Skip to contentSkip to conversation

powerpc/64: fast interrupt exits

PATCH In review Nicholas Piggin 3 participants 20 messages Jun 17, 2021 – May 16, 2022

Subsystems (4)
lore
Share & tools
lore

Nicholas Piggin

author reviewer
#0 lore raw

v4 cover powerpc/64: fast interrupt exits

This series attempts to improve the speed of interrupts and system calls in three major ways.

Firstly, the SRR/HSRR registers do not need to be reloaded if they were clobbered for the duration of the interrupt and the return NIP and MSR did not changed. 64e does not implement this part, but it could quite easily.

Secondly, mtmsrd and mtspr are reduced by various means, this is mostly specific to 64s.

Lastly, an alternate interrupt return location facility is added for soft-masked asynchronous interrupts, and then that's used to set everything up for return without having to disable MSR RI or EE.

After this series, the entire system call / interrupt handler fast path can execute with no mtsprs and one mtmsrd to enable interrupts initially, and the system call vectored path doesn't even need to do that. This gives a decent performance benefit. On POWER9 with a powernv_defconfig without VIRT_CPU_ACCOUNTING_NATIVE, no meltdown workarounds, gettid sc system call goes from 481 -> 344 cycles, gettid scv 345->299 cycles, and page fault 1225->1064 cycles.

Since v3:
- Added Christophe's series to the end (with one broken out patch at the start and minor tweaks).
- Fix lbz of SOFTE that broke on BE, should be loaded with ld.
- Fix case where fast_interrupt_return was not marking SRRs as clobbered correctly, causing random userspace segfaults etc.
- Fix cpu lock inversion problem in the workaround static branch switching code.
- Removed the .L local name from some labels for the restart handlers, which makes the disassembly easier to read.
- Fixed several nitpicks found by kbuild robot (duplicate include, missing prototypes, etc).
- Fix 64e compile bug [from mpe].
- Fix KUAP re-locking in some restart cases.

Christophe Leroy (6):
powerpc/interrupt: Fix CONFIG ifdef typo powerpc/interrupt: Rename and lightly change
syscall_exit_prepare_main()
powerpc/interrupt: Refactor interrupt_exit_user_prepare() powerpc/interrupt: Interchange
prep_irq_for_{kernel_enabled/user}_exit()
powerpc/interrupt: Refactor prep_irq_for_{user/kernel_enabled}_exit() powerpc/interrupt: Remove prep_irq_for_user_exit()

Nicholas Piggin (11):
powerpc: remove interrupt exit helpers unused argument powerpc/64s: introduce different functions to return from SRR vs HSRR
interrupts
powerpc/64s: avoid reloading (H)SRR registers if they are still valid powerpc/64: handle MSR EE and RI in interrupt entry wrapper powerpc/64: move interrupt return asm to interrupt_64.S powerpc/64s: system call avoid setting MSR[RI] until we set MSR[EE] powerpc/64s: save one more register in the masked interrupt handler powerpc/64: allow alternate return locations for soft-masked
interrupts
powerpc/64: interrupt soft-enable race fix
powerpc/64: treat low kernel text as irqs soft-masked powerpc/64: use interrupt restart table to speed up return from
interrupt

 arch/powerpc/Kconfig.debug                 |   5 +
 arch/powerpc/include/asm/asm-prototypes.h  |   9 +-
 arch/powerpc/include/asm/head-64.h         |   2 +-
 arch/powerpc/include/asm/hw_irq.h          |  23 +-
 arch/powerpc/include/asm/interrupt.h       |  60 +-
 arch/powerpc/include/asm/paca.h            |   7 +
 arch/powerpc/include/asm/ppc_asm.h         |   8 +
 arch/powerpc/include/asm/ptrace.h          |  75 ++-
 arch/powerpc/kernel/asm-offsets.c          |   7 +
 arch/powerpc/kernel/entry_64.S             | 516 --------------
 arch/powerpc/kernel/exceptions-64e.S       |  53 +-
 arch/powerpc/kernel/exceptions-64s.S       | 219 +++---
 arch/powerpc/kernel/fpu.S                  |   4 +
 arch/powerpc/kernel/head_64.S              |   5 +-
 arch/powerpc/kernel/interrupt.c            | 413 +++++++-----
 arch/powerpc/kernel/interrupt_64.S         | 742 +++++++++++++++++++++
 arch/powerpc/kernel/irq.c                  |  95 +++
 arch/powerpc/kernel/kgdb.c                 |   2 +-
 arch/powerpc/kernel/kprobes-ftrace.c       |   2 +-
 arch/powerpc/kernel/kprobes.c              |  10 +-
 arch/powerpc/kernel/process.c              |  26 +-
 arch/powerpc/kernel/prom_init.c            |   3 +
 arch/powerpc/kernel/rtas.c                 |  14 +-
 arch/powerpc/kernel/signal.c               |   2 +-
 arch/powerpc/kernel/signal_64.c            |   9 +
 arch/powerpc/kernel/syscalls.c             |   3 +-
 arch/powerpc/kernel/traps.c                |  18 +-
 arch/powerpc/kernel/vector.S               |   8 +-
 arch/powerpc/kernel/vmlinux.lds.S          |  10 +
 arch/powerpc/kvm/book3s_hv.c               |   3 +
 arch/powerpc/kvm/book3s_pr.c               |   2 +
 arch/powerpc/lib/Makefile                  |   2 +-
 arch/powerpc/lib/feature-fixups.c          |  52 +-
 arch/powerpc/lib/restart_table.c           |  30 +
 arch/powerpc/lib/sstep.c                   |   4 +-
 arch/powerpc/math-emu/math.c               |   2 +-
 arch/powerpc/platforms/powernv/opal-call.c |   4 +
 arch/powerpc/platforms/pseries/hvCall.S    |  29 +
 arch/powerpc/sysdev/fsl_pci.c              |   2 +-
 39 files changed, 1622 insertions(+), 858 deletions(-)
 create mode 100644 arch/powerpc/kernel/interrupt_64.S
 create mode 100644 arch/powerpc/lib/restart_table.c
Signature · 1 line
2.23.0
author reviewer
#1 lore raw

PATCH v4 1/17 powerpc/interrupt: Fix CONFIG ifdef typo

From: Christophe Leroy <[email protected]>

CONFIG_PPC_BOOK3S should be CONFIG_PPC_BOOK3S_64. restore_math is a no-op for other configurations.

[np: split from another patch]
---
 arch/powerpc/kernel/interrupt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/powerpc/kernel/interrupt.c +1 -1
index e0938ba298f2..f2baeee437f7 100644
@@ -303,7 +303,7 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
 		ti_flags = READ_ONCE(current_thread_info()->flags);
 	}
 
-	if (IS_ENABLED(CONFIG_PPC_BOOK3S) && IS_ENABLED(CONFIG_PPC_FPU)) {
+	if (IS_ENABLED(CONFIG_PPC_BOOK3S_64) && IS_ENABLED(CONFIG_PPC_FPU)) {
 		if (IS_ENABLED(CONFIG_PPC_TRANSACTIONAL_MEM) &&
 				unlikely((ti_flags & _TIF_RESTORE_TM))) {
 			restore_tm_state(regs);
Signature · 1 line
2.23.0
author reviewer
#2 lore raw

PATCH v4 2/17 powerpc: remove interrupt exit helpers unused argument

In reply to Nicholas Piggin

The msr argument is not used, remove it.

---
 arch/powerpc/include/asm/asm-prototypes.h | 4 ++--
 arch/powerpc/kernel/interrupt.c           | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)
arch/powerpc/include/asm/asm-prototypes.h +2 -2
index 1c7b75834e04..95492655462e 100644
@@ -71,8 +71,8 @@ void __init machine_init(u64 dt_ptr);
 #endif
 long system_call_exception(long r3, long r4, long r5, long r6, long r7, long r8, unsigned long r0, struct pt_regs *regs);
 notrace unsigned long syscall_exit_prepare(unsigned long r3, struct pt_regs *regs, long scv);
-notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned long msr);
-notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsigned long msr);
+notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs);
+notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs);
 
 long ppc_fadvise64_64(int fd, int advice, u32 offset_high, u32 offset_low,
 		      u32 len_high, u32 len_low);
arch/powerpc/kernel/interrupt.c +2 -2
index f2baeee437f7..05fa3ae56e25 100644
@@ -352,7 +352,7 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
 	return ret;
 }
 
-notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned long msr)
+notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs)
 {
 	unsigned long ti_flags;
 	unsigned long flags;
@@ -431,7 +431,7 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned
 
 void preempt_schedule_irq(void);
 
-notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsigned long msr)
+notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs)
 {
 	unsigned long flags;
 	unsigned long ret = 0;
Signature · 1 line
2.23.0
author reviewer
#3 lore raw

PATCH v4 3/17 powerpc/64s: introduce different functions to return from SRR vs HSRR interrupts

In reply to Nicholas Piggin

This makes no real difference yet except that HSRR type interrupts will use hrfid to return. This is important for the next patch.

---
 arch/powerpc/kernel/entry_64.S       | 65 +++++++++++++++++-------
 arch/powerpc/kernel/exceptions-64e.S |  4 ++
 arch/powerpc/kernel/exceptions-64s.S | 76 +++++++++++++++-------------
 arch/powerpc/kernel/vector.S         |  2 +-
 4 files changed, 92 insertions(+), 55 deletions(-)
arch/powerpc/kernel/entry_64.S +47 -18
index 03727308d8cc..03a45a88b4b8 100644
@@ -635,51 +635,57 @@ END_FTR_SECTION_IFCLR(CPU_FTR_ARCH_207S)
 	 * touched, no exit work created, then this can be used.
 	 */
 	.balign IFETCH_ALIGN_BYTES
-	.globl fast_interrupt_return
-fast_interrupt_return:
-_ASM_NOKPROBE_SYMBOL(fast_interrupt_return)
+	.globl fast_interrupt_return_srr
+fast_interrupt_return_srr:
+_ASM_NOKPROBE_SYMBOL(fast_interrupt_return_srr)
 	kuap_check_amr r3, r4
 	ld	r5,_MSR(r1)
 	andi.	r0,r5,MSR_PR
 #ifdef CONFIG_PPC_BOOK3S
-	bne	.Lfast_user_interrupt_return_amr
+	bne	.Lfast_user_interrupt_return_amr_srr
 	kuap_kernel_restore r3, r4
 	andi.	r0,r5,MSR_RI
 	li	r3,0 /* 0 return value, no EMULATE_STACK_STORE */
-	bne+	.Lfast_kernel_interrupt_return
+	bne+	.Lfast_kernel_interrupt_return_srr
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	unrecoverable_exception
 	b	. /* should not get here */
 #else
-	bne	.Lfast_user_interrupt_return
-	b	.Lfast_kernel_interrupt_return
+	bne	.Lfast_user_interrupt_return_srr
+	b	.Lfast_kernel_interrupt_return_srr
 #endif
 
+.macro interrupt_return_macro srr
 	.balign IFETCH_ALIGN_BYTES
-	.globl interrupt_return
-interrupt_return:
-_ASM_NOKPROBE_SYMBOL(interrupt_return)
+	.globl interrupt_return_\srr
+interrupt_return_\srr\():
+_ASM_NOKPROBE_SYMBOL(interrupt_return_\srr\())
 	ld	r4,_MSR(r1)
 	andi.	r0,r4,MSR_PR
-	beq	.Lkernel_interrupt_return
+	beq	.Lkernel_interrupt_return_\srr
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	interrupt_exit_user_prepare
 	cmpdi	r3,0
-	bne-	.Lrestore_nvgprs
+	bne-	.Lrestore_nvgprs_\srr
 
 #ifdef CONFIG_PPC_BOOK3S
-.Lfast_user_interrupt_return_amr:
+.Lfast_user_interrupt_return_amr_\srr\():
 	kuap_user_restore r3, r4
 #endif
-.Lfast_user_interrupt_return:
+.Lfast_user_interrupt_return_\srr\():
 	ld	r11,_NIP(r1)
 	ld	r12,_MSR(r1)
 BEGIN_FTR_SECTION
 	ld	r10,_PPR(r1)
 	mtspr	SPRN_PPR,r10
 END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
+	.ifc \srr,srr
 	mtspr	SPRN_SRR0,r11
 	mtspr	SPRN_SRR1,r12
+	.else
+	mtspr	SPRN_HSRR0,r11
+	mtspr	SPRN_HSRR1,r12
+	.endif
 
 BEGIN_FTR_SECTION
 	stdcx.	r0,0,r1		/* to clear the reservation */
@@ -706,24 +712,33 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_STCX_CHECKS_ADDRESS)
 	REST_GPR(6, r1)
 	REST_GPR(0, r1)
 	REST_GPR(1, r1)
+	.ifc \srr,srr
 	RFI_TO_USER
+	.else
+	HRFI_TO_USER
+	.endif
 	b	.	/* prevent speculative execution */
 
-.Lrestore_nvgprs:
+.Lrestore_nvgprs_\srr\():
 	REST_NVGPRS(r1)
-	b	.Lfast_user_interrupt_return
+	b	.Lfast_user_interrupt_return_\srr
 
 	.balign IFETCH_ALIGN_BYTES
-.Lkernel_interrupt_return:
+.Lkernel_interrupt_return_\srr\():
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	interrupt_exit_kernel_prepare
 
-.Lfast_kernel_interrupt_return:
+.Lfast_kernel_interrupt_return_\srr\():
 	cmpdi	cr1,r3,0
 	ld	r11,_NIP(r1)
 	ld	r12,_MSR(r1)
+	.ifc \srr,srr
 	mtspr	SPRN_SRR0,r11
 	mtspr	SPRN_SRR1,r12
+	.else
+	mtspr	SPRN_HSRR0,r11
+	mtspr	SPRN_HSRR1,r12
+	.endif
 
 BEGIN_FTR_SECTION
 	stdcx.	r0,0,r1		/* to clear the reservation */
@@ -757,7 +772,11 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_STCX_CHECKS_ADDRESS)
 	REST_GPR(6, r1)
 	REST_GPR(0, r1)
 	REST_GPR(1, r1)
+	.ifc \srr,srr
 	RFI_TO_KERNEL
+	.else
+	HRFI_TO_KERNEL
+	.endif
 	b	.	/* prevent speculative execution */
 
 1:	/*
@@ -777,8 +796,18 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_STCX_CHECKS_ADDRESS)
 	std	r9,0(r1) /* perform store component of stdu */
 	ld	r9,PACA_EXGEN+0(r13)
 
+	.ifc \srr,srr
 	RFI_TO_KERNEL
+	.else
+	HRFI_TO_KERNEL
+	.endif
 	b	.	/* prevent speculative execution */
+.endm
+
+interrupt_return_macro srr
+#ifdef CONFIG_PPC_BOOK3S
+interrupt_return_macro hsrr
+#endif
 
 #ifdef CONFIG_PPC_RTAS
 /*
arch/powerpc/kernel/exceptions-64e.S +4 -0
index f1ae710274bc..b35c97c7082f 100644
@@ -26,6 +26,10 @@
 #include <asm/feature-fixups.h>
 #include <asm/context_tracking.h>
 
+/* 64e interrupt returns always use SRR registers */
+#define fast_interrupt_return fast_interrupt_return_srr
+#define interrupt_return interrupt_return_srr
+
 /* XXX This will ultimately add space for a special exception save
  *     structure used to save things like SRR0/SRR1, SPRGs, MAS, etc...
  *     when taking special interrupts. For now we don't support that,
arch/powerpc/kernel/exceptions-64s.S +40 -36

288 lines not on this page · raw message

arch/powerpc/kernel/vector.S +1 -1
index f5a52f444e36..54dbefcb4cde 100644
@@ -131,7 +131,7 @@ _GLOBAL(load_up_vsx)
 	/* enable use of VSX after return */
 	oris	r12,r12,MSR_VSX@h
 	std	r12,_MSR(r1)
-	b	fast_interrupt_return
+	b	fast_interrupt_return_srr
 
 #endif /* CONFIG_VSX */
 
Signature · 1 line
2.23.0
author reviewer
#4 lore raw

PATCH v4 4/17 powerpc/64s: avoid reloading (H)SRR registers if they are still valid

In reply to Nicholas Piggin

When an interrupt is taken, the SRR registers are set to return to where it left off. Unless they are modified in the meantime, or the return address or MSR are modified, there is no need to reload these registers when returning from interrupt.

Introduce per-CPU flags that track the validity of SRR and HSRR registers. These are cleared when returning from interrupt, when using the registers for something else (e.g., OPAL calls), when adjusting the return address or MSR of a context, and when context switching (which changes the return address and MSR).

This improves the performance of interrupt returns.

---
 arch/powerpc/Kconfig.debug                 |  5 ++
 arch/powerpc/include/asm/hw_irq.h          | 10 ++-
 arch/powerpc/include/asm/interrupt.h       | 12 +++
 arch/powerpc/include/asm/paca.h            |  4 +
 arch/powerpc/include/asm/ptrace.h          | 74 ++++++++++++-----
 arch/powerpc/kernel/asm-offsets.c          |  4 +
 arch/powerpc/kernel/entry_64.S             | 92 ++++++++++++++++++++--
 arch/powerpc/kernel/exceptions-64s.S       | 27 +++++++
 arch/powerpc/kernel/fpu.S                  |  4 +
 arch/powerpc/kernel/kgdb.c                 |  2 +-
 arch/powerpc/kernel/kprobes-ftrace.c       |  2 +-
 arch/powerpc/kernel/kprobes.c              | 10 +--
 arch/powerpc/kernel/process.c              | 26 +++---
 arch/powerpc/kernel/prom_init.c            |  3 +
 arch/powerpc/kernel/rtas.c                 | 14 +++-
 arch/powerpc/kernel/signal.c               |  2 +-
 arch/powerpc/kernel/signal_64.c            |  9 +++
 arch/powerpc/kernel/syscalls.c             |  3 +-
 arch/powerpc/kernel/traps.c                | 18 ++---
 arch/powerpc/kernel/vector.S               |  6 ++
 arch/powerpc/kvm/book3s_hv.c               |  3 +
 arch/powerpc/kvm/book3s_pr.c               |  2 +
 arch/powerpc/lib/sstep.c                   |  4 +-
 arch/powerpc/math-emu/math.c               |  2 +-
 arch/powerpc/platforms/powernv/opal-call.c |  4 +
 arch/powerpc/platforms/pseries/hvCall.S    | 29 +++++++
 arch/powerpc/sysdev/fsl_pci.c              |  2 +-
 27 files changed, 310 insertions(+), 63 deletions(-)
arch/powerpc/Kconfig.debug +5 -0
index 6342f9da4545..205cd77f321f 100644
@@ -84,6 +84,11 @@ config MSI_BITMAP_SELFTEST
 
 config PPC_IRQ_SOFT_MASK_DEBUG
 	bool "Include extra checks for powerpc irq soft masking"
+	depends on PPC64
+
+config PPC_RFI_SRR_DEBUG
+	bool "Include extra checks for RFI SRR register validity"
+	depends on PPC_BOOK3S_64
 
 config XMON
 	bool "Include xmon kernel debugger"
arch/powerpc/include/asm/hw_irq.h +9 -1
index 56a98936a6a9..19bcef666cf6 100644
@@ -389,7 +389,15 @@ static inline bool arch_irq_disabled_regs(struct pt_regs *regs)
 	return !(regs->msr & MSR_EE);
 }
 
-static inline void may_hard_irq_enable(void) { }
+static inline bool may_hard_irq_enable(void)
+{
+	return false;
+}
+
+static inline void do_hard_irq_enable(void)
+{
+	BUILD_BUG();
+}
 
 static inline void irq_soft_mask_regs_set_state(struct pt_regs *regs, unsigned long val)
 {
arch/powerpc/include/asm/interrupt.h +12 -0
index 59f704408d65..6e9d18838d56 100644
@@ -73,6 +73,18 @@
 #include <asm/kprobes.h>
 #include <asm/runlatch.h>
 
+#ifdef CONFIG_PPC_BOOK3S_64
+static inline void srr_regs_clobbered(void)
+{
+	local_paca->srr_valid = 0;
+	local_paca->hsrr_valid = 0;
+}
+#else
+static inline void srr_regs_clobbered(void)
+{
+}
+#endif
+
 static inline void nap_adjust_return(struct pt_regs *regs)
 {
 #ifdef CONFIG_PPC_970_NAP
arch/powerpc/include/asm/paca.h +4 -0
index ec18ac818e3a..dfc984b0e640 100644
@@ -169,6 +169,10 @@ struct paca_struct {
 	u64 saved_msr;			/* MSR saved here by enter_rtas */
 #ifdef CONFIG_PPC_BOOK3E
 	u16 trap_save;			/* Used when bad stack is encountered */
+#endif
+#ifdef CONFIG_PPC_BOOK3S_64
+	u8 hsrr_valid;			/* HSRRs set for HRFID */
+	u8 srr_valid;			/* SRRs set for RFID */
 #endif
 	u8 irq_soft_mask;		/* mask for irq soft masking */
 	u8 irq_happened;		/* irq happened while soft-disabled */
arch/powerpc/include/asm/ptrace.h +53 -21

86 lines not on this page · raw message

arch/powerpc/kernel/asm-offsets.c +4 -0
index 28af4efb4587..ebe9afef619f 100644
@@ -209,6 +209,10 @@ int main(void)
 	OFFSET(PACATOC, paca_struct, kernel_toc);
 	OFFSET(PACAKBASE, paca_struct, kernelbase);
 	OFFSET(PACAKMSR, paca_struct, kernel_msr);
+#ifdef CONFIG_PPC_BOOK3S_64
+	OFFSET(PACAHSRR_VALID, paca_struct, hsrr_valid);
+	OFFSET(PACASRR_VALID, paca_struct, srr_valid);
+#endif
 	OFFSET(PACAIRQSOFTMASK, paca_struct, irq_soft_mask);
 	OFFSET(PACAIRQHAPPENED, paca_struct, irq_happened);
 	OFFSET(PACA_FTRACE_ENABLED, paca_struct, ftrace_enabled);
arch/powerpc/kernel/entry_64.S +85 -7

149 lines not on this page · raw message

arch/powerpc/kernel/exceptions-64s.S +27 -0
index 3a63feedae0b..5c18a2a3058d 100644
@@ -567,6 +567,20 @@ DEFINE_FIXED_SYMBOL(\name\()_common_real)
 	std	r0,GPR0(r1)		/* save r0 in stackframe	*/
 	std	r10,GPR1(r1)		/* save r1 in stackframe	*/
 
+	/* Mark our [H]SRRs valid for return */
+	li	r10,1
+	.if IHSRR_IF_HVMODE
+	BEGIN_FTR_SECTION
+	stb	r10,PACAHSRR_VALID(r13)
+	FTR_SECTION_ELSE
+	stb	r10,PACASRR_VALID(r13)
+	ALT_FTR_SECTION_END_IFSET(CPU_FTR_HVMODE | CPU_FTR_ARCH_206)
+	.elseif IHSRR
+	stb	r10,PACAHSRR_VALID(r13)
+	.else
+	stb	r10,PACASRR_VALID(r13)
+	.endif
+
 	.if ISET_RI
 	li	r10,MSR_RI
 	mtmsrd	r10,1			/* Set MSR_RI */
@@ -666,10 +680,13 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
 .macro EXCEPTION_RESTORE_REGS hsrr=0
 	/* Move original SRR0 and SRR1 into the respective regs */
 	ld	r9,_MSR(r1)
+	li	r10,0
 	.if \hsrr
 	mtspr	SPRN_HSRR1,r9
+	stb	r10,PACAHSRR_VALID(r13)
 	.else
 	mtspr	SPRN_SRR1,r9
+	stb	r10,PACASRR_VALID(r13)
 	.endif
 	ld	r9,_NIP(r1)
 	.if \hsrr
@@ -1825,6 +1842,8 @@ EXC_COMMON_BEGIN(hdecrementer_common)
 	 *
 	 * Be careful to avoid touching the kernel stack.
 	 */
+	li	r10,0
+	stb	r10,PACAHSRR_VALID(r13)
 	ld	r10,PACA_EXGEN+EX_CTR(r13)
 	mtctr	r10
 	mtcrf	0x80,r9
@@ -2660,6 +2679,8 @@ BEGIN_FTR_SECTION
 	ld	r10,PACA_EXGEN+EX_CFAR(r13)
 	mtspr	SPRN_CFAR,r10
 END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
+	li	r10,0
+	stb	r10,PACAHSRR_VALID(r13)
 	ld	r10,PACA_EXGEN+EX_R10(r13)
 	ld	r11,PACA_EXGEN+EX_R11(r13)
 	ld	r12,PACA_EXGEN+EX_R12(r13)
@@ -2828,6 +2849,12 @@ masked_interrupt:
 	ori	r11,r11,PACA_IRQ_HARD_DIS
 	stb	r11,PACAIRQHAPPENED(r13)
 2:	/* done */
+	li	r10,0
+	.if \hsrr
+	stb	r10,PACAHSRR_VALID(r13)
+	.else
+	stb	r10,PACASRR_VALID(r13)
+	.endif
 	ld	r10,PACA_EXGEN+EX_CTR(r13)
 	mtctr	r10
 	mtcrf	0x80,r9
arch/powerpc/kernel/fpu.S +4 -0
index 2c57ece6671c..6010adcee16e 100644
@@ -103,6 +103,10 @@ END_FTR_SECTION_IFSET(CPU_FTR_VSX)
 	ori	r12,r12,MSR_FP
 	or	r12,r12,r4
 	std	r12,_MSR(r1)
+#ifdef CONFIG_PPC_BOOK3S_64
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+#endif
 #endif
 	li	r4,1
 	stb	r4,THREAD_LOAD_FP(r5)
arch/powerpc/kernel/kgdb.c +1 -1
index 7dd2ad3603ad..26163497d69e 100644
@@ -147,7 +147,7 @@ static int kgdb_handle_breakpoint(struct pt_regs *regs)
 		return 0;
 
 	if (*(u32 *)regs->nip == BREAK_INSTR)
-		regs->nip += BREAK_INSTR_SIZE;
+		regs_add_return_ip(regs, BREAK_INSTR_SIZE);
 
 	return 1;
 }
arch/powerpc/kernel/kprobes-ftrace.c +1 -1
index 660138f6c4b2..a4965a32628a 100644
@@ -48,7 +48,7 @@ void kprobe_ftrace_handler(unsigned long nip, unsigned long parent_nip,
 			 * Emulate singlestep (and also recover regs->nip)
 			 * as if there is a nop
 			 */
-			regs->nip += MCOUNT_INSN_SIZE;
+			regs_add_return_ip(regs, MCOUNT_INSN_SIZE);
 			if (unlikely(p->post_handler)) {
 				kcb->kprobe_status = KPROBE_HIT_SSDONE;
 				p->post_handler(p, regs, 0);
arch/powerpc/kernel/kprobes.c +5 -5
index 01ab2163659e..8165ed71ab51 100644
@@ -178,7 +178,7 @@ static nokprobe_inline void prepare_singlestep(struct kprobe *p, struct pt_regs
 	 * variant as values in regs could play a part in
 	 * if the trap is taken or not
 	 */
-	regs->nip = (unsigned long)p->ainsn.insn;
+	regs_set_return_ip(regs, (unsigned long)p->ainsn.insn);
 }
 
 static nokprobe_inline void save_previous_kprobe(struct kprobe_ctlblk *kcb)
@@ -415,7 +415,7 @@ static int trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
 	 * we end up emulating it in kprobe_handler(), which increments the nip
 	 * again.
 	 */
-	regs->nip = orig_ret_address - 4;
+	regs_set_return_ip(regs, orig_ret_address - 4);
 	regs->link = orig_ret_address;
 
 	return 0;
@@ -450,7 +450,7 @@ int kprobe_post_handler(struct pt_regs *regs)
 	}
 
 	/* Adjust nip to after the single-stepped instruction */
-	regs->nip = (unsigned long)cur->addr + len;
+	regs_set_return_ip(regs, (unsigned long)cur->addr + len);
 	regs->msr |= kcb->kprobe_saved_msr;
 
 	/*Restore back the original saved kprobes variables and continue. */
@@ -490,7 +490,7 @@ int kprobe_fault_handler(struct pt_regs *regs, int trapnr)
 		 * and allow the page fault handler to continue as a
 		 * normal page fault.
 		 */
-		regs->nip = (unsigned long)cur->addr;
+		regs_set_return_ip(regs, (unsigned long)cur->addr);
 		regs->msr &= ~MSR_SINGLESTEP; /* Turn off 'trace' bits */
 		regs->msr |= kcb->kprobe_saved_msr;
 		if (kcb->kprobe_status == KPROBE_REENTER)
@@ -523,7 +523,7 @@ int kprobe_fault_handler(struct pt_regs *regs, int trapnr)
 		 * zero, try to fix up.
 		 */
 		if ((entry = search_exception_tables(regs->nip)) != NULL) {
-			regs->nip = extable_fixup(entry);
+			regs_set_return_ip(regs, extable_fixup(entry));
 			return 1;
 		}
 
arch/powerpc/kernel/process.c +16 -10
index 89e34aa273e2..fea5c68daeef 100644
@@ -96,7 +96,8 @@ static void check_if_tm_restore_required(struct task_struct *tsk)
 	if (tsk == current && tsk->thread.regs &&
 	    MSR_TM_ACTIVE(tsk->thread.regs->msr) &&
 	    !test_thread_flag(TIF_RESTORE_TM)) {
-		tsk->thread.ckpt_regs.msr = tsk->thread.regs->msr;
+		regs_set_return_msr(&tsk->thread.ckpt_regs,
+						tsk->thread.regs->msr);
 		set_thread_flag(TIF_RESTORE_TM);
 	}
 }
@@ -161,7 +162,7 @@ static void __giveup_fpu(struct task_struct *tsk)
 	msr &= ~(MSR_FP|MSR_FE0|MSR_FE1);
 	if (cpu_has_feature(CPU_FTR_VSX))
 		msr &= ~MSR_VSX;
-	tsk->thread.regs->msr = msr;
+	regs_set_return_msr(tsk->thread.regs, msr);
 }
 
 void giveup_fpu(struct task_struct *tsk)
@@ -244,7 +245,7 @@ static void __giveup_altivec(struct task_struct *tsk)
 	msr &= ~MSR_VEC;
 	if (cpu_has_feature(CPU_FTR_VSX))
 		msr &= ~MSR_VSX;
-	tsk->thread.regs->msr = msr;
+	regs_set_return_msr(tsk->thread.regs, msr);
 }
 
 void giveup_altivec(struct task_struct *tsk)
@@ -559,7 +560,7 @@ void notrace restore_math(struct pt_regs *regs)
 
 		msr_check_and_clear(new_msr);
 
-		regs->msr |= new_msr | fpexc_mode;
+		regs_set_return_msr(regs, regs->msr | new_msr | fpexc_mode);
 	}
 }
 #endif /* CONFIG_PPC_BOOK3S_64 */
@@ -1287,6 +1288,8 @@ struct task_struct *__switch_to(struct task_struct *prev,
 	}
 #endif /* CONFIG_PPC_BOOK3S_64 */
 
+	return_ip_or_msr_changed();
+
 	return last;
 }
 
@@ -1845,6 +1848,9 @@ void start_thread(struct pt_regs *regs, unsigned long start, unsigned long sp)
 		regs->gpr[2] = 0;
 		regs->msr = MSR_USER32;
 	}
+
+	return_ip_or_msr_changed();
+
 #endif
 #ifdef CONFIG_VSX
 	current->thread.used_vsr = 0;
@@ -1875,7 +1881,6 @@ void start_thread(struct pt_regs *regs, unsigned long start, unsigned long sp)
 	current->thread.tm_tfiar = 0;
 	current->thread.load_tm = 0;
 #endif /* CONFIG_PPC_TRANSACTIONAL_MEM */
-
 }
 EXPORT_SYMBOL(start_thread);
 
@@ -1923,9 +1928,10 @@ int set_fpexc_mode(struct task_struct *tsk, unsigned int val)
 	if (val > PR_FP_EXC_PRECISE)
 		return -EINVAL;
 	tsk->thread.fpexc_mode = __pack_fe01(val);
-	if (regs != NULL && (regs->msr & MSR_FP) != 0)
-		regs->msr = (regs->msr & ~(MSR_FE0|MSR_FE1))
-			| tsk->thread.fpexc_mode;
+	if (regs != NULL && (regs->msr & MSR_FP) != 0) {
+		regs_set_return_msr(regs, (regs->msr & ~(MSR_FE0|MSR_FE1))
+						| tsk->thread.fpexc_mode);
+	}
 	return 0;
 }
 
@@ -1971,9 +1977,9 @@ int set_endian(struct task_struct *tsk, unsigned int val)
 		return -EINVAL;
 
 	if (val == PR_ENDIAN_BIG)
-		regs->msr &= ~MSR_LE;
+		regs_set_return_msr(regs, regs->msr & ~MSR_LE);
 	else if (val == PR_ENDIAN_LITTLE || val == PR_ENDIAN_PPC_LITTLE)
-		regs->msr |= MSR_LE;
+		regs_set_return_msr(regs, regs->msr | MSR_LE);
 	else
 		return -EINVAL;
 
arch/powerpc/kernel/prom_init.c +3 -0
index 41ed7e33d897..f4774614e7eb 100644
@@ -31,6 +31,7 @@
 #include <asm/rtas.h>
 #include <asm/page.h>
 #include <asm/processor.h>
+#include <asm/interrupt.h>
 #include <asm/irq.h>
 #include <asm/io.h>
 #include <asm/smp.h>
@@ -1762,6 +1763,8 @@ static int prom_rtas_hcall(uint64_t args)
 	asm volatile("sc 1\n" : "=r" (arg1) :
 			"r" (arg1),
 			"r" (arg2) :);
+	srr_regs_clobbered();
+
 	return arg1;
 }
 
arch/powerpc/kernel/rtas.c +11 -3
index 6bada744402b..99f2cce635fb 100644
@@ -25,6 +25,7 @@
 #include <linux/reboot.h>
 #include <linux/syscalls.h>
 
+#include <asm/interrupt.h>
 #include <asm/prom.h>
 #include <asm/rtas.h>
 #include <asm/hvcall.h>
@@ -46,6 +47,13 @@
 /* This is here deliberately so it's only used in this file */
 void enter_rtas(unsigned long);
 
+static inline void do_enter_rtas(unsigned long args)
+{
+	enter_rtas(args);
+
+	srr_regs_clobbered(); /* rtas uses SRRs, invalidate */
+}
+
 struct rtas_t rtas = {
 	.lock = __ARCH_SPIN_LOCK_UNLOCKED
 };
@@ -384,7 +392,7 @@ static char *__fetch_rtas_last_error(char *altbuf)
 	save_args = rtas.args;
 	rtas.args = err_args;
 
-	enter_rtas(__pa(&rtas.args));
+	do_enter_rtas(__pa(&rtas.args));
 
 	err_args = rtas.args;
 	rtas.args = save_args;
@@ -430,7 +438,7 @@ va_rtas_call_unlocked(struct rtas_args *args, int token, int nargs, int nret,
 	for (i = 0; i < nret; ++i)
 		args->rets[i] = 0;
 
-	enter_rtas(__pa(args));
+	do_enter_rtas(__pa(args));
 }
 
 void rtas_call_unlocked(struct rtas_args *args, int token, int nargs, int nret, ...)
@@ -1138,7 +1146,7 @@ SYSCALL_DEFINE1(rtas, struct rtas_args __user *, uargs)
 	flags = lock_rtas();
 
 	rtas.args = args;
-	enter_rtas(__pa(&rtas.args));
+	do_enter_rtas(__pa(&rtas.args));
 	args = rtas.args;
 
 	/* A -1 return code indicates that the last command couldn't
arch/powerpc/kernel/signal.c +1 -1
index 9ded046edb0e..3f851fa8f5f1 100644
@@ -214,7 +214,7 @@ static void check_syscall_restart(struct pt_regs *regs, struct k_sigaction *ka,
 			regs->gpr[0] = __NR_restart_syscall;
 		else
 			regs->gpr[3] = regs->orig_gpr3;
-		regs->nip -= 4;
+		regs_add_return_ip(regs, -4);
 		regs->result = 0;
 	} else {
 		if (trap_is_scv(regs)) {
arch/powerpc/kernel/signal_64.c +9 -0
index dca66481d0c2..4a62faefba0f 100644
@@ -720,6 +720,9 @@ SYSCALL_DEFINE3(swapcontext, struct ucontext __user *, old_ctx,
 
 	/* This returns like rt_sigreturn */
 	set_thread_flag(TIF_RESTOREALL);
+
+	return_ip_or_msr_changed();
+
 	return 0;
 
 efault_out:
@@ -832,6 +835,9 @@ SYSCALL_DEFINE0(rt_sigreturn)
 		goto badframe;
 
 	set_thread_flag(TIF_RESTOREALL);
+
+	return_ip_or_msr_changed();
+
 	return 0;
 
 badframe_block:
@@ -957,12 +963,15 @@ int handle_rt_signal64(struct ksignal *ksig, sigset_t *set,
 	if (err)
 		goto badframe;
 
+	return_ip_or_msr_changed();
+
 	return 0;
 
 badframe_block:
 	user_write_access_end();
 badframe:
 	signal_fault(current, regs, "handle_rt_signal64", frame);
+	return_ip_or_msr_changed();
 
 	return 1;
 }
arch/powerpc/kernel/syscalls.c +2 -1
index a552c9e68d7e..bf4ae0f0e36c 100644
@@ -114,7 +114,8 @@ SYSCALL_DEFINE0(switch_endian)
 {
 	struct thread_info *ti;
 
-	current->thread.regs->msr ^= MSR_LE;
+	regs_set_return_msr(current->thread.regs,
+				current->thread.regs->msr ^ MSR_LE);
 
 	/*
 	 * Set TIF_RESTOREALL so that r3 isn't clobbered on return to
arch/powerpc/kernel/traps.c +9 -9
index b4ab95c9e94a..e413f9908664 100644
@@ -1031,7 +1031,7 @@ static void p9_hmi_special_emu(struct pt_regs *regs)
 #endif /* !__LITTLE_ENDIAN__ */
 
 	/* Go to next instruction */
-	regs->nip += 4;
+	regs_add_return_ip(regs, 4);
 }
 #endif /* CONFIG_VSX */
 
@@ -1476,7 +1476,7 @@ static void do_program_check(struct pt_regs *regs)
 
 		if (!(regs->msr & MSR_PR) &&  /* not user-mode */
 		    report_bug(bugaddr, regs) == BUG_TRAP_TYPE_WARN) {
-			regs->nip += 4;
+			regs_add_return_ip(regs, 4);
 			return;
 		}
 		_exception(SIGTRAP, regs, TRAP_BRKPT, regs->nip);
@@ -1538,7 +1538,7 @@ static void do_program_check(struct pt_regs *regs)
 	if (reason & (REASON_ILLEGAL | REASON_PRIVILEGED)) {
 		switch (emulate_instruction(regs)) {
 		case 0:
-			regs->nip += 4;
+			regs_add_return_ip(regs, 4);
 			emulate_single_step(regs);
 			return;
 		case -EFAULT:
@@ -1593,7 +1593,7 @@ DEFINE_INTERRUPT_HANDLER(alignment_exception)
 
 	if (fixed == 1) {
 		/* skip over emulated instruction */
-		regs->nip += inst_length(reason);
+		regs_add_return_ip(regs, inst_length(reason));
 		emulate_single_step(regs);
 		return;
 	}
@@ -1751,7 +1751,7 @@ DEFINE_INTERRUPT_HANDLER(facility_unavailable_exception)
 				pr_err("DSCR based mfspr emulation failed\n");
 				return;
 			}
-			regs->nip += 4;
+			regs_add_return_ip(regs, 4);
 			emulate_single_step(regs);
 		}
 		return;
@@ -2044,7 +2044,7 @@ DEFINE_INTERRUPT_HANDLER(altivec_assist_exception)
 	PPC_WARN_EMULATED(altivec, regs);
 	err = emulate_altivec(regs);
 	if (err == 0) {
-		regs->nip += 4;		/* skip emulated instruction */
+		regs_add_return_ip(regs, 4); /* skip emulated instruction */
 		emulate_single_step(regs);
 		return;
 	}
@@ -2109,7 +2109,7 @@ DEFINE_INTERRUPT_HANDLER(SPEFloatingPointException)
 
 	err = do_spe_mathemu(regs);
 	if (err == 0) {
-		regs->nip += 4;		/* skip emulated instruction */
+		regs_add_return_ip(regs, 4); /* skip emulated instruction */
 		emulate_single_step(regs);
 		return;
 	}
@@ -2140,10 +2140,10 @@ DEFINE_INTERRUPT_HANDLER(SPEFloatingPointRoundException)
 		giveup_spe(current);
 	preempt_enable();
 
-	regs->nip -= 4;
+	regs_add_return_ip(regs, -4);
 	err = speround_handler(regs);
 	if (err == 0) {
-		regs->nip += 4;		/* skip emulated instruction */
+		regs_add_return_ip(regs, 4); /* skip emulated instruction */
 		emulate_single_step(regs);
 		return;
 	}
arch/powerpc/kernel/vector.S +6 -0
index 54dbefcb4cde..fc120fac1910 100644
@@ -73,6 +73,10 @@ _GLOBAL(load_up_altivec)
 	addi	r5,r4,THREAD		/* Get THREAD */
 	oris	r12,r12,MSR_VEC@h
 	std	r12,_MSR(r1)
+#ifdef CONFIG_PPC_BOOK3S_64
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+#endif
 #endif
 	li	r4,1
 	stb	r4,THREAD_LOAD_VEC(r5)
@@ -131,6 +135,8 @@ _GLOBAL(load_up_vsx)
 	/* enable use of VSX after return */
 	oris	r12,r12,MSR_VSX@h
 	std	r12,_MSR(r1)
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
 	b	fast_interrupt_return_srr
 
 #endif /* CONFIG_VSX */
arch/powerpc/kvm/book3s_hv.c +3 -0
index 28a80d240b76..efd95b867210 100644
@@ -4461,6 +4461,9 @@ static int kvmppc_vcpu_run_hv(struct kvm_vcpu *vcpu)
 
 	vcpu->arch.state = KVMPPC_VCPU_NOTREADY;
 	atomic_dec(&kvm->arch.vcpus_running);
+
+	srr_regs_clobbered();
+
 	return r;
 }
 
arch/powerpc/kvm/book3s_pr.c +2 -0
index d7733b07f489..1ed5ceee73eb 100644
@@ -25,6 +25,7 @@
 #include <asm/cputable.h>
 #include <asm/cacheflush.h>
 #include <linux/uaccess.h>
+#include <asm/interrupt.h>
 #include <asm/io.h>
 #include <asm/kvm_ppc.h>
 #include <asm/kvm_book3s.h>
@@ -1848,6 +1849,7 @@ static int kvmppc_vcpu_run_pr(struct kvm_vcpu *vcpu)
 	/* Make sure we save the guest TAR/EBB/DSCR state */
 	kvmppc_giveup_fac(vcpu, FSCR_TAR_LG);
 
+	srr_regs_clobbered();
 out:
 	vcpu->mode = OUTSIDE_GUEST_MODE;
 	return ret;
arch/powerpc/lib/sstep.c +3 -1
index 45bda2520755..90fa3878299a 100644
@@ -3203,7 +3203,7 @@ void emulate_update_regs(struct pt_regs *regs, struct instruction_op *op)
 	default:
 		WARN_ON_ONCE(1);
 	}
-	regs->nip = next_pc;
+	regs_set_return_ip(regs, next_pc);
 }
 NOKPROBE_SYMBOL(emulate_update_regs);
 
@@ -3480,6 +3480,8 @@ int emulate_step(struct pt_regs *regs, struct ppc_inst instr)
 	unsigned long val;
 	unsigned long ea;
 
+	return_ip_or_msr_changed();
+
 	r = analyse_instr(&op, regs, instr);
 	if (r < 0)
 		return r;
arch/powerpc/math-emu/math.c +1 -1
index 327165f26ca6..36761bd00f38 100644
@@ -453,7 +453,7 @@ do_mathemu(struct pt_regs *regs)
 		break;
 	}
 
-	regs->nip += 4;
+	regs_add_return_ip(regs, 4);
 	return 0;
 
 illegal:
arch/powerpc/platforms/powernv/opal-call.c +4 -0
index 01401e3da7ca..f812c74c61e5 100644
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0
 #include <linux/percpu.h>
 #include <linux/jump_label.h>
+#include <asm/interrupt.h>
 #include <asm/opal-api.h>
 #include <asm/trace.h>
 #include <asm/asm-prototypes.h>
@@ -100,6 +101,9 @@ static int64_t opal_call(int64_t a0, int64_t a1, int64_t a2, int64_t a3,
 	bool mmu = (msr & (MSR_IR|MSR_DR));
 	int64_t ret;
 
+	/* OPAL call / firmware may use SRR and/or HSRR */
+	srr_regs_clobbered();
+
 	msr &= ~MSR_EE;
 
 	if (unlikely(!mmu))
arch/powerpc/platforms/pseries/hvCall.S +29 -0
index 8a2b8d64265b..ab9fc6506861 100644
@@ -108,6 +108,10 @@ _GLOBAL_TOC(plpar_hcall_norets_notrace)
 	mfcr	r0
 	stw	r0,8(r1)
 	HVSC				/* invoke the hypervisor */
+
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 	blr				/* return r3 = status */
@@ -120,6 +124,9 @@ _GLOBAL_TOC(plpar_hcall_norets)
 	HCALL_BRANCH(plpar_hcall_norets_trace)
 	HVSC				/* invoke the hypervisor */
 
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 	blr				/* return r3 = status */
@@ -129,6 +136,10 @@ plpar_hcall_norets_trace:
 	HCALL_INST_PRECALL(R4)
 	HVSC
 	HCALL_INST_POSTCALL_NORETS
+
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 	blr
@@ -159,6 +170,9 @@ _GLOBAL_TOC(plpar_hcall)
 	std	r6, 16(r12)
 	std	r7, 24(r12)
 
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 
@@ -188,6 +202,9 @@ plpar_hcall_trace:
 
 	HCALL_INST_POSTCALL(r12)
 
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 
@@ -223,6 +240,9 @@ _GLOBAL(plpar_hcall_raw)
 	std	r6, 16(r12)
 	std	r7, 24(r12)
 
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 
@@ -262,6 +282,9 @@ _GLOBAL_TOC(plpar_hcall9)
 	std	r11,56(r12)
 	std	r0, 64(r12)
 
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 
@@ -300,6 +323,9 @@ plpar_hcall9_trace:
 
 	HCALL_INST_POSTCALL(r12)
 
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 
@@ -339,6 +365,9 @@ _GLOBAL(plpar_hcall9_raw)
 	std	r11,56(r12)
 	std	r0, 64(r12)
 
+	li	r4,0
+	stb	r4,PACASRR_VALID(r13)
+
 	lwz	r0,8(r1)
 	mtcrf	0xff,r0
 
arch/powerpc/sysdev/fsl_pci.c +1 -1
index 69af73765783..b8f76f3fd994 100644
@@ -1072,7 +1072,7 @@ int fsl_pci_mcheck_exception(struct pt_regs *regs)
 			ret = get_kernel_nofault(inst, (void *)regs->nip);
 
 		if (!ret && mcheck_handle_load(regs, inst)) {
-			regs->nip += 4;
+			regs_add_return_ip(regs, 4);
 			return 1;
 		}
 	}
Signature · 1 line
2.23.0
author reviewer
#5 lore raw

PATCH v4 5/17 powerpc/64: handle MSR EE and RI in interrupt entry wrapper

In reply to Nicholas Piggin

Similarly to the system call change in the previous patch, the mtmsrd to enable RI can be combined with the mtmsrd to enable EE for interrupts which enable the latter, which tends to be the important synchronous interrupts (i.e., page faults).

Do this by enabling EE and RI together at the beginning of the entry wrapper if PACA_IRQ_HARD_DIS is clear, and just enabling RI if it is set (which means something wanted EE=0).

Asynchronous interrupts set PACA_IRQ_HARD_DIS, but synchronous ones leave it unchanged, so by default they always get EE=1 unless they interrupt a caller that has hard disabled. When the sync interrupt later calls interrupt_cond_local_irq_enable(), that will not require another mtmsrd because we already enabled here.

This tends to save one mtmsrd L=1 for synchronous interrupts on 64s. 64e is conceptually unchanged, but it also sets MSR[EE]=1 now in the interrupt wrapper for synchronous interrupts with the same code.

From: Nicholas Piggin <[email protected]>

---
 arch/powerpc/include/asm/interrupt.h | 22 ++++++++++++++++++--
 arch/powerpc/kernel/exceptions-64s.S | 30 ----------------------------
 2 files changed, 20 insertions(+), 32 deletions(-)
arch/powerpc/include/asm/interrupt.h +20 -2
index 6e9d18838d56..b9c510187b58 100644
@@ -126,9 +126,21 @@ static inline void interrupt_enter_prepare(struct pt_regs *regs, struct interrup
 #endif
 
 #ifdef CONFIG_PPC64
-	if (irq_soft_mask_set_return(IRQS_ALL_DISABLED) == IRQS_ENABLED)
+	bool trace_enable = false;
+
+	if (IS_ENABLED(CONFIG_TRACE_IRQFLAGS)) {
+		if (irq_soft_mask_set_return(IRQS_DISABLED) == IRQS_ENABLED)
+			trace_enable = true;
+	} else {
+		irq_soft_mask_set(IRQS_DISABLED);
+	}
+	/* If the interrupt was taken with HARD_DIS set, don't enable MSR[EE] */
+	if (local_paca->irq_happened & PACA_IRQ_HARD_DIS)
+		__hard_RI_enable();
+	else
+		__hard_irq_enable();
+	if (trace_enable)
 		trace_hardirqs_off();
-	local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
 
 	if (user_mode(regs)) {
 		CT_WARN_ON(ct_state() != CONTEXT_USER);
@@ -175,6 +187,10 @@ static inline void interrupt_async_enter_prepare(struct pt_regs *regs, struct in
 		__ppc64_runlatch_on();
 #endif
 
+#ifdef CONFIG_PPC64
+	/* Ensure interrupt_enter_prepare does not enable MSR[EE] */
+	local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
+#endif
 	interrupt_enter_prepare(regs, state);
 	irq_enter();
 }
@@ -239,6 +255,8 @@ static inline void interrupt_nmi_enter_prepare(struct pt_regs *regs, struct inte
 		regs->softe = IRQS_ALL_DISABLED;
 	}
 
+	__hard_RI_enable();
+
 	/* Don't do any per-CPU operations until interrupt state is fixed */
 
 	if (nmi_disables_ftrace(regs)) {
arch/powerpc/kernel/exceptions-64s.S +0 -30
index 5c18a2a3058d..a2ae14d0600e 100644
@@ -129,7 +129,6 @@ name:
 #define IISIDE		.L_IISIDE_\name\()	/* Uses SRR0/1 not DAR/DSISR */
 #define IDAR		.L_IDAR_\name\()	/* Uses DAR (or SRR0) */
 #define IDSISR		.L_IDSISR_\name\()	/* Uses DSISR (or SRR1) */
-#define ISET_RI		.L_ISET_RI_\name\()	/* Run common code w/ MSR[RI]=1 */
 #define IBRANCH_TO_COMMON	.L_IBRANCH_TO_COMMON_\name\() /* ENTRY branch to common */
 #define IREALMODE_COMMON	.L_IREALMODE_COMMON_\name\() /* Common runs in realmode */
 #define IMASK		.L_IMASK_\name\()	/* IRQ soft-mask bit */
@@ -174,9 +173,6 @@ do_define_int n
 	.ifndef IDSISR
 		IDSISR=0
 	.endif
-	.ifndef ISET_RI
-		ISET_RI=1
-	.endif
 	.ifndef IBRANCH_TO_COMMON
 		IBRANCH_TO_COMMON=1
 	.endif
@@ -581,11 +577,6 @@ DEFINE_FIXED_SYMBOL(\name\()_common_real)
 	stb	r10,PACASRR_VALID(r13)
 	.endif
 
-	.if ISET_RI
-	li	r10,MSR_RI
-	mtmsrd	r10,1			/* Set MSR_RI */
-	.endif
-
 	.if ISTACK
 	.if IKUAP
 	kuap_save_amr_and_lock r9, r10, cr1, cr0
@@ -906,11 +897,6 @@ INT_DEFINE_BEGIN(system_reset)
 	IVEC=0x100
 	IAREA=PACA_EXNMI
 	IVIRT=0 /* no virt entry point */
-	/*
-	 * MSR_RI is not enabled, because PACA_EXNMI and nmi stack is
-	 * being used, so a nested NMI exception would corrupt it.
-	 */
-	ISET_RI=0
 	ISTACK=0
 	IKVM_REAL=1
 INT_DEFINE_END(system_reset)
@@ -991,8 +977,6 @@ EXC_COMMON_BEGIN(system_reset_common)
 	lhz	r10,PACA_IN_NMI(r13)
 	addi	r10,r10,1
 	sth	r10,PACA_IN_NMI(r13)
-	li	r10,MSR_RI
-	mtmsrd 	r10,1
 
 	mr	r10,r1
 	ld	r1,PACA_NMI_EMERG_SP(r13)
@@ -1068,12 +1052,6 @@ INT_DEFINE_BEGIN(machine_check_early)
 	IAREA=PACA_EXMC
 	IVIRT=0 /* no virt entry point */
 	IREALMODE_COMMON=1
-	/*
-	 * MSR_RI is not enabled, because PACA_EXMC is being used, so a
-	 * nested machine check corrupts it. machine_check_common enables
-	 * MSR_RI.
-	 */
-	ISET_RI=0
 	ISTACK=0
 	IDAR=1
 	IDSISR=1
@@ -1084,7 +1062,6 @@ INT_DEFINE_BEGIN(machine_check)
 	IVEC=0x200
 	IAREA=PACA_EXMC
 	IVIRT=0 /* no virt entry point */
-	ISET_RI=0
 	IDAR=1
 	IDSISR=1
 	IKVM_SKIP=1
@@ -1155,9 +1132,6 @@ EXC_COMMON_BEGIN(machine_check_early_common)
 BEGIN_FTR_SECTION
 	bl	enable_machine_check
 END_FTR_SECTION_IFSET(CPU_FTR_HVMODE)
-	li	r10,MSR_RI
-	mtmsrd	r10,1
-
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	machine_check_early
 	std	r3,RESULT(r1)	/* Save result */
@@ -1245,10 +1219,6 @@ EXC_COMMON_BEGIN(machine_check_common)
 	 * save area: PACA_EXMC instead of PACA_EXGEN.
 	 */
 	GEN_COMMON machine_check
-
-	/* Enable MSR_RI when finished with PACA_EXMC */
-	li	r10,MSR_RI
-	mtmsrd 	r10,1
 	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	machine_check_exception
 	b	interrupt_return_srr
Signature · 1 line
2.23.0
author reviewer
#6 lore raw

PATCH v4 6/17 powerpc/64: move interrupt return asm to interrupt_64.S

In reply to Nicholas Piggin

The next patch would like to move interrupt return assembly code to a low location before general text, so move it into its own file and include via head_64.S

---
 arch/powerpc/include/asm/head-64.h |   2 +-
 arch/powerpc/kernel/entry_64.S     | 623 ----------------------------
 arch/powerpc/kernel/head_64.S      |   5 +-
 arch/powerpc/kernel/interrupt_64.S | 635 +++++++++++++++++++++++++++++
 4 files changed, 640 insertions(+), 625 deletions(-)
 create mode 100644 arch/powerpc/kernel/interrupt_64.S
arch/powerpc/include/asm/head-64.h +1 -1
index 4cb9efa2eb21..242204e12993 100644
@@ -16,7 +16,7 @@
 	.section ".head.data.\name\()","a",@progbits
 .endm
 .macro use_ftsec name
-	.section ".head.text.\name\()"
+	.section ".head.text.\name\()","ax",@progbits
 .endm
 
 /*
arch/powerpc/kernel/entry_64.S +0 -623

642 lines not on this page · raw message

arch/powerpc/kernel/head_64.S +4 -1
index ece7f97bafff..d49c25daf1c0 100644
@@ -194,8 +194,9 @@ CLOSE_FIXED_SECTION(first_256B)
 
 /* This value is used to mark exception frames on the stack. */
 	.section ".toc","aw"
+/* This value is used to mark exception frames on the stack. */
 exception_marker:
-	.tc	ID_72656773_68657265[TC],0x7265677368657265
+	.tc	ID_EXC_MARKER[TC],STACK_FRAME_REGS_MARKER
 	.previous
 
 /*
@@ -211,6 +212,8 @@ OPEN_TEXT_SECTION(0x100)
 
 USE_TEXT_SECTION()
 
+#include "interrupt_64.S"
+
 #ifdef CONFIG_PPC_BOOK3E
 /*
  * The booting_thread_hwid holds the thread id we want to boot in cpu
arch/powerpc/kernel/interrupt_64.S new file +635 -0

635 lines not on this page · raw message

Signature · 1 line
2.23.0
author reviewer
#7 lore raw

PATCH v4 7/17 powerpc/64s: system call avoid setting MSR[RI] until we set MSR[EE]

In reply to Nicholas Piggin

This extends the MSR[RI]=0 window a little further into the system call in order to pair RI and EE enabling with a single mtmsrd.

---
 arch/powerpc/kernel/exceptions-64s.S | 2 --
 arch/powerpc/kernel/interrupt_64.S   | 4 ++--
 2 files changed, 2 insertions(+), 4 deletions(-)
arch/powerpc/kernel/exceptions-64s.S +0 -2
index a2ae14d0600e..b6e1c46c97d0 100644
@@ -1942,8 +1942,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_REAL_LE)
 	mtctr	r10
 	bctr
 	.else
-	li	r10,MSR_RI
-	mtmsrd 	r10,1			/* Set RI (EE=0) */
 #ifdef CONFIG_RELOCATABLE
 	__LOAD_HANDLER(r10, system_call_common)
 	mtctr	r10
arch/powerpc/kernel/interrupt_64.S +2 -2
index e17a77a2c1a1..ab6b99609d0e 100644
@@ -283,9 +283,9 @@ END_BTB_FLUSH_SECTION
 	 * trace_hardirqs_off().
 	 */
 	li	r11,IRQS_ALL_DISABLED
-	li	r12,PACA_IRQ_HARD_DIS
+	li	r12,-1 /* Set MSR_EE and MSR_RI */
 	stb	r11,PACAIRQSOFTMASK(r13)
-	stb	r12,PACAIRQHAPPENED(r13)
+	mtmsrd	r12,1
 
 	/* Calling convention has r9 = orig r0, r10 = regs */
 	mr	r9,r0
Signature · 1 line
2.23.0
author reviewer
#8 lore raw

PATCH v4 8/17 powerpc/64s: save one more register in the masked interrupt handler

In reply to Nicholas Piggin

This frees up one more register (and takes advantage of that to clean things up a little bit).

This register will be used in the following patch.

---
 arch/powerpc/kernel/exceptions-64s.S | 34 ++++++++++++++++------------
 1 file changed, 20 insertions(+), 14 deletions(-)
arch/powerpc/kernel/exceptions-64s.S +20 -14
index b6e1c46c97d0..0ba8c2387aac 100644
@@ -2758,7 +2758,6 @@ INT_DEFINE_END(soft_nmi)
  * and run it entirely with interrupts hard disabled.
  */
 EXC_COMMON_BEGIN(soft_nmi_common)
-	mfspr	r11,SPRN_SRR0
 	mr	r10,r1
 	ld	r1,PACAEMERGSP(r13)
 	subi	r1,r1,INT_FRAME_SIZE
@@ -2793,19 +2792,24 @@ masked_Hinterrupt:
 	.else
 masked_interrupt:
 	.endif
-	lbz	r11,PACAIRQHAPPENED(r13)
-	or	r11,r11,r10
-	stb	r11,PACAIRQHAPPENED(r13)
+	stw	r9,PACA_EXGEN+EX_CCR(r13)
+	lbz	r9,PACAIRQHAPPENED(r13)
+	or	r9,r9,r10
+	stb	r9,PACAIRQHAPPENED(r13)
+
+	.if ! \hsrr
 	cmpwi	r10,PACA_IRQ_DEC
 	bne	1f
-	lis	r10,0x7fff
-	ori	r10,r10,0xffff
-	mtspr	SPRN_DEC,r10
+	LOAD_REG_IMMEDIATE(r9, 0x7fffffff)
+	mtspr	SPRN_DEC,r9
 #ifdef CONFIG_PPC_WATCHDOG
+	lwz	r9,PACA_EXGEN+EX_CCR(r13)
 	b	soft_nmi_common
 #else
 	b	2f
 #endif
+	.endif
+
 1:	andi.	r10,r10,PACA_IRQ_MUST_HARD_MASK
 	beq	2f
 	xori	r12,r12,MSR_EE	/* clear MSR_EE */
@@ -2814,17 +2818,19 @@ masked_interrupt:
 	.else
 	mtspr	SPRN_SRR1,r12
 	.endif
-	ori	r11,r11,PACA_IRQ_HARD_DIS
-	stb	r11,PACAIRQHAPPENED(r13)
+	ori	r9,r9,PACA_IRQ_HARD_DIS
+	stb	r9,PACAIRQHAPPENED(r13)
 2:	/* done */
-	li	r10,0
+	li	r9,0
 	.if \hsrr
-	stb	r10,PACAHSRR_VALID(r13)
+	stb	r9,PACAHSRR_VALID(r13)
 	.else
-	stb	r10,PACASRR_VALID(r13)
+	stb	r9,PACASRR_VALID(r13)
 	.endif
-	ld	r10,PACA_EXGEN+EX_CTR(r13)
-	mtctr	r10
+
+	ld	r9,PACA_EXGEN+EX_CTR(r13)
+	mtctr	r9
+	lwz	r9,PACA_EXGEN+EX_CCR(r13)
 	mtcrf	0x80,r9
 	std	r1,PACAR1(r13)
 	ld	r9,PACA_EXGEN+EX_R9(r13)
Signature · 1 line
2.23.0
author reviewer
#9 lore raw

PATCH v4 9/17 powerpc/64: allow alternate return locations for soft-masked interrupts

In reply to Nicholas Piggin

The exception table fixup adjusts a failed page fault's interrupt return location if it was taken at an address specified in the exception table, to a corresponding fixup handler address.

Introduce a variation of that idea which adds a fixup table for NMIs and soft-masked asynchronous interrupts. This will be used to protect certain critical sections that are sensitive to being clobbered by interrupts coming in (due to using the same SPRs and/or irq soft-mask state).

---
 arch/powerpc/include/asm/interrupt.h | 13 +++++++++
 arch/powerpc/include/asm/ppc_asm.h   |  8 ++++++
 arch/powerpc/kernel/exceptions-64e.S | 37 +++++++++++++++++++++++--
 arch/powerpc/kernel/exceptions-64s.S | 41 ++++++++++++++++++++++++++++
 arch/powerpc/kernel/vmlinux.lds.S    | 10 +++++++
 arch/powerpc/lib/Makefile            |  2 +-
 arch/powerpc/lib/restart_table.c     | 30 ++++++++++++++++++++
 7 files changed, 138 insertions(+), 3 deletions(-)
 create mode 100644 arch/powerpc/lib/restart_table.c
arch/powerpc/include/asm/interrupt.h +13 -0
index b9c510187b58..49d9a6fd1bb9 100644
@@ -73,6 +73,11 @@
 #include <asm/kprobes.h>
 #include <asm/runlatch.h>
 
+#ifdef CONFIG_PPC64
+extern char __end_soft_masked[];
+unsigned long search_kernel_restart_table(unsigned long addr);
+#endif
+
 #ifdef CONFIG_PPC_BOOK3S_64
 static inline void srr_regs_clobbered(void)
 {
@@ -287,6 +292,14 @@ static inline void interrupt_nmi_exit_prepare(struct pt_regs *regs, struct inter
 	 * new work to do (must use irq_work for that).
 	 */
 
+#ifdef CONFIG_PPC64
+	if (arch_irq_disabled_regs(regs)) {
+		unsigned long rst = search_kernel_restart_table(regs->nip);
+		if (rst)
+			regs_set_return_ip(regs, rst);
+	}
+#endif
+
 #ifdef CONFIG_PPC64
 	if (nmi_disables_ftrace(regs))
 		this_cpu_set_ftrace_enabled(state->ftrace_enabled);
arch/powerpc/include/asm/ppc_asm.h +8 -0
index d6739d700f0a..c9c2c36c1f8f 100644
@@ -762,6 +762,14 @@ END_FTR_SECTION_NESTED(CPU_FTR_CELL_TB_BUG, CPU_FTR_CELL_TB_BUG, 96)
 	stringify_in_c(.long (_target) - . ;)	\
 	stringify_in_c(.previous)
 
+#define RESTART_TABLE(_start, _end, _target)	\
+	stringify_in_c(.section __restart_table,"a";)\
+	stringify_in_c(.balign 8;)		\
+	stringify_in_c(.llong (_start);)	\
+	stringify_in_c(.llong (_end);)		\
+	stringify_in_c(.llong (_target);)	\
+	stringify_in_c(.previous)
+
 #ifdef CONFIG_PPC_FSL_BOOK3E
 #define BTB_FLUSH(reg)			\
 	lis reg,BUCSR_INIT@h;		\
arch/powerpc/kernel/exceptions-64e.S +35 -2
index b35c97c7082f..1b79f8a75298 100644
@@ -901,6 +901,28 @@ kernel_dbg_exc:
 	bl	unknown_exception
 	b	interrupt_return
 
+.macro SEARCH_RESTART_TABLE
+	LOAD_REG_IMMEDIATE_SYM(r14, r11, __start___restart_table)
+	LOAD_REG_IMMEDIATE_SYM(r15, r11, __stop___restart_table)
+300:
+	cmpd	r14,r15
+	beq	302f
+	ld	r11,0(r14)
+	cmpld	r10,r11
+	blt	301f
+	ld	r11,8(r14)
+	cmpld	r10,r11
+	bge	301f
+	ld	r11,16(r14)
+	b	303f
+301:
+	addi	r14,r14,24
+	b	300b
+302:
+	li	r11,0
+303:
+.endm
+
 /*
  * An interrupt came in while soft-disabled; We mark paca->irq_happened
  * accordingly and if the interrupt is level sensitive, we hard disable
@@ -909,6 +931,9 @@ kernel_dbg_exc:
  */
 
 .macro masked_interrupt_book3e paca_irq full_mask
+	std	r14,PACA_EXGEN+EX_R14(r13)
+	std	r15,PACA_EXGEN+EX_R15(r13)
+
 	lbz	r10,PACAIRQHAPPENED(r13)
 	.if \full_mask == 1
 	ori	r10,r10,\paca_irq | PACA_IRQ_HARD_DIS
@@ -918,15 +943,23 @@ kernel_dbg_exc:
 	stb	r10,PACAIRQHAPPENED(r13)
 
 	.if \full_mask == 1
-	rldicl	r10,r11,48,1		/* clear MSR_EE */
-	rotldi	r11,r10,16
+	xori	r11,r11,MSR_EE		/* clear MSR_EE */
 	mtspr	SPRN_SRR1,r11
 	.endif
 
+	mfspr	r10,SPRN_SRR0
+	SEARCH_RESTART_TABLE
+	cmpdi	r11,0
+	beq	1f
+	mtspr	SPRN_SRR0,r11		/* return to restart address */
+1:
+
 	lwz	r11,PACA_EXGEN+EX_CR(r13)
 	mtcr	r11
 	ld	r10,PACA_EXGEN+EX_R10(r13)
 	ld	r11,PACA_EXGEN+EX_R11(r13)
+	ld	r14,PACA_EXGEN+EX_R14(r13)
+	ld	r15,PACA_EXGEN+EX_R15(r13)
 	mfspr	r13,SPRN_SPRG_GEN_SCRATCH
 	rfi
 	b	.
arch/powerpc/kernel/exceptions-64s.S +41 -0

59 lines not on this page · raw message

arch/powerpc/kernel/vmlinux.lds.S +10 -0
index 72fa3c00229a..16c5e13e00c4 100644
@@ -9,6 +9,14 @@
 #define EMITS_PT_NOTE
 #define RO_EXCEPTION_TABLE_ALIGN	0
 
+#define RESTART_TABLE(align)						\
+	. = ALIGN(align);						\
+	__restart_table : AT(ADDR(__restart_table) - LOAD_OFFSET) {	\
+		__start___restart_table = .;				\
+		KEEP(*(__restart_table))				\
+		__stop___restart_table = .;				\
+	}
+
 #include <asm/page.h>
 #include <asm-generic/vmlinux.lds.h>
 #include <asm/cache.h>
@@ -124,6 +132,8 @@ SECTIONS
 	RO_DATA(PAGE_SIZE)
 
 #ifdef CONFIG_PPC64
+	RESTART_TABLE(8)
+
 	. = ALIGN(8);
 	__stf_entry_barrier_fixup : AT(ADDR(__stf_entry_barrier_fixup) - LOAD_OFFSET) {
 		__start___stf_entry_barrier_fixup = .;
arch/powerpc/lib/Makefile +1 -1
index cc1a8a0f311e..4c92c80454f3 100644
@@ -42,7 +42,7 @@ obj-$(CONFIG_PPC_BOOK3S_64) += copyuser_power7.o copypage_power7.o \
 			       memcpy_power7.o
 
 obj64-y	+= copypage_64.o copyuser_64.o mem_64.o hweight_64.o \
-	   memcpy_64.o copy_mc_64.o
+	   memcpy_64.o copy_mc_64.o restart_table.o
 
 ifndef CONFIG_PPC_QUEUED_SPINLOCKS
 obj64-$(CONFIG_SMP)	+= locks.o
arch/powerpc/lib/restart_table.c new file +30 -0
new file mode 100644
index 000000000000..7cd20757cc33
@@ -0,0 +1,30 @@
+#include <asm/interrupt.h>
+#include <asm/kprobes.h>
+
+struct restart_table_entry {
+	unsigned long start;
+	unsigned long end;
+	unsigned long fixup;
+};
+
+extern struct restart_table_entry __start___restart_table[];
+extern struct restart_table_entry __stop___restart_table[];
+
+/* Given an address, look for it in the kernel exception table */
+unsigned long search_kernel_restart_table(unsigned long addr)
+{
+	struct restart_table_entry *rte = __start___restart_table;
+
+	while (rte < __stop___restart_table) {
+		unsigned long start = rte->start;
+		unsigned long end = rte->end;
+		unsigned long fixup = rte->fixup;
+
+		if (addr >= start && addr < end)
+			return fixup;
+
+		rte++;
+	}
+	return 0;
+}
+NOKPROBE_SYMBOL(search_kernel_restart_table);
Signature · 1 line
2.23.0
author reviewer
#10 lore raw

PATCH v4 10/17 powerpc/64: interrupt soft-enable race fix

In reply to Nicholas Piggin

Prevent interrupt restore from allowing racing hard interrupts going ahead of previous soft-pending ones, by using the soft-masked restart handler to allow a store to clear the soft-mask while knowing nothing is soft-pending.

This probably doesn't matter much in practice, but it's a simple demonstrator / test case to exercise the restart table logic.

---
 arch/powerpc/kernel/irq.c | 95 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 95 insertions(+)
arch/powerpc/kernel/irq.c +95 -0
index 72cb45393ef2..8428caf3194e 100644
@@ -217,6 +217,100 @@ static inline void replay_soft_interrupts_irqrestore(void)
 #define replay_soft_interrupts_irqrestore() replay_soft_interrupts()
 #endif
 
+#ifdef CONFIG_CC_HAS_ASM_GOTO
+notrace void arch_local_irq_restore(unsigned long mask)
+{
+	unsigned char irq_happened;
+
+	/* Write the new soft-enabled value if it is a disable */
+	if (mask) {
+		irq_soft_mask_set(mask);
+		return;
+	}
+
+	/*
+	 * After the stb, interrupts are unmasked and there are no interrupts
+	 * pending replay. The restart sequence makes this atomic with
+	 * respect to soft-masked interrupts. If this was just a simple code
+	 * sequence, a soft-masked interrupt could become pending right after
+	 * the comparison and before the stb.
+	 *
+	 * This allows interrupts to be unmasked without hard disabling, and
+	 * also without new hard interrupts coming in ahead of pending ones.
+	 */
+	asm_volatile_goto(
+"1:					\n"
+"		lbz	9,%0(13)	\n"
+"		cmpwi	9,0		\n"
+"		bne	%l[happened]	\n"
+"		stb	9,%1(13)	\n"
+"2:					\n"
+		RESTART_TABLE(1b, 2b, 1b)
+	: : "i" (offsetof(struct paca_struct, irq_happened)),
+	    "i" (offsetof(struct paca_struct, irq_soft_mask))
+	: "cr0", "r9"
+	: happened);
+
+	if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG))
+		WARN_ON_ONCE(!(mfmsr() & MSR_EE));
+
+	return;
+
+happened:
+	irq_happened = get_irq_happened();
+	if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG))
+		WARN_ON_ONCE(!irq_happened);
+
+	if (irq_happened == PACA_IRQ_HARD_DIS) {
+		if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG))
+			WARN_ON_ONCE(mfmsr() & MSR_EE);
+		irq_soft_mask_set(IRQS_ENABLED);
+		local_paca->irq_happened = 0;
+		__hard_irq_enable();
+		return;
+	}
+
+	/* Have interrupts to replay, need to hard disable first */
+	if (!(irq_happened & PACA_IRQ_HARD_DIS)) {
+		if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG)) {
+			if (!(mfmsr() & MSR_EE)) {
+				/*
+				 * An interrupt could have come in and cleared
+				 * MSR[EE] and set IRQ_HARD_DIS, so check
+				 * IRQ_HARD_DIS again and warn if it is still
+				 * clear.
+				 */
+				irq_happened = get_irq_happened();
+				WARN_ON_ONCE(!(irq_happened & PACA_IRQ_HARD_DIS));
+			}
+		}
+		__hard_irq_disable();
+		local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
+	} else {
+		if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG)) {
+			if (WARN_ON_ONCE(mfmsr() & MSR_EE))
+				__hard_irq_disable();
+		}
+	}
+
+	/*
+	 * Disable preempt here, so that the below preempt_enable will
+	 * perform resched if required (a replayed interrupt may set
+	 * need_resched).
+	 */
+	preempt_disable();
+	irq_soft_mask_set(IRQS_ALL_DISABLED);
+	trace_hardirqs_off();
+
+	replay_soft_interrupts_irqrestore();
+	local_paca->irq_happened = 0;
+
+	trace_hardirqs_on();
+	irq_soft_mask_set(IRQS_ENABLED);
+	__hard_irq_enable();
+	preempt_enable();
+}
+#else
 notrace void arch_local_irq_restore(unsigned long mask)
 {
 	unsigned char irq_happened;
@@ -288,6 +382,7 @@ notrace void arch_local_irq_restore(unsigned long mask)
 	__hard_irq_enable();
 	preempt_enable();
 }
+#endif
 EXPORT_SYMBOL(arch_local_irq_restore);
 
 /*
Signature · 1 line
2.23.0
author reviewer
#12 lore raw

PATCH v4 11/17 powerpc/64: treat low kernel text as irqs soft-masked

In reply to Nicholas Piggin

Treat code below __end_soft_masked as soft-masked for the purpose of alternate return. 64s already mostly does this for scv entry.

This will be used to exit from interrupts without disabling MSR[EE].

---
 arch/powerpc/include/asm/interrupt.h | 11 ++++++++---
 arch/powerpc/kernel/exceptions-64e.S | 12 +++++++++++-
 arch/powerpc/kernel/exceptions-64s.S | 19 +++++++++++--------
 arch/powerpc/kernel/interrupt_64.S   |  6 +++++-
 4 files changed, 35 insertions(+), 13 deletions(-)
arch/powerpc/include/asm/interrupt.h +8 -3
index 49d9a6fd1bb9..88043e46442a 100644
@@ -158,8 +158,13 @@ static inline void interrupt_enter_prepare(struct pt_regs *regs, struct interrup
 		 * CT_WARN_ON comes here via program_check_exception,
 		 * so avoid recursion.
 		 */
-		if (TRAP(regs) != INTERRUPT_PROGRAM)
+		if (TRAP(regs) != INTERRUPT_PROGRAM) {
 			CT_WARN_ON(ct_state() != CONTEXT_KERNEL);
+			BUG_ON(regs->nip < (unsigned long)__end_soft_masked);
+		}
+		/* Move this under a debugging check */
+		if (arch_irq_disabled_regs(regs))
+			BUG_ON(search_kernel_restart_table(regs->nip));
 	}
 #endif
 
@@ -254,8 +259,8 @@ static inline void interrupt_nmi_enter_prepare(struct pt_regs *regs, struct inte
 	local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
 
 	if (IS_ENABLED(CONFIG_PPC_BOOK3S_64) && !(regs->msr & MSR_PR) &&
-				regs->nip < (unsigned long)__end_interrupts) {
-		// Kernel code running below __end_interrupts is
+				regs->nip < (unsigned long)__end_soft_masked) {
+		// Kernel code running below __end_soft_masked is
 		// implicitly soft-masked.
 		regs->softe = IRQS_ALL_DISABLED;
 	}
arch/powerpc/kernel/exceptions-64e.S +11 -1
index 1b79f8a75298..22fcd95dd8dc 100644
@@ -342,7 +342,17 @@ ret_from_mc_except:
 #define PROLOG_ADDITION_MASKABLE_GEN(n)					    \
 	lbz	r10,PACAIRQSOFTMASK(r13);	/* are irqs soft-masked? */ \
 	andi.	r10,r10,IRQS_DISABLED;	/* yes -> go out of line */ \
-	bne	masked_interrupt_book3e_##n
+	bne	masked_interrupt_book3e_##n;				    \
+	/* Kernel code below __end_soft_masked is implicitly masked */	    \
+	andi.	r10,r11,MSR_PR;						    \
+	bne	1f;			/* user -> not masked */	    \
+	std	r14,PACA_EXGEN+EX_R14(r13);				    \
+	LOAD_REG_IMMEDIATE_SYM(r14, r10, __end_soft_masked);		    \
+	mfspr	r10,SPRN_SRR0;						    \
+	cmpld	r10,r14;						    \
+	ld	r14,PACA_EXGEN+EX_R14(r13);				    \
+	blt	masked_interrupt_book3e_##n;				    \
+1:
 
 /*
  * Additional regs must be re-loaded from paca before EXCEPTION_COMMON* is
arch/powerpc/kernel/exceptions-64s.S +11 -8
index 17a213f25c92..2d980addc88c 100644
@@ -508,10 +508,13 @@ DEFINE_FIXED_SYMBOL(\name\()_common_real)
 		andi.	r10,r12,MSR_PR
 		bne	2f
 
-		/* Kernel code running below __end_interrupts is implicitly
-		 * soft-masked */
-		LOAD_HANDLER(r10, __end_interrupts)
+		/*
+		 * Kernel code running below __end_soft_masked is implicitly
+		 * soft-masked
+		 */
+		LOAD_HANDLER(r10, __end_soft_masked)
 		cmpld	r11,r10
+
 		li	r10,IMASK
 		blt-	1f
 
@@ -824,17 +827,17 @@ __start_interrupts:
  * scv instructions enter the kernel without changing EE, RI, ME, or HV.
  * In particular, this means we can take a maskable interrupt at any point
  * in the scv handler, which is unlike any other interrupt. This is solved
- * by treating the instruction addresses below __end_interrupts as being
+ * by treating the instruction addresses below __end_soft_masked as being
  * soft-masked.
  *
  * AIL-0 mode scv exceptions go to 0x17000-0x17fff, but we set AIL-3 and
  * ensure scv is never executed with relocation off, which means AIL-0
  * should never happen.
  *
- * Before leaving the below __end_interrupts text, at least of the following
- * must be true:
+ * Before leaving the following inside-__end_soft_masked text, at least of the
+ * following must be true:
  * - MSR[PR]=1 (i.e., return to userspace)
- * - MSR_EE|MSR_RI is set (no reentrant exceptions)
+ * - MSR_EE|MSR_RI is clear (no reentrant exceptions)
  * - Standard kernel environment is set up (stack, paca, etc)
  *
  * Call convention:
@@ -3099,7 +3102,7 @@ kvmppc_skip_Hinterrupt:
 
 USE_FIXED_SECTION(virt_trampolines)
 	/*
-	 * All code below __end_interrupts is treated as soft-masked. If
+	 * All code below __end_soft_masked is treated as soft-masked. If
 	 * any code runs here with MSR[EE]=1, it must then cope with pending
 	 * soft interrupt being raised (i.e., by ensuring it is replayed).
 	 *
arch/powerpc/kernel/interrupt_64.S +5 -1
index ab6b99609d0e..33ee83791487 100644
@@ -632,4 +632,8 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_STCX_CHECKS_ADDRESS)
 interrupt_return_macro srr
 #ifdef CONFIG_PPC_BOOK3S
 interrupt_return_macro hsrr
-#endif
+#endif /* CONFIG_PPC_BOOK3S */
+
+	.globl __end_soft_masked
+__end_soft_masked:
+DEFINE_FIXED_SYMBOL(__end_soft_masked)
Signature · 1 line
2.23.0
author reviewer
#13 lore raw

PATCH v4 12/17 powerpc/64: use interrupt restart table to speed up return from interrupt

In reply to Nicholas Piggin

Use the restart table facility to return from interrupt or system calls without disabling MSR[EE] or MSR[RI].

Interrupt return asm is put into the low soft-masked region, to prevent interrupts being processed here, although they are still taken as masked interrupts which causes SRRs to be clobbered, and a pending soft-masked interrupt to require replaying.

The return code uses restart table regions to redirct to a fixup handler rather than continue with the exit, if such an interrupt happens. In this case the interrupt return is redirected to a fixup handler which reloads r1 for the interrupt stack and reloads registers and sets state up to replay the soft-masked interrupt and try the exit again.

Some types of security exit fallback flushes and barriers are currently unable to cope with reentrant interrupts, e.g., because they store some state in the scratch SPR which would be clobbered even by masked interrupts. For now the interrupts-enabled exits are disabled when these flushes are used.

---
 arch/powerpc/include/asm/asm-prototypes.h |   5 +
 arch/powerpc/include/asm/hw_irq.h         |  13 +-
 arch/powerpc/include/asm/interrupt.h      |   2 +
 arch/powerpc/include/asm/paca.h           |   3 +
 arch/powerpc/include/asm/ptrace.h         |   1 +
 arch/powerpc/kernel/asm-offsets.c         |   3 +
 arch/powerpc/kernel/interrupt.c           | 393 ++++++++++++++--------
 arch/powerpc/kernel/interrupt_64.S        | 143 ++++++--
 arch/powerpc/lib/feature-fixups.c         |  52 ++-
 9 files changed, 457 insertions(+), 158 deletions(-)
arch/powerpc/include/asm/asm-prototypes.h +5 -0
index 95492655462e..89765b2b3be0 100644
@@ -73,6 +73,11 @@ long system_call_exception(long r3, long r4, long r5, long r6, long r7, long r8,
 notrace unsigned long syscall_exit_prepare(unsigned long r3, struct pt_regs *regs, long scv);
 notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs);
 notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs);
+#ifdef CONFIG_PPC64
+unsigned long syscall_exit_restart(unsigned long r3, struct pt_regs *regs);
+unsigned long interrupt_exit_user_restart(struct pt_regs *regs);
+unsigned long interrupt_exit_kernel_restart(struct pt_regs *regs);
+#endif
 
 long ppc_fadvise64_64(int fd, int advice, u32 offset_high, u32 offset_low,
 		      u32 len_high, u32 len_low);
arch/powerpc/include/asm/hw_irq.h +11 -2
index 19bcef666cf6..21cc571ea9c2 100644
@@ -18,8 +18,17 @@
  * PACA flags in paca->irq_happened.
  *
  * This bits are set when interrupts occur while soft-disabled
- * and allow a proper replay. Additionally, PACA_IRQ_HARD_DIS
- * is set whenever we manually hard disable.
+ * and allow a proper replay.
+ *
+ * The PACA_IRQ_HARD_DIS is set whenever we hard disable. It is almost
+ * always in synch with the MSR[EE] state, except:
+ * - A window in interrupt entry, where hardware disables MSR[EE] and that
+ *   must be "reconciled" with the soft mask state.
+ * - NMI interrupts that hit in awkward places, until they fix the state.
+ * - When local irqs are being enabled and state is being fixed up.
+ * - When returning from an interrupt there are some windows where this
+ *   can become out of synch, but gets fixed before the RFI or before
+ *   executing the next user instruction (see arch/powerpc/kernel/interrupt.c).
  */
 #define PACA_IRQ_HARD_DIS	0x01
 #define PACA_IRQ_DBELL		0x02
arch/powerpc/include/asm/interrupt.h +2 -0
index 88043e46442a..c7b4adbe2dfe 100644
@@ -79,6 +79,8 @@ unsigned long search_kernel_restart_table(unsigned long addr);
 #endif
 
 #ifdef CONFIG_PPC_BOOK3S_64
+DECLARE_STATIC_KEY_FALSE(interrupt_exit_not_reentrant);
+
 static inline void srr_regs_clobbered(void)
 {
 	local_paca->srr_valid = 0;
arch/powerpc/include/asm/paca.h +3 -0
index dfc984b0e640..ca0b52e107a8 100644
@@ -167,6 +167,9 @@ struct paca_struct {
 	u64 kstack;			/* Saved Kernel stack addr */
 	u64 saved_r1;			/* r1 save for RTAS calls or PM or EE=0 */
 	u64 saved_msr;			/* MSR saved here by enter_rtas */
+#ifdef CONFIG_PPC64
+	u64 exit_save_r1;		/* Syscall/interrupt R1 save */
+#endif
 #ifdef CONFIG_PPC_BOOK3E
 	u16 trap_save;			/* Used when bad stack is encountered */
 #endif
arch/powerpc/include/asm/ptrace.h +1 -0
index 516117bba4e6..cb2b093dee13 100644
@@ -52,6 +52,7 @@ struct pt_regs
 		struct {
 #ifdef CONFIG_PPC64
 			unsigned long ppr;
+			unsigned long exit_result;
 #endif
 			union {
 #ifdef CONFIG_PPC_KUAP
arch/powerpc/kernel/asm-offsets.c +3 -0
index ebe9afef619f..f90a8876e622 100644
@@ -286,6 +286,9 @@ int main(void)
 	OFFSET(PACAHWCPUID, paca_struct, hw_cpu_id);
 	OFFSET(PACAKEXECSTATE, paca_struct, kexec_state);
 	OFFSET(PACA_DSCR_DEFAULT, paca_struct, dscr_default);
+#ifdef CONFIG_PPC64
+	OFFSET(PACA_EXIT_SAVE_R1, paca_struct, exit_save_r1);
+#endif
 #ifdef CONFIG_PPC_BOOK3E
 	OFFSET(PACA_TRAP_SAVE, paca_struct, trap_save);
 #endif
arch/powerpc/kernel/interrupt.c +261 -132

465 lines not on this page · raw message

arch/powerpc/kernel/interrupt_64.S +123 -20

235 lines not on this page · raw message

arch/powerpc/lib/feature-fixups.c +48 -4

85 lines not on this page · raw message

Signature · 1 line
2.23.0
author reviewer
#14 lore raw

PATCH v4 13/17 powerpc/interrupt: Rename and lightly change syscall_exit_prepare_main()

In reply to Nicholas Piggin

From: Christophe Leroy <[email protected]>

Rename syscall_exit_prepare_main() into interrupt_exit_prepare_main()

Pass it the 'ret' so that it can 'or' it directly instead of oring twice, once inside the function and once outside.

And remove 'r3' parameter which is not used.

[np: split out some changes into other patches]
---
 arch/powerpc/kernel/interrupt.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
arch/powerpc/kernel/interrupt.c +4 -5
index 6dc9b7536511..5bfdf8f9b130 100644
@@ -243,11 +243,10 @@ static notrace void booke_load_dbcr0(void)
 #endif
 }
 
-static notrace unsigned long syscall_exit_prepare_main(unsigned long r3,
-						       struct pt_regs *regs)
+static notrace unsigned long
+interrupt_exit_user_prepare_main(unsigned long ret, struct pt_regs *regs)
 {
 	unsigned long ti_flags;
-	unsigned long ret = 0;
 
 again:
 	ti_flags = READ_ONCE(current_thread_info()->flags);
@@ -365,7 +364,7 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3,
 	}
 
 	local_irq_disable();
-	ret |= syscall_exit_prepare_main(r3, regs);
+	ret = interrupt_exit_user_prepare_main(ret, regs);
 
 #ifdef CONFIG_PPC64
 	regs->exit_result = ret;
@@ -397,7 +396,7 @@ notrace unsigned long syscall_exit_restart(unsigned long r3, struct pt_regs *reg
 
 	BUG_ON(!user_mode(regs));
 
-	regs->exit_result |= syscall_exit_prepare_main(r3, regs);
+	regs->exit_result = interrupt_exit_user_prepare_main(regs->exit_result, regs);
 
 	return regs->exit_result;
 }
Signature · 1 line
2.23.0
author reviewer
#15 lore raw

PATCH v4 14/17 powerpc/interrupt: Refactor interrupt_exit_user_prepare()

In reply to Nicholas Piggin

From: Christophe Leroy <[email protected]>

interrupt_exit_user_prepare() is a superset of
interrupt_exit_user_prepare_main().

Refactor to avoid code duplication.

---
 arch/powerpc/kernel/interrupt.c | 57 ++-------------------------------
 1 file changed, 3 insertions(+), 54 deletions(-)
arch/powerpc/kernel/interrupt.c +3 -54
index 5bfdf8f9b130..e946084d4906 100644
@@ -404,9 +404,7 @@ notrace unsigned long syscall_exit_restart(unsigned long r3, struct pt_regs *reg
 
 notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs)
 {
-	unsigned long ti_flags;
-	unsigned long flags;
-	unsigned long ret = 0;
+	unsigned long ret;
 
 	if (!IS_ENABLED(CONFIG_BOOKE) && !IS_ENABLED(CONFIG_40x))
 		BUG_ON(!(regs->msr & MSR_RI));
@@ -420,63 +418,14 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs)
 	 */
 	kuap_assert_locked();
 
-	local_irq_save(flags);
-
-again:
-	ti_flags = READ_ONCE(current_thread_info()->flags);
-	while (unlikely(ti_flags & (_TIF_USER_WORK_MASK & ~_TIF_RESTORE_TM))) {
-		local_irq_enable(); /* returning to user: may enable */
-		if (ti_flags & _TIF_NEED_RESCHED) {
-			schedule();
-		} else {
-			if (ti_flags & _TIF_SIGPENDING)
-				ret |= _TIF_RESTOREALL;
-			do_notify_resume(regs, ti_flags);
-		}
-		local_irq_disable();
-		ti_flags = READ_ONCE(current_thread_info()->flags);
-	}
-
-	if (IS_ENABLED(CONFIG_PPC_BOOK3S_64) && IS_ENABLED(CONFIG_PPC_FPU)) {
-		if (IS_ENABLED(CONFIG_PPC_TRANSACTIONAL_MEM) &&
-				unlikely((ti_flags & _TIF_RESTORE_TM))) {
-			restore_tm_state(regs);
-		} else {
-			unsigned long mathflags = MSR_FP;
-
-			if (cpu_has_feature(CPU_FTR_VSX))
-				mathflags |= MSR_VEC | MSR_VSX;
-			else if (cpu_has_feature(CPU_FTR_ALTIVEC))
-				mathflags |= MSR_VEC;
-
-			/* See above restore_math comment */
-			if ((regs->msr & mathflags) != mathflags)
-				restore_math(regs);
-		}
-	}
-
-	if (!prep_irq_for_user_exit()) {
-		local_irq_enable();
-		local_irq_disable();
-		goto again;
-	}
-
-	booke_load_dbcr0();
-
-#ifdef CONFIG_PPC_TRANSACTIONAL_MEM
-	local_paca->tm_scratch = regs->msr;
-#endif
+	local_irq_disable();
 
-	account_cpu_user_exit();
+	ret = interrupt_exit_user_prepare_main(0, regs);
 
 #ifdef CONFIG_PPC64
 	regs->exit_result = ret;
 #endif
 
-	/* Restore user access locks last */
-	kuap_user_restore(regs);
-	kuep_unlock();
-
 	return ret;
 }
 
Signature · 1 line
2.23.0
author reviewer
#16 lore raw

PATCH v4 15/17 powerpc/interrupt: Interchange prep_irq_for_{kernel_enabled/user}_exit()

In reply to Nicholas Piggin

From: Christophe Leroy <[email protected]>

prep_irq_for_user_exit() is a superset of prep_irq_for_kernel_enabled_exit(). In order to allow refactoring in following patch, interchange the two. This will allow prep_irq_for_user_exit() to call a renamed version of prep_irq_for_kernel_enabled_exit().

---
 arch/powerpc/kernel/interrupt.c | 23 +++++++++++------------
 1 file changed, 11 insertions(+), 12 deletions(-)
arch/powerpc/kernel/interrupt.c +11 -12
index e946084d4906..6ba693f99e5a 100644
@@ -46,27 +46,28 @@ static inline bool exit_must_hard_disable(void)
  * This should be called with local irqs disabled, but if they were previously
  * enabled when the interrupt handler returns (indicating a process-context /
  * synchronous interrupt) then irqs_enabled should be true.
+ *
+ * restartable is true then EE/RI can be left on because interrupts are handled
+ * with a restart sequence.
  */
-static notrace __always_inline bool prep_irq_for_user_exit(void)
+static notrace __always_inline bool prep_irq_for_kernel_enabled_exit(bool restartable)
 {
-	user_enter_irqoff();
 	/* This must be done with RI=1 because tracing may touch vmaps */
 	trace_hardirqs_on();
 
 #ifdef CONFIG_PPC32
 	__hard_EE_RI_disable();
 #else
-	if (exit_must_hard_disable())
+	if (exit_must_hard_disable() || !restartable)
 		__hard_EE_RI_disable();
 
 	/* This pattern matches prep_irq_for_idle */
 	if (unlikely(lazy_irq_pending_nocheck())) {
-		if (exit_must_hard_disable()) {
+		if (exit_must_hard_disable() || !restartable) {
 			local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
 			__hard_RI_enable();
 		}
 		trace_hardirqs_off();
-		user_exit_irqoff();
 
 		return false;
 	}
@@ -74,28 +75,26 @@ static notrace __always_inline bool prep_irq_for_user_exit(void)
 	return true;
 }
 
-/*
- * restartable is true then EE/RI can be left on because interrupts are handled
- * with a restart sequence.
- */
-static notrace __always_inline bool prep_irq_for_kernel_enabled_exit(bool restartable)
+static notrace __always_inline bool prep_irq_for_user_exit(void)
 {
+	user_enter_irqoff();
 	/* This must be done with RI=1 because tracing may touch vmaps */
 	trace_hardirqs_on();
 
 #ifdef CONFIG_PPC32
 	__hard_EE_RI_disable();
 #else
-	if (exit_must_hard_disable() || !restartable)
+	if (exit_must_hard_disable())
 		__hard_EE_RI_disable();
 
 	/* This pattern matches prep_irq_for_idle */
 	if (unlikely(lazy_irq_pending_nocheck())) {
-		if (exit_must_hard_disable() || !restartable) {
+		if (exit_must_hard_disable()) {
 			local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
 			__hard_RI_enable();
 		}
 		trace_hardirqs_off();
+		user_exit_irqoff();
 
 		return false;
 	}
Signature · 1 line
2.23.0
author reviewer
#17 lore raw

PATCH v4 16/17 powerpc/interrupt: Refactor prep_irq_for_{user/kernel_enabled}_exit()

In reply to Nicholas Piggin

From: Christophe Leroy <[email protected]>

prep_irq_for_user_exit() is a superset of
prep_irq_for_kernel_enabled_exit().

Rename prep_irq_for_kernel_enabled_exit() as prep_irq_for_enabled_exit() and have prep_irq_for_user_exit() use it.

---
 arch/powerpc/kernel/interrupt.c | 29 +++++++----------------------
 1 file changed, 7 insertions(+), 22 deletions(-)
arch/powerpc/kernel/interrupt.c +7 -22
index 6ba693f99e5a..138c450b01bd 100644
@@ -50,7 +50,7 @@ static inline bool exit_must_hard_disable(void)
  * restartable is true then EE/RI can be left on because interrupts are handled
  * with a restart sequence.
  */
-static notrace __always_inline bool prep_irq_for_kernel_enabled_exit(bool restartable)
+static notrace __always_inline bool prep_irq_for_enabled_exit(bool restartable)
 {
 	/* This must be done with RI=1 because tracing may touch vmaps */
 	trace_hardirqs_on();
@@ -77,29 +77,14 @@ static notrace __always_inline bool prep_irq_for_kernel_enabled_exit(bool restar
 
 static notrace __always_inline bool prep_irq_for_user_exit(void)
 {
-	user_enter_irqoff();
-	/* This must be done with RI=1 because tracing may touch vmaps */
-	trace_hardirqs_on();
-
-#ifdef CONFIG_PPC32
-	__hard_EE_RI_disable();
-#else
-	if (exit_must_hard_disable())
-		__hard_EE_RI_disable();
+	bool ret;
 
-	/* This pattern matches prep_irq_for_idle */
-	if (unlikely(lazy_irq_pending_nocheck())) {
-		if (exit_must_hard_disable()) {
-			local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
-			__hard_RI_enable();
-		}
-		trace_hardirqs_off();
+	user_enter_irqoff();
+	ret = prep_irq_for_enabled_exit(true);
+	if (!ret)
 		user_exit_irqoff();
 
-		return false;
-	}
-#endif
-	return true;
+	return ret;
 }
 
 /* Has to run notrace because it is entered not completely "reconciled" */
@@ -469,7 +454,7 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs)
 		 * Stack store exit can't be restarted because the interrupt
 		 * stack frame might have been clobbered.
 		 */
-		if (!prep_irq_for_kernel_enabled_exit(unlikely(stack_store))) {
+		if (!prep_irq_for_enabled_exit(unlikely(stack_store))) {
 			/*
 			 * Replay pending soft-masked interrupts now. Don't
 			 * just local_irq_enabe(); local_irq_disable(); because
Signature · 1 line
2.23.0
author reviewer
#18 lore raw

PATCH v4 17/17 powerpc/interrupt: Remove prep_irq_for_user_exit()

In reply to Nicholas Piggin

From: Christophe Leroy <[email protected]>

prep_irq_for_user_exit() has only one caller, squash it inside that caller.

---
 arch/powerpc/kernel/interrupt.c | 16 +++-------------
 1 file changed, 3 insertions(+), 13 deletions(-)
arch/powerpc/kernel/interrupt.c +3 -13
index 138c450b01bd..09b8304a7011 100644
@@ -75,18 +75,6 @@ static notrace __always_inline bool prep_irq_for_enabled_exit(bool restartable)
 	return true;
 }
 
-static notrace __always_inline bool prep_irq_for_user_exit(void)
-{
-	bool ret;
-
-	user_enter_irqoff();
-	ret = prep_irq_for_enabled_exit(true);
-	if (!ret)
-		user_exit_irqoff();
-
-	return ret;
-}
-
 /* Has to run notrace because it is entered not completely "reconciled" */
 notrace long system_call_exception(long r3, long r4, long r5,
 				   long r6, long r7, long r8,
@@ -276,7 +264,9 @@ interrupt_exit_user_prepare_main(unsigned long ret, struct pt_regs *regs)
 		}
 	}
 
-	if (!prep_irq_for_user_exit()) {
+	user_enter_irqoff();
+	if (!prep_irq_for_enabled_exit(true)) {
+		user_exit_irqoff();
 		local_irq_enable();
 		local_irq_disable();
 		goto again;
Signature · 1 line
2.23.0
#19 lore raw
In reply to Nicholas Piggin
Quoted 9 lines · Nicholas Piggin

This series attempts to improve the speed of interrupts and system calls in three major ways.

Firstly, the SRR/HSRR registers do not need to be reloaded if they were clobbered for the duration of the interrupt and the return NIP and MSR did not changed. 64e does not implement this part, but it could quite easily.

[...]

Patches 1-4, and 6-17 applied to powerpc/next.

[01/17] powerpc/interrupt: Fix CONFIG ifdef typo

        https://git.kernel.org/powerpc/c/9a3ed7adcabce24a85fbe05f54e762b18756ec22
[02/17] powerpc: remove interrupt exit helpers unused argument
        https://git.kernel.org/powerpc/c/bf9155f1970c4dbf9ec6b87d3688433bd494a4e1
[03/17] powerpc/64s: introduce different functions to return from SRR vs HSRR interrupts
        https://git.kernel.org/powerpc/c/1df7d5e4baeac74d14c1bee18b2dff9302b3efbc
[04/17] powerpc/64s: avoid reloading (H)SRR registers if they are still valid
        https://git.kernel.org/powerpc/c/59dc5bfca0cb6a29db1a50847684eb5c19f8f400
[06/17] powerpc/64: move interrupt return asm to interrupt_64.S
        https://git.kernel.org/powerpc/c/e754f4d13e3919aafa485657599907aa63b9a40c
[07/17] powerpc/64s: system call avoid setting MSR[RI] until we set MSR[EE]
        https://git.kernel.org/powerpc/c/dd152f70bdc1b91445b10c65ac874b90c93fb3b5
[08/17] powerpc/64s: save one more register in the masked interrupt handler
        https://git.kernel.org/powerpc/c/63e40806eea984f770c992120bbfd71b589ea580
[09/17] powerpc/64: allow alternate return locations for soft-masked interrupts
        https://git.kernel.org/powerpc/c/f23699c93becd746295aaa506537882a46a62219
[10/17] powerpc/64: interrupt soft-enable race fix
        https://git.kernel.org/powerpc/c/862fa563524b9f92d7e89fe332732bd3421772db
[11/17] powerpc/64: treat low kernel text as irqs soft-masked
        https://git.kernel.org/powerpc/c/9d1988ca87dd90ecf80a0601c7fd13071fbb1a83
[12/17] powerpc/64: use interrupt restart table to speed up return from interrupt
        https://git.kernel.org/powerpc/c/13799748b957bc5659f97c036224b0f4b42172e2
[13/17] powerpc/interrupt: Rename and lightly change syscall_exit_prepare_main()
        https://git.kernel.org/powerpc/c/f84aa284947f325c5697d35b92abd2047224f24b
[14/17] powerpc/interrupt: Refactor interrupt_exit_user_prepare()
        https://git.kernel.org/powerpc/c/a214ee8802adb864d175ea6ca4176223bcc11d2b
[15/17] powerpc/interrupt: Interchange prep_irq_for_{kernel_enabled/user}_exit()
        https://git.kernel.org/powerpc/c/99f98f849cf13e5fac532979ccdb77dff07665db
[16/17] powerpc/interrupt: Refactor prep_irq_for_{user/kernel_enabled}_exit()
        https://git.kernel.org/powerpc/c/61eece2d1707796fd45225ea3d20e9289251311c
[17/17] powerpc/interrupt: Remove prep_irq_for_user_exit()
        https://git.kernel.org/powerpc/c/ae58b1c645895c28ca155843db6788d57ea99e11

cheers

#11 lore raw

Re: [PATCH v4 10/17] powerpc/64: interrupt soft-enable race fix

In reply to Nicholas Piggin

Hi Nick.

Le 17/06/2021 à 17:51, Nicholas Piggin a écrit :

Quoted 18 lines

Prevent interrupt restore from allowing racing hard interrupts going ahead of previous soft-pending ones, by using the soft-masked restart handler to allow a store to clear the soft-mask while knowing nothing is soft-pending.

This probably doesn't matter much in practice, but it's a simple demonstrator / test case to exercise the restart table logic.

---
  arch/powerpc/kernel/irq.c | 95 +++++++++++++++++++++++++++++++++++++++
  1 file changed, 95 insertions(+)
arch/powerpc/kernel/irq.c +0 -0
index 72cb45393ef2..8428caf3194e 100644

arch/powerpc/kernel/irq.c static inline void replay_soft_interrupts_irqrestore(void) · line 220

Quoted 4 lines
arch/powerpc/kernel/irq.c +1 -0
@@ -217,6 +217,100 @@ static inline void replay_soft_interrupts_irqrestore(void)
  #define replay_soft_interrupts_irqrestore() replay_soft_interrupts()
  #endif
 
+#ifdef CONFIG_CC_HAS_ASM_GOTO

Wondering why you added that #ifdef CONFIG_CC_HAS_ASM_GOTO ? In 2021 the minimum GCC version was already supporting asm goto and we were already using asm goto unconditionnaly in uaccess.

I just sent a patch to remove it.

Quoted 104 lines
arch/powerpc/kernel/irq.c +94 -0
+notrace void arch_local_irq_restore(unsigned long mask)
+{
+	unsigned char irq_happened;
+
+	/* Write the new soft-enabled value if it is a disable */
+	if (mask) {
+		irq_soft_mask_set(mask);
+		return;
+	}
+
+	/*
+	 * After the stb, interrupts are unmasked and there are no interrupts
+	 * pending replay. The restart sequence makes this atomic with
+	 * respect to soft-masked interrupts. If this was just a simple code
+	 * sequence, a soft-masked interrupt could become pending right after
+	 * the comparison and before the stb.
+	 *
+	 * This allows interrupts to be unmasked without hard disabling, and
+	 * also without new hard interrupts coming in ahead of pending ones.
+	 */
+	asm_volatile_goto(
+"1:					\n"
+"		lbz	9,%0(13)	\n"
+"		cmpwi	9,0		\n"
+"		bne	%l[happened]	\n"
+"		stb	9,%1(13)	\n"
+"2:					\n"
+		RESTART_TABLE(1b, 2b, 1b)
+	: : "i" (offsetof(struct paca_struct, irq_happened)),
+	    "i" (offsetof(struct paca_struct, irq_soft_mask))
+	: "cr0", "r9"
+	: happened);
+
+	if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG))
+		WARN_ON_ONCE(!(mfmsr() & MSR_EE));
+
+	return;
+
+happened:
+	irq_happened = get_irq_happened();
+	if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG))
+		WARN_ON_ONCE(!irq_happened);
+
+	if (irq_happened == PACA_IRQ_HARD_DIS) {
+		if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG))
+			WARN_ON_ONCE(mfmsr() & MSR_EE);
+		irq_soft_mask_set(IRQS_ENABLED);
+		local_paca->irq_happened = 0;
+		__hard_irq_enable();
+		return;
+	}
+
+	/* Have interrupts to replay, need to hard disable first */
+	if (!(irq_happened & PACA_IRQ_HARD_DIS)) {
+		if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG)) {
+			if (!(mfmsr() & MSR_EE)) {
+				/*
+				 * An interrupt could have come in and cleared
+				 * MSR[EE] and set IRQ_HARD_DIS, so check
+				 * IRQ_HARD_DIS again and warn if it is still
+				 * clear.
+				 */
+				irq_happened = get_irq_happened();
+				WARN_ON_ONCE(!(irq_happened & PACA_IRQ_HARD_DIS));
+			}
+		}
+		__hard_irq_disable();
+		local_paca->irq_happened |= PACA_IRQ_HARD_DIS;
+	} else {
+		if (IS_ENABLED(CONFIG_PPC_IRQ_SOFT_MASK_DEBUG)) {
+			if (WARN_ON_ONCE(mfmsr() & MSR_EE))
+				__hard_irq_disable();
+		}
+	}
+
+	/*
+	 * Disable preempt here, so that the below preempt_enable will
+	 * perform resched if required (a replayed interrupt may set
+	 * need_resched).
+	 */
+	preempt_disable();
+	irq_soft_mask_set(IRQS_ALL_DISABLED);
+	trace_hardirqs_off();
+
+	replay_soft_interrupts_irqrestore();
+	local_paca->irq_happened = 0;
+
+	trace_hardirqs_on();
+	irq_soft_mask_set(IRQS_ENABLED);
+	__hard_irq_enable();
+	preempt_enable();
+}
+#else
  notrace void arch_local_irq_restore(unsigned long mask)
  {
  	unsigned char irq_happened;
@@ -288,6 +382,7 @@ notrace void arch_local_irq_restore(unsigned long mask)
  	__hard_irq_enable();
  	preempt_enable();
  }
+#endif
  EXPORT_SYMBOL(arch_local_irq_restore);
 
  /*