Excerpts from Christophe Leroy's message of March 16, 2021 5:16 pm:
>
>
> Le 12/03/2021 à 02:20, Nicholas Piggin a écrit :
>> There is no need for this to be in asm, use the new intrrupt entry wrapper.
>>
>> Signed-off-by: Nicholas Piggin <npig...@gmail.com>
>> ---
>> Hopefully this works on a real G5 now, but I couldn't reproduce the
>> problem with QEMU.
>>
>> Thanks,
>> Nick
>>
>> arch/powerpc/include/asm/interrupt.h | 19 +++++++++++
>> arch/powerpc/include/asm/processor.h | 1 +
>> arch/powerpc/include/asm/thread_info.h | 6 ++++
>> arch/powerpc/kernel/exceptions-64s.S | 45 --------------------------
>> arch/powerpc/kernel/idle_book3s.S | 4 +++
>> 5 files changed, 30 insertions(+), 45 deletions(-)
>>
>> diff --git a/arch/powerpc/include/asm/interrupt.h
>> b/arch/powerpc/include/asm/interrupt.h
>> index aedfba29e43a..ef015d3b5e39 100644
>> --- a/arch/powerpc/include/asm/interrupt.h
>> +++ b/arch/powerpc/include/asm/interrupt.h
>> @@ -9,6 +9,17 @@
>> #include <asm/kprobes.h>
>> #include <asm/runlatch.h>
>>
>> +static inline void nap_adjust_return(struct pt_regs *regs)
>> +{
>> +#ifdef CONFIG_PPC_970_NAP
>> + if (unlikely(test_thread_local_flags(_TLF_NAPPING))) {
>> + /* Can avoid a test-and-clear because NMIs do not call this */
>> + clear_thread_local_flags(_TLF_NAPPING);
>> + regs->nip = (unsigned long)power4_idle_nap_return;
>
> Why don't you do regs->nip = regs->link like PPC32 instead of going via an
> intermediate symbol that
> does nothing else than branching to LR ?
It is supposed to keep the return branch predictor balanced.
I don't know if these CPUs have one, if it gets lost during nap, or if
nap latency is so high it really doesn't matter. But I think it is good
practice to make a habit of keeping things balanced.
Thanks,
Nick
