr[rt], t, t2, t,
+                       tcg_constant_tl(1 << (memop_size(memop) * 8 - 1)));
+}
+
+/*
+ * Fixed-Point Atomic Load/Store Instructions
+ *
+ * In the X-form encoding the RB field carries the Function Code (FC)
+ * that selects the atomic operation.  EA is computed from RA alone.
+ */
+static bool do_ld_atomic(DisasContext *ctx, arg_X *a, MemOp memop)
+{
+    TCGv EA, dst, src;
+    bool need_serial;
+
+    REQUIRE_INSNS_FLAGS2(ctx, ISA300);
+
+    EA = do_ea_calc(ctx, a->ra, tcg_constant_tl(0));
+    dst = cpu_gpr[a->rt];
+    src = cpu_gpr[(a->rt + 1) & 31];
+
+    need_serial = false;
+    memop |= MO_ALIGN;
+    switch (a->rb) {
+    case 0: /* Fetch and add */
+        tcg_gen_atomic_fetch_add_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 1: /* Fetch and xor */
+        tcg_gen_atomic_fetch_xor_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 2: /* Fetch and or */
+        tcg_gen_atomic_fetch_or_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 3: /* Fetch and 'and' */
+        tcg_gen_atomic_fetch_and_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 4: /* Fetch and max unsigned */
+        tcg_gen_atomic_fetch_umax_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 5: /* Fetch and max signed */
+        tcg_gen_atomic_fetch_smax_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 6: /* Fetch and min unsigned */
+        tcg_gen_atomic_fetch_umin_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 7: /* Fetch and min signed */
+        tcg_gen_atomic_fetch_smin_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+    case 8: /* Swap */
+        tcg_gen_atomic_xchg_tl(dst, EA, src, ctx->mem_idx, memop);
+        break;
+
+    case 16: /* Compare and swap not equal */
+        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
+            need_serial = true;
+        } else {
+            TCGv t0 = tcg_temp_new();
+            TCGv t1 = tcg_temp_new();
+
+            tcg_gen_qemu_ld_tl(t0, EA, ctx->mem_idx, memop);
+            if ((memop & MO_SIZE) == MO_64 || TARGET_LONG_BITS == 32) {
+                tcg_gen_mov_tl(t1, src);
+            } else {
+                tcg_gen_ext32u_tl(t1, src);
+            }
+            tcg_gen_movcond_tl(TCG_COND_NE, t1, t0, t1,
+                               cpu_gpr[(a->rt + 2) & 31], t0);
+            tcg_gen_qemu_st_tl(t1, EA, ctx->mem_idx, memop);
+            tcg_gen_mov_tl(dst, t0);
+        }
+        break;
+
+    case 24: /* Fetch and increment bounded */
+        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
+            need_serial = true;
+        } else {
+            do_fetch_inc_conditional(ctx, memop, EA, a->rt, TCG_COND_NE, 1);
+        }
+        break;
+    case 25: /* Fetch and increment equal */
+        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
+            need_serial = true;
+        } else {
+            do_fetch_inc_conditional(ctx, memop, EA, a->rt, TCG_COND_EQ, 1);
+        }
+        break;
+    case 28: /* Fetch and decrement bounded */
+        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_SRAWI(DisasContext *ctx, arg_SRAWI *a) · line 1765

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +10 -0
+            need_serial = true;
+        } else {
+            do_fetch_inc_conditional(ctx, memop, EA, a->rt, TCG_COND_NE, -1);
+        }
+        break;
+
+    default:
+        /* invoke data storage error handler */
+        gen_exception_err(ctx, POWERPC_EXCP_DSI, POWERPC_EXCP_INVAL);
+        return true;

Nit: This bypasses the need_serial block (harmless since gen_exception_err sets DISAS_NORETURN), but it's inconsistent with do_st_atomic which falls through to the final return true. The early return true here is unnecessary — dropping it makes both functions structurally uniform and keeps a single exit point.

Otherwise, LGTM.

Quoted 101 lines
target/ppc/translate/fixedpoint-impl.c.inc +98 -0
+    }
+
+    if (need_serial) {
+        /* Restart with exclusive lock.  */
+        gen_helper_exit_atomic(tcg_env);
+        ctx->base.is_jmp = DISAS_NORETURN;
+    }
+    return true;
+}
+
+static bool do_st_atomic(DisasContext *ctx, arg_X *a, MemOp memop)
+{
+    TCGv EA, src, discard;
+
+    REQUIRE_INSNS_FLAGS2(ctx, ISA300);
+
+    EA = do_ea_calc(ctx, a->ra, tcg_constant_tl(0));
+    src = cpu_gpr[a->rt];
+    discard = tcg_temp_new();
+
+    memop |= MO_ALIGN;
+    switch (a->rb) {
+    case 0: /* add and Store */
+        tcg_gen_atomic_add_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 1: /* xor and Store */
+        tcg_gen_atomic_xor_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 2: /* Or and Store */
+        tcg_gen_atomic_or_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 3: /* 'and' and Store */
+        tcg_gen_atomic_and_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 4: /* Store max unsigned */
+        tcg_gen_atomic_umax_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 5: /* Store max signed */
+        tcg_gen_atomic_smax_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 6: /* Store min unsigned */
+        tcg_gen_atomic_umin_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 7: /* Store min signed */
+        tcg_gen_atomic_smin_fetch_tl(discard, EA, src, ctx->mem_idx, memop);
+        break;
+    case 24: /* Store twin */
+        if (tb_cflags(ctx->base.tb) & CF_PARALLEL) {
+            /* Restart with exclusive lock.  */
+            gen_helper_exit_atomic(tcg_env);
+            ctx->base.is_jmp = DISAS_NORETURN;
+        } else {
+            TCGv t = tcg_temp_new();
+            TCGv t2 = tcg_temp_new();
+            TCGv s = tcg_temp_new();
+            TCGv s2 = tcg_temp_new();
+            TCGv ea_plus_s = tcg_temp_new();
+
+            tcg_gen_qemu_ld_tl(t, EA, ctx->mem_idx, memop);
+            tcg_gen_addi_tl(ea_plus_s, EA, memop_size(memop));
+            tcg_gen_qemu_ld_tl(t2, ea_plus_s, ctx->mem_idx, memop);
+            tcg_gen_movcond_tl(TCG_COND_EQ, s, t, t2, src, t);
+            tcg_gen_movcond_tl(TCG_COND_EQ, s2, t, t2, src, t2);
+            tcg_gen_qemu_st_tl(s, EA, ctx->mem_idx, memop);
+            tcg_gen_qemu_st_tl(s2, ea_plus_s, ctx->mem_idx, memop);
+        }
+        break;
+    default:
+        /* invoke data storage error handler */
+        gen_exception_err(ctx, POWERPC_EXCP_DSI, POWERPC_EXCP_INVAL);
+    }
+    return true;
+}
+
+TRANS(LWAT, do_ld_atomic, DEF_MEMOP(MO_UL))
+TRANS(STWAT, do_st_atomic, DEF_MEMOP(MO_UL))
+
+static bool trans_LDAT(DisasContext *ctx, arg_LDAT *a)
+{
+    REQUIRE_64BIT(ctx);
+#if defined(TARGET_PPC64)
+    return do_ld_atomic(ctx, a, DEF_MEMOP(MO_UQ));
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+
+static bool trans_STDAT(DisasContext *ctx, arg_STDAT *a)
+{
+    REQUIRE_64BIT(ctx);
+#if defined(TARGET_PPC64)
+    return do_st_atomic(ctx, a, DEF_MEMOP(MO_UQ));
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
Signature · 1 line
2.55.0
#30 lore raw

PATCH v2 13/37 target/ppc: Move isync instruction to decodetree.

In reply to Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Moving the following instructions to decodetree specification :
isync

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/insn32.decode                     |  1 +
 target/ppc/translate.c                       | 15 ---------------
 target/ppc/translate/storage-ctrl-impl.c.inc | 17 +++++++++++++++++
 3 files changed, 18 insertions(+), 15 deletions(-)
target/ppc/insn32.decode +1 -0
index a1c2f549bd..91ece88a70 100644
@@ -1472,6 +1472,7 @@ MSGSYNC         011111 ----- ----- ----- 1101110110 -
 @X_sync         ...... .. l:3 ... sc:2 ..... .......... .           &X_sync
 SYNC            011111 -- ... --- ..   ----- 1001010110 -           @X_sync
 EIEIO           011111 ----- ----- ----- 1101010110 -
+ISYNC           010011 ----- ----- ----- 0010010110 -
 
 # Condition Register Field Instruction
 
target/ppc/translate.c +0 -15
index 179cb69dc5..a31cda4490 100644
@@ -2727,20 +2727,6 @@ static inline void gen_check_tlb_flush(DisasContext *ctx, bool global)
 static inline void gen_check_tlb_flush(DisasContext *ctx, bool global) { }
 #endif
 
-/* isync */
-static void gen_isync(DisasContext *ctx)
-{
-    /*
-     * We need to check for a pending TLB flush. This can only happen in
-     * kernel mode however so check MSR_PR
-     */
-    if (!ctx->pr) {
-        gen_check_tlb_flush(ctx, false);
-    }
-    tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
-    ctx->base.is_jmp = DISAS_EXIT_UPDATE;
-}
-
 /* wait */
 static void gen_wait(DisasContext *ctx)
 {
@@ -5377,7 +5363,6 @@ GEN_HANDLER(lswi, 0x1F, 0x15, 0x12, 0x00000001, PPC_STRING),
 GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
-GEN_HANDLER(isync, 0x13, 0x16, 0x04, 0x03FFF801, PPC_MEM),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
 GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
target/ppc/translate/storage-ctrl-impl.c.inc +17 -0
index b8b4454663..943229c438 100644
@@ -155,6 +155,23 @@ static bool trans_SLBFEE(DisasContext *ctx, arg_SLBFEE *a)
     return true;
 }
 
+static bool trans_ISYNC(DisasContext *ctx, arg_ISYNC *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, MEM);
+
+    /*
+     * We need to check for a pending TLB flush. This can only happen in
+     * kernel mode however so check MSR_PR
+     */
+    if (!ctx->pr) {
+        gen_check_tlb_flush(ctx, false);
+    }
+    tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
+    ctx->base.is_jmp = DISAS_EXIT_UPDATE;
+
+    return true;
+}
+
 static bool trans_SLBSYNC(DisasContext *ctx, arg_SLBSYNC *a)
 {
     REQUIRE_64BIT(ctx);
Signature · 1 line
2.55.0
Reviewed-by on v2 13/37
#31 lore raw

Re: [PATCH v2 13/37] target/ppc: Move isync instruction to decodetree.

Quoted 8 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Moving the following instructions to decodetree specification :
isync

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

Nit: Trailing period in subject line. Otherwise,

Quoted 87 lines
---
 target/ppc/insn32.decode                     |  1 +
 target/ppc/translate.c                       | 15 ---------------
 target/ppc/translate/storage-ctrl-impl.c.inc | 17 +++++++++++++++++
 3 files changed, 18 insertions(+), 15 deletions(-)
target/ppc/insn32.decode +1 -0
index a1c2f549bd..91ece88a70 100644
@@ -1472,6 +1472,7 @@ MSGSYNC         011111 ----- ----- ----- 1101110110 -
 @X_sync         ...... .. l:3 ... sc:2 ..... .......... .           &X_sync
 SYNC            011111 -- ... --- ..   ----- 1001010110 -           @X_sync
 EIEIO           011111 ----- ----- ----- 1101010110 -
+ISYNC           010011 ----- ----- ----- 0010010110 -
 
 # Condition Register Field Instruction
 
target/ppc/translate.c +0 -15
index 179cb69dc5..a31cda4490 100644
@@ -2727,20 +2727,6 @@ static inline void gen_check_tlb_flush(DisasContext *ctx, bool global)
 static inline void gen_check_tlb_flush(DisasContext *ctx, bool global) { }
 #endif
 
-/* isync */
-static void gen_isync(DisasContext *ctx)
-{
-    /*
-     * We need to check for a pending TLB flush. This can only happen in
-     * kernel mode however so check MSR_PR
-     */
-    if (!ctx->pr) {
-        gen_check_tlb_flush(ctx, false);
-    }
-    tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
-    ctx->base.is_jmp = DISAS_EXIT_UPDATE;
-}
-
 /* wait */
 static void gen_wait(DisasContext *ctx)
 {
@@ -5377,7 +5363,6 @@ GEN_HANDLER(lswi, 0x1F, 0x15, 0x12, 0x00000001, PPC_STRING),
 GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
-GEN_HANDLER(isync, 0x13, 0x16, 0x04, 0x03FFF801, PPC_MEM),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
 GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
target/ppc/translate/storage-ctrl-impl.c.inc +17 -0
index b8b4454663..943229c438 100644
@@ -155,6 +155,23 @@ static bool trans_SLBFEE(DisasContext *ctx, arg_SLBFEE *a)
     return true;
 }
 
+static bool trans_ISYNC(DisasContext *ctx, arg_ISYNC *a)
+{
+    REQUIRE_INSNS_FLAGS(ctx, MEM);
+
+    /*
+     * We need to check for a pending TLB flush. This can only happen in
+     * kernel mode however so check MSR_PR
+     */
+    if (!ctx->pr) {
+        gen_check_tlb_flush(ctx, false);
+    }
+    tcg_gen_mb(TCG_MO_ALL | TCG_BAR_SC);
+    ctx->base.is_jmp = DISAS_EXIT_UPDATE;
+
+    return true;
+}
+
 static bool trans_SLBSYNC(DisasContext *ctx, arg_SLBSYNC *a)
 {
     REQUIRE_64BIT(ctx);
Signature · 1 line
2.55.0
#32 lore raw

PATCH v2 14/37 target/ppc: Convert b{a, l, la} to decode tree

In reply to Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Convert b, ba, bl and bla to decode tree specification. The functionality was tested by comparing the qemu -D log -d op,in_asm output as well as single stepping gdb to confirm LR was correctly populated only when LK is set.

---
 target/ppc/insn32.decode               |  7 +++++++
 target/ppc/internal.h                  |  7 -------
 target/ppc/translate.c                 | 24 ------------------------
 target/ppc/translate/branch-impl.c.inc | 24 ++++++++++++++++++++++++
 4 files changed, 31 insertions(+), 31 deletions(-)
target/ppc/insn32.decode +7 -0
index 91ece88a70..6a3254aaf6 100644
@@ -1488,6 +1488,13 @@ MCRF            010011 ... -- ... -- ----- 00000 00000 -       @XL_bfa
 MFBHRBE         011111 ..... ..... ..... 0100101110 -   @XFX_bhrbe
 CLRBHRB         011111 ----- ----- ----- 0110101110 -
 
+# Branch Instructions
+%li             2:24 !function=times_4
+&I_b            li aa:bool lk:bool
+@I_b            ...... ........................ aa:1 lk:1 &I_b li=%li
+
+B               010010 ........................ . .  @I_b
+
 ## Misc POWER instructions
 
 ATTN            000000 00000 00000 00000 0100000000 0
target/ppc/internal.h +0 -7
index a3bb12afd9..a20a0f9f7e 100644
@@ -197,13 +197,6 @@ EXTRACT_HELPER(L, 16, 2);
 EXTRACT_HELPER(WC, 21, 2);
 EXTRACT_HELPER(PL, 16, 2);
 
-/***                            Jump target decoding                       ***/
-/* Immediate address */
-static inline target_ulong LI(uint32_t opcode)
-{
-    return (opcode >> 0) & 0x03FFFFFC;
-}
-
 static inline uint32_t BD(uint32_t opcode)
 {
     return (opcode >> 0) & 0xFFFC;
target/ppc/translate.c +0 -24
index a31cda4490..b528cab17d 100644
@@ -3089,29 +3089,6 @@ static inline void gen_setlr(DisasContext *ctx, target_ulong nip)
     tcg_gen_movi_tl(cpu_lr, nip);
 }
 
-/* b ba bl bla */
-static void gen_b(DisasContext *ctx)
-{
-    target_ulong li, target;
-
-    /* sign extend LI */
-    li = LI(ctx->opcode);
-    li = (li ^ 0x02000000) - 0x02000000;
-    if (likely(AA(ctx->opcode) == 0)) {
-        target = ctx->cia + li;
-    } else {
-        target = li;
-    }
-    if (LK(ctx->opcode)) {
-        gen_setlr(ctx, ctx->base.pc_next);
-        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_CALL);
-    } else {
-        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_OTHER);
-    }
-    gen_goto_tb(ctx, 0, target);
-    ctx->base.is_jmp = DISAS_NORETURN;
-}
-
 #define BCOND_IM  0
 #define BCOND_LR  1
 #define BCOND_CTR 2
@@ -5366,7 +5343,6 @@ GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
 GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(b, 0x12, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bc, 0x10, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bcctr, 0x13, 0x10, 0x10, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bclr, 0x13, 0x10, 0x00, 0x00000000, PPC_FLOW),
target/ppc/translate/branch-impl.c.inc +24 -0
index 9ade0c659a..745f71afd1 100644
@@ -31,3 +31,27 @@ static bool trans_RFEBB(DisasContext *ctx, arg_XL_s *arg)
     return true;
 }
 #endif
+
+static bool trans_B(DisasContext *ctx, arg_I_b *a)
+{
+    target_ulong target, li;
+
+    /* sign extend LI */
+    li = (a->li ^ 0x02000000) - 0x02000000;
+
+    if (likely(a->aa == 0)) {
+        target = ctx->cia + li;
+    } else {
+        target = li;
+    }
+    if (a->lk) {
+        gen_setlr(ctx, ctx->base.pc_next);
+        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_CALL);
+    } else {
+        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_OTHER);
+    }
+    gen_goto_tb(ctx, 0, target);
+    ctx->base.is_jmp = DISAS_NORETURN;
+
+    return true;
+}
Signature · 1 line
2.55.0
Reviewed-by on v2 14/37
#33 lore raw

Re: [PATCH v2 14/37] target/ppc: Convert b{a, l, la} to decode tree

Quoted 11 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Convert b, ba, bl and bla to decode tree specification. The functionality was tested by comparing the qemu -D log -d op,in_asm output as well as single stepping gdb to confirm LR was correctly populated only when LK is set.

Quoted 6 lines
---
 target/ppc/insn32.decode               |  7 +++++++
 target/ppc/internal.h                  |  7 -------
 target/ppc/translate.c                 | 24 ------------------------
 target/ppc/translate/branch-impl.c.inc | 24 ++++++++++++++++++++++++
 4 files changed, 31 insertions(+), 31 deletions(-)
Quoted 114 lines
target/ppc/insn32.decode +7 -0
index 91ece88a70..6a3254aaf6 100644
@@ -1488,6 +1488,13 @@ MCRF            010011 ... -- ... -- ----- 00000 00000 -       @XL_bfa
 MFBHRBE         011111 ..... ..... ..... 0100101110 -   @XFX_bhrbe
 CLRBHRB         011111 ----- ----- ----- 0110101110 -
 
+# Branch Instructions
+%li             2:24 !function=times_4
+&I_b            li aa:bool lk:bool
+@I_b            ...... ........................ aa:1 lk:1 &I_b li=%li
+
+B               010010 ........................ . .  @I_b
+
 ## Misc POWER instructions
 
 ATTN            000000 00000 00000 00000 0100000000 0
target/ppc/internal.h +0 -7
index a3bb12afd9..a20a0f9f7e 100644
@@ -197,13 +197,6 @@ EXTRACT_HELPER(L, 16, 2);
 EXTRACT_HELPER(WC, 21, 2);
 EXTRACT_HELPER(PL, 16, 2);
 
-/***                            Jump target decoding                       ***/
-/* Immediate address */
-static inline target_ulong LI(uint32_t opcode)
-{
-    return (opcode >> 0) & 0x03FFFFFC;
-}
-
 static inline uint32_t BD(uint32_t opcode)
 {
     return (opcode >> 0) & 0xFFFC;
target/ppc/translate.c +0 -24
index a31cda4490..b528cab17d 100644
@@ -3089,29 +3089,6 @@ static inline void gen_setlr(DisasContext *ctx, target_ulong nip)
     tcg_gen_movi_tl(cpu_lr, nip);
 }
 
-/* b ba bl bla */
-static void gen_b(DisasContext *ctx)
-{
-    target_ulong li, target;
-
-    /* sign extend LI */
-    li = LI(ctx->opcode);
-    li = (li ^ 0x02000000) - 0x02000000;
-    if (likely(AA(ctx->opcode) == 0)) {
-        target = ctx->cia + li;
-    } else {
-        target = li;
-    }
-    if (LK(ctx->opcode)) {
-        gen_setlr(ctx, ctx->base.pc_next);
-        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_CALL);
-    } else {
-        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_OTHER);
-    }
-    gen_goto_tb(ctx, 0, target);
-    ctx->base.is_jmp = DISAS_NORETURN;
-}
-
 #define BCOND_IM  0
 #define BCOND_LR  1
 #define BCOND_CTR 2
@@ -5366,7 +5343,6 @@ GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
 GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(b, 0x12, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bc, 0x10, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bcctr, 0x13, 0x10, 0x10, 0x00000000, PPC_FLOW),
 GEN_HANDLER(bclr, 0x13, 0x10, 0x00, 0x00000000, PPC_FLOW),
target/ppc/translate/branch-impl.c.inc +24 -0
index 9ade0c659a..745f71afd1 100644
@@ -31,3 +31,27 @@ static bool trans_RFEBB(DisasContext *ctx, arg_XL_s *arg)
     return true;
 }
 #endif
+
+static bool trans_B(DisasContext *ctx, arg_I_b *a)
+{
+    target_ulong target, li;
+
+    /* sign extend LI */
+    li = (a->li ^ 0x02000000) - 0x02000000;
+
+    if (likely(a->aa == 0)) {
+        target = ctx->cia + li;
+    } else {
+        target = li;
+    }
+    if (a->lk) {
+        gen_setlr(ctx, ctx->base.pc_next);
+        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_CALL);
+    } else {
+        gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_OTHER);
+    }
+    gen_goto_tb(ctx, 0, target);
+    ctx->base.is_jmp = DISAS_NORETURN;
+
+    return true;
+}
Signature · 1 line
2.55.0
#34 lore raw

PATCH v2 15/37 target/ppc: move various conditional branch insns to decodetree

In reply to Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Convert the following instructions to decodetree specification:

bc
bca
bcl
bcla

bclr
bclrl

bcctr
bcctrl

bctar
bctarl

The branch was tested by comparing the qemu -D log -d op,in_asm output for each instruction and also for the various combinations of branch option (BO) bits for the conditional branch instructions. Additionally, gdb was used to confirm the LR and other register's behavior is consistent to legacy behavior.

Further, the changes also pass a boot test into Fedora distro.

---
 target/ppc/insn32.decode               |  13 ++
 target/ppc/translate.c                 | 148 -----------------------
 target/ppc/translate/branch-impl.c.inc | 157 +++++++++++++++++++++++++
 3 files changed, 170 insertions(+), 148 deletions(-)
target/ppc/insn32.decode +13 -0
index 6a3254aaf6..67fb181542 100644
@@ -1495,6 +1495,19 @@ CLRBHRB         011111 ----- ----- ----- 0110101110 -
 
 B               010010 ........................ . .  @I_b
 
+%bd             2:14 !function=times_4
+&bcond          bo:uint32_t bi bd aa:bool lk:bool
+@B_bcond        ...... bo:5 bi:5 .............. aa:1 lk:1 &bcond bd=%bd
+
+BC              010000 ..... ..... .............. . .  @B_bcond
+
+&bclr          bo bi bh lk:bool
+@XL_bclr       ...... bo:5 bi:5 --- bh:2 .......... lk:1 &bclr
+
+BCLR            010011 ..... ..... ---.. 0000010000 . @XL_bclr
+BCCTR           010011 ..... ..... ---.. 1000010000 . @XL_bclr
+BCTAR           010011 ..... ..... ---.. 1000110000 . @XL_bclr
+
 ## Misc POWER instructions
 
 ATTN            000000 00000 00000 00000 0100000000 0
target/ppc/translate.c +0 -148

160 lines not on this page · raw message

target/ppc/translate/branch-impl.c.inc +157 -0

160 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 15/37
#35 lore raw

Re: [PATCH v2 15/37] target/ppc: move various conditional branch insns to decodetree

Quoted 26 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Convert the following instructions to decodetree specification:

bc
bca
bcl
bcla

bclr
bclrl

bcctr
bcctrl

bctar
bctarl

The branch was tested by comparing the qemu -D log -d op,in_asm output for each instruction and also for the various combinations of branch option (BO) bits for the conditional branch instructions. Additionally, gdb was used to confirm the LR and other register's behavior is consistent to legacy behavior.

Further, the changes also pass a boot test into Fedora distro.

Quoted 8 lines
---
 target/ppc/insn32.decode               |  13 ++
 target/ppc/translate.c                 | 148 -----------------------
 target/ppc/translate/branch-impl.c.inc | 157 +++++++++++++++++++++++++
 3 files changed, 170 insertions(+), 148 deletions(-)
Quoted 359 lines
target/ppc/insn32.decode +13 -0
index 6a3254aaf6..67fb181542 100644
@@ -1495,6 +1495,19 @@ CLRBHRB         011111 ----- ----- ----- 0110101110 -
 
 B               010010 ........................ . .  @I_b
 
+%bd             2:14 !function=times_4
+&bcond          bo:uint32_t bi bd aa:bool lk:bool
+@B_bcond        ...... bo:5 bi:5 .............. aa:1 lk:1 &bcond bd=%bd
+
+BC              010000 ..... ..... .............. . .  @B_bcond
+
+&bclr          bo bi bh lk:bool
+@XL_bclr       ...... bo:5 bi:5 --- bh:2 .......... lk:1 &bclr
+
+BCLR            010011 ..... ..... ---.. 0000010000 . @XL_bclr
+BCCTR           010011 ..... ..... ---.. 1000010000 . @XL_bclr
+BCTAR           010011 ..... ..... ---.. 1000110000 . @XL_bclr
+
 ## Misc POWER instructions
 
 ATTN            000000 00000 00000 00000 0100000000 0
target/ppc/translate.c +0 -148
index b528cab17d..7100787a25 100644
@@ -3089,150 +3089,6 @@ static inline void gen_setlr(DisasContext *ctx, target_ulong nip)
     tcg_gen_movi_tl(cpu_lr, nip);
 }
 
-#define BCOND_IM  0
-#define BCOND_LR  1
-#define BCOND_CTR 2
-#define BCOND_TAR 3
-
-static void gen_bcond(DisasContext *ctx, int type)
-{
-    uint32_t bo = BO(ctx->opcode);
-    TCGLabel *l1;
-    TCGv target;
-    target_long bhrb_type = BHRB_TYPE_OTHER;
-
-    if (type == BCOND_LR || type == BCOND_CTR || type == BCOND_TAR) {
-        target = tcg_temp_new();
-        if (type == BCOND_CTR) {
-            tcg_gen_mov_tl(target, cpu_ctr);
-        } else if (type == BCOND_TAR) {
-            gen_load_spr(target, SPR_TAR);
-        } else {
-            tcg_gen_mov_tl(target, cpu_lr);
-        }
-        if (!LK(ctx->opcode)) {
-            bhrb_type |= BHRB_TYPE_INDIRECT;
-        }
-        bhrb_type |= BHRB_TYPE_XL_FORM;
-    } else {
-        target = NULL;
-    }
-    if (LK(ctx->opcode)) {
-        gen_setlr(ctx, ctx->base.pc_next);
-        bhrb_type |= BHRB_TYPE_CALL;
-    }
-    l1 = gen_new_label();
-    if ((bo & 0x4) == 0) {
-        /* Decrement and test CTR */
-        TCGv temp = tcg_temp_new();
-
-        if (type == BCOND_CTR) {
-            /*
-             * All ISAs up to v3 describe this form of bcctr as invalid but
-             * some processors, ie. 64-bit server processors compliant with
-             * arch 2.x, do implement a "test and decrement" logic instead,
-             * as described in their respective UMs. This logic involves CTR
-             * to act as both the branch target and a counter, which makes
-             * it basically useless and thus never used in real code.
-             *
-             * This form was hence chosen to trigger extra micro-architectural
-             * side-effect on real HW needed for the Spectre v2 workaround.
-             * It is up to guests that implement such workaround, ie. linux, to
-             * use this form in a way it just triggers the side-effect without
-             * doing anything else harmful.
-             */
-            if (unlikely(!is_book3s_arch2x(ctx))) {
-                gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-                return;
-            }
-
-            if (NARROW_MODE(ctx)) {
-                tcg_gen_ext32u_tl(temp, cpu_ctr);
-            } else {
-                tcg_gen_mov_tl(temp, cpu_ctr);
-            }
-            if (bo & 0x2) {
-                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
-            } else {
-                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
-            }
-            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
-        } else {
-            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
-            if (NARROW_MODE(ctx)) {
-                tcg_gen_ext32u_tl(temp, cpu_ctr);
-            } else {
-                tcg_gen_mov_tl(temp, cpu_ctr);
-            }
-            if (bo & 0x2) {
-                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
-            } else {
-                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
-            }
-        }
-        bhrb_type |= BHRB_TYPE_COND;
-    }
-    if ((bo & 0x10) == 0) {
-        /* Test CR */
-        uint32_t bi = BI(ctx->opcode);
-        uint32_t mask = 0x08 >> (bi & 0x03);
-        TCGv_i32 temp = tcg_temp_new_i32();
-
-        if (bo & 0x8) {
-            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
-            tcg_gen_brcondi_i32(TCG_COND_EQ, temp, 0, l1);
-        } else {
-            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
-            tcg_gen_brcondi_i32(TCG_COND_NE, temp, 0, l1);
-        }
-        bhrb_type |= BHRB_TYPE_COND;
-    }
-
-    gen_update_branch_history(ctx, ctx->cia, target, bhrb_type);
-
-    if (type == BCOND_IM) {
-        target_ulong li = (target_long)((int16_t)(BD(ctx->opcode)));
-        if (likely(AA(ctx->opcode) == 0)) {
-            gen_goto_tb(ctx, 0, ctx->cia + li);
-        } else {
-            gen_goto_tb(ctx, 0, li);
-        }
-    } else {
-        if (NARROW_MODE(ctx)) {
-            tcg_gen_andi_tl(cpu_nip, target, (uint32_t)~3);
-        } else {
-            tcg_gen_andi_tl(cpu_nip, target, ~3);
-        }
-        gen_lookup_and_goto_ptr(ctx);
-    }
-    if ((bo & 0x14) != 0x14) {
-        /* fallthrough case */
-        gen_set_label(l1);
-        gen_goto_tb(ctx, 1, ctx->base.pc_next);
-    }
-    ctx->base.is_jmp = DISAS_NORETURN;
-}
-
-static void gen_bc(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_IM);
-}
-
-static void gen_bcctr(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_CTR);
-}
-
-static void gen_bclr(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_LR);
-}
-
-static void gen_bctar(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_TAR);
-}
-
 /***                      Condition register logical                       ***/
 #define GEN_CRLOGIC(name, tcg_op, opc)                                        \
 static void glue(gen_, name)(DisasContext *ctx)                               \
@@ -5343,10 +5199,6 @@ GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
 GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(bc, 0x10, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
-GEN_HANDLER(bcctr, 0x13, 0x10, 0x10, 0x00000000, PPC_FLOW),
-GEN_HANDLER(bclr, 0x13, 0x10, 0x00, 0x00000000, PPC_FLOW),
-GEN_HANDLER_E(bctar, 0x13, 0x10, 0x11, 0x0000E000, PPC_NONE, PPC2_BCTAR_ISA207),
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(rfid, 0x13, 0x12, 0x00, 0x03FF8001, PPC_64B),
target/ppc/translate/branch-impl.c.inc +157 -0
index 745f71afd1..c2889c1d12 100644
@@ -55,3 +55,160 @@ static bool trans_B(DisasContext *ctx, arg_I_b *a)
 
     return true;
 }
+
+#define BCOND_IM  0
+#define BCOND_LR  1
+#define BCOND_CTR 2
+#define BCOND_TAR 3
+
+static bool bcond_helper(DisasContext *ctx, int type, uint32_t bo, int bi,
+                         int bd, int bh, bool aa, bool lk)
+{
+    TCGLabel *l1;
+    TCGv target;
+    target_long bhrb_type = BHRB_TYPE_OTHER;
+
+    if (type == BCOND_IM && bh != -1) {
+        /* BCOND_IM should never use bh */
+        return false;
+    } else if (type != BCOND_IM && (bd != -1 || aa != 0)) {
+        /* Other BCOND types should never use bd or aa */
+        return false;
+    }
+
+    if (type == BCOND_LR || type == BCOND_CTR || type == BCOND_TAR) {
+        target = tcg_temp_new();
+        if (type == BCOND_CTR) {
+            tcg_gen_mov_tl(target, cpu_ctr);
+        } else if (type == BCOND_TAR) {
+            gen_load_spr(target, SPR_TAR);
+        } else {
+            tcg_gen_mov_tl(target, cpu_lr);
+        }
+        if (!lk) {
+            bhrb_type |= BHRB_TYPE_INDIRECT;
+        }
+        bhrb_type |= BHRB_TYPE_XL_FORM;
+    } else {
+        target = NULL;
+    }
+    if (lk) {
+        gen_setlr(ctx, ctx->base.pc_next);
+        bhrb_type |= BHRB_TYPE_CALL;
+    }
+    l1 = gen_new_label();
+    if ((bo & 0x4) == 0) {
+        /* Decrement and test CTR */
+        TCGv temp = tcg_temp_new();
+
+        if (type == BCOND_CTR) {
+            /*
+             * All ISAs up to v3 describe this form of bcctr as invalid but
+             * some processors, ie. 64-bit server processors compliant with
+             * arch 2.x, do implement a "test and decrement" logic instead,
+             * as described in their respective UMs. This logic involves CTR
+             * to act as both the branch target and a counter, which makes
+             * it basically useless and thus never used in real code.
+             *
+             * This form was hence chosen to trigger extra micro-architectural
+             * side-effect on real HW needed for the Spectre v2 workaround.
+             * It is up to guests that implement such workaround, ie. linux, to
+             * use this form in a way it just triggers the side-effect without
+             * doing anything else harmful.
+             */
+            if (unlikely(!is_book3s_arch2x(ctx))) {
+                gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+                return true;
+            }
+
+            if (NARROW_MODE(ctx)) {
+                tcg_gen_ext32u_tl(temp, cpu_ctr);
+            } else {
+                tcg_gen_mov_tl(temp, cpu_ctr);
+            }
+            if (bo & 0x2) {
+                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
+            } else {
+                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
+            }
+            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
+        } else {
+            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
+            if (NARROW_MODE(ctx)) {
+                tcg_gen_ext32u_tl(temp, cpu_ctr);
+            } else {
+                tcg_gen_mov_tl(temp, cpu_ctr);
+            }
+            if (bo & 0x2) {
+                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
+            } else {
+                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
+            }
+        }
+        bhrb_type |= BHRB_TYPE_COND;
+    }
+    if ((bo & 0x10) == 0) {
+        /* Test CR */
+        uint32_t mask = 0x08 >> (bi & 0x03);
+        TCGv_i32 temp = tcg_temp_new_i32();
+
+        if (bo & 0x8) {
+            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
+            tcg_gen_brcondi_i32(TCG_COND_EQ, temp, 0, l1);
+        } else {
+            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
+            tcg_gen_brcondi_i32(TCG_COND_NE, temp, 0, l1);
+        }
+        bhrb_type |= BHRB_TYPE_COND;
+    }
+
+    gen_update_branch_history(ctx, ctx->cia, target, bhrb_type);
+
+    if (type == BCOND_IM) {
+        target_ulong li = (target_long)((int16_t)(bd));
+        if (likely(aa == 0)) {
+            gen_goto_tb(ctx, 0, ctx->cia + li);
+        } else {
+            gen_goto_tb(ctx, 0, li);
+        }
+    } else {
+        if (NARROW_MODE(ctx)) {
+            tcg_gen_andi_tl(cpu_nip, target, (uint32_t)~3);
+        } else {
+            tcg_gen_andi_tl(cpu_nip, target, ~3);
+        }
+        gen_lookup_and_goto_ptr(ctx);
+    }
+    if ((bo & 0x14) != 0x14) {
+        /* fallthrough case */
+        gen_set_label(l1);
+        gen_goto_tb(ctx, 1, ctx->base.pc_next);
+    }
+    ctx->base.is_jmp = DISAS_NORETURN;
+
+    return true;
+}
+
+static bool trans_BC(DisasContext *ctx, arg_bcond *a)
+{
+    /*
+     * bh is not used for bc variants hence we pass -1
+     */
+    return bcond_helper(ctx, BCOND_IM, a->bo, a->bi, a->bd, -1, a->aa, a->lk);
+}
+
+/*
+ * This helper is shared by bclr, bcctr and bctar.
+ */
+static bool bclr_helper(DisasContext *ctx, arg_bclr *a, int type)
+{
+    /*
+     * bd and aa is not used for bc variants hence we pass -1 and 0 respectively
+     */
+    return bcond_helper(ctx, type, a->bo, a->bi, -1, a->bh, 0, a->lk);
+}
+
+TRANS(BCLR, bclr_helper, BCOND_LR)
+TRANS(BCCTR, bclr_helper, BCOND_CTR)
+TRANS_FLAGS2(BCTAR_ISA207, BCTAR, bclr_helper, BCOND_TAR)
+
Signature · 1 line
2.55.0
#36 lore raw

Re: [PATCH v2 15/37] target/ppc: move various conditional branch insns to decodetree

In reply to Chinmay Rath
Quoted 380 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Convert the following instructions to decodetree specification:

bc
bca
bcl
bcla

bclr
bclrl

bcctr
bcctrl

bctar
bctarl

The branch was tested by comparing the qemu -D log -d op,in_asm output for each instruction and also for the various combinations of branch option (BO) bits for the conditional branch instructions. Additionally, gdb was used to confirm the LR and other register's behavior is consistent to legacy behavior.

Further, the changes also pass a boot test into Fedora distro.

---
 target/ppc/insn32.decode               |  13 ++
 target/ppc/translate.c                 | 148 -----------------------
 target/ppc/translate/branch-impl.c.inc | 157 +++++++++++++++++++++++++
 3 files changed, 170 insertions(+), 148 deletions(-)
target/ppc/insn32.decode +13 -0
index 6a3254aaf6..67fb181542 100644
@@ -1495,6 +1495,19 @@ CLRBHRB         011111 ----- ----- ----- 0110101110 -
 
 B               010010 ........................ . .  @I_b
 
+%bd             2:14 !function=times_4
+&bcond          bo:uint32_t bi bd aa:bool lk:bool
+@B_bcond        ...... bo:5 bi:5 .............. aa:1 lk:1 &bcond bd=%bd
+
+BC              010000 ..... ..... .............. . .  @B_bcond
+
+&bclr          bo bi bh lk:bool
+@XL_bclr       ...... bo:5 bi:5 --- bh:2 .......... lk:1 &bclr
+
+BCLR            010011 ..... ..... ---.. 0000010000 . @XL_bclr
+BCCTR           010011 ..... ..... ---.. 1000010000 . @XL_bclr
+BCTAR           010011 ..... ..... ---.. 1000110000 . @XL_bclr
+
 ## Misc POWER instructions
 
 ATTN            000000 00000 00000 00000 0100000000 0
target/ppc/translate.c +0 -148
index b528cab17d..7100787a25 100644
@@ -3089,150 +3089,6 @@ static inline void gen_setlr(DisasContext *ctx, target_ulong nip)
     tcg_gen_movi_tl(cpu_lr, nip);
 }
 
-#define BCOND_IM  0
-#define BCOND_LR  1
-#define BCOND_CTR 2
-#define BCOND_TAR 3
-
-static void gen_bcond(DisasContext *ctx, int type)
-{
-    uint32_t bo = BO(ctx->opcode);
-    TCGLabel *l1;
-    TCGv target;
-    target_long bhrb_type = BHRB_TYPE_OTHER;
-
-    if (type == BCOND_LR || type == BCOND_CTR || type == BCOND_TAR) {
-        target = tcg_temp_new();
-        if (type == BCOND_CTR) {
-            tcg_gen_mov_tl(target, cpu_ctr);
-        } else if (type == BCOND_TAR) {
-            gen_load_spr(target, SPR_TAR);
-        } else {
-            tcg_gen_mov_tl(target, cpu_lr);
-        }
-        if (!LK(ctx->opcode)) {
-            bhrb_type |= BHRB_TYPE_INDIRECT;
-        }
-        bhrb_type |= BHRB_TYPE_XL_FORM;
-    } else {
-        target = NULL;
-    }
-    if (LK(ctx->opcode)) {
-        gen_setlr(ctx, ctx->base.pc_next);
-        bhrb_type |= BHRB_TYPE_CALL;
-    }
-    l1 = gen_new_label();
-    if ((bo & 0x4) == 0) {
-        /* Decrement and test CTR */
-        TCGv temp = tcg_temp_new();
-
-        if (type == BCOND_CTR) {
-            /*
-             * All ISAs up to v3 describe this form of bcctr as invalid but
-             * some processors, ie. 64-bit server processors compliant with
-             * arch 2.x, do implement a "test and decrement" logic instead,
-             * as described in their respective UMs. This logic involves CTR
-             * to act as both the branch target and a counter, which makes
-             * it basically useless and thus never used in real code.
-             *
-             * This form was hence chosen to trigger extra micro-architectural
-             * side-effect on real HW needed for the Spectre v2 workaround.
-             * It is up to guests that implement such workaround, ie. linux, to
-             * use this form in a way it just triggers the side-effect without
-             * doing anything else harmful.
-             */
-            if (unlikely(!is_book3s_arch2x(ctx))) {
-                gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-                return;
-            }
-
-            if (NARROW_MODE(ctx)) {
-                tcg_gen_ext32u_tl(temp, cpu_ctr);
-            } else {
-                tcg_gen_mov_tl(temp, cpu_ctr);
-            }
-            if (bo & 0x2) {
-                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
-            } else {
-                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
-            }
-            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
-        } else {
-            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
-            if (NARROW_MODE(ctx)) {
-                tcg_gen_ext32u_tl(temp, cpu_ctr);
-            } else {
-                tcg_gen_mov_tl(temp, cpu_ctr);
-            }
-            if (bo & 0x2) {
-                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
-            } else {
-                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
-            }
-        }
-        bhrb_type |= BHRB_TYPE_COND;
-    }
-    if ((bo & 0x10) == 0) {
-        /* Test CR */
-        uint32_t bi = BI(ctx->opcode);
-        uint32_t mask = 0x08 >> (bi & 0x03);
-        TCGv_i32 temp = tcg_temp_new_i32();
-
-        if (bo & 0x8) {
-            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
-            tcg_gen_brcondi_i32(TCG_COND_EQ, temp, 0, l1);
-        } else {
-            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
-            tcg_gen_brcondi_i32(TCG_COND_NE, temp, 0, l1);
-        }
-        bhrb_type |= BHRB_TYPE_COND;
-    }
-
-    gen_update_branch_history(ctx, ctx->cia, target, bhrb_type);
-
-    if (type == BCOND_IM) {
-        target_ulong li = (target_long)((int16_t)(BD(ctx->opcode)));
-        if (likely(AA(ctx->opcode) == 0)) {
-            gen_goto_tb(ctx, 0, ctx->cia + li);
-        } else {
-            gen_goto_tb(ctx, 0, li);
-        }
-    } else {
-        if (NARROW_MODE(ctx)) {
-            tcg_gen_andi_tl(cpu_nip, target, (uint32_t)~3);
-        } else {
-            tcg_gen_andi_tl(cpu_nip, target, ~3);
-        }
-        gen_lookup_and_goto_ptr(ctx);
-    }
-    if ((bo & 0x14) != 0x14) {
-        /* fallthrough case */
-        gen_set_label(l1);
-        gen_goto_tb(ctx, 1, ctx->base.pc_next);
-    }
-    ctx->base.is_jmp = DISAS_NORETURN;
-}
-
-static void gen_bc(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_IM);
-}
-
-static void gen_bcctr(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_CTR);
-}
-
-static void gen_bclr(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_LR);
-}
-
-static void gen_bctar(DisasContext *ctx)
-{
-    gen_bcond(ctx, BCOND_TAR);
-}
-
 /***                      Condition register logical                       ***/
 #define GEN_CRLOGIC(name, tcg_op, opc)                                        \
 static void glue(gen_, name)(DisasContext *ctx)                               \
@@ -5343,10 +5199,6 @@ GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
 GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(bc, 0x10, 0xFF, 0xFF, 0x00000000, PPC_FLOW),
-GEN_HANDLER(bcctr, 0x13, 0x10, 0x10, 0x00000000, PPC_FLOW),
-GEN_HANDLER(bclr, 0x13, 0x10, 0x00, 0x00000000, PPC_FLOW),
-GEN_HANDLER_E(bctar, 0x13, 0x10, 0x11, 0x0000E000, PPC_NONE, PPC2_BCTAR_ISA207),
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(rfid, 0x13, 0x12, 0x00, 0x03FF8001, PPC_64B),
target/ppc/translate/branch-impl.c.inc +147 -0
index 745f71afd1..c2889c1d12 100644
@@ -55,3 +55,160 @@ static bool trans_B(DisasContext *ctx, arg_I_b *a)
 
     return true;
 }
+
+#define BCOND_IM  0
+#define BCOND_LR  1
+#define BCOND_CTR 2
+#define BCOND_TAR 3
+
+static bool bcond_helper(DisasContext *ctx, int type, uint32_t bo, int bi,
+                         int bd, int bh, bool aa, bool lk)
+{
+    TCGLabel *l1;
+    TCGv target;
+    target_long bhrb_type = BHRB_TYPE_OTHER;
+
+    if (type == BCOND_IM && bh != -1) {
+        /* BCOND_IM should never use bh */
+        return false;
+    } else if (type != BCOND_IM && (bd != -1 || aa != 0)) {
+        /* Other BCOND types should never use bd or aa */
+        return false;
+    }
+
+    if (type == BCOND_LR || type == BCOND_CTR || type == BCOND_TAR) {
+        target = tcg_temp_new();
+        if (type == BCOND_CTR) {
+            tcg_gen_mov_tl(target, cpu_ctr);
+        } else if (type == BCOND_TAR) {
+            gen_load_spr(target, SPR_TAR);
+        } else {
+            tcg_gen_mov_tl(target, cpu_lr);
+        }
+        if (!lk) {
+            bhrb_type |= BHRB_TYPE_INDIRECT;
+        }
+        bhrb_type |= BHRB_TYPE_XL_FORM;
+    } else {
+        target = NULL;
+    }
+    if (lk) {
+        gen_setlr(ctx, ctx->base.pc_next);
+        bhrb_type |= BHRB_TYPE_CALL;
+    }
+    l1 = gen_new_label();
+    if ((bo & 0x4) == 0) {
+        /* Decrement and test CTR */
+        TCGv temp = tcg_temp_new();
+
+        if (type == BCOND_CTR) {
+            /*
+             * All ISAs up to v3 describe this form of bcctr as invalid but
+             * some processors, ie. 64-bit server processors compliant with
+             * arch 2.x, do implement a "test and decrement" logic instead,
+             * as described in their respective UMs. This logic involves CTR
+             * to act as both the branch target and a counter, which makes
+             * it basically useless and thus never used in real code.
+             *
+             * This form was hence chosen to trigger extra micro-architectural
+             * side-effect on real HW needed for the Spectre v2 workaround.
+             * It is up to guests that implement such workaround, ie. linux, to
+             * use this form in a way it just triggers the side-effect without
+             * doing anything else harmful.
+             */
+            if (unlikely(!is_book3s_arch2x(ctx))) {
+                gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
+                return true;
+            }
+
+            if (NARROW_MODE(ctx)) {
+                tcg_gen_ext32u_tl(temp, cpu_ctr);
+            } else {
+                tcg_gen_mov_tl(temp, cpu_ctr);
+            }
+            if (bo & 0x2) {
+                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
+            } else {
+                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
+            }
+            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
+        } else {
+            tcg_gen_subi_tl(cpu_ctr, cpu_ctr, 1);
+            if (NARROW_MODE(ctx)) {
+                tcg_gen_ext32u_tl(temp, cpu_ctr);
+            } else {
+                tcg_gen_mov_tl(temp, cpu_ctr);
+            }
+            if (bo & 0x2) {
+                tcg_gen_brcondi_tl(TCG_COND_NE, temp, 0, l1);
+            } else {
+                tcg_gen_brcondi_tl(TCG_COND_EQ, temp, 0, l1);
+            }
+        }
+        bhrb_type |= BHRB_TYPE_COND;
+    }
+    if ((bo & 0x10) == 0) {
+        /* Test CR */
+        uint32_t mask = 0x08 >> (bi & 0x03);
+        TCGv_i32 temp = tcg_temp_new_i32();
+
+        if (bo & 0x8) {
+            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
+            tcg_gen_brcondi_i32(TCG_COND_EQ, temp, 0, l1);
+        } else {
+            tcg_gen_andi_i32(temp, cpu_crf[bi >> 2], mask);
+            tcg_gen_brcondi_i32(TCG_COND_NE, temp, 0, l1);
+        }
+        bhrb_type |= BHRB_TYPE_COND;
+    }
+
+    gen_update_branch_history(ctx, ctx->cia, target, bhrb_type);
+
+    if (type == BCOND_IM) {
+        target_ulong li = (target_long)((int16_t)(bd));
+        if (likely(aa == 0)) {
+            gen_goto_tb(ctx, 0, ctx->cia + li);
+        } else {
+            gen_goto_tb(ctx, 0, li);
+        }
+    } else {
+        if (NARROW_MODE(ctx)) {
+            tcg_gen_andi_tl(cpu_nip, target, (uint32_t)~3);
+        } else {
+            tcg_gen_andi_tl(cpu_nip, target, ~3);
+        }
+        gen_lookup_and_goto_ptr(ctx);
+    }
+    if ((bo & 0x14) != 0x14) {
+        /* fallthrough case */
+        gen_set_label(l1);
+        gen_goto_tb(ctx, 1, ctx->base.pc_next);
+    }
+    ctx->base.is_jmp = DISAS_NORETURN;
+
+    return true;
+}
+
+static bool trans_BC(DisasContext *ctx, arg_bcond *a)
+{
+    /*
+     * bh is not used for bc variants hence we pass -1
+     */
+    return bcond_helper(ctx, BCOND_IM, a->bo, a->bi, a->bd, -1, a->aa, a->lk);
+}
+
+/*
+ * This helper is shared by bclr, bcctr and bctar.
+ */
+static bool bclr_helper(DisasContext *ctx, arg_bclr *a, int type)
+{

target/ppc/translate/branch-impl.c.inc static bool trans_B(DisasContext *ctx, arg_I_b *a) · line 214

Quoted 10 lines
target/ppc/translate/branch-impl.c.inc +10 -0
+    /*
+     * bd and aa is not used for bc variants hence we pass -1 and 0 respectively
+     */
+    return bcond_helper(ctx, type, a->bo, a->bi, -1, a->bh, 0, a->lk);
+}
+
+TRANS(BCLR, bclr_helper, BCOND_LR)
+TRANS(BCCTR, bclr_helper, BCOND_CTR)
+TRANS_FLAGS2(BCTAR_ISA207, BCTAR, bclr_helper, BCOND_TAR)
+

This extra line causes the below warning on patch application. Please remove it.

warning: 1 line adds whitespace errors.

Quoted 3 lines
Signature · 1 line
2.55.0
#37 lore raw

PATCH v2 16/37 target/ppc: Fix TRANS* macro variadic arguments handling

In reply to Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Use ##__VA_ARGS__ in TRANS* macros to allow variadic arguments to be optional instead of mandatory. "##" removes the preceding comma when __VA_ARGS__ is empty, enabling macros to work with functions that don't need extra parameters.

This avoid compilation errors when using a pattern like below:

static bool do_wait(DisasContext *ctx, arg_X_wait *a) {...}

TRANS_FLAGS(WAIT, WAIT_ISA_2_X, do_wait)

Compilation Error:
../target/ppc/translate.c:5526:40: error: expected expression before ‘)’ token
5526 | return FUNC(ctx, a, __VA_ARGS__); \
| ^
../target/ppc/translate/storage-ctrl-impl.c.inc:368:1: note: in expansion of macro ‘TRANS_FLAGS’
368 | TRANS_FLAGS(WAIT, WAIT_ISA_2_X, do_wait)

---
 target/ppc/translate.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
target/ppc/translate.c +5 -5
index 7100787a25..f7c7f53d82 100644
@@ -4719,29 +4719,29 @@ static int64_t dw_compose_ea(DisasContext *ctx, int x)
  */
 #define TRANS(NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
-    { return FUNC(ctx, a, __VA_ARGS__); }
+    { return FUNC(ctx, a, ##__VA_ARGS__); }
 #define TRANS_FLAGS(FLAGS, NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     {                                                          \
         REQUIRE_INSNS_FLAGS(ctx, FLAGS);                       \
-        return FUNC(ctx, a, __VA_ARGS__);                      \
+        return FUNC(ctx, a, ##__VA_ARGS__);                    \
     }
 #define TRANS_FLAGS2(FLAGS2, NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     {                                                          \
         REQUIRE_INSNS_FLAGS2(ctx, FLAGS2);                     \
-        return FUNC(ctx, a, __VA_ARGS__);                      \
+        return FUNC(ctx, a, ##__VA_ARGS__);                    \
     }
 
 #define TRANS64(NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
-    { REQUIRE_64BIT(ctx); return FUNC(ctx, a, __VA_ARGS__); }
+    { REQUIRE_64BIT(ctx); return FUNC(ctx, a, ##__VA_ARGS__); }
 #define TRANS64_FLAGS2(FLAGS2, NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     {                                                          \
         REQUIRE_64BIT(ctx);                                    \
         REQUIRE_INSNS_FLAGS2(ctx, FLAGS2);                     \
-        return FUNC(ctx, a, __VA_ARGS__);                      \
+        return FUNC(ctx, a, ##__VA_ARGS__);                    \
     }
 
 /* TODO: More TRANS* helpers for extra insn_flags checks. */
Signature · 1 line
2.55.0
Reviewed-by on v2 16/37
#38 lore raw

Re: [PATCH v2 16/37] target/ppc: Fix TRANS* macro variadic arguments handling

Quoted 21 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Use ##__VA_ARGS__ in TRANS* macros to allow variadic arguments to be optional instead of mandatory. "##" removes the preceding comma when __VA_ARGS__ is empty, enabling macros to work with functions that don't need extra parameters.

This avoid compilation errors when using a pattern like below:

static bool do_wait(DisasContext *ctx, arg_X_wait *a) {...}

TRANS_FLAGS(WAIT, WAIT_ISA_2_X, do_wait)

Compilation Error:
../target/ppc/translate.c:5526:40: error: expected expression before ‘)’ token
5526 | return FUNC(ctx, a, __VA_ARGS__); \
| ^
../target/ppc/translate/storage-ctrl-impl.c.inc:368:1: note: in expansion of macro ‘TRANS_FLAGS’
368 | TRANS_FLAGS(WAIT, WAIT_ISA_2_X, do_wait)

Quoted 6 lines
---
 target/ppc/translate.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
Quoted 44 lines
target/ppc/translate.c +5 -5
index 7100787a25..f7c7f53d82 100644
@@ -4719,29 +4719,29 @@ static int64_t dw_compose_ea(DisasContext *ctx, int x)
  */
 #define TRANS(NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
-    { return FUNC(ctx, a, __VA_ARGS__); }
+    { return FUNC(ctx, a, ##__VA_ARGS__); }
 #define TRANS_FLAGS(FLAGS, NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     {                                                          \
         REQUIRE_INSNS_FLAGS(ctx, FLAGS);                       \
-        return FUNC(ctx, a, __VA_ARGS__);                      \
+        return FUNC(ctx, a, ##__VA_ARGS__);                    \
     }
 #define TRANS_FLAGS2(FLAGS2, NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     {                                                          \
         REQUIRE_INSNS_FLAGS2(ctx, FLAGS2);                     \
-        return FUNC(ctx, a, __VA_ARGS__);                      \
+        return FUNC(ctx, a, ##__VA_ARGS__);                    \
     }
 
 #define TRANS64(NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
-    { REQUIRE_64BIT(ctx); return FUNC(ctx, a, __VA_ARGS__); }
+    { REQUIRE_64BIT(ctx); return FUNC(ctx, a, ##__VA_ARGS__); }
 #define TRANS64_FLAGS2(FLAGS2, NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     {                                                          \
         REQUIRE_64BIT(ctx);                                    \
         REQUIRE_INSNS_FLAGS2(ctx, FLAGS2);                     \
-        return FUNC(ctx, a, __VA_ARGS__);                      \
+        return FUNC(ctx, a, ##__VA_ARGS__);                    \
     }
 
 /* TODO: More TRANS* helpers for extra insn_flags checks. */
Signature · 1 line
2.55.0
#39 lore raw

PATCH v2 17/37 target/ppc: Move wait instruction to decodetree

In reply to Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Implement wait instruction using decodetree with proper handling of WC and PL fields across different ISA versions. Since the opcode changed between ISA 2.x and 3.x we have defined 2 different decodetree instructions that are handled by the same helper.

The changes have been tested by comparing the tcg generated before and after the change for wait instruction. The test was repeated for 2 types of machines:

Power10 - which should use the newer wait implementation e500mc - which should use older PPC_WAIT implementation

Both generated the same tcg ops with and without the change. Further, with the changes we were able to boot into Fedora distro as well.

---
 target/ppc/insn32.decode                     |  15 +++
 target/ppc/translate.c                       |  91 -----------------
 target/ppc/translate/storage-ctrl-impl.c.inc | 100 +++++++++++++++++++
 3 files changed, 115 insertions(+), 91 deletions(-)
target/ppc/insn32.decode +15 -0
index 67fb181542..3b8387b60c 100644
@@ -1480,6 +1480,21 @@ ISYNC           010011 ----- ----- ----- 0010010110 -
 @XL_bfa         ...... bf:3 .. bfa:3 .. ..... ..... ..... .
 MCRF            010011 ... -- ... -- ----- 00000 00000 -       @XL_bfa
 
+# Wait Instructions
+
+# We can use the same X_wait argument set for both WAIT and older ISA2.x
+# compliant WAIT_2_x. Depending on the version, we may or may not use the
+# arguments. For more information, check do_wait() helper.
+
+&X_wait         wc pl
+@X_wait         ...... ..- wc:2 --- pl:2 ----- .......... -  &X_wait
+
+# According to ISA v3.1, bit 6 and 7 are implementation dependent and "unless
+# the intention is to use the implementation-dependent field, these bits must be
+# coded zero". Hence, we code them to 0.
+WAIT            011111 00-.. ---.. ----- 0000011110 -  @X_wait
+WAIT_ISA_2_X    011111 ---.. ---.. ----- 0000111110 -  @X_wait
+
 # Branch History Rolling Buffer (BHRB) Instructions
 
 &XFX_bhrbe      rt bhrbe
target/ppc/translate.c +0 -91

103 lines not on this page · raw message

target/ppc/translate/storage-ctrl-impl.c.inc +100 -0

103 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 17/37
#40 lore raw

Re: [PATCH v2 17/37] target/ppc: Move wait instruction to decodetree

Quoted 256 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Implement wait instruction using decodetree with proper handling of WC and PL fields across different ISA versions. Since the opcode changed between ISA 2.x and 3.x we have defined 2 different decodetree instructions that are handled by the same helper.

The changes have been tested by comparing the tcg generated before and after the change for wait instruction. The test was repeated for 2 types of machines:

Power10 - which should use the newer wait implementation e500mc - which should use older PPC_WAIT implementation

Both generated the same tcg ops with and without the change. Further, with the changes we were able to boot into Fedora distro as well.

---
 target/ppc/insn32.decode                     |  15 +++
 target/ppc/translate.c                       |  91 -----------------
 target/ppc/translate/storage-ctrl-impl.c.inc | 100 +++++++++++++++++++
 3 files changed, 115 insertions(+), 91 deletions(-)
target/ppc/insn32.decode +15 -0
index 67fb181542..3b8387b60c 100644
@@ -1480,6 +1480,21 @@ ISYNC           010011 ----- ----- ----- 0010010110 -
 @XL_bfa         ...... bf:3 .. bfa:3 .. ..... ..... ..... .
 MCRF            010011 ... -- ... -- ----- 00000 00000 -       @XL_bfa
 
+# Wait Instructions
+
+# We can use the same X_wait argument set for both WAIT and older ISA2.x
+# compliant WAIT_2_x. Depending on the version, we may or may not use the
+# arguments. For more information, check do_wait() helper.
+
+&X_wait         wc pl
+@X_wait         ...... ..- wc:2 --- pl:2 ----- .......... -  &X_wait
+
+# According to ISA v3.1, bit 6 and 7 are implementation dependent and "unless
+# the intention is to use the implementation-dependent field, these bits must be
+# coded zero". Hence, we code them to 0.
+WAIT            011111 00-.. ---.. ----- 0000011110 -  @X_wait
+WAIT_ISA_2_X    011111 ---.. ---.. ----- 0000111110 -  @X_wait
+
 # Branch History Rolling Buffer (BHRB) Instructions
 
 &XFX_bhrbe      rt bhrbe
target/ppc/translate.c +0 -91
index f7c7f53d82..2d5157080c 100644
@@ -2727,95 +2727,6 @@ static inline void gen_check_tlb_flush(DisasContext *ctx, bool global)
 static inline void gen_check_tlb_flush(DisasContext *ctx, bool global) { }
 #endif
 
-/* wait */
-static void gen_wait(DisasContext *ctx)
-{
-    uint32_t wc;
-
-    if (ctx->insns_flags & PPC_WAIT) {
-        /* v2.03-v2.07 define an older incompatible 'wait' encoding. */
-
-        if (ctx->insns_flags2 & PPC2_PM_ISA206) {
-            /* v2.06 introduced the WC field. WC > 0 may be treated as no-op. */
-            wc = WC(ctx->opcode);
-        } else {
-            wc = 0;
-        }
-
-    } else if (ctx->insns_flags2 & PPC2_ISA300) {
-        /* v3.0 defines a new 'wait' encoding. */
-        wc = WC(ctx->opcode);
-        if (ctx->insns_flags2 & PPC2_ISA310) {
-            uint32_t pl = PL(ctx->opcode);
-
-            /* WC 1,2 may be treated as no-op. WC 3 is reserved. */
-            if (wc == 3) {
-                gen_invalid(ctx);
-                return;
-            }
-
-            /* PL 1-3 are reserved. If WC=2 then the insn is treated as noop. */
-            if (pl > 0 && wc != 2) {
-                gen_invalid(ctx);
-                return;
-            }
-
-        } else { /* ISA300 */
-            /* WC 1-3 are reserved */
-            if (wc > 0) {
-                gen_invalid(ctx);
-                return;
-            }
-        }
-
-    } else {
-        warn_report("wait instruction decoded with wrong ISA flags.");
-        gen_invalid(ctx);
-        return;
-    }
-
-    /*
-     * wait without WC field or with WC=0 waits for an exception / interrupt
-     * to occur.
-     */
-    if (wc == 0) {
-        TCGv_i32 t0 = tcg_constant_i32(1);
-        tcg_gen_st_i32(t0, tcg_env,
-                       -offsetof(PowerPCCPU, env) + offsetof(CPUState, halted));
-        /* Stop translation, as the CPU is supposed to sleep from now */
-        gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-    }
-
-    /*
-     * Other wait types must not just wait until an exception occurs because
-     * ignoring their other wake-up conditions could cause a hang.
-     *
-     * For v2.06 and 2.07, wc=1,2,3 are architected but may be implemented as
-     * no-ops.
-     *
-     * wc=1 and wc=3 explicitly allow the instruction to be treated as a no-op.
-     *
-     * wc=2 waits for an implementation-specific condition, such could be
-     * always true, so it can be implemented as a no-op.
-     *
-     * For v3.1, wc=1,2 are architected but may be implemented as no-ops.
-     *
-     * wc=1 (waitrsv) waits for an exception or a reservation to be lost.
-     * Reservation-loss may have implementation-specific conditions, so it
-     * can be implemented as a no-op.
-     *
-     * wc=2 waits for an exception or an amount of time to pass. This
-     * amount is implementation-specific so it can be implemented as a
-     * no-op.
-     *
-     * ISA v3.1 allows for execution to resume "in the rare case of
-     * an implementation-dependent event", so in any case software must
-     * not depend on the architected resumption condition to become
-     * true, so no-op implementations should be architecturally correct
-     * (if suboptimal).
-     */
-}
-
 #if defined(TARGET_PPC64)
 static void gen_doze(DisasContext *ctx)
 {
@@ -5197,8 +5108,6 @@ GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
 GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
-GEN_HANDLER(wait, 0x1F, 0x1E, 0x01, 0x039FF801, PPC_WAIT),
-GEN_HANDLER_E(wait, 0x1F, 0x1E, 0x00, 0x039CF801, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(rfid, 0x13, 0x12, 0x00, 0x03FF8001, PPC_64B),
target/ppc/translate/storage-ctrl-impl.c.inc +85 -0
index 943229c438..21c7999d5d 100644
@@ -270,3 +270,103 @@ static bool do_tlbie(DisasContext *ctx, arg_X_tlbie *a, bool local)
 
 TRANS_FLAGS(MEM_TLBIE, TLBIE, do_tlbie, false)
 TRANS_FLAGS(MEM_TLBIE, TLBIEL, do_tlbie, true)
+
+/*
+ * Decodetree populates a->wc and a->pl based on ISA v3.1, however they may
+ * or may not be used based on the ISA:
+ *   For ISA < 2.06 - both wc and pl are ignored
+ *   For ISA == 2.06, 2.07, 3.0 - wc is used and pl is ignored
+ *   For ISA 3.1 - both wc and pl are used
+ */
+static bool do_wait(DisasContext *ctx, arg_X_wait *a)
+{
+    uint32_t wc;
+
+    if (ctx->insns_flags & PPC_WAIT) {
+        /* v2.03-v2.07 define an older incompatible 'wait' encoding. */
+
+        if (ctx->insns_flags2 & PPC2_PM_ISA206) {
+            /* v2.06 introduced the WC field. WC > 0 may be treated as no-op. */
+            wc = a->wc;
+        } else {
+            wc = 0;
+        }
+
+    } else if (ctx->insns_flags2 & PPC2_ISA300) {
+        /* v3.0 defines a new 'wait' encoding. */
+        wc = a->wc;
+        if (ctx->insns_flags2 & PPC2_ISA310) {
+            uint32_t pl = a->pl;
+
+            /* WC 1,2 may be treated as no-op. WC 3 is reserved. */
+            if (wc == 3) {
+                gen_invalid(ctx);
+                return true;
+            }
+
+            /* PL 1-3 are reserved. If WC=2 then the insn is treated as noop. */
+            if (pl > 0 && wc != 2) {
+                gen_invalid(ctx);
+                return true;
+            }
+
+        } else { /* ISA300 */
+            /* WC 1-3 are reserved */
+            if (wc > 0) {
+                gen_invalid(ctx);
+                return true;
+            }
+        }
+
+    } else {
+        warn_report("wait instruction decoded with wrong ISA flags.");
+        gen_invalid(ctx);
+        return true;
+    }
+
+    /*
+     * wait without WC field or with WC=0 waits for an exception / interrupt
+     * to occur.
+     */
+    if (wc == 0) {
+        TCGv_i32 t0 = tcg_constant_i32(1);
+        tcg_gen_st_i32(t0, tcg_env,
+                       -offsetof(PowerPCCPU, env) + offsetof(CPUState, halted));
+        /* Stop translation, as the CPU is supposed to sleep from now */
+        gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+    }
+
+    /*
+     * Other wait types must not just wait until an exception occurs because
+     * ignoring their other wake-up conditions could cause a hang.
+     *
+     * For v2.06 and 2.07, wc=1,2,3 are architected but may be implemented as
+     * no-ops.
+     *
+     * wc=1 and wc=3 explicitly allow the instruction to be treated as a no-op.
+     *
+     * wc=2 waits for an implementation-specific condition, such could be
+     * always true, so it can be implemented as a no-op.
+     *
+     * For v3.1, wc=1,2 are architected but may be implemented as no-ops.
+     *
+     * wc=1 (waitrsv) waits for an exception or a reservation to be lost.
+     * Reservation-loss may have implementation-specific conditions, so it
+     * can be implemented as a no-op.
+     *
+     * wc=2 waits for an exception or an amount of time to pass. This

target/ppc/translate/storage-ctrl-impl.c.inc static bool do_tlbie(DisasContext *ctx, arg_X_tlbie *a, bool local) · line 367

Quoted 10 lines
target/ppc/translate/storage-ctrl-impl.c.inc +10 -0
+     * amount is implementation-specific so it can be implemented as a
+     * no-op.
+     *
+     * ISA v3.1 allows for execution to resume "in the rare case of
+     * an implementation-dependent event", so in any case software must
+     * not depend on the architected resumption condition to become
+     * true, so no-op implementations should be architecturally correct
+     * (if suboptimal).
+     */
+     return true;

The return true is aligned under the comment's */ instead of the function body indent level.

target/ppc/translate/storage-ctrl-impl.c.inc static bool do_tlbie(DisasContext *ctx, arg_X_tlbie *a, bool local) · line 372

Quoted 5 lines
target/ppc/translate/storage-ctrl-impl.c.inc +5 -0
+}
+
+TRANS_FLAGS(WAIT, WAIT_ISA_2_X, do_wait)
+TRANS_FLAGS2(ISA300, WAIT, do_wait)
+

This extra new line needs to be dropped to avoid git warning.

Rest looks good to me.

Quoted 3 lines
Signature · 1 line
2.55.0
#41 lore raw

PATCH v2 18/37 target/ppc: Move sleep & friends to decodetree

In reply to Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Convert the following instructions to decode tree specification:

doze
nap
sleep
rvwinkle
stop

The functionality was tested by comparing the qemu -D log -d op,in_asm output as well as single stepping gdb to confirm behavior was not changed for ISA3.x (stop) as well as ISA206 (doze, nap, sleep, rvwinkle).

---
 target/ppc/insn32.decode                      |   8 ++
 target/ppc/translate.c                        |  85 ---------------
 .../ppc/translate/processor-ctrl-impl.c.inc   | 101 ++++++++++++++++++
 3 files changed, 109 insertions(+), 85 deletions(-)
target/ppc/insn32.decode +8 -0
index 3b8387b60c..d23168a303 100644
@@ -913,6 +913,14 @@ VPERMR          000100 ..... ..... ..... ..... 111011   @VA
 
 VSEL            000100 ..... ..... ..... ..... 101010   @VA
 
+## Power Management Instructions
+
+DOZE            010011 ----- ----- ----- 0110010010 -
+NAP             010011 ----- ----- ----- 0110110010 -
+SLEEP           010011 ----- ----- ----- 0111010010 -
+RVWINKLE        010011 ----- ----- ----- 0111110010 -
+STOP            010011 ----- ----- ----- 0101110010 -
+
 ## Vector Integer Shift Instruction
 
 VSLB            000100 ..... ..... ..... 00100000100    @VX
target/ppc/translate.c +0 -85

97 lines not on this page · raw message

target/ppc/translate/processor-ctrl-impl.c.inc +101 -0

104 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 18/37
#42 lore raw

Re: [PATCH v2 18/37] target/ppc: Move sleep & friends to decodetree

Quoted 162 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

Convert the following instructions to decode tree specification:

doze
nap
sleep
rvwinkle
stop

The functionality was tested by comparing the qemu -D log -d op,in_asm output as well as single stepping gdb to confirm behavior was not changed for ISA3.x (stop) as well as ISA206 (doze, nap, sleep, rvwinkle).

---
 target/ppc/insn32.decode                      |   8 ++
 target/ppc/translate.c                        |  85 ---------------
 .../ppc/translate/processor-ctrl-impl.c.inc   | 101 ++++++++++++++++++
 3 files changed, 109 insertions(+), 85 deletions(-)
target/ppc/insn32.decode +8 -0
index 3b8387b60c..d23168a303 100644
@@ -913,6 +913,14 @@ VPERMR          000100 ..... ..... ..... ..... 111011   @VA
 
 VSEL            000100 ..... ..... ..... ..... 101010   @VA
 
+## Power Management Instructions
+
+DOZE            010011 ----- ----- ----- 0110010010 -
+NAP             010011 ----- ----- ----- 0110110010 -
+SLEEP           010011 ----- ----- ----- 0111010010 -
+RVWINKLE        010011 ----- ----- ----- 0111110010 -
+STOP            010011 ----- ----- ----- 0101110010 -
+
 ## Vector Integer Shift Instruction
 
 VSLB            000100 ..... ..... ..... 00100000100    @VX
target/ppc/translate.c +0 -85
index 2d5157080c..f6f7bb4e6b 100644
@@ -2728,86 +2728,6 @@ static inline void gen_check_tlb_flush(DisasContext *ctx, bool global) { }
 #endif
 
 #if defined(TARGET_PPC64)
-static void gen_doze(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv_i32 t;
-
-    CHK_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_DOZE);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-static void gen_nap(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv_i32 t;
-
-    CHK_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_NAP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-static void gen_stop(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv_i32 t;
-
-    CHK_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_STOP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-static void gen_sleep(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv_i32 t;
-
-    CHK_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_SLEEP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
-static void gen_rvwinkle(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    TCGv_i32 t;
-
-    CHK_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_RVWINKLE);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-}
-
 static inline TCGv gen_write_bhrb(TCGv_ptr base, TCGv offset, TCGv mask, TCGv value)
 {
     TCGv_ptr tmp = tcg_temp_new_ptr();
@@ -5117,11 +5037,6 @@ GEN_HANDLER_E(scv, 0x11, 0x10, 0xFF, 0x03FFF01E, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(scv, 0x11, 0x00, 0xFF, 0x03FFF01E, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(rfscv, 0x13, 0x12, 0x02, 0x03FF8001, PPC_NONE, PPC2_ISA300),
 #endif
-GEN_HANDLER_E(stop, 0x13, 0x12, 0x0b, 0x03FFF801, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER_E(doze, 0x13, 0x12, 0x0c, 0x03FFF801, PPC_NONE, PPC2_PM_ISA206),
-GEN_HANDLER_E(nap, 0x13, 0x12, 0x0d, 0x03FFF801, PPC_NONE, PPC2_PM_ISA206),
-GEN_HANDLER_E(sleep, 0x13, 0x12, 0x0e, 0x03FFF801, PPC_NONE, PPC2_PM_ISA206),
-GEN_HANDLER_E(rvwinkle, 0x13, 0x12, 0x0f, 0x03FFF801, PPC_NONE, PPC2_PM_ISA206),
 GEN_HANDLER(hrfid, 0x13, 0x12, 0x08, 0x03FF8001, PPC_64H),
 #endif
 /* Top bit of opc2 corresponds with low bit of LEV, so use two handlers */
target/ppc/translate/processor-ctrl-impl.c.inc +8 -0
index 8abbb89630..bdbc195294 100644
@@ -103,3 +103,104 @@ static bool trans_MSGSYNC(DisasContext *ctx, arg_MSGSYNC *a)
     /* interpreted as no-op */
     return true;
 }
+
+static bool do_doze(DisasContext *ctx, arg_DOZE *a)
+{
+    REQUIRE_64BIT(ctx);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+#else

target/ppc/translate/processor-ctrl-impl.c.inc static bool trans_MSGSYNC(DisasContext *ctx, arg_MSGSYNC *a) · line 123

Quoted 10 lines
target/ppc/translate/processor-ctrl-impl.c.inc +10 -0
+    TCGv_i32 t;
+
+    REQUIRE_HV(ctx);
+    translator_io_start(&ctx->base);
+    t = tcg_constant_i32(PPC_PM_DOZE);
+    gen_helper_pminsn(tcg_env, t);
+    /* Stop translation, as the CPU is supposed to sleep from now */
+    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+#endif /* defined(CONFIG_USER_ONLY) */
+

Nit: Inconsistent blank line

Quoted 70 lines
target/ppc/translate/processor-ctrl-impl.c.inc +70 -0
+    return true;
+}
+TRANS_FLAGS2(PM_ISA206, DOZE, do_doze);
+
+static bool do_nap(DisasContext *ctx, arg_NAP *a)
+{
+    REQUIRE_64BIT(ctx);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+#else
+    TCGv_i32 t;
+
+    REQUIRE_HV(ctx);
+    translator_io_start(&ctx->base);
+    t = tcg_constant_i32(PPC_PM_NAP);
+    gen_helper_pminsn(tcg_env, t);
+    /* Stop translation, as the CPU is supposed to sleep from now */
+    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+#endif /* defined(CONFIG_USER_ONLY) */
+    return true;
+}
+TRANS_FLAGS2(PM_ISA206, NAP, do_nap);
+
+static bool do_sleep(DisasContext *ctx, arg_SLEEP *a)
+{
+    REQUIRE_64BIT(ctx);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+#else
+    TCGv_i32 t;
+
+    REQUIRE_HV(ctx);
+    translator_io_start(&ctx->base);
+    t = tcg_constant_i32(PPC_PM_SLEEP);
+    gen_helper_pminsn(tcg_env, t);
+    /* Stop translation, as the CPU is supposed to sleep from now */
+    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+#endif /* defined(CONFIG_USER_ONLY) */
+    return true;
+}
+TRANS_FLAGS2(PM_ISA206, SLEEP, do_sleep);
+
+static bool do_rvwinkle(DisasContext *ctx, arg_RVWINKLE *a)
+{
+    REQUIRE_64BIT(ctx);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+#else
+    TCGv_i32 t;
+
+    REQUIRE_HV(ctx);
+    translator_io_start(&ctx->base);
+    t = tcg_constant_i32(PPC_PM_RVWINKLE);
+    gen_helper_pminsn(tcg_env, t);
+    /* Stop translation, as the CPU is supposed to sleep from now */
+    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+#endif /* defined(CONFIG_USER_ONLY) */
+    return true;
+}
+TRANS_FLAGS2(PM_ISA206, RVWINKLE, do_rvwinkle);
+
+static bool do_stop(DisasContext *ctx, arg_STOP *a)
+{
+    REQUIRE_64BIT(ctx);
+
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);

target/ppc/translate/processor-ctrl-impl.c.inc static bool trans_MSGSYNC(DisasContext *ctx, arg_MSGSYNC *a) · line 203

Quoted 10 lines
target/ppc/translate/processor-ctrl-impl.c.inc +10 -0
+#else
+    TCGv_i32 t;
+
+    REQUIRE_HV(ctx);
+    translator_io_start(&ctx->base);
+    t = tcg_constant_i32(PPC_PM_STOP);
+    gen_helper_pminsn(tcg_env, t);
+    /* Stop translation, as the CPU is supposed to sleep from now */
+    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+#endif

Nit: Can be replaced with:

#endif /* defined(CONFIG_USER_ONLY) */

..consistent with other 3 patches and for better readability. Rest LGTM.

Quoted 6 lines
target/ppc/translate/processor-ctrl-impl.c.inc +3 -0
+    return true;
+}
+TRANS_FLAGS2(ISA300, STOP, do_stop);
Signature · 1 line
2.55.0
#43 lore raw

PATCH v2 19/37 target/ppc: Refactor sleep and its variants to use a common helper

In reply to Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

With sleep, doze, nap, rvwinkle and stop instructions moved to decodetree, the helper these instructions used : gen_helper_pminsn, should now be renamed to gen_helper_PMINSN, abiding by the decodetree standards. Do the same.
Also refactor their translation routine to use a common do_sleep routine since they share majority of the code.

---
 target/ppc/helper.h                           |  2 +-
 target/ppc/tcg-excp_helper.c                  |  2 +-
 .../ppc/translate/processor-ctrl-impl.c.inc   | 99 +++----------------
 3 files changed, 16 insertions(+), 87 deletions(-)
target/ppc/helper.h +1 -1
index cd2a0948a3..7e15676122 100644
@@ -18,7 +18,7 @@ DEF_HELPER_1(rfdi, void, env)
 DEF_HELPER_1(rfmci, void, env)
 #if defined(TARGET_PPC64)
 DEF_HELPER_2(scv, noreturn, env, i32)
-DEF_HELPER_2(pminsn, void, env, i32)
+DEF_HELPER_2(PMINSN, void, env, i32)
 DEF_HELPER_1(rfid, void, env)
 DEF_HELPER_1(rfscv, void, env)
 DEF_HELPER_1(hrfid, void, env)
target/ppc/tcg-excp_helper.c +1 -1
index b04f07a637..c4ffa2dfbb 100644
@@ -459,7 +459,7 @@ void helper_scv(CPUPPCState *env, uint32_t lev)
     }
 }
 
-void helper_pminsn(CPUPPCState *env, uint32_t insn)
+void helper_PMINSN(CPUPPCState *env, uint32_t insn)
 {
     CPUState *cs = env_cpu(env);
 
target/ppc/translate/processor-ctrl-impl.c.inc +14 -85

117 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 19/37
#44 lore raw

Re: [PATCH v2 19/37] target/ppc: Refactor sleep and its variants to use a common helper

Quoted 9 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

With sleep, doze, nap, rvwinkle and stop instructions moved to decodetree, the helper these instructions used : gen_helper_pminsn, should now be renamed to gen_helper_PMINSN, abiding by the decodetree standards. Do the same.
Also refactor their translation routine to use a common do_sleep routine since they share majority of the code.

Quoted 7 lines
---
 target/ppc/helper.h                           |  2 +-
 target/ppc/tcg-excp_helper.c                  |  2 +-
 .../ppc/translate/processor-ctrl-impl.c.inc   | 99 +++----------------
 3 files changed, 16 insertions(+), 87 deletions(-)
Quoted 152 lines
target/ppc/helper.h +1 -1
index cd2a0948a3..7e15676122 100644
@@ -18,7 +18,7 @@ DEF_HELPER_1(rfdi, void, env)
 DEF_HELPER_1(rfmci, void, env)
 #if defined(TARGET_PPC64)
 DEF_HELPER_2(scv, noreturn, env, i32)
-DEF_HELPER_2(pminsn, void, env, i32)
+DEF_HELPER_2(PMINSN, void, env, i32)
 DEF_HELPER_1(rfid, void, env)
 DEF_HELPER_1(rfscv, void, env)
 DEF_HELPER_1(hrfid, void, env)
target/ppc/tcg-excp_helper.c +1 -1
index b04f07a637..c4ffa2dfbb 100644
@@ -459,7 +459,7 @@ void helper_scv(CPUPPCState *env, uint32_t lev)
     }
 }
 
-void helper_pminsn(CPUPPCState *env, uint32_t insn)
+void helper_PMINSN(CPUPPCState *env, uint32_t insn)
 {
     CPUState *cs = env_cpu(env);
 
target/ppc/translate/processor-ctrl-impl.c.inc +14 -85
index bdbc195294..3b3ed3019a 100644
@@ -104,103 +104,32 @@ static bool trans_MSGSYNC(DisasContext *ctx, arg_MSGSYNC *a)
     return true;
 }
 
-static bool do_doze(DisasContext *ctx, arg_DOZE *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
-#else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_DOZE);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, DOZE, do_doze);
-
-static bool do_nap(DisasContext *ctx, arg_NAP *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
-#else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_NAP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, NAP, do_nap);
-
-static bool do_sleep(DisasContext *ctx, arg_SLEEP *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
-#else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_SLEEP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, SLEEP, do_sleep);
-
-static bool do_rvwinkle(DisasContext *ctx, arg_RVWINKLE *a)
+/*
+ * Helper to handle DOZE, NAP, SLEEP, RVWINKLE & STOP. Since none of them use
+ * any arguments just use a placeholder arg_SLEEP.
+ */
+static bool do_sleep(DisasContext *ctx, arg_SLEEP *a, powerpc_pm_insn_t type)
 {
     REQUIRE_64BIT(ctx);
 
 #if defined(CONFIG_USER_ONLY)
     gen_priv_opc(ctx);
-#else
+#elif defined(TARGET_PPC64)
     TCGv_i32 t;
 
     REQUIRE_HV(ctx);
     translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_RVWINKLE);
-    gen_helper_pminsn(tcg_env, t);
+    t = tcg_constant_i32(type);
+    gen_helper_PMINSN(tcg_env, t);
     /* Stop translation, as the CPU is supposed to sleep from now */
     gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, RVWINKLE, do_rvwinkle);
-
-static bool do_stop(DisasContext *ctx, arg_STOP *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
 #else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_STOP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+    qemu_build_not_reached();
 #endif
     return true;
 }
-TRANS_FLAGS2(ISA300, STOP, do_stop);
+TRANS_FLAGS2(PM_ISA206, DOZE, do_sleep, PPC_PM_DOZE);
+TRANS_FLAGS2(PM_ISA206, NAP, do_sleep, PPC_PM_NAP);
+TRANS_FLAGS2(PM_ISA206, SLEEP, do_sleep, PPC_PM_SLEEP);
+TRANS_FLAGS2(PM_ISA206, RVWINKLE, do_sleep, PPC_PM_RVWINKLE);
+TRANS_FLAGS2(ISA300, STOP, do_sleep, PPC_PM_STOP);
Signature · 1 line
2.55.0
Reviewed-by on v2 19/37
#45 lore raw

Re: [PATCH v2 19/37] target/ppc: Refactor sleep and its variants to use a common helper

In reply to Chinmay Rath

Thanks,

Glenn

Quoted 165 lines · Chinmay Rath

From: Ojaswin Mujoo <[email protected]>

With sleep, doze, nap, rvwinkle and stop instructions moved to decodetree, the helper these instructions used : gen_helper_pminsn, should now be renamed to gen_helper_PMINSN, abiding by the decodetree standards. Do the same.
Also refactor their translation routine to use a common do_sleep routine since they share majority of the code.

---
 target/ppc/helper.h                           |  2 +-
 target/ppc/tcg-excp_helper.c                  |  2 +-
 .../ppc/translate/processor-ctrl-impl.c.inc   | 99 +++----------------
 3 files changed, 16 insertions(+), 87 deletions(-)
target/ppc/helper.h +1 -1
index cd2a0948a3..7e15676122 100644
@@ -18,7 +18,7 @@ DEF_HELPER_1(rfdi, void, env)
 DEF_HELPER_1(rfmci, void, env)
 #if defined(TARGET_PPC64)
 DEF_HELPER_2(scv, noreturn, env, i32)
-DEF_HELPER_2(pminsn, void, env, i32)
+DEF_HELPER_2(PMINSN, void, env, i32)
 DEF_HELPER_1(rfid, void, env)
 DEF_HELPER_1(rfscv, void, env)
 DEF_HELPER_1(hrfid, void, env)
target/ppc/tcg-excp_helper.c +1 -1
index b04f07a637..c4ffa2dfbb 100644
@@ -459,7 +459,7 @@ void helper_scv(CPUPPCState *env, uint32_t lev)
     }
 }
 
-void helper_pminsn(CPUPPCState *env, uint32_t insn)
+void helper_PMINSN(CPUPPCState *env, uint32_t insn)
 {
     CPUState *cs = env_cpu(env);
 
target/ppc/translate/processor-ctrl-impl.c.inc +14 -85
index bdbc195294..3b3ed3019a 100644
@@ -104,103 +104,32 @@ static bool trans_MSGSYNC(DisasContext *ctx, arg_MSGSYNC *a)
     return true;
 }
 
-static bool do_doze(DisasContext *ctx, arg_DOZE *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
-#else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_DOZE);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, DOZE, do_doze);
-
-static bool do_nap(DisasContext *ctx, arg_NAP *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
-#else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_NAP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, NAP, do_nap);
-
-static bool do_sleep(DisasContext *ctx, arg_SLEEP *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
-#else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_SLEEP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, SLEEP, do_sleep);
-
-static bool do_rvwinkle(DisasContext *ctx, arg_RVWINKLE *a)
+/*
+ * Helper to handle DOZE, NAP, SLEEP, RVWINKLE & STOP. Since none of them use
+ * any arguments just use a placeholder arg_SLEEP.
+ */
+static bool do_sleep(DisasContext *ctx, arg_SLEEP *a, powerpc_pm_insn_t type)
 {
     REQUIRE_64BIT(ctx);
 
 #if defined(CONFIG_USER_ONLY)
     gen_priv_opc(ctx);
-#else
+#elif defined(TARGET_PPC64)
     TCGv_i32 t;
 
     REQUIRE_HV(ctx);
     translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_RVWINKLE);
-    gen_helper_pminsn(tcg_env, t);
+    t = tcg_constant_i32(type);
+    gen_helper_PMINSN(tcg_env, t);
     /* Stop translation, as the CPU is supposed to sleep from now */
     gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
-#endif /* defined(CONFIG_USER_ONLY) */
-    return true;
-}
-TRANS_FLAGS2(PM_ISA206, RVWINKLE, do_rvwinkle);
-
-static bool do_stop(DisasContext *ctx, arg_STOP *a)
-{
-    REQUIRE_64BIT(ctx);
-
-#if defined(CONFIG_USER_ONLY)
-    gen_priv_opc(ctx);
 #else
-    TCGv_i32 t;
-
-    REQUIRE_HV(ctx);
-    translator_io_start(&ctx->base);
-    t = tcg_constant_i32(PPC_PM_STOP);
-    gen_helper_pminsn(tcg_env, t);
-    /* Stop translation, as the CPU is supposed to sleep from now */
-    gen_exception_nip(ctx, EXCP_HLT, ctx->base.pc_next);
+    qemu_build_not_reached();
 #endif
     return true;
 }
-TRANS_FLAGS2(ISA300, STOP, do_stop);
+TRANS_FLAGS2(PM_ISA206, DOZE, do_sleep, PPC_PM_DOZE);
+TRANS_FLAGS2(PM_ISA206, NAP, do_sleep, PPC_PM_NAP);
+TRANS_FLAGS2(PM_ISA206, SLEEP, do_sleep, PPC_PM_SLEEP);
+TRANS_FLAGS2(PM_ISA206, RVWINKLE, do_sleep, PPC_PM_RVWINKLE);
+TRANS_FLAGS2(ISA300, STOP, do_sleep, PPC_PM_STOP);
#46 lore raw

PATCH v2 20/37 target/ppc: Move Condition Register access instructions to decodetree.

In reply to Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

mfcr, mfocr,
mtcrf, mtocrf : XFX-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/insn32.decode                   |  8 +++
 target/ppc/translate.c                     | 62 -----------------
 target/ppc/translate/fixedpoint-impl.c.inc | 79 ++++++++++++++++++++++
 3 files changed, 87 insertions(+), 62 deletions(-)
target/ppc/insn32.decode +8 -0
index d23168a303..dc708b5442 100644
@@ -70,6 +70,9 @@
 %dx_d           6:s10 16:5 0:1
 @DX             ...... rt:5  ..... .......... ..... .           &DX d=%dx_d
 
+&XFX            rt fxm
+@XFX            ...... rt:5  . fxm:8 ...... ..... .             &XFX
+
 %md_sh          1:1 11:5
 %md_mb          5:1 6:5
 &MD             rs ra sh mb rc
@@ -648,6 +651,11 @@ FSEL            111111 ..... ..... ..... ..... 10111 .  @A
 
 ### Move To/From System Register Instructions
 
+MFCR            011111 ..... 0.... ....- 0000010011 -   @XFX
+MFOCR           011111 ..... 1.... ....- 0000010011 -   @XFX
+MTCRF           011111 ..... 0.... ....- 0010010000 -   @XFX
+MTOCRF          011111 ..... 1.... ....- 0010010000 -   @XFX
+
 SETBC           011111 ..... ..... ----- 0110000000 -   @X_bi
 SETBCR          011111 ..... ..... ----- 0110100000 -   @X_bi
 SETNBC          011111 ..... ..... ----- 0111000000 -   @X_bi
target/ppc/translate.c +0 -62

83 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +79 -0

85 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 20/37
#47 lore raw

Re: [PATCH v2 20/37] target/ppc: Move Condition Register access instructions to decodetree.

Quoted 141 lines · Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

mfcr, mfocr,
mtcrf, mtocrf : XFX-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/insn32.decode                   |  8 +++
 target/ppc/translate.c                     | 62 -----------------
 target/ppc/translate/fixedpoint-impl.c.inc | 79 ++++++++++++++++++++++
 3 files changed, 87 insertions(+), 62 deletions(-)
target/ppc/insn32.decode +8 -0
index d23168a303..dc708b5442 100644
@@ -70,6 +70,9 @@
 %dx_d           6:s10 16:5 0:1
 @DX             ...... rt:5  ..... .......... ..... .           &DX d=%dx_d
 
+&XFX            rt fxm
+@XFX            ...... rt:5  . fxm:8 ...... ..... .             &XFX
+
 %md_sh          1:1 11:5
 %md_mb          5:1 6:5
 &MD             rs ra sh mb rc
@@ -648,6 +651,11 @@ FSEL            111111 ..... ..... ..... ..... 10111 .  @A
 
 ### Move To/From System Register Instructions
 
+MFCR            011111 ..... 0.... ....- 0000010011 -   @XFX
+MFOCR           011111 ..... 1.... ....- 0000010011 -   @XFX
+MTCRF           011111 ..... 0.... ....- 0010010000 -   @XFX
+MTOCRF          011111 ..... 1.... ....- 0010010000 -   @XFX
+
 SETBC           011111 ..... ..... ----- 0110000000 -   @X_bi
 SETBCR          011111 ..... ..... ----- 0110100000 -   @X_bi
 SETNBC          011111 ..... ..... ----- 0111000000 -   @X_bi
target/ppc/translate.c +0 -62
index f6f7bb4e6b..d01981de73 100644
@@ -3131,40 +3131,6 @@ static void gen_mcrxrx(DisasContext *ctx)
 }
 #endif
 
-/* mfcr mfocrf */
-static void gen_mfcr(DisasContext *ctx)
-{
-    uint32_t crm, crn;
-
-    if (likely(ctx->opcode & 0x00100000)) {
-        crm = CRM(ctx->opcode);
-        if (likely(crm && ((crm & (crm - 1)) == 0))) {
-            crn = ctz32(crm);
-            tcg_gen_extu_i32_tl(cpu_gpr[rD(ctx->opcode)], cpu_crf[7 - crn]);
-            tcg_gen_shli_tl(cpu_gpr[rD(ctx->opcode)],
-                            cpu_gpr[rD(ctx->opcode)], crn * 4);
-        }
-    } else {
-        TCGv_i32 t0 = tcg_temp_new_i32();
-        tcg_gen_mov_i32(t0, cpu_crf[0]);
-        tcg_gen_shli_i32(t0, t0, 4);
-        tcg_gen_or_i32(t0, t0, cpu_crf[1]);
-        tcg_gen_shli_i32(t0, t0, 4);
-        tcg_gen_or_i32(t0, t0, cpu_crf[2]);
-        tcg_gen_shli_i32(t0, t0, 4);
-        tcg_gen_or_i32(t0, t0, cpu_crf[3]);
-        tcg_gen_shli_i32(t0, t0, 4);
-        tcg_gen_or_i32(t0, t0, cpu_crf[4]);
-        tcg_gen_shli_i32(t0, t0, 4);
-        tcg_gen_or_i32(t0, t0, cpu_crf[5]);
-        tcg_gen_shli_i32(t0, t0, 4);
-        tcg_gen_or_i32(t0, t0, cpu_crf[6]);
-        tcg_gen_shli_i32(t0, t0, 4);
-        tcg_gen_or_i32(t0, t0, cpu_crf[7]);
-        tcg_gen_extu_i32_tl(cpu_gpr[rD(ctx->opcode)], t0);
-    }
-}
-
 /* mfmsr */
 static void gen_mfmsr(DisasContext *ctx)
 {
@@ -3245,32 +3211,6 @@ static void gen_mftb(DisasContext *ctx)
     gen_op_mfspr(ctx);
 }
 
-/* mtcrf mtocrf*/
-static void gen_mtcrf(DisasContext *ctx)
-{
-    uint32_t crm, crn;
-
-    crm = CRM(ctx->opcode);
-    if (likely((ctx->opcode & 0x00100000))) {
-        if (crm && ((crm & (crm - 1)) == 0)) {
-            TCGv_i32 temp = tcg_temp_new_i32();
-            crn = ctz32(crm);
-            tcg_gen_trunc_tl_i32(temp, cpu_gpr[rS(ctx->opcode)]);
-            tcg_gen_shri_i32(temp, temp, crn * 4);
-            tcg_gen_andi_i32(cpu_crf[7 - crn], temp, 0xf);
-        }
-    } else {
-        TCGv_i32 temp = tcg_temp_new_i32();
-        tcg_gen_trunc_tl_i32(temp, cpu_gpr[rS(ctx->opcode)]);
-        for (crn = 0 ; crn < 8 ; crn++) {
-            if (crm & (1 << crn)) {
-                    tcg_gen_shri_i32(cpu_crf[7 - crn], temp, crn * 4);
-                    tcg_gen_andi_i32(cpu_crf[7 - crn], cpu_crf[7 - crn], 0xf);
-            }
-        }
-    }
-}
-
 /* mtmsr */
 #if defined(TARGET_PPC64)
 static void gen_mtmsrd(DisasContext *ctx)
@@ -5043,11 +4983,9 @@ GEN_HANDLER(hrfid, 0x13, 0x12, 0x08, 0x03FF8001, PPC_64H),
 GEN_HANDLER(sc, 0x11, 0x11, 0xFF, 0x03FFF01D, PPC_FLOW),
 GEN_HANDLER(sc, 0x11, 0x01, 0xFF, 0x03FFF01D, PPC_FLOW),
 GEN_HANDLER(mcrxr, 0x1F, 0x00, 0x10, 0x007FF801, PPC_MISC),
-GEN_HANDLER(mfcr, 0x1F, 0x13, 0x00, 0x00000801, PPC_MISC),
 GEN_HANDLER(mfmsr, 0x1F, 0x13, 0x02, 0x001FF801, PPC_MISC),
 GEN_HANDLER(mfspr, 0x1F, 0x13, 0x0A, 0x00000001, PPC_MISC),
 GEN_HANDLER(mftb, 0x1F, 0x13, 0x0B, 0x00000001, PPC_MFTB),
-GEN_HANDLER(mtcrf, 0x1F, 0x10, 0x04, 0x00000801, PPC_MISC),
 #if defined(TARGET_PPC64)
 GEN_HANDLER(mtmsrd, 0x1F, 0x12, 0x05, 0x001EF801, PPC_64B),
 GEN_HANDLER_E(setb, 0x1F, 0x00, 0x04, 0x0003F801, PPC_NONE, PPC2_ISA300),
target/ppc/translate/fixedpoint-impl.c.inc +0 -0
index 4d35133adc..d768a0454d 100644

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_ADDEX(DisasContext *ctx, arg_X *a) · line 470

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +7 -0
@@ -461,6 +461,85 @@ static bool trans_ADDEX(DisasContext *ctx, arg_X *a)
     return true;
 }
 
+static bool trans_MTCRF(DisasContext *ctx, arg_MTCRF *a)
+{
+    uint32_t crm, crn;
+
+    crm = a->fxm;
+
+    TCGv_i32 temp = tcg_temp_new_i32();

Declarations should precede statements.

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_ADDEX(DisasContext *ctx, arg_X *a) · line 476

Quoted 6 lines
target/ppc/translate/fixedpoint-impl.c.inc +6 -0
+    tcg_gen_trunc_tl_i32(temp, cpu_gpr[a->rt]);
+
+    for (crn = 0 ; crn < 8 ; crn++) {
+        if (crm & (1 << crn)) {
+                tcg_gen_shri_i32(cpu_crf[7 - crn], temp, crn * 4);
+                tcg_gen_andi_i32(cpu_crf[7 - crn], cpu_crf[7 - crn], 0xf);

These statements are indented by an extra tab.

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_ADDEX(DisasContext *ctx, arg_X *a) · line 485

Quoted 9 lines
target/ppc/translate/fixedpoint-impl.c.inc +9 -0
+        }
+    }
+
+    return true;
+}
+
+static bool trans_MTOCRF(DisasContext *ctx, arg_MTOCRF *a)
+{
+

Please remove this blank line.

Quoted 26 lines
target/ppc/translate/fixedpoint-impl.c.inc +26 -0
+    uint32_t crm, crn;
+
+    crm = a->fxm;
+
+    /* Checking crm > 0 and set_bits(crm) == 1 */
+    if (crm && ((crm & (crm - 1)) == 0)) {
+        TCGv_i32 temp = tcg_temp_new_i32();
+        crn = ctz32(crm);
+        tcg_gen_trunc_tl_i32(temp, cpu_gpr[a->rt]);
+        tcg_gen_shri_i32(temp, temp, crn * 4);
+        tcg_gen_andi_i32(cpu_crf[7 - crn], temp, 0xf);
+    }
+
+    return true;
+}
+
+static bool trans_MFOCR(DisasContext *ctx, arg_MFOCR *a)
+{
+    uint32_t crm, crn;
+
+    crm = a->fxm;
+
+    /* Checking crm > 0 and set_bits(crm) == 1 */
+    if (likely(crm && ((crm & (crm - 1)) == 0))) {
+        crn = ctz32(crm);
+        tcg_gen_extu_i32_tl(cpu_gpr[a->rt], cpu_crf[7 - crn]);

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_ADDEX(DisasContext *ctx, arg_X *a) · line 521

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +10 -0
+        tcg_gen_shli_tl(cpu_gpr[a->rt],
+                        cpu_gpr[a->rt], crn * 4);
+    }
+
+    return true;
+}
+
+static bool trans_MFCR(DisasContext *ctx, arg_MFCR *a)
+{
+

Ditto.

With above fixed:

Quoted 27 lines
target/ppc/translate/fixedpoint-impl.c.inc +21 -0
+    TCGv_i32 t0 = tcg_temp_new_i32();
+    tcg_gen_mov_i32(t0, cpu_crf[0]);
+    tcg_gen_shli_i32(t0, t0, 4);
+    tcg_gen_or_i32(t0, t0, cpu_crf[1]);
+    tcg_gen_shli_i32(t0, t0, 4);
+    tcg_gen_or_i32(t0, t0, cpu_crf[2]);
+    tcg_gen_shli_i32(t0, t0, 4);
+    tcg_gen_or_i32(t0, t0, cpu_crf[3]);
+    tcg_gen_shli_i32(t0, t0, 4);
+    tcg_gen_or_i32(t0, t0, cpu_crf[4]);
+    tcg_gen_shli_i32(t0, t0, 4);
+    tcg_gen_or_i32(t0, t0, cpu_crf[5]);
+    tcg_gen_shli_i32(t0, t0, 4);
+    tcg_gen_or_i32(t0, t0, cpu_crf[6]);
+    tcg_gen_shli_i32(t0, t0, 4);
+    tcg_gen_or_i32(t0, t0, cpu_crf[7]);
+    tcg_gen_extu_i32_tl(cpu_gpr[a->rt], t0);
+
+    return true;
+}
+
 static bool do_add_D(DisasContext *ctx, arg_D *a, bool add_ca, bool compute_ca,
                      bool compute_ov, bool compute_rc0)
 {
Signature · 1 line
2.55.0
#48 lore raw

PATCH v2 21/37 target/ppc: Move Condition Register logical instructions to decodetree.

In reply to Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

cr{and, or, xor, nand, nor, eqv, andc, orc} : XL-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/insn32.decode                   | 11 ++++
 target/ppc/translate.c                     | 59 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 45 +++++++++++++++++
 3 files changed, 56 insertions(+), 59 deletions(-)
target/ppc/insn32.decode +11 -0
index dc708b5442..a3973e042d 100644
@@ -516,6 +516,17 @@ CFUGED          011111 ..... ..... ..... 0011011100 -   @X
 PDEPD           011111 ..... ..... ..... 0010011100 -   @X
 PEXTD           011111 ..... ..... ..... 0010111100 -   @X
 
+## Condition Register Instructions
+
+CRAND           010011 ..... ..... ..... 0100000001 -  @X
+CROR            010011 ..... ..... ..... 0111000001 -  @X
+CRNAND          010011 ..... ..... ..... 0011100001 -  @X
+CRXOR           010011 ..... ..... ..... 0011000001 -  @X
+CRNOR           010011 ..... ..... ..... 0000100001 -  @X
+CREQV           010011 ..... ..... ..... 0100100001 -  @X
+CRANDC          010011 ..... ..... ..... 0010000001 -  @X
+CRORC           010011 ..... ..... ..... 0110100001 -  @X
+
 # Fixed-Point Hash Instructions
 
 HASHST          011111 ..... ..... ..... 1011010010 .   @X_DW
target/ppc/translate.c +0 -59

71 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +45 -0

51 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 21/37
#49 lore raw

Re: [PATCH v2 21/37] target/ppc: Move Condition Register logical instructions to decodetree.

Quoted 123 lines · Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

cr{and, or, xor, nand, nor, eqv, andc, orc} : XL-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/insn32.decode                   | 11 ++++
 target/ppc/translate.c                     | 59 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 45 +++++++++++++++++
 3 files changed, 56 insertions(+), 59 deletions(-)
target/ppc/insn32.decode +11 -0
index dc708b5442..a3973e042d 100644
@@ -516,6 +516,17 @@ CFUGED          011111 ..... ..... ..... 0011011100 -   @X
 PDEPD           011111 ..... ..... ..... 0010011100 -   @X
 PEXTD           011111 ..... ..... ..... 0010111100 -   @X
 
+## Condition Register Instructions
+
+CRAND           010011 ..... ..... ..... 0100000001 -  @X
+CROR            010011 ..... ..... ..... 0111000001 -  @X
+CRNAND          010011 ..... ..... ..... 0011100001 -  @X
+CRXOR           010011 ..... ..... ..... 0011000001 -  @X
+CRNOR           010011 ..... ..... ..... 0000100001 -  @X
+CREQV           010011 ..... ..... ..... 0100100001 -  @X
+CRANDC          010011 ..... ..... ..... 0010000001 -  @X
+CRORC           010011 ..... ..... ..... 0110100001 -  @X
+
 # Fixed-Point Hash Instructions
 
 HASHST          011111 ..... ..... ..... 1011010010 .   @X_DW
target/ppc/translate.c +0 -59
index d01981de73..5b08d9b282 100644
@@ -2920,53 +2920,6 @@ static inline void gen_setlr(DisasContext *ctx, target_ulong nip)
     tcg_gen_movi_tl(cpu_lr, nip);
 }
 
-/***                      Condition register logical                       ***/
-#define GEN_CRLOGIC(name, tcg_op, opc)                                        \
-static void glue(gen_, name)(DisasContext *ctx)                               \
-{                                                                             \
-    uint8_t bitmask;                                                          \
-    int sh;                                                                   \
-    TCGv_i32 t0, t1;                                                          \
-    sh = (crbD(ctx->opcode) & 0x03) - (crbA(ctx->opcode) & 0x03);             \
-    t0 = tcg_temp_new_i32();                                                  \
-    if (sh > 0)                                                               \
-        tcg_gen_shri_i32(t0, cpu_crf[crbA(ctx->opcode) >> 2], sh);            \
-    else if (sh < 0)                                                          \
-        tcg_gen_shli_i32(t0, cpu_crf[crbA(ctx->opcode) >> 2], -sh);           \
-    else                                                                      \
-        tcg_gen_mov_i32(t0, cpu_crf[crbA(ctx->opcode) >> 2]);                 \
-    t1 = tcg_temp_new_i32();                                                  \
-    sh = (crbD(ctx->opcode) & 0x03) - (crbB(ctx->opcode) & 0x03);             \
-    if (sh > 0)                                                               \
-        tcg_gen_shri_i32(t1, cpu_crf[crbB(ctx->opcode) >> 2], sh);            \
-    else if (sh < 0)                                                          \
-        tcg_gen_shli_i32(t1, cpu_crf[crbB(ctx->opcode) >> 2], -sh);           \
-    else                                                                      \
-        tcg_gen_mov_i32(t1, cpu_crf[crbB(ctx->opcode) >> 2]);                 \
-    tcg_op(t0, t0, t1);                                                       \
-    bitmask = 0x08 >> (crbD(ctx->opcode) & 0x03);                             \
-    tcg_gen_andi_i32(t0, t0, bitmask);                                        \
-    tcg_gen_andi_i32(t1, cpu_crf[crbD(ctx->opcode) >> 2], ~bitmask);          \
-    tcg_gen_or_i32(cpu_crf[crbD(ctx->opcode) >> 2], t0, t1);                  \
-}
-
-/* crand */
-GEN_CRLOGIC(crand, tcg_gen_and_i32, 0x08);
-/* crandc */
-GEN_CRLOGIC(crandc, tcg_gen_andc_i32, 0x04);
-/* creqv */
-GEN_CRLOGIC(creqv, tcg_gen_eqv_i32, 0x09);
-/* crnand */
-GEN_CRLOGIC(crnand, tcg_gen_nand_i32, 0x07);
-/* crnor */
-GEN_CRLOGIC(crnor, tcg_gen_nor_i32, 0x01);
-/* cror */
-GEN_CRLOGIC(cror, tcg_gen_or_i32, 0x0E);
-/* crorc */
-GEN_CRLOGIC(crorc, tcg_gen_orc_i32, 0x0D);
-/* crxor */
-GEN_CRLOGIC(crxor, tcg_gen_xor_i32, 0x06);
-
 /***                           System linkage                              ***/
 
 /* rfi (supervisor only) */
@@ -5120,18 +5073,6 @@ GEN_STEPX(stw, DEF_MEMOP(MO_UL), 0x1F, 0x04)
 GEN_STEPX(std, DEF_MEMOP(MO_UQ), 0x1D, 0x04)
 #endif
 
-#undef GEN_CRLOGIC
-#define GEN_CRLOGIC(name, tcg_op, opc)                                        \
-GEN_HANDLER(name, 0x13, 0x01, opc, 0x00000001, PPC_INTEGER)
-GEN_CRLOGIC(crand, tcg_gen_and_i32, 0x08),
-GEN_CRLOGIC(crandc, tcg_gen_andc_i32, 0x04),
-GEN_CRLOGIC(creqv, tcg_gen_eqv_i32, 0x09),
-GEN_CRLOGIC(crnand, tcg_gen_nand_i32, 0x07),
-GEN_CRLOGIC(crnor, tcg_gen_nor_i32, 0x01),
-GEN_CRLOGIC(cror, tcg_gen_or_i32, 0x0E),
-GEN_CRLOGIC(crorc, tcg_gen_orc_i32, 0x0D),
-GEN_CRLOGIC(crxor, tcg_gen_xor_i32, 0x06),
-
 #undef GEN_MAC_HANDLER
 #define GEN_MAC_HANDLER(name, opc2, opc3)                                     \
 GEN_HANDLER(name, 0x04, opc2, opc3, 0x00000000, PPC_405_MAC)
target/ppc/translate/fixedpoint-impl.c.inc +0 -0
index d768a0454d..26b308e435 100644

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_MFCR(DisasContext *ctx, arg_MFCR *a) · line 546

Quoted 7 lines
target/ppc/translate/fixedpoint-impl.c.inc +4 -0
@@ -540,6 +540,51 @@ static bool trans_MFCR(DisasContext *ctx, arg_MFCR *a)
     return true;
 }
 
+static bool cr_logic_common(DisasContext *ctx, arg_X *a,
+                            void (*tcg_op)(TCGv_i32, TCGv_i32, TCGv_i32))
+{
+

Extra new line - please remove it.

Quoted 4 lines
target/ppc/translate/fixedpoint-impl.c.inc +4 -0
+    uint8_t bitmask;
+    int sh;
+    TCGv_i32 t0, t1;
+    sh = (a->rt & 0x03) - (a->ra & 0x03);

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_MFCR(DisasContext *ctx, arg_MFCR *a) · line 560

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +10 -0
+    t0 = tcg_temp_new_i32();
+    if (sh > 0) {
+        tcg_gen_shri_i32(t0, cpu_crf[a->ra >> 2], sh);
+    } else if (sh < 0) {
+        tcg_gen_shli_i32(t0, cpu_crf[a->ra >> 2], -sh);
+    } else {
+        tcg_gen_mov_i32(t0, cpu_crf[a->ra >> 2]);
+    }
+
+    t1 = tcg_temp_new_i32();

It maybe a good idea to keep these allocations together, consistent with other places in this file.

    t0 = tcg_temp_new_i32();
    t1 = tcg_temp_new_i32();
    ...

With these nits fixed:

Quoted 33 lines
target/ppc/translate/fixedpoint-impl.c.inc +27 -0
+    sh = (a->rt & 0x03) - (a->rb & 0x03);
+    if (sh > 0) {
+        tcg_gen_shri_i32(t1, cpu_crf[a->rb >> 2], sh);
+    } else if (sh < 0) {
+        tcg_gen_shli_i32(t1, cpu_crf[a->rb >> 2], -sh);
+    } else {
+        tcg_gen_mov_i32(t1, cpu_crf[a->rb >> 2]);
+    }
+    tcg_op(t0, t0, t1);
+
+    bitmask = 0x08 >> (a->rt & 0x03);
+    tcg_gen_andi_i32(t0, t0, bitmask);
+    tcg_gen_andi_i32(t1, cpu_crf[a->rt >> 2], ~bitmask);
+    tcg_gen_or_i32(cpu_crf[a->rt >> 2], t0, t1);
+
+    return true;
+}
+
+TRANS(CRAND, cr_logic_common, tcg_gen_and_i32);
+TRANS(CROR, cr_logic_common, tcg_gen_or_i32);
+TRANS(CRXOR, cr_logic_common, tcg_gen_xor_i32);
+TRANS(CRNAND, cr_logic_common, tcg_gen_nand_i32);
+TRANS(CRNOR, cr_logic_common, tcg_gen_nor_i32);
+TRANS(CRANDC, cr_logic_common, tcg_gen_andc_i32);
+TRANS(CREQV, cr_logic_common, tcg_gen_eqv_i32);
+TRANS(CRORC, cr_logic_common, tcg_gen_orc_i32);
+
 static bool do_add_D(DisasContext *ctx, arg_D *a, bool add_ca, bool compute_ca,
                      bool compute_ov, bool compute_rc0)
 {
Signature · 1 line
2.55.0
#50 lore raw

PATCH v2 22/37 target/ppc: make do_ea_calc_ra available for 32 bit builds

In reply to Chinmay Rath

There is no reason to keep do_ea_calc_ra exclusive to TARGET_PPC64. Make it available for PPC32 builds.

---
 target/ppc/translate.c | 2 --
 1 file changed, 2 deletions(-)
target/ppc/translate.c +0 -2
index 5b08d9b282..e40841b68e 100644
@@ -2397,7 +2397,6 @@ static TCGv do_ea_calc(DisasContext *ctx, int ra, TCGv displ)
     return ea;
 }
 
-#if defined(TARGET_PPC64)
 /* EA <- (ra == 0) ? 0 : GPR[ra] */
 static TCGv do_ea_calc_ra(DisasContext *ctx, int ra)
 {
@@ -2411,7 +2410,6 @@ static TCGv do_ea_calc_ra(DisasContext *ctx, int ra)
     }
     return EA;
 }
-#endif
 
 /***                             Integer load                              ***/
 #define DEF_MEMOP(op) ((op) | ctx->default_tcg_memop_mask)
Signature · 1 line
2.55.0
Reviewed-by on v2 22/37
#51 lore raw

Re: [PATCH v2 22/37] target/ppc: make do_ea_calc_ra available for 32 bit builds

Quoted 8 lines · Chinmay Rath

There is no reason to keep do_ea_calc_ra exclusive to TARGET_PPC64. Make it available for PPC32 builds.

---
 target/ppc/translate.c | 2 --
 1 file changed, 2 deletions(-)
Quoted 24 lines
target/ppc/translate.c +0 -2
index 5b08d9b282..e40841b68e 100644
@@ -2397,7 +2397,6 @@ static TCGv do_ea_calc(DisasContext *ctx, int ra, TCGv displ)
     return ea;
 }
 
-#if defined(TARGET_PPC64)
 /* EA <- (ra == 0) ? 0 : GPR[ra] */
 static TCGv do_ea_calc_ra(DisasContext *ctx, int ra)
 {
@@ -2411,7 +2410,6 @@ static TCGv do_ea_calc_ra(DisasContext *ctx, int ra)
     }
     return EA;
 }
-#endif
 
 /***                             Integer load                              ***/
 #define DEF_MEMOP(op) ((op) | ctx->default_tcg_memop_mask)
Signature · 1 line
2.55.0
#52 lore raw

PATCH v2 23/37 target/ppc: Move Fixed-Point Load/Store String instructions to decodetree.

In reply to Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

lsw{i, x},
stsw{i, x} : X-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/helper.h                        |   6 +-
 target/ppc/insn32.decode                   |   7 ++
 target/ppc/mem_helper.c                    |   6 +-
 target/ppc/translate.c                     | 102 --------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 103 +++++++++++++++++++++
 5 files changed, 116 insertions(+), 108 deletions(-)
target/ppc/helper.h +3 -3
index 7e15676122..0e8b96d0ec 100644
@@ -45,9 +45,9 @@ DEF_HELPER_1(check_tlb_flush_global, void, env)
 
 DEF_HELPER_3(lmw, void, env, tl, i32)
 DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
-DEF_HELPER_4(lsw, void, env, tl, i32, i32)
-DEF_HELPER_5(lswx, void, env, tl, i32, i32, i32)
-DEF_HELPER_FLAGS_4(stsw, TCG_CALL_NO_WG, void, env, tl, i32, i32)
+DEF_HELPER_4(LSW, void, env, tl, i32, i32)
+DEF_HELPER_5(LSWX, void, env, tl, i32, i32, i32)
+DEF_HELPER_FLAGS_4(STSW, TCG_CALL_NO_WG, void, env, tl, i32, i32)
 DEF_HELPER_FLAGS_3(DCBZ, TCG_CALL_NO_WG, void, env, tl, int)
 #ifdef TARGET_PPC64
 DEF_HELPER_FLAGS_2(DCBZL, TCG_CALL_NO_WG, void, env, tl)
target/ppc/insn32.decode +7 -0
index a3973e042d..e7b5d19cde 100644
@@ -527,6 +527,13 @@ CREQV           010011 ..... ..... ..... 0100100001 -  @X
 CRANDC          010011 ..... ..... ..... 0010000001 -  @X
 CRORC           010011 ..... ..... ..... 0110100001 -  @X
 
+# Fixed Point Move Assist Instructions
+
+LSWI            011111 ..... ..... ..... 1 001010101 -  @X
+LSWX            011111 ..... ..... ..... 1 000010101 -  @X
+STSWI           011111 ..... ..... ..... 1 011010101 -  @X
+STSWX           011111 ..... ..... ..... 1 010010101 -  @X
+
 # Fixed-Point Hash Instructions
 
 HASHST          011111 ..... ..... ..... 1011010010 .   @X_DW
target/ppc/mem_helper.c +3 -3
index 95a224e74c..582180cf88 100644
@@ -190,7 +190,7 @@ static void do_lsw(CPUPPCState *env, target_ulong addr, uint32_t nb,
     env->gpr[reg] = val;
 }
 
-void helper_lsw(CPUPPCState *env, target_ulong addr,
+void helper_LSW(CPUPPCState *env, target_ulong addr,
                 uint32_t nb, uint32_t reg)
 {
     do_lsw(env, addr, nb, reg, GETPC());
@@ -202,7 +202,7 @@ void helper_lsw(CPUPPCState *env, target_ulong addr,
  * this is valid, but rA won't be loaded.  For now, I'll follow the
  * spec...
  */
-void helper_lswx(CPUPPCState *env, target_ulong addr, uint32_t reg,
+void helper_LSWX(CPUPPCState *env, target_ulong addr, uint32_t reg,
                  uint32_t ra, uint32_t rb)
 {
     if (likely(xer_bc != 0)) {
@@ -218,7 +218,7 @@ void helper_lswx(CPUPPCState *env, target_ulong addr, uint32_t reg,
     }
 }
 
-void helper_stsw(CPUPPCState *env, target_ulong addr, uint32_t nb,
+void helper_STSW(CPUPPCState *env, target_ulong addr, uint32_t nb,
                  uint32_t reg)
 {
     uintptr_t raddr = GETPC();
target/ppc/translate.c +0 -102

114 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +103 -0

109 lines not on this page · raw message

Signature · 1 line
2.55.0
#53 lore raw

Re: [PATCH v2 23/37] target/ppc: Move Fixed-Point Load/Store String instructions to decodetree.

Quoted 213 lines · Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

lsw{i, x},
stsw{i, x} : X-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/helper.h                        |   6 +-
 target/ppc/insn32.decode                   |   7 ++
 target/ppc/mem_helper.c                    |   6 +-
 target/ppc/translate.c                     | 102 --------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 103 +++++++++++++++++++++
 5 files changed, 116 insertions(+), 108 deletions(-)
target/ppc/helper.h +3 -3
index 7e15676122..0e8b96d0ec 100644
@@ -45,9 +45,9 @@ DEF_HELPER_1(check_tlb_flush_global, void, env)
 
 DEF_HELPER_3(lmw, void, env, tl, i32)
 DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
-DEF_HELPER_4(lsw, void, env, tl, i32, i32)
-DEF_HELPER_5(lswx, void, env, tl, i32, i32, i32)
-DEF_HELPER_FLAGS_4(stsw, TCG_CALL_NO_WG, void, env, tl, i32, i32)
+DEF_HELPER_4(LSW, void, env, tl, i32, i32)
+DEF_HELPER_5(LSWX, void, env, tl, i32, i32, i32)
+DEF_HELPER_FLAGS_4(STSW, TCG_CALL_NO_WG, void, env, tl, i32, i32)
 DEF_HELPER_FLAGS_3(DCBZ, TCG_CALL_NO_WG, void, env, tl, int)
 #ifdef TARGET_PPC64
 DEF_HELPER_FLAGS_2(DCBZL, TCG_CALL_NO_WG, void, env, tl)
target/ppc/insn32.decode +7 -0
index a3973e042d..e7b5d19cde 100644
@@ -527,6 +527,13 @@ CREQV           010011 ..... ..... ..... 0100100001 -  @X
 CRANDC          010011 ..... ..... ..... 0010000001 -  @X
 CRORC           010011 ..... ..... ..... 0110100001 -  @X
 
+# Fixed Point Move Assist Instructions
+
+LSWI            011111 ..... ..... ..... 1 001010101 -  @X
+LSWX            011111 ..... ..... ..... 1 000010101 -  @X
+STSWI           011111 ..... ..... ..... 1 011010101 -  @X
+STSWX           011111 ..... ..... ..... 1 010010101 -  @X
+
 # Fixed-Point Hash Instructions
 
 HASHST          011111 ..... ..... ..... 1011010010 .   @X_DW
target/ppc/mem_helper.c +3 -3
index 95a224e74c..582180cf88 100644
@@ -190,7 +190,7 @@ static void do_lsw(CPUPPCState *env, target_ulong addr, uint32_t nb,
     env->gpr[reg] = val;
 }
 
-void helper_lsw(CPUPPCState *env, target_ulong addr,
+void helper_LSW(CPUPPCState *env, target_ulong addr,
                 uint32_t nb, uint32_t reg)
 {
     do_lsw(env, addr, nb, reg, GETPC());
@@ -202,7 +202,7 @@ void helper_lsw(CPUPPCState *env, target_ulong addr,
  * this is valid, but rA won't be loaded.  For now, I'll follow the
  * spec...
  */
-void helper_lswx(CPUPPCState *env, target_ulong addr, uint32_t reg,
+void helper_LSWX(CPUPPCState *env, target_ulong addr, uint32_t reg,
                  uint32_t ra, uint32_t rb)
 {
     if (likely(xer_bc != 0)) {
@@ -218,7 +218,7 @@ void helper_lswx(CPUPPCState *env, target_ulong addr, uint32_t reg,
     }
 }
 
-void helper_stsw(CPUPPCState *env, target_ulong addr, uint32_t nb,
+void helper_STSW(CPUPPCState *env, target_ulong addr, uint32_t nb,
                  uint32_t reg)
 {
     uintptr_t raddr = GETPC();
target/ppc/translate.c +0 -102
index e40841b68e..712990cf2d 100644
@@ -2595,104 +2595,6 @@ static void gen_stmw(DisasContext *ctx)
     gen_helper_stmw(tcg_env, t0, t1);
 }
 
-/***                    Integer load and store strings                     ***/
-
-/* lswi */
-/*
- * PowerPC32 specification says we must generate an exception if rA is
- * in the range of registers to be loaded.  In an other hand, IBM says
- * this is valid, but rA won't be loaded.  For now, I'll follow the
- * spec...
- */
-static void gen_lswi(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1, t2;
-    int nb = NB(ctx->opcode);
-    int start = rD(ctx->opcode);
-    int ra = rA(ctx->opcode);
-    int nr;
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    if (nb == 0) {
-        nb = 32;
-    }
-    nr = DIV_ROUND_UP(nb, 4);
-    if (unlikely(lsw_reg_in_range(start, nr, ra))) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_LSWX);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    gen_addr_register(ctx, t0);
-    t1 = tcg_constant_i32(nb);
-    t2 = tcg_constant_i32(start);
-    gen_helper_lsw(tcg_env, t0, t1, t2);
-}
-
-/* lswx */
-static void gen_lswx(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1, t2, t3;
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    t1 = tcg_constant_i32(rD(ctx->opcode));
-    t2 = tcg_constant_i32(rA(ctx->opcode));
-    t3 = tcg_constant_i32(rB(ctx->opcode));
-    gen_helper_lswx(tcg_env, t0, t1, t2, t3);
-}
-
-/* stswi */
-static void gen_stswi(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1, t2;
-    int nb = NB(ctx->opcode);
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    gen_addr_register(ctx, t0);
-    if (nb == 0) {
-        nb = 32;
-    }
-    t1 = tcg_constant_i32(nb);
-    t2 = tcg_constant_i32(rS(ctx->opcode));
-    gen_helper_stsw(tcg_env, t0, t1, t2);
-}
-
-/* stswx */
-static void gen_stswx(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1, t2;
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    gen_addr_reg_index(ctx, t0);
-    t1 = tcg_temp_new_i32();
-    tcg_gen_trunc_tl_i32(t1, cpu_xer);
-    tcg_gen_andi_i32(t1, t1, 0x7F);
-    t2 = tcg_constant_i32(rS(ctx->opcode));
-    gen_helper_stsw(tcg_env, t0, t1, t2);
-}
 
 #if !defined(CONFIG_USER_ONLY)
 static inline void gen_check_tlb_flush(DisasContext *ctx, bool global)
@@ -4914,10 +4816,6 @@ GEN_HANDLER_E(dform39, 0x39, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
 GEN_HANDLER_E(dform3D, 0x3D, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
 GEN_HANDLER(lmw, 0x2E, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(stmw, 0x2F, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(lswi, 0x1F, 0x15, 0x12, 0x00000001, PPC_STRING),
-GEN_HANDLER(lswx, 0x1F, 0x15, 0x10, 0x00000001, PPC_STRING),
-GEN_HANDLER(stswi, 0x1F, 0x15, 0x16, 0x00000001, PPC_STRING),
-GEN_HANDLER(stswx, 0x1F, 0x15, 0x14, 0x00000001, PPC_STRING),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
target/ppc/translate/fixedpoint-impl.c.inc +0 -0
index 26b308e435..82fb645279 100644

target/ppc/translate/fixedpoint-impl.c.inc TRANS(CRANDC, cr_logic_common, tcg_gen_andc_i32); · line 588

Quoted 4 lines
target/ppc/translate/fixedpoint-impl.c.inc +1 -0
@@ -585,6 +585,109 @@ TRANS(CRANDC, cr_logic_common, tcg_gen_andc_i32);
 TRANS(CREQV, cr_logic_common, tcg_gen_eqv_i32);
 TRANS(CRORC, cr_logic_common, tcg_gen_orc_i32);
 
+

This is extra blank line. Please remove.

target/ppc/translate/fixedpoint-impl.c.inc TRANS(CRANDC, cr_logic_common, tcg_gen_andc_i32); · line 589

Quoted 1 line
target/ppc/translate/fixedpoint-impl.c.inc +1 -0
+/***                    Integer load and store strings                     ***/

This banner-comment might not needed. But don't have any strong feelings about it.

Quoted 82 lines
target/ppc/translate/fixedpoint-impl.c.inc +82 -0
+
+/* lswi */
+static bool trans_LSWI(DisasContext *ctx, arg_LSWI *a)
+{
+    TCGv t0;
+    TCGv_i32 t1, t2;
+    int nb = a->rb;
+    int start = a->rt;
+    int ra = a->ra;
+    int nr;
+
+    REQUIRE_INSNS_FLAGS(ctx, STRING);
+    if (ctx->le_mode) {
+        gen_align_no_le(ctx);
+        return true;
+    }
+    if (nb == 0) {
+        nb = 32;
+    }
+    nr = DIV_ROUND_UP(nb, 4);
+    if (unlikely(lsw_reg_in_range(start, nr, ra))) {
+        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_LSWX);
+        return true;
+    }
+    gen_set_access_type(ctx, ACCESS_INT);
+    t0 = do_ea_calc_ra(ctx, ra);
+    t1 = tcg_constant_i32(nb);
+    t2 = tcg_constant_i32(start);
+    gen_helper_LSW(tcg_env, t0, t1, t2);
+
+    return true;
+}
+
+/* lswx */
+static bool trans_LSWX(DisasContext *ctx, arg_LSWX *a)
+{
+    TCGv t0;
+    TCGv_i32 t1, t2, t3;
+
+    REQUIRE_INSNS_FLAGS(ctx, STRING);
+    if (ctx->le_mode) {
+        gen_align_no_le(ctx);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_INT);
+    t0 = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
+    t1 = tcg_constant_i32(a->rt);
+    t2 = tcg_constant_i32(a->ra);
+    t3 = tcg_constant_i32(a->rb);
+    gen_helper_LSWX(tcg_env, t0, t1, t2, t3);
+
+    return true;
+}
+
+/* stswi */
+static bool trans_STSWI(DisasContext *ctx, arg_STSWI *a)
+{
+    TCGv t0;
+    TCGv_i32 t1, t2;
+    int nb = a->rb;
+
+    REQUIRE_INSNS_FLAGS(ctx, STRING);
+    if (ctx->le_mode) {
+        gen_align_no_le(ctx);
+        return true;
+    }
+    gen_set_access_type(ctx, ACCESS_INT);
+    t0 = do_ea_calc_ra(ctx, a->ra);
+    if (nb == 0) {
+        nb = 32;
+    }
+    t1 = tcg_constant_i32(nb);
+    t2 = tcg_constant_i32(a->rt);
+    gen_helper_STSW(tcg_env, t0, t1, t2);
+
+    return true;
+}
+
+/* stswx */
+static bool trans_STSWX(DisasContext *ctx, arg_STSWX *a)
+{

target/ppc/translate/fixedpoint-impl.c.inc TRANS(CRANDC, cr_logic_common, tcg_gen_andc_i32); · line 681

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +10 -0
+    TCGv t0;
+    TCGv_i32 t1, t2;
+
+    REQUIRE_INSNS_FLAGS(ctx, STRING);
+    if (ctx->le_mode) {
+        gen_align_no_le(ctx);
+        return true;
+    }
+    gen_set_access_type(ctx, ACCESS_INT);
+    t0 = do_ea_calc_ra(ctx, a->ra);

STSWX is X-form indexed. As per Power ISA:

EA ← (RA=0 ? 0 : (RA)) + (RB)

do_ea_calc_ra only computes (RA=0 ? 0 : GPR[RA]) — GPR[RB] is never added. Any stswx with a non-zero RB will store to the wrong address.

The old gen_stswx correctly called gen_addr_reg_index() which includes the RB addition. trans_LSWX (three functions above) also gets this right. Fix:

- t0 = do_ea_calc_ra(ctx, a->ra);
+ t0 = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);

Note: do_ea_calc_ra is correct for trans_LSWI and trans_STSWI because those are the immediate forms — their ISA EA is also (RA==0 ? 0 :
GPR[RA]) with no RB term; the RB field carries NB, not a register EA operand.

Quoted 15 lines
target/ppc/translate/fixedpoint-impl.c.inc +9 -0
+    t1 = tcg_temp_new_i32();
+    tcg_gen_trunc_tl_i32(t1, cpu_xer);
+    tcg_gen_andi_i32(t1, t1, 0x7F);
+    t2 = tcg_constant_i32(a->rt);
+    gen_helper_STSW(tcg_env, t0, t1, t2);
+
+    return true;
+}
+
 static bool do_add_D(DisasContext *ctx, arg_D *a, bool add_ca, bool compute_ca,
                      bool compute_ov, bool compute_rc0)
 {
Signature · 1 line
2.55.0
#54 lore raw

PATCH v2 24/37 target/ppc: Move VMX integer arithmetic and BCD instructions to decodetree.

In reply to Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

        vadd{u, s}{b, h, w, d}m[.],
        vsub{u, s}{b, h, w, d}m[.],
        vmuluwm[.],
        vmul10{u, eu, ecu, cu}q[.]       : VX-form

        bcd{add, sub}[.],
        bcd{s, us}[.],
        bcdcpsgn[.],
        bcdct{sq, n}[.],
        bcdcf{sq, z, n}[.],
        bcdctz[.],
        bcdsetsgn[.],
        bcdsr[.],
        bcd{u,}trunc[.]                  : VX-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/helper.h                 |  30 +--
 target/ppc/insn32.decode            |  39 ++++
 target/ppc/int_helper.c             |  32 +--
 target/ppc/translate/vmx-impl.c.inc | 299 ++++++++++++----------------
 target/ppc/translate/vmx-ops.c.inc  |  22 --
 5 files changed, 199 insertions(+), 223 deletions(-)
target/ppc/helper.h +15 -15
index 0e8b96d0ec..8425eceeb3 100644
@@ -349,21 +349,21 @@ DEF_HELPER_FLAGS_3(vshasigmaw, TCG_CALL_NO_RWG, void, avr, avr, i32)
 DEF_HELPER_FLAGS_3(vshasigmad, TCG_CALL_NO_RWG, void, avr, avr, i32)
 DEF_HELPER_FLAGS_4(vpermxor, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 
-DEF_HELPER_FLAGS_4(bcdadd, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdsub, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdcfn, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdctn, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdcfz, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdctz, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdcfsq, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdctsq, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdcpsgn, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdsetsgn, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcds, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdus, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdsr, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdtrunc, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdutrunc, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDADD, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDSUB, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCFN, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCTN, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCFZ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCTZ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCFSQ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCTSQ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDCPSGN, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDSETSGN, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDS, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDUS, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDSR, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDTRUNC, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDUTRUNC, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
 
 DEF_HELPER_4(XSADDDP, void, env, vsr, vsr, vsr)
 DEF_HELPER_5(xsaddqp, void, env, i32, vsr, vsr, vsr)
target/ppc/insn32.decode +39 -0
index e7b5d19cde..eed963bd71 100644
@@ -90,6 +90,9 @@
 &VX             vrt vra vrb
 @VX             ...... vrt:5 vra:5 vrb:5 .......... .           &VX
 
+&VX_ps             vrt vra vrb ps
+@VX_ps             ...... vrt:5 vra:5 vrb:5 . ps:1 ........ .           &VX_ps
+
 &VX_bf          bf vra vrb
 @VX_bf          ...... bf:3 .. vra:5 vrb:5 ...........          &VX_bf
 
@@ -108,6 +111,9 @@
 &VX_tb          vrt vrb
 @VX_tb          ...... vrt:5 ..... vrb:5 ...........            &VX_tb
 
+&VX_tb_ps       vrt vrb ps
+@VX_tb_ps       ...... vrt:5 ..... vrb:5 . ps:1 .........            &VX_tb_ps
+
 &X              rt ra rb
 @X              ...... rt:5 ra:5 rb:5 .......... .              &X
 
@@ -991,6 +997,10 @@ VADDSBS         000100 ..... ..... ..... 01100000000    @VX
 VADDSHS         000100 ..... ..... ..... 01101000000    @VX
 VADDSWS         000100 ..... ..... ..... 01110000000    @VX
 
+VADDUBM         000100 ..... ..... ..... 00000000000    @VX
+VADDUHM         000100 ..... ..... ..... 00001000000    @VX
+VADDUWM         000100 ..... ..... ..... 00010000000    @VX
+VADDUDM         000100 ..... ..... ..... 00011000000    @VX
 VADDUBS         000100 ..... ..... ..... 01000000000    @VX
 VADDUHS         000100 ..... ..... ..... 01001000000    @VX
 VADDUWS         000100 ..... ..... ..... 01010000000    @VX
@@ -1006,6 +1016,10 @@ VSUBSBS         000100 ..... ..... ..... 11100000000    @VX
 VSUBSHS         000100 ..... ..... ..... 11101000000    @VX
 VSUBSWS         000100 ..... ..... ..... 11110000000    @VX
 
+VSUBUBM         000100 ..... ..... ..... 10000000000    @VX
+VSUBUHM         000100 ..... ..... ..... 10001000000    @VX
+VSUBUWM         000100 ..... ..... ..... 10010000000    @VX
+VSUBUDM         000100 ..... ..... ..... 10011000000    @VX
 VSUBUBS         000100 ..... ..... ..... 11000000000    @VX
 VSUBUHS         000100 ..... ..... ..... 11001000000    @VX
 VSUBUWS         000100 ..... ..... ..... 11010000000    @VX
@@ -1099,6 +1113,13 @@ VMULHSD         000100 ..... ..... ..... 01111001001    @VX
 VMULHUD         000100 ..... ..... ..... 01011001001    @VX
 VMULLD          000100 ..... ..... ..... 00111001001    @VX
 
+VMULUWM         000100 ..... ..... ..... 00010001001    @VX
+
+VMUL10UQ        000100 ..... ..... ..... 01000000001    @VX
+VMUL10EUQ       000100 ..... ..... ..... 01001000001    @VX
+VMUL10ECUQ      000100 ..... ..... ..... 00001000001    @VX
+VMUL10CUQ       000100 ..... ..... ..... 00000000001    @VX
+
 ## Vector Multiply-Sum Instructions
 
 VMSUMUBM        000100 ..... ..... ..... ..... 100100   @VA
@@ -1141,6 +1162,24 @@ VPKSWSS         000100 ..... ..... ..... 0011100111 0    @VX
 VPKSDSS         000100 ..... ..... ..... 1011100111 0    @VX
 VPKPX           000100 ..... ..... ..... 0110000111 0    @VX
 
+## Decimal Integer Arithmetic Instructions
+
+BCDADD          000100 ..... ..... ..... 1.000000001    @VX_ps
+BCDSUB          000100 ..... ..... ..... 1.001000001    @VX_ps
+BCDUS           000100 ..... ..... ..... 10010000001    @VX
+BCDS            000100 ..... ..... ..... 1.011000001    @VX_ps
+BCDCTSQ         000100 ..... 00000 ..... 10110000001    @VX_tb
+BCDCTN          000100 ..... 00101 ..... 10110000001    @VX_tb
+BCDCFSQ         000100 ..... 00010 ..... 1.110000001    @VX_tb_ps
+BCDCTZ          000100 ..... 00100 ..... 1.110000001    @VX_tb_ps
+BCDCFZ          000100 ..... 00110 ..... 1.110000001    @VX_tb_ps
+BCDCFN          000100 ..... 00111 ..... 1.110000001    @VX_tb_ps
+BCDSETSGN       000100 ..... 11111 ..... 1.110000001    @VX_tb_ps
+BCDCPSGN        000100 ..... ..... ..... 01101000001    @VX
+BCDSR           000100 ..... ..... ..... 1.111000001    @VX_ps
+BCDTRUNC        000100 ..... ..... ..... 1.100000001    @VX_ps
+BCDUTRUNC       000100 ..... ..... ..... 10101000001    @VX
+
 # VSX Load/Store Instructions
 
 LXSD            111001 ..... ..... .............. 10    @DS
target/ppc/int_helper.c +16 -16
index 8e8076a92c..86346f5b9a 100644
@@ -2348,7 +2348,7 @@ static void bcd_sub_mag(ppc_avr_t *t, ppc_avr_t *a, ppc_avr_t *b, int *invalid,
     *overflow = carry;
 }
 
-uint32_t helper_bcdadd(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDADD(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
 
     int sgna = bcd_get_sgn(a);
@@ -2395,7 +2395,7 @@ uint32_t helper_bcdadd(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdsub(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDSUB(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     ppc_avr_t bcopy = *b;
     int sgnb = bcd_get_sgn(b);
@@ -2406,10 +2406,10 @@ uint32_t helper_bcdsub(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     }
     /* else invalid ... defer to bcdadd code for proper handling */
 
-    return helper_bcdadd(r, a, &bcopy, ps);
+    return helper_BCDADD(r, a, &bcopy, ps);
 }
 
-uint32_t helper_bcdcfn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCFN(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2445,7 +2445,7 @@ uint32_t helper_bcdcfn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdctn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCTN(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2479,7 +2479,7 @@ uint32_t helper_bcdctn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdcfz(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCFZ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2523,7 +2523,7 @@ uint32_t helper_bcdcfz(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdctz(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCTZ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2582,7 +2582,7 @@ static inline int ucmp128(uint64_t alo, uint64_t ahi,
         (ahi > bhi ? 1 : -1);
 }
 
-uint32_t helper_bcdcfsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCFSQ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr;
@@ -2640,7 +2640,7 @@ uint32_t helper_bcdcfsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdctsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCTSQ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     uint8_t i;
     int cr;
@@ -2680,7 +2680,7 @@ uint32_t helper_bcdctsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdcpsgn(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCPSGN(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int invalid = 0;
@@ -2703,7 +2703,7 @@ uint32_t helper_bcdcpsgn(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return bcd_cmp_zero(r);
 }
 
-uint32_t helper_bcdsetsgn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDSETSGN(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int sgnb = bcd_get_sgn(b);
 
@@ -2717,7 +2717,7 @@ uint32_t helper_bcdsetsgn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return bcd_cmp_zero(r);
 }
 
-uint32_t helper_bcds(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDS(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int cr;
     int i = a->VsrSB(7);
@@ -2753,7 +2753,7 @@ uint32_t helper_bcds(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdus(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDUS(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int cr;
     int i;
@@ -2790,7 +2790,7 @@ uint32_t helper_bcdus(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdsr(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDSR(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int cr;
     int unused = 0;
@@ -2836,7 +2836,7 @@ uint32_t helper_bcdsr(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdtrunc(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDTRUNC(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     uint64_t mask;
     uint32_t ox_flag = 0;
@@ -2869,7 +2869,7 @@ uint32_t helper_bcdtrunc(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return bcd_cmp_zero(&ret) | ox_flag;
 }
 
-uint32_t helper_bcdutrunc(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDUTRUNC(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     uint64_t mask;
target/ppc/translate/vmx-impl.c.inc +129 -170

363 lines not on this page · raw message

target/ppc/translate/vmx-ops.c.inc +0 -22
index 2820976c0f..c1951c6975 100644
@@ -24,39 +24,17 @@ GEN_HANDLER_E(name0##_##name1, 0x4, opc2, opc3, 0x00000000, type0, type1)
 GEN_HANDLER_E(name0##_##name1, 0x4, opc2, opc3, 0x00000000, tp0, tp1), \
 GEN_HANDLER_E(name0##_##name1, 0x4, opc2, (opc3 | 0x10), 0x00000000, tp0, tp1),
 
-GEN_VXFORM_DUAL(vaddubm, vmul10cuq, 0, 0, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM_DUAL(vadduhm, vmul10ecuq, 0, 1, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM(vadduwm, 0, 2),
-GEN_VXFORM_207(vaddudm, 0, 3),
-GEN_VXFORM_DUAL(vsububm, bcdadd, 0, 16, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM_DUAL(vsubuhm, bcdsub, 0, 17, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM_DUAL(vsubuwm, bcdus, 0, 18, PPC_ALTIVEC, PPC2_ISA300),
-GEN_VXFORM_DUAL(vsubudm, bcds, 0, 19, PPC2_ALTIVEC_207, PPC2_ISA300),
-GEN_VXFORM_300(bcds, 0, 27),
 GEN_VXFORM_300(vextublx, 6, 24),
 GEN_VXFORM_300(vextuhlx, 6, 25),
 GEN_VXFORM_DUAL(vmrgow, vextuwlx, 6, 26, PPC_NONE, PPC2_ALTIVEC_207),
 GEN_VXFORM_300(vextubrx, 6, 28),
 GEN_VXFORM_300(vextuhrx, 6, 29),
 GEN_VXFORM_DUAL(vmrgew, vextuwrx, 6, 30, PPC_NONE, PPC2_ALTIVEC_207),
-GEN_VXFORM_207(vmuluwm, 4, 2),
 GEN_VXFORM_300(vsrv, 2, 28),
 GEN_VXFORM_300(vslv, 2, 29),
 GEN_VXFORM(vslo, 6, 16),
 GEN_VXFORM(vsro, 6, 17),
 
-GEN_VXFORM(xpnd04_1, 0, 22),
-GEN_VXFORM_300(bcdsr, 0, 23),
-GEN_VXFORM_300(bcdsr, 0, 31),
-GEN_VXFORM_300_EXT(vmul10uq, 0, 8, 0x0000F800),
-GEN_VXFORM_300(vmul10euq, 0, 9),
-GEN_VXFORM_300(bcdcpsgn, 0, 13),
-GEN_VXFORM_207(bcdadd, 0, 24),
-GEN_VXFORM_207(bcdsub, 0, 25),
-GEN_VXFORM_300(bcdtrunc, 0, 28),
-GEN_VXFORM_300(xpnd04_2, 0, 30),
-GEN_VXFORM_300(bcdtrunc, 0, 20),
-GEN_VXFORM_300(bcdutrunc, 0, 21),
 GEN_VXFORM(vsl, 2, 7),
 GEN_VXFORM(vsr, 2, 11),
 GEN_VXFORM(vsum4ubs, 4, 24),
Signature · 1 line
2.55.0
Reviewed-by on v2 24/37
#55 lore raw

Re: [PATCH v2 24/37] target/ppc: Move VMX integer arithmetic and BCD instructions to decodetree.

Quoted 80 lines · Chinmay Rath

From: Shivang Upadhyay <[email protected]>

Move below instructions to decodetree specification :

        vadd{u, s}{b, h, w, d}m[.],
        vsub{u, s}{b, h, w, d}m[.],
        vmuluwm[.],
        vmul10{u, eu, ecu, cu}q[.]       : VX-form

        bcd{add, sub}[.],
        bcd{s, us}[.],
        bcdcpsgn[.],
        bcdct{sq, n}[.],
        bcdcf{sq, z, n}[.],
        bcdctz[.],
        bcdsetsgn[.],
        bcdsr[.],
        bcd{u,}trunc[.]                  : VX-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the '-d in_asm,op' flag.

---
 target/ppc/helper.h                 |  30 +--
 target/ppc/insn32.decode            |  39 ++++
 target/ppc/int_helper.c             |  32 +--
 target/ppc/translate/vmx-impl.c.inc | 299 ++++++++++++----------------
 target/ppc/translate/vmx-ops.c.inc  |  22 --
 5 files changed, 199 insertions(+), 223 deletions(-)
target/ppc/helper.h +15 -15
index 0e8b96d0ec..8425eceeb3 100644
@@ -349,21 +349,21 @@ DEF_HELPER_FLAGS_3(vshasigmaw, TCG_CALL_NO_RWG, void, avr, avr, i32)
 DEF_HELPER_FLAGS_3(vshasigmad, TCG_CALL_NO_RWG, void, avr, avr, i32)
 DEF_HELPER_FLAGS_4(vpermxor, TCG_CALL_NO_RWG, void, avr, avr, avr, avr)
 
-DEF_HELPER_FLAGS_4(bcdadd, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdsub, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdcfn, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdctn, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdcfz, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdctz, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdcfsq, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdctsq, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdcpsgn, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_3(bcdsetsgn, TCG_CALL_NO_RWG, i32, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcds, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdus, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdsr, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdtrunc, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
-DEF_HELPER_FLAGS_4(bcdutrunc, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDADD, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDSUB, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCFN, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCTN, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCFZ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCTZ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCFSQ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDCTSQ, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDCPSGN, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_3(BCDSETSGN, TCG_CALL_NO_RWG, i32, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDS, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDUS, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDSR, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDTRUNC, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
+DEF_HELPER_FLAGS_4(BCDUTRUNC, TCG_CALL_NO_RWG, i32, avr, avr, avr, i32)
 
 DEF_HELPER_4(XSADDDP, void, env, vsr, vsr, vsr)
 DEF_HELPER_5(xsaddqp, void, env, i32, vsr, vsr, vsr)
target/ppc/insn32.decode +0 -0
index e7b5d19cde..eed963bd71 100644

target/ppc/insn32.decode line 94

Quoted 5 lines
target/ppc/insn32.decode +2 -0
@@ -90,6 +90,9 @@
 &VX             vrt vra vrb
 @VX             ...... vrt:5 vra:5 vrb:5 .......... .           &VX
 
+&VX_ps             vrt vra vrb ps
+@VX_ps             ...... vrt:5 vra:5 vrb:5 . ps:1 ........ .           &VX_ps

Alignment looks a bit off. Otherwise,

Quoted 640 lines
target/ppc/insn32.decode +37 -0
+
 &VX_bf          bf vra vrb
 @VX_bf          ...... bf:3 .. vra:5 vrb:5 ...........          &VX_bf
 
@@ -108,6 +111,9 @@
 &VX_tb          vrt vrb
 @VX_tb          ...... vrt:5 ..... vrb:5 ...........            &VX_tb
 
+&VX_tb_ps       vrt vrb ps
+@VX_tb_ps       ...... vrt:5 ..... vrb:5 . ps:1 .........            &VX_tb_ps
+
 &X              rt ra rb
 @X              ...... rt:5 ra:5 rb:5 .......... .              &X
 
@@ -991,6 +997,10 @@ VADDSBS         000100 ..... ..... ..... 01100000000    @VX
 VADDSHS         000100 ..... ..... ..... 01101000000    @VX
 VADDSWS         000100 ..... ..... ..... 01110000000    @VX
 
+VADDUBM         000100 ..... ..... ..... 00000000000    @VX
+VADDUHM         000100 ..... ..... ..... 00001000000    @VX
+VADDUWM         000100 ..... ..... ..... 00010000000    @VX
+VADDUDM         000100 ..... ..... ..... 00011000000    @VX
 VADDUBS         000100 ..... ..... ..... 01000000000    @VX
 VADDUHS         000100 ..... ..... ..... 01001000000    @VX
 VADDUWS         000100 ..... ..... ..... 01010000000    @VX
@@ -1006,6 +1016,10 @@ VSUBSBS         000100 ..... ..... ..... 11100000000    @VX
 VSUBSHS         000100 ..... ..... ..... 11101000000    @VX
 VSUBSWS         000100 ..... ..... ..... 11110000000    @VX
 
+VSUBUBM         000100 ..... ..... ..... 10000000000    @VX
+VSUBUHM         000100 ..... ..... ..... 10001000000    @VX
+VSUBUWM         000100 ..... ..... ..... 10010000000    @VX
+VSUBUDM         000100 ..... ..... ..... 10011000000    @VX
 VSUBUBS         000100 ..... ..... ..... 11000000000    @VX
 VSUBUHS         000100 ..... ..... ..... 11001000000    @VX
 VSUBUWS         000100 ..... ..... ..... 11010000000    @VX
@@ -1099,6 +1113,13 @@ VMULHSD         000100 ..... ..... ..... 01111001001    @VX
 VMULHUD         000100 ..... ..... ..... 01011001001    @VX
 VMULLD          000100 ..... ..... ..... 00111001001    @VX
 
+VMULUWM         000100 ..... ..... ..... 00010001001    @VX
+
+VMUL10UQ        000100 ..... ..... ..... 01000000001    @VX
+VMUL10EUQ       000100 ..... ..... ..... 01001000001    @VX
+VMUL10ECUQ      000100 ..... ..... ..... 00001000001    @VX
+VMUL10CUQ       000100 ..... ..... ..... 00000000001    @VX
+
 ## Vector Multiply-Sum Instructions
 
 VMSUMUBM        000100 ..... ..... ..... ..... 100100   @VA
@@ -1141,6 +1162,24 @@ VPKSWSS         000100 ..... ..... ..... 0011100111 0    @VX
 VPKSDSS         000100 ..... ..... ..... 1011100111 0    @VX
 VPKPX           000100 ..... ..... ..... 0110000111 0    @VX
 
+## Decimal Integer Arithmetic Instructions
+
+BCDADD          000100 ..... ..... ..... 1.000000001    @VX_ps
+BCDSUB          000100 ..... ..... ..... 1.001000001    @VX_ps
+BCDUS           000100 ..... ..... ..... 10010000001    @VX
+BCDS            000100 ..... ..... ..... 1.011000001    @VX_ps
+BCDCTSQ         000100 ..... 00000 ..... 10110000001    @VX_tb
+BCDCTN          000100 ..... 00101 ..... 10110000001    @VX_tb
+BCDCFSQ         000100 ..... 00010 ..... 1.110000001    @VX_tb_ps
+BCDCTZ          000100 ..... 00100 ..... 1.110000001    @VX_tb_ps
+BCDCFZ          000100 ..... 00110 ..... 1.110000001    @VX_tb_ps
+BCDCFN          000100 ..... 00111 ..... 1.110000001    @VX_tb_ps
+BCDSETSGN       000100 ..... 11111 ..... 1.110000001    @VX_tb_ps
+BCDCPSGN        000100 ..... ..... ..... 01101000001    @VX
+BCDSR           000100 ..... ..... ..... 1.111000001    @VX_ps
+BCDTRUNC        000100 ..... ..... ..... 1.100000001    @VX_ps
+BCDUTRUNC       000100 ..... ..... ..... 10101000001    @VX
+
 # VSX Load/Store Instructions
 
 LXSD            111001 ..... ..... .............. 10    @DS
target/ppc/int_helper.c +16 -16
index 8e8076a92c..86346f5b9a 100644
@@ -2348,7 +2348,7 @@ static void bcd_sub_mag(ppc_avr_t *t, ppc_avr_t *a, ppc_avr_t *b, int *invalid,
     *overflow = carry;
 }
 
-uint32_t helper_bcdadd(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDADD(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
 
     int sgna = bcd_get_sgn(a);
@@ -2395,7 +2395,7 @@ uint32_t helper_bcdadd(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdsub(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDSUB(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     ppc_avr_t bcopy = *b;
     int sgnb = bcd_get_sgn(b);
@@ -2406,10 +2406,10 @@ uint32_t helper_bcdsub(ppc_avr_t *r,  ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     }
     /* else invalid ... defer to bcdadd code for proper handling */
 
-    return helper_bcdadd(r, a, &bcopy, ps);
+    return helper_BCDADD(r, a, &bcopy, ps);
 }
 
-uint32_t helper_bcdcfn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCFN(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2445,7 +2445,7 @@ uint32_t helper_bcdcfn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdctn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCTN(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2479,7 +2479,7 @@ uint32_t helper_bcdctn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdcfz(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCFZ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2523,7 +2523,7 @@ uint32_t helper_bcdcfz(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdctz(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCTZ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr = 0;
@@ -2582,7 +2582,7 @@ static inline int ucmp128(uint64_t alo, uint64_t ahi,
         (ahi > bhi ? 1 : -1);
 }
 
-uint32_t helper_bcdcfsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCFSQ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int cr;
@@ -2640,7 +2640,7 @@ uint32_t helper_bcdcfsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdctsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCTSQ(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     uint8_t i;
     int cr;
@@ -2680,7 +2680,7 @@ uint32_t helper_bcdctsq(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdcpsgn(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDCPSGN(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     int invalid = 0;
@@ -2703,7 +2703,7 @@ uint32_t helper_bcdcpsgn(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return bcd_cmp_zero(r);
 }
 
-uint32_t helper_bcdsetsgn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDSETSGN(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
 {
     int sgnb = bcd_get_sgn(b);
 
@@ -2717,7 +2717,7 @@ uint32_t helper_bcdsetsgn(ppc_avr_t *r, ppc_avr_t *b, uint32_t ps)
     return bcd_cmp_zero(r);
 }
 
-uint32_t helper_bcds(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDS(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int cr;
     int i = a->VsrSB(7);
@@ -2753,7 +2753,7 @@ uint32_t helper_bcds(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdus(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDUS(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int cr;
     int i;
@@ -2790,7 +2790,7 @@ uint32_t helper_bcdus(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdsr(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDSR(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int cr;
     int unused = 0;
@@ -2836,7 +2836,7 @@ uint32_t helper_bcdsr(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return cr;
 }
 
-uint32_t helper_bcdtrunc(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDTRUNC(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     uint64_t mask;
     uint32_t ox_flag = 0;
@@ -2869,7 +2869,7 @@ uint32_t helper_bcdtrunc(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
     return bcd_cmp_zero(&ret) | ox_flag;
 }
 
-uint32_t helper_bcdutrunc(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
+uint32_t helper_BCDUTRUNC(ppc_avr_t *r, ppc_avr_t *a, ppc_avr_t *b, uint32_t ps)
 {
     int i;
     uint64_t mask;
target/ppc/translate/vmx-impl.c.inc +129 -170
index 65230f3461..4e7fe22156 100644
@@ -142,7 +142,8 @@ static void gen_mtvscr(DisasContext *ctx)
     gen_helper_mtvscr(tcg_env, val);
 }
 
-static void gen_vx_vmul10(DisasContext *ctx, bool add_cin, bool ret_carry)
+static bool do_vx_vmul10(DisasContext *ctx, arg_VX *a,
+                         bool add_cin, bool ret_carry)
 {
     TCGv_i64 t0;
     TCGv_i64 t1;
@@ -150,10 +151,7 @@ static void gen_vx_vmul10(DisasContext *ctx, bool add_cin, bool ret_carry)
     TCGv_i64 avr;
     TCGv_i64 ten, z;
 
-    if (unlikely(!ctx->altivec_enabled)) {
-        gen_exception(ctx, POWERPC_EXCP_VPU);
-        return;
-    }
+    REQUIRE_VECTOR(ctx);
 
     t0 = tcg_temp_new_i64();
     t1 = tcg_temp_new_i64();
@@ -163,54 +161,32 @@ static void gen_vx_vmul10(DisasContext *ctx, bool add_cin, bool ret_carry)
     z = tcg_constant_i64(0);
 
     if (add_cin) {
-        get_avr64(avr, rA(ctx->opcode), false);
+        get_avr64(avr, a->vra, false);
         tcg_gen_mulu2_i64(t0, t1, avr, ten);
-        get_avr64(avr, rB(ctx->opcode), false);
+        get_avr64(avr, a->vrb, false);
         tcg_gen_andi_i64(t2, avr, 0xF);
         tcg_gen_add2_i64(avr, t2, t0, t1, t2, z);
-        set_avr64(rD(ctx->opcode), avr, false);
+        set_avr64(a->vrt, avr, false);
     } else {
-        get_avr64(avr, rA(ctx->opcode), false);
+        get_avr64(avr, a->vra, false);
         tcg_gen_mulu2_i64(avr, t2, avr, ten);
-        set_avr64(rD(ctx->opcode), avr, false);
+        set_avr64(a->vrt, avr, false);
     }
 
     if (ret_carry) {
-        get_avr64(avr, rA(ctx->opcode), true);
+        get_avr64(avr, a->vra, true);
         tcg_gen_mulu2_i64(t0, t1, avr, ten);
         tcg_gen_add2_i64(t0, avr, t0, t1, t2, z);
-        set_avr64(rD(ctx->opcode), avr, false);
-        set_avr64(rD(ctx->opcode), z, true);
+        set_avr64(a->vrt, avr, false);
+        set_avr64(a->vrt, z, true);
     } else {
-        get_avr64(avr, rA(ctx->opcode), true);
+        get_avr64(avr, a->vra, true);
         tcg_gen_mul_i64(t0, avr, ten);
         tcg_gen_add_i64(avr, t0, t2);
-        set_avr64(rD(ctx->opcode), avr, true);
+        set_avr64(a->vrt, avr, true);
     }
-}
-
-#define GEN_VX_VMUL10(name, add_cin, ret_carry)                         \
-    static void glue(gen_, name)(DisasContext *ctx)                     \
-    { gen_vx_vmul10(ctx, add_cin, ret_carry); }
-
-GEN_VX_VMUL10(vmul10uq, 0, 0);
-GEN_VX_VMUL10(vmul10euq, 1, 0);
-GEN_VX_VMUL10(vmul10cuq, 0, 1);
-GEN_VX_VMUL10(vmul10ecuq, 1, 1);
 
-#define GEN_VXFORM_V(name, vece, tcg_op, opc2, opc3)                    \
-static void glue(gen_, name)(DisasContext *ctx)                         \
-{                                                                       \
-    if (unlikely(!ctx->altivec_enabled)) {                              \
-        gen_exception(ctx, POWERPC_EXCP_VPU);                           \
-        return;                                                         \
-    }                                                                   \
-                                                                        \
-    tcg_op(vece,                                                        \
-           avr_full_offset(rD(ctx->opcode)),                            \
-           avr_full_offset(rA(ctx->opcode)),                            \
-           avr_full_offset(rB(ctx->opcode)),                            \
-           16, 16);                                                     \
+    return true;
 }
 
 #define GEN_VXFORM(name, opc2, opc3)                                    \
@@ -338,19 +314,6 @@ static void glue(gen_, name)(DisasContext *ctx)                         \
     gen_helper_##name(cpu_gpr[rD(ctx->opcode)], cpu_gpr[rA(ctx->opcode)], rb); \
 }
 
-GEN_VXFORM_V(vaddubm, MO_8, tcg_gen_gvec_add, 0, 0);
-GEN_VXFORM_DUAL_EXT(vaddubm, PPC_ALTIVEC, PPC_NONE, 0,       \
-                    vmul10cuq, PPC_NONE, PPC2_ISA300, 0x0000F800)
-GEN_VXFORM_V(vadduhm, MO_16, tcg_gen_gvec_add, 0, 1);
-GEN_VXFORM_DUAL(vadduhm, PPC_ALTIVEC, PPC_NONE,  \
-                vmul10ecuq, PPC_NONE, PPC2_ISA300)
-GEN_VXFORM_V(vadduwm, MO_32, tcg_gen_gvec_add, 0, 2);
-GEN_VXFORM_V(vaddudm, MO_64, tcg_gen_gvec_add, 0, 3);
-GEN_VXFORM_V(vsububm, MO_8, tcg_gen_gvec_sub, 0, 16);
-GEN_VXFORM_V(vsubuhm, MO_16, tcg_gen_gvec_sub, 0, 17);
-GEN_VXFORM_V(vsubuwm, MO_32, tcg_gen_gvec_sub, 0, 18);
-GEN_VXFORM_V(vsubudm, MO_64, tcg_gen_gvec_sub, 0, 19);
-
 static bool do_vmrg(DisasContext *ctx, arg_VX *a,
                     void (*helper)(TCGv_ptr, TCGv_ptr, TCGv_ptr))
 {
@@ -370,6 +333,38 @@ TRANS_FLAGS(ALTIVEC, VMRGLB, do_vmrg, gen_helper_VMRGLB);
 TRANS_FLAGS(ALTIVEC, VMRGLH, do_vmrg, gen_helper_VMRGLH);
 TRANS_FLAGS(ALTIVEC, VMRGLW, do_vmrg, gen_helper_VMRGLW);
 
+static bool do_vx_vaddsubm(DisasContext *ctx, arg_VX *a, MemOp vece,
+                           void (*gen_op)(unsigned, uint32_t, uint32_t,
+                                          uint32_t, uint32_t, uint32_t))
+{
+    REQUIRE_VECTOR(ctx);
+
+    gen_op(vece,
+           avr_full_offset(a->vrt),
+           avr_full_offset(a->vra),
+           avr_full_offset(a->vrb),
+           16, 16);
+
+    return true;
+}
+
+TRANS_FLAGS(ALTIVEC, VADDUBM, do_vx_vaddsubm, MO_8, tcg_gen_gvec_add)
+TRANS_FLAGS(ALTIVEC, VADDUHM, do_vx_vaddsubm, MO_16, tcg_gen_gvec_add)
+TRANS_FLAGS(ALTIVEC, VADDUWM, do_vx_vaddsubm, MO_32, tcg_gen_gvec_add)
+TRANS_FLAGS2(ALTIVEC_207, VADDUDM, do_vx_vaddsubm, MO_64, tcg_gen_gvec_add)
+
+TRANS_FLAGS(ALTIVEC, VSUBUBM, do_vx_vaddsubm, MO_8, tcg_gen_gvec_sub)
+TRANS_FLAGS(ALTIVEC, VSUBUHM, do_vx_vaddsubm, MO_16, tcg_gen_gvec_sub)
+TRANS_FLAGS(ALTIVEC, VSUBUWM, do_vx_vaddsubm, MO_32, tcg_gen_gvec_sub)
+TRANS_FLAGS2(ALTIVEC_207, VSUBUDM, do_vx_vaddsubm, MO_64, tcg_gen_gvec_sub)
+
+TRANS_FLAGS2(ALTIVEC_207, VMULUWM, do_vx_vaddsubm, MO_32, tcg_gen_gvec_mul)
+
+TRANS_FLAGS2(ISA300, VMUL10CUQ, do_vx_vmul10, false, true)
+TRANS_FLAGS2(ISA300, VMUL10ECUQ, do_vx_vmul10, true, true)
+TRANS_FLAGS2(ISA300, VMUL10UQ, do_vx_vmul10, false, false)
+TRANS_FLAGS2(ISA300, VMUL10EUQ, do_vx_vmul10, true, false)
+
 static void trans_vmrgew(DisasContext *ctx)
 {
     int VT = rD(ctx->opcode);
@@ -684,7 +679,6 @@ static void trans_vclzd(DisasContext *ctx)
     set_avr64(VT, avr, false);
 }
 
-GEN_VXFORM_V(vmuluwm, MO_32, tcg_gen_gvec_mul, 4, 2);
 GEN_VXFORM(vsrv, 2, 28);
 GEN_VXFORM(vslv, 2, 29);
 GEN_VXFORM(vslo, 6, 16);
@@ -2523,126 +2517,93 @@ GEN_VXFORM(vpmsumb, 4, 16)
 GEN_VXFORM(vpmsumh, 4, 17)
 GEN_VXFORM(vpmsumw, 4, 18)
 
-#define GEN_BCD(op)                                 \
-static void gen_##op(DisasContext *ctx)             \
-{                                                   \
-    TCGv_ptr ra, rb, rd;                            \
-    TCGv_i32 ps;                                    \
-                                                    \
-    if (unlikely(!ctx->altivec_enabled)) {          \
-        gen_exception(ctx, POWERPC_EXCP_VPU);       \
-        return;                                     \
-    }                                               \
-                                                    \
-    ra = gen_avr_ptr(rA(ctx->opcode));              \
-    rb = gen_avr_ptr(rB(ctx->opcode));              \
-    rd = gen_avr_ptr(rD(ctx->opcode));              \
-                                                    \
-    ps = tcg_constant_i32((ctx->opcode & 0x200) != 0); \
-                                                    \
-    gen_helper_##op(cpu_crf[6], rd, ra, rb, ps);    \
-}
-
-#define GEN_BCD2(op)                                \
-static void gen_##op(DisasContext *ctx)             \
-{                                                   \
-    TCGv_ptr rd, rb;                                \
-    TCGv_i32 ps;                                    \
-                                                    \
-    if (unlikely(!ctx->altivec_enabled)) {          \
-        gen_exception(ctx, POWERPC_EXCP_VPU);       \
-        return;                                     \
-    }                                               \
-                                                    \
-    rb = gen_avr_ptr(rB(ctx->opcode));              \
-    rd = gen_avr_ptr(rD(ctx->opcode));              \
-                                                    \
-    ps = tcg_constant_i32((ctx->opcode & 0x200) != 0); \
-                                                    \
-    gen_helper_##op(cpu_crf[6], rd, rb, ps);        \
-}
-
-GEN_BCD(bcdadd)
-GEN_BCD(bcdsub)
-GEN_BCD2(bcdcfn)
-GEN_BCD2(bcdctn)
-GEN_BCD2(bcdcfz)
-GEN_BCD2(bcdctz)
-GEN_BCD2(bcdcfsq)
-GEN_BCD2(bcdctsq)
-GEN_BCD2(bcdsetsgn)
-GEN_BCD(bcdcpsgn);
-GEN_BCD(bcds);
-GEN_BCD(bcdus);
-GEN_BCD(bcdsr);
-GEN_BCD(bcdtrunc);
-GEN_BCD(bcdutrunc);
-
-static void gen_xpnd04_1(DisasContext *ctx)
-{
-    switch (opc4(ctx->opcode)) {
-    case 0:
-        gen_bcdctsq(ctx);
-        break;
-    case 2:
-        gen_bcdcfsq(ctx);
-        break;
-    case 4:
-        gen_bcdctz(ctx);
-        break;
-    case 5:
-        gen_bcdctn(ctx);
-        break;
-    case 6:
-        gen_bcdcfz(ctx);
-        break;
-    case 7:
-        gen_bcdcfn(ctx);
-        break;
-    case 31:
-        gen_bcdsetsgn(ctx);
-        break;
-    default:
-        gen_invalid(ctx);
-        break;
-    }
+static bool do_bcd_ps(DisasContext *ctx, arg_VX_ps *a,
+                   void (*helper)(TCGv_i32, TCGv_ptr, TCGv_ptr,
+                                  TCGv_ptr, TCGv_i32))
+{
+    TCGv_ptr ra, rb, rd;
+    TCGv_i32 ps;
+
+    REQUIRE_VECTOR(ctx);
+
+    ra = gen_avr_ptr(a->vra);
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+    ps = tcg_constant_i32(a->ps);
+
+    helper(cpu_crf[6], rd, ra, rb, ps);
+    return true;
 }
 
-static void gen_xpnd04_2(DisasContext *ctx)
-{
-    switch (opc4(ctx->opcode)) {
-    case 0:
-        gen_bcdctsq(ctx);
-        break;
-    case 2:
-        gen_bcdcfsq(ctx);
-        break;
-    case 4:
-        gen_bcdctz(ctx);
-        break;
-    case 6:
-        gen_bcdcfz(ctx);
-        break;
-    case 7:
-        gen_bcdcfn(ctx);
-        break;
-    case 31:
-        gen_bcdsetsgn(ctx);
-        break;
-    default:
-        gen_invalid(ctx);
-        break;
-    }
+static bool do_bcd(DisasContext *ctx, arg_VX *a,
+                   void (*helper)(TCGv_i32, TCGv_ptr, TCGv_ptr,
+                                  TCGv_ptr, TCGv_i32))
+{
+    TCGv_ptr ra, rb, rd;
+    TCGv_i32 ps;
+
+    REQUIRE_VECTOR(ctx);
+
+    ra = gen_avr_ptr(a->vra);
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+    ps = tcg_constant_i32(0);
+
+    helper(cpu_crf[6], rd, ra, rb, ps);
+    return true;
 }
 
-GEN_VXFORM_DUAL(vsububm, PPC_ALTIVEC, PPC_NONE, \
-                bcdadd, PPC_NONE, PPC2_ALTIVEC_207)
-GEN_VXFORM_DUAL(vsubuhm, PPC_ALTIVEC, PPC_NONE, \
-                bcdsub, PPC_NONE, PPC2_ALTIVEC_207)
-GEN_VXFORM_DUAL(vsubudm, PPC2_ALTIVEC_207, PPC_NONE, \
-                bcds, PPC_NONE, PPC2_ISA300)
-GEN_VXFORM_DUAL(vsubuwm, PPC_ALTIVEC, PPC_NONE, \
-                bcdus, PPC_NONE, PPC2_ISA300)
+static bool do_bcd_tb(DisasContext *ctx, arg_VX_tb *a,
+                    void (*helper)(TCGv_i32, TCGv_ptr, TCGv_ptr, TCGv_i32))
+{
+    TCGv_ptr rd, rb;
+    TCGv_i32 ps;
+
+    REQUIRE_VECTOR(ctx);
+
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+
+    ps = tcg_constant_i32(0);
+
+    helper(cpu_crf[6], rd, rb, ps);
+
+    return true;
+}
+
+static bool do_bcd_tb_ps(DisasContext *ctx, arg_VX_tb_ps *a,
+                    void (*helper)(TCGv_i32, TCGv_ptr, TCGv_ptr, TCGv_i32))
+{
+    TCGv_ptr rd, rb;
+    TCGv_i32 ps;
+
+    REQUIRE_VECTOR(ctx);
+
+    rb = gen_avr_ptr(a->vrb);
+    rd = gen_avr_ptr(a->vrt);
+
+    ps = tcg_constant_i32(a->ps);
+
+    helper(cpu_crf[6], rd, rb, ps);
+
+    return true;
+}
+
+TRANS_FLAGS2(ALTIVEC_207, BCDADD, do_bcd_ps, gen_helper_BCDADD)
+TRANS_FLAGS2(ALTIVEC_207, BCDSUB, do_bcd_ps, gen_helper_BCDSUB)
+TRANS_FLAGS2(ISA300, BCDUS, do_bcd, gen_helper_BCDUS)
+TRANS_FLAGS2(ISA300, BCDS, do_bcd_ps, gen_helper_BCDS)
+TRANS_FLAGS2(ISA300, BCDCFN, do_bcd_tb_ps, gen_helper_BCDCFN)
+TRANS_FLAGS2(ISA300, BCDCTN, do_bcd_tb, gen_helper_BCDCTN)
+TRANS_FLAGS2(ISA300, BCDCFZ, do_bcd_tb_ps, gen_helper_BCDCFZ)
+TRANS_FLAGS2(ISA300, BCDCTZ, do_bcd_tb_ps, gen_helper_BCDCTZ)
+TRANS_FLAGS2(ISA300, BCDCFSQ, do_bcd_tb_ps, gen_helper_BCDCFSQ)
+TRANS_FLAGS2(ISA300, BCDCTSQ, do_bcd_tb, gen_helper_BCDCTSQ)
+TRANS_FLAGS2(ISA300, BCDSETSGN, do_bcd_tb_ps, gen_helper_BCDSETSGN)
+TRANS_FLAGS2(ISA300, BCDCPSGN, do_bcd, gen_helper_BCDCPSGN)
+TRANS_FLAGS2(ISA300, BCDTRUNC, do_bcd_ps, gen_helper_BCDTRUNC)
+TRANS_FLAGS2(ISA300, BCDUTRUNC, do_bcd, gen_helper_BCDUTRUNC)
+TRANS_FLAGS2(ISA300, BCDSR, do_bcd_ps, gen_helper_BCDSR)
 
 static void gen_vsbox(DisasContext *ctx)
 {
@@ -3502,5 +3463,3 @@ TRANS_FLAGS2(ISA310, VMODUQ, do_vx_helper, gen_helper_VMODUQ)
 #undef GEN_VXFORM_NOA
 #undef GEN_VXFORM_UIMM
 #undef GEN_VAFORM_PAIRED
-
-#undef GEN_BCD2
target/ppc/translate/vmx-ops.c.inc +0 -22
index 2820976c0f..c1951c6975 100644
@@ -24,39 +24,17 @@ GEN_HANDLER_E(name0##_##name1, 0x4, opc2, opc3, 0x00000000, type0, type1)
 GEN_HANDLER_E(name0##_##name1, 0x4, opc2, opc3, 0x00000000, tp0, tp1), \
 GEN_HANDLER_E(name0##_##name1, 0x4, opc2, (opc3 | 0x10), 0x00000000, tp0, tp1),
 
-GEN_VXFORM_DUAL(vaddubm, vmul10cuq, 0, 0, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM_DUAL(vadduhm, vmul10ecuq, 0, 1, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM(vadduwm, 0, 2),
-GEN_VXFORM_207(vaddudm, 0, 3),
-GEN_VXFORM_DUAL(vsububm, bcdadd, 0, 16, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM_DUAL(vsubuhm, bcdsub, 0, 17, PPC_ALTIVEC, PPC_NONE),
-GEN_VXFORM_DUAL(vsubuwm, bcdus, 0, 18, PPC_ALTIVEC, PPC2_ISA300),
-GEN_VXFORM_DUAL(vsubudm, bcds, 0, 19, PPC2_ALTIVEC_207, PPC2_ISA300),
-GEN_VXFORM_300(bcds, 0, 27),
 GEN_VXFORM_300(vextublx, 6, 24),
 GEN_VXFORM_300(vextuhlx, 6, 25),
 GEN_VXFORM_DUAL(vmrgow, vextuwlx, 6, 26, PPC_NONE, PPC2_ALTIVEC_207),
 GEN_VXFORM_300(vextubrx, 6, 28),
 GEN_VXFORM_300(vextuhrx, 6, 29),
 GEN_VXFORM_DUAL(vmrgew, vextuwrx, 6, 30, PPC_NONE, PPC2_ALTIVEC_207),
-GEN_VXFORM_207(vmuluwm, 4, 2),
 GEN_VXFORM_300(vsrv, 2, 28),
 GEN_VXFORM_300(vslv, 2, 29),
 GEN_VXFORM(vslo, 6, 16),
 GEN_VXFORM(vsro, 6, 17),
 
-GEN_VXFORM(xpnd04_1, 0, 22),
-GEN_VXFORM_300(bcdsr, 0, 23),
-GEN_VXFORM_300(bcdsr, 0, 31),
-GEN_VXFORM_300_EXT(vmul10uq, 0, 8, 0x0000F800),
-GEN_VXFORM_300(vmul10euq, 0, 9),
-GEN_VXFORM_300(bcdcpsgn, 0, 13),
-GEN_VXFORM_207(bcdadd, 0, 24),
-GEN_VXFORM_207(bcdsub, 0, 25),
-GEN_VXFORM_300(bcdtrunc, 0, 28),
-GEN_VXFORM_300(xpnd04_2, 0, 30),
-GEN_VXFORM_300(bcdtrunc, 0, 20),
-GEN_VXFORM_300(bcdutrunc, 0, 21),
 GEN_VXFORM(vsl, 2, 7),
 GEN_VXFORM(vsr, 2, 11),
 GEN_VXFORM(vsum4ubs, 4, 24),
Signature · 1 line
2.55.0
#56 lore raw

PATCH v2 25/37 target/ppc: Move rlwimi, rlwinm instructions to decodetree

In reply to Chinmay Rath

From: Tanushree Shah <[email protected]>

-Moving the following instructions to decodetree specification:

            rlwimi                           : M-form
            rlwimi.                          : M-form
            rlwinm                           : M-form
            rlwinm.                          : M-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the "-d in_asm,op" flag.

Additionally, validated using small assembly tests confirming the destination register is correctly updated based on rotated and masked source values and confirmed that the value was same before and after the change

---
 target/ppc/insn32.decode                   |   6 ++
 target/ppc/translate.c                     | 106 ---------------------
 target/ppc/translate/fixedpoint-impl.c.inc |  98 +++++++++++++++++++
 3 files changed, 104 insertions(+), 106 deletions(-)
target/ppc/insn32.decode +6 -0
index eed963bd71..8f027e7ff5 100644
@@ -312,6 +312,9 @@
 
 @Z23_te_tbp     ...... ....0 te:5 ....0 rmc:2 ........ rc:1     &Z23_te_tb frt=%z23_frtp frb=%z23_frbp
 
+&M              ra rs sh mb me rc:bool
+@M              ...... rs:5 ra:5 sh:5 mb:5 me:5 rc:1 &M
+
 ### Fixed-Point Load Instructions
 
 LBZ             100010 ..... ..... ................     @D
@@ -561,6 +564,9 @@ SRADI           011111 ..... ..... ..... 110011101 . .  @XS
 
 EXTSWSLI        011111 ..... ..... ..... 110111101 . .  @XS
 
+RLWIMI          010100 ..... ..... ..... ..... ......   @M
+RLWINM          010101 ..... ..... ..... ..... ......   @M
+
 ## BCD Assist
 
 ADDG6S          011111 ..... ..... ..... - 001001010 -  @X
target/ppc/translate.c +0 -106

118 lines not on this page · raw message

target/ppc/translate/fixedpoint-impl.c.inc +98 -0

104 lines not on this page · raw message

Signature · 1 line
2.55.0
#57 lore raw

Re: [PATCH v2 25/37] target/ppc: Move rlwimi, rlwinm instructions to decodetree

Quoted 186 lines · Chinmay Rath

From: Tanushree Shah <[email protected]>

-Moving the following instructions to decodetree specification:

            rlwimi                           : M-form
            rlwimi.                          : M-form
            rlwinm                           : M-form
            rlwinm.                          : M-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the "-d in_asm,op" flag.

Additionally, validated using small assembly tests confirming the destination register is correctly updated based on rotated and masked source values and confirmed that the value was same before and after the change

---
 target/ppc/insn32.decode                   |   6 ++
 target/ppc/translate.c                     | 106 ---------------------
 target/ppc/translate/fixedpoint-impl.c.inc |  98 +++++++++++++++++++
 3 files changed, 104 insertions(+), 106 deletions(-)
target/ppc/insn32.decode +6 -0
index eed963bd71..8f027e7ff5 100644
@@ -312,6 +312,9 @@
 
 @Z23_te_tbp     ...... ....0 te:5 ....0 rmc:2 ........ rc:1     &Z23_te_tb frt=%z23_frtp frb=%z23_frbp
 
+&M              ra rs sh mb me rc:bool
+@M              ...... rs:5 ra:5 sh:5 mb:5 me:5 rc:1 &M
+
 ### Fixed-Point Load Instructions
 
 LBZ             100010 ..... ..... ................     @D
@@ -561,6 +564,9 @@ SRADI           011111 ..... ..... ..... 110011101 . .  @XS
 
 EXTSWSLI        011111 ..... ..... ..... 110111101 . .  @XS
 
+RLWIMI          010100 ..... ..... ..... ..... ......   @M
+RLWINM          010101 ..... ..... ..... ..... ......   @M
+
 ## BCD Assist
 
 ADDG6S          011111 ..... ..... ..... - 001001010 -  @X
target/ppc/translate.c +0 -106
index 712990cf2d..d2c51b4d59 100644
@@ -2012,110 +2012,6 @@ static void gen_pause(DisasContext *ctx)
 
 /***                             Integer rotate                            ***/
 
-/* rlwimi & rlwimi. */
-static void gen_rlwimi(DisasContext *ctx)
-{
-    TCGv t_ra = cpu_gpr[rA(ctx->opcode)];
-    TCGv t_rs = cpu_gpr[rS(ctx->opcode)];
-    uint32_t sh = SH(ctx->opcode);
-    uint32_t mb = MB(ctx->opcode);
-    uint32_t me = ME(ctx->opcode);
-
-    if (sh == (31 - me) && mb <= me) {
-        tcg_gen_deposit_tl(t_ra, t_ra, t_rs, sh, me - mb + 1);
-    } else {
-        target_ulong mask;
-        bool mask_in_32b = true;
-        TCGv t1;
-
-#if defined(TARGET_PPC64)
-        mb += 32;
-        me += 32;
-#endif
-        mask = MASK(mb, me);
-
-#if defined(TARGET_PPC64)
-        if (mask > 0xffffffffu) {
-            mask_in_32b = false;
-        }
-#endif
-        t1 = tcg_temp_new();
-        if (mask_in_32b) {
-            TCGv_i32 t0 = tcg_temp_new_i32();
-            tcg_gen_trunc_tl_i32(t0, t_rs);
-            tcg_gen_rotli_i32(t0, t0, sh);
-            tcg_gen_extu_i32_tl(t1, t0);
-        } else {
-#if defined(TARGET_PPC64)
-            tcg_gen_deposit_i64(t1, t_rs, t_rs, 32, 32);
-            tcg_gen_rotli_i64(t1, t1, sh);
-#else
-            g_assert_not_reached();
-#endif
-        }
-
-        tcg_gen_andi_tl(t1, t1, mask);
-        tcg_gen_andi_tl(t_ra, t_ra, ~mask);
-        tcg_gen_or_tl(t_ra, t_ra, t1);
-    }
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, t_ra);
-    }
-}
-
-/* rlwinm & rlwinm. */
-static void gen_rlwinm(DisasContext *ctx)
-{
-    TCGv t_ra = cpu_gpr[rA(ctx->opcode)];
-    TCGv t_rs = cpu_gpr[rS(ctx->opcode)];
-    int sh = SH(ctx->opcode);
-    int mb = MB(ctx->opcode);
-    int me = ME(ctx->opcode);
-    int len = me - mb + 1;
-    int rsh = (32 - sh) & 31;
-
-    if (sh != 0 && len > 0 && me == (31 - sh)) {
-        tcg_gen_deposit_z_tl(t_ra, t_rs, sh, len);
-    } else if (me == 31 && rsh + len <= 32) {
-        tcg_gen_extract_tl(t_ra, t_rs, rsh, len);
-    } else {
-        target_ulong mask;
-        bool mask_in_32b = true;
-#if defined(TARGET_PPC64)
-        mb += 32;
-        me += 32;
-#endif
-        mask = MASK(mb, me);
-#if defined(TARGET_PPC64)
-        if (mask > 0xffffffffu) {
-            mask_in_32b = false;
-        }
-#endif
-        if (mask_in_32b) {
-            if (sh == 0) {
-                tcg_gen_andi_tl(t_ra, t_rs, mask);
-            } else {
-                TCGv_i32 t0 = tcg_temp_new_i32();
-                tcg_gen_trunc_tl_i32(t0, t_rs);
-                tcg_gen_rotli_i32(t0, t0, sh);
-                tcg_gen_andi_i32(t0, t0, mask);
-                tcg_gen_extu_i32_tl(t_ra, t0);
-            }
-        } else {
-#if defined(TARGET_PPC64)
-            tcg_gen_deposit_i64(t_ra, t_rs, t_rs, 32, 32);
-            tcg_gen_rotli_i64(t_ra, t_ra, sh);
-            tcg_gen_andi_i64(t_ra, t_ra, mask);
-#else
-            g_assert_not_reached();
-#endif
-        }
-    }
-    if (unlikely(Rc(ctx->opcode) != 0)) {
-        gen_set_Rc0(ctx, t_ra);
-    }
-}
-
 /* rlwnm & rlwnm. */
 static void gen_rlwnm(DisasContext *ctx)
 {
@@ -4807,8 +4703,6 @@ GEN_HANDLER(invalid, 0x00, 0x00, 0x00, 0xFFFFFFFF, PPC_NONE),
 GEN_HANDLER_E(copy, 0x1F, 0x06, 0x18, 0x03C00001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(cp_abort, 0x1F, 0x06, 0x1A, 0x03FFF801, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(paste, 0x1F, 0x06, 0x1C, 0x03C00000, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER(rlwimi, 0x14, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(rlwinm, 0x15, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER(rlwnm, 0x17, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 /* handles lfdp, lxsd, lxssp */
 GEN_HANDLER_E(dform39, 0x39, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
target/ppc/translate/fixedpoint-impl.c.inc +5 -0
index 82fb645279..8fcee7a89d 100644
@@ -1874,6 +1874,104 @@ static bool trans_SRAWI(DisasContext *ctx, arg_SRAWI *a)
     return true;
 }
 
+static bool trans_RLWIMI(DisasContext *ctx, arg_RLWIMI *a)
+{
+    TCGv t_ra = cpu_gpr[a->ra];
+    TCGv t_rs = cpu_gpr[a->rs];
+

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_SRAWI(DisasContext *ctx, arg_SRAWI *a) · line 1891

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +10 -0
+    if (a->sh == (31 - a->me) && a->mb <= a->me) {
+        tcg_gen_deposit_tl(t_ra, t_ra, t_rs, a->sh, a->me - a->mb + 1);
+    } else {
+        target_ulong mask;
+        bool mask_in_32b = true;
+        TCGv t1;
+
+#if defined(TARGET_PPC64)
+        a->mb += 32;
+        a->me += 32;

The old gen_rlwimi/gen_rlwinm used local int mb/int me for this adjustment, keeping the opcode struct immutable. Writing back into a-> fields is contrary to normal decodetree style — the arg_* struct is an input, not a scratch pad. Please use local copies instead:

target/ppc/translate/fixedpoint-impl.c.inc static bool trans_SRAWI(DisasContext *ctx, arg_SRAWI *a) · line 1893

Quoted 2 lines
target/ppc/translate/fixedpoint-impl.c.inc +2 -0
+#endif
+        mask = MASK((uint32_t)a->mb, (uint32_t)a->me);

MASK() takes uint64_t on PPC64 builds. Both translators in this patch should use the same form. If the local-variable suggestion above is adopted, the cast becomes unnecessary in both (the local int will promote cleanly), and both can just use MASK(mb, me).

Quoted 87 lines
target/ppc/translate/fixedpoint-impl.c.inc +81 -0
+
+#if defined(TARGET_PPC64)
+        if (mask > 0xffffffffu) {
+            mask_in_32b = false;
+        }
+#endif
+        t1 = tcg_temp_new();
+        if (mask_in_32b) {
+            TCGv_i32 t0 = tcg_temp_new_i32();
+            tcg_gen_trunc_tl_i32(t0, t_rs);
+            tcg_gen_rotli_i32(t0, t0, a->sh);
+            tcg_gen_extu_i32_tl(t1, t0);
+        } else {
+#if defined(TARGET_PPC64)
+            tcg_gen_deposit_i64(t1, t_rs, t_rs, 32, 32);
+            tcg_gen_rotli_i64(t1, t1, a->sh);
+#else
+            g_assert_not_reached();
+#endif
+        }
+
+        tcg_gen_andi_tl(t1, t1, mask);
+        tcg_gen_andi_tl(t_ra, t_ra, ~mask);
+        tcg_gen_or_tl(t_ra, t_ra, t1);
+    }
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, t_ra);
+    }
+    return true;
+}
+
+static bool trans_RLWINM(DisasContext *ctx, arg_RLWINM *a)
+{
+    TCGv t_ra = cpu_gpr[a->ra];
+    TCGv t_rs = cpu_gpr[a->rs];
+    int len = a->me - a->mb + 1;
+    int rsh = (32 - a->sh) & 31;
+
+    if (a->sh != 0 && len > 0 && a->me == (31 - a->sh)) {
+        tcg_gen_deposit_z_tl(t_ra, t_rs, a->sh, len);
+    } else if (a->me == 31 && rsh + len <= 32) {
+        tcg_gen_extract_tl(t_ra, t_rs, rsh, len);
+    } else {
+        target_ulong mask;
+        bool mask_in_32b = true;
+#if defined(TARGET_PPC64)
+        a->mb += 32;
+        a->me += 32;
+#endif
+        mask = MASK(a->mb, a->me);
+#if defined(TARGET_PPC64)
+        if (mask > 0xffffffffu) {
+            mask_in_32b = false;
+        }
+#endif
+        if (mask_in_32b) {
+            if (a->sh == 0) {
+                tcg_gen_andi_tl(t_ra, t_rs, mask);
+            } else {
+                TCGv_i32 t0 = tcg_temp_new_i32();
+                tcg_gen_trunc_tl_i32(t0, t_rs);
+                tcg_gen_rotli_i32(t0, t0, a->sh);
+                tcg_gen_andi_i32(t0, t0, mask);
+                tcg_gen_extu_i32_tl(t_ra, t0);
+            }
+        } else {
+#if defined(TARGET_PPC64)
+            tcg_gen_deposit_i64(t_ra, t_rs, t_rs, 32, 32);
+            tcg_gen_rotli_i64(t_ra, t_ra, a->sh);
+            tcg_gen_andi_i64(t_ra, t_ra, mask);
+#else
+            g_assert_not_reached();
+#endif
+        }
+    }
+    if (unlikely(a->rc)) {
+        gen_set_Rc0(ctx, t_ra);
+    }
+    return true;
+}
+
 static void do_fetch_inc_conditional(DisasContext *ctx, MemOp memop,
                                      TCGv EA, int rt,
                                      TCGCond cond, int addend)
Signature · 1 line
2.55.0
#58 lore raw

PATCH v2 26/37 target/ppc: Move lmw, stmw instructions to decodetree

In reply to Chinmay Rath

From: Tanushree Shah <[email protected]>

Moving the following instructions to decodetree specification:
lmw : D-form
stmw : D-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the "-d in_asm,op" flag.

Additionally, validated using small assembly tests confirming the the correct register store/load behavior. Also, confirmed the register values are same before and after the change.

---
 target/ppc/helper.h                        |  4 +--
 target/ppc/insn32.decode                   |  3 ++
 target/ppc/mem_helper.c                    |  4 +--
 target/ppc/translate.c                     | 39 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 32 ++++++++++++++++++
 5 files changed, 39 insertions(+), 43 deletions(-)
target/ppc/helper.h +2 -2
index 8425eceeb3..3659408c46 100644
@@ -43,8 +43,8 @@ DEF_HELPER_1(check_tlb_flush_local, void, env)
 DEF_HELPER_1(check_tlb_flush_global, void, env)
 #endif
 
-DEF_HELPER_3(lmw, void, env, tl, i32)
-DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
+DEF_HELPER_3(LMW, void, env, tl, i32)
+DEF_HELPER_FLAGS_3(STMW, TCG_CALL_NO_WG, void, env, tl, i32)
 DEF_HELPER_4(LSW, void, env, tl, i32, i32)
 DEF_HELPER_5(LSWX, void, env, tl, i32, i32, i32)
 DEF_HELPER_FLAGS_4(STSW, TCG_CALL_NO_WG, void, env, tl, i32, i32)
target/ppc/insn32.decode +3 -0
index 8f027e7ff5..b0fa85b93e 100644
@@ -354,6 +354,7 @@ LVDX            011111 ..... ..... ..... 0000010001 -   @X
 LSKU            111010 ..... ..... ............. 0 11   @DD
 LCXU            111010 ..... ..... ............. 1 11   @DD
 
+LMW             101110 ..... ..... ................     @D
 
 ### Fixed-Point Atomic Load/Store Instructions
 
@@ -399,6 +400,8 @@ STBCIX          011111 ..... ..... ..... 1111010101 -   @X
 STHBRX          011111 ..... ..... ..... 1110010110 -   @X
 STWBRX          011111 ..... ..... ..... 1010010110 -   @X
 
+STMW            101111 ..... ..... ................     @D
+
 ### Fixed-Point Compare Instructions
 
 CMP             011111 ... - . ..... ..... 0000000000 - @X_bfl
target/ppc/mem_helper.c +2 -2
index 582180cf88..7517b3954b 100644
@@ -73,7 +73,7 @@ static void *probe_contiguous(CPUPPCState *env, target_ulong addr, uint32_t nb,
     return NULL;
 }
 
-void helper_lmw(CPUPPCState *env, target_ulong addr, uint32_t reg)
+void helper_LMW(CPUPPCState *env, target_ulong addr, uint32_t reg)
 {
     uintptr_t raddr = GETPC();
     int mmu_idx = ppc_env_mmu_index(env, false);
@@ -98,7 +98,7 @@ void helper_lmw(CPUPPCState *env, target_ulong addr, uint32_t reg)
     }
 }
 
-void helper_stmw(CPUPPCState *env, target_ulong addr, uint32_t reg)
+void helper_STMW(CPUPPCState *env, target_ulong addr, uint32_t reg)
 {
     uintptr_t raddr = GETPC();
     int mmu_idx = ppc_env_mmu_index(env, false);
target/ppc/translate.c +0 -39
index d2c51b4d59..9db9d1e070 100644
@@ -2455,43 +2455,6 @@ GEN_LDX(lwbr, ld32ur, 0x16, 0x10, PPC_INTEGER);
 GEN_LDX_E(ldbr, ld64ur_i64, 0x14, 0x10, PPC_NONE, PPC2_DBRX, CHK_NONE);
 #endif  /* TARGET_PPC64 */
 
-/***                    Integer load and store multiple                    ***/
-
-/* lmw */
-static void gen_lmw(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1;
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    t1 = tcg_constant_i32(rD(ctx->opcode));
-    gen_addr_imm_index(ctx, t0, 0);
-    gen_helper_lmw(tcg_env, t0, t1);
-}
-
-/* stmw */
-static void gen_stmw(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1;
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    t1 = tcg_constant_i32(rS(ctx->opcode));
-    gen_addr_imm_index(ctx, t0, 0);
-    gen_helper_stmw(tcg_env, t0, t1);
-}
-
-
 #if !defined(CONFIG_USER_ONLY)
 static inline void gen_check_tlb_flush(DisasContext *ctx, bool global)
 {
@@ -4708,8 +4671,6 @@ GEN_HANDLER(rlwnm, 0x17, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER_E(dform39, 0x39, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
 /* handles stfdp, stxsd, stxssp */
 GEN_HANDLER_E(dform3D, 0x3D, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
-GEN_HANDLER(lmw, 0x2E, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(stmw, 0x2F, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
target/ppc/translate/fixedpoint-impl.c.inc +32 -0
index 8fcee7a89d..a75279022c 100644
@@ -288,6 +288,38 @@ static bool trans_STQCX(DisasContext *ctx, arg_STQCX *a)
     return true;
 }
 
+/* Load/Store Multiple Word */
+static bool do_ldst_multiple(DisasContext *ctx, arg_D *a, bool store)
+{
+    TCGv ea;
+    TCGv_i32 reg;
+
+    REQUIRE_INSNS_FLAGS(ctx, INTEGER);
+
+    /* Little-endian mode is not supported for multiple word operations */
+    if (ctx->le_mode) {
+        gen_align_no_le(ctx);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_INT);
+
+    reg = tcg_constant_i32(a->rt);
+    ea = do_ea_calc(ctx, a->ra, tcg_constant_tl(a->si));
+
+    /* Call the appropriate helper function */
+    if (store) {
+        gen_helper_STMW(tcg_env, ea, reg);
+    } else {
+        gen_helper_LMW(tcg_env, ea, reg);
+    }
+
+    return true;
+}
+
+TRANS(LMW, do_ldst_multiple, false)
+TRANS(STMW, do_ldst_multiple, true)
+
 /*
  * Fixed-Point Compare Instructions
  */
Signature · 1 line
2.55.0
Reviewed-by on v2 26/37
#59 lore raw

Re: [PATCH v2 26/37] target/ppc: Move lmw, stmw instructions to decodetree

Quoted 17 lines · Chinmay Rath

From: Tanushree Shah <[email protected]>

Moving the following instructions to decodetree specification:
lmw : D-form
stmw : D-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the "-d in_asm,op" flag.

Additionally, validated using small assembly tests confirming the the correct register store/load behavior. Also, confirmed the register values are same before and after the change.

Quoted 7 lines
---
 target/ppc/helper.h                        |  4 +--
 target/ppc/insn32.decode                   |  3 ++
 target/ppc/mem_helper.c                    |  4 +--
 target/ppc/translate.c                     | 39 ----------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 32 ++++++++++++++++++
 5 files changed, 39 insertions(+), 43 deletions(-)
Quoted 162 lines
target/ppc/helper.h +2 -2
index 8425eceeb3..3659408c46 100644
@@ -43,8 +43,8 @@ DEF_HELPER_1(check_tlb_flush_local, void, env)
 DEF_HELPER_1(check_tlb_flush_global, void, env)
 #endif
 
-DEF_HELPER_3(lmw, void, env, tl, i32)
-DEF_HELPER_FLAGS_3(stmw, TCG_CALL_NO_WG, void, env, tl, i32)
+DEF_HELPER_3(LMW, void, env, tl, i32)
+DEF_HELPER_FLAGS_3(STMW, TCG_CALL_NO_WG, void, env, tl, i32)
 DEF_HELPER_4(LSW, void, env, tl, i32, i32)
 DEF_HELPER_5(LSWX, void, env, tl, i32, i32, i32)
 DEF_HELPER_FLAGS_4(STSW, TCG_CALL_NO_WG, void, env, tl, i32, i32)
target/ppc/insn32.decode +3 -0
index 8f027e7ff5..b0fa85b93e 100644
@@ -354,6 +354,7 @@ LVDX            011111 ..... ..... ..... 0000010001 -   @X
 LSKU            111010 ..... ..... ............. 0 11   @DD
 LCXU            111010 ..... ..... ............. 1 11   @DD
 
+LMW             101110 ..... ..... ................     @D
 
 ### Fixed-Point Atomic Load/Store Instructions
 
@@ -399,6 +400,8 @@ STBCIX          011111 ..... ..... ..... 1111010101 -   @X
 STHBRX          011111 ..... ..... ..... 1110010110 -   @X
 STWBRX          011111 ..... ..... ..... 1010010110 -   @X
 
+STMW            101111 ..... ..... ................     @D
+
 ### Fixed-Point Compare Instructions
 
 CMP             011111 ... - . ..... ..... 0000000000 - @X_bfl
target/ppc/mem_helper.c +2 -2
index 582180cf88..7517b3954b 100644
@@ -73,7 +73,7 @@ static void *probe_contiguous(CPUPPCState *env, target_ulong addr, uint32_t nb,
     return NULL;
 }
 
-void helper_lmw(CPUPPCState *env, target_ulong addr, uint32_t reg)
+void helper_LMW(CPUPPCState *env, target_ulong addr, uint32_t reg)
 {
     uintptr_t raddr = GETPC();
     int mmu_idx = ppc_env_mmu_index(env, false);
@@ -98,7 +98,7 @@ void helper_lmw(CPUPPCState *env, target_ulong addr, uint32_t reg)
     }
 }
 
-void helper_stmw(CPUPPCState *env, target_ulong addr, uint32_t reg)
+void helper_STMW(CPUPPCState *env, target_ulong addr, uint32_t reg)
 {
     uintptr_t raddr = GETPC();
     int mmu_idx = ppc_env_mmu_index(env, false);
target/ppc/translate.c +0 -39
index d2c51b4d59..9db9d1e070 100644
@@ -2455,43 +2455,6 @@ GEN_LDX(lwbr, ld32ur, 0x16, 0x10, PPC_INTEGER);
 GEN_LDX_E(ldbr, ld64ur_i64, 0x14, 0x10, PPC_NONE, PPC2_DBRX, CHK_NONE);
 #endif  /* TARGET_PPC64 */
 
-/***                    Integer load and store multiple                    ***/
-
-/* lmw */
-static void gen_lmw(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1;
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    t1 = tcg_constant_i32(rD(ctx->opcode));
-    gen_addr_imm_index(ctx, t0, 0);
-    gen_helper_lmw(tcg_env, t0, t1);
-}
-
-/* stmw */
-static void gen_stmw(DisasContext *ctx)
-{
-    TCGv t0;
-    TCGv_i32 t1;
-
-    if (ctx->le_mode) {
-        gen_align_no_le(ctx);
-        return;
-    }
-    gen_set_access_type(ctx, ACCESS_INT);
-    t0 = tcg_temp_new();
-    t1 = tcg_constant_i32(rS(ctx->opcode));
-    gen_addr_imm_index(ctx, t0, 0);
-    gen_helper_stmw(tcg_env, t0, t1);
-}
-
-
 #if !defined(CONFIG_USER_ONLY)
 static inline void gen_check_tlb_flush(DisasContext *ctx, bool global)
 {
@@ -4708,8 +4671,6 @@ GEN_HANDLER(rlwnm, 0x17, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER_E(dform39, 0x39, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
 /* handles stfdp, stxsd, stxssp */
 GEN_HANDLER_E(dform3D, 0x3D, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
-GEN_HANDLER(lmw, 0x2E, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
-GEN_HANDLER(stmw, 0x2F, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 /* ISA v3.0 changed the extended opcode from 62 to 30 */
 GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
 #if defined(TARGET_PPC64)
target/ppc/translate/fixedpoint-impl.c.inc +32 -0
index 8fcee7a89d..a75279022c 100644
@@ -288,6 +288,38 @@ static bool trans_STQCX(DisasContext *ctx, arg_STQCX *a)
     return true;
 }
 
+/* Load/Store Multiple Word */
+static bool do_ldst_multiple(DisasContext *ctx, arg_D *a, bool store)
+{
+    TCGv ea;
+    TCGv_i32 reg;
+
+    REQUIRE_INSNS_FLAGS(ctx, INTEGER);
+
+    /* Little-endian mode is not supported for multiple word operations */
+    if (ctx->le_mode) {
+        gen_align_no_le(ctx);
+        return true;
+    }
+
+    gen_set_access_type(ctx, ACCESS_INT);
+
+    reg = tcg_constant_i32(a->rt);
+    ea = do_ea_calc(ctx, a->ra, tcg_constant_tl(a->si));
+
+    /* Call the appropriate helper function */
+    if (store) {
+        gen_helper_STMW(tcg_env, ea, reg);
+    } else {
+        gen_helper_LMW(tcg_env, ea, reg);
+    }
+
+    return true;
+}
+
+TRANS(LMW, do_ldst_multiple, false)
+TRANS(STMW, do_ldst_multiple, true)
+
 /*
  * Fixed-Point Compare Instructions
  */
Signature · 1 line
2.55.0
#60 lore raw

PATCH v2 27/37 target/ppc: Move mfmsr, mtmsr[d] instructions to decodetree

In reply to Chinmay Rath

From: Vishal Chourasia <[email protected]>

Moving the following instructions to decodetree specification:
mfmsr : X-form
mtmsr[d] : X-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the -d in_asm,op flag.

This also includes improvements from review feedback:
- Gate MFMSR/MTMSR on PPC_MISC via TRANS_FLAGS()
- Use UINT32_MAX instead of 0xFFFFFFFF for clarity

---
 target/ppc/insn32.decode             |  5 ++
 target/ppc/translate.c               | 97 ----------------------------
 target/ppc/translate/misc-impl.c.inc | 64 ++++++++++++++++++
 3 files changed, 69 insertions(+), 97 deletions(-)
target/ppc/insn32.decode +5 -0
index b0fa85b93e..f7ab08704f 100644
@@ -315,6 +315,11 @@
 &M              ra rs sh mb me rc:bool
 @M              ...... rs:5 ra:5 sh:5 mb:5 me:5 rc:1 &M
 
+### MSR
+MFMSR           011111 ..... ----- ----- 0001010011 -   @X_t
+MTMSR           011111 ..... ---- . ----- 0010010010 -  @X_rs_l
+MTMSRD          011111 ..... ---- . ----- 0010110010 -  @X_rs_l
+
 ### Fixed-Point Load Instructions
 
 LBZ             100010 ..... ..... ................     @D
target/ppc/translate.c +0 -97

121 lines not on this page · raw message

target/ppc/translate/misc-impl.c.inc +64 -0

67 lines not on this page · raw message

Signature · 1 line
2.55.0
Reviewed-by on v2 27/37
#61 lore raw

Re: [PATCH v2 27/37] target/ppc: Move mfmsr, mtmsr[d] instructions to decodetree

Quoted 14 lines · Chinmay Rath

From: Vishal Chourasia <[email protected]>

Moving the following instructions to decodetree specification:
mfmsr : X-form
mtmsr[d] : X-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the -d in_asm,op flag.

This also includes improvements from review feedback:
- Gate MFMSR/MTMSR on PPC_MISC via TRANS_FLAGS()
- Use UINT32_MAX instead of 0xFFFFFFFF for clarity

Quoted 8 lines
---
 target/ppc/insn32.decode             |  5 ++
 target/ppc/translate.c               | 97 ----------------------------
 target/ppc/translate/misc-impl.c.inc | 64 ++++++++++++++++++
 3 files changed, 69 insertions(+), 97 deletions(-)
Quoted 220 lines
target/ppc/insn32.decode +5 -0
index b0fa85b93e..f7ab08704f 100644
@@ -315,6 +315,11 @@
 &M              ra rs sh mb me rc:bool
 @M              ...... rs:5 ra:5 sh:5 mb:5 me:5 rc:1 &M
 
+### MSR
+MFMSR           011111 ..... ----- ----- 0001010011 -   @X_t
+MTMSR           011111 ..... ---- . ----- 0010010010 -  @X_rs_l
+MTMSRD          011111 ..... ---- . ----- 0010110010 -  @X_rs_l
+
 ### Fixed-Point Load Instructions
 
 LBZ             100010 ..... ..... ................     @D
target/ppc/translate.c +0 -97
index 9db9d1e070..9d29134fa1 100644
@@ -2843,13 +2843,6 @@ static void gen_mcrxrx(DisasContext *ctx)
 }
 #endif
 
-/* mfmsr */
-static void gen_mfmsr(DisasContext *ctx)
-{
-    CHK_SV(ctx);
-    tcg_gen_mov_tl(cpu_gpr[rD(ctx->opcode)], cpu_msr);
-}
-
 /* mfspr */
 static inline void gen_op_mfspr(DisasContext *ctx)
 {
@@ -2923,93 +2916,6 @@ static void gen_mftb(DisasContext *ctx)
     gen_op_mfspr(ctx);
 }
 
-/* mtmsr */
-#if defined(TARGET_PPC64)
-static void gen_mtmsrd(DisasContext *ctx)
-{
-    if (unlikely(!is_book3s_arch2x(ctx))) {
-        gen_invalid(ctx);
-        return;
-    }
-
-    CHK_SV(ctx);
-
-#if !defined(CONFIG_USER_ONLY)
-    TCGv t0, t1;
-    target_ulong mask;
-
-    t0 = tcg_temp_new();
-    t1 = tcg_temp_new();
-
-    translator_io_start(&ctx->base);
-
-    if (ctx->opcode & 0x00010000) {
-        /* L=1 form only updates EE and RI */
-        mask = (1ULL << MSR_RI) | (1ULL << MSR_EE);
-    } else {
-        /* mtmsrd does not alter HV, S, ME, or LE */
-        mask = ~((1ULL << MSR_LE) | (1ULL << MSR_ME) | (1ULL << MSR_S) |
-                 (1ULL << MSR_HV));
-        /*
-         * XXX: we need to update nip before the store if we enter
-         *      power saving mode, we will exit the loop directly from
-         *      ppc_store_msr
-         */
-        gen_update_nip(ctx, ctx->base.pc_next);
-    }
-
-    tcg_gen_andi_tl(t0, cpu_gpr[rS(ctx->opcode)], mask);
-    tcg_gen_andi_tl(t1, cpu_msr, ~mask);
-    tcg_gen_or_tl(t0, t0, t1);
-
-    gen_helper_store_msr(tcg_env, t0);
-
-    /* Must stop the translation as machine state (may have) changed */
-    ctx->base.is_jmp = DISAS_EXIT_UPDATE;
-#endif /* !defined(CONFIG_USER_ONLY) */
-}
-#endif /* defined(TARGET_PPC64) */
-
-static void gen_mtmsr(DisasContext *ctx)
-{
-    CHK_SV(ctx);
-
-#if !defined(CONFIG_USER_ONLY)
-    TCGv t0, t1;
-    target_ulong mask = 0xFFFFFFFF;
-
-    t0 = tcg_temp_new();
-    t1 = tcg_temp_new();
-
-    translator_io_start(&ctx->base);
-    if (ctx->opcode & 0x00010000) {
-        /* L=1 form only updates EE and RI */
-        mask &= (1ULL << MSR_RI) | (1ULL << MSR_EE);
-    } else {
-        if (likely(!(ctx->insns_flags2 & PPC2_PPE42))) {
-            /* mtmsr does not alter S, ME, or LE */
-            mask &= ~((1ULL << MSR_LE) | (1ULL << MSR_ME) | (1ULL << MSR_S));
-        }
-
-        /*
-         * XXX: we need to update nip before the store if we enter
-         *      power saving mode, we will exit the loop directly from
-         *      ppc_store_msr
-         */
-        gen_update_nip(ctx, ctx->base.pc_next);
-    }
-
-    tcg_gen_andi_tl(t0, cpu_gpr[rS(ctx->opcode)], mask);
-    tcg_gen_andi_tl(t1, cpu_msr, ~mask);
-    tcg_gen_or_tl(t0, t0, t1);
-
-    gen_helper_store_msr(tcg_env, t0);
-
-    /* Must stop the translation as machine state (may have) changed */
-    ctx->base.is_jmp = DISAS_EXIT_UPDATE;
-#endif
-}
-
 /* mtspr */
 static void gen_mtspr(DisasContext *ctx)
 {
@@ -4687,15 +4593,12 @@ GEN_HANDLER(hrfid, 0x13, 0x12, 0x08, 0x03FF8001, PPC_64H),
 GEN_HANDLER(sc, 0x11, 0x11, 0xFF, 0x03FFF01D, PPC_FLOW),
 GEN_HANDLER(sc, 0x11, 0x01, 0xFF, 0x03FFF01D, PPC_FLOW),
 GEN_HANDLER(mcrxr, 0x1F, 0x00, 0x10, 0x007FF801, PPC_MISC),
-GEN_HANDLER(mfmsr, 0x1F, 0x13, 0x02, 0x001FF801, PPC_MISC),
 GEN_HANDLER(mfspr, 0x1F, 0x13, 0x0A, 0x00000001, PPC_MISC),
 GEN_HANDLER(mftb, 0x1F, 0x13, 0x0B, 0x00000001, PPC_MFTB),
 #if defined(TARGET_PPC64)
-GEN_HANDLER(mtmsrd, 0x1F, 0x12, 0x05, 0x001EF801, PPC_64B),
 GEN_HANDLER_E(setb, 0x1F, 0x00, 0x04, 0x0003F801, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(mcrxrx, 0x1F, 0x00, 0x12, 0x007FF801, PPC_NONE, PPC2_ISA300),
 #endif
-GEN_HANDLER(mtmsr, 0x1F, 0x12, 0x04, 0x001EF801, PPC_MISC),
 GEN_HANDLER(mtspr, 0x1F, 0x13, 0x0E, 0x00000000, PPC_MISC),
 GEN_HANDLER(mfsr, 0x1F, 0x13, 0x12, 0x0010F801, PPC_SEGMENT),
 GEN_HANDLER(mfsrin, 0x1F, 0x13, 0x14, 0x001F0001, PPC_SEGMENT),
target/ppc/translate/misc-impl.c.inc +64 -0
index 890fbb209e..54712f9b73 100644
@@ -161,3 +161,67 @@ static bool trans_MCRF(DisasContext *ctx, arg_MCRF *a)
     tcg_gen_mov_i32(cpu_crf[a->bf], cpu_crf[a->bfa]);
     return true;
 }
+
+static bool do_mfmsr(DisasContext *ctx, arg_X_t *a)
+{
+    REQUIRE_SV(ctx);
+    tcg_gen_mov_tl(cpu_gpr[a->rt], cpu_msr);
+    return true;
+}
+TRANS_FLAGS(MISC, MFMSR, do_mfmsr)
+
+static bool do_mtmsr(DisasContext *ctx, arg_X_rs_l *a, bool is_mtmsrd)
+{
+    REQUIRE_SV(ctx);
+#if !defined(CONFIG_USER_ONLY)
+    TCGv t0, t1;
+    target_ulong mask;
+
+    if (is_mtmsrd) {
+        if (unlikely(!is_book3s_arch2x(ctx))) {
+            gen_invalid(ctx);
+            return true;
+        }
+        mask = ~(target_ulong)0;
+    } else {
+        mask = UINT32_MAX;
+    }
+
+    t0 = tcg_temp_new();
+    t1 = tcg_temp_new();
+    translator_io_start(&ctx->base);
+
+    if (a->l) {
+        /* L=1 form only updates EE and RI */
+        mask &= (1ULL << MSR_RI) | (1ULL << MSR_EE);
+    } else {
+        if (is_mtmsrd) {
+            /* mtmsrd does not alter HV, S, ME, or LE */
+            mask &= ~((1ULL << MSR_LE) | (1ULL << MSR_ME) |
+                      (1ULL << MSR_S)  | (1ULL << MSR_HV));
+        } else if (likely(!(ctx->insns_flags2 & PPC2_PPE42))) {
+            /* mtmsr does not alter S, ME, or LE */
+            mask &= ~((1ULL << MSR_LE) | (1ULL << MSR_ME) | (1ULL << MSR_S));
+        }
+        /*
+         * XXX: we need to update nip before the store if we enter
+         *      power saving mode, we will exit the loop directly from
+         *      ppc_store_msr
+         */
+        gen_update_nip(ctx, ctx->base.pc_next);
+    }
+
+    tcg_gen_andi_tl(t0, cpu_gpr[a->rs], mask);
+    tcg_gen_andi_tl(t1, cpu_msr, ~mask);
+    tcg_gen_or_tl(t0, t0, t1);
+
+    gen_helper_store_msr(tcg_env, t0);
+
+    /* Must stop the translation as machine state (may have) changed */
+    ctx->base.is_jmp = DISAS_EXIT_UPDATE;
+#endif /* !CONFIG_USER_ONLY */
+    return true;
+}
+
+TRANS_FLAGS(MISC, MTMSR, do_mtmsr, false)
+TRANS64(MTMSRD, do_mtmsr, true)
Signature · 1 line
2.55.0
#62 lore raw

PATCH v2 28/37 target/ppc: Move byte-reverse instructions to decodetree

In reply to Chinmay Rath

From: Vishal Chourasia <[email protected]>

Moving the following instructions to decodetree specification:
brd, brw, brh : X-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the -d in_asm,op flag.

This also includes improvements from review feedback:
- Add TARGET_PPC64 guards with qemu_build_not_reached() for 32-bit builds

---
 target/ppc/insn32.decode                   |  5 +++
 target/ppc/translate.c                     | 35 -------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 40 ++++++++++++++++++++++
 3 files changed, 45 insertions(+), 35 deletions(-)
target/ppc/insn32.decode +5 -0
index f7ab08704f..7614655f72 100644
@@ -320,6 +320,11 @@ MFMSR           011111 ..... ----- ----- 0001010011 -   @X_t
 MTMSR           011111 ..... ---- . ----- 0010010010 -  @X_rs_l
 MTMSRD          011111 ..... ---- . ----- 0010110010 -  @X_rs_l
 
+### Fixed-Point Byte-Reverse Instructions
+BRW             011111 ..... ..... ----- 0010011011 -   @X_sa
+BRD             011111 ..... ..... ----- 0010111011 -   @X_sa
+BRH             011111 ..... ..... ----- 0011011011 -   @X_sa
+
 ### Fixed-Point Load Instructions
 
 LBZ             100010 ..... ..... ................     @D
target/ppc/translate.c +0 -35
index 9d29134fa1..8950476be7 100644
@@ -4532,42 +4532,7 @@ static void gen_dform3D(DisasContext *ctx)
     return gen_invalid(ctx);
 }
 
-#if defined(TARGET_PPC64)
-/* brd */
-static void gen_brd(DisasContext *ctx)
-{
-    tcg_gen_bswap64_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rS(ctx->opcode)]);
-}
-
-/* brw */
-static void gen_brw(DisasContext *ctx)
-{
-    tcg_gen_bswap64_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rS(ctx->opcode)]);
-    tcg_gen_rotli_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rA(ctx->opcode)], 32);
-
-}
-
-/* brh */
-static void gen_brh(DisasContext *ctx)
-{
-    TCGv_i64 mask = tcg_constant_i64(0x00ff00ff00ff00ffull);
-    TCGv_i64 t1 = tcg_temp_new_i64();
-    TCGv_i64 t2 = tcg_temp_new_i64();
-
-    tcg_gen_shri_i64(t1, cpu_gpr[rS(ctx->opcode)], 8);
-    tcg_gen_and_i64(t2, t1, mask);
-    tcg_gen_and_i64(t1, cpu_gpr[rS(ctx->opcode)], mask);
-    tcg_gen_shli_i64(t1, t1, 8);
-    tcg_gen_or_i64(cpu_gpr[rA(ctx->opcode)], t1, t2);
-}
-#endif
-
 static opcode_t opcodes[] = {
-#if defined(TARGET_PPC64)
-GEN_HANDLER_E(brd, 0x1F, 0x1B, 0x05, 0x0000F801, PPC_NONE, PPC2_ISA310),
-GEN_HANDLER_E(brw, 0x1F, 0x1B, 0x04, 0x0000F801, PPC_NONE, PPC2_ISA310),
-GEN_HANDLER_E(brh, 0x1F, 0x1B, 0x06, 0x0000F801, PPC_NONE, PPC2_ISA310),
-#endif
 GEN_HANDLER(invalid, 0x00, 0x00, 0x00, 0xFFFFFFFF, PPC_NONE),
 GEN_HANDLER_E(copy, 0x1F, 0x06, 0x18, 0x03C00001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(cp_abort, 0x1F, 0x06, 0x1A, 0x03FFF801, PPC_NONE, PPC2_ISA300),
target/ppc/translate/fixedpoint-impl.c.inc +40 -0
index a75279022c..cef67d6431 100644
@@ -17,6 +17,46 @@
  * License along with this library; if not, see <http://www.gnu.org/licenses/>.
  */
 
+/*
+ * Byte-Reverse Instructions
+ */
+static bool do_byte_reverse_X_sa(DisasContext *ctx, arg_X_sa *a, int kind)
+{
+#if defined(TARGET_PPC64)
+    switch (kind) {
+    case 0: /* brd */
+        tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
+        break;
+    case 1: /* brw */
+        tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
+        tcg_gen_rotli_i64(cpu_gpr[a->ra], cpu_gpr[a->ra], 32);
+        break;
+    case 2: /* brh */
+    {
+        TCGv_i64 mask = tcg_constant_i64(0x00ff00ff00ff00ffull);
+        TCGv_i64 t1 = tcg_temp_new_i64();
+        TCGv_i64 t2 = tcg_temp_new_i64();
+
+        tcg_gen_shri_i64(t1, cpu_gpr[a->rs], 8);
+        tcg_gen_and_i64(t2, t1, mask);
+        tcg_gen_and_i64(t1, cpu_gpr[a->rs], mask);
+        tcg_gen_shli_i64(t1, t1, 8);
+        tcg_gen_or_i64(cpu_gpr[a->ra], t1, t2);
+        break;
+    }
+    default:
+        g_assert_not_reached();
+    }
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+
+TRANS64_FLAGS2(ISA310, BRD, do_byte_reverse_X_sa, 0)
+TRANS64_FLAGS2(ISA310, BRW, do_byte_reverse_X_sa, 1)
+TRANS64_FLAGS2(ISA310, BRH, do_byte_reverse_X_sa, 2)
+
 /*
  * Fixed-Point Load/Store Instructions
  */
Signature · 1 line
2.55.0
#63 lore raw

Re: [PATCH v2 28/37] target/ppc: Move byte-reverse instructions to decodetree

Quoted 121 lines · Chinmay Rath

From: Vishal Chourasia <[email protected]>

Moving the following instructions to decodetree specification:
brd, brw, brh : X-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the -d in_asm,op flag.

This also includes improvements from review feedback:
- Add TARGET_PPC64 guards with qemu_build_not_reached() for 32-bit builds

---
 target/ppc/insn32.decode                   |  5 +++
 target/ppc/translate.c                     | 35 -------------------
 target/ppc/translate/fixedpoint-impl.c.inc | 40 ++++++++++++++++++++++
 3 files changed, 45 insertions(+), 35 deletions(-)
target/ppc/insn32.decode +5 -0
index f7ab08704f..7614655f72 100644
@@ -320,6 +320,11 @@ MFMSR           011111 ..... ----- ----- 0001010011 -   @X_t
 MTMSR           011111 ..... ---- . ----- 0010010010 -  @X_rs_l
 MTMSRD          011111 ..... ---- . ----- 0010110010 -  @X_rs_l
 
+### Fixed-Point Byte-Reverse Instructions
+BRW             011111 ..... ..... ----- 0010011011 -   @X_sa
+BRD             011111 ..... ..... ----- 0010111011 -   @X_sa
+BRH             011111 ..... ..... ----- 0011011011 -   @X_sa
+
 ### Fixed-Point Load Instructions
 
 LBZ             100010 ..... ..... ................     @D
target/ppc/translate.c +0 -35
index 9d29134fa1..8950476be7 100644
@@ -4532,42 +4532,7 @@ static void gen_dform3D(DisasContext *ctx)
     return gen_invalid(ctx);
 }
 
-#if defined(TARGET_PPC64)
-/* brd */
-static void gen_brd(DisasContext *ctx)
-{
-    tcg_gen_bswap64_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rS(ctx->opcode)]);
-}
-
-/* brw */
-static void gen_brw(DisasContext *ctx)
-{
-    tcg_gen_bswap64_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rS(ctx->opcode)]);
-    tcg_gen_rotli_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rA(ctx->opcode)], 32);
-
-}
-
-/* brh */
-static void gen_brh(DisasContext *ctx)
-{
-    TCGv_i64 mask = tcg_constant_i64(0x00ff00ff00ff00ffull);
-    TCGv_i64 t1 = tcg_temp_new_i64();
-    TCGv_i64 t2 = tcg_temp_new_i64();
-
-    tcg_gen_shri_i64(t1, cpu_gpr[rS(ctx->opcode)], 8);
-    tcg_gen_and_i64(t2, t1, mask);
-    tcg_gen_and_i64(t1, cpu_gpr[rS(ctx->opcode)], mask);
-    tcg_gen_shli_i64(t1, t1, 8);
-    tcg_gen_or_i64(cpu_gpr[rA(ctx->opcode)], t1, t2);
-}
-#endif
-
 static opcode_t opcodes[] = {
-#if defined(TARGET_PPC64)
-GEN_HANDLER_E(brd, 0x1F, 0x1B, 0x05, 0x0000F801, PPC_NONE, PPC2_ISA310),
-GEN_HANDLER_E(brw, 0x1F, 0x1B, 0x04, 0x0000F801, PPC_NONE, PPC2_ISA310),
-GEN_HANDLER_E(brh, 0x1F, 0x1B, 0x06, 0x0000F801, PPC_NONE, PPC2_ISA310),
-#endif
 GEN_HANDLER(invalid, 0x00, 0x00, 0x00, 0xFFFFFFFF, PPC_NONE),
 GEN_HANDLER_E(copy, 0x1F, 0x06, 0x18, 0x03C00001, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(cp_abort, 0x1F, 0x06, 0x1A, 0x03FFF801, PPC_NONE, PPC2_ISA300),
target/ppc/translate/fixedpoint-impl.c.inc +29 -0
index a75279022c..cef67d6431 100644
@@ -17,6 +17,46 @@
  * License along with this library; if not, see <http://www.gnu.org/licenses/>.
  */
 
+/*
+ * Byte-Reverse Instructions
+ */
+static bool do_byte_reverse_X_sa(DisasContext *ctx, arg_X_sa *a, int kind)
+{
+#if defined(TARGET_PPC64)
+    switch (kind) {
+    case 0: /* brd */
+        tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
+        break;
+    case 1: /* brw */
+        tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
+        tcg_gen_rotli_i64(cpu_gpr[a->ra], cpu_gpr[a->ra], 32);
+        break;
+    case 2: /* brh */
+    {
+        TCGv_i64 mask = tcg_constant_i64(0x00ff00ff00ff00ffull);
+        TCGv_i64 t1 = tcg_temp_new_i64();
+        TCGv_i64 t2 = tcg_temp_new_i64();
+
+        tcg_gen_shri_i64(t1, cpu_gpr[a->rs], 8);
+        tcg_gen_and_i64(t2, t1, mask);
+        tcg_gen_and_i64(t1, cpu_gpr[a->rs], mask);
+        tcg_gen_shli_i64(t1, t1, 8);
+        tcg_gen_or_i64(cpu_gpr[a->ra], t1, t2);
+        break;
+    }
+    default:
+        g_assert_not_reached();

target/ppc/translate/fixedpoint-impl.c.inc line 58

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +10 -0
+    }
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+
+TRANS64_FLAGS2(ISA310, BRD, do_byte_reverse_X_sa, 0)
+TRANS64_FLAGS2(ISA310, BRW, do_byte_reverse_X_sa, 1)
+TRANS64_FLAGS2(ISA310, BRH, do_byte_reverse_X_sa, 2)

The kind values 0, 1, 2 are opaque at the call site — a reader has to cross-reference the switch inside do_byte_reverse_X_sa to know which is which. The more common pattern in this series is three small separate translators, which avoids the switch entirely and is self-documenting:

static bool trans_BRD(DisasContext *ctx, arg_BRD *a)

{
    REQUIRE_64BIT(ctx);
    REQUIRE_INSNS_FLAGS2(ctx, ISA310);
    tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
    return true;
}

...and similarly for BRW and BRH. This also eliminates the do_byte_reverse_X_sa wrapper and the #else qemu_build_not_reached() block entirely, since REQUIRE_64BIT already handles 32-bit builds.

~Amit

#64 lore raw

Re: [PATCH v2 28/37] target/ppc: Move byte-reverse instructions to decodetree

Quoted 145 lines · Amit Machhiwal
Quoted 121 lines · Chinmay Rath

From: Vishal Chourasia <[email protected]>

Moving the following instructions to decodetree specification:
brd, brw, brh : X-form

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the -d in_asm,op flag.

This also includes improvements from review feedback:
- Add TARGET_PPC64 guards with qemu_build_not_reached() for 32-bit builds

---
  target/ppc/insn32.decode                   |  5 +++
  target/ppc/translate.c                     | 35 -------------------
  target/ppc/translate/fixedpoint-impl.c.inc | 40 ++++++++++++++++++++++
  3 files changed, 45 insertions(+), 35 deletions(-)
target/ppc/insn32.decode +5 -0
index f7ab08704f..7614655f72 100644
@@ -320,6 +320,11 @@ MFMSR           011111 ..... ----- ----- 0001010011 -   @X_t
  MTMSR           011111 ..... ---- . ----- 0010010010 -  @X_rs_l
  MTMSRD          011111 ..... ---- . ----- 0010110010 -  @X_rs_l
 
+### Fixed-Point Byte-Reverse Instructions
+BRW             011111 ..... ..... ----- 0010011011 -   @X_sa
+BRD             011111 ..... ..... ----- 0010111011 -   @X_sa
+BRH             011111 ..... ..... ----- 0011011011 -   @X_sa
+
  ### Fixed-Point Load Instructions
 
  LBZ             100010 ..... ..... ................     @D
target/ppc/translate.c +0 -35
index 9d29134fa1..8950476be7 100644
@@ -4532,42 +4532,7 @@ static void gen_dform3D(DisasContext *ctx)
      return gen_invalid(ctx);
  }
 
-#if defined(TARGET_PPC64)
-/* brd */
-static void gen_brd(DisasContext *ctx)
-{
-    tcg_gen_bswap64_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rS(ctx->opcode)]);
-}
-
-/* brw */
-static void gen_brw(DisasContext *ctx)
-{
-    tcg_gen_bswap64_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rS(ctx->opcode)]);
-    tcg_gen_rotli_i64(cpu_gpr[rA(ctx->opcode)], cpu_gpr[rA(ctx->opcode)], 32);
-
-}
-
-/* brh */
-static void gen_brh(DisasContext *ctx)
-{
-    TCGv_i64 mask = tcg_constant_i64(0x00ff00ff00ff00ffull);
-    TCGv_i64 t1 = tcg_temp_new_i64();
-    TCGv_i64 t2 = tcg_temp_new_i64();
-
-    tcg_gen_shri_i64(t1, cpu_gpr[rS(ctx->opcode)], 8);
-    tcg_gen_and_i64(t2, t1, mask);
-    tcg_gen_and_i64(t1, cpu_gpr[rS(ctx->opcode)], mask);
-    tcg_gen_shli_i64(t1, t1, 8);
-    tcg_gen_or_i64(cpu_gpr[rA(ctx->opcode)], t1, t2);
-}
-#endif
-
  static opcode_t opcodes[] = {
-#if defined(TARGET_PPC64)
-GEN_HANDLER_E(brd, 0x1F, 0x1B, 0x05, 0x0000F801, PPC_NONE, PPC2_ISA310),
-GEN_HANDLER_E(brw, 0x1F, 0x1B, 0x04, 0x0000F801, PPC_NONE, PPC2_ISA310),
-GEN_HANDLER_E(brh, 0x1F, 0x1B, 0x06, 0x0000F801, PPC_NONE, PPC2_ISA310),
-#endif
  GEN_HANDLER(invalid, 0x00, 0x00, 0x00, 0xFFFFFFFF, PPC_NONE),
  GEN_HANDLER_E(copy, 0x1F, 0x06, 0x18, 0x03C00001, PPC_NONE, PPC2_ISA300),
  GEN_HANDLER_E(cp_abort, 0x1F, 0x06, 0x1A, 0x03FFF801, PPC_NONE, PPC2_ISA300),
target/ppc/translate/fixedpoint-impl.c.inc +29 -0
index a75279022c..cef67d6431 100644
@@ -17,6 +17,46 @@
   * License along with this library; if not, see <http://www.gnu.org/licenses/>.
   */
 
+/*
+ * Byte-Reverse Instructions
+ */
+static bool do_byte_reverse_X_sa(DisasContext *ctx, arg_X_sa *a, int kind)
+{
+#if defined(TARGET_PPC64)
+    switch (kind) {
+    case 0: /* brd */
+        tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
+        break;
+    case 1: /* brw */
+        tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
+        tcg_gen_rotli_i64(cpu_gpr[a->ra], cpu_gpr[a->ra], 32);
+        break;
+    case 2: /* brh */
+    {
+        TCGv_i64 mask = tcg_constant_i64(0x00ff00ff00ff00ffull);
+        TCGv_i64 t1 = tcg_temp_new_i64();
+        TCGv_i64 t2 = tcg_temp_new_i64();
+
+        tcg_gen_shri_i64(t1, cpu_gpr[a->rs], 8);
+        tcg_gen_and_i64(t2, t1, mask);
+        tcg_gen_and_i64(t1, cpu_gpr[a->rs], mask);
+        tcg_gen_shli_i64(t1, t1, 8);
+        tcg_gen_or_i64(cpu_gpr[a->ra], t1, t2);
+        break;
+    }
+    default:
+        g_assert_not_reached();

target/ppc/translate/fixedpoint-impl.c.inc line 58

Quoted 10 lines
target/ppc/translate/fixedpoint-impl.c.inc +10 -0
+    }
+#else
+    qemu_build_not_reached();
+#endif
+    return true;
+}
+
+TRANS64_FLAGS2(ISA310, BRD, do_byte_reverse_X_sa, 0)
+TRANS64_FLAGS2(ISA310, BRW, do_byte_reverse_X_sa, 1)
+TRANS64_FLAGS2(ISA310, BRH, do_byte_reverse_X_sa, 2)

The kind values 0, 1, 2 are opaque at the call site — a reader has to cross-reference the switch inside do_byte_reverse_X_sa to know which is which. The more common pattern in this series is three small separate translators, which avoids the switch entirely and is self-documenting:

static bool trans_BRD(DisasContext *ctx, arg_BRD *a)

{
     REQUIRE_64BIT(ctx);
     REQUIRE_INSNS_FLAGS2(ctx, ISA310);
     tcg_gen_bswap64_i64(cpu_gpr[a->ra], cpu_gpr[a->rs]);
     return true;
}
Quoted 3 lines

...and similarly for BRW and BRH. This also eliminates the do_byte_reverse_X_sa wrapper and the #else qemu_build_not_reached() block entirely, since REQUIRE_64BIT already handles 32-bit builds.

Separating the trans_XXX handlers as suggested but preserving #if defined (TARGET_PPC64) and #else qemu_build_not_reached() blocks in them, since REQUIRE_64B is a run time check. #if defined (TARGET_PPC64) is required during compilation.

Quoted 3 lines

~Amit

#65 lore raw

PATCH v2 29/37 target/ppc: Move system call and rfi instructions to decodetree

In reply to Chinmay Rath

From: Vishal Chourasia <[email protected]>

Moving the following instructions to decodetree specification:
sc, scv : SC-form
rfi, rfid, rfscv, hrfid : XL-form

This builds upon the previous work that moved mfmsr and mtmsr[d] instructions to decodetree.

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the -d in_asm,op flag, and also by booting a pseries qemu guest.

This also includes improvements from review feedback:
- Rename helpers to uppercase (RFI/RFID/RFSCV/HRFID) to match ISA mnemonics
- Add TRANS64_FLAGS() macro variants
- Add REQUIRE_INSNS_FLAGS_NOT() check for flag exclusion
- Specify lev field as uint8_t in SC instruction format
- Remove redundant masking in SCV since lev is already 7-bit
- Replace TARGET_PPC64 ifdefs with REQUIRE_64BIT() macro
- Gate SC, SCV, RFI, RFID, RFSCV, and HRFID with appropriate flags
- Consolidate CONFIG_USER_ONLY guards into single block

---
 target/ppc/helper.h                  |   8 +-
 target/ppc/insn32.decode             |  11 +++
 target/ppc/tcg-excp_helper.c         |   8 +-
 target/ppc/translate.c               | 126 ++-------------------------
 target/ppc/translate/misc-impl.c.inc | 100 +++++++++++++++++++++
 5 files changed, 126 insertions(+), 127 deletions(-)
target/ppc/helper.h +4 -4
index 3659408c46..f24a34cfa3 100644
@@ -11,7 +11,7 @@ DEF_HELPER_4(HASHCHKP, void, env, tl, tl, tl)
 #if !defined(CONFIG_USER_ONLY)
 DEF_HELPER_2(store_msr, void, env, tl)
 DEF_HELPER_1(ppc_maybe_interrupt, void, env)
-DEF_HELPER_1(rfi, void, env)
+DEF_HELPER_1(RFI, void, env)
 DEF_HELPER_1(40x_rfci, void, env)
 DEF_HELPER_1(rfci, void, env)
 DEF_HELPER_1(rfdi, void, env)
@@ -19,9 +19,9 @@ DEF_HELPER_1(rfmci, void, env)
 #if defined(TARGET_PPC64)
 DEF_HELPER_2(scv, noreturn, env, i32)
 DEF_HELPER_2(PMINSN, void, env, i32)
-DEF_HELPER_1(rfid, void, env)
-DEF_HELPER_1(rfscv, void, env)
-DEF_HELPER_1(hrfid, void, env)
+DEF_HELPER_1(RFID, void, env)
+DEF_HELPER_1(RFSCV, void, env)
+DEF_HELPER_1(HRFID, void, env)
 DEF_HELPER_2(rfebb, void, env, tl)
 DEF_HELPER_2(store_lpcr, void, env, tl)
 DEF_HELPER_2(store_pcr, void, env, tl)
target/ppc/insn32.decode +11 -0
index 7614655f72..9d6c4b550d 100644
@@ -199,6 +199,9 @@
 &X_rs_l         rs l:bool
 @X_rs_l         ...... rs:5 .... l:1 ..... .......... .         &X_rs_l
 
+&SC             lev:uint8_t
+@SC             ...... ..... ..... .... lev:7 ... . .           &SC
+
 &X_uim5         xt uim:uint8_t
 @X_uim5         ...... ..... ..... uim:5 .......... .           &X_uim5 xt=%x_xt
 
@@ -320,6 +323,14 @@ MFMSR           011111 ..... ----- ----- 0001010011 -   @X_t
 MTMSR           011111 ..... ---- . ----- 0010010010 -  @X_rs_l
 MTMSRD          011111 ..... ---- . ----- 0010110010 -  @X_rs_l
 
+### System Call and Return from Interrupt
+SC              010001 ----- ----- ---- ....... --- 1 -         @SC
+SCV             010001 ----- ----- ---- ....... --- 0 1         @SC
+RFI             010011 ----- ----- ----- 0000110010 -
+RFID            010011 ----- ----- ----- 0000010010 -
+RFSCV           010011 ----- ----- ----- 0001010010 -
+HRFID           010011 ----- ----- ----- 0100010010 -
+
 ### Fixed-Point Byte-Reverse Instructions
 BRW             011111 ..... ..... ----- 0010011011 -   @X_sa
 BRD             011111 ..... ..... ----- 0010111011 -   @X_sa
target/ppc/tcg-excp_helper.c +4 -4
index c4ffa2dfbb..d06d3f8642 100644
@@ -528,13 +528,13 @@ static void do_rfi(CPUPPCState *env, target_ulong nip, target_ulong msr)
     check_tlb_flush(env, false);
 }
 
-void helper_rfi(CPUPPCState *env)
+void helper_RFI(CPUPPCState *env)
 {
     do_rfi(env, env->spr[SPR_SRR0], env->spr[SPR_SRR1] & 0xfffffffful);
 }
 
 #ifdef TARGET_PPC64
-void helper_rfid(CPUPPCState *env)
+void helper_RFID(CPUPPCState *env)
 {
     /*
      * The architecture defines a number of rules for which bits can
@@ -545,12 +545,12 @@ void helper_rfid(CPUPPCState *env)
     do_rfi(env, env->spr[SPR_SRR0], env->spr[SPR_SRR1]);
 }
 
-void helper_rfscv(CPUPPCState *env)
+void helper_RFSCV(CPUPPCState *env)
 {
     do_rfi(env, env->lr, env->ctr);
 }
 
-void helper_hrfid(CPUPPCState *env)
+void helper_HRFID(CPUPPCState *env)
 {
     do_rfi(env, env->spr[SPR_HSRR0], env->spr[SPR_HSRR1]);
 }
target/ppc/translate.c +7 -119

144 lines not on this page · raw message

target/ppc/translate/misc-impl.c.inc +100 -0

103 lines not on this page · raw message

Signature · 1 line
2.55.0
#66 lore raw

Re: [PATCH v2 29/37] target/ppc: Move system call and rfi instructions to decodetree

Quoted 316 lines · Chinmay Rath

From: Vishal Chourasia <[email protected]>

Moving the following instructions to decodetree specification:
sc, scv : SC-form
rfi, rfid, rfscv, hrfid : XL-form

This builds upon the previous work that moved mfmsr and mtmsr[d] instructions to decodetree.

The changes were verified by validating that the tcg ops generated by those instructions remain the same, which were captured with the -d in_asm,op flag, and also by booting a pseries qemu guest.

This also includes improvements from review feedback:
- Rename helpers to uppercase (RFI/RFID/RFSCV/HRFID) to match ISA mnemonics
- Add TRANS64_FLAGS() macro variants
- Add REQUIRE_INSNS_FLAGS_NOT() check for flag exclusion
- Specify lev field as uint8_t in SC instruction format
- Remove redundant masking in SCV since lev is already 7-bit
- Replace TARGET_PPC64 ifdefs with REQUIRE_64BIT() macro
- Gate SC, SCV, RFI, RFID, RFSCV, and HRFID with appropriate flags
- Consolidate CONFIG_USER_ONLY guards into single block

---
 target/ppc/helper.h                  |   8 +-
 target/ppc/insn32.decode             |  11 +++
 target/ppc/tcg-excp_helper.c         |   8 +-
 target/ppc/translate.c               | 126 ++-------------------------
 target/ppc/translate/misc-impl.c.inc | 100 +++++++++++++++++++++
 5 files changed, 126 insertions(+), 127 deletions(-)
target/ppc/helper.h +4 -4
index 3659408c46..f24a34cfa3 100644
@@ -11,7 +11,7 @@ DEF_HELPER_4(HASHCHKP, void, env, tl, tl, tl)
 #if !defined(CONFIG_USER_ONLY)
 DEF_HELPER_2(store_msr, void, env, tl)
 DEF_HELPER_1(ppc_maybe_interrupt, void, env)
-DEF_HELPER_1(rfi, void, env)
+DEF_HELPER_1(RFI, void, env)
 DEF_HELPER_1(40x_rfci, void, env)
 DEF_HELPER_1(rfci, void, env)
 DEF_HELPER_1(rfdi, void, env)
@@ -19,9 +19,9 @@ DEF_HELPER_1(rfmci, void, env)
 #if defined(TARGET_PPC64)
 DEF_HELPER_2(scv, noreturn, env, i32)
 DEF_HELPER_2(PMINSN, void, env, i32)
-DEF_HELPER_1(rfid, void, env)
-DEF_HELPER_1(rfscv, void, env)
-DEF_HELPER_1(hrfid, void, env)
+DEF_HELPER_1(RFID, void, env)
+DEF_HELPER_1(RFSCV, void, env)
+DEF_HELPER_1(HRFID, void, env)
 DEF_HELPER_2(rfebb, void, env, tl)
 DEF_HELPER_2(store_lpcr, void, env, tl)
 DEF_HELPER_2(store_pcr, void, env, tl)
target/ppc/insn32.decode +11 -0
index 7614655f72..9d6c4b550d 100644
@@ -199,6 +199,9 @@
 &X_rs_l         rs l:bool
 @X_rs_l         ...... rs:5 .... l:1 ..... .......... .         &X_rs_l
 
+&SC             lev:uint8_t
+@SC             ...... ..... ..... .... lev:7 ... . .           &SC
+
 &X_uim5         xt uim:uint8_t
 @X_uim5         ...... ..... ..... uim:5 .......... .           &X_uim5 xt=%x_xt
 
@@ -320,6 +323,14 @@ MFMSR           011111 ..... ----- ----- 0001010011 -   @X_t
 MTMSR           011111 ..... ---- . ----- 0010010010 -  @X_rs_l
 MTMSRD          011111 ..... ---- . ----- 0010110010 -  @X_rs_l
 
+### System Call and Return from Interrupt
+SC              010001 ----- ----- ---- ....... --- 1 -         @SC
+SCV             010001 ----- ----- ---- ....... --- 0 1         @SC
+RFI             010011 ----- ----- ----- 0000110010 -
+RFID            010011 ----- ----- ----- 0000010010 -
+RFSCV           010011 ----- ----- ----- 0001010010 -
+HRFID           010011 ----- ----- ----- 0100010010 -
+
 ### Fixed-Point Byte-Reverse Instructions
 BRW             011111 ..... ..... ----- 0010011011 -   @X_sa
 BRD             011111 ..... ..... ----- 0010111011 -   @X_sa
target/ppc/tcg-excp_helper.c +4 -4
index c4ffa2dfbb..d06d3f8642 100644
@@ -528,13 +528,13 @@ static void do_rfi(CPUPPCState *env, target_ulong nip, target_ulong msr)
     check_tlb_flush(env, false);
 }
 
-void helper_rfi(CPUPPCState *env)
+void helper_RFI(CPUPPCState *env)
 {
     do_rfi(env, env->spr[SPR_SRR0], env->spr[SPR_SRR1] & 0xfffffffful);
 }
 
 #ifdef TARGET_PPC64
-void helper_rfid(CPUPPCState *env)
+void helper_RFID(CPUPPCState *env)
 {
     /*
      * The architecture defines a number of rules for which bits can
@@ -545,12 +545,12 @@ void helper_rfid(CPUPPCState *env)
     do_rfi(env, env->spr[SPR_SRR0], env->spr[SPR_SRR1]);
 }
 
-void helper_rfscv(CPUPPCState *env)
+void helper_RFSCV(CPUPPCState *env)
 {
     do_rfi(env, env->lr, env->ctr);
 }
 
-void helper_hrfid(CPUPPCState *env)
+void helper_HRFID(CPUPPCState *env)
 {
     do_rfi(env, env->spr[SPR_HSRR0], env->spr[SPR_HSRR1]);
 }
target/ppc/translate.c +7 -119
index 8950476be7..c6c87d3e33 100644
@@ -2679,110 +2679,6 @@ static inline void gen_setlr(DisasContext *ctx, target_ulong nip)
     tcg_gen_movi_tl(cpu_lr, nip);
 }
 
-/***                           System linkage                              ***/
-
-/* rfi (supervisor only) */
-static void gen_rfi(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    /*
-     * This instruction doesn't exist anymore on 64-bit server
-     * processors compliant with arch 2.x
-     */
-    if (is_book3s_arch2x(ctx)) {
-        gen_inval_exception(ctx, POWERPC_EXCP_INVAL_INVAL);
-        return;
-    }
-    /* Restore CPU state */
-    CHK_SV(ctx);
-    translator_io_start(&ctx->base);
-    gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_NORECORD);
-    gen_helper_rfi(tcg_env);
-    ctx->base.is_jmp = DISAS_EXIT;
-#endif
-}
-
-#if defined(TARGET_PPC64)
-static void gen_rfid(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    /* Restore CPU state */
-    CHK_SV(ctx);
-    translator_io_start(&ctx->base);
-    gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_NORECORD);
-    gen_helper_rfid(tcg_env);
-    ctx->base.is_jmp = DISAS_EXIT;
-#endif
-}
-
-#if !defined(CONFIG_USER_ONLY)
-static void gen_rfscv(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    /* Restore CPU state */
-    CHK_SV(ctx);
-    translator_io_start(&ctx->base);
-    gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_NORECORD);
-    gen_helper_rfscv(tcg_env);
-    ctx->base.is_jmp = DISAS_EXIT;
-#endif
-}
-#endif
-
-static void gen_hrfid(DisasContext *ctx)
-{
-#if defined(CONFIG_USER_ONLY)
-    GEN_PRIV(ctx);
-#else
-    /* Restore CPU state */
-    CHK_HV(ctx);
-    translator_io_start(&ctx->base);
-    gen_helper_hrfid(tcg_env);
-    ctx->base.is_jmp = DISAS_EXIT;
-#endif
-}
-#endif
-
-/* sc */
-#if defined(CONFIG_USER_ONLY)
-#define POWERPC_SYSCALL POWERPC_EXCP_SYSCALL_USER
-#else
-#define POWERPC_SYSCALL POWERPC_EXCP_SYSCALL
-#endif
-static void gen_sc(DisasContext *ctx)
-{
-    uint32_t lev;
-
-    /*
-     * LEV is a 7-bit field, but the top 6 bits are treated as a reserved
-     * field (i.e., ignored). ISA v3.1 changes that to 5 bits, but that is
-     * for Ultravisor which TCG does not support, so just ignore the top 6.
-     */
-    lev = (ctx->opcode >> 5) & 0x1;
-    gen_exception_err(ctx, POWERPC_SYSCALL, lev);
-}
-
-#if defined(TARGET_PPC64)
-#if !defined(CONFIG_USER_ONLY)
-static void gen_scv(DisasContext *ctx)
-{
-    uint32_t lev = (ctx->opcode >> 5) & 0x7F;
-
-    /* Set the PC back to the faulting instruction. */
-    gen_update_nip(ctx, ctx->cia);
-    gen_helper_scv(tcg_env, tcg_constant_i32(lev));
-
-    ctx->base.is_jmp = DISAS_NORETURN;
-}
-#endif
-#endif
-
 /***                                Trap                                   ***/
 
 /* Check for unconditional traps (always or never) */
@@ -4125,6 +4021,13 @@ static int64_t dw_compose_ea(DisasContext *ctx, int x)
 #define TRANS64(NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     { REQUIRE_64BIT(ctx); return FUNC(ctx, a, ##__VA_ARGS__); }
+#define TRANS64_FLAGS(FLAGS, NAME, FUNC, ...) \
+    static bool trans_##NAME(DisasContext *ctx, arg_##NAME * a) \
+    {                                                           \
+        REQUIRE_64BIT(ctx);                                     \
+        REQUIRE_INSNS_FLAGS(ctx, FLAGS);                        \
+        return FUNC(ctx, a, ##__VA_ARGS__);                     \
+    }
 #define TRANS64_FLAGS2(FLAGS2, NAME, FUNC, ...) \
     static bool trans_##NAME(DisasContext *ctx, arg_##NAME *a) \
     {                                                          \
@@ -4542,21 +4445,6 @@ GEN_HANDLER(rlwnm, 0x17, 0xFF, 0xFF, 0x00000000, PPC_INTEGER),
 GEN_HANDLER_E(dform39, 0x39, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
 /* handles stfdp, stxsd, stxssp */
 GEN_HANDLER_E(dform3D, 0x3D, 0xFF, 0xFF, 0x00000000, PPC_NONE, PPC2_ISA205),
-/* ISA v3.0 changed the extended opcode from 62 to 30 */
-GEN_HANDLER(rfi, 0x13, 0x12, 0x01, 0x03FF8001, PPC_FLOW),
-#if defined(TARGET_PPC64)
-GEN_HANDLER(rfid, 0x13, 0x12, 0x00, 0x03FF8001, PPC_64B),
-#if !defined(CONFIG_USER_ONLY)
-/* Top bit of opc2 corresponds with low bit of LEV, so use two handlers */
-GEN_HANDLER_E(scv, 0x11, 0x10, 0xFF, 0x03FFF01E, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER_E(scv, 0x11, 0x00, 0xFF, 0x03FFF01E, PPC_NONE, PPC2_ISA300),
-GEN_HANDLER_E(rfscv, 0x13, 0x12, 0x02, 0x03FF8001, PPC_NONE, PPC2_ISA300),
-#endif
-GEN_HANDLER(hrfid, 0x13, 0x12, 0x08, 0x03FF8001, PPC_64H),
-#endif
-/* Top bit of opc2 corresponds with low bit of LEV, so use two handlers */
-GEN_HANDLER(sc, 0x11, 0x11, 0xFF, 0x03FFF01D, PPC_FLOW),
-GEN_HANDLER(sc, 0x11, 0x01, 0xFF, 0x03FFF01D, PPC_FLOW),
 GEN_HANDLER(mcrxr, 0x1F, 0x00, 0x10, 0x007FF801, PPC_MISC),
 GEN_HANDLER(mfspr, 0x1F, 0x13, 0x0A, 0x00000001, PPC_MISC),
 GEN_HANDLER(mftb, 0x1F, 0x13, 0x0B, 0x00000001, PPC_MFTB),
target/ppc/translate/misc-impl.c.inc +33 -0
index 54712f9b73..05757e8190 100644
@@ -225,3 +225,103 @@ static bool do_mtmsr(DisasContext *ctx, arg_X_rs_l *a, bool is_mtmsrd)
 
 TRANS_FLAGS(MISC, MTMSR, do_mtmsr, false)
 TRANS64(MTMSRD, do_mtmsr, true)
+
+/*
+ * System Call and Return from Interrupt Instructions
+ */
+
+static bool do_SC(DisasContext *ctx, arg_SC * a, uint32_t excp)
+{
+    uint32_t lev;
+
+    /*
+     * LEV is a 7-bit field, but the top 6 bits are treated as a reserved
+     * field (i.e., ignored). ISA v3.1 changes that to 5 bits, but that is
+     * for Ultravisor which TCG does not support, so just ignore the top 6.
+     */
+    lev = a->lev & 0x1;
+    gen_exception_err(ctx, excp, lev);
+    return true;
+}
+
+static bool do_SCV(DisasContext *ctx, arg_SC *a)
+{
+#if defined(TARGET_PPC64) && !defined(CONFIG_USER_ONLY)
+    /* Set the PC back to the faulting instruction. */
+    gen_update_nip(ctx, ctx->cia);
+    gen_helper_scv(tcg_env, tcg_constant_i32(a->lev));
+
+    ctx->base.is_jmp = DISAS_NORETURN;
+    return true;
+#else
+    gen_invalid(ctx);
+    return true;
+#endif
+}

target/ppc/translate/misc-impl.c.inc static bool do_mtmsr(DisasContext *ctx, arg_X_rs_l *a, bool is_mtmsrd) · line 270

Quoted 10 lines
target/ppc/translate/misc-impl.c.inc +10 -0
+
+/*
+ * Common helper for return-from-interrupt instructions
+ */
+enum {
+    RFI = 0,
+    RFID = 1,
+    HRFID = 2,
+    RFSCV = 3,
+};

These names seem shadow the decodetree-generated arg_RFI, arg_RFID struct names and instruction mnemonics. The line TRANS_FLAGS(FLOW, RFI, do_rfi, RFI) is particularly confusing — the trailing RFI is the enum value 0, but it looks like the instruction name. Prefer a distinct prefix, e.g. RFI_KIND_RFI = 0, RFI_KIND_RFID = 1, etc.

Quoted 29 lines
target/ppc/translate/misc-impl.c.inc +29 -0
+
+static bool do_rfi(DisasContext *ctx, arg_RFID *a, int kind)
+{
+#if defined(CONFIG_USER_ONLY)
+    gen_priv_opc(ctx);
+    return true;
+#else
+    void (*helper)(TCGv_ptr);
+
+    switch (kind) {
+    case RFI:
+        if (is_book3s_arch2x(ctx)) {
+            gen_invalid(ctx);
+            return true;
+        }
+        REQUIRE_SV(ctx);
+        helper = gen_helper_RFI;
+        break;
+#if defined(TARGET_PPC64)
+    case HRFID:
+        REQUIRE_HV(ctx);
+        helper = gen_helper_HRFID;
+        break;
+    case RFID:
+        REQUIRE_SV(ctx);
+        helper = gen_helper_RFID;
+        break;
+    case RFSCV:
+        REQUIRE_SV(ctx);

target/ppc/translate/misc-impl.c.inc static bool do_mtmsr(DisasContext *ctx, arg_X_rs_l *a, bool is_mtmsrd) · line 309

Quoted 10 lines
target/ppc/translate/misc-impl.c.inc +10 -0
+        helper = gen_helper_RFSCV;
+        break;
+#endif
+    default:
+        gen_invalid(ctx);
+        return true;
+    }
+
+    translator_io_start(&ctx->base);
+    gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_NORECORD);

This gen_update_branch_history call sits after the switch and executes unconditionally for all four kind values. The old gen_hrfid did not call it.

Inside gen_update_branch_history, if ctx->has_cfar is true, cpu_cfar is written unconditionally before the BHRB-recording check. So hrfid now updates CFAR on every execution, where it did not before. The fix is to guard the call:

  if (kind != HRFID) {
      gen_update_branch_history(ctx, ctx->cia, NULL, BHRB_TYPE_NORECORD);
  }

Having said that, I think the root cause is the consolidation itself. With four separate translators (the pattern used throughout this series), each would contain exactly what its old gen_ function had, and the HRFID omission would have been naturally preserved.

~Amit

Quoted 21 lines
target/ppc/translate/misc-impl.c.inc +18 -0
+    helper(tcg_env);
+    ctx->base.is_jmp = DISAS_EXIT;
+    return true;
+#endif
+}
+
+#if defined(CONFIG_USER_ONLY)
+TRANS_FLAGS(FLOW, SC, do_SC, POWERPC_EXCP_SYSCALL_USER)
+#else
+TRANS_FLAGS(FLOW, SC, do_SC, POWERPC_EXCP_SYSCALL)
+#endif
+
+TRANS64_FLAGS2(ISA300, SCV, do_SCV)
+
+TRANS_FLAGS(FLOW, RFI, do_rfi, RFI)
+TRANS64(RFID, do_rfi, RFID)
+TRANS64_FLAGS(64H, HRFID, do_rfi, HRFID)
+TRANS64_FLAGS2(ISA300, RFSCV, do_rfi, RFSCV)
Signature · 1 line
2.55.0
#67 lore raw

PATCH v2 30/37 target/ppc: Replace PPC2_VSX207 flag with PPC2_ISA207

In reply to Chinmay Rath

There are several flags for Power ISA 2.07 currently in use, namely :
PPC2_VSX207, PPC2_BCTAR_ISA207, PPC2_LSQ_ISA207, etc, but most of them are used together without any individual use.

Create a generic ISA 2.07 flag and switch usage of PPC2_VSX207 to it instead. Next commits would consolidate the remaining flags.

---
 target/ppc/cpu.h                    |  6 ++--
 target/ppc/cpu_init.c               |  2 +-
 target/ppc/cpu_init.h               |  2 +-
 target/ppc/translate/fp-impl.c.inc  |  4 +--
 target/ppc/translate/vsx-impl.c.inc | 56 ++++++++++++++---------------
 target/ppc/translate/vsx-ops.c.inc  | 24 ++++++-------
 6 files changed, 47 insertions(+), 47 deletions(-)
target/ppc/cpu.h +3 -3
index cbd5964b1a..91ef34cb21 100644
@@ -2565,8 +2565,8 @@ enum {
     PPC2_DBRX          = 0x0000000000000010ULL,
     /* Book I 2.05 PowerPC specification                                     */
     PPC2_ISA205        = 0x0000000000000020ULL,
-    /* VSX additions in ISA 2.07                                             */
-    PPC2_VSX207        = 0x0000000000000040ULL,
+    /* POWER ISA 2.07                                                        */
+    PPC2_ISA207        = 0x0000000000000040ULL,
     /* ISA 2.06B bpermd                                                      */
     PPC2_PERM_ISA206   = 0x0000000000000080ULL,
     /* ISA 2.06B divide extended variants                                    */
@@ -2607,7 +2607,7 @@ enum {
     PPC2_PPE42XM       = 0x0000000002000000ULL,
 
 #define PPC_TCG_INSNS2 (PPC2_BOOKE206 | PPC2_VSX | PPC2_PRCNTL | PPC2_DBRX | \
-                        PPC2_ISA205 | PPC2_VSX207 | PPC2_PERM_ISA206 | \
+                        PPC2_ISA205 | PPC2_ISA207 | PPC2_PERM_ISA206 | \
                         PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 | \
                         PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | \
                         PPC2_BCTAR_ISA207 | PPC2_LSQ_ISA207 | \
target/ppc/cpu_init.c +1 -1
index e2533481d8..b913edd243 100644
@@ -6346,7 +6346,7 @@ POWERPC_FAMILY(POWER8)(ObjectClass *oc, const void *data)
                        PPC_SEGMENT_64B | PPC_SLBI |
                        PPC_POPCNTB | PPC_POPCNTWD |
                        PPC_CILDST;
-    pcc->insns_flags2 = PPC2_VSX | PPC2_VSX207 | PPC2_DFP | PPC2_DBRX |
+    pcc->insns_flags2 = PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |
                         PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 |
                         PPC2_ATOMIC_ISA206 | PPC2_FP_CVT_ISA206 |
                         PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |
target/ppc/cpu_init.h +1 -1
index f8fd6ff5cd..1c41c0d349 100644
@@ -15,7 +15,7 @@
 #define PPC_INSNS_FLAGS_POWER11 PPC_INSNS_FLAGS_POWER10
 
 #define PPC_INSNS_FLAGS2_POWER_COMMON                                \
-    (PPC2_VSX | PPC2_VSX207 | PPC2_DFP | PPC2_DBRX |                 \
+    (PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |                 \
      PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 |      \
      PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |   \
      PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 | PPC2_ISA205 |              \
target/ppc/translate/fp-impl.c.inc +2 -2
index 464fb1d90f..e0d9fa2e24 100644
@@ -342,7 +342,7 @@ static bool trans_FCPSGN(DisasContext *ctx, arg_FCPSGN *a)
 static bool trans_FMRGEW(DisasContext *ctx, arg_FMRGEW *a)
 {
     TCGv_i64 t0, t1, t2;
-    REQUIRE_INSNS_FLAGS2(ctx, VSX207);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207);
     REQUIRE_FPU(ctx);
     t0 = tcg_temp_new_i64();
     t1 = tcg_temp_new_i64();
@@ -358,7 +358,7 @@ static bool trans_FMRGEW(DisasContext *ctx, arg_FMRGEW *a)
 static bool trans_FMRGOW(DisasContext *ctx, arg_FMRGOW *a)
 {
     TCGv_i64 t0, t1, t2;
-    REQUIRE_INSNS_FLAGS2(ctx, VSX207);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207);
     REQUIRE_FPU(ctx);
     t0 = tcg_temp_new_i64();
     t1 = tcg_temp_new_i64();
target/ppc/translate/vsx-impl.c.inc +28 -28

110 lines not on this page · raw message

target/ppc/translate/vsx-ops.c.inc +12 -12
index e553b5b8fa..ae3df2c13d 100644
@@ -1,9 +1,9 @@
-GEN_HANDLER_E(mfvsrwz, 0x1F, 0x13, 0x03, 0x0000F800, PPC_NONE, PPC2_VSX207),
-GEN_HANDLER_E(mtvsrwa, 0x1F, 0x13, 0x06, 0x0000F800, PPC_NONE, PPC2_VSX207),
-GEN_HANDLER_E(mtvsrwz, 0x1F, 0x13, 0x07, 0x0000F800, PPC_NONE, PPC2_VSX207),
+GEN_HANDLER_E(mfvsrwz, 0x1F, 0x13, 0x03, 0x0000F800, PPC_NONE, PPC2_ISA207),
+GEN_HANDLER_E(mtvsrwa, 0x1F, 0x13, 0x06, 0x0000F800, PPC_NONE, PPC2_ISA207),
+GEN_HANDLER_E(mtvsrwz, 0x1F, 0x13, 0x07, 0x0000F800, PPC_NONE, PPC2_ISA207),
 #if defined(TARGET_PPC64)
-GEN_HANDLER_E(mfvsrd, 0x1F, 0x13, 0x01, 0x0000F800, PPC_NONE, PPC2_VSX207),
-GEN_HANDLER_E(mtvsrd, 0x1F, 0x13, 0x05, 0x0000F800, PPC_NONE, PPC2_VSX207),
+GEN_HANDLER_E(mfvsrd, 0x1F, 0x13, 0x01, 0x0000F800, PPC_NONE, PPC2_ISA207),
+GEN_HANDLER_E(mtvsrd, 0x1F, 0x13, 0x05, 0x0000F800, PPC_NONE, PPC2_ISA207),
 GEN_HANDLER_E(mfvsrld, 0X1F, 0x13, 0x09, 0x0000F800, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(mtvsrdd, 0X1F, 0x13, 0x0D, 0x0, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(mtvsrws, 0x1F, 0x13, 0x0C, 0x0000F800, PPC_NONE, PPC2_ISA300),
@@ -127,7 +127,7 @@ GEN_VSX_XFORM_300(xscmpoqp, 0x04, 0x04, 0x00600001),
 GEN_VSX_XFORM_300(xscmpuqp, 0x04, 0x14, 0x00600001),
 GEN_XX2FORM_EO(xscvdphp, 0x16, 0x15, 0x11, PPC2_ISA300),
 GEN_XX2FORM(xscvdpsp, 0x12, 0x10, PPC2_VSX),
-GEN_XX2FORM(xscvdpspn, 0x16, 0x10, PPC2_VSX207),
+GEN_XX2FORM(xscvdpspn, 0x16, 0x10, PPC2_ISA207),
 GEN_XX2FORM_EO(xscvhpdp, 0x16, 0x15, 0x10, PPC2_ISA300),
 GEN_VSX_XFORM_300_EO(xscvsdqp, 0x04, 0x1A, 0x0A, 0x00000001),
 GEN_XX2FORM(xscvspdp, 0x12, 0x14, PPC2_VSX),
@@ -145,12 +145,12 @@ GEN_XX2FORM(xsrdpip, 0x12, 0x06, PPC2_VSX),
 GEN_XX2FORM(xsrdpiz, 0x12, 0x05, PPC2_VSX),
 
 GEN_VSX_XFORM_300(xsdivqp, 0x04, 0x11, 0x0),
-GEN_XX2FORM(xsresp,  0x14, 0x01, PPC2_VSX207),
-GEN_XX2FORM(xsrsp, 0x12, 0x11, PPC2_VSX207),
-GEN_XX2FORM(xssqrtsp,  0x16, 0x00, PPC2_VSX207),
-GEN_XX2FORM(xsrsqrtesp,  0x14, 0x00, PPC2_VSX207),
-GEN_XX2FORM(xscvsxdsp, 0x10, 0x13, PPC2_VSX207),
-GEN_XX2FORM(xscvuxdsp, 0x10, 0x12, PPC2_VSX207),
+GEN_XX2FORM(xsresp,  0x14, 0x01, PPC2_ISA207),
+GEN_XX2FORM(xsrsp, 0x12, 0x11, PPC2_ISA207),
+GEN_XX2FORM(xssqrtsp,  0x16, 0x00, PPC2_ISA207),
+GEN_XX2FORM(xsrsqrtesp,  0x14, 0x00, PPC2_ISA207),
+GEN_XX2FORM(xscvsxdsp, 0x10, 0x13, PPC2_ISA207),
+GEN_XX2FORM(xscvuxdsp, 0x10, 0x12, PPC2_ISA207),
 
 GEN_XX2FORM(xvredp,  0x14, 0x0D, PPC2_VSX),
 GEN_XX2FORM(xvsqrtdp,  0x16, 0x0C, PPC2_VSX),
Signature · 1 line
2.55.0
Reviewed-by on v2 30/37
#68 lore raw

Re: [PATCH v2 30/37] target/ppc: Replace PPC2_VSX207 flag with PPC2_ISA207

Quoted 10 lines · Chinmay Rath

There are several flags for Power ISA 2.07 currently in use, namely :
PPC2_VSX207, PPC2_BCTAR_ISA207, PPC2_LSQ_ISA207, etc, but most of them are used together without any individual use.

Create a generic ISA 2.07 flag and switch usage of PPC2_VSX207 to it instead. Next commits would consolidate the remaining flags.

Quoted 8 lines
---
 target/ppc/cpu.h                    |  6 ++--
 target/ppc/cpu_init.c               |  2 +-
 target/ppc/cpu_init.h               |  2 +-
 target/ppc/translate/fp-impl.c.inc  |  4 +--
 target/ppc/translate/vsx-impl.c.inc | 56 ++++++++++++++---------------
 target/ppc/translate/vsx-ops.c.inc  | 24 ++++++-------
 6 files changed, 47 insertions(+), 47 deletions(-)
Quoted 245 lines
target/ppc/cpu.h +3 -3
index cbd5964b1a..91ef34cb21 100644
@@ -2565,8 +2565,8 @@ enum {
     PPC2_DBRX          = 0x0000000000000010ULL,
     /* Book I 2.05 PowerPC specification                                     */
     PPC2_ISA205        = 0x0000000000000020ULL,
-    /* VSX additions in ISA 2.07                                             */
-    PPC2_VSX207        = 0x0000000000000040ULL,
+    /* POWER ISA 2.07                                                        */
+    PPC2_ISA207        = 0x0000000000000040ULL,
     /* ISA 2.06B bpermd                                                      */
     PPC2_PERM_ISA206   = 0x0000000000000080ULL,
     /* ISA 2.06B divide extended variants                                    */
@@ -2607,7 +2607,7 @@ enum {
     PPC2_PPE42XM       = 0x0000000002000000ULL,
 
 #define PPC_TCG_INSNS2 (PPC2_BOOKE206 | PPC2_VSX | PPC2_PRCNTL | PPC2_DBRX | \
-                        PPC2_ISA205 | PPC2_VSX207 | PPC2_PERM_ISA206 | \
+                        PPC2_ISA205 | PPC2_ISA207 | PPC2_PERM_ISA206 | \
                         PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 | \
                         PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | \
                         PPC2_BCTAR_ISA207 | PPC2_LSQ_ISA207 | \
target/ppc/cpu_init.c +1 -1
index e2533481d8..b913edd243 100644
@@ -6346,7 +6346,7 @@ POWERPC_FAMILY(POWER8)(ObjectClass *oc, const void *data)
                        PPC_SEGMENT_64B | PPC_SLBI |
                        PPC_POPCNTB | PPC_POPCNTWD |
                        PPC_CILDST;
-    pcc->insns_flags2 = PPC2_VSX | PPC2_VSX207 | PPC2_DFP | PPC2_DBRX |
+    pcc->insns_flags2 = PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |
                         PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 |
                         PPC2_ATOMIC_ISA206 | PPC2_FP_CVT_ISA206 |
                         PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |
target/ppc/cpu_init.h +1 -1
index f8fd6ff5cd..1c41c0d349 100644
@@ -15,7 +15,7 @@
 #define PPC_INSNS_FLAGS_POWER11 PPC_INSNS_FLAGS_POWER10
 
 #define PPC_INSNS_FLAGS2_POWER_COMMON                                \
-    (PPC2_VSX | PPC2_VSX207 | PPC2_DFP | PPC2_DBRX |                 \
+    (PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |                 \
      PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 |      \
      PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |   \
      PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 | PPC2_ISA205 |              \
target/ppc/translate/fp-impl.c.inc +2 -2
index 464fb1d90f..e0d9fa2e24 100644
@@ -342,7 +342,7 @@ static bool trans_FCPSGN(DisasContext *ctx, arg_FCPSGN *a)
 static bool trans_FMRGEW(DisasContext *ctx, arg_FMRGEW *a)
 {
     TCGv_i64 t0, t1, t2;
-    REQUIRE_INSNS_FLAGS2(ctx, VSX207);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207);
     REQUIRE_FPU(ctx);
     t0 = tcg_temp_new_i64();
     t1 = tcg_temp_new_i64();
@@ -358,7 +358,7 @@ static bool trans_FMRGEW(DisasContext *ctx, arg_FMRGEW *a)
 static bool trans_FMRGOW(DisasContext *ctx, arg_FMRGOW *a)
 {
     TCGv_i64 t0, t1, t2;
-    REQUIRE_INSNS_FLAGS2(ctx, VSX207);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207);
     REQUIRE_FPU(ctx);
     t0 = tcg_temp_new_i64();
     t1 = tcg_temp_new_i64();
target/ppc/translate/vsx-impl.c.inc +28 -28
index 00ad57c628..78311dcba9 100644
@@ -50,11 +50,11 @@ static bool do_lxs(DisasContext *ctx, arg_X *a,
 }
 
 TRANS_FLAGS2(VSX, LXSDX, do_lxs, gen_qemu_ld64_i64);
-TRANS_FLAGS2(VSX207, LXSIWAX, do_lxs, gen_qemu_ld32s_i64);
+TRANS_FLAGS2(ISA207, LXSIWAX, do_lxs, gen_qemu_ld32s_i64);
 TRANS_FLAGS2(ISA300, LXSIBZX, do_lxs, gen_qemu_ld8u_i64);
 TRANS_FLAGS2(ISA300, LXSIHZX, do_lxs, gen_qemu_ld16u_i64);
-TRANS_FLAGS2(VSX207, LXSIWZX, do_lxs, gen_qemu_ld32u_i64);
-TRANS_FLAGS2(VSX207, LXSSPX, do_lxs, gen_qemu_ld32fs);
+TRANS_FLAGS2(ISA207, LXSIWZX, do_lxs, gen_qemu_ld32u_i64);
+TRANS_FLAGS2(ISA207, LXSSPX, do_lxs, gen_qemu_ld32fs);
 
 static bool trans_LXVD2X(DisasContext *ctx, arg_LXVD2X *a)
 {
@@ -304,8 +304,8 @@ static bool do_stxs(DisasContext *ctx, arg_X *a,
 TRANS_FLAGS2(VSX, STXSDX, do_stxs, gen_qemu_st64_i64);
 TRANS_FLAGS2(ISA300, STXSIBX, do_stxs, gen_qemu_st8_i64);
 TRANS_FLAGS2(ISA300, STXSIHX, do_stxs, gen_qemu_st16_i64);
-TRANS_FLAGS2(VSX207, STXSIWX, do_stxs, gen_qemu_st32_i64);
-TRANS_FLAGS2(VSX207, STXSSPX, do_stxs, gen_qemu_st32fs);
+TRANS_FLAGS2(ISA207, STXSIWX, do_stxs, gen_qemu_st32_i64);
+TRANS_FLAGS2(ISA207, STXSSPX, do_stxs, gen_qemu_st32fs);
 
 static bool trans_STXVD2X(DisasContext *ctx, arg_STXVD2X *a)
 {
@@ -988,7 +988,7 @@ GEN_VSX_HELPER_R2_AB(xscmpuqp, 0x04, 0x14, 0, PPC2_VSX)
 GEN_VSX_HELPER_X2(xscvdphp, 0x16, 0x15, 0x11, PPC2_ISA300)
 GEN_VSX_HELPER_X2(xscvdpsp, 0x12, 0x10, 0, PPC2_VSX)
 GEN_VSX_HELPER_R2(xscvdpqp, 0x04, 0x1A, 0x16, PPC2_ISA300)
-GEN_VSX_HELPER_XT_XB_ENV(xscvdpspn, 0x16, 0x10, 0, PPC2_VSX207)
+GEN_VSX_HELPER_XT_XB_ENV(xscvdpspn, 0x16, 0x10, 0, PPC2_ISA207)
 GEN_VSX_HELPER_R2(xscvqpsdz, 0x04, 0x1A, 0x19, PPC2_ISA300)
 GEN_VSX_HELPER_R2(xscvqpswz, 0x04, 0x1A, 0x09, PPC2_ISA300)
 GEN_VSX_HELPER_R2(xscvqpudz, 0x04, 0x1A, 0x11, PPC2_ISA300)
@@ -1185,7 +1185,7 @@ bool trans_XSCVSPDPN(DisasContext *ctx, arg_XX2 *a)
 {
     TCGv_i64 tmp;
 
-    REQUIRE_INSNS_FLAGS2(ctx, VSX207);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207);
     REQUIRE_VSX(ctx);
 
     tmp = tcg_temp_new_i64();
@@ -1210,16 +1210,16 @@ GEN_VSX_HELPER_X2(xsrdpic, 0x16, 0x06, 0, PPC2_VSX)
 GEN_VSX_HELPER_X2(xsrdpim, 0x12, 0x07, 0, PPC2_VSX)
 GEN_VSX_HELPER_X2(xsrdpip, 0x12, 0x06, 0, PPC2_VSX)
 GEN_VSX_HELPER_X2(xsrdpiz, 0x12, 0x05, 0, PPC2_VSX)
-GEN_VSX_HELPER_XT_XB_ENV(xsrsp, 0x12, 0x11, 0, PPC2_VSX207)
+GEN_VSX_HELPER_XT_XB_ENV(xsrsp, 0x12, 0x11, 0, PPC2_ISA207)
 GEN_VSX_HELPER_R2(xsrqpi, 0x05, 0x00, 0, PPC2_ISA300)
 GEN_VSX_HELPER_R2(xsrqpxp, 0x05, 0x01, 0, PPC2_ISA300)
 GEN_VSX_HELPER_R2(xssqrtqp, 0x04, 0x19, 0x1B, PPC2_ISA300)
 GEN_VSX_HELPER_R3(xssubqp, 0x04, 0x10, 0, PPC2_ISA300)
-GEN_VSX_HELPER_X2(xsresp, 0x14, 0x01, 0, PPC2_VSX207)
-GEN_VSX_HELPER_X2(xssqrtsp, 0x16, 0x00, 0, PPC2_VSX207)
-GEN_VSX_HELPER_X2(xsrsqrtesp, 0x14, 0x00, 0, PPC2_VSX207)
-GEN_VSX_HELPER_X2(xscvsxdsp, 0x10, 0x13, 0, PPC2_VSX207)
-GEN_VSX_HELPER_X2(xscvuxdsp, 0x10, 0x12, 0, PPC2_VSX207)
+GEN_VSX_HELPER_X2(xsresp, 0x14, 0x01, 0, PPC2_ISA207)
+GEN_VSX_HELPER_X2(xssqrtsp, 0x16, 0x00, 0, PPC2_ISA207)
+GEN_VSX_HELPER_X2(xsrsqrtesp, 0x14, 0x00, 0, PPC2_ISA207)
+GEN_VSX_HELPER_X2(xscvsxdsp, 0x10, 0x13, 0, PPC2_ISA207)
+GEN_VSX_HELPER_X2(xscvuxdsp, 0x10, 0x12, 0, PPC2_ISA207)
 
 GEN_VSX_HELPER_X2(xvredp, 0x14, 0x0D, 0, PPC2_VSX)
 GEN_VSX_HELPER_X2(xvsqrtdp, 0x16, 0x0C, 0, PPC2_VSX)
@@ -1409,14 +1409,14 @@ TRANS_FLAGS2(VSX, XSNMADDADP, do_xsmadd_XX3, true, gen_helper_XSNMADDDP)
 TRANS_FLAGS2(VSX, XSNMADDMDP, do_xsmadd_XX3, false, gen_helper_XSNMADDDP)
 TRANS_FLAGS2(VSX, XSNMSUBADP, do_xsmadd_XX3, true, gen_helper_XSNMSUBDP)
 TRANS_FLAGS2(VSX, XSNMSUBMDP, do_xsmadd_XX3, false, gen_helper_XSNMSUBDP)
-TRANS_FLAGS2(VSX207, XSMADDASP, do_xsmadd_XX3, true, gen_helper_XSMADDSP)
-TRANS_FLAGS2(VSX207, XSMADDMSP, do_xsmadd_XX3, false, gen_helper_XSMADDSP)
-TRANS_FLAGS2(VSX207, XSMSUBASP, do_xsmadd_XX3, true, gen_helper_XSMSUBSP)
-TRANS_FLAGS2(VSX207, XSMSUBMSP, do_xsmadd_XX3, false, gen_helper_XSMSUBSP)
-TRANS_FLAGS2(VSX207, XSNMADDASP, do_xsmadd_XX3, true, gen_helper_XSNMADDSP)
-TRANS_FLAGS2(VSX207, XSNMADDMSP, do_xsmadd_XX3, false, gen_helper_XSNMADDSP)
-TRANS_FLAGS2(VSX207, XSNMSUBASP, do_xsmadd_XX3, true, gen_helper_XSNMSUBSP)
-TRANS_FLAGS2(VSX207, XSNMSUBMSP, do_xsmadd_XX3, false, gen_helper_XSNMSUBSP)
+TRANS_FLAGS2(ISA207, XSMADDASP, do_xsmadd_XX3, true, gen_helper_XSMADDSP)
+TRANS_FLAGS2(ISA207, XSMADDMSP, do_xsmadd_XX3, false, gen_helper_XSMADDSP)
+TRANS_FLAGS2(ISA207, XSMSUBASP, do_xsmadd_XX3, true, gen_helper_XSMSUBSP)
+TRANS_FLAGS2(ISA207, XSMSUBMSP, do_xsmadd_XX3, false, gen_helper_XSMSUBSP)
+TRANS_FLAGS2(ISA207, XSNMADDASP, do_xsmadd_XX3, true, gen_helper_XSNMADDSP)
+TRANS_FLAGS2(ISA207, XSNMADDMSP, do_xsmadd_XX3, false, gen_helper_XSNMADDSP)
+TRANS_FLAGS2(ISA207, XSNMSUBASP, do_xsmadd_XX3, true, gen_helper_XSNMSUBSP)
+TRANS_FLAGS2(ISA207, XSNMSUBMSP, do_xsmadd_XX3, false, gen_helper_XSNMSUBSP)
 
 static bool do_xsmadd_X(DisasContext *ctx, arg_X_rc *a,
         void (*gen_helper)(TCGv_ptr, TCGv_ptr, TCGv_ptr, TCGv_ptr, TCGv_ptr),
@@ -1590,9 +1590,9 @@ TRANS_FLAGS2(VSX, XXLANDC, do_logical_op, MO_64, tcg_gen_gvec_andc);
 TRANS_FLAGS2(VSX, XXLOR, do_logical_op, MO_64, tcg_gen_gvec_or);
 TRANS_FLAGS2(VSX, XXLXOR, do_logical_op, MO_64, tcg_gen_gvec_xor);
 TRANS_FLAGS2(VSX, XXLNOR, do_logical_op, MO_64, tcg_gen_gvec_nor);
-TRANS_FLAGS2(VSX207, XXLEQV, do_logical_op, MO_64, tcg_gen_gvec_eqv);
-TRANS_FLAGS2(VSX207, XXLNAND, do_logical_op, MO_64, tcg_gen_gvec_nand);
-TRANS_FLAGS2(VSX207, XXLORC, do_logical_op, MO_64, tcg_gen_gvec_orc);
+TRANS_FLAGS2(ISA207, XXLEQV, do_logical_op, MO_64, tcg_gen_gvec_eqv);
+TRANS_FLAGS2(ISA207, XXLNAND, do_logical_op, MO_64, tcg_gen_gvec_nand);
+TRANS_FLAGS2(ISA207, XXLORC, do_logical_op, MO_64, tcg_gen_gvec_orc);
 
 #define VSX_XXMRG(name, high)                               \
 static void glue(gen_, name)(DisasContext *ctx)             \
@@ -2682,10 +2682,10 @@ TRANS_FLAGS2(ISA300, XSMINCDP, do_helper_XX3, gen_helper_XSMINCDP)
 TRANS_FLAGS2(ISA300, XSMAXJDP, do_helper_XX3, gen_helper_XSMAXJDP)
 TRANS_FLAGS2(ISA300, XSMINJDP, do_helper_XX3, gen_helper_XSMINJDP)
 
-TRANS_FLAGS2(VSX207, XSADDSP, do_helper_XX3, gen_helper_XSADDSP)
-TRANS_FLAGS2(VSX207, XSSUBSP, do_helper_XX3, gen_helper_XSSUBSP)
-TRANS_FLAGS2(VSX207, XSMULSP, do_helper_XX3, gen_helper_XSMULSP)
-TRANS_FLAGS2(VSX207, XSDIVSP, do_helper_XX3, gen_helper_XSDIVSP)
+TRANS_FLAGS2(ISA207, XSADDSP, do_helper_XX3, gen_helper_XSADDSP)
+TRANS_FLAGS2(ISA207, XSSUBSP, do_helper_XX3, gen_helper_XSSUBSP)
+TRANS_FLAGS2(ISA207, XSMULSP, do_helper_XX3, gen_helper_XSMULSP)
+TRANS_FLAGS2(ISA207, XSDIVSP, do_helper_XX3, gen_helper_XSDIVSP)
 
 TRANS_FLAGS2(VSX, XSADDDP, do_helper_XX3, gen_helper_XSADDDP)
 TRANS_FLAGS2(VSX, XSSUBDP, do_helper_XX3, gen_helper_XSSUBDP)
target/ppc/translate/vsx-ops.c.inc +12 -12
index e553b5b8fa..ae3df2c13d 100644
@@ -1,9 +1,9 @@
-GEN_HANDLER_E(mfvsrwz, 0x1F, 0x13, 0x03, 0x0000F800, PPC_NONE, PPC2_VSX207),
-GEN_HANDLER_E(mtvsrwa, 0x1F, 0x13, 0x06, 0x0000F800, PPC_NONE, PPC2_VSX207),
-GEN_HANDLER_E(mtvsrwz, 0x1F, 0x13, 0x07, 0x0000F800, PPC_NONE, PPC2_VSX207),
+GEN_HANDLER_E(mfvsrwz, 0x1F, 0x13, 0x03, 0x0000F800, PPC_NONE, PPC2_ISA207),
+GEN_HANDLER_E(mtvsrwa, 0x1F, 0x13, 0x06, 0x0000F800, PPC_NONE, PPC2_ISA207),
+GEN_HANDLER_E(mtvsrwz, 0x1F, 0x13, 0x07, 0x0000F800, PPC_NONE, PPC2_ISA207),
 #if defined(TARGET_PPC64)
-GEN_HANDLER_E(mfvsrd, 0x1F, 0x13, 0x01, 0x0000F800, PPC_NONE, PPC2_VSX207),
-GEN_HANDLER_E(mtvsrd, 0x1F, 0x13, 0x05, 0x0000F800, PPC_NONE, PPC2_VSX207),
+GEN_HANDLER_E(mfvsrd, 0x1F, 0x13, 0x01, 0x0000F800, PPC_NONE, PPC2_ISA207),
+GEN_HANDLER_E(mtvsrd, 0x1F, 0x13, 0x05, 0x0000F800, PPC_NONE, PPC2_ISA207),
 GEN_HANDLER_E(mfvsrld, 0X1F, 0x13, 0x09, 0x0000F800, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(mtvsrdd, 0X1F, 0x13, 0x0D, 0x0, PPC_NONE, PPC2_ISA300),
 GEN_HANDLER_E(mtvsrws, 0x1F, 0x13, 0x0C, 0x0000F800, PPC_NONE, PPC2_ISA300),
@@ -127,7 +127,7 @@ GEN_VSX_XFORM_300(xscmpoqp, 0x04, 0x04, 0x00600001),
 GEN_VSX_XFORM_300(xscmpuqp, 0x04, 0x14, 0x00600001),
 GEN_XX2FORM_EO(xscvdphp, 0x16, 0x15, 0x11, PPC2_ISA300),
 GEN_XX2FORM(xscvdpsp, 0x12, 0x10, PPC2_VSX),
-GEN_XX2FORM(xscvdpspn, 0x16, 0x10, PPC2_VSX207),
+GEN_XX2FORM(xscvdpspn, 0x16, 0x10, PPC2_ISA207),
 GEN_XX2FORM_EO(xscvhpdp, 0x16, 0x15, 0x10, PPC2_ISA300),
 GEN_VSX_XFORM_300_EO(xscvsdqp, 0x04, 0x1A, 0x0A, 0x00000001),
 GEN_XX2FORM(xscvspdp, 0x12, 0x14, PPC2_VSX),
@@ -145,12 +145,12 @@ GEN_XX2FORM(xsrdpip, 0x12, 0x06, PPC2_VSX),
 GEN_XX2FORM(xsrdpiz, 0x12, 0x05, PPC2_VSX),
 
 GEN_VSX_XFORM_300(xsdivqp, 0x04, 0x11, 0x0),
-GEN_XX2FORM(xsresp,  0x14, 0x01, PPC2_VSX207),
-GEN_XX2FORM(xsrsp, 0x12, 0x11, PPC2_VSX207),
-GEN_XX2FORM(xssqrtsp,  0x16, 0x00, PPC2_VSX207),
-GEN_XX2FORM(xsrsqrtesp,  0x14, 0x00, PPC2_VSX207),
-GEN_XX2FORM(xscvsxdsp, 0x10, 0x13, PPC2_VSX207),
-GEN_XX2FORM(xscvuxdsp, 0x10, 0x12, PPC2_VSX207),
+GEN_XX2FORM(xsresp,  0x14, 0x01, PPC2_ISA207),
+GEN_XX2FORM(xsrsp, 0x12, 0x11, PPC2_ISA207),
+GEN_XX2FORM(xssqrtsp,  0x16, 0x00, PPC2_ISA207),
+GEN_XX2FORM(xsrsqrtesp,  0x14, 0x00, PPC2_ISA207),
+GEN_XX2FORM(xscvsxdsp, 0x10, 0x13, PPC2_ISA207),
+GEN_XX2FORM(xscvuxdsp, 0x10, 0x12, PPC2_ISA207),
 
 GEN_XX2FORM(xvredp,  0x14, 0x0D, PPC2_VSX),
 GEN_XX2FORM(xvsqrtdp,  0x16, 0x0C, PPC2_VSX),
Signature · 1 line
2.55.0
#69 lore raw

PATCH v2 31/37 target/ppc: Use PPC2_ISA207 instead of PPC2_BCTAR_ISA207

In reply to Chinmay Rath

PPC2_BCTAR_ISA207 is only ever set in the CPUPPCState's insns_flags2 alongside PPC2_ISA207. Checks made by PPC2_BCTAR_ISA207 could be replaced with PPC2_ISA207, hence rendering it useless and apt for removal. This patch does the same.

---
 linux-user/ppc/elfload.c               | 7 +++----
 target/ppc/cpu.h                       | 4 +---
 target/ppc/cpu_init.c                  | 2 +-
 target/ppc/cpu_init.h                  | 2 +-
 target/ppc/translate/branch-impl.c.inc | 2 +-
 5 files changed, 7 insertions(+), 10 deletions(-)
linux-user/ppc/elfload.c +3 -4
index 0d54da9803..5c6065e08a 100644
@@ -116,10 +116,9 @@ abi_ulong get_elf_hwcap2(CPUState *cs)
     do { if (cpu->env.insns_flags2 & flag) { features |= feature; } } while (0)
 
     GET_FEATURE(PPC_ISEL, QEMU_PPC_FEATURE2_HAS_ISEL);
-    GET_FEATURE2(PPC2_BCTAR_ISA207, QEMU_PPC_FEATURE2_HAS_TAR);
-    GET_FEATURE2((PPC2_BCTAR_ISA207 | PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
-                  PPC2_ISA207S), QEMU_PPC_FEATURE2_ARCH_2_07 |
-                  QEMU_PPC_FEATURE2_VEC_CRYPTO);
+    GET_FEATURE2((PPC2_ISA207 | PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
+                  PPC2_ISA207S), (QEMU_PPC_FEATURE2_ARCH_2_07 |
+                  QEMU_PPC_FEATURE2_VEC_CRYPTO | QEMU_PPC_FEATURE2_HAS_TAR));
     GET_FEATURE2(PPC2_ISA300, QEMU_PPC_FEATURE2_ARCH_3_00 |
                  QEMU_PPC_FEATURE2_DARN | QEMU_PPC_FEATURE2_HAS_IEEE128);
     GET_FEATURE2(PPC2_ISA310, QEMU_PPC_FEATURE2_ARCH_3_1 |
target/ppc/cpu.h +1 -3
index 91ef34cb21..df1e943d8d 100644
@@ -2577,8 +2577,6 @@ enum {
     PPC2_FP_CVT_ISA206 = 0x0000000000000400ULL,
     /* ISA 2.06B floating point test instructions                            */
     PPC2_FP_TST_ISA206 = 0x0000000000000800ULL,
-    /* ISA 2.07 bctar instruction                                            */
-    PPC2_BCTAR_ISA207  = 0x0000000000001000ULL,
     /* ISA 2.07 load/store quadword                                          */
     PPC2_LSQ_ISA207    = 0x0000000000002000ULL,
     /* ISA 2.07 Altivec                                                      */
@@ -2610,7 +2608,7 @@ enum {
                         PPC2_ISA205 | PPC2_ISA207 | PPC2_PERM_ISA206 | \
                         PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 | \
                         PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | \
-                        PPC2_BCTAR_ISA207 | PPC2_LSQ_ISA207 | \
+                        PPC2_LSQ_ISA207 | \
                         PPC2_ALTIVEC_207 | PPC2_ISA207S | PPC2_DFP | \
                         PPC2_FP_CVT_S64 | PPC2_TM | PPC2_PM_ISA206 | \
                         PPC2_ISA300 | PPC2_ISA310 | PPC2_MEM_LWSYNC | \
target/ppc/cpu_init.c +1 -1
index b913edd243..c250f805a7 100644
@@ -6349,7 +6349,7 @@ POWERPC_FAMILY(POWER8)(ObjectClass *oc, const void *data)
     pcc->insns_flags2 = PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |
                         PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 |
                         PPC2_ATOMIC_ISA206 | PPC2_FP_CVT_ISA206 |
-                        PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |
+                        PPC2_FP_TST_ISA206 |
                         PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
                         PPC2_ISA205 | PPC2_ISA207S | PPC2_FP_CVT_S64 |
                         PPC2_TM | PPC2_PM_ISA206 | PPC2_MEM_LWSYNC |
target/ppc/cpu_init.h +1 -1
index 1c41c0d349..7dd587908e 100644
@@ -17,7 +17,7 @@
 #define PPC_INSNS_FLAGS2_POWER_COMMON                                \
     (PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |                 \
      PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 |      \
-     PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |   \
+     PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 |                       \
      PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 | PPC2_ISA205 |              \
      PPC2_ISA207S | PPC2_FP_CVT_S64 | PPC2_ISA300 | PPC2_PRCNTL |    \
      PPC2_MEM_LWSYNC | PPC2_BCDA_ISA206)
target/ppc/translate/branch-impl.c.inc +1 -1
index c2889c1d12..77eac0a239 100644
@@ -210,5 +210,5 @@ static bool bclr_helper(DisasContext *ctx, arg_bclr *a, int type)
 
 TRANS(BCLR, bclr_helper, BCOND_LR)
 TRANS(BCCTR, bclr_helper, BCOND_CTR)
-TRANS_FLAGS2(BCTAR_ISA207, BCTAR, bclr_helper, BCOND_TAR)
+TRANS_FLAGS2(ISA207, BCTAR, bclr_helper, BCOND_TAR)
 
Signature · 1 line
2.55.0
Reviewed-by on v2 31/37
#70 lore raw

Re: [PATCH v2 31/37] target/ppc: Use PPC2_ISA207 instead of PPC2_BCTAR_ISA207

Quoted 8 lines · Chinmay Rath

PPC2_BCTAR_ISA207 is only ever set in the CPUPPCState's insns_flags2 alongside PPC2_ISA207. Checks made by PPC2_BCTAR_ISA207 could be replaced with PPC2_ISA207, hence rendering it useless and apt for removal. This patch does the same.

Quoted 7 lines
---
 linux-user/ppc/elfload.c               | 7 +++----
 target/ppc/cpu.h                       | 4 +---
 target/ppc/cpu_init.c                  | 2 +-
 target/ppc/cpu_init.h                  | 2 +-
 target/ppc/translate/branch-impl.c.inc | 2 +-
 5 files changed, 7 insertions(+), 10 deletions(-)
Quoted 81 lines
linux-user/ppc/elfload.c +3 -4
index 0d54da9803..5c6065e08a 100644
@@ -116,10 +116,9 @@ abi_ulong get_elf_hwcap2(CPUState *cs)
     do { if (cpu->env.insns_flags2 & flag) { features |= feature; } } while (0)
 
     GET_FEATURE(PPC_ISEL, QEMU_PPC_FEATURE2_HAS_ISEL);
-    GET_FEATURE2(PPC2_BCTAR_ISA207, QEMU_PPC_FEATURE2_HAS_TAR);
-    GET_FEATURE2((PPC2_BCTAR_ISA207 | PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
-                  PPC2_ISA207S), QEMU_PPC_FEATURE2_ARCH_2_07 |
-                  QEMU_PPC_FEATURE2_VEC_CRYPTO);
+    GET_FEATURE2((PPC2_ISA207 | PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
+                  PPC2_ISA207S), (QEMU_PPC_FEATURE2_ARCH_2_07 |
+                  QEMU_PPC_FEATURE2_VEC_CRYPTO | QEMU_PPC_FEATURE2_HAS_TAR));
     GET_FEATURE2(PPC2_ISA300, QEMU_PPC_FEATURE2_ARCH_3_00 |
                  QEMU_PPC_FEATURE2_DARN | QEMU_PPC_FEATURE2_HAS_IEEE128);
     GET_FEATURE2(PPC2_ISA310, QEMU_PPC_FEATURE2_ARCH_3_1 |
target/ppc/cpu.h +1 -3
index 91ef34cb21..df1e943d8d 100644
@@ -2577,8 +2577,6 @@ enum {
     PPC2_FP_CVT_ISA206 = 0x0000000000000400ULL,
     /* ISA 2.06B floating point test instructions                            */
     PPC2_FP_TST_ISA206 = 0x0000000000000800ULL,
-    /* ISA 2.07 bctar instruction                                            */
-    PPC2_BCTAR_ISA207  = 0x0000000000001000ULL,
     /* ISA 2.07 load/store quadword                                          */
     PPC2_LSQ_ISA207    = 0x0000000000002000ULL,
     /* ISA 2.07 Altivec                                                      */
@@ -2610,7 +2608,7 @@ enum {
                         PPC2_ISA205 | PPC2_ISA207 | PPC2_PERM_ISA206 | \
                         PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 | \
                         PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | \
-                        PPC2_BCTAR_ISA207 | PPC2_LSQ_ISA207 | \
+                        PPC2_LSQ_ISA207 | \
                         PPC2_ALTIVEC_207 | PPC2_ISA207S | PPC2_DFP | \
                         PPC2_FP_CVT_S64 | PPC2_TM | PPC2_PM_ISA206 | \
                         PPC2_ISA300 | PPC2_ISA310 | PPC2_MEM_LWSYNC | \
target/ppc/cpu_init.c +1 -1
index b913edd243..c250f805a7 100644
@@ -6349,7 +6349,7 @@ POWERPC_FAMILY(POWER8)(ObjectClass *oc, const void *data)
     pcc->insns_flags2 = PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |
                         PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 |
                         PPC2_ATOMIC_ISA206 | PPC2_FP_CVT_ISA206 |
-                        PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |
+                        PPC2_FP_TST_ISA206 |
                         PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
                         PPC2_ISA205 | PPC2_ISA207S | PPC2_FP_CVT_S64 |
                         PPC2_TM | PPC2_PM_ISA206 | PPC2_MEM_LWSYNC |
target/ppc/cpu_init.h +1 -1
index 1c41c0d349..7dd587908e 100644
@@ -17,7 +17,7 @@
 #define PPC_INSNS_FLAGS2_POWER_COMMON                                \
     (PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |                 \
      PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 |      \
-     PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | PPC2_BCTAR_ISA207 |   \
+     PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 |                       \
      PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 | PPC2_ISA205 |              \
      PPC2_ISA207S | PPC2_FP_CVT_S64 | PPC2_ISA300 | PPC2_PRCNTL |    \
      PPC2_MEM_LWSYNC | PPC2_BCDA_ISA206)
target/ppc/translate/branch-impl.c.inc +1 -1
index c2889c1d12..77eac0a239 100644
@@ -210,5 +210,5 @@ static bool bclr_helper(DisasContext *ctx, arg_bclr *a, int type)
 
 TRANS(BCLR, bclr_helper, BCOND_LR)
 TRANS(BCCTR, bclr_helper, BCOND_CTR)
-TRANS_FLAGS2(BCTAR_ISA207, BCTAR, bclr_helper, BCOND_TAR)
+TRANS_FLAGS2(ISA207, BCTAR, bclr_helper, BCOND_TAR)
 
Signature · 1 line
2.55.0
#71 lore raw

PATCH v2 32/37 target/ppc: Use PPC2_ISA207 instead of PPC2_LSQ_ISA207

In reply to Chinmay Rath

PPC2_LSQ_ISA207 is only ever set in the CPUPPCState's insns_flags2 alongside PPC2_ISA207. Checks made by PPC2_LSQ_ISA207 could be replaced with PPC2_ISA207, hence rendering it useless and apt for removal. This patch does the same.

---
 linux-user/ppc/elfload.c                   | 2 +-
 target/ppc/cpu.h                           | 3 ---
 target/ppc/cpu_init.c                      | 3 +--
 target/ppc/cpu_init.h                      | 2 +-
 target/ppc/translate/fixedpoint-impl.c.inc | 4 ++--
 5 files changed, 5 insertions(+), 9 deletions(-)
linux-user/ppc/elfload.c +1 -1
index 5c6065e08a..76ab9f43de 100644
@@ -116,7 +116,7 @@ abi_ulong get_elf_hwcap2(CPUState *cs)
     do { if (cpu->env.insns_flags2 & flag) { features |= feature; } } while (0)
 
     GET_FEATURE(PPC_ISEL, QEMU_PPC_FEATURE2_HAS_ISEL);
-    GET_FEATURE2((PPC2_ISA207 | PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
+    GET_FEATURE2((PPC2_ISA207 | PPC2_ALTIVEC_207 |
                   PPC2_ISA207S), (QEMU_PPC_FEATURE2_ARCH_2_07 |
                   QEMU_PPC_FEATURE2_VEC_CRYPTO | QEMU_PPC_FEATURE2_HAS_TAR));
     GET_FEATURE2(PPC2_ISA300, QEMU_PPC_FEATURE2_ARCH_3_00 |
target/ppc/cpu.h +0 -3
index df1e943d8d..b3cd8d85df 100644
@@ -2577,8 +2577,6 @@ enum {
     PPC2_FP_CVT_ISA206 = 0x0000000000000400ULL,
     /* ISA 2.06B floating point test instructions                            */
     PPC2_FP_TST_ISA206 = 0x0000000000000800ULL,
-    /* ISA 2.07 load/store quadword                                          */
-    PPC2_LSQ_ISA207    = 0x0000000000002000ULL,
     /* ISA 2.07 Altivec                                                      */
     PPC2_ALTIVEC_207   = 0x0000000000004000ULL,
     /* PowerISA 2.07 Book3s specification                                    */
@@ -2608,7 +2606,6 @@ enum {
                         PPC2_ISA205 | PPC2_ISA207 | PPC2_PERM_ISA206 | \
                         PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 | \
                         PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 | \
-                        PPC2_LSQ_ISA207 | \
                         PPC2_ALTIVEC_207 | PPC2_ISA207S | PPC2_DFP | \
                         PPC2_FP_CVT_S64 | PPC2_TM | PPC2_PM_ISA206 | \
                         PPC2_ISA300 | PPC2_ISA310 | PPC2_MEM_LWSYNC | \
target/ppc/cpu_init.c +1 -2
index c250f805a7..4cc8b93dca 100644
@@ -6349,8 +6349,7 @@ POWERPC_FAMILY(POWER8)(ObjectClass *oc, const void *data)
     pcc->insns_flags2 = PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |
                         PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 |
                         PPC2_ATOMIC_ISA206 | PPC2_FP_CVT_ISA206 |
-                        PPC2_FP_TST_ISA206 |
-                        PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 |
+                        PPC2_FP_TST_ISA206 | PPC2_ALTIVEC_207 |
                         PPC2_ISA205 | PPC2_ISA207S | PPC2_FP_CVT_S64 |
                         PPC2_TM | PPC2_PM_ISA206 | PPC2_MEM_LWSYNC |
                         PPC2_BCDA_ISA206;
target/ppc/cpu_init.h +1 -1
index 7dd587908e..6961007926 100644
@@ -18,7 +18,7 @@
     (PPC2_VSX | PPC2_ISA207 | PPC2_DFP | PPC2_DBRX |                 \
      PPC2_PERM_ISA206 | PPC2_DIVE_ISA206 | PPC2_ATOMIC_ISA206 |      \
      PPC2_FP_CVT_ISA206 | PPC2_FP_TST_ISA206 |                       \
-     PPC2_LSQ_ISA207 | PPC2_ALTIVEC_207 | PPC2_ISA205 |              \
+     PPC2_ALTIVEC_207 | PPC2_ISA205 |                                \
      PPC2_ISA207S | PPC2_FP_CVT_S64 | PPC2_ISA300 | PPC2_PRCNTL |    \
      PPC2_MEM_LWSYNC | PPC2_BCDA_ISA206)
 
target/ppc/translate/fixedpoint-impl.c.inc +2 -2
index cef67d6431..b9c20b66dc 100644
@@ -116,7 +116,7 @@ static bool do_ldst_quad(DisasContext *ctx, arg_D *a, bool store, bool prefixed)
 
     REQUIRE_INSNS_FLAGS(ctx, 64BX);
 
-    if (!prefixed && !(ctx->insns_flags2 & PPC2_LSQ_ISA207)) {
+    if (!prefixed && !(ctx->insns_flags2 & PPC2_ISA207)) {
         /* lq and stq were privileged prior to V. 2.07 */
         REQUIRE_SV(ctx);
 
@@ -280,7 +280,7 @@ TRANS64(STDCX, do_store_cond, MO_UQ);
 static bool trans_STQCX(DisasContext *ctx, arg_STQCX *a)
 {
     REQUIRE_64BIT(ctx);
-    REQUIRE_INSNS_FLAGS2(ctx, LSQ_ISA207);
+    REQUIRE_INSNS_FLAGS2(ctx, ISA207);
 #if defined(TARGET_PPC64)
     TCGLabel *lfail = gen_new_label();
     TCGv ea = do_ea_calc(ctx, a->ra, cpu_gpr[a->rb]);
Signature · 1 line
2.55.0