From patchwork Wed Jun 17 09:01:53 2009 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Benjamin Herrenschmidt X-Patchwork-Id: 28768 Return-Path: X-Original-To: patchwork-incoming@bilbo.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from ozlabs.org (ozlabs.org [203.10.76.45]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "mx.ozlabs.org", Issuer "CA Cert Signing Authority" (verified OK)) by bilbo.ozlabs.org (Postfix) with ESMTPS id C0404B71B8 for ; Wed, 17 Jun 2009 19:03:12 +1000 (EST) Received: by ozlabs.org (Postfix) id AEB26DDDA1; Wed, 17 Jun 2009 19:03:12 +1000 (EST) Delivered-To: patchwork-incoming@ozlabs.org Received: from bilbo.ozlabs.org (bilbo.ozlabs.org [203.10.76.25]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "bilbo.ozlabs.org", Issuer "CAcert Class 3 Root" (verified OK)) by ozlabs.org (Postfix) with ESMTPS id A2DCDDDDA0 for ; Wed, 17 Jun 2009 19:03:12 +1000 (EST) Received: from bilbo.ozlabs.org (localhost [127.0.0.1]) by bilbo.ozlabs.org (Postfix) with ESMTP id 8C5EAB7582 for ; Wed, 17 Jun 2009 19:02:37 +1000 (EST) Received: from ozlabs.org (ozlabs.org [203.10.76.45]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client CN "mx.ozlabs.org", Issuer "CA Cert Signing Authority" (verified OK)) by bilbo.ozlabs.org (Postfix) with ESMTPS id CE964B7177 for ; Wed, 17 Jun 2009 19:02:31 +1000 (EST) Received: by ozlabs.org (Postfix) id BE7E6DDD1C; Wed, 17 Jun 2009 19:02:31 +1000 (EST) Delivered-To: linuxppc-dev@ozlabs.org Received: by ozlabs.org (Postfix, from userid 1030) id BCD5EDDDA0; Wed, 17 Jun 2009 19:02:31 +1000 (EST) To: From: Benjamin Herrenschmidt Date: Wed, 17 Jun 2009 19:01:53 +1000 Subject: [PATCH] powerpc: Add irqtrace support for 32-bit powerpc Message-Id: <20090617090231.BCD5EDDDA0@ozlabs.org> X-BeenThere: linuxppc-dev@lists.ozlabs.org X-Mailman-Version: 2.1.11 Precedence: list List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Sender: linuxppc-dev-bounces+patchwork-incoming=ozlabs.org@lists.ozlabs.org Errors-To: linuxppc-dev-bounces+patchwork-incoming=ozlabs.org@lists.ozlabs.org Based on initial work from: Dale Farnsworth Add the low level irq tracing hooks for 32-bit powerpc needed to enable full lockdep functionality. The approach taken to deal with the code in entry_32.S is that we don't trace all the transitions of MSR:EE when we just turn it off to peek at TI_FLAGS without races. Only when we are calling into C code or returning from exceptions with a state that have changed from what lockdep thinks. There's a little bugger though: If we take an exception that keeps interrupts enabled (such as an alignment exception) while interrupts are enabled, we will call trace_hardirqs_on() on the way back spurriously. Not a big deal, but to get rid of it would require remembering in pt_regs that the exception was one of the type that kept interrupts enabled which we don't know at this stage. (Well, we could test all cases for regs->trap but that sucks too much). Signed-off-by: Benjamin Herrenschmidt --- Not very well tested ... got issues with my u-boot on the 440 board and I didn't have much time to test elsewhere. FSL welcome to test on funky configs. It's likely to still be subtely broken :-) arch/powerpc/Kconfig | 1 arch/powerpc/include/asm/hw_irq.h | 20 +++--- arch/powerpc/kernel/entry_32.S | 126 ++++++++++++++++++++++++++++++++++++-- arch/powerpc/kernel/setup_32.c | 2 arch/powerpc/kernel/udbg.c | 1 5 files changed, 135 insertions(+), 15 deletions(-) --- linux-work.orig/arch/powerpc/Kconfig 2009-06-17 16:45:15.000000000 +1000 +++ linux-work/arch/powerpc/Kconfig 2009-06-17 16:45:36.000000000 +1000 @@ -62,7 +62,6 @@ config HAVE_LATENCYTOP_SUPPORT config TRACE_IRQFLAGS_SUPPORT bool - depends on PPC64 default y config LOCKDEP_SUPPORT Index: linux-work/arch/powerpc/kernel/entry_32.S =================================================================== --- linux-work.orig/arch/powerpc/kernel/entry_32.S 2009-06-17 16:45:15.000000000 +1000 +++ linux-work/arch/powerpc/kernel/entry_32.S 2009-06-17 18:29:23.000000000 +1000 @@ -191,11 +191,49 @@ transfer_to_handler_cont: mflr r9 lwz r11,0(r9) /* virtual address of handler */ lwz r9,4(r9) /* where to go when done */ +#ifdef CONFIG_TRACE_IRQFLAGS + lis r12,reenable_mmu@h + ori r12,r12,reenable_mmu@l + mtspr SPRN_SRR0,r12 + mtspr SPRN_SRR1,r10 + SYNC + RFI +reenable_mmu: /* re-enable mmu so we can */ + mfmsr r10 + lwz r12,_MSR(r1) + xor r10,r10,r12 + andi. r10,r10,MSR_EE /* Did EE change? */ + beq 1f + + /* Save handler and return address into the 2 unused words + * of the STACK_FRAME_OVERHEAD (sneak sneak sneak). Everything + * else can be recovered from the pt_regs except r3 which for + * normal interrupts has been set to pt_regs and for syscalls + * is an argument, so we temporarily use ORIG_GPR3 to save it + */ + stw r9,8(r1) + stw r11,12(r1) + stw r3,ORIG_GPR3(r1) + bl trace_hardirqs_off + lwz r0,GPR0(r1) + lwz r3,ORIG_GPR3(r1) + lwz r4,GPR4(r1) + lwz r5,GPR5(r1) + lwz r6,GPR6(r1) + lwz r7,GPR7(r1) + lwz r8,GPR8(r1) + lwz r9,8(r1) + lwz r11,12(r1) +1: mtctr r11 + mtlr r9 + bctr /* jump to handler */ +#else /* CONFIG_TRACE_IRQFLAGS */ mtspr SPRN_SRR0,r11 mtspr SPRN_SRR1,r10 mtlr r9 SYNC RFI /* jump to handler, enable MMU */ +#endif /* CONFIG_TRACE_IRQFLAGS */ #if defined (CONFIG_6xx) || defined(CONFIG_E500) 4: rlwinm r12,r12,0,~_TLF_NAPPING @@ -251,6 +289,31 @@ _GLOBAL(DoSyscall) #ifdef SHOW_SYSCALLS bl do_show_syscall #endif /* SHOW_SYSCALLS */ +#ifdef CONFIG_TRACE_IRQFLAGS + /* Return from syscalls can (and generally will) hard enable + * interrupts. You aren't supposed to call a syscall with + * interrupts disabled in the first place. However, to ensure + * that we get it right vs. lockdep if it happens, we force + * that hard enable here with appropriate tracing if we see + * that we have been called with interrupts off + */ + mfmsr r11 + andi. r12,r11,MSR_EE + bne+ 1f + /* We came in with interrupts disabled, we enable them now */ + bl trace_hardirqs_on + mfmsr r11 + lwz r0,GPR0(r1) + lwz r3,GPR3(r1) + lwz r4,GPR4(r1) + ori r11,r11,MSR_EE + lwz r5,GPR5(r1) + lwz r6,GPR6(r1) + lwz r7,GPR7(r1) + lwz r8,GPR8(r1) + mtmsr r11 +1: +#endif /* CONFIG_TRACE_IRQFLAGS */ rlwinm r10,r1,0,0,(31-THREAD_SHIFT) /* current_thread_info() */ lwz r11,TI_FLAGS(r10) andi. r11,r11,_TIF_SYSCALL_T_OR_A @@ -275,6 +338,7 @@ ret_from_syscall: rlwinm r12,r1,0,0,(31-THREAD_SHIFT) /* current_thread_info() */ /* disable interrupts so current_thread_info()->flags can't change */ LOAD_MSR_KERNEL(r10,MSR_KERNEL) /* doesn't include MSR_EE */ + /* Note: We don't bother telling lockdep about it */ SYNC MTMSRD(r10) lwz r9,TI_FLAGS(r12) @@ -288,6 +352,19 @@ ret_from_syscall: oris r11,r11,0x1000 /* Set SO bit in CR */ stw r11,_CCR(r1) syscall_exit_cont: + lwz r8,_MSR(r1) +#ifdef CONFIG_TRACE_IRQFLAGS + /* If we are going to return from the syscall with interrupts + * off, we trace that here. It shouldn't happen though but we + * want to catch the bugger if it does right ? + */ + andi. r10,r8,MSR_EE + bne+ 1f + stw r3,GPR3(r1) + bl trace_hardirqs_off + lwz r3,GPR3(r1) +1: +#endif /* CONFIG_TRACE_IRQFLAGS */ #if defined(CONFIG_4xx) || defined(CONFIG_BOOKE) /* If the process has its own DBCR0 value, load it up. The internal debug mode bit tells us that dbcr0 should be loaded. */ @@ -311,7 +388,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_NEED_PAIRE mtlr r4 mtcr r5 lwz r7,_NIP(r1) - lwz r8,_MSR(r1) FIX_SRR1(r8, r0) lwz r2,GPR2(r1) lwz r1,GPR1(r1) @@ -394,7 +470,9 @@ syscall_exit_work: andi. r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP) beq ret_from_except - /* Re-enable interrupts */ + /* Re-enable interrupts. There is no need to trace that with + * lockdep as we are supposed to have IRQs on at this point + */ ori r10,r10,MSR_EE SYNC MTMSRD(r10) @@ -705,6 +783,7 @@ ret_from_except: /* Hard-disable interrupts so that current_thread_info()->flags * can't change between when we test it and when we return * from the interrupt. */ + /* Note: We don't bother telling lockdep about it */ LOAD_MSR_KERNEL(r10,MSR_KERNEL) SYNC /* Some chip revs have problems here... */ MTMSRD(r10) /* disable interrupts */ @@ -744,11 +823,24 @@ resume_kernel: beq+ restore andi. r0,r3,MSR_EE /* interrupts off? */ beq restore /* don't schedule if so */ +#ifdef CONFIG_TRACE_IRQFLAGS + /* Lockdep thinks irqs are enabled, we need to call + * preempt_schedule_irq with IRQs off, so we inform lockdep + * now that we -did- turn them off already + */ + bl trace_hardirqs_off +#endif 1: bl preempt_schedule_irq rlwinm r9,r1,0,0,(31-THREAD_SHIFT) lwz r3,TI_FLAGS(r9) andi. r0,r3,_TIF_NEED_RESCHED bne- 1b +#ifdef CONFIG_TRACE_IRQFLAGS + /* And now, to properly rebalance the above, we tell lockdep they + * are being turned back on, which will happen when we return + */ + bl trace_hardirqs_on +#endif #else resume_kernel: #endif /* CONFIG_PREEMPT */ @@ -781,8 +873,28 @@ BEGIN_FTR_SECTION END_FTR_SECTION_IFSET(CPU_FTR_NEED_PAIRED_STWCX) stwcx. r0,0,r1 /* to clear the reservation */ -#if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE)) lwz r9,_MSR(r1) +#ifdef CONFIG_TRACE_IRQFLAGS + /* Lockdep doesn't know about the fact that IRQs are temporarily turned + * off in this assembly code while peeking at TI_FLAGS() and such. However + * we need to inform it if the exception turned interrupts off, and we + * are about to trun them back on. + * + * The problem here sadly is that we don't know whether the exceptions was + * one that turned interrupts off or not. So we always tell lockdep about + * turning them on here when we go back to wherever we came from with EE + * on, even if that may meen some redudant calls being tracked. Maybe later + * we could encode what the exception did somewhere or test the exception + * type in the pt_regs but that sounds overkill + */ + andi. r10,r9,MSR_EE + beq 1f + bl trace_hardirqs_on + lwz r9,_MSR(r1) +1: +#endif /* CONFIG_TRACE_IRQFLAGS */ + +#if !(defined(CONFIG_4xx) || defined(CONFIG_BOOKE)) andi. r10,r9,MSR_RI /* check if this exception occurred */ beql nonrecoverable /* at a bad place (MSR:RI = 0) */ @@ -805,7 +917,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_NEED_PAIRE MTMSRD(r10) /* clear the RI bit */ .globl exc_exit_restart exc_exit_restart: - lwz r9,_MSR(r1) lwz r12,_NIP(r1) FIX_SRR1(r9,r10) mtspr SPRN_SRR0,r12 @@ -1035,11 +1146,18 @@ do_work: /* r10 contains MSR_KERNEL he beq do_user_signal do_resched: /* r10 contains MSR_KERNEL here */ + /* Note: We don't need to inform lockdep that we are enabling + * interrupts here. As far as it knows, they are already enabled + */ ori r10,r10,MSR_EE SYNC MTMSRD(r10) /* hard-enable interrupts */ bl schedule recheck: + /* Note: And we don't tell it we are disabling them again + * neither. Those disable/enable cycles used to peek at + * TI_FLAGS aren't advertised. + */ LOAD_MSR_KERNEL(r10,MSR_KERNEL) SYNC MTMSRD(r10) /* disable interrupts */ Index: linux-work/arch/powerpc/kernel/setup_32.c =================================================================== --- linux-work.orig/arch/powerpc/kernel/setup_32.c 2009-06-17 16:45:15.000000000 +1000 +++ linux-work/arch/powerpc/kernel/setup_32.c 2009-06-17 18:29:37.000000000 +1000 @@ -119,6 +119,8 @@ notrace unsigned long __init early_init( */ notrace void __init machine_init(unsigned long dt_ptr) { + lockdep_init(); + /* Enable early debugging if any specified (see udbg.h) */ udbg_early_init(); Index: linux-work/arch/powerpc/include/asm/hw_irq.h =================================================================== --- linux-work.orig/arch/powerpc/include/asm/hw_irq.h 2009-06-17 16:45:15.000000000 +1000 +++ linux-work/arch/powerpc/include/asm/hw_irq.h 2009-06-17 16:45:36.000000000 +1000 @@ -68,13 +68,13 @@ static inline int irqs_disabled_flags(un #if defined(CONFIG_BOOKE) #define SET_MSR_EE(x) mtmsr(x) -#define local_irq_restore(flags) __asm__ __volatile__("wrtee %0" : : "r" (flags) : "memory") +#define raw_local_irq_restore(flags) __asm__ __volatile__("wrtee %0" : : "r" (flags) : "memory") #else #define SET_MSR_EE(x) mtmsr(x) -#define local_irq_restore(flags) mtmsr(flags) +#define raw_local_irq_restore(flags) mtmsr(flags) #endif -static inline void local_irq_disable(void) +static inline void raw_local_irq_disable(void) { #ifdef CONFIG_BOOKE __asm__ __volatile__("wrteei 0": : :"memory"); @@ -86,7 +86,7 @@ static inline void local_irq_disable(voi #endif } -static inline void local_irq_enable(void) +static inline void raw_local_irq_enable(void) { #ifdef CONFIG_BOOKE __asm__ __volatile__("wrteei 1": : :"memory"); @@ -98,7 +98,7 @@ static inline void local_irq_enable(void #endif } -static inline void local_irq_save_ptr(unsigned long *flags) +static inline void raw_local_irq_save_ptr(unsigned long *flags) { unsigned long msr; msr = mfmsr(); @@ -110,12 +110,12 @@ static inline void local_irq_save_ptr(un #endif } -#define local_save_flags(flags) ((flags) = mfmsr()) -#define local_irq_save(flags) local_irq_save_ptr(&flags) -#define irqs_disabled() ((mfmsr() & MSR_EE) == 0) +#define raw_local_save_flags(flags) ((flags) = mfmsr()) +#define raw_local_irq_save(flags) raw_local_irq_save_ptr(&flags) +#define raw_irqs_disabled() ((mfmsr() & MSR_EE) == 0) +#define raw_irqs_disabled_flags(flags) (((flags) & MSR_EE) == 0) -#define hard_irq_enable() local_irq_enable() -#define hard_irq_disable() local_irq_disable() +#define hard_irq_disable() raw_local_irq_disable() static inline int irqs_disabled_flags(unsigned long flags) { Index: linux-work/arch/powerpc/kernel/udbg.c =================================================================== --- linux-work.orig/arch/powerpc/kernel/udbg.c 2009-06-17 17:14:41.000000000 +1000 +++ linux-work/arch/powerpc/kernel/udbg.c 2009-06-17 17:14:43.000000000 +1000 @@ -65,6 +65,7 @@ void __init udbg_early_init(void) #ifdef CONFIG_PPC_EARLY_DEBUG console_loglevel = 10; #endif + register_early_udbg_console(); } /* udbg library, used by xmon et al */