* [PATCH v3 0/3] Support Soft NMI
@ 2017-01-05 4:55 Balbir Singh
2017-01-05 4:55 ` [PATCH v3 1/3] Merge IPI and DEFAULT priorities Balbir Singh
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Balbir Singh @ 2017-01-05 4:55 UTC (permalink / raw)
To: linuxppc-dev
Cc: Michael Ellerman, Benjamin Herrenschmidt, Paul Mackerras,
Nicholas Piggin
This patch is based on suggestions from paulus and benh.
The bugs are all mine. The idea was to implement soft
NMI(s) by keeping interrupts enabled in the soft-disabled
state, but to use the interrupt controller to gate posting
of new interrupts to the processor. This is still work in
progress and a preliminary RFC that needs testing.
Nick posted a more comprehensive version for soft NMI at
https://patchwork.ozlabs.org/patch/704605/, but it does
not work when interrupts are disabled
TODOs:
1. Integrate with
https://patchwork.ozlabs.org/patch/706983/
2. Remove the duplicate XIRR between KVM and the
paca's irq field. Will do this in a follow on
series, since that requires extensive testing
Changelog v3:
- Try to address kbuild failure
- Initialize paca->irq to a sane value
Changelog v2:
- Rename functions as suggested by Nick
- Move irq in paca down for better alignment
- Break the patches up further into interrupt disabling
and soft nmi patches
- Support only PPC_XICS for now
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Nicholas Piggin <npiggin@gmail.com>
Balbir Singh (3):
Merge IPI and DEFAULT priorities
Keep interrupts enabled even on soft disable
Enable soft nmi handler
arch/powerpc/include/asm/paca.h | 1 +
arch/powerpc/include/asm/xics.h | 8 ++----
arch/powerpc/kernel/exceptions-64s.S | 33 ++++++++++++++++------
arch/powerpc/kernel/irq.c | 55 ++++++++++++++++++++++++++++++++++--
arch/powerpc/kernel/paca.c | 1 +
arch/powerpc/kernel/time.c | 27 +++++++++++++++++-
6 files changed, 108 insertions(+), 17 deletions(-)
--
2.9.3
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCH v3 1/3] Merge IPI and DEFAULT priorities 2017-01-05 4:55 [PATCH v3 0/3] Support Soft NMI Balbir Singh @ 2017-01-05 4:55 ` Balbir Singh 2017-01-05 4:55 ` [PATCH v3 2/3] Keep interrupts enabled even on soft disable Balbir Singh 2017-01-05 4:55 ` [PATCH v3 3/3] Enable soft nmi handler Balbir Singh 2 siblings, 0 replies; 6+ messages in thread From: Balbir Singh @ 2017-01-05 4:55 UTC (permalink / raw) To: linuxppc-dev Cc: Balbir Singh, Michael Ellerman, Benjamin Herrenschmidt, Paul Mackerras, Nicholas Piggin We merge IPI and DEFAULT priorities to the same value. The idea is to keep interrupts enabled even in lazy soft-disabled mode. Instead of storing the IPI and irq separately, we keep the levels same so that we deal with only one of them Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Balbir Singh <bsingharora@gmail.com> --- arch/powerpc/include/asm/xics.h | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/arch/powerpc/include/asm/xics.h b/arch/powerpc/include/asm/xics.h index f0b2385..d6cb9a0 100644 --- a/arch/powerpc/include/asm/xics.h +++ b/arch/powerpc/include/asm/xics.h @@ -14,17 +14,13 @@ /* Want a priority other than 0. Various HW issues require this. */ #define DEFAULT_PRIORITY 5 -/* - * Mark IPIs as higher priority so we can take them inside interrupts - * FIXME: still true now? - */ -#define IPI_PRIORITY 4 +#define IPI_PRIORITY 5 /* The least favored priority */ #define LOWEST_PRIORITY 0xFF /* The number of priorities defined above */ -#define MAX_NUM_PRIORITIES 3 +#define MAX_NUM_PRIORITIES 2 /* Native ICP */ #ifdef CONFIG_PPC_ICP_NATIVE -- 2.9.3 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v3 2/3] Keep interrupts enabled even on soft disable 2017-01-05 4:55 [PATCH v3 0/3] Support Soft NMI Balbir Singh 2017-01-05 4:55 ` [PATCH v3 1/3] Merge IPI and DEFAULT priorities Balbir Singh @ 2017-01-05 4:55 ` Balbir Singh 2017-01-05 5:12 ` Madhavan Srinivasan 2017-01-06 5:18 ` kbuild test robot 2017-01-05 4:55 ` [PATCH v3 3/3] Enable soft nmi handler Balbir Singh 2 siblings, 2 replies; 6+ messages in thread From: Balbir Singh @ 2017-01-05 4:55 UTC (permalink / raw) To: linuxppc-dev Cc: Balbir Singh, Michael Ellerman, Benjamin Herrenschmidt, Paul Mackerras, Nicholas Piggin This patch removes the disabling of interrupts in soft-disable mode, when interrupts are received (in lazy mode). The new scheme keeps the interrupts enabled when we receive an interrupt and does the following: When an external interrupt is received, we store the interrupt in local_paca via ppc_md.get_irq(). Later when interrupts are enabled and replayed, we reuse the stored interrupt and process it via generic_handle_irq NOTE: This works only for PPC_XICS at the moment and we'll enable it for XIVE in the future. MPIC/ OpenPIC is not supported due to the requirement that external interrupts/IPIs need unique priorities. At the cost of adding more space in the PACA, we can store multiple priorities and support more controllers, but I think we can live with supporting only XICS for now and XIVE in the future. Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Balbir Singh <bsingharora@gmail.com> --- arch/powerpc/include/asm/paca.h | 1 + arch/powerpc/kernel/exceptions-64s.S | 25 +++++++++++++--- arch/powerpc/kernel/irq.c | 55 ++++++++++++++++++++++++++++++++++-- arch/powerpc/kernel/paca.c | 1 + 4 files changed, 76 insertions(+), 6 deletions(-) diff --git a/arch/powerpc/include/asm/paca.h b/arch/powerpc/include/asm/paca.h index 6a6792b..dcbcaa6 100644 --- a/arch/powerpc/include/asm/paca.h +++ b/arch/powerpc/include/asm/paca.h @@ -163,6 +163,7 @@ struct paca_struct { #ifdef CONFIG_PPC_TRANSACTIONAL_MEM u64 tm_scratch; /* TM scratch area for reclaim */ #endif + u32 irq; /* IRQ pending */ #ifdef CONFIG_PPC_POWERNV /* Per-core mask tracking idle threads and a lock bit-[L][TTTTTTTT] */ diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S index d39d611..cf64bc4 100644 --- a/arch/powerpc/kernel/exceptions-64s.S +++ b/arch/powerpc/kernel/exceptions-64s.S @@ -1273,6 +1273,23 @@ EXC_REAL_NONE(0x1800, 0x1900) EXC_VIRT_NONE(0x5800, 0x5900) #endif +/* + * Currently we support keeping interrupts + * enabled only for XICS. We can enhance this + * as we add support for other controllers + */ +#ifdef CONFIG_PPC_XICS +#define MASKED_INTERRUPT_DISABLE(_H) \ + GET_SCRATCH0(r10); \ + std r13,PACA_EXGEN+EX_R13(r13); \ + EXCEPTION_PROLOG_PSERIES_1(handle_irq_mask, _H); +#else +#define MASKED_INTERRUPT_DISABLE(_H) \ + mfspr r10,SPRN_##_H##SRR1; \ + rldicl r10,r10,48,1; /* clear MSR_EE */ \ + rotldi r10,r10,16; \ + mtspr SPRN_##_H##SRR1,r10; +#endif /* * An interrupt came in while soft-disabled. We set paca->irq_happened, then: @@ -1287,6 +1304,7 @@ EXC_VIRT_NONE(0x5800, 0x5900) #define MASKED_INTERRUPT(_H) \ masked_##_H##interrupt: \ std r11,PACA_EXGEN+EX_R11(r13); \ + std r12,PACA_EXGEN+EX_R12(r13); \ lbz r11,PACAIRQHAPPENED(r13); \ or r11,r11,r10; \ stb r11,PACAIRQHAPPENED(r13); \ @@ -1300,10 +1318,7 @@ masked_##_H##interrupt: \ beq 2f; \ cmpwi r10,PACA_IRQ_HMI; \ beq 2f; \ - mfspr r10,SPRN_##_H##SRR1; \ - rldicl r10,r10,48,1; /* clear MSR_EE */ \ - rotldi r10,r10,16; \ - mtspr SPRN_##_H##SRR1,r10; \ + MASKED_INTERRUPT_DISABLE(_H) \ 2: mtcrf 0x80,r9; \ ld r9,PACA_EXGEN+EX_R9(r13); \ ld r10,PACA_EXGEN+EX_R10(r13); \ @@ -1321,6 +1336,8 @@ USE_FIXED_SECTION(virt_trampolines) MASKED_INTERRUPT() MASKED_INTERRUPT(H) +EXC_COMMON(handle_irq_mask, 0x500, handle_masked_irq) + #ifdef CONFIG_KVM_BOOK3S_64_HANDLER TRAMP_REAL_BEGIN(kvmppc_skip_interrupt) /* diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c index a018f5c..f34f06b 100644 --- a/arch/powerpc/kernel/irq.c +++ b/arch/powerpc/kernel/irq.c @@ -96,7 +96,6 @@ extern int tau_interrupts(int); #ifdef CONFIG_PPC64 int distribute_irqs = 1; - static inline notrace unsigned long get_irq_happened(void) { unsigned long happened; @@ -498,6 +497,54 @@ static inline void check_stack_overflow(void) #endif } +#ifdef CONFIG_PPC_XICS +static inline notrace int get_paca_irq(void) +{ + int irq; + + __asm__ __volatile__("lbz %0,%1(13)" + : "=r" (irq) : "i" (offsetof(struct paca_struct, irq))); + + return irq; +} + +static inline notrace void set_paca_irq(int irq) +{ + __asm__ __volatile__("stb %0,%1(13)" + : : "r" (irq), "i" (offsetof(struct paca_struct, irq))); +} + +void handle_masked_irq(struct pt_regs *regs) +{ + /* + * TODO: Add support for XIVE as applicable + */ + unsigned int irq; + /* + * NOTE, we don't use irq_enter/exit, otherwise + * our accounting and tracing might be incorrect. + */ + irq = ppc_md.get_irq(); + + /* + * Store away irq in PACA for replay later + */ + set_paca_irq(irq); +} + +#else + +static inline notrace int get_paca_irq(void) +{ + return -1; +} + +static inline notrace void set_paca_irq(int irq) +{ +} + +#endif /* CONFIG_PPC_XICS */ + void __do_irq(struct pt_regs *regs) { unsigned int irq; @@ -513,7 +560,11 @@ void __do_irq(struct pt_regs *regs) * * This will typically lower the interrupt line to the CPU */ - irq = ppc_md.get_irq(); + irq = get_paca_irq(); + if (irq != -1) + set_paca_irq(-1); + else + irq = ppc_md.get_irq(); /* We can hard enable interrupts now to allow perf interrupts */ may_hard_irq_enable(); diff --git a/arch/powerpc/kernel/paca.c b/arch/powerpc/kernel/paca.c index fa20060..2ad6108 100644 --- a/arch/powerpc/kernel/paca.c +++ b/arch/powerpc/kernel/paca.c @@ -168,6 +168,7 @@ void __init initialise_paca(struct paca_struct *new_paca, int cpu) /* For now -- if we have threads this will be adjusted later */ new_paca->tcd_ptr = &new_paca->tcd; #endif + new_paca->irq = -1; } /* Put the paca pointer into r13 and SPRG_PACA */ -- 2.9.3 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v3 2/3] Keep interrupts enabled even on soft disable 2017-01-05 4:55 ` [PATCH v3 2/3] Keep interrupts enabled even on soft disable Balbir Singh @ 2017-01-05 5:12 ` Madhavan Srinivasan 2017-01-06 5:18 ` kbuild test robot 1 sibling, 0 replies; 6+ messages in thread From: Madhavan Srinivasan @ 2017-01-05 5:12 UTC (permalink / raw) To: linuxppc-dev On Thursday 05 January 2017 10:25 AM, Balbir Singh wrote: > This patch removes the disabling of interrupts > in soft-disable mode, when interrupts are received > (in lazy mode). The new scheme keeps the interrupts > enabled when we receive an interrupt and does the > following: > > When an external interrupt is received, we > store the interrupt in local_paca via > ppc_md.get_irq(). Later when interrupts are > enabled and replayed, we reuse the stored > interrupt and process it via generic_handle_irq > > NOTE: This works only for PPC_XICS at the moment > and we'll enable it for XIVE in the future. MPIC/ > OpenPIC is not supported due to the requirement > that external interrupts/IPIs need unique priorities. > At the cost of adding more space in the PACA, we > can store multiple priorities and support more > controllers, but I think we can live with supporting > only XICS for now and XIVE in the future. > > Cc: Michael Ellerman <mpe@ellerman.id.au> > Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> > Cc: Paul Mackerras <paulus@samba.org> > Cc: Nicholas Piggin <npiggin@gmail.com> > > Signed-off-by: Balbir Singh <bsingharora@gmail.com> > --- > arch/powerpc/include/asm/paca.h | 1 + > arch/powerpc/kernel/exceptions-64s.S | 25 +++++++++++++--- > arch/powerpc/kernel/irq.c | 55 ++++++++++++++++++++++++++++++++++-- > arch/powerpc/kernel/paca.c | 1 + > 4 files changed, 76 insertions(+), 6 deletions(-) > > diff --git a/arch/powerpc/include/asm/paca.h b/arch/powerpc/include/asm/paca.h > index 6a6792b..dcbcaa6 100644 > --- a/arch/powerpc/include/asm/paca.h > +++ b/arch/powerpc/include/asm/paca.h > @@ -163,6 +163,7 @@ struct paca_struct { > #ifdef CONFIG_PPC_TRANSACTIONAL_MEM > u64 tm_scratch; /* TM scratch area for reclaim */ > #endif > + u32 irq; /* IRQ pending */ > > #ifdef CONFIG_PPC_POWERNV > /* Per-core mask tracking idle threads and a lock bit-[L][TTTTTTTT] */ > diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S > index d39d611..cf64bc4 100644 > --- a/arch/powerpc/kernel/exceptions-64s.S > +++ b/arch/powerpc/kernel/exceptions-64s.S > @@ -1273,6 +1273,23 @@ EXC_REAL_NONE(0x1800, 0x1900) > EXC_VIRT_NONE(0x5800, 0x5900) > #endif > > +/* > + * Currently we support keeping interrupts > + * enabled only for XICS. We can enhance this > + * as we add support for other controllers > + */ > +#ifdef CONFIG_PPC_XICS > +#define MASKED_INTERRUPT_DISABLE(_H) \ > + GET_SCRATCH0(r10); \ > + std r13,PACA_EXGEN+EX_R13(r13); \ > + EXCEPTION_PROLOG_PSERIES_1(handle_irq_mask, _H); > +#else > +#define MASKED_INTERRUPT_DISABLE(_H) \ > + mfspr r10,SPRN_##_H##SRR1; \ > + rldicl r10,r10,48,1; /* clear MSR_EE */ \ > + rotldi r10,r10,16; \ > + mtspr SPRN_##_H##SRR1,r10; > +#endif > > /* > * An interrupt came in while soft-disabled. We set paca->irq_happened, then: > @@ -1287,6 +1304,7 @@ EXC_VIRT_NONE(0x5800, 0x5900) > #define MASKED_INTERRUPT(_H) \ > masked_##_H##interrupt: \ > std r11,PACA_EXGEN+EX_R11(r13); \ > + std r12,PACA_EXGEN+EX_R12(r13); \ > lbz r11,PACAIRQHAPPENED(r13); \ > or r11,r11,r10; \ > stb r11,PACAIRQHAPPENED(r13); \ > @@ -1300,10 +1318,7 @@ masked_##_H##interrupt: \ > beq 2f; \ > cmpwi r10,PACA_IRQ_HMI; \ > beq 2f; \ > - mfspr r10,SPRN_##_H##SRR1; \ > - rldicl r10,r10,48,1; /* clear MSR_EE */ \ > - rotldi r10,r10,16; \ > - mtspr SPRN_##_H##SRR1,r10; \ > + MASKED_INTERRUPT_DISABLE(_H) \ > 2: mtcrf 0x80,r9; \ > ld r9,PACA_EXGEN+EX_R9(r13); \ > ld r10,PACA_EXGEN+EX_R10(r13); \ > @@ -1321,6 +1336,8 @@ USE_FIXED_SECTION(virt_trampolines) > MASKED_INTERRUPT() > MASKED_INTERRUPT(H) > > +EXC_COMMON(handle_irq_mask, 0x500, handle_masked_irq) > + > #ifdef CONFIG_KVM_BOOK3S_64_HANDLER > TRAMP_REAL_BEGIN(kvmppc_skip_interrupt) > /* > diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c > index a018f5c..f34f06b 100644 > --- a/arch/powerpc/kernel/irq.c > +++ b/arch/powerpc/kernel/irq.c > @@ -96,7 +96,6 @@ extern int tau_interrupts(int); > #ifdef CONFIG_PPC64 > > int distribute_irqs = 1; > - > static inline notrace unsigned long get_irq_happened(void) > { > unsigned long happened; > @@ -498,6 +497,54 @@ static inline void check_stack_overflow(void) > #endif > } > > +#ifdef CONFIG_PPC_XICS > +static inline notrace int get_paca_irq(void) > +{ > + int irq; > + > + __asm__ __volatile__("lbz %0,%1(13)" > + : "=r" (irq) : "i" (offsetof(struct paca_struct, irq))); > + > + return irq; > +} > + > +static inline notrace void set_paca_irq(int irq) > +{ > + __asm__ __volatile__("stb %0,%1(13)" > + : : "r" (irq), "i" (offsetof(struct paca_struct, irq))); Can we add "memory" to the clobber list, since you are writing to PACA? > +} > + > +void handle_masked_irq(struct pt_regs *regs) > +{ > + /* > + * TODO: Add support for XIVE as applicable > + */ > + unsigned int irq; > + /* > + * NOTE, we don't use irq_enter/exit, otherwise > + * our accounting and tracing might be incorrect. > + */ > + irq = ppc_md.get_irq(); > + > + /* > + * Store away irq in PACA for replay later > + */ > + set_paca_irq(irq); > +} > + > +#else > + > +static inline notrace int get_paca_irq(void) > +{ > + return -1; > +} > + > +static inline notrace void set_paca_irq(int irq) > +{ > +} > + > +#endif /* CONFIG_PPC_XICS */ > + > void __do_irq(struct pt_regs *regs) > { > unsigned int irq; > @@ -513,7 +560,11 @@ void __do_irq(struct pt_regs *regs) > * > * This will typically lower the interrupt line to the CPU > */ > - irq = ppc_md.get_irq(); > + irq = get_paca_irq(); > + if (irq != -1) > + set_paca_irq(-1); > + else > + irq = ppc_md.get_irq(); > > /* We can hard enable interrupts now to allow perf interrupts */ > may_hard_irq_enable(); > diff --git a/arch/powerpc/kernel/paca.c b/arch/powerpc/kernel/paca.c > index fa20060..2ad6108 100644 > --- a/arch/powerpc/kernel/paca.c > +++ b/arch/powerpc/kernel/paca.c > @@ -168,6 +168,7 @@ void __init initialise_paca(struct paca_struct *new_paca, int cpu) > /* For now -- if we have threads this will be adjusted later */ > new_paca->tcd_ptr = &new_paca->tcd; > #endif > + new_paca->irq = -1; > } > > /* Put the paca pointer into r13 and SPRG_PACA */ ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v3 2/3] Keep interrupts enabled even on soft disable 2017-01-05 4:55 ` [PATCH v3 2/3] Keep interrupts enabled even on soft disable Balbir Singh 2017-01-05 5:12 ` Madhavan Srinivasan @ 2017-01-06 5:18 ` kbuild test robot 1 sibling, 0 replies; 6+ messages in thread From: kbuild test robot @ 2017-01-06 5:18 UTC (permalink / raw) To: Balbir Singh; +Cc: kbuild-all, linuxppc-dev, Paul Mackerras, Nicholas Piggin [-- Attachment #1: Type: text/plain, Size: 1092 bytes --] Hi Balbir, [auto build test ERROR on powerpc/next] [also build test ERROR on v4.10-rc2 next-20170105] [if your patch is applied to the wrong git tree, please drop us a note to help improve the system] url: https://github.com/0day-ci/linux/commits/Balbir-Singh/Merge-IPI-and-DEFAULT-priorities/20170106-101307 base: https://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git next config: powerpc-cell_defconfig (attached as .config) compiler: powerpc64-linux-gnu-gcc (Debian 6.1.1-9) 6.1.1 20160705 reproduce: wget https://git.kernel.org/cgit/linux/kernel/git/wfg/lkp-tests.git/plain/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross # save the attached .config to linux build tree make.cross ARCH=powerpc All errors (new ones prefixed by >>): arch/powerpc/kernel/head_64.o: In function `handle_irq_mask': >> (.text+0x2ae0): undefined reference to `handle_masked_irq' --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation [-- Attachment #2: .config.gz --] [-- Type: application/gzip, Size: 18432 bytes --] ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v3 3/3] Enable soft nmi handler 2017-01-05 4:55 [PATCH v3 0/3] Support Soft NMI Balbir Singh 2017-01-05 4:55 ` [PATCH v3 1/3] Merge IPI and DEFAULT priorities Balbir Singh 2017-01-05 4:55 ` [PATCH v3 2/3] Keep interrupts enabled even on soft disable Balbir Singh @ 2017-01-05 4:55 ` Balbir Singh 2 siblings, 0 replies; 6+ messages in thread From: Balbir Singh @ 2017-01-05 4:55 UTC (permalink / raw) To: linuxppc-dev Cc: Balbir Singh, Michael Ellerman, Benjamin Herrenschmidt, Paul Mackerras, Nicholas Piggin a. On decrementer interrupt, instead of setting dec to maximum and returning, we do the following i. Call a function handle_nmi_dec, which in turn calls handle_soft_nmi ii. handle_soft_nmi sets the decrementer value to 1 second and checks if more than 30 seconds have passed since starting it. If so it calls BUG_ON(1), we can do an NMI panic as well. This patch relies on the previous patch to keep interrupts enabled and held at the previous priority (not EOI'd) when we disable interrupts. Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Balbir Singh <bsingharora@gmail.com> --- arch/powerpc/kernel/exceptions-64s.S | 8 ++++---- arch/powerpc/kernel/time.c | 25 +++++++++++++++++++++++++ 2 files changed, 29 insertions(+), 4 deletions(-) diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S index cf64bc4..7c6bbe8a 100644 --- a/arch/powerpc/kernel/exceptions-64s.S +++ b/arch/powerpc/kernel/exceptions-64s.S @@ -1310,10 +1310,9 @@ masked_##_H##interrupt: \ stb r11,PACAIRQHAPPENED(r13); \ cmpwi r10,PACA_IRQ_DEC; \ bne 1f; \ - lis r10,0x7fff; \ - ori r10,r10,0xffff; \ - mtspr SPRN_DEC,r10; \ - b 2f; \ + GET_SCRATCH0(r10); \ + std r13,PACA_EXGEN+EX_R13(r13); \ + EXCEPTION_PROLOG_PSERIES_1(handle_nmi_dec, _H); \ 1: cmpwi r10,PACA_IRQ_DBELL; \ beq 2f; \ cmpwi r10,PACA_IRQ_HMI; \ @@ -1337,6 +1336,7 @@ USE_FIXED_SECTION(virt_trampolines) MASKED_INTERRUPT(H) EXC_COMMON(handle_irq_mask, 0x500, handle_masked_irq) +EXC_COMMON(handle_nmi_dec, 0x900, handle_soft_nmi) #ifdef CONFIG_KVM_BOOK3S_64_HANDLER TRAMP_REAL_BEGIN(kvmppc_skip_interrupt) diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c index bc2e08d..cf814ea 100644 --- a/arch/powerpc/kernel/time.c +++ b/arch/powerpc/kernel/time.c @@ -118,6 +118,7 @@ struct clock_event_device decrementer_clockevent = { EXPORT_SYMBOL(decrementer_clockevent); DEFINE_PER_CPU(u64, decrementers_next_tb); +DEFINE_PER_CPU(u64, nmi_started_tb); static DEFINE_PER_CPU(struct clock_event_device, decrementers); #define XSEC_PER_SEC (1024*1024) @@ -520,6 +521,7 @@ static void __timer_interrupt(void) u64 now; trace_timer_interrupt_entry(regs); + __this_cpu_write(nmi_started_tb, 0); if (test_irq_work_pending()) { clear_irq_work_pending(); @@ -566,6 +568,7 @@ void timer_interrupt(struct pt_regs * regs) * some CPUs will continue to take decrementer exceptions. */ set_dec(decrementer_max); + __this_cpu_write(nmi_started_tb, 0); /* Some implementations of hotplug will get timer interrupts while * offline, just ignore these and we also need to set @@ -599,6 +602,28 @@ void timer_interrupt(struct pt_regs * regs) EXPORT_SYMBOL(timer_interrupt); /* + * If we have watchdog enabled, we do expect to hit this + * at-least once per sample_frequence. + * We set the decrement to 30 seconds and if we hit it + * again.. it's a BUG, it can be made a NMI panic as well. + */ +void handle_soft_nmi(struct pt_regs *regs) +{ + unsigned long long now = get_tb_or_rtc(); + unsigned long long nmi_started_time = __this_cpu_read(nmi_started_tb); + + if (!nmi_started_time) { + set_dec(ppc_tb_freq); + __this_cpu_write(nmi_started_tb, now); + } else { + if ((now - nmi_started_time) >= (30 * ppc_tb_freq)) + BUG(); + else + set_dec(ppc_tb_freq); + } +} + +/* * Hypervisor decrementer interrupts shouldn't occur but are sometimes * left pending on exit from a KVM guest. We don't need to do anything * to clear them, as they are edge-triggered. -- 2.9.3 ^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2017-01-06 5:19 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2017-01-05 4:55 [PATCH v3 0/3] Support Soft NMI Balbir Singh 2017-01-05 4:55 ` [PATCH v3 1/3] Merge IPI and DEFAULT priorities Balbir Singh 2017-01-05 4:55 ` [PATCH v3 2/3] Keep interrupts enabled even on soft disable Balbir Singh 2017-01-05 5:12 ` Madhavan Srinivasan 2017-01-06 5:18 ` kbuild test robot 2017-01-05 4:55 ` [PATCH v3 3/3] Enable soft nmi handler Balbir Singh
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).