Please backport commit 190ce8693c23 ("powerpc/tm: Avoid SLB faults in treclaim/trecheckpoint when RI=0") to v3.11+. Attached below if it helps. cheers >From 190ce8693c23eae09ba5f303a83bf2fbeb6478b1 Mon Sep 17 00:00:00 2001 From: Michael Neuling <mikey@xxxxxxxxxxx> Date: Tue, 28 Jun 2016 13:01:04 +1000 Subject: [PATCH] powerpc/tm: Avoid SLB faults in treclaim/trecheckpoint when RI=0 Currently we have 2 segments that are bolted for the kernel linear mapping (ie 0xc000... addresses). This is 0 to 1TB and also the kernel stacks. Anything accessed outside of these regions may need to be faulted in. (In practice machines with TM always have 1T segments) If a machine has < 2TB of memory we never fault on the kernel linear mapping as these two segments cover all physical memory. If a machine has > 2TB of memory, there may be structures outside of these two segments that need to be faulted in. This faulting can occur when running as a guest as the hypervisor may remove any SLB that's not bolted. When we treclaim and trecheckpoint we have a window where we need to run with the userspace GPRs. This means that we no longer have a valid stack pointer in r1. For this window we therefore clear MSR RI to indicate that any exceptions taken at this point won't be able to be handled. This means that we can't take segment misses in this RI=0 window. In this RI=0 region, we currently access the thread_struct for the process being context switched to or from. This thread_struct access may cause a segment fault since it's not guaranteed to be covered by the two bolted segment entries described above. We've seen this with a crash when running as a guest with > 2TB of memory on PowerVM: Unrecoverable exception 4100 at c00000000004f138 Oops: Unrecoverable exception, sig: 6 [#1] SMP NR_CPUS=2048 NUMA pSeries CPU: 1280 PID: 7755 Comm: kworker/1280:1 Tainted: G X 4.4.13-46-default #1 task: c000189001df4210 ti: c000189001d5c000 task.ti: c000189001d5c000 NIP: c00000000004f138 LR: 0000000010003a24 CTR: 0000000010001b20 REGS: c000189001d5f730 TRAP: 4100 Tainted: G X (4.4.13-46-default) MSR: 8000000100001031 <SF,ME,IR,DR,LE> CR: 24000048 XER: 00000000 CFAR: c00000000004ed18 SOFTE: 0 GPR00: ffffffffc58d7b60 c000189001d5f9b0 00000000100d7d00 000000003a738288 GPR04: 0000000000002781 0000000000000006 0000000000000000 c0000d1f4d889620 GPR08: 000000000000c350 00000000000008ab 00000000000008ab 00000000100d7af0 GPR12: 00000000100d7ae8 00003ffe787e67a0 0000000000000000 0000000000000211 GPR16: 0000000010001b20 0000000000000000 0000000000800000 00003ffe787df110 GPR20: 0000000000000001 00000000100d1e10 0000000000000000 00003ffe787df050 GPR24: 0000000000000003 0000000000010000 0000000000000000 00003fffe79e2e30 GPR28: 00003fffe79e2e68 00000000003d0f00 00003ffe787e67a0 00003ffe787de680 NIP [c00000000004f138] restore_gprs+0xd0/0x16c LR [0000000010003a24] 0x10003a24 Call Trace: [c000189001d5f9b0] [c000189001d5f9f0] 0xc000189001d5f9f0 (unreliable) [c000189001d5fb90] [c00000000001583c] tm_recheckpoint+0x6c/0xa0 [c000189001d5fbd0] [c000000000015c40] __switch_to+0x2c0/0x350 [c000189001d5fc30] [c0000000007e647c] __schedule+0x32c/0x9c0 [c000189001d5fcb0] [c0000000007e6b58] schedule+0x48/0xc0 [c000189001d5fce0] [c0000000000deabc] worker_thread+0x22c/0x5b0 [c000189001d5fd80] [c0000000000e7000] kthread+0x110/0x130 [c000189001d5fe30] [c000000000009538] ret_from_kernel_thread+0x5c/0xa4 Instruction dump: 7cb103a6 7cc0e3a6 7ca222a6 78a58402 38c00800 7cc62838 08860000 7cc000a6 38a00006 78c60022 7cc62838 0b060000 <e8c701a0> 7ccff120 e8270078 e8a70098 ---[ end trace 602126d0a1dedd54 ]--- This fixes this by copying the required data from the thread_struct to the stack before we clear MSR RI. Then once we clear RI, we only access the stack, guaranteeing there's no segment miss. We also tighten the region over which we set RI=0 on the treclaim() path. This may have a slight performance impact since we're adding an mtmsr instruction. Fixes: 090b9284d725 ("powerpc/tm: Clear MSR RI in non-recoverable TM code") Signed-off-by: Michael Neuling <mikey@xxxxxxxxxxx> Reviewed-by: Cyril Bur <cyrilbur@xxxxxxxxx> Signed-off-by: Michael Ellerman <mpe@xxxxxxxxxxxxxx> --- arch/powerpc/kernel/tm.S | 61 ++++++++++++++++++++++++++++++++++-------------- 1 file changed, 44 insertions(+), 17 deletions(-) diff --git a/arch/powerpc/kernel/tm.S b/arch/powerpc/kernel/tm.S index bf8f34a58670..b7019b559ddb 100644 --- a/arch/powerpc/kernel/tm.S +++ b/arch/powerpc/kernel/tm.S @@ -110,17 +110,11 @@ _GLOBAL(tm_reclaim) std r3, STK_PARAM(R3)(r1) SAVE_NVGPRS(r1) - /* We need to setup MSR for VSX register save instructions. Here we - * also clear the MSR RI since when we do the treclaim, we won't have a - * valid kernel pointer for a while. We clear RI here as it avoids - * adding another mtmsr closer to the treclaim. This makes the region - * maked as non-recoverable wider than it needs to be but it saves on - * inserting another mtmsrd later. - */ + /* We need to setup MSR for VSX register save instructions. */ mfmsr r14 mr r15, r14 ori r15, r15, MSR_FP - li r16, MSR_RI + li r16, 0 ori r16, r16, MSR_EE /* IRQs hard off */ andc r15, r15, r16 oris r15, r15, MSR_VEC@h @@ -176,7 +170,17 @@ dont_backup_fp: 1: tdeqi r6, 0 EMIT_BUG_ENTRY 1b,__FILE__,__LINE__,0 - /* The moment we treclaim, ALL of our GPRs will switch + /* Clear MSR RI since we are about to change r1, EE is already off. */ + li r4, 0 + mtmsrd r4, 1 + + /* + * BE CAREFUL HERE: + * At this point we can't take an SLB miss since we have MSR_RI + * off. Load only to/from the stack/paca which are in SLB bolted regions + * until we turn MSR RI back on. + * + * The moment we treclaim, ALL of our GPRs will switch * to user register state. (FPRs, CCR etc. also!) * Use an sprg and a tm_scratch in the PACA to shuffle. */ @@ -197,6 +201,11 @@ dont_backup_fp: /* Store the PPR in r11 and reset to decent value */ std r11, GPR11(r1) /* Temporary stash */ + + /* Reset MSR RI so we can take SLB faults again */ + li r11, MSR_RI + mtmsrd r11, 1 + mfspr r11, SPRN_PPR HMT_MEDIUM @@ -397,11 +406,6 @@ restore_gprs: ld r5, THREAD_TM_DSCR(r3) ld r6, THREAD_TM_PPR(r3) - /* Clear the MSR RI since we are about to change R1. EE is already off - */ - li r4, 0 - mtmsrd r4, 1 - REST_GPR(0, r7) /* GPR0 */ REST_2GPRS(2, r7) /* GPR2-3 */ REST_GPR(4, r7) /* GPR4 */ @@ -439,10 +443,33 @@ restore_gprs: ld r6, _CCR(r7) mtcr r6 - REST_GPR(1, r7) /* GPR1 */ - REST_GPR(5, r7) /* GPR5-7 */ REST_GPR(6, r7) - ld r7, GPR7(r7) + + /* + * Store r1 and r5 on the stack so that we can access them + * after we clear MSR RI. + */ + + REST_GPR(5, r7) + std r5, -8(r1) + ld r5, GPR1(r7) + std r5, -16(r1) + + REST_GPR(7, r7) + + /* Clear MSR RI since we are about to change r1. EE is already off */ + li r5, 0 + mtmsrd r5, 1 + + /* + * BE CAREFUL HERE: + * At this point we can't take an SLB miss since we have MSR_RI + * off. Load only to/from the stack/paca which are in SLB bolted regions + * until we turn MSR RI back on. + */ + + ld r5, -8(r1) + ld r1, -16(r1) /* Commit register state as checkpointed state: */ TRECHKPT -- 2.7.4 -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html