Re: [PATCH 09/14] perf annotate-data: Handle call instructions

[Namhyung Kim <namhyung@xxxxxxxxxx>]

On Tue, Feb 6, 2024 at 3:44 PM Arnaldo Carvalho de Melo
<arnaldo.melo@xxxxxxxxx> wrote:
>
>
>
> On Tue, Feb 6, 2024, 8:36 PM Ian Rogers <irogers@xxxxxxxxxx> wrote:
>>
>> On Tue, Feb 6, 2024 at 3:17 PM Namhyung Kim <namhyung@xxxxxxxxxx> wrote:
>> >
>> > On Fri, Feb 2, 2024 at 7:09 PM Ian Rogers <irogers@xxxxxxxxxx> wrote:
>> > >
>> > > On Fri, Feb 2, 2024 at 2:05 PM Namhyung Kim <namhyung@xxxxxxxxxx> wrote:
>> > > >
>> > > > When updating instruction states, the call instruction should play a
>> > > > role since it can change the register states.  For simplicity, mark some
>> > > > registers as scratch registers (should be arch-dependent), and
>> > > > invalidate them all after a function call.
>> > >
>> > > nit: Volatile or caller-save would be a more conventional name than scratch.
>> >
>> > 'volatile' is a keyword and 'caller_saved' seems somewhat verbose.
>> > Maybe 'temporary'?
>>
>> Sgtm, perhaps temp for brevity and the documentation to call them caller save?
>
>
>
> "caller_saved" seems to be the conventional name doesn't look too long to use to help in reading this code by new people that have read the literature.

Ok, as you both requested, I will use "caller_saved". :)

Thanks,
Namhyung


>
> For instance, from a quick Google search:
>
> https://stackoverflow.com/questions/9268586/what-are-callee-and-caller-saved-registers
>
> - Arnaldo