Re: [PATCH] perf: riscv: fix broken build due to struct redefinition

From: Conor.Dooley
Date: Fri Aug 12 2022 - 18:37:01 EST


On 12/08/2022 23:33, Palmer Dabbelt wrote:
> On Fri, 12 Aug 2022 07:19:54 PDT (-0700), Conor.Dooley@xxxxxxxxxxxxx wrote:
>> On 12/08/2022 15:13, Jessica Clarke wrote:
>>> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
>>>
>>> On 12 Aug 2022, at 14:51, Conor Dooley <conor.dooley@xxxxxxxxxxxxx> wrote:
>>>>
>>>> Building riscv/for-next produces following error:
>>>> drivers/perf/riscv_pmu_sbi.c:44:7: error: redefinition of 'sbi_pmu_ctr_info'
>>>> union sbi_pmu_ctr_info {
>>>>       ^
>>>> arch/riscv/include/asm/sbi.h:125:7: note: previous definition is here
>>>> union sbi_pmu_ctr_info {
>>>>
>>>> This appears to have been caused by a merge conflict resolution between
>>>> riscv/for-next & riscv/fixes, causing the struct define not being
>>>
>>> union, not struct
>>
>> autopilot, thanks :)
>>
>>>
>>> Jess
>>>
>>>> properly moved to its header.
>>>>
>>>> Fixes: 9a7ccac63f9c ("perf: riscv_pmu{,_sbi}: Miscallenous improvement & fixes")
>>>> Signed-off-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>
>>>> ---
>>>> drivers/perf/riscv_pmu_sbi.c | 14 --------------
>>>> 1 file changed, 14 deletions(-)
>>>>
>>>> diff --git a/drivers/perf/riscv_pmu_sbi.c b/drivers/perf/riscv_pmu_sbi.c
>>>> index e7c6fecbf061..6f6681bbfd36 100644
>>>> --- a/drivers/perf/riscv_pmu_sbi.c
>>>> +++ b/drivers/perf/riscv_pmu_sbi.c
>>>> @@ -41,20 +41,6 @@ static const struct attribute_group *riscv_pmu_attr_groups[] = {
>>>>        NULL,
>>>> };
>>>>
>>>> -union sbi_pmu_ctr_info {
>>>> -     unsigned long value;
>>>> -     struct {
>>>> -             unsigned long csr:12;
>>>> -             unsigned long width:6;
>>>> -#if __riscv_xlen == 32
>>>> -             unsigned long reserved:13;
>>>> -#else
>>>> -             unsigned long reserved:45;
>>>> -#endif
>>>> -             unsigned long type:1;
>>>> -     };
>>>> -};
>>>> -
>>>> /*
>>>>   * RISC-V doesn't have hetergenous harts yet. This need to be part of
>>>>   * per_cpu in case of harts with different pmu counters
>
> Atish just poked me about this as I was tagging my PR, I squashed it in.
> I don't see a message on lore, not sure if something went off the rails
> or if it's just in the outbox on my box at home.

I assume stuck in the outbox. I checked out for-next after I got home
from work and the error was gone so I assumed you'd done so.

:)