| Message ID | 20191107203722.8814-1-matthew.d.roper@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drm/i915: Don't test plane stride with !INTEL_DISPLAY_ENABLED | expand |
On Thu, Nov 07, 2019 at 12:37:22PM -0800, Matt Roper wrote: > If INTEL_DISPLAY_ENABLED is false, then the modesetting resources were > never initialized. Userspace may still call DRM_IOCTL_MODE_CREATE_DUMB > which will eventually lead i915_gem_dumb_create() to try to dereference > a non-existent pipe A primary plane while figuring out a proper pitch. > > We could force dumb buffer creation to fail in cases where display isn't > enabled (since there isn't really a definition of a "suitable for > scanout" buffer in that case), but it's easier (and probably less > invasive to IGT tests) to just drop the attempt to align to plane max > stride in cases where the display isn't enabled. I just pushed my earlier patch which does it the other way around. But yeah, we should perhaps just reject the entire thing. > > Cc: Lucas De Marchi <lucas.demarchi@intel.com> > Cc: Jani Nikula <jani.nikula@intel.com> > Signed-off-by: Matt Roper <matthew.d.roper@intel.com> > --- > drivers/gpu/drm/i915/i915_gem.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c > index 143a8952b736..f022a17328b0 100644 > --- a/drivers/gpu/drm/i915/i915_gem.c > +++ b/drivers/gpu/drm/i915/i915_gem.c > @@ -229,7 +229,8 @@ i915_gem_dumb_create(struct drm_file *file, > args->pitch = ALIGN(args->width * cpp, 64); > > /* align stride to page size so that we can remap */ > - if (args->pitch > intel_plane_fb_max_stride(to_i915(dev), format, > + if (INTEL_DISPLAY_ENABLED(to_i915(dev)) && > + args->pitch > intel_plane_fb_max_stride(to_i915(dev), format, > DRM_FORMAT_MOD_LINEAR)) > args->pitch = ALIGN(args->pitch, 4096); > > -- > 2.21.0 > > _______________________________________________ > Intel-gfx mailing list > Intel-gfx@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 143a8952b736..f022a17328b0 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -229,7 +229,8 @@ i915_gem_dumb_create(struct drm_file *file, args->pitch = ALIGN(args->width * cpp, 64); /* align stride to page size so that we can remap */ - if (args->pitch > intel_plane_fb_max_stride(to_i915(dev), format, + if (INTEL_DISPLAY_ENABLED(to_i915(dev)) && + args->pitch > intel_plane_fb_max_stride(to_i915(dev), format, DRM_FORMAT_MOD_LINEAR)) args->pitch = ALIGN(args->pitch, 4096);
If INTEL_DISPLAY_ENABLED is false, then the modesetting resources were never initialized. Userspace may still call DRM_IOCTL_MODE_CREATE_DUMB which will eventually lead i915_gem_dumb_create() to try to dereference a non-existent pipe A primary plane while figuring out a proper pitch. We could force dumb buffer creation to fail in cases where display isn't enabled (since there isn't really a definition of a "suitable for scanout" buffer in that case), but it's easier (and probably less invasive to IGT tests) to just drop the attempt to align to plane max stride in cases where the display isn't enabled. Cc: Lucas De Marchi <lucas.demarchi@intel.com> Cc: Jani Nikula <jani.nikula@intel.com> Signed-off-by: Matt Roper <matthew.d.roper@intel.com> --- drivers/gpu/drm/i915/i915_gem.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)