On Tue, Mar 23, 2021 at 04:50:06PM +0100, Maarten Lankhorst wrote: > With userptr fixed, there is no need for all separate lockdep classes > now, and we can remove all lockdep tricks used. A trylock in the > shrinker is all we need now to flatten the locking hierarchy. > > Signed-off-by: Maarten Lankhorst <maarten.lankho...@linux.intel.com> > Reviewed-by: Thomas Hellström <thomas.hellst...@linux.intel.com>
Minor conflict with this one here when applying, due to lack of https://lore.kernel.org/intel-gfx/20210202154318.19246-1-ch...@chris-wilson.co. It's not a bad idea, but if we do this we really should move this helper into the right headers, and ideally get linux-mm/-fs folks to ack this as a good idea. Especially Dave Chinner very much wants to be in the loop for anything that touches fs_reclaim lockdep annotations. -Daniel > --- > drivers/gpu/drm/i915/gem/i915_gem_object.c | 5 +-- > drivers/gpu/drm/i915/gem/i915_gem_object.h | 20 ++---------- > drivers/gpu/drm/i915/gem/i915_gem_pages.c | 34 ++++++++++---------- > drivers/gpu/drm/i915/gem/i915_gem_phys.c | 2 +- > drivers/gpu/drm/i915/gem/i915_gem_shrinker.c | 10 +++--- > drivers/gpu/drm/i915/gem/i915_gem_userptr.c | 2 +- > 6 files changed, 27 insertions(+), 46 deletions(-) > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c > b/drivers/gpu/drm/i915/gem/i915_gem_object.c > index 6083b9c14be6..821cb40f8d73 100644 > --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c > +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c > @@ -62,7 +62,7 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj, > const struct drm_i915_gem_object_ops *ops, > struct lock_class_key *key, unsigned flags) > { > - __mutex_init(&obj->mm.lock, ops->name ?: "obj->mm.lock", key); > + mutex_init(&obj->mm.lock); > > spin_lock_init(&obj->vma.lock); > INIT_LIST_HEAD(&obj->vma.list); > @@ -86,9 +86,6 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj, > mutex_init(&obj->mm.get_page.lock); > INIT_RADIX_TREE(&obj->mm.get_dma_page.radix, GFP_KERNEL | __GFP_NOWARN); > mutex_init(&obj->mm.get_dma_page.lock); > - > - if (IS_ENABLED(CONFIG_LOCKDEP) && i915_gem_object_is_shrinkable(obj)) > - fs_reclaim_taints_mutex(&obj->mm.lock); > } > > /** > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.h > b/drivers/gpu/drm/i915/gem/i915_gem_object.h > index b5af9c440ac5..a0e1c4ff0de4 100644 > --- a/drivers/gpu/drm/i915/gem/i915_gem_object.h > +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.h > @@ -372,27 +372,10 @@ void __i915_gem_object_set_pages(struct > drm_i915_gem_object *obj, > int ____i915_gem_object_get_pages(struct drm_i915_gem_object *obj); > int __i915_gem_object_get_pages(struct drm_i915_gem_object *obj); > > -enum i915_mm_subclass { /* lockdep subclass for obj->mm.lock/struct_mutex */ > - I915_MM_NORMAL = 0, > - /* > - * Only used by struct_mutex, when called "recursively" from > - * direct-reclaim-esque. Safe because there is only every one > - * struct_mutex in the entire system. > - */ > - I915_MM_SHRINKER = 1, > - /* > - * Used for obj->mm.lock when allocating pages. Safe because the object > - * isn't yet on any LRU, and therefore the shrinker can't deadlock on > - * it. As soon as the object has pages, obj->mm.lock nests within > - * fs_reclaim. > - */ > - I915_MM_GET_PAGES = 1, > -}; > - > static inline int __must_check > i915_gem_object_pin_pages(struct drm_i915_gem_object *obj) > { > - might_lock_nested(&obj->mm.lock, I915_MM_GET_PAGES); > + might_lock(&obj->mm.lock); > > if (atomic_inc_not_zero(&obj->mm.pages_pin_count)) > return 0; > @@ -436,6 +419,7 @@ i915_gem_object_unpin_pages(struct drm_i915_gem_object > *obj) > } > > int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj); > +int __i915_gem_object_put_pages_locked(struct drm_i915_gem_object *obj); > void i915_gem_object_truncate(struct drm_i915_gem_object *obj); > void i915_gem_object_writeback(struct drm_i915_gem_object *obj); > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_pages.c > b/drivers/gpu/drm/i915/gem/i915_gem_pages.c > index e7d7650072c5..e947d4c0da1f 100644 > --- a/drivers/gpu/drm/i915/gem/i915_gem_pages.c > +++ b/drivers/gpu/drm/i915/gem/i915_gem_pages.c > @@ -114,7 +114,7 @@ int __i915_gem_object_get_pages(struct > drm_i915_gem_object *obj) > { > int err; > > - err = mutex_lock_interruptible_nested(&obj->mm.lock, I915_MM_GET_PAGES); > + err = mutex_lock_interruptible(&obj->mm.lock); > if (err) > return err; > > @@ -196,21 +196,13 @@ __i915_gem_object_unset_pages(struct > drm_i915_gem_object *obj) > return pages; > } > > -int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj) > +int __i915_gem_object_put_pages_locked(struct drm_i915_gem_object *obj) > { > struct sg_table *pages; > - int err; > > if (i915_gem_object_has_pinned_pages(obj)) > return -EBUSY; > > - /* May be called by shrinker from within get_pages() (on another bo) */ > - mutex_lock(&obj->mm.lock); > - if (unlikely(atomic_read(&obj->mm.pages_pin_count))) { > - err = -EBUSY; > - goto unlock; > - } > - > i915_gem_object_release_mmap_offset(obj); > > /* > @@ -226,14 +218,22 @@ int __i915_gem_object_put_pages(struct > drm_i915_gem_object *obj) > * get_pages backends we should be better able to handle the > * cancellation of the async task in a more uniform manner. > */ > - if (!pages) > - pages = ERR_PTR(-EINVAL); > - > - if (!IS_ERR(pages)) > + if (!IS_ERR_OR_NULL(pages)) > obj->ops->put_pages(obj, pages); > > - err = 0; > -unlock: > + return 0; > +} > + > +int __i915_gem_object_put_pages(struct drm_i915_gem_object *obj) > +{ > + int err; > + > + if (i915_gem_object_has_pinned_pages(obj)) > + return -EBUSY; > + > + /* May be called by shrinker from within get_pages() (on another bo) */ > + mutex_lock(&obj->mm.lock); > + err = __i915_gem_object_put_pages_locked(obj); > mutex_unlock(&obj->mm.lock); > > return err; > @@ -341,7 +341,7 @@ void *i915_gem_object_pin_map(struct drm_i915_gem_object > *obj, > !i915_gem_object_type_has(obj, I915_GEM_OBJECT_HAS_IOMEM)) > return ERR_PTR(-ENXIO); > > - err = mutex_lock_interruptible_nested(&obj->mm.lock, I915_MM_GET_PAGES); > + err = mutex_lock_interruptible(&obj->mm.lock); > if (err) > return ERR_PTR(err); > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_phys.c > b/drivers/gpu/drm/i915/gem/i915_gem_phys.c > index 06c481ff79d8..44329c435cf1 100644 > --- a/drivers/gpu/drm/i915/gem/i915_gem_phys.c > +++ b/drivers/gpu/drm/i915/gem/i915_gem_phys.c > @@ -236,7 +236,7 @@ int i915_gem_object_attach_phys(struct > drm_i915_gem_object *obj, int align) > if (err) > return err; > > - err = mutex_lock_interruptible_nested(&obj->mm.lock, I915_MM_GET_PAGES); > + err = mutex_lock_interruptible(&obj->mm.lock); > if (err) > goto err_unlock; > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c > b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c > index b64a0788381f..3052ef5ad89d 100644 > --- a/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c > +++ b/drivers/gpu/drm/i915/gem/i915_gem_shrinker.c > @@ -49,9 +49,9 @@ static bool unsafe_drop_pages(struct drm_i915_gem_object > *obj, > flags = I915_GEM_OBJECT_UNBIND_TEST; > > if (i915_gem_object_unbind(obj, flags) == 0) > - __i915_gem_object_put_pages(obj); > + return true; > > - return !i915_gem_object_has_pages(obj); > + return false; > } > > static void try_to_writeback(struct drm_i915_gem_object *obj, > @@ -200,10 +200,10 @@ i915_gem_shrink(struct drm_i915_private *i915, > > spin_unlock_irqrestore(&i915->mm.obj_lock, flags); > > - if (unsafe_drop_pages(obj, shrink)) { > + if (unsafe_drop_pages(obj, shrink) && > + mutex_trylock(&obj->mm.lock)) { > /* May arrive from get_pages on another bo */ > - mutex_lock(&obj->mm.lock); > - if (!i915_gem_object_has_pages(obj)) { > + if (!__i915_gem_object_put_pages_locked(obj)) { > try_to_writeback(obj, shrink); > count += obj->base.size >> PAGE_SHIFT; > } > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c > b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c > index 1e42fbc68697..503325e74eff 100644 > --- a/drivers/gpu/drm/i915/gem/i915_gem_userptr.c > +++ b/drivers/gpu/drm/i915/gem/i915_gem_userptr.c > @@ -253,7 +253,7 @@ static int i915_gem_object_userptr_unbind(struct > drm_i915_gem_object *obj, bool > if (GEM_WARN_ON(i915_gem_object_has_pinned_pages(obj))) > return -EBUSY; > > - mutex_lock_nested(&obj->mm.lock, I915_MM_GET_PAGES); > + mutex_lock(&obj->mm.lock); > > pages = __i915_gem_object_unset_pages(obj); > if (!IS_ERR_OR_NULL(pages)) > -- > 2.31.0 > > _______________________________________________ > Intel-gfx mailing list > intel-...@lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/intel-gfx -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
