Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Mika Kuoppala <mika.kuoppala@linux.intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: Ben Widawsky <ben@bwidawsk.net>, miku@iki.fi
Subject: [PATCH 1/2] drm/i915: put context upon switching
Date: Thu,  2 May 2013 16:48:07 +0300	[thread overview]
Message-ID: <1367502488-4108-1-git-send-email-mika.kuoppala@intel.com> (raw)

From: Chris Wilson <chris@chris-wilson.co.uk>

In order to be notified of when the context and all of its associated
objects is idle (for if the context maps to a ppgtt) we need a callback
from the retire handler. We can arrange this by using the kref_get/put
of the context for request tracking and by inserting a request to
demarque the switch away from the old context.

[Ben: fixed minor error to patch compile, AND s/last_context/from/]
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
---
 drivers/gpu/drm/i915/i915_gem_context.c |   37 ++++++++++++++++++++-----------
 drivers/gpu/drm/i915/intel_ringbuffer.h |    2 +-
 2 files changed, 25 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index 9e8c685..817222a 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -353,13 +353,13 @@ mi_set_context(struct intel_ring_buffer *ring,
 static int do_switch(struct i915_hw_context *to)
 {
 	struct intel_ring_buffer *ring = to->ring;
-	struct drm_i915_gem_object *from_obj = ring->last_context_obj;
+	struct i915_hw_context *from = ring->last_context;
 	u32 hw_flags = 0;
 	int ret;
 
-	BUG_ON(from_obj != NULL && from_obj->pin_count == 0);
+	BUG_ON(from != NULL && from->obj != NULL && from->obj->pin_count == 0);
 
-	if (from_obj == to->obj)
+	if (from == to)
 		return 0;
 
 	ret = i915_gem_object_pin(to->obj, CONTEXT_ALIGN, false, false);
@@ -382,7 +382,7 @@ static int do_switch(struct i915_hw_context *to)
 
 	if (!to->is_initialized || is_default_context(to))
 		hw_flags |= MI_RESTORE_INHIBIT;
-	else if (WARN_ON_ONCE(from_obj == to->obj)) /* not yet expected */
+	else if (WARN_ON_ONCE(from == to)) /* not yet expected */
 		hw_flags |= MI_FORCE_RESTORE;
 
 	ret = mi_set_context(ring, to, hw_flags);
@@ -397,9 +397,9 @@ static int do_switch(struct i915_hw_context *to)
 	 * is a bit suboptimal because the retiring can occur simply after the
 	 * MI_SET_CONTEXT instead of when the next seqno has completed.
 	 */
-	if (from_obj != NULL) {
-		from_obj->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
-		i915_gem_object_move_to_active(from_obj, ring);
+	if (from != NULL) {
+		from->obj->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
+		i915_gem_object_move_to_active(from->obj, ring);
 		/* As long as MI_SET_CONTEXT is serializing, ie. it flushes the
 		 * whole damn pipeline, we don't need to explicitly mark the
 		 * object dirty. The only exception is that the context must be
@@ -407,15 +407,26 @@ static int do_switch(struct i915_hw_context *to)
 		 * able to defer doing this until we know the object would be
 		 * swapped, but there is no way to do that yet.
 		 */
-		from_obj->dirty = 1;
-		BUG_ON(from_obj->ring != ring);
-		i915_gem_object_unpin(from_obj);
+		from->obj->dirty = 1;
+		BUG_ON(from->obj->ring != ring);
+
+		ret = i915_add_request(ring, NULL, NULL);
+		if (ret) {
+			/* Too late, we've already scheduled a context switch.
+			 * Try to undo the change so that the hw state is
+			 * consistent with out tracking. In case of emergency,
+			 * scream.
+			 */
+			WARN_ON(mi_set_context(ring, from, MI_RESTORE_INHIBIT));
+			return ret;
+		}
 
-		drm_gem_object_unreference(&from_obj->base);
+		i915_gem_object_unpin(from->obj);
+		i915_gem_context_unreference(from);
 	}
 
-	drm_gem_object_reference(&to->obj->base);
-	ring->last_context_obj = to->obj;
+	i915_gem_context_reference(to);
+	ring->last_context = to;
 	to->is_initialized = true;
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index d66208c..dac1614 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -135,7 +135,7 @@ struct  intel_ring_buffer {
 	 */
 	bool itlb_before_ctx_switch;
 	struct i915_hw_context *default_context;
-	struct drm_i915_gem_object *last_context_obj;
+	struct i915_hw_context *last_context;
 
 	void *private;
 };
-- 
1.7.9.5

             reply	other threads:[~2013-05-02 13:48 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-05-02 13:48 Mika Kuoppala [this message]
2013-05-02 13:48 ` [PATCH 2/2] drm/i915: add context into request struct Mika Kuoppala
2013-05-04  0:07   ` Ben Widawsky
2013-05-06  9:22     ` Daniel Vetter
2013-05-06  9:28     ` Mika Kuoppala
2013-05-07  5:12       ` Ben Widawsky

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=1367502488-4108-1-git-send-email-mika.kuoppala@intel.com \
    --to=mika.kuoppala@linux.intel.com \
    --cc=ben@bwidawsk.net \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=miku@iki.fi \
    /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