diff options
| author | Benjamin Herrenschmidt <benh@kernel.crashing.org> | 2012-03-16 09:26:59 +1100 |
|---|---|---|
| committer | Benjamin Herrenschmidt <benh@kernel.crashing.org> | 2012-03-16 09:28:17 +1100 |
| commit | 01e8ec4417d3c484986af0adaa0ae6632e0a59cd (patch) | |
| tree | 46c760b18c2c5f089ac4043f95de4fd420c1b082 | |
| parent | 7d3d897a4697e4ff746e5e82f116b2346ed28150 (diff) | |
powerpc: Fix power4/970 idle code regression with lockdep
in commit 7230c5644188cd9e3fb380cc97dde00c464a3ba7
"powerpc: Rework lazy-interrupt handling"
I introduced a regression, accidentally calling irq tracing twice
and not properly restoring a clobbered register (r7) later used
for writing to the MSR.
This caused lockups when booting on a G5 with lockdep enabled.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
| -rw-r--r-- | arch/powerpc/kernel/idle_power4.S | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/arch/powerpc/kernel/idle_power4.S b/arch/powerpc/kernel/idle_power4.S index d8cdba4c28b..2c71b0fc9f9 100644 --- a/arch/powerpc/kernel/idle_power4.S +++ b/arch/powerpc/kernel/idle_power4.S @@ -50,9 +50,9 @@ END_FTR_SECTION_IFCLR(CPU_FTR_CAN_NAP) addi r1,r1,128 ld r0,16(r1) mtlr r0 + mfmsr r7 #endif /* CONFIG_TRACE_IRQFLAGS */ - TRACE_ENABLE_INTS li r0,1 stb r0,PACASOFTIRQEN(r13) /* we'll hard-enable shortly */ BEGIN_FTR_SECTION |