intel-gfx.lists.freedesktop.org archive mirror
 help / color / mirror / Atom feed
From: Sagar Arun Kamble <sagar.a.kamble@intel.com>
To: intel-gfx@lists.freedesktop.org
Subject: [PATCH 1/5] drm/i915/guc: Separate GuC doorbell destroy function into doorbell unit and GuC task
Date: Sun,  5 Nov 2017 19:09:39 +0530	[thread overview]
Message-ID: <1509889183-2094-1-git-send-email-sagar.a.kamble@intel.com> (raw)

Releasing GuC doorbell involves stopping the doorbell unit snooping and
executing doorbell deallocation flow in GuC firmware. When GuC is in sane
state, both steps will be necessary to release doorbell.
If GuC is hung (possibly leading to i915 reset), only doorbell unit
snooping is to be stopped. destroy_doorbell will be called in suspend and
reset flows differently in upcoming patches.

Signed-off-by: Sagar Arun Kamble <sagar.a.kamble@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Arkadiusz Hiler <arkadiusz.hiler@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_guc_submission.c | 23 +++++++++++++++--------
 1 file changed, 15 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index d14c134..b6486dc 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -207,7 +207,7 @@ static int __create_doorbell(struct i915_guc_client *client)
 	return err;
 }
 
-static int __destroy_doorbell(struct i915_guc_client *client)
+static int __destroy_doorbell(struct i915_guc_client *client, bool notify_guc)
 {
 	struct drm_i915_private *dev_priv = guc_to_i915(client->guc);
 	struct guc_doorbell_info *doorbell;
@@ -225,7 +225,10 @@ static int __destroy_doorbell(struct i915_guc_client *client)
 	if (wait_for_us(!(I915_READ(GEN8_DRBREGL(db_id)) & GEN8_DRB_VALID), 10))
 		WARN_ONCE(true, "Doorbell never became invalid after disable\n");
 
-	return __guc_deallocate_doorbell(client->guc, client->stage_id);
+	if (notify_guc)
+		return __guc_deallocate_doorbell(client->guc, client->stage_id);
+
+	return 0;
 }
 
 static int create_doorbell(struct i915_guc_client *client)
@@ -250,7 +253,11 @@ static int create_doorbell(struct i915_guc_client *client)
 	return ret;
 }
 
-static int destroy_doorbell(struct i915_guc_client *client)
+/*
+ * This function deactivates doorbell monitoring through doorbell unit and
+ * optionally notifies GuC to deallocate the doorbell.
+ */
+static int destroy_doorbell(struct i915_guc_client *client, bool notify_guc)
 {
 	int err;
 
@@ -259,7 +266,7 @@ static int destroy_doorbell(struct i915_guc_client *client)
 	/* XXX: wait for any interrupts */
 	/* XXX: wait for workqueue to drain */
 
-	err = __destroy_doorbell(client);
+	err = __destroy_doorbell(client, notify_guc);
 	if (err)
 		return err;
 
@@ -873,7 +880,7 @@ static int __reset_doorbell(struct i915_guc_client* client, u16 db_id)
 	__update_doorbell_desc(client, db_id);
 	err = __create_doorbell(client);
 	if (!err)
-		err = __destroy_doorbell(client);
+		err = __destroy_doorbell(client, true);
 
 	return err;
 }
@@ -899,7 +906,7 @@ static int guc_init_doorbell_hw(struct intel_guc *guc)
 
 		if (has_doorbell(client)) {
 			/* Borrow execbuf_client (we will recreate it later) */
-			destroy_doorbell(client);
+			destroy_doorbell(client, true);
 			recreate_first_client = true;
 		}
 
@@ -925,7 +932,7 @@ static int guc_init_doorbell_hw(struct intel_guc *guc)
 
 	ret = __create_doorbell(guc->preempt_client);
 	if (ret) {
-		__destroy_doorbell(guc->execbuf_client);
+		__destroy_doorbell(guc->execbuf_client, true);
 		return ret;
 	}
 
@@ -1043,7 +1050,7 @@ static void guc_client_free(struct i915_guc_client *client)
 	/* FIXME: in many cases, by the time we get here the GuC has been
 	 * reset, so we cannot destroy the doorbell properly. Ignore the
 	 * error message for now */
-	destroy_doorbell(client);
+	destroy_doorbell(client, true);
 	guc_stage_desc_fini(client->guc, client);
 	i915_gem_object_unpin_map(client->vma->obj);
 	i915_vma_unpin_and_release(&client->vma);
-- 
1.9.1

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

             reply	other threads:[~2017-11-05 13:36 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-11-05 13:39 Sagar Arun Kamble [this message]
2017-11-05 13:39 ` [PATCH 2/5] drm/i915/guc: Release all client doorbells on suspend and acquire on resume Sagar Arun Kamble
2017-11-06 12:13   ` Chris Wilson
2017-11-07  6:05     ` Sagar Arun Kamble
2017-11-07  9:21       ` Chris Wilson
2017-11-07 10:27         ` Michal Wajdeczko
2017-11-07 10:33           ` Sagar Arun Kamble
2017-11-05 13:39 ` [PATCH 3/5] drm/i915/guc: Deactivate all client doorbells before reset Sagar Arun Kamble
2017-11-05 13:39 ` [PATCH 4/5] drm/i915/guc: Increase wait timeout for doorbell release status update Sagar Arun Kamble
2017-11-05 15:55   ` Michal Wajdeczko
2017-11-06  5:43     ` Sagar Arun Kamble
2017-11-05 13:39 ` [PATCH 5/5] HAX enable GuC submission for CI Sagar Arun Kamble
2017-11-05 14:18 ` ✗ Fi.CI.BAT: warning for series starting with [1/5] drm/i915/guc: Separate GuC doorbell destroy function into doorbell unit and GuC task Patchwork
2017-11-08 12:32 ` ✓ Fi.CI.BAT: success " 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=1509889183-2094-1-git-send-email-sagar.a.kamble@intel.com \
    --to=sagar.a.kamble@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 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).