Re: [Valgrind-developers] extra syscall_cancel frames

[Adhemerval Z. N. <adh...@li...>]

On 21/03/25 08:54, Florian Weimer wrote:
> * Adhemerval Zanella Netto:
> 
>> On 20/03/25 19:10, Florian Weimer wrote:
>>> * Mark Wielaard:
>>>
>>>>> I think __internal_syscall_cancel should get inlined into
>>>>> __syscall_cancel.
>>>>
>>>> It isn't, I double checked with gdb and there are always two extra
>>>> frames on top of the call stack.
>>>
>>> Oh, right, it's not static.
>>
>> And we need __always_inline to proper inline it. I will send a patch
>> to change it.
> 
> I think there are expected to be external callers from the rest of
> glibc, so we can't inline it.

We can add a common implementation and implement __syscall_cancel and 
__internal_syscall_cancel on top of it.