s390x/tcg: Implement Vector-Enhancements Facility 2
v6 cover s390x/tcg: Implement Vector-Enhancements Facility 2
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER)
VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR)
VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG)
VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ) VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)
v5 -> v6:
* Move fix to #1
* Include max CPU model cleanups
* "target/s390x: add S390_FEAT_VECTOR_ENH2 to qemu CPU model" -> Take care of compat machines
* "tests/tcg/s390x: Tests for Vector Enhancements Facility 2" -> Add missing newline to end of header file
-> Resolve simple conflict in Makefile
David Hildenbrand (2):
s390x/cpu_models: drop "msa5" from the TCG "max" model s390x/cpu_models: make "max" match the unmodified "qemu" CPU model
under TCG
David Miller (9):
target/s390x: vxeh2: vector convert short/32b target/s390x: vxeh2: vector string search target/s390x: vxeh2: Update for changes to vector shifts target/s390x: vxeh2: vector shift double by bit target/s390x: vxeh2: vector {load, store} elements reversed target/s390x: vxeh2: vector {load, store} byte reversed elements target/s390x: vxeh2: vector {load, store} byte reversed element target/s390x: add S390_FEAT_VECTOR_ENH2 to qemu CPU model tests/tcg/s390x: Tests for Vector Enhancements Facility 2
Richard Henderson (2):
target/s390x: Fix writeback to v1 in helper_vstl
tcg: Implement tcg_gen_{h,w}swap_{i32,i64}
hw/s390x/s390-virtio-ccw.c | 3 + include/tcg/tcg-op.h | 6 + target/s390x/cpu_models.c | 26 +- target/s390x/gen-features.c | 14 +- target/s390x/helper.h | 13 + target/s390x/tcg/insn-data.def | 40 ++- target/s390x/tcg/translate.c | 3 +- target/s390x/tcg/translate_vx.c.inc | 461 ++++++++++++++++++++++++--- target/s390x/tcg/vec_fpu_helper.c | 31 ++ target/s390x/tcg/vec_helper.c | 2 - target/s390x/tcg/vec_int_helper.c | 55 ++++ target/s390x/tcg/vec_string_helper.c | 99 ++++++ tcg/tcg-op.c | 30 ++ tests/tcg/s390x/Makefile.target | 8 + tests/tcg/s390x/vx.h | 19 ++ tests/tcg/s390x/vxeh2_vcvt.c | 88 +++++ tests/tcg/s390x/vxeh2_vlstr.c | 139 ++++++++ tests/tcg/s390x/vxeh2_vs.c | 93 ++++++ 18 files changed, 1051 insertions(+), 79 deletions(-) create mode 100644 tests/tcg/s390x/vx.h create mode 100644 tests/tcg/s390x/vxeh2_vcvt.c create mode 100644 tests/tcg/s390x/vxeh2_vlstr.c create mode 100644 tests/tcg/s390x/vxeh2_vs.c
Signature · 1 line
2.35.1
PATCH v6 1/13 target/s390x: Fix writeback to v1 in helper_vstl
From: Richard Henderson <[email protected]>
--- target/s390x/tcg/vec_helper.c | 2 -- 1 file changed, 2 deletions(-)
target/s390x/tcg/vec_helper.c +0 -2
@@ -200,7 +200,6 @@ void HELPER(vstl)(CPUS390XState *env, const void *v1, uint64_t addr, addr =wrap_address (env, addr +8 );cpu_stq_data_ra (env, addr,s390_vec_read_element64 (v1,1 ),GETPC ()); }else { -S390Vector tmp = {};int i;for (i =0 ; i < bytes; i++) {
@@ -209,6 +208,5 @@ void HELPER(vstl)(CPUS390XState *env, const void *v1, uint64_t addr,cpu_stb_data_ra (env, addr, byte,GETPC ()); addr =wrap_address (env, addr +1 ); } - *(S390Vector *)v1 = tmp; } }
Signature · 1 line
2.35.1
PATCH v6 2/13 s390x/cpu_models: drop "msa5" from the TCG "max" model
In reply to David HildenbrandWe don't include the "msa5" feature in the "qemu" model because it generates a warning. The PoP states:
"The message-security-assist extension 5 requires the secure-hash-algorithm (SHA-512) capabilities of the message-security-assist extension 2 as a prereq- uisite. (March, 2015)"
As SHA-512 won't be supported in the near future, let's just drop the feature from the "max" model. This avoids the warning and allows us for making the "max" model match the "qemu" model (except for compat machines). We don't lose much, as we only implement the function stubs for MSA, excluding any real subfunctions.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/897
--- target/s390x/gen-features.c | 2 -- 1 file changed, 2 deletions(-)
target/s390x/gen-features.c +0 -2
@@ -743,8 +743,6 @@ static uint16_t qemu_LATEST[] = { };/* add all new definitions before this point */ static uint16_t qemu_MAX[] = { -/* generates a dependency warning, leave it out for now */ -S390_FEAT_MSA_EXT_5 , };/****** END FEATURE DEFS ******/
Signature · 1 line
2.35.1
PATCH v6 3/13 s390x/cpu_models: make "max" match the unmodified "qemu" CPU model under TCG
In reply to David HildenbrandBefore we were able to bump up the qemu CPU model to a z13, we included some experimental features during development in the "max" model only. Nowadays, the "max" model corresponds exactly to the "qemu" CPU model of the latest QEMU machine under TCG.
Let's remove all the special casing, effectively making both models match completely from now on, and clean up.
--- target/s390x/cpu_models.c | 26 +++++++------------------- target/s390x/gen-features.c | 11 ++++++----- 2 files changed, 13 insertions(+), 24 deletions(-)
target/s390x/cpu_models.c +7 -19
@@ -89,7 +89,6 @@ static S390CPUDef s390_cpu_defs[] = {#define QEMU_MAX_CPU_TYPE 0x8561#define QEMU_MAX_CPU_GEN 15#define QEMU_MAX_CPU_EC_GA 1 -static const S390FeatInit qemu_max_cpu_feat_init = {S390_FEAT_LIST_QEMU_MAX };static S390FeatBitmap qemu_max_cpu_feat;/* features part of a base model but not relevant for finding a base model */
@@ -728,7 +727,6 @@ static void s390_cpu_model_initfn(Object *obj) } } -static S390CPUDef s390_qemu_cpu_def;static S390CPUModel s390_qemu_cpu_model;/* Set the qemu CPU model (on machine initialization). Must not be called
@@ -742,17 +740,8 @@ void s390_set_qemu_cpu_model(uint16_t type, uint8_t gen, uint8_t ec_ga,g_assert (def);g_assert (QTAILQ_EMPTY_RCU (&cpus)); -/* TCG emulates some features that can usually not be enabled with -* the emulated machine generation. Make sure they can be enabled -* when using the QEMU model by adding them to full_feat. We have -* to copy the definition to do that. -*/ -memcpy (&s390_qemu_cpu_def, def,sizeof (s390_qemu_cpu_def)); -bitmap_or (s390_qemu_cpu_def.full_feat, s390_qemu_cpu_def.full_feat, - qemu_max_cpu_feat,S390_FEAT_MAX ); -/* build the CPU model */ - s390_qemu_cpu_model.def = &s390_qemu_cpu_def; + s390_qemu_cpu_model.def = def;bitmap_zero (s390_qemu_cpu_model.features,S390_FEAT_MAX );s390_init_feat_bitmap (feat_init, s390_qemu_cpu_model.features); }
@@ -885,9 +874,8 @@ static void s390_max_cpu_model_class_init(ObjectClass *oc, void *data)/* * The "max" model is neither static nor migration safe. Under KVM -* it represents the "host" model. Under TCG it represents some kind of -* "qemu" CPU model without compat handling and maybe with some additional -* CPU features that are not yet unlocked in the "qemu" model. +* it represents the "host" model. Under TCG it represents the "qemu" CPU +* model of the latest QEMU machine. */ xcc->desc ="Enables all features supported by the accelerator in the current host" ;
@@ -966,13 +954,13 @@ static void init_ignored_base_feat(void)static void register_types (void ) { -static const S390FeatInit qemu_latest_init = {S390_FEAT_LIST_QEMU_LATEST }; +static const S390FeatInit qemu_max_init = {S390_FEAT_LIST_QEMU_MAX };int i;init_ignored_base_feat ();/* init all bitmaps from gnerated data initially */ -s390_init_feat_bitmap (qemu_max_cpu_feat_init, qemu_max_cpu_feat); +s390_init_feat_bitmap (qemu_max_init, qemu_max_cpu_feat);for (i =0 ; i <ARRAY_SIZE (s390_cpu_defs); i++) {s390_init_feat_bitmap (s390_cpu_defs[i].base_init, s390_cpu_defs[i].base_feat);
@@ -982,9 +970,9 @@ static void register_types(void) s390_cpu_defs[i].full_feat); } -/* initialize the qemu model with latest definition */ +/* initialize the qemu model with the maximum definition ("max" model) */ s390_set_qemu_cpu_model (QEMU_MAX_CPU_TYPE ,QEMU_MAX_CPU_GEN , -QEMU_MAX_CPU_EC_GA , qemu_latest_init); +QEMU_MAX_CPU_EC_GA , qemu_max_init);for (i =0 ; i <ARRAY_SIZE (s390_cpu_defs); i++) {char *base_name =s390_base_cpu_type_name (s390_cpu_defs[i].name);
target/s390x/gen-features.c +6 -5
@@ -738,11 +738,13 @@ static uint16_t qemu_V6_2[] = { S390_FEAT_VECTOR_ENH, }; -static uint16_t qemu_LATEST[] = { -S390_FEAT_MISC_INSTRUCTION_EXT3 , -}; -/* add all new definitions before this point */ +/* +* Features for the "qemu" CPU model of the latest QEMU machine and the "max" +* CPU model under TCG. Don't include features that are not part of the full +* feature set of the current "max" CPU model generation. +*/ static uint16_t qemu_MAX[] = { +S390_FEAT_MISC_INSTRUCTION_EXT3 , };/****** END FEATURE DEFS ******/
@@ -864,7 +866,6 @@ static FeatGroupDefSpec QemuFeatDef[] = {QEMU_FEAT_INITIALIZER (V4_1 ),QEMU_FEAT_INITIALIZER (V6_0 ),QEMU_FEAT_INITIALIZER (V6_2 ), -QEMU_FEAT_INITIALIZER (LATEST ), QEMU_FEAT_INITIALIZER(MAX), };
Signature · 1 line
2.35.1
PATCH v6 4/13 tcg: Implement tcg_gen_{h,w}swap_{i32,i64}
In reply to David HildenbrandFrom: Richard Henderson <[email protected]>
Swap half-words (16-bit) and words (32-bit) within a larger value. Mirrors functions of the same names within include/qemu/bitops.h.
--- include/tcg/tcg-op.h | 6 ++++++ tcg/tcg-op.c | 30 ++++++++++++++++++++++++++++++ 2 files changed, 36 insertions(+)
include/tcg/tcg-op.h +6 -0
@@ -332,6 +332,7 @@ void tcg_gen_ext8u_i32(TCGv_i32 ret, TCGv_i32 arg);void tcg_gen_ext16u_i32 (TCGv_i32 ret,TCGv_i32 arg);void tcg_gen_bswap16_i32 (TCGv_i32 ret,TCGv_i32 arg,int flags);void tcg_gen_bswap32_i32 (TCGv_i32 ret,TCGv_i32 arg); +void tcg_gen_hswap_i32 (TCGv_i32 ret,TCGv_i32 arg);void tcg_gen_smin_i32 (TCGv_i32 ,TCGv_i32 arg1,TCGv_i32 arg2);void tcg_gen_smax_i32 (TCGv_i32 ,TCGv_i32 arg1,TCGv_i32 arg2);void tcg_gen_umin_i32 (TCGv_i32 ,TCGv_i32 arg1,TCGv_i32 arg2);
@@ -531,6 +532,8 @@ void tcg_gen_ext32u_i64(TCGv_i64 ret, TCGv_i64 arg);void tcg_gen_bswap16_i64 (TCGv_i64 ret,TCGv_i64 arg,int flags);void tcg_gen_bswap32_i64 (TCGv_i64 ret,TCGv_i64 arg,int flags);void tcg_gen_bswap64_i64 (TCGv_i64 ret,TCGv_i64 arg); +void tcg_gen_hswap_i64 (TCGv_i64 ret,TCGv_i64 arg); +void tcg_gen_wswap_i64 (TCGv_i64 ret,TCGv_i64 arg);void tcg_gen_smin_i64 (TCGv_i64 ,TCGv_i64 arg1,TCGv_i64 arg2);void tcg_gen_smax_i64 (TCGv_i64 ,TCGv_i64 arg1,TCGv_i64 arg2);void tcg_gen_umin_i64 (TCGv_i64 ,TCGv_i64 arg1,TCGv_i64 arg2);
@@ -1077,6 +1080,8 @@ void tcg_gen_stl_vec(TCGv_vec r, TCGv_ptr base, TCGArg offset, TCGType t);#define tcg_gen_bswap32_tl tcg_gen_bswap32_i64#define tcg_gen_bswap64_tl tcg_gen_bswap64_i64#define tcg_gen_bswap_tl tcg_gen_bswap64_i64 +#define tcg_gen_hswap_tl tcg_gen_hswap_i64 +#define tcg_gen_wswap_tl tcg_gen_wswap_i64#define tcg_gen_concat_tl_i64 tcg_gen_concat32_i64#define tcg_gen_extr_i64_tl tcg_gen_extr32_i64#define tcg_gen_andc_tl tcg_gen_andc_i64
@@ -1192,6 +1197,7 @@ void tcg_gen_stl_vec(TCGv_vec r, TCGv_ptr base, TCGArg offset, TCGType t);#define tcg_gen_bswap16_tl tcg_gen_bswap16_i32#define tcg_gen_bswap32_tl (D ,S ,F ) tcg_gen_bswap32_i32(D, S)#define tcg_gen_bswap_tl tcg_gen_bswap32_i32 +#define tcg_gen_hswap_tl tcg_gen_hswap_i32#define tcg_gen_concat_tl_i64 tcg_gen_concat_i32_i64#define tcg_gen_extr_i64_tl tcg_gen_extr_i64_i32#define tcg_gen_andc_tl tcg_gen_andc_i32
tcg/tcg-op.c +30 -0
@@ -1056,6 +1056,12 @@ void tcg_gen_bswap32_i32(TCGv_i32 ret, TCGv_i32 arg) } } +void tcg_gen_hswap_i32 (TCGv_i32 ret,TCGv_i32 arg) +{ +/* Swapping 2 16-bit elements is a rotate. */ +tcg_gen_rotli_i32 (ret, arg,16 ); +} +void tcg_gen_smin_i32 (TCGv_i32 ret,TCGv_i32 a,TCGv_i32 b) {tcg_gen_movcond_i32 (TCG_COND_LT , ret, a, b, a, b);
@@ -1792,6 +1798,30 @@ void tcg_gen_bswap64_i64(TCGv_i64 ret, TCGv_i64 arg) } } +void tcg_gen_hswap_i64 (TCGv_i64 ret,TCGv_i64 arg) +{ +uint64_t m =0x0000ffff0000ffffull ; +TCGv_i64 t0 =tcg_temp_new_i64 (); +TCGv_i64 t1 =tcg_temp_new_i64 (); + +/* See include/qemu/bitops.h, hswap64. */ +tcg_gen_rotli_i64 (t1, arg,32 ); +tcg_gen_andi_i64 (t0, t1, m); +tcg_gen_shli_i64 (t0, t0,16 ); +tcg_gen_shri_i64 (t1, t1,16 ); +tcg_gen_andi_i64 (t1, t1, m); +tcg_gen_or_i64 (ret, t0, t1); + +tcg_temp_free_i64 (t0); +tcg_temp_free_i64 (t1); +} + +void tcg_gen_wswap_i64 (TCGv_i64 ret,TCGv_i64 arg) +{ +/* Swapping 2 32-bit elements is a rotate. */ +tcg_gen_rotli_i64 (ret, arg,32 ); +} +void tcg_gen_not_i64 (TCGv_i64 ret,TCGv_i64 arg) {if (TCG_TARGET_REG_BITS ==32 ) {
Signature · 1 line
2.35.1
PATCH v6 5/13 target/s390x: vxeh2: vector convert short/32b
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- target/s390x/helper.h | 4 +++ target/s390x/tcg/translate_vx.c.inc | 44 ++++++++++++++++++++++++++--- target/s390x/tcg/vec_fpu_helper.c | 31 ++++++++++++++++++++ 3 files changed, 75 insertions(+), 4 deletions(-)
target/s390x/helper.h +4 -0
@@ -275,6 +275,10 @@ DEF_HELPER_FLAGS_5(gvec_vfche64, TCG_CALL_NO_WG, void, ptr, cptr, cptr, env, i32 DEF_HELPER_5(gvec_vfche64_cc, void, ptr, cptr, cptr, env, i32) DEF_HELPER_FLAGS_5(gvec_vfche128, TCG_CALL_NO_WG, void, ptr, cptr, cptr, env, i32) DEF_HELPER_5(gvec_vfche128_cc, void, ptr, cptr, cptr, env, i32) +DEF_HELPER_FLAGS_4(gvec_vcdg32, TCG_CALL_NO_WG, void, ptr, cptr, env, i32) +DEF_HELPER_FLAGS_4(gvec_vcdlg32, TCG_CALL_NO_WG, void, ptr, cptr, env, i32) +DEF_HELPER_FLAGS_4(gvec_vcgd32, TCG_CALL_NO_WG, void, ptr, cptr, env, i32) +DEF_HELPER_FLAGS_4(gvec_vclgd32, TCG_CALL_NO_WG, void, ptr, cptr, env, i32) DEF_HELPER_FLAGS_4(gvec_vcdg64, TCG_CALL_NO_WG, void, ptr, cptr, env, i32) DEF_HELPER_FLAGS_4(gvec_vcdlg64, TCG_CALL_NO_WG, void, ptr, cptr, env, i32)DEF_HELPER_FLAGS_4 (gvec_vcgd64,TCG_CALL_NO_WG , void, ptr, cptr, env, i32)
target/s390x/tcg/translate_vx.c.inc +40 -4
63 lines not on this page · raw message
target/s390x/tcg/vec_fpu_helper.c +31 -0
@@ -175,6 +175,30 @@ static void vop128_2(S390Vector *v1, const S390Vector *v2, CPUS390XState *env, *v1 = tmp; } +static float32 vcdg32 (float32 a,float_status *s) +{ +return int32_to_float32 (a, s); +} + +static float32 vcdlg32 (float32 a,float_status *s) +{ +return uint32_to_float32 (a, s); +} + +static float32 vcgd32 (float32 a,float_status *s) +{ +const float32 tmp =float32_to_int32 (a, s); + +return float32_is_any_nan (a) ?INT32_MIN : tmp; +} + +static float32 vclgd32 (float32 a,float_status *s) +{ +const float32 tmp =float32_to_uint32 (a, s); + +return float32_is_any_nan (a) ?0 : tmp; +} +static float64 vcdg64 (float64 a,float_status *s) {return int64_to_float64 (a, s);
@@ -210,6 +234,9 @@ void HELPER(gvec_##NAME##BITS)(void *v1, const void *v2, CPUS390XState *env, \ vop##BITS##_2(v1, v2, env, se, XxC, erm, FN, GETPC()); \ } +#define DEF_GVEC_VOP2_32 (NAME ) \ +DEF_GVEC_VOP2_FN(NAME, NAME##32, 32) +#define DEF_GVEC_VOP2_64 (NAME ) \ DEF_GVEC_VOP2_FN(NAME, NAME##64, 64)
@@ -218,6 +245,10 @@ DEF_GVEC_VOP2_FN(NAME, float32_##OP, 32) \ DEF_GVEC_VOP2_FN(NAME, float64_##OP, 64) \ DEF_GVEC_VOP2_FN(NAME, float128_##OP, 128) +DEF_GVEC_VOP2_32(vcdg) +DEF_GVEC_VOP2_32(vcdlg) +DEF_GVEC_VOP2_32(vcgd) +DEF_GVEC_VOP2_32(vclgd) DEF_GVEC_VOP2_64(vcdg) DEF_GVEC_VOP2_64(vcdlg) DEF_GVEC_VOP2_64(vcgd)
Signature · 1 line
2.35.1
PATCH v6 6/13 target/s390x: vxeh2: vector string search
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- target/s390x/helper.h | 6 ++ target/s390x/tcg/insn-data.def | 2 + target/s390x/tcg/translate.c | 3 +- target/s390x/tcg/translate_vx.c.inc | 25 +++++++ target/s390x/tcg/vec_string_helper.c | 99 ++++++++++++++++++++++++++++ 5 files changed, 134 insertions(+), 1 deletion(-)
target/s390x/helper.h +6 -0
@@ -246,6 +246,12 @@ DEF_HELPER_6(gvec_vstrc_cc32, void, ptr, cptr, cptr, cptr, env, i32) DEF_HELPER_6(gvec_vstrc_cc_rt8, void, ptr, cptr, cptr, cptr, env, i32) DEF_HELPER_6(gvec_vstrc_cc_rt16, void, ptr, cptr, cptr, cptr, env, i32) DEF_HELPER_6(gvec_vstrc_cc_rt32, void, ptr, cptr, cptr, cptr, env, i32) +DEF_HELPER_6(gvec_vstrs_8, void, ptr, cptr, cptr, cptr, env, i32) +DEF_HELPER_6(gvec_vstrs_16, void, ptr, cptr, cptr, cptr, env, i32) +DEF_HELPER_6(gvec_vstrs_32, void, ptr, cptr, cptr, cptr, env, i32) +DEF_HELPER_6(gvec_vstrs_zs8, void, ptr, cptr, cptr, cptr, env, i32) +DEF_HELPER_6(gvec_vstrs_zs16, void, ptr, cptr, cptr, cptr, env, i32) +DEF_HELPER_6 (gvec_vstrs_zs32, void, ptr, cptr, cptr, cptr, env, i32) /* === Vector Floating-Point Instructions */DEF_HELPER_FLAGS_5 (gvec_vfa32,TCG_CALL_NO_WG , void, ptr, cptr, cptr, env, i32)
target/s390x/tcg/insn-data.def +2 -0
@@ -1246,6 +1246,8 @@ F(0xe75c, VISTR, VRR_a, V, 0, 0, 0, 0, vistr, 0, IF_VEC) /* VECTOR STRING RANGE COMPARE */ F(0xe78a, VSTRC, VRR_d, V, 0, 0, 0, 0, vstrc, 0, IF_VEC) +/* VECTOR STRING SEARCH */ + F(0xe78b, VSTRS, VRR_d, VE2, 0, 0, 0, 0, vstrs, 0, IF_VEC) /* === Vector Floating-Point Instructions */
target/s390x/tcg/translate.c +2 -1
@@ -6222,7 +6222,8 @@ enum DisasInsnEnum {#define FAC_PCI S390_FEAT_ZPCI/* z/PCI facility */ #define FAC_AIS S390_FEAT_ADAPTER_INT_SUPPRESSION#define FAC_V S390_FEAT_VECTOR/* vector facility */ -#define FAC_VE S390_FEAT_VECTOR_ENH/* vector enhancements facility 1 */ +#define FAC_VE S390_FEAT_VECTOR_ENH/* vector enhancements facility 1 */ +#define FAC_VE2 S390_FEAT_VECTOR_ENH2/* vector enhancements facility 2 */ #define FAC_MIE2 S390_FEAT_MISC_INSTRUCTION_EXT2/* miscellaneous-instruction-extensions facility 2 */ #define FAC_MIE3 S390_FEAT_MISC_INSTRUCTION_EXT3/* miscellaneous-instruction-extensions facility 3 */
target/s390x/tcg/translate_vx.c.inc +25 -0
@@ -2497,6 +2497,31 @@ static DisasJumpType op_vstrc(DisasContext *s, DisasOps *o) return DISAS_NEXT; } +static DisasJumpType op_vstrs(DisasContext *s, DisasOps *o) +{ + typedef void (*helper_vstrs)(TCGv_ptr, TCGv_ptr, TCGv_ptr, + TCGv_ptr, TCGv_ptr, TCGv_i32); + static const helper_vstrs fns[3][2] = { + { gen_helper_gvec_vstrs_8, gen_helper_gvec_vstrs_zs8 }, + { gen_helper_gvec_vstrs_16, gen_helper_gvec_vstrs_zs16 }, + { gen_helper_gvec_vstrs_32, gen_helper_gvec_vstrs_zs32 }, + }; + const uint8_t es = get_field(s, m5); + const uint8_t m6 = get_field(s, m6); + const bool zs = extract32(m6, 1, 1); + + if (es > ES_32 || m6 & ~2) { + gen_program_exception(s, PGM_SPECIFICATION); + return DISAS_NORETURN; + } + + gen_gvec_4_ptr(get_field(s, v1), get_field(s, v2), + get_field(s, v3), get_field(s, v4), + cpu_env, 0, fns[es][zs]); + set_cc_static(s); + return DISAS_NEXT; +} + static DisasJumpType op_vfa(DisasContext *s, DisasOps *o) { const uint8_t fpf = get_field(s, m4);
target/s390x/tcg/vec_string_helper.c +99 -0
102 lines not on this page · raw message
Signature · 1 line
2.35.1
PATCH v6 7/13 target/s390x: vxeh2: Update for changes to vector shifts
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- target/s390x/helper.h | 3 ++ target/s390x/tcg/insn-data.def | 12 ++--- target/s390x/tcg/translate_vx.c.inc | 75 ++++++++++++----------------- target/s390x/tcg/vec_int_helper.c | 55 +++++++++++++++++++++ 4 files changed, 95 insertions(+), 50 deletions(-)
target/s390x/helper.h +3 -0
@@ -203,8 +203,11 @@ DEF_HELPER_FLAGS_3(gvec_vpopct16, TCG_CALL_NO_RWG, void, ptr, cptr, i32) DEF_HELPER_FLAGS_4(gvec_verim8, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32) DEF_HELPER_FLAGS_4(gvec_verim16, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32) DEF_HELPER_FLAGS_4(gvec_vsl, TCG_CALL_NO_RWG, void, ptr, cptr, i64, i32) +DEF_HELPER_FLAGS_4(gvec_vsl_ve2, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32) DEF_HELPER_FLAGS_4(gvec_vsra, TCG_CALL_NO_RWG, void, ptr, cptr, i64, i32) +DEF_HELPER_FLAGS_4(gvec_vsra_ve2, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32) DEF_HELPER_FLAGS_4(gvec_vsrl, TCG_CALL_NO_RWG, void, ptr, cptr, i64, i32) +DEF_HELPER_FLAGS_4(gvec_vsrl_ve2, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32) DEF_HELPER_FLAGS_4(gvec_vscbi8, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32) DEF_HELPER_FLAGS_4(gvec_vscbi16, TCG_CALL_NO_RWG, void, ptr, cptr, cptr, i32)DEF_HELPER_4 (gvec_vtm, void, ptr, cptr, env, i32)
target/s390x/tcg/insn-data.def +6 -6
@@ -1204,19 +1204,19 @@ F(0xe778, VESRLV, VRR_c, V, 0, 0, 0, 0, vesv, 0, IF_VEC) F(0xe738, VESRL, VRS_a, V, la2, 0, 0, 0, ves, 0, IF_VEC) /* VECTOR SHIFT LEFT */ - F(0xe774, VSL, VRR_c, V, 0, 0, 0, 0, vsl, 0, IF_VEC) + E(0xe774, VSL, VRR_c, V, 0, 0, 0, 0, vsl, 0, 0, IF_VEC) /* VECTOR SHIFT LEFT BY BYTE */ - F(0xe775, VSLB, VRR_c, V, 0, 0, 0, 0, vsl, 0, IF_VEC) + E(0xe775, VSLB, VRR_c, V, 0, 0, 0, 0, vsl, 0, 1, IF_VEC) /* VECTOR SHIFT LEFT DOUBLE BY BYTE */ F(0xe777, VSLDB, VRI_d, V, 0, 0, 0, 0, vsldb, 0, IF_VEC) /* VECTOR SHIFT RIGHT ARITHMETIC */ - F(0xe77e, VSRA, VRR_c, V, 0, 0, 0, 0, vsra, 0, IF_VEC) + E(0xe77e, VSRA, VRR_c, V, 0, 0, 0, 0, vsra, 0, 0, IF_VEC) /* VECTOR SHIFT RIGHT ARITHMETIC BY BYTE */ - F(0xe77f, VSRAB, VRR_c, V, 0, 0, 0, 0, vsra, 0, IF_VEC) + E(0xe77f, VSRAB, VRR_c, V, 0, 0, 0, 0, vsra, 0, 1, IF_VEC) /* VECTOR SHIFT RIGHT LOGICAL */ - F(0xe77c, VSRL, VRR_c, V, 0, 0, 0, 0, vsrl, 0, IF_VEC) + E(0xe77c, VSRL, VRR_c, V, 0, 0, 0, 0, vsrl, 0, 0, IF_VEC) /* VECTOR SHIFT RIGHT LOGICAL BY BYTE */ - F(0xe77d, VSRLB, VRR_c, V, 0, 0, 0, 0, vsrl, 0, IF_VEC) + E(0xe77d, VSRLB, VRR_c, V, 0, 0, 0, 0, vsrl, 0, 1, IF_VEC) /* VECTOR SUBTRACT */ F(0xe7f7, VS, VRR_c, V, 0, 0, 0, 0, vs, 0, IF_VEC) /* VECTOR SUBTRACT COMPUTE BORROW INDICATION */
target/s390x/tcg/translate_vx.c.inc +31 -44
94 lines not on this page · raw message
target/s390x/tcg/vec_int_helper.c +55 -0
73 lines not on this page · raw message
Signature · 1 line
2.35.1
PATCH v6 8/13 target/s390x: vxeh2: vector shift double by bit
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- target/s390x/tcg/insn-data.def | 6 +++- target/s390x/tcg/translate_vx.c.inc | 55 +++++++++++++++++++++++++---- 2 files changed, 53 insertions(+), 8 deletions(-)
target/s390x/tcg/insn-data.def +5 -1
@@ -1207,12 +1207,16 @@ E(0xe774, VSL, VRR_c, V, 0, 0, 0, 0, vsl, 0, 0, IF_VEC) /* VECTOR SHIFT LEFT BY BYTE */ E(0xe775, VSLB, VRR_c, V, 0, 0, 0, 0, vsl, 0, 1, IF_VEC) +/* VECTOR SHIFT LEFT DOUBLE BY BIT */ + E(0xe786, VSLD, VRI_d, VE2, 0, 0, 0, 0, vsld, 0, 0, IF_VEC) /* VECTOR SHIFT LEFT DOUBLE BY BYTE */ - F(0xe777, VSLDB, VRI_d, V, 0, 0, 0, 0, vsldb, 0, IF_VEC) + E(0xe777, VSLDB, VRI_d, V, 0, 0, 0, 0, vsld, 0, 1, IF_VEC) /* VECTOR SHIFT RIGHT ARITHMETIC */ E(0xe77e, VSRA, VRR_c, V, 0, 0, 0, 0, vsra, 0, 0, IF_VEC) /* VECTOR SHIFT RIGHT ARITHMETIC BY BYTE */ E(0xe77f, VSRAB, VRR_c, V, 0, 0, 0, 0, vsra, 0, 1, IF_VEC) +/* VECTOR SHIFT RIGHT DOUBLE BY BIT */ + F(0xe787, VSRD, VRI_d, VE2, 0, 0, 0, 0, vsrd, 0, IF_VEC) /* VECTOR SHIFT RIGHT LOGICAL */ E(0xe77c, VSRL, VRR_c, V, 0, 0, 0, 0, vsrl, 0, 0, IF_VEC) /* VECTOR SHIFT RIGHT LOGICAL BY BYTE */
target/s390x/tcg/translate_vx.c.inc +48 -7
70 lines not on this page · raw message
Signature · 1 line
2.35.1
PATCH v6 9/13 target/s390x: vxeh2: vector {load, store} elements reversed
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- target/s390x/tcg/insn-data.def | 4 ++ target/s390x/tcg/translate_vx.c.inc | 84 +++++++++++++++++++++++++++++ 2 files changed, 88 insertions(+)
target/s390x/tcg/insn-data.def +4 -0
@@ -1037,6 +1037,8 @@ E(0xe741, VLEIH, VRI_a, V, 0, 0, 0, 0, vlei, 0, ES_16, IF_VEC) E(0xe743, VLEIF, VRI_a, V, 0, 0, 0, 0, vlei, 0, ES_32, IF_VEC) E(0xe742, VLEIG, VRI_a, V, 0, 0, 0, 0, vlei, 0, ES_64, IF_VEC) +/* VECTOR LOAD ELEMENTS REVERSED */ + F(0xe607, VLER, VRX, VE2, la2, 0, 0, 0, vler, 0, IF_VEC) /* VECTOR LOAD GR FROM VR ELEMENT */ F(0xe721, VLGV, VRS_c, V, la2, 0, r1, 0, vlgv, 0, IF_VEC) /* VECTOR LOAD LOGICAL ELEMENT AND ZERO */
@@ -1082,6 +1084,8 @@ E(0xe709, VSTEH, VRX, V, la2, 0, 0, 0, vste, 0, ES_16, IF_VEC) E(0xe70b, VSTEF, VRX, V, la2, 0, 0, 0, vste, 0, ES_32, IF_VEC) E(0xe70a, VSTEG, VRX, V, la2, 0, 0, 0, vste, 0, ES_64, IF_VEC) +/* VECTOR STORE ELEMENTS REVERSED */ + F(0xe60f, VSTER, VRX, VE2, la2, 0, 0, 0, vster, 0, IF_VEC) /* VECTOR STORE MULTIPLE */ F(0xe73e, VSTM, VRS_a, V, la2, 0, 0, 0, vstm, 0, IF_VEC) /* VECTOR STORE WITH LENGTH */
target/s390x/tcg/translate_vx.c.inc +84 -0
96 lines not on this page · raw message
Signature · 1 line
2.35.1
PATCH v6 10/13 target/s390x: vxeh2: vector {load, store} byte reversed elements
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- target/s390x/tcg/insn-data.def | 4 + target/s390x/tcg/translate_vx.c.inc | 113 ++++++++++++++++++++++++++++ 2 files changed, 117 insertions(+)
target/s390x/tcg/insn-data.def +4 -0
@@ -1027,6 +1027,8 @@ F(0xe756, VLR, VRR_a, V, 0, 0, 0, 0, vlr, 0, IF_VEC) /* VECTOR LOAD AND REPLICATE */ F(0xe705, VLREP, VRX, V, la2, 0, 0, 0, vlrep, 0, IF_VEC) +/* VECTOR LOAD BYTE REVERSED ELEMENTS */ + F(0xe606, VLBR, VRX, VE2, la2, 0, 0, 0, vlbr, 0, IF_VEC) /* VECTOR LOAD ELEMENT */ E(0xe700, VLEB, VRX, V, la2, 0, 0, 0, vle, 0, ES_8, IF_VEC) E(0xe701, VLEH, VRX, V, la2, 0, 0, 0, vle, 0, ES_16, IF_VEC)
@@ -1079,6 +1081,8 @@ F(0xe75f, VSEG, VRR_a, V, 0, 0, 0, 0, vseg, 0, IF_VEC) /* VECTOR STORE */ F(0xe70e, VST, VRX, V, la2, 0, 0, 0, vst, 0, IF_VEC) +/* VECTOR STORE BYTE REVERSED ELEMENTS */ + F(0xe60e, VSTBR, VRX, VE2, la2, 0, 0, 0, vstbr, 0, IF_VEC) /* VECTOR STORE ELEMENT */ E(0xe708, VSTEB, VRX, V, la2, 0, 0, 0, vste, 0, ES_8, IF_VEC) E(0xe709, VSTEH, VRX, V, la2, 0, 0, 0, vste, 0, ES_16, IF_VEC)
target/s390x/tcg/translate_vx.c.inc +113 -0
125 lines not on this page · raw message
Signature · 1 line
2.35.1
PATCH v6 11/13 target/s390x: vxeh2: vector {load, store} byte reversed element
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- target/s390x/tcg/insn-data.def | 12 ++++ target/s390x/tcg/translate_vx.c.inc | 85 +++++++++++++++++++++++++++++ 2 files changed, 97 insertions(+)
target/s390x/tcg/insn-data.def +12 -0
@@ -1027,6 +1027,14 @@ F(0xe756, VLR, VRR_a, V, 0, 0, 0, 0, vlr, 0, IF_VEC) /* VECTOR LOAD AND REPLICATE */ F(0xe705, VLREP, VRX, V, la2, 0, 0, 0, vlrep, 0, IF_VEC) +/* VECTOR LOAD BYTE REVERSED ELEMENT */ + E(0xe601, VLEBRH, VRX, VE2, la2, 0, 0, 0, vlebr, 0, ES_16, IF_VEC) + E(0xe603, VLEBRF, VRX, VE2, la2, 0, 0, 0, vlebr, 0, ES_32, IF_VEC) + E(0xe602, VLEBRG, VRX, VE2, la2, 0, 0, 0, vlebr, 0, ES_64, IF_VEC) +/* VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE */ + F(0xe605, VLBRREP, VRX, VE2, la2, 0, 0, 0, vlbrrep, 0, IF_VEC) +/* VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO */ + F(0xe604, VLLEBRZ, VRX, VE2, la2, 0, 0, 0, vllebrz, 0, IF_VEC) /* VECTOR LOAD BYTE REVERSED ELEMENTS */ F(0xe606, VLBR, VRX, VE2, la2, 0, 0, 0, vlbr, 0, IF_VEC) /* VECTOR LOAD ELEMENT */
@@ -1081,6 +1089,10 @@ F(0xe75f, VSEG, VRR_a, V, 0, 0, 0, 0, vseg, 0, IF_VEC) /* VECTOR STORE */ F(0xe70e, VST, VRX, V, la2, 0, 0, 0, vst, 0, IF_VEC) +/* VECTOR STORE BYTE REVERSED ELEMENT */ + E(0xe609, VSTEBRH, VRX, VE2, la2, 0, 0, 0, vstebr, 0, ES_16, IF_VEC) + E(0xe60b, VSTEBRF, VRX, VE2, la2, 0, 0, 0, vstebr, 0, ES_32, IF_VEC) + E(0xe60a, VSTEBRG, VRX, VE2, la2, 0, 0, 0, vstebr, 0, ES_64, IF_VEC) /* VECTOR STORE BYTE REVERSED ELEMENTS */ F(0xe60e, VSTBR, VRX, VE2, la2, 0, 0, 0, vstbr, 0, IF_VEC) /* VECTOR STORE ELEMENT */
target/s390x/tcg/translate_vx.c.inc +85 -0
97 lines not on this page · raw message
Signature · 1 line
2.35.1
PATCH v6 12/13 target/s390x: add S390_FEAT_VECTOR_ENH2 to qemu CPU model
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- hw/s390x/s390-virtio-ccw.c | 3 +++ target/s390x/gen-features.c | 7 ++++++- 2 files changed, 9 insertions(+), 1 deletion(-)
hw/s390x/s390-virtio-ccw.c +3 -0
@@ -803,7 +803,10 @@ DEFINE_CCW_MACHINE(7_1, "7.1", true);static void ccw_machine_7_0_instance_options (MachineState *machine) { +static const S390FeatInit qemu_cpu_feat = {S390_FEAT_LIST_QEMU_V7_0 }; +ccw_machine_7_1_instance_options (machine); +s390_set_qemu_cpu_model (0x8561 ,15 ,1 , qemu_cpu_feat); }static void ccw_machine_7_0_class_options (MachineClass *mc)
target/s390x/gen-features.c +6 -1
@@ -738,13 +738,17 @@ static uint16_t qemu_V6_2[] = { S390_FEAT_VECTOR_ENH, }; +static uint16_t qemu_V7_0[] = { +S390_FEAT_MISC_INSTRUCTION_EXT3 , +}; +/* * Features for the "qemu" CPU model of the latest QEMU machine and the "max" * CPU model under TCG. Don't include features that are not part of the full * feature set of the current "max" CPU model generation. */ static uint16_t qemu_MAX[] = { -S390_FEAT_MISC_INSTRUCTION_EXT3 , +S390_FEAT_VECTOR_ENH2 , };/****** END FEATURE DEFS ******/
@@ -866,6 +870,7 @@ static FeatGroupDefSpec QemuFeatDef[] = {QEMU_FEAT_INITIALIZER (V4_1 ),QEMU_FEAT_INITIALIZER (V6_0 ),QEMU_FEAT_INITIALIZER (V6_2 ), +QEMU_FEAT_INITIALIZER (V7_0 ), QEMU_FEAT_INITIALIZER(MAX), };
Signature · 1 line
2.35.1
PATCH v6 13/13 tests/tcg/s390x: Tests for Vector Enhancements Facility 2
In reply to David HildenbrandFrom: David Miller <[email protected]>
--- tests/tcg/s390x/Makefile.target | 8 ++ tests/tcg/s390x/vx.h | 19 +++++ tests/tcg/s390x/vxeh2_vcvt.c | 88 ++++++++++++++++++++ tests/tcg/s390x/vxeh2_vlstr.c | 139 ++++++++++++++++++++++++++++++++ tests/tcg/s390x/vxeh2_vs.c | 93 +++++++++++++++++++++ 5 files changed, 347 insertions(+) create mode 100644 tests/tcg/s390x/vx.h create mode 100644 tests/tcg/s390x/vxeh2_vcvt.c create mode 100644 tests/tcg/s390x/vxeh2_vlstr.c create mode 100644 tests/tcg/s390x/vxeh2_vs.c
tests/tcg/s390x/Makefile.target +8 -0
@@ -17,6 +17,14 @@ TESTS+=trap TESTS+=signals-s390x TESTS+=branch-relative-long +VECTOR_TESTS=vxeh2_vs +VECTOR_TESTS+=vxeh2_vcvt +VECTOR_TESTS+=vxeh2_vlstr + +TESTS+=$(VECTOR_TESTS) + +$(VECTOR_TESTS): CFLAGS+=-march=z15 -O2 + ifneq ($(HAVE_GDB_BIN),) GDB_SCRIPT=$(SRC_PATH)/tests/guest-debug/run-test.py
tests/tcg/s390x/vx.h new file +19 -0
@@ -0,0 +1,19 @@ +#ifndef QEMU_TESTS_S390X_VX_H +#define QEMU_TESTS_S390X_VX_H + +typedef union S390Vector { +uint64_t d[2 ];/* doubleword */ +uint32_t w[4 ];/* word */ +uint16_t h[8 ];/* halfword */ +uint8_t b[16 ];/* byte */ +float f[4 ];/* float32 */ +double fd[2 ];/* float64 */ +__uint128_t v; +}S390Vector ; + +#define ES8 0 +#define ES16 1 +#define ES32 2 +#define ES64 3 + +#endif /* QEMU_TESTS_S390X_VX_H */
tests/tcg/s390x/vxeh2_vcvt.c new file +88 -0
88 lines not on this page · raw message
tests/tcg/s390x/vxeh2_vlstr.c new file +139 -0
139 lines not on this page · raw message
tests/tcg/s390x/vxeh2_vs.c new file +93 -0
93 lines not on this page · raw message
Signature · 1 line
2.35.1
Re: [PATCH v6 13/13] tests/tcg/s390x: Tests for Vector Enhancements Facility 2
Quoted 23 lines · David Hildenbrand
From: David Miller <[email protected]>
Signed-off-by David Miller <[email protected]>Signed-off-by Richard Henderson <[email protected]>Tested-by Thomas Huth <[email protected]>Signed-off-by David Hildenbrand <[email protected]>--- tests/tcg/s390x/Makefile.target | 8 ++ tests/tcg/s390x/vx.h | 19 +++++ tests/tcg/s390x/vxeh2_vcvt.c | 88 ++++++++++++++++++++ tests/tcg/s390x/vxeh2_vlstr.c | 139 ++++++++++++++++++++++++++++++++ tests/tcg/s390x/vxeh2_vs.c | 93 +++++++++++++++++++++ 5 files changed, 347 insertions(+)create mode 100644
tests/tcg/s390x/vx.hcreate mode 100644tests/tcg/s390x/vxeh2_vcvt.ccreate mode 100644tests/tcg/s390x/vxeh2_vlstr.ccreate mode 100644tests/tcg/s390x/vxeh2_vs.ctests/tcg/s390x/Makefile.target +0
-0
tests/tcg/s390x/Makefile.target TESTS+=trap · line 26
Quoted 10 lines
tests/tcg/s390x/Makefile.target +7
-0@@ -17,6 +17,14 @@ TESTS+=trap TESTS+=signals-s390x TESTS+=branch-relative-long +VECTOR_TESTS=vxeh2_vs +VECTOR_TESTS+=vxeh2_vcvt +VECTOR_TESTS+=vxeh2_vlstr + +TESTS+=$(VECTOR_TESTS) + +$(VECTOR_TESTS): CFLAGS+=-march=z15 -O2
I'm sorry, but this still fails in the QEMU CI:
https://gitlab.com/thuth/qemu/-/jobs/2401500348
s390x-linux-gnu-gcc: error: unrecognized argument in option '-march=z15'
I think we either have to switch to manually encoded instructions again, or add a check to the Makefile and only add the tests if the compiler supports -march=z15 ...? Opinions? Preferences?
Thomas
Re: [PATCH v6 13/13] tests/tcg/s390x: Tests for Vector Enhancements Facility 2
Quoted 36 lines · Thomas Huth
Quoted 25 lines · David Hildenbrand
From: David Miller <[email protected]>
Signed-off-by David Miller <[email protected]>Signed-off-by Richard Henderson <[email protected]>Tested-by Thomas Huth <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
tests/tcg/s390x/Makefile.target | 8 ++ tests/tcg/s390x/vx.h | 19 +++++ tests/tcg/s390x/vxeh2_vcvt.c | 88 ++++++++++++++++++++ tests/tcg/s390x/vxeh2_vlstr.c | 139 ++++++++++++++++++++++++++++++++ tests/tcg/s390x/vxeh2_vs.c | 93 +++++++++++++++++++++5 files changed, 347 insertions(+)create mode 100644
tests/tcg/s390x/vx.hcreate mode 100644tests/tcg/s390x/vxeh2_vcvt.ccreate mode 100644tests/tcg/s390x/vxeh2_vlstr.ccreate mode 100644tests/tcg/s390x/vxeh2_vs.ctests/tcg/s390x/Makefile.target +0
-0@@ -17,6 +17,14 @@ TESTS+=trapTESTS+=signals-s390x
TESTS+=branch-relative-longtests/tcg/s390x/Makefile.target
Quoted 7 lines
tests/tcg/s390x/Makefile.target +7
-0+VECTOR_TESTS=vxeh2_vs +VECTOR_TESTS+=vxeh2_vcvt +VECTOR_TESTS+=vxeh2_vlstr + +TESTS+=$(VECTOR_TESTS) + +$(VECTOR_TESTS): CFLAGS+=-march=z15 -O2I'm sorry, but this still fails in the QEMU CI:
Quoted 6 lines
https://gitlab.com/thuth/qemu/-/jobs/2401500348
s390x-linux-gnu-gcc: error: unrecognized argument in option '-march=z15'
I think we either have to switch to manually encoded instructions again, or add a check to the
Makefileand only add the tests if the compiler supports -march=z15 ...? Opinions? Preferences?
I just tried, and seems like something like this should do the job, I think:
tests/tcg/s390x/Makefile.target +3 -3
@@ -20,11 +20,11 @@ TESTS+=branch-relative-long VECTOR_TESTS=vxeh2_vs VECTOR_TESTS+=vxeh2_vcvt VECTOR_TESTS+=vxeh2_vlstr - -TESTS+=$(VECTOR_TESTS) - $(VECTOR_TESTS): CFLAGS+=-march=z15 -O2 +TESTS+=$(if $(shell $(CC) -march=z15 -S -o /dev/null -xc /dev/null \ + >/dev/null 2>&1 && echo OK),$(VECTOR_TESTS)) + ifneq ($(HAVE_GDB_BIN),) GDB_SCRIPT=$(SRC_PATH)/tests/guest-debug/run-test.py
Does that look reasonable?
Thomas
Re: [PATCH v6 13/13] tests/tcg/s390x: Tests for Vector Enhancements Facility 2
In reply to David HildenbrandQuoted 7 lines · David Hildenbrand
From: David Miller <[email protected]>
Signed-off-by David Miller <[email protected]>Signed-off-by Richard Henderson <[email protected]>Tested-by Thomas Huth <[email protected]>Signed-off-by David Hildenbrand <[email protected]>---
[...]
Quoted 15 lines
tests/tcg/s390x/vxeh2_vlstr.c new file +9
-0@@ -0,0 +1,139 @@ +/* +* vxeh2_vlstr: vector-enhancements facility 2 vector load/store reversed * +*/ +#include <stdint.h> +#include "vx.h" + +#define vtst (v1, v2) \ + if (v1.d[0] != v2.d[0] || v1.d[1] != v2.d[1]) { \ + return 1; \
tests/tcg/s390x/vxeh2_vlstr.c line 19
Quoted 10 lines
tests/tcg/s390x/vxeh2_vlstr.c new file +10
-0+ } + +static inline void vler (S390Vector *v1,const void *va,uint8_t m3) +{ + asmvolatile ("vler %[v1], 0(%[va]), %[m3]\n" + : [v1]"+v" (v1->v) + : [va]"d" (va) + , [m3]"i" (m3) + :"memory" ); +}
The vxeh2_vlstr test fails when compiling with Clang instead of GCC ... seems like it enjoys using register r0 in the spots that use the "d" constraints in the inline assembly in here. The fix is easy:
tests/tcg/s390x/vxeh2_vlstr.c +8 -8
@@ -13,7 +13,7 @@ static inline void vler(S390Vector *v1, const void *va, uint8_t m3) { asmvolatile ("vler %[v1], 0(%[va]), %[m3]\n" : [v1]"+v" (v1->v) - : [va]"d" (va) + : [va]"a" (va) , [m3]"i" (m3) :"memory" ); }
@@ -21,7 +21,7 @@ static inline void vler(S390Vector *v1, const void *va, uint8_t m3)static inline void vster (S390Vector *v1,const void *va,uint8_t m3) { asmvolatile ("vster %[v1], 0(%[va]), %[m3]\n" - : [va]"+d" (va) + : [va]"+a" (va) : [v1]"v" (v1->v) , [m3]"i" (m3) :"memory" );
@@ -31,7 +31,7 @@ static inline void vlbr(S390Vector *v1, void *va, const uint8_t m3) { asmvolatile ("vlbr %[v1], 0(%[va]), %[m3]\n" : [v1]"+v" (v1->v) - : [va]"d" (va) + : [va]"a" (va) , [m3]"i" (m3) :"memory" ); }
@@ -39,7 +39,7 @@ static inline void vlbr(S390Vector *v1, void *va, const uint8_t m3)static inline void vstbr (S390Vector *v1,void *va,const uint8_t m3) { asmvolatile ("vstbr %[v1], 0(%[va]), %[m3]\n" - : [va]"+d" (va) + : [va]"+a" (va) : [v1]"v" (v1->v) , [m3]"i" (m3) :"memory" );
@@ -50,7 +50,7 @@ static inline void vlebrh(S390Vector *v1, void *va, const uint8_t m3) { asmvolatile ("vlebrh %[v1], 0(%[va]), %[m3]\n" : [v1]"+v" (v1->v) - : [va]"d" (va) + : [va]"a" (va) , [m3]"i" (m3) :"memory" ); }
@@ -58,7 +58,7 @@ static inline void vlebrh(S390Vector *v1, void *va, const uint8_t m3)static inline void vstebrh (S390Vector *v1,void *va,const uint8_t m3) { asmvolatile ("vstebrh %[v1], 0(%[va]), %[m3]\n" - : [va]"+d" (va) + : [va]"+a" (va) : [v1]"v" (v1->v) , [m3]"i" (m3) :"memory" );
@@ -68,7 +68,7 @@ static inline void vllebrz(S390Vector *v1, void *va, const uint8_t m3) { asmvolatile ("vllebrz %[v1], 0(%[va]), %[m3]\n" : [v1]"+v" (v1->v) - : [va]"d" (va) + : [va]"a" (va) , [m3]"i" (m3) :"memory" ); }
@@ -77,7 +77,7 @@ static inline void vlbrrep(S390Vector *v1, void *va, const uint8_t m3) { asmvolatile ("vlbrrep %[v1], 0(%[va]), %[m3]\n" : [v1]"+v" (v1->v) - : [va]"d" (va) + : [va]"a" (va) , [m3]"i" (m3) :"memory" ); }
I'll fix it up in my queue, so no need to resend.
Thomas
Quoted 17 lines · David Hildenbrand
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER) VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR) VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG) VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ)VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)
Quoted 8 lines
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)
Thanks, queued to my s390x-next branch now:
https://gitlab.com/thuth/qemu/-/commits/s390x-next/
Thomas
Quoted 27 lines · Thomas Huth
Quoted 17 lines · David Hildenbrand
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER) VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR) VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG) VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ)VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)Quoted 8 lines
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)
Quoted 4 lines
Thanks, queued to my s390x-next branch now:
Thanks for fixing up. At this point I would have suggested to exclude the tests for now.
Signature · 3 lines
Thanks, David / dhildenb
There was also the patch that had them as .insn in the other series of emails.
Quoted 41 lines · David Hildenbrand
Quoted 27 lines · Thomas Huth
Quoted 17 lines · David Hildenbrand
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER) VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR) VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG) VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ)VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)Quoted 8 lines
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)Quoted 4 lines
Thanks, queued to my s390x-next branch now:
Thanks for fixing up. At this point I would have suggested to exclude the tests for now.
--
Thanks,David / dhildenb
Hi!
Quoted 1 line · David Miller
There was also the patch that had them as .insn in the other series of emails.
Sorry, I missed that patch, could you please point me to the mail on https://lore.kernel.org/qemu-devel/ ? I remember that there was a discussion about the vri-d encoding, but I apparently missed the patch that came out of this discussion...
Thomas
Quoted 43 lines · David Hildenbrand
Quoted 41 lines · Thomas Huth
Quoted 27 lines · David Hildenbrand
Quoted 17 lines
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER) VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR) VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG) VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ)VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)Quoted 8 lines
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)Quoted 4 lines
Thanks, queued to my s390x-next branch now:
https://gitlab.com/thuth/qemu/-/commits/s390x-next/Thanks for fixing up. At this point I would have suggested to exclude the tests for now.
--
Thanks,David / dhildenb
Sorry, It was in the discussion for v4 patches, as an attachment . mail thread:
[PATCH v4 10/11] tests/tcg/s390x: Tests for Vector Enhancements Facility 2 So it likely never made it to the mailing list.
I've reattached and will forward the patch (by itself) to the mailing list.
I think the other solution works just as well by ignoring if compiler doesn't support z15.
I just thought I'd bring it back up as I saw discussion about it.
Thanks
- David Miller
Quoted 4 lines · Thomas Huth
Hi!
On 02/05/2022 18.06, David Miller wrote:
Quoted 53 lines
Quoted 1 line
There was also the patch that had them as .insn in the other series of emails.
Sorry, I missed that patch, could you please point me to the mail on
https://lore.kernel.org/qemu-devel/ ? I remember that there was a discussion about the vri-d encoding, but I apparently missed the patch that came out of this discussion...Thomas
Quoted 43 lines · David Hildenbrand
Quoted 41 lines · Thomas Huth
Quoted 27 lines · David Hildenbrand
Quoted 17 lines
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER) VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR) VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG) VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ)VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)Quoted 8 lines
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)Quoted 4 lines
Thanks, queued to my s390x-next branch now:
https://gitlab.com/thuth/qemu/-/commits/s390x-next/Thanks for fixing up. At this point I would have suggested to exclude the tests for now.
--
Thanks,David / dhildenb
From bb6bf2f9529c4d76db9a9eff2ff7fa1235657103 Mon Sep 17 00:00:00 2001
From: David Miller <[email protected]>
Date: Mon, 21 Mar 2022 16:58:57 -0400
Subject: [PATCH v5 10/11] tests/tcg/s390x: Tests for Vector Enhancements
Facility 2
--- tests/tcg/s390x/Makefile.target | 8 ++ tests/tcg/s390x/vx.h | 19 +++++ tests/tcg/s390x/vxeh2_vcvt.c | 88 ++++++++++++++++++++ tests/tcg/s390x/vxeh2_vlstr.c | 139 ++++++++++++++++++++++++++++++++ tests/tcg/s390x/vxeh2_vs.c | 95 ++++++++++++++++++++++ 5 files changed, 349 insertions(+) create mode 100644 tests/tcg/s390x/vx.h create mode 100644 tests/tcg/s390x/vxeh2_vcvt.c create mode 100644 tests/tcg/s390x/vxeh2_vlstr.c create mode 100644 tests/tcg/s390x/vxeh2_vs.c
tests/tcg/s390x/Makefile.target +8 -0
@@ -16,6 +16,14 @@ TESTS+=shift TESTS+=trap TESTS+=signals-s390x +VECTOR_TESTS=vxeh2_vs +VECTOR_TESTS+=vxeh2_vcvt +VECTOR_TESTS+=vxeh2_vlstr + +TESTS+=$(VECTOR_TESTS) + +$(VECTOR_TESTS): CFLAGS+=-march=z15 -O2 + ifneq ($(HAVE_GDB_BIN),) GDB_SCRIPT=$(SRC_PATH)/tests/guest-debug/run-test.py
tests/tcg/s390x/vx.h new file +19 -0
@@ -0,0 +1,19 @@ +#ifndef QEMU_TESTS_S390X_VX_H +#define QEMU_TESTS_S390X_VX_H + +typedef union S390Vector { +uint64_t d[2 ];/* doubleword */ +uint32_t w[4 ];/* word */ +uint16_t h[8 ];/* halfword */ +uint8_t b[16 ];/* byte */ +float f[4 ];/* float32 */ +double fd[2 ];/* float64 */ +__uint128_t v; +}S390Vector ; + +#define ES8 0 +#define ES16 1 +#define ES32 2 +#define ES64 3 + +#endif
tests/tcg/s390x/vxeh2_vcvt.c new file +88 -0
88 lines not on this page · raw message
tests/tcg/s390x/vxeh2_vlstr.c new file +139 -0
139 lines not on this page · raw message
tests/tcg/s390x/vxeh2_vs.c new file +95 -0
95 lines not on this page · raw message
Signature · 1 line
2.32.0
It looks like google killed allowing password access early, nothing makes it work anymore.
They had plans to disable 'less secure app' in may, but it thought it was the end of the month.
I'll try copy/paste as plain text as well though I Know it will likely screw it up..
Quoted 79 lines · David Miller
Sorry, It was in the discussion for v4 patches, as an attachment . mail thread:
[PATCH v4 10/11] tests/tcg/s390x: Tests for Vector Enhancements Facility 2 So it likely never made it to the mailing list.I've reattached and will forward the patch (by itself) to the mailing list.
I think the other solution works just as well by ignoring if compiler doesn't support z15.
I just thought I'd bring it back up as I saw discussion about it.
Thanks
- David MillerQuoted 57 lines · Thomas Huth
Hi!
Quoted 1 line · David Miller
There was also the patch that had them as .insn in the other series of emails.
Sorry, I missed that patch, could you please point me to the mail on
https://lore.kernel.org/qemu-devel/ ? I remember that there was a discussion about the vri-d encoding, but I apparently missed the patch that came out of this discussion...Thomas
Quoted 43 lines · David Hildenbrand
Quoted 41 lines · Thomas Huth
Quoted 27 lines · David Hildenbrand
Quoted 17 lines
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER) VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR) VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG) VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ)VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)Quoted 8 lines
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)Quoted 4 lines
Thanks, queued to my s390x-next branch now:
https://gitlab.com/thuth/qemu/-/commits/s390x-next/Thanks for fixing up. At this point I would have suggested to exclude the tests for now.
--
Thanks,David / dhildenb
From bb6bf2f9529c4d76db9a9eff2ff7fa1235657103 Mon Sep 17 00:00:00 2001
From: David Miller <[email protected]>
Date: Mon, 21 Mar 2022 16:58:57 -0400
Subject: [PATCH v5 10/11] tests/tcg/s390x: Tests for Vector Enhancements
Facility 2
--- tests/tcg/s390x/Makefile.target | 8 ++ tests/tcg/s390x/vx.h | 19 +++++ tests/tcg/s390x/vxeh2_vcvt.c | 88 ++++++++++++++++++++ tests/tcg/s390x/vxeh2_vlstr.c | 139 ++++++++++++++++++++++++++++++++ tests/tcg/s390x/vxeh2_vs.c | 95 ++++++++++++++++++++++ 5 files changed, 349 insertions(+) create mode 100644 tests/tcg/s390x/vx.h create mode 100644 tests/tcg/s390x/vxeh2_vcvt.c create mode 100644 tests/tcg/s390x/vxeh2_vlstr.c create mode 100644 tests/tcg/s390x/vxeh2_vs.c
tests/tcg/s390x/Makefile.target +8 -0
@@ -16,6 +16,14 @@ TESTS+=shift TESTS+=trap TESTS+=signals-s390x +VECTOR_TESTS=vxeh2_vs +VECTOR_TESTS+=vxeh2_vcvt +VECTOR_TESTS+=vxeh2_vlstr + +TESTS+=$(VECTOR_TESTS) + +$(VECTOR_TESTS): CFLAGS+=-march=z15 -O2 + ifneq ($(HAVE_GDB_BIN),) GDB_SCRIPT=$(SRC_PATH)/tests/guest-debug/run-test.py
tests/tcg/s390x/vx.h new file +19 -0
@@ -0,0 +1,19 @@ +#ifndef QEMU_TESTS_S390X_VX_H +#define QEMU_TESTS_S390X_VX_H + +typedef union S390Vector { +uint64_t d[2 ];/* doubleword */ +uint32_t w[4 ];/* word */ +uint16_t h[8 ];/* halfword */ +uint8_t b[16 ];/* byte */ +float f[4 ];/* float32 */ +double fd[2 ];/* float64 */ +__uint128_t v; +}S390Vector ; + +#define ES8 0 +#define ES16 1 +#define ES32 2 +#define ES64 3 + +#endif
tests/tcg/s390x/vxeh2_vcvt.c new file +14 -0
@@ -0,0 +1,88 @@ +/* +* vxeh2_vcvt: vector-enhancements facility 2 vector convert * +*/ +#include <stdint.h> +#include "vx.h" + +#define M_S 8 +#define M4_XxC 4 +#define M4_def M4_XxC + +static inline void vcfps (S390Vector *v1,S390Vector *v2, +const uint8_t m3,const uint8_t m4,const uint8_t m5) +{ + asm volatile(".insn vrr, 0xE700000000C3, %[v1], %[v2], 0, %[m3],
%[m4], %[m5]\n"
+ : [v1] "=v" (v1->v) + : [v2] "v" (v2->v) + , [m3] "i" (m3) + , [m4] "i" (m4) + , [m5] "i" (m5));
+}
+
+static inline void vcfpl(S390Vector *v1, S390Vector *v2,
+ const uint8_t m3, const uint8_t m4, const uint8_t m5)
+{
+ asm volatile(".insn vrr, 0xE700000000C1, %[v1], %[v2], 0, %[m3],
%[m4], %[m5]\n"
+ : [v1] "=v" (v1->v) + : [v2] "v" (v2->v) + , [m3] "i" (m3) + , [m4] "i" (m4) + , [m5] "i" (m5));
+}
+
+static inline void vcsfp(S390Vector *v1, S390Vector *v2,
+ const uint8_t m3, const uint8_t m4, const uint8_t m5)
+{
+ asm volatile(".insn vrr, 0xE700000000C2, %[v1], %[v2], 0, %[m3],
%[m4], %[m5]\n"
+ : [v1] "=v" (v1->v) + : [v2] "v" (v2->v) + , [m3] "i" (m3) + , [m4] "i" (m4) + , [m5] "i" (m5));
+}
+
+static inline void vclfp(S390Vector *v1, S390Vector *v2,
+ const uint8_t m3, const uint8_t m4, const uint8_t m5)
+{
+ asm volatile(".insn vrr, 0xE700000000C0, %[v1], %[v2], 0, %[m3],
%[m4], %[m5]\n"
+ : [v1] "=v" (v1->v) + : [v2] "v" (v2->v) + , [m3] "i" (m3) + , [m4] "i" (m4) + , [m5] "i" (m5));
+}
+
+int main(int argc, char *argv[])
+{
+ S390Vector vd;
+ S390Vector vs_i32 = { .w[0] = 1, .w[1] = 64, .w[2] = 1024, .w[3] = -10 };
+ S390Vector vs_u32 = { .w[0] = 2, .w[1] = 32, .w[2] = 4096, .w[3] = 8888 };
+ S390Vector vs_f32 = { .f[0] = 3.987, .f[1] = 5.123,
+ .f[2] = 4.499, .f[3] = 0.512 };
+
+ vd.d[0] = vd.d[1] = 0;
+ vcfps(&vd, &vs_i32, 2, M4_def, 0);
+ if (1 != vd.f[0] || 1024 != vd.f[2] || 64 != vd.f[1] || -10 != vd.f[3]) {
+ return 1;
+ }
+
+ vd.d[0] = vd.d[1] = 0;
+ vcfpl(&vd, &vs_u32, 2, M4_def, 0);
+ if (2 != vd.f[0] || 4096 != vd.f[2] || 32 != vd.f[1] || 8888 != vd.f[3]) {
+ return 1;
+ }
+
+ vd.d[0] = vd.d[1] = 0;
+ vcsfp(&vd, &vs_f32, 2, M4_def, 0);
+ if (4 != vd.w[0] || 4 != vd.w[2] || 5 != vd.w[1] || 1 != vd.w[3]) {
+ return 1;
+ }
+
+ vd.d[0] = vd.d[1] = 0;
+ vclfp(&vd, &vs_f32, 2, M4_def, 0);
+ if (4 != vd.w[0] || 4 != vd.w[2] || 5 != vd.w[1] || 1 != vd.w[3]) {+ return 1;
+ }
+
+ return 0;
+}
tests/tcg/s390x/vxeh2_vlstr.c new file +139 -0
139 lines not on this page · raw message
tests/tcg/s390x/vxeh2_vs.c new file +95 -0
95 lines not on this page · raw message
Signature · 1 line
2.32.0
Quoted 88 lines · David Miller
It looks like google killed allowing password access early, nothing makes it work anymore.
They had plans to disable 'less secure app' in may, but it thought it was the end of the month.
I'll try copy/paste as plain text as well though I Know it will likely screw it up..Quoted 79 lines · David Miller
Sorry, It was in the discussion for v4 patches, as an attachment . mail thread:
[PATCH v4 10/11] tests/tcg/s390x: Tests for Vector Enhancements Facility 2 So it likely never made it to the mailing list.I've reattached and will forward the patch (by itself) to the mailing list.
I think the other solution works just as well by ignoring if compiler doesn't support z15.
I just thought I'd bring it back up as I saw discussion about it.
Thanks
- David MillerQuoted 57 lines · Thomas Huth
Hi!
Quoted 1 line · David Miller
There was also the patch that had them as .insn in the other series of emails.
Sorry, I missed that patch, could you please point me to the mail on
https://lore.kernel.org/qemu-devel/ ? I remember that there was a discussion about the vri-d encoding, but I apparently missed the patch that came out of this discussion...Thomas
Quoted 43 lines · David Hildenbrand
Quoted 41 lines · Thomas Huth
Quoted 27 lines · David Hildenbrand
Quoted 17 lines
Implement Vector-Enhancements Facility 2 for s390x
resolves: https://gitlab.com/qemu-project/qemu/-/issues/738
implements:
VECTOR LOAD ELEMENTS REVERSED (VLER) VECTOR LOAD BYTE REVERSED ELEMENTS (VLBR) VECTOR LOAD BYTE REVERSED ELEMENT (VLEBRH, VLEBRF, VLEBRG) VECTOR LOAD BYTE REVERSED ELEMENT AND ZERO (VLLEBRZ)VECTOR LOAD BYTE REVERSED ELEMENT AND REPLICATE (VLBRREP)
VECTOR STORE ELEMENTS REVERSED (VSTER)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTBR)
VECTOR STORE BYTE REVERSED ELEMENTS (VSTEBRH, VSTEBRF, VSTEBRG)
VECTOR SHIFT LEFT DOUBLE BY BIT (VSLD)
VECTOR SHIFT RIGHT DOUBLE BY BIT (VSRD)
VECTOR STRING SEARCH (VSTRS)Quoted 8 lines
modifies:
VECTOR FP CONVERT FROM FIXED (VCFPS)
VECTOR FP CONVERT FROM LOGICAL (VCFPL)
VECTOR FP CONVERT TO FIXED (VCSFP)
VECTOR FP CONVERT TO LOGICAL (VCLFP)
VECTOR SHIFT LEFT (VSL)
VECTOR SHIFT RIGHT ARITHMETIC (VSRA)
VECTOR SHIFT RIGHT LOGICAL (VSRL)Quoted 4 lines
Thanks, queued to my s390x-next branch now:
https://gitlab.com/thuth/qemu/-/commits/s390x-next/Thanks for fixing up. At this point I would have suggested to exclude the tests for now.
--
Thanks,David / dhildenb
Quoted 2 lines · David Miller
It looks like google killed allowing password access early, nothing makes it work anymore.
Uh, that's ugly! I hope you'll figure out a way to work-around that problem!
Quoted 4 lines
They had plans to disable 'less secure app' in may, but it thought it was the end of the month.
I'll try copy/paste as plain text as well though I Know it will likely screw it up..
Yup, that plain text patch didn't apply anymore - so I went for the attachment from your previous mail this time instead (hoping that you'll find another way for using git-send-email again in the future).
Thomas
Quoted 11 lines · David Miller
Sorry, It was in the discussion for v4 patches, as an attachment . mail thread:
[PATCH v4 10/11] tests/tcg/s390x: Tests for Vector Enhancements Facility 2 So it likely never made it to the mailing list.I've reattached and will forward the patch (by itself) to the mailing list.
I think the other solution works just as well by ignoring if compiler doesn't support z15.
I just thought I'd bring it back up as I saw discussion about it.
Ok, I now gave this a try ... and while this should now work fine with older versions of gcc/binutils, it's failing with Clang now:
BUILD s390x-linux-user guest-tests/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vs.c:14:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE70000000074, %[v1], %[v2], %[v3], 0,0,0\n"
^
/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vs.c:22:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE7000000007E, %[v1], %[v2], %[v3], 0,0,0\n"
^
/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vs.c:30:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE7000000007C, %[v1], %[v2], %[v3], 0,0,0\n"
^
/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vs.c:40:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE70000000086, %[v1], %[v2], %[v3], 0, %[I], 0\n"
^
/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vs.c:51:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE70000000087, %[v1], %[v2], %[v3], 0, %[I], 0\n"
^5 errors generated.
make[1]: *** [../Makefile.target:109: vxeh2_vs] Error 1
make[1]: *** Waiting for unfinished jobs.... /home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vcvt.c:14:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE700000000C3, %[v1], %[v2], 0, %[m3], %[m4],
%[m5]\n"
^
/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vcvt.c:25:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE700000000C1, %[v1], %[v2], 0, %[m3], %[m4],
%[m5]\n"
^
/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vcvt.c:36:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE700000000C2, %[v1], %[v2], 0, %[m3], %[m4],
%[m5]\n"
^
/home/thuth/devel/qemu/tests/tcg/s390x/vxeh2_vcvt.c:47:18: error: couldn't allocate output register for constraint 'v'
asm volatile(".insn vrr, 0xE700000000C0, %[v1], %[v2], 0, %[m3], %[m4],
%[m5]\n"
^4 errors generated.
...
Thus I think I'll rather go with the other approach instead that checks for the availability of -march=z15.
Thomas