On Thu, Jul 22, 2021 at 11:02:55AM +0100, Tvrtko Ursulin wrote:
> On 21/07/2021 19:32, Daniel Vetter wrote:
> > This essentially reverts
> >
> > commit 84a1074920523430f9dc30ff907f4801b4820072
> > Author: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> > 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.
>
> This is somewhat incomplete statement which ignores a couple of angles so I
> wish there was a bit more time to respond before steam rolling it in. :(
>
> The removed code was not a hand rolled shrinker, but about managing slab
> sizes in face of bursty workloads. Core code does not know when i915 is
> active and when it is idle, so calling kmem_cache_shrink() after going idle
> wass supposed to help with house keeping by doing house keeping work outside
> of the latency sensitive phase.
>
> To "fix" (improve really) it in core as you suggest, would need some method
> of signaling when a slab user feels is an opportunte moment to do this house
> keeping. And kmem_cache_shrink is just that so I don't see the problem.
>
> Granted, argument kmem_cache_shrink is not much used is a valid one so
> discussion overall is definitely valid. Becuase on the higher level we could
> definitely talk about which workloads actually benefit from this code and
> how much which probably no one knows at this point.
>
> But in general I think you needed to leave more time for discussion. 12
> hours is way too short.
It's 500+ users of kmem_cache_create vs i915 doing kmem_cache_shrink. And
I guarantee you there's slab users that churn through more allocations
than we do, and are more bursty.
An extraordinary claim like this needs extraordinary evidence. And then a
discussion with core mm/ folks so that we can figure out how to solve the
discovered problem best for the other 500+ users of slabs in-tree, so that
everyone benefits. Not just i915 gpu workloads.
-Daniel
> > 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.
> >
> > Cc: David Airlie <airlied@xxxxxxxx>
> > Cc: Jason Ekstrand <jason@xxxxxxxxxxxxxx>
> > Signed-off-by: Daniel Vetter <daniel.vetter@xxxxxxxxx>
> > ---
> > drivers/gpu/drm/i915/gem/i915_gem_context.c | 6 --
> > drivers/gpu/drm/i915/gem/i915_gem_object.c | 6 --
> > drivers/gpu/drm/i915/gt/intel_context.c | 6 --
> > drivers/gpu/drm/i915/gt/intel_gt_pm.c | 4 -
> > drivers/gpu/drm/i915/i915_active.c | 6 --
> > drivers/gpu/drm/i915/i915_globals.c | 95 ---------------------
> > drivers/gpu/drm/i915/i915_globals.h | 3 -
> > drivers/gpu/drm/i915/i915_request.c | 7 --
> > drivers/gpu/drm/i915/i915_scheduler.c | 7 --
> > drivers/gpu/drm/i915/i915_vma.c | 6 --
> > 10 files changed, 146 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
> > index 7d6f52d8a801..bf2a2319353a 100644
> > --- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
> > +++ b/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,
> > } };
> > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > index 9da7b288b7ed..5c21cff33199 100644
> > --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c
> > +++ b/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,
> > } };
> > diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
> > index bd63813c8a80..c1338441cc1d 100644
> > --- a/drivers/gpu/drm/i915/gt/intel_context.c
> > +++ b/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,
> > } };
> > diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm.c b/drivers/gpu/drm/i915/gt/intel_gt_pm.c
> > index aef3084e8b16..d86825437516 100644
> > --- a/drivers/gpu/drm/i915/gt/intel_gt_pm.c
> > +++ b/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;
> > }
> > diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
> > index b1aa1c482c32..91723123ae9f 100644
> > --- a/drivers/gpu/drm/i915/i915_active.c
> > +++ b/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,
> > } };
> > diff --git a/drivers/gpu/drm/i915/i915_globals.c b/drivers/gpu/drm/i915/i915_globals.c
> > index 77f1911c463b..7fe2e503897b 100644
> > --- a/drivers/gpu/drm/i915/i915_globals.c
> > +++ b/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();
> > }
> > diff --git a/drivers/gpu/drm/i915/i915_globals.h b/drivers/gpu/drm/i915/i915_globals.h
> > index 2d199f411a4a..9e6b4fd07528 100644
> > --- a/drivers/gpu/drm/i915/i915_globals.h
> > +++ b/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 */
> > diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
> > index 09ebea9a0090..d3de9f60e03a 100644
> > --- a/drivers/gpu/drm/i915/i915_request.c
> > +++ b/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,
> > } };
> > diff --git a/drivers/gpu/drm/i915/i915_scheduler.c b/drivers/gpu/drm/i915/i915_scheduler.c
> > index 3a58a9130309..561c649e59f7 100644
> > --- a/drivers/gpu/drm/i915/i915_scheduler.c
> > +++ b/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,
> > } };
> > diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
> > index 5b9dce0f443b..09a7c47926f7 100644
> > --- a/drivers/gpu/drm/i915/i915_vma.c
> > +++ b/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,
> > } };
> >
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
