intel-gfx.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: Chris Wilson <chris@chris-wilson.co.uk>
To: intel-gfx@lists.freedesktop.org
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Subject: [Intel-gfx] [PATCH 08/22] drm/i915: Improve execute_cb struct packing
Date: Wed, 20 May 2020 08:54:49 +0100	[thread overview]
Message-ID: <20200520075503.10388-8-chris@chris-wilson.co.uk> (raw)
In-Reply-To: <20200520075503.10388-1-chris@chris-wilson.co.uk>

Reduce the irq_work llist for attaching the callbacks to the signal for
both smaller structs (two fewer pointers!) and simpler [debug] code:

Function                                     old     new   delta
irq_execute_cb                                35      34      -1
__igt_breadcrumbs_smoketest                 1684    1682      -2
i915_request_retire                         2003    1996      -7
__i915_request_create                       1047    1040      -7
__notify_execute_cb                          135     126      -9
__i915_request_ctor                          188     178     -10
__await_execution.part.constprop             451     440     -11
igt_wait_request                             924     714    -210

One minor artifact is that the order of cb exection is reversed. No
current use cases are affected by that change.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_request.c | 18 +++++++++---------
 drivers/gpu/drm/i915/i915_request.h |  2 +-
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 8e5511bc0f22..6f8b5794b73c 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -42,7 +42,6 @@
 #include "intel_pm.h"
 
 struct execute_cb {
-	struct list_head link;
 	struct irq_work work;
 	struct i915_sw_fence *fence;
 	void (*hook)(struct i915_request *rq, struct dma_fence *signal);
@@ -179,14 +178,14 @@ static void irq_execute_cb_hook(struct irq_work *wrk)
 
 static void __notify_execute_cb(struct i915_request *rq)
 {
-	struct execute_cb *cb;
+	struct execute_cb *cb, *cn;
 
 	lockdep_assert_held(&rq->lock);
 
-	if (list_empty(&rq->execute_cb))
+	if (llist_empty(&rq->execute_cb))
 		return;
 
-	list_for_each_entry(cb, &rq->execute_cb, link)
+	llist_for_each_entry_safe(cb, cn, rq->execute_cb.first, work.llnode)
 		irq_work_queue(&cb->work);
 
 	/*
@@ -199,7 +198,7 @@ static void __notify_execute_cb(struct i915_request *rq)
 	 * preempt-to-idle cycle on the target engine, all the while the
 	 * master execute_cb may refire.
 	 */
-	INIT_LIST_HEAD(&rq->execute_cb);
+	rq->execute_cb.first = NULL;
 }
 
 static inline void
@@ -317,7 +316,7 @@ bool i915_request_retire(struct i915_request *rq)
 		set_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags);
 		__notify_execute_cb(rq);
 	}
-	GEM_BUG_ON(!list_empty(&rq->execute_cb));
+	GEM_BUG_ON(!llist_empty(&rq->execute_cb));
 	spin_unlock_irq(&rq->lock);
 
 	remove_from_client(rq);
@@ -385,7 +384,8 @@ __await_execution(struct i915_request *rq,
 		i915_sw_fence_complete(cb->fence);
 		kmem_cache_free(global.slab_execute_cbs, cb);
 	} else {
-		list_add_tail(&cb->link, &signal->execute_cb);
+		cb->work.llnode.next = signal->execute_cb.first;
+		signal->execute_cb.first = &cb->work.llnode;
 	}
 	spin_unlock_irq(&signal->lock);
 
@@ -694,7 +694,7 @@ static void __i915_request_ctor(void *arg)
 	rq->file_priv = NULL;
 	rq->capture_list = NULL;
 
-	INIT_LIST_HEAD(&rq->execute_cb);
+	init_llist_head(&rq->execute_cb);
 }
 
 struct i915_request *
@@ -784,7 +784,7 @@ __i915_request_create(struct intel_context *ce, gfp_t gfp)
 	rq->batch = NULL;
 	GEM_BUG_ON(rq->file_priv);
 	GEM_BUG_ON(rq->capture_list);
-	GEM_BUG_ON(!list_empty(&rq->execute_cb));
+	GEM_BUG_ON(!llist_empty(&rq->execute_cb));
 
 	/*
 	 * Reserve space in the ring buffer for all the commands required to
diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h
index 8ec7ee4dbadc..5d4709a3dace 100644
--- a/drivers/gpu/drm/i915/i915_request.h
+++ b/drivers/gpu/drm/i915/i915_request.h
@@ -214,7 +214,7 @@ struct i915_request {
 			ktime_t emitted;
 		} duration;
 	};
-	struct list_head execute_cb;
+	struct llist_head execute_cb;
 	struct i915_sw_fence semaphore;
 
 	/*
-- 
2.20.1

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

  parent reply	other threads:[~2020-05-20  7:55 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-20  7:54 [Intel-gfx] [PATCH 01/22] drm/i915/gem: Suppress some random warnings Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 02/22] drm/i915/execlists: Shortcircuit queue_prio() for no internal levels Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 03/22] drm/i915: Avoid using rq->engine after free during i915_fence_release Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 04/22] drm/i915: Move saturated workload detection back to the context Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 05/22] drm/i915/gt: Use virtual_engine during execlists_dequeue Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 06/22] drm/i915/gt: Decouple inflight virtual engines Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 07/22] drm/i915/gt: Resubmit the virtual engine on schedule-out Chris Wilson
2020-05-20  7:54 ` Chris Wilson [this message]
2020-05-20  7:54 ` [Intel-gfx] [PATCH 09/22] dma-buf: Proxy fence, an unsignaled fence placeholder Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 10/22] drm/syncobj: Allow use of dma-fence-proxy Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 11/22] drm/i915/gem: Teach execbuf how to wait on future syncobj Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 12/22] drm/i915/gem: Allow combining submit-fences with syncobj Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 13/22] drm/i915/gt: Declare when we enabled timeslicing Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 14/22] drm/i915/gt: Use built-in active intel_context reference Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 15/22] drm/i915: Drop I915_IDLE_ENGINES_TIMEOUT Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 16/22] drm/i915: Always defer fenced work to the worker Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 17/22] drm/i915/gem: Assign context id for async work Chris Wilson
2020-05-20  7:54 ` [Intel-gfx] [PATCH 18/22] drm/i915: Export a preallocate variant of i915_active_acquire() Chris Wilson
2020-05-20  7:55 ` [Intel-gfx] [PATCH 19/22] drm/i915/gem: Separate the ww_mutex walker into its own list Chris Wilson
2020-05-20  7:55 ` [Intel-gfx] [PATCH 20/22] drm/i915/gem: Asynchronous GTT unbinding Chris Wilson
2020-05-20  7:55 ` [Intel-gfx] [PATCH 21/22] drm/i915/gem: Bind the fence async for execbuf Chris Wilson
2020-05-20  7:55 ` [Intel-gfx] [PATCH 22/22] drm/i915: Micro-optimise i915_request_completed() Chris Wilson
2020-05-20  8:37 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for series starting with [01/22] drm/i915/gem: Suppress some random warnings Patchwork
2020-05-20  8:38 ` [Intel-gfx] ✗ Fi.CI.SPARSE: " Patchwork
2020-05-20  9:03 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2020-05-20 20:47 ` [Intel-gfx] ✗ Fi.CI.IGT: failure " 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=20200520075503.10388-8-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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).