On Thu, Aug 03, 2017 at 02:23:42PM +0200, Daniel Vetter wrote:
> pm_rpm is one of the main offenders for long runtime in our full igt
> run. And hey, there's a quick option already, so make it the default.
>
> This means there's yet another way to enable stress tests, we really
> need to standardize this somehow. Or shred the stress-tests we really
> don't want/need anymore.
Having a standarized way would be a good thing.
We should look into that.
> v2: Fixup the entire help text (David).
>
> Cc: David Weinehall <david.weinehall@xxxxxxxxxxxxxxx>
> Cc: Paulo Zanoni <paulo.r.zanoni@xxxxxxxxx>
> Cc: Arkadiusz Hiler <arkadiusz.hiler@xxxxxxxxx>
> Signed-off-by: Daniel Vetter <daniel.vetter@xxxxxxxxx>
> ---
> tests/pm_rpm.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/tests/pm_rpm.c b/tests/pm_rpm.c
> index 1873c9017b6e..aaf610321edd 100644
> --- a/tests/pm_rpm.c
> +++ b/tests/pm_rpm.c
> @@ -1852,14 +1852,14 @@ static void fences_subtest(bool dpms)
> igt_assert(munmap(buf_ptr, params.fb.size) == 0);
> }
>
> -int rounds = 40;
> +int rounds = 10;
> bool stay = false;
>
> static int opt_handler(int opt, int opt_index, void *data)
> {
> switch (opt) {
> case 'q':
> - rounds = 10;
> + rounds = 50;
> break;
> case 's':
> stay = true;
> @@ -1874,10 +1874,10 @@ static int opt_handler(int opt, int opt_index, void *data)
> int main(int argc, char *argv[])
> {
> const char *help_str =
> - " --quick\t\tMake the stress-tests not stressful, for quick regression testing.\n"
> + " --stress\t\tMake the stress-tests more stressful.\n"
> " --stay\t\tDisable all screen and try to go into runtime pm. Useful for debugging.";
> static struct option long_options[] = {
> - {"quick", 0, 0, 'q'},
> + {"stress", 0, 0, 'q'},
> {"stay", 0, 0, 's'},
I don't like that we repurpose 'q' here, as it is confusing in context
of this patch.
I get that we don't have short options so it's not affecting the outside world.
I would be more happy with other letter, but 's', which would be best is
already taken, then maybe 'l' for "long"?
Anyway, either with 'q' or 'l':
Reviewed-by: Arkadiusz Hiler <arkadiusz.hiler@xxxxxxxxx>
--
Cheers,
Arek
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
