All of lore.kernel.org
 help / color / mirror / Atom feed
From: Chris Wilson <chris@chris-wilson.co.uk>
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH v2] drm/i915: Shrink the GEM kmem_caches upon idling
Date: Tue, 16 Jan 2018 13:05:19 +0000	[thread overview]
Message-ID: <20180116130519.16255-1-chris@chris-wilson.co.uk> (raw)
In-Reply-To: <c945dbeb-05b3-36f9-d06f-abf254f51d51@linux.intel.com>

When we finally decide the gpu is idle, that is a good time to shrink
our kmem_caches.

v2: Comment upon the random sprinkling of rcu_barrier() inside the idle
worker.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_gem.c | 30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 335731c93b4a..61b13fdfaa71 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -4716,6 +4716,21 @@ i915_gem_retire_work_handler(struct work_struct *work)
 	}
 }
 
+static void shrink_caches(struct drm_i915_private *i915)
+{
+	/*
+	 * 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.
+	 */
+	kmem_cache_shrink(i915->priorities);
+	kmem_cache_shrink(i915->dependencies);
+	kmem_cache_shrink(i915->requests);
+	kmem_cache_shrink(i915->luts);
+	kmem_cache_shrink(i915->vmas);
+	kmem_cache_shrink(i915->objects);
+}
+
 static inline bool
 new_requests_since_last_retire(const struct drm_i915_private *i915)
 {
@@ -4803,6 +4818,21 @@ i915_gem_idle_work_handler(struct work_struct *work)
 		GEM_BUG_ON(!dev_priv->gt.awake);
 		i915_queue_hangcheck(dev_priv);
 	}
+
+	/*
+	 * We use magical TYPESAFE_BY_RCU kmem_caches whose pages are not
+	 * returned to the system imediately but only after an RCU grace
+	 * period. We want to encourage such pages to be returned and so
+	 * incorporate a RCU barrier here to provide some rate limiting
+	 * of the driver and flush the old pages before we free a new batch
+	 * from the next round of shrinking.
+	 */
+	rcu_barrier();
+
+	if (!new_requests_since_last_retire(dev_priv)) {
+		__i915_gem_free_work(&dev_priv->mm.free_work);
+		shrink_caches(dev_priv);
+	}
 }
 
 int i915_gem_suspend(struct drm_i915_private *dev_priv)
-- 
2.15.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  parent reply	other threads:[~2018-01-16 13:06 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-01-15 21:24 Prevent trivial oom from gem_exec_nop/sequential Chris Wilson
2018-01-15 21:24 ` [PATCH 01/10] drm/i915: Only attempt to scan the requested number of shrinker slabs Chris Wilson
2018-01-17 10:29   ` Tvrtko Ursulin
2018-01-18  9:16     ` Chris Wilson
2018-01-18  9:19     ` Chris Wilson
2018-01-15 21:24 ` [PATCH 02/10] drm/i915: Move i915_gem_retire_work_handler Chris Wilson
2018-01-17 10:33   ` Tvrtko Ursulin
2018-01-15 21:24 ` [PATCH 03/10] drm/i915: Shrink the GEM kmem_caches upon idling Chris Wilson
2018-01-16 10:00   ` Tvrtko Ursulin
2018-01-16 10:19     ` Chris Wilson
2018-01-16 13:05     ` Chris Wilson [this message]
2018-01-16 15:12       ` [PATCH v2] " Tvrtko Ursulin
2018-01-16 15:16         ` Tvrtko Ursulin
2018-01-16 15:21         ` Chris Wilson
2018-01-16 17:25           ` Tvrtko Ursulin
2018-01-16 17:36             ` Chris Wilson
2018-01-17 10:18               ` Tvrtko Ursulin
2018-01-18 18:06                 ` Chris Wilson
2018-01-15 21:24 ` [PATCH 04/10] drm/i915: Shrink the request kmem_cache on allocation error Chris Wilson
2018-01-16 10:10   ` Tvrtko Ursulin
2018-01-16 10:26     ` Chris Wilson
2018-01-16 13:15     ` [PATCH v2] " Chris Wilson
2018-01-16 15:19       ` Tvrtko Ursulin
2018-01-15 21:24 ` [PATCH 05/10] drm/i915: Trim the retired request queue after submitting Chris Wilson
2018-01-16 10:18   ` Tvrtko Ursulin
2018-01-16 10:32     ` Chris Wilson
2018-01-17 10:23       ` Tvrtko Ursulin
2018-01-16 13:30     ` [PATCH v2] " Chris Wilson
2018-01-15 21:24 ` [PATCH 06/10] drm/i915/breadcrumbs: Drop request reference for the signaler thread Chris Wilson
2018-01-15 21:24 ` [PATCH 07/10] drm/i915: Reduce spinlock hold time during notify_ring() interrupt Chris Wilson
2018-01-17 10:45   ` Tvrtko Ursulin
2018-01-18 18:08     ` Chris Wilson
2018-01-18 18:10     ` Chris Wilson
2018-01-15 21:24 ` [PATCH 08/10] drm/i915: Move the irq_counter inside the spinlock Chris Wilson
2018-01-17 12:12   ` Tvrtko Ursulin
2018-01-15 21:24 ` [PATCH 09/10] drm/i915: Only signal from interrupt when requested Chris Wilson
2018-01-17 12:22   ` Tvrtko Ursulin
2018-01-18 18:12     ` Chris Wilson
2018-01-15 21:24 ` [PATCH 10/10] drm/i915/breadcrumbs: Reduce signaler rbtree to a sorted list Chris Wilson
2018-01-15 22:04 ` ✓ Fi.CI.BAT: success for series starting with [01/10] drm/i915: Only attempt to scan the requested number of shrinker slabs Patchwork
2018-01-16  9:21 ` ✓ Fi.CI.IGT: " Patchwork
2018-01-16  9:52 ` Prevent trivial oom from gem_exec_nop/sequential Tvrtko Ursulin
2018-01-16 10:02   ` Chris Wilson
2018-01-16 13:10   ` Chris Wilson
2018-01-16 13:42 ` ✓ Fi.CI.BAT: success for series starting with [01/10] drm/i915: Only attempt to scan the requested number of shrinker slabs (rev3) Patchwork
2018-01-16 14:02 ` ✓ Fi.CI.BAT: success for series starting with [01/10] drm/i915: Only attempt to scan the requested number of shrinker slabs (rev4) Patchwork
2018-01-16 15:29 ` ✓ Fi.CI.IGT: " Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180116130519.16255-1-chris@chris-wilson.co.uk \
    --to=chris@chris-wilson.co.uk \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.