Hey,

Den 2025-11-12 kl. 16:56, skrev Jani Nikula:
> i915 has the Kconfig option DRM_I915_FENCE_TIMEOUT, defaulting to 10
> seconds. xe doesn't use it, instead defaulting to
> MAX_SCHEDULE_TIMEOUT. Do the same for i915.
> 
> Cc: Maarten Lankhorst <[email protected]>
> Cc: Ville Syrjälä <[email protected]>
> Signed-off-by: Jani Nikula <[email protected]>
> 
> ---
> 
> I am not sure about all the implications here. I'm just wondering if we
> need to go through all the trouble of adding an interface to query this
> from the parent driver.
> ---
>  drivers/gpu/drm/i915/display/intel_display.c  |  4 +---
>  .../drm/xe/compat-i915-headers/i915_config.h  | 19 -------------------
>  2 files changed, 1 insertion(+), 22 deletions(-)
>  delete mode 100644 drivers/gpu/drm/xe/compat-i915-headers/i915_config.h
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_display.c 
> b/drivers/gpu/drm/i915/display/intel_display.c
> index 6bca186608ce..cc502dfe8aff 100644
> --- a/drivers/gpu/drm/i915/display/intel_display.c
> +++ b/drivers/gpu/drm/i915/display/intel_display.c
> @@ -49,7 +49,6 @@
>  #include "g4x_dp.h"
>  #include "g4x_hdmi.h"
>  #include "hsw_ips.h"
> -#include "i915_config.h"
>  #include "i915_drv.h"
>  #include "i915_reg.h"
>  #include "i9xx_plane.h"
> @@ -7150,7 +7149,6 @@ static void skl_commit_modeset_enables(struct 
> intel_atomic_state *state)
>  
>  static void intel_atomic_commit_fence_wait(struct intel_atomic_state 
> *intel_state)
>  {
> -     struct drm_i915_private *i915 = to_i915(intel_state->base.dev);
>       struct drm_plane *plane;
>       struct drm_plane_state *new_plane_state;
>       long ret;
> @@ -7159,7 +7157,7 @@ static void intel_atomic_commit_fence_wait(struct 
> intel_atomic_state *intel_stat
>       for_each_new_plane_in_state(&intel_state->base, plane, new_plane_state, 
> i) {
>               if (new_plane_state->fence) {
>                       ret = dma_fence_wait_timeout(new_plane_state->fence, 
> false,
> -                                                  i915_fence_timeout(i915));
> +                                                  MAX_SCHEDULE_TIMEOUT);
>                       if (ret <= 0)
>                               break;
>  
You should probably use dma_fence_wait instead now. ret == 0 should never 
happen any more either. :-)j

Otherwise looks good.

Reviewed-by: Maarten Lankhorst <[email protected]>

Reply via email to