Quoting Matthew Auld (2018-03-01 10:18:55)
> Add some onion to populate_lr_context.
>
> Fixes: d2b4b97933f5 ("drm/i915: Record the default hw state after reset upon
> load")
> Signed-off-by: Matthew Auld <[email protected]>
> Cc: Chris Wilson <[email protected]>
Worth backporting? Not sure, as the leak will be forcibly freed when
the context is freed. We should spit a warning in that case, which as
you haven't quoted I presume this was found by inspection rather than
hit?
> drivers/gpu/drm/i915/intel_lrc.c | 10 ++++++----
> 1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_lrc.c
> b/drivers/gpu/drm/i915/intel_lrc.c
> index 14288743909f..9dd6440eb0e1 100644
> --- a/drivers/gpu/drm/i915/intel_lrc.c
> +++ b/drivers/gpu/drm/i915/intel_lrc.c
> @@ -2318,8 +2318,10 @@ populate_lr_context(struct i915_gem_context *ctx,
>
> defaults = i915_gem_object_pin_map(engine->default_state,
> I915_MAP_WB);
> - if (IS_ERR(defaults))
> - return PTR_ERR(defaults);
> + if (IS_ERR(defaults)) {
> + ret = PTR_ERR(defaults);
> + goto error;
> + }
>
> memcpy(vaddr + start, defaults + start, engine->context_size);
> i915_gem_object_unpin_map(engine->default_state);
> @@ -2337,9 +2339,9 @@ populate_lr_context(struct i915_gem_context *ctx,
>
> _MASKED_BIT_ENABLE(CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT |
> CTX_CTRL_ENGINE_CTX_SAVE_INHIBIT);
>
> +error:
err_unpin_ctx: or out_unpin_ctx: if Joonas is listening.
Reviewed-by: Chris Wilson <[email protected]>
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
