Hi Matt,
On Thursday, 15 September 2022 00:04:24 CEST Matt Roper wrote:
> From: Chris Wilson <chris.p.wilson@xxxxxxxxx>
>
> If we abort driver initialisation in the middle of gt/engine discovery,
> some engines will be fully setup and some not. Those incompletely setup
> engines only have 'engine->release == NULL' and so will leak any of the
> common objects allocated.
>
> Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> Cc: Janusz Krzysztofik <janusz.krzysztofik@xxxxxxxxxxxxxxx>
> Signed-off-by: Matt Roper <matthew.d.roper@xxxxxxxxx>
> ---
> drivers/gpu/drm/i915/gt/intel_engine_cs.c | 16 +++++++++++++---
> 1 file changed, 13 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
> index 1f7188129cd1..bff12b4ec314 100644
> --- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
> +++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
> @@ -1196,6 +1196,12 @@ void intel_engine_destroy_pinned_context(struct intel_context *ce)
> intel_context_put(ce);
> }
>
> +static void destroy_pinned_context(struct intel_context *ce)
> +{
> + if (ce)
> + intel_engine_destroy_pinned_context(ce);
> +}
> +
> static struct intel_context *
> create_kernel_context(struct intel_engine_cs *engine)
> {
> @@ -1274,8 +1280,13 @@ int intel_engines_init(struct intel_gt *gt)
> return err;
>
> err = setup(engine);
> - if (err)
> + if (err) {
> + intel_engine_cleanup_common(engine);
> return err;
> + }
> +
> + /* The backend should now be responsible for cleanup */
> + GEM_BUG_ON(engine->release == NULL);
LGTM
>
> err = engine_init_common(engine);
> if (err)
> @@ -1307,8 +1318,7 @@ void intel_engine_cleanup_common(struct intel_engine_cs *engine)
> if (engine->default_state)
> fput(engine->default_state);
>
> - if (engine->kernel_context)
> - intel_engine_destroy_pinned_context(engine->kernel_context);
> + destroy_pinned_context(engine->kernel_context);
I think there is no point for this hunk, and consequently for the first one.
A version of intel_engine_destroy_pinned_context() that accepts NULL can be
added later, when we see need for similar code optimisation in more places, if
ever.
Thanks,
Janusz
>
> GEM_BUG_ON(!llist_empty(&engine->barrier_tasks));
> cleanup_status_page(engine);
>
