All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dave Gordon <david.s.gordon@intel.com>
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH 1/4] drm/i915: teardown default context in reverse, update comments
Date: Wed, 16 Dec 2015 18:36:48 +0000	[thread overview]
Message-ID: <1450291011-31486-2-git-send-email-david.s.gordon@intel.com> (raw)
In-Reply-To: <1450291011-31486-1-git-send-email-david.s.gordon@intel.com>

We set up engines in forwards order, so some things (notably the
default context) are "owned" by engine 0 (the render engine, aka "RCS").
For symmetry and to make sure such shared objects don't disappear too
early, we should generally run teardown loops in the reverse order,
so that engine 0 is processed last.

This patch changes i915_gem_context_fini() to do that, and clarifies the
comments in i915_gem_context_{init,fini}() about the refcounting of the
default {struct intel_)context: the refcount is just ONE, no matter how
many rings exist or are active, and this refcount is nominally ascribed
to the render ring (RCS), which is set up first and now torn down last.

Signed-off-by: Dave Gordon <david.s.gordon@intel.com>
---
 drivers/gpu/drm/i915/i915_gem_context.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index 900ffd0..e143ea5 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -391,7 +391,13 @@ int i915_gem_context_init(struct drm_device *dev)
 	for (i = 0; i < I915_NUM_RINGS; i++) {
 		struct intel_engine_cs *ring = &dev_priv->ring[i];
 
-		/* NB: RCS will hold a ref for all rings */
+		/*
+		 * Although each engine has a pointer to the global default
+		 * context, they don't contribute to the refcount on the
+		 * context. We consider that RCS (which is set up first and
+		 * torn down last) holds this reference on behalf of all the
+		 * other engines
+		 */
 		ring->default_context = ctx;
 	}
 
@@ -431,14 +437,21 @@ void i915_gem_context_fini(struct drm_device *dev)
 		i915_gem_object_ggtt_unpin(dctx->legacy_hw_ctx.rcs_state);
 	}
 
-	for (i = 0; i < I915_NUM_RINGS; i++) {
+	for (i = I915_NUM_RINGS; --i >= 0;) {
 		struct intel_engine_cs *ring = &dev_priv->ring[i];
 
-		if (ring->last_context)
+		if (ring->last_context) {
 			i915_gem_context_unreference(ring->last_context);
+			ring->last_context = NULL;
+		}
 
+		/*
+		 * These default_context pointers don't contribute to the
+		 * refcount on the context. We consider that RCS holds its
+		 * reference on behalf of all the other engines, so there's
+		 * just a single unreference() call below.
+		 */
 		ring->default_context = NULL;
-		ring->last_context = NULL;
 	}
 
 	i915_gem_context_unreference(dctx);
-- 
1.9.1

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

  reply	other threads:[~2015-12-16 18:37 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-11 14:59 [PATCH v2] drm/i915: Fix context/engine cleanup order Nick Hoath
2015-12-11 16:26 ` Daniel Vetter
2015-12-11 18:58   ` Daniel Vetter
2015-12-16 18:36 ` tidy up and fix init-fail and teardown paths Dave Gordon
2015-12-16 18:36   ` Dave Gordon [this message]
2015-12-17 10:43     ` [PATCH 1/4] drm/i915: teardown default context in reverse, update comments Nick Hoath
2015-12-21 10:48     ` Daniel Vetter
2015-12-21 11:01       ` Chris Wilson
2015-12-21 11:38         ` Dave Gordon
2015-12-16 18:36   ` [PATCH 2/4] drm/i915: mark the global default (intel_)context as such Dave Gordon
2015-12-16 18:57     ` Chris Wilson
2015-12-16 19:22       ` Dave Gordon
2015-12-16 19:30         ` Chris Wilson
2015-12-17 11:09           ` Nick Hoath
2015-12-17 12:27             ` Chris Wilson
2015-12-17 19:00               ` Dave Gordon
2015-12-18 16:02                 ` Dave Gordon
2015-12-16 18:36   ` [PATCH 3/4] drm/i915: tidy up initialisation failure paths (legacy) Dave Gordon
2015-12-17 11:36     ` Nick Hoath
2015-12-16 18:36   ` [PATCH 4/4] drm/i915: tidy up initialisation failure paths (GEM & LRC) Dave Gordon
2015-12-17 11:37     ` Nick Hoath

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=1450291011-31486-2-git-send-email-david.s.gordon@intel.com \
    --to=david.s.gordon@intel.com \
    --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.