Re: [PATCH v3] perf record: Add support for arch_sdt_arg_parse_op() on s390

[Mi, Dapeng]

On 3/18/2026 2:27 PM, Thomas Richter wrote:
> On 3/18/26 02:52, Mi, Dapeng wrote:
>> On 3/17/2026 7:06 PM, Thomas Richter wrote:
>>> commit e5e66adfe45a6 ("perf regs: Remove __weak attributive arch_sdt_arg_parse_op() function")
>>> removes arch_sdt_arg_parse_op() functions and s390 support is lost.
>>> The following warning is printed:
>> Not sure if I miss something, but it looks there was also no s390 specific
>> support for arch_sdt_arg_parse_op() before the commit e5e66adfe45a6 ("perf
>> regs: Remove __weak attributive arch_sdt_arg_parse_op() function") and we
>> would see same warning even without the commit e5e66adfe45a6, right?
>>
>>
> Absolutely Correct, but in my opinion it does not matter if it was your patch or if
> you just remove the __weak attribute. Your patch revealed the missing s390 support, which triggered
> this patch.
> If you do not like the wording, what do  you suggest?

I see. The original words lead me think the commit e5e66adfe45a6 ("perf
regs: Remove __weak attributive arch_sdt_arg_parse_op() function") drops
the s390 specific support unexpectedly. :)

So precisely speaking, we may say "the commit e5e66adfe45a6 ("perf regs:
Remove __weak attributive arch_sdt_arg_parse_op() function") introducing
perf_sdt_arg_parse_op() to support architecture-specific argument parsing,
but s390 specific argument parsing is still not supported. So this patch
adds the missing support for s390 ..."

Thanks.


>
> Thanks Thomas