When I "fixed" the CONFIG_TRACE_IRQFLAGS case on interrupt entry, I screwed up a little bit with the test for user space vs. kernel.
The code is fine, there's just some dead code around it. I basically removed the test and always create the added stack frame whether coming from user or kernel since in any case we do need to save a bunch of volatile registers or bad things would happen (we can take page faults in the kernel for example). Signed-off-by: Benjamin Herrenschmidt <b...@kernel.crashing.org> --- diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S index ba3aeb4..477ec2c 100644 --- a/arch/powerpc/kernel/entry_32.S +++ b/arch/powerpc/kernel/entry_32.S @@ -226,13 +226,7 @@ reenable_mmu: /* re-enable mmu so we can */ stw r3,16(r1) stw r4,20(r1) stw r5,24(r1) - andi. r12,r12,MSR_PR - b 11f bl trace_hardirqs_off - b 12f -11: - bl trace_hardirqs_off -12: lwz r5,24(r1) lwz r4,20(r1) lwz r3,16(r1) _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@lists.ozlabs.org https://lists.ozlabs.org/listinfo/linuxppc-dev