From: Nicholas Piggin npiggin@gmail.com
[ Upstream commit 8a5054d8cbbe03c68dcb0957c291c942132e4101 ]
The architecture allows for machine check exceptions to cause idle wakeups which resume at the 0x200 address which has to return via the idle wakeup code, but the early machine check handler is run first.
The case of a no state-loss sleep is broken because the early handler uses non-volatile register r1 , which is needed for the wakeup protocol, but it is not restored.
Fix this by loading r1 from the MCE exception frame before returning to the idle wakeup code. Also update the comment which has become stale since the idle rewrite in C.
This crash was found and fix confirmed with a machine check injection test in qemu powernv model (which is not upstream in qemu yet).
Fixes: 10d91611f426d ("powerpc/64s: Reimplement book3s idle code in C") Signed-off-by: Nicholas Piggin npiggin@gmail.com Signed-off-by: Michael Ellerman mpe@ellerman.id.au Link: https://lore.kernel.org/r/20200508043408.886394-2-npiggin@gmail.com Signed-off-by: Sasha Levin sashal@kernel.org --- arch/powerpc/kernel/exceptions-64s.S | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S index d0018dd17e0a6..70ac8a6ba0c18 100644 --- a/arch/powerpc/kernel/exceptions-64s.S +++ b/arch/powerpc/kernel/exceptions-64s.S @@ -1090,17 +1090,19 @@ EXC_COMMON_BEGIN(machine_check_idle_common) bl machine_check_queue_event
/* - * We have not used any non-volatile GPRs here, and as a rule - * most exception code including machine check does not. - * Therefore PACA_NAPSTATELOST does not need to be set. Idle - * wakeup will restore volatile registers. + * GPR-loss wakeups are relatively straightforward, because the + * idle sleep code has saved all non-volatile registers on its + * own stack, and r1 in PACAR1. * - * Load the original SRR1 into r3 for pnv_powersave_wakeup_mce. + * For no-loss wakeups the r1 and lr registers used by the + * early machine check handler have to be restored first. r2 is + * the kernel TOC, so no need to restore it. * * Then decrement MCE nesting after finishing with the stack. */ ld r3,_MSR(r1) ld r4,_LINK(r1) + ld r1,GPR1(r1)
lhz r11,PACA_IN_MCE(r13) subi r11,r11,1 @@ -1109,7 +1111,7 @@ EXC_COMMON_BEGIN(machine_check_idle_common) mtlr r4 rlwinm r10,r3,47-31,30,31 cmpwi cr1,r10,2 - bltlr cr1 /* no state loss, return to idle caller */ + bltlr cr1 /* no state loss, return to idle caller with r3=SRR1 */ b idle_return_gpr_loss #endif