From 0094368e3bb97a710ce163f9c06d290c1c308621 Mon Sep 17 00:00:00 2001 From: Michael Ellerman Date: Sun, 3 May 2020 00:33:16 +1000 Subject: powerpc/64s: Fix unrecoverable SLB crashes due to preemption check Hugh reported that his trusty G5 crashed after a few hours under load with an "Unrecoverable exception 380". The crash is in interrupt_return() where we check lazy_irq_pending(), which calls get_paca() and with CONFIG_DEBUG_PREEMPT=y that goes to check_preemption_disabled() via debug_smp_processor_id(). As Nick explained on the list: Problem is MSR[RI] is cleared here, ready to do the last few things for interrupt return where we're not allowed to take any other interrupts. SLB interrupts can happen just about anywhere aside from kernel text, global variables, and stack. When that hits, it appears to be unrecoverable due to RI=0. The problematic access is in preempt_count() which is: return READ_ONCE(current_thread_info()->preempt_count); Because of THREAD_INFO_IN_TASK, current_thread_info() just points to current, so the access is to somewhere in kernel memory, but not on the stack or in .data, which means it can cause an SLB miss. If we take an SLB miss with RI=0 it is fatal. The easiest solution is to add a version of lazy_irq_pending() that doesn't do the preemption check and call it from the interrupt return path. Fixes: 68b34588e202 ("powerpc/64/sycall: Implement syscall entry/exit logic in C") Reported-by: Hugh Dickins Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20200502143316.929341-1-mpe@ellerman.id.au --- arch/powerpc/kernel/syscall_64.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/powerpc/kernel/syscall_64.c') diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c index c74295a7765b..1fe94dd9de32 100644 --- a/arch/powerpc/kernel/syscall_64.c +++ b/arch/powerpc/kernel/syscall_64.c @@ -189,7 +189,7 @@ again: /* This pattern matches prep_irq_for_idle */ __hard_EE_RI_disable(); - if (unlikely(lazy_irq_pending())) { + if (unlikely(lazy_irq_pending_nocheck())) { __hard_RI_enable(); trace_hardirqs_off(); local_paca->irq_happened |= PACA_IRQ_HARD_DIS; @@ -264,7 +264,7 @@ again: trace_hardirqs_on(); __hard_EE_RI_disable(); - if (unlikely(lazy_irq_pending())) { + if (unlikely(lazy_irq_pending_nocheck())) { __hard_RI_enable(); trace_hardirqs_off(); local_paca->irq_happened |= PACA_IRQ_HARD_DIS; @@ -334,7 +334,7 @@ again: trace_hardirqs_on(); __hard_EE_RI_disable(); - if (unlikely(lazy_irq_pending())) { + if (unlikely(lazy_irq_pending_nocheck())) { __hard_RI_enable(); irq_soft_mask_set(IRQS_ALL_DISABLED); trace_hardirqs_off(); -- cgit v1.2.3 From c0d7dcf89e5151b2259d1c2c1b922da3b881d02e Mon Sep 17 00:00:00 2001 From: Nicholas Piggin Date: Wed, 29 Apr 2020 16:56:49 +1000 Subject: powerpc/64/kuap: Move kuap checks out of MSR[RI]=0 regions of exit code Any kind of WARN causes a program check that will crash with unrecoverable exception if it occurs when RI is clear. Fixes: 68b34588e202 ("powerpc/64/sycall: Implement syscall entry/exit logic in C") Signed-off-by: Nicholas Piggin Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20200429065654.1677541-2-npiggin@gmail.com --- arch/powerpc/kernel/syscall_64.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'arch/powerpc/kernel/syscall_64.c') diff --git a/arch/powerpc/kernel/syscall_64.c b/arch/powerpc/kernel/syscall_64.c index 1fe94dd9de32..7b7c89cad901 100644 --- a/arch/powerpc/kernel/syscall_64.c +++ b/arch/powerpc/kernel/syscall_64.c @@ -35,6 +35,8 @@ notrace long system_call_exception(long r3, long r4, long r5, BUG_ON(!FULL_REGS(regs)); BUG_ON(regs->softe != IRQS_ENABLED); + kuap_check_amr(); + account_cpu_user_entry(); #ifdef CONFIG_PPC_SPLPAR @@ -47,8 +49,6 @@ notrace long system_call_exception(long r3, long r4, long r5, } #endif - kuap_check_amr(); - /* * This is not required for the syscall exit path, but makes the * stack frame look nicer. If this was initialised in the first stack @@ -117,6 +117,8 @@ notrace unsigned long syscall_exit_prepare(unsigned long r3, unsigned long ti_flags; unsigned long ret = 0; + kuap_check_amr(); + regs->result = r3; /* Check whether the syscall is issued inside a restartable sequence */ @@ -204,8 +206,6 @@ again: local_paca->tm_scratch = regs->msr; #endif - kuap_check_amr(); - account_cpu_user_exit(); return ret; @@ -228,6 +228,8 @@ notrace unsigned long interrupt_exit_user_prepare(struct pt_regs *regs, unsigned BUG_ON(!FULL_REGS(regs)); BUG_ON(regs->softe != IRQS_ENABLED); + kuap_check_amr(); + local_irq_save(flags); again: @@ -292,8 +294,6 @@ again: local_paca->tm_scratch = regs->msr; #endif - kuap_check_amr(); - account_cpu_user_exit(); return ret; @@ -313,6 +313,8 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs, unsign BUG_ON(regs->msr & MSR_PR); BUG_ON(!FULL_REGS(regs)); + kuap_check_amr(); + if (unlikely(*ti_flagsp & _TIF_EMULATE_STACK_STORE)) { clear_bits(_TIF_EMULATE_STACK_STORE, ti_flagsp); ret = 1; -- cgit v1.2.3