[igt-dev] [PATCH igt] lib: Use C99 initialisers to clear context parameters

[Chris Wilson]

Quoting Tvrtko Ursulin (2018-03-07 15:06:37)
> 
> On 07/03/2018 14:40, Chris Wilson wrote:
> > Quoting Tvrtko Ursulin (2018-03-07 14:36:34)
> >>
> >> On 07/03/2018 12:17, Chris Wilson wrote:
> >>> valgrind complains we feed uninitialised stack into the CONTEXT_SETPARAM
> >>> ioctl. It is unused by the kernel, but valgrind doesn't know that and
> >>> it's easy enough to clear the struct to prevent the warning.
> >>>
> >>> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> >>> ---
> >>>    lib/igt_gt.c | 29 ++++++++++++-----------------
> >>>    1 file changed, 12 insertions(+), 17 deletions(-)
> >>>
> >>> diff --git a/lib/igt_gt.c b/lib/igt_gt.c
> >>> index 799ca1ae..168c5a07 100644
> >>> --- a/lib/igt_gt.c
> >>> +++ b/lib/igt_gt.c
> >>> @@ -126,11 +126,10 @@ void igt_require_hang_ring(int fd, int ring)
> >>>    
> >>>    static unsigned context_get_ban(int fd, unsigned ctx)
> >>>    {
> >>> -     struct drm_i915_gem_context_param param;
> >>> -
> >>> -     param.param = I915_CONTEXT_PARAM_BANNABLE;
> >>> -     param.value = 0;
> >>> -     param.size = 0;
> >>> +     struct drm_i915_gem_context_param param = {
> >>> +             .ctx_id = ctx,
> >>
> >> Where was this assignment before? Was it all broken as well as upsetting
> >> Valgrind?
> > 
> > It existed only in the callers imagination. Broken is an
> > overstatement, since no one varied from the defaults, i.e. the caller
> > has to mix in using PARAM_BANNABLE with igt_hang_ctx.
> 
> gem_reset_stats is doing that, both with default and own context. But I 
> don't feel like untangling this right now. Whatever it is supposed to 
> be, if it breaks, it was broken anyway and not a fault of this patch. so:

gem_reset_stats doesn't exist... We basically need to rewrite the test
from scratch to check expected behaviour.
-Chris