drm/i915: Ditch i915 globals shrink infrastructure

This essentially reverts

commit 84a10749
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Jan 24 11:36:08 2018 +0000

    drm/i915: Shrink the GEM kmem_caches upon idling

mm/vmscan.c:do_shrink_slab() is a thing, if there's an issue with it
then we need to fix that there, not hand-roll our own slab shrinking
code in i915.

Also when this was added there was only one other caller of
kmem_cache_shrink (added 2005 to the acpi code). Now there's a 2nd one
outside of i915 code in a kunit test, which seems legit since that
wants to very carefully control what's in the kmem_cache. This out of
a total of over 500 calls to kmem_cache_create. This alone should have
been warning sign enough that we're doing something silly.

Noticed while reviewing a patch set from Jason to fix up some issues
in our i915_init() and i915_exit() module load/cleanup code. Now that
i915_globals.c isn't any different than normal init/exit functions, we
should convert them over to one unified table and remove
i915_globals.[hc] entirely.

v2: Improve commit message (Jason)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Cc: David Airlie <airlied@linux.ie>
Cc: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210721183229.4136488-1-daniel.vetter@ffwll.ch

drivers/gpu/drm/i915/gem/i915_gem_context.c

@@ -2280,18 +2280,12 @@ i915_gem_engines_iter_next(struct i915_gem_engines_iter *it)
 #include "selftests/i915_gem_context.c"
 #endif
 
-static void i915_global_gem_context_shrink(void)
-{
-	kmem_cache_shrink(global.slab_luts);
-}
-
 static void i915_global_gem_context_exit(void)
 {
 	kmem_cache_destroy(global.slab_luts);
 }
 
 static struct i915_global_gem_context global = { {
-	.shrink = i915_global_gem_context_shrink,
 	.exit = i915_global_gem_context_exit,
 } };
 

drivers/gpu/drm/i915/gem/i915_gem_object.c

@@ -664,18 +664,12 @@ void i915_gem_init__objects(struct drm_i915_private *i915)
 	INIT_WORK(&i915->mm.free_work, __i915_gem_free_work);
 }
 
-static void i915_global_objects_shrink(void)
-{
-	kmem_cache_shrink(global.slab_objects);
-}
-
 static void i915_global_objects_exit(void)
 {
 	kmem_cache_destroy(global.slab_objects);
 }
 
 static struct i915_global_object global = { {
-	.shrink = i915_global_objects_shrink,
 	.exit = i915_global_objects_exit,
 } };
 

drivers/gpu/drm/i915/gt/intel_context.c

@@ -398,18 +398,12 @@ void intel_context_fini(struct intel_context *ce)
 	i915_active_fini(&ce->active);
 }
 
-static void i915_global_context_shrink(void)
-{
-	kmem_cache_shrink(global.slab_ce);
-}
-
 static void i915_global_context_exit(void)
 {
 	kmem_cache_destroy(global.slab_ce);
 }
 
 static struct i915_global_context global = { {
-	.shrink = i915_global_context_shrink,
 	.exit = i915_global_context_exit,
 } };
 

drivers/gpu/drm/i915/gt/intel_gt_pm.c

@@ -67,8 +67,6 @@ static int __gt_unpark(struct intel_wakeref *wf)
 
 	GT_TRACE(gt, "\n");
 
-	i915_globals_unpark();
-
 	/*
 	 * It seems that the DMC likes to transition between the DC states a lot
 	 * when there are no connected displays (no active power domains) during
@@ -116,8 +114,6 @@ static int __gt_park(struct intel_wakeref *wf)
 	GEM_BUG_ON(!wakeref);
 	intel_display_power_put_async(i915, POWER_DOMAIN_GT_IRQ, wakeref);
 
-	i915_globals_park();
-
 	return 0;
 }
 

drivers/gpu/drm/i915/i915_active.c

@@ -1176,18 +1176,12 @@ struct i915_active *i915_active_create(void)
 #include "selftests/i915_active.c"
 #endif
 
-static void i915_global_active_shrink(void)
-{
-	kmem_cache_shrink(global.slab_cache);
-}
-
 static void i915_global_active_exit(void)
 {
 	kmem_cache_destroy(global.slab_cache);
 }
 
 static struct i915_global_active global = { {
-	.shrink = i915_global_active_shrink,
 	.exit = i915_global_active_exit,
 } };
 

drivers/gpu/drm/i915/i915_globals.c

@@ -17,61 +17,8 @@
 
 static LIST_HEAD(globals);
 
-static atomic_t active;
-static atomic_t epoch;
-static struct park_work {
-	struct delayed_work work;
-	struct rcu_head rcu;
-	unsigned long flags;
-#define PENDING 0
-	int epoch;
-} park;
-
-static void i915_globals_shrink(void)
-{
-	struct i915_global *global;
-
-	/*
-	 * kmem_cache_shrink() discards empty slabs and reorders partially
-	 * filled slabs to prioritise allocating from the mostly full slabs,
-	 * with the aim of reducing fragmentation.
-	 */
-	list_for_each_entry(global, &globals, link)
-		global->shrink();
-}
-
-static void __i915_globals_grace(struct rcu_head *rcu)
-{
-	/* Ratelimit parking as shrinking is quite slow */
-	schedule_delayed_work(&park.work, round_jiffies_up_relative(2 * HZ));
-}
-
-static void __i915_globals_queue_rcu(void)
-{
-	park.epoch = atomic_inc_return(&epoch);
-	if (!atomic_read(&active)) {
-		init_rcu_head(&park.rcu);
-		call_rcu(&park.rcu, __i915_globals_grace);
-	}
-}
-
-static void __i915_globals_park(struct work_struct *work)
-{
-	destroy_rcu_head(&park.rcu);
-
-	/* Confirm nothing woke up in the last grace period */
-	if (park.epoch != atomic_read(&epoch)) {
-		__i915_globals_queue_rcu();
-		return;
-	}
-
-	clear_bit(PENDING, &park.flags);
-	i915_globals_shrink();
-}
-
 void __init i915_global_register(struct i915_global *global)
 {
-	GEM_BUG_ON(!global->shrink);
 	GEM_BUG_ON(!global->exit);
 
 	list_add_tail(&global->link, &globals);
@@ -109,52 +56,10 @@ int __init i915_globals_init(void)
 		}
 	}
 
-	INIT_DELAYED_WORK(&park.work, __i915_globals_park);
 	return 0;
 }
 
-void i915_globals_park(void)
-{
-	/*
-	 * Defer shrinking the global slab caches (and other work) until
-	 * after a RCU grace period has completed with no activity. This
-	 * is to try and reduce the latency impact on the consumers caused
-	 * by us shrinking the caches the same time as they are trying to
-	 * allocate, with the assumption being that if we idle long enough
-	 * for an RCU grace period to elapse since the last use, it is likely
-	 * to be longer until we need the caches again.
-	 */
-	if (!atomic_dec_and_test(&active))
-		return;
-
-	/* Queue cleanup after the next RCU grace period has freed slabs */
-	if (!test_and_set_bit(PENDING, &park.flags))
-		__i915_globals_queue_rcu();
-}
-
-void i915_globals_unpark(void)
-{
-	atomic_inc(&epoch);
-	atomic_inc(&active);
-}
-
-static void __exit __i915_globals_flush(void)
-{
-	atomic_inc(&active); /* skip shrinking */
-
-	rcu_barrier(); /* wait for the work to be queued */
-	flush_delayed_work(&park.work);
-
-	atomic_dec(&active);
-}
-
 void __exit i915_globals_exit(void)
 {
-	GEM_BUG_ON(atomic_read(&active));
-
-	__i915_globals_flush();
 	__i915_globals_cleanup();
-
-	/* And ensure that our DESTROY_BY_RCU slabs are truly destroyed */
-	rcu_barrier();
 }

drivers/gpu/drm/i915/i915_globals.h

@@ -14,15 +14,12 @@ typedef void (*i915_global_func_t)(void);
 struct i915_global {
 	struct list_head link;
 
-	i915_global_func_t shrink;
 	i915_global_func_t exit;
 };
 
 void i915_global_register(struct i915_global *global);
 
 int i915_globals_init(void);
-void i915_globals_park(void);
-void i915_globals_unpark(void);
 void i915_globals_exit(void);
 
 /* constructors */

drivers/gpu/drm/i915/i915_request.c

@@ -2077,12 +2077,6 @@ void i915_request_show(struct drm_printer *m,
 #include "selftests/i915_request.c"
 #endif
 
-static void i915_global_request_shrink(void)
-{
-	kmem_cache_shrink(global.slab_execute_cbs);
-	kmem_cache_shrink(global.slab_requests);
-}
-
 static void i915_global_request_exit(void)
 {
 	kmem_cache_destroy(global.slab_execute_cbs);
@@ -2090,7 +2084,6 @@ static void i915_global_request_exit(void)
 }
 
 static struct i915_global_request global = { {
-	.shrink = i915_global_request_shrink,
 	.exit = i915_global_request_exit,
 } };
 

drivers/gpu/drm/i915/i915_scheduler.c

@@ -475,12 +475,6 @@ i915_sched_engine_create(unsigned int subclass)
 	return sched_engine;
 }
 
-static void i915_global_scheduler_shrink(void)
-{
-	kmem_cache_shrink(global.slab_dependencies);
-	kmem_cache_shrink(global.slab_priorities);
-}
-
 static void i915_global_scheduler_exit(void)
 {
 	kmem_cache_destroy(global.slab_dependencies);
@@ -488,7 +482,6 @@ static void i915_global_scheduler_exit(void)
 }
 
 static struct i915_global_scheduler global = { {
-	.shrink = i915_global_scheduler_shrink,
 	.exit = i915_global_scheduler_exit,
 } };
 

drivers/gpu/drm/i915/i915_vma.c

@@ -1414,18 +1414,12 @@ void i915_vma_make_purgeable(struct i915_vma *vma)
 #include "selftests/i915_vma.c"
 #endif
 
-static void i915_global_vma_shrink(void)
-{
-	kmem_cache_shrink(global.slab_vmas);
-}
-
 static void i915_global_vma_exit(void)
 {
 	kmem_cache_destroy(global.slab_vmas);
 }
 
 static struct i915_global_vma global = { {
-	.shrink = i915_global_vma_shrink,
 	.exit = i915_global_vma_exit,
 } };