[PATCH i-g-t v2 5/7] runner: Free settings at the end

[Lucas De Marchi]

On Wed, Jan 22, 2025 at 08:46:33AM -0300, Gustavo Sousa wrote:
>Quoting Lucas De Marchi (2025-01-21 19:57:31-03:00)
>>Keep valgrind happy with the normal allocations so we can find the real
>>leaks. Avoid this kind of leak:
>>
>>        ==806592== 8 bytes in 1 blocks are definitely lost in loss record 46 of 188
>>        ==806592==    at 0x4846828: malloc (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
>>        ==806592==    by 0x4BD534E: strdup (strdup.c:42)
>>        ==806592==    by 0x10E6AA: parse_options (settings.c:863)
>>        ==806592==    by 0x10D2DD: main (runner.c:19)
>>
>>Signed-off-by: Lucas De Marchi <lucas.demarchi at intel.com>
>
>This looks correct, so:
>
>Reviewed-by: Gustavo Sousa <gustavo.sousa at intel.com>
>
>, but it seems we also need to make initialize_execute_state() and
>initialize_execute_state_from_resume() properly clear_settings() when
>things go wrong internally.

Looking at runner/runner.c, there are more places to do it when things
go bad, example create_job_list(). Maybe replace the returns on error
path with `goto cleanup`?

Lucas De Marchi

>
>--
>Gustavo Sousa
>
>>---
>> runner/resume.c | 2 ++
>> runner/runner.c | 2 ++
>> 2 files changed, 4 insertions(+)
>>
>>diff --git a/runner/resume.c b/runner/resume.c
>>index 0f4e42bfa..ed17351c6 100644
>>--- a/runner/resume.c
>>+++ b/runner/resume.c
>>@@ -55,6 +55,8 @@ int main(int argc, char **argv)
>>                 exitcode = 3;
>>         }
>>
>>+        clear_settings(&settings);
>>+
>>         printf("Done.\n");
>>         return exitcode;
>> }
>>diff --git a/runner/runner.c b/runner/runner.c
>>index 4855ad641..258b30b36 100644
>>--- a/runner/runner.c
>>+++ b/runner/runner.c
>>@@ -49,6 +49,8 @@ int main(int argc, char **argv)
>>                 exitcode = 1;
>>         }
>>
>>+        clear_settings(&settings);
>>+
>>         printf("Done.\n");
>>         return exitcode;
>> }
>>--
>>2.48.0
>>