Skip to contentSkip to conversation

drm/i915: fix the dequeue logic for single_port_submission context

PATCH In review Min He 4 participants 5 messages Nov 16, 2016 – Nov 17, 2016

Subsystems (2)
lore
Share & tools
lore

Min He

author
#0 lore raw

PATCH v3 1/1 drm/i915: fix the dequeue logic for single_port_submission context

For a singl_port_submission context, it can only be submitted to port 0, and there shouldn't be any other context in port 1 at the same time. This is required by GVT-g context to have an opportunity to save/restore some non-hw context render registers.
This patch is to implement the correct logic in execlists_dequeue.

v2: optimized code by following Chris's advice, and added more comments to explain the patch.
v3: followed the coding style.

---
 drivers/gpu/drm/i915/intel_lrc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_lrc.c +2 -1
index f50feaa..b2c0d50 100644
@@ -499,7 +499,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 			 * context (even though a different request) to
 			 * the second port.
 			 */
-			if (ctx_single_port_submission(cursor->ctx))
+			if (ctx_single_port_submission(last->ctx) ||
+			    ctx_single_port_submission(cursor->ctx))
 				break;
 
 			GEM_BUG_ON(last->ctx == cursor->ctx);
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

✓ Fi.CI.BAT: success for drm/i915: fix the dequeue logic for single_port_submission context (rev3)

== Series Details ==

== Series Details ==

Series: drm/i915: fix the dequeue logic for single_port_submission context (rev3)
URL : https://patchwork.freedesktop.org/series/15391/
State : success

== Summary ==

Series 15391v3 drm/i915: fix the dequeue logic for single_port_submission context
https://patchwork.freedesktop.org/api/1.0/series/15391/revisions/3/mbox/

fi-bdw-5557u     total:244  pass:229  dwarn:0   dfail:0   fail:0   skip:15 
fi-bsw-n3050     total:244  pass:204  dwarn:0   dfail:0   fail:0   skip:40 
fi-byt-j1900     total:244  pass:216  dwarn:0   dfail:0   fail:0   skip:28 
fi-byt-n2820     total:244  pass:212  dwarn:0   dfail:0   fail:0   skip:32 
fi-hsw-4770      total:244  pass:224  dwarn:0   dfail:0   fail:0   skip:20 
fi-hsw-4770r     total:244  pass:224  dwarn:0   dfail:0   fail:0   skip:20 
fi-ilk-650       total:244  pass:191  dwarn:0   dfail:0   fail:0   skip:53 
fi-ivb-3520m     total:244  pass:222  dwarn:0   dfail:0   fail:0   skip:22 
fi-ivb-3770      total:244  pass:222  dwarn:0   dfail:0   fail:0   skip:22 
fi-kbl-7200u     total:244  pass:222  dwarn:0   dfail:0   fail:0   skip:22 
fi-skl-6260u     total:244  pass:230  dwarn:0   dfail:0   fail:0   skip:14 
fi-skl-6700hq    total:244  pass:223  dwarn:0   dfail:0   fail:0   skip:21 
fi-skl-6700k     total:244  pass:222  dwarn:1   dfail:0   fail:0   skip:21 
fi-skl-6770hq    total:244  pass:230  dwarn:0   dfail:0   fail:0   skip:14 
fi-snb-2520m     total:244  pass:212  dwarn:0   dfail:0   fail:0   skip:32 
fi-snb-2600      total:244  pass:211  dwarn:0   dfail:0   fail:0   skip:33 

6294f67611ebe69006c0e85c372efadcac8e9d66 drm-intel-nightly: 2016y-11m-16d-09h-57m-25s UTC integration manifest d8671cd drm/i915: fix the dequeue logic for single_port_submission context

== Logs ==

For more details see: https://intel-gfx-ci.01.org/CI/Patchwork_3022/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Reviewed-by on v3 1/1
#2 lore raw
In reply to Min He
Quoted 6 lines · Min He

For a singl_port_submission context, it can only be submitted to port 0, and there shouldn't be any other context in port 1 at the same time. This is required by GVT-g context to have an opportunity to save/restore some non-hw context render registers.
This patch is to implement the correct logic in execlists_dequeue.

Quoted 7 lines

v2: optimized code by following Chris's advice, and added more comments to explain the patch.
v3: followed the coding style.

---

Hi, Min, like Daniel said not need to add my s-o-b.

Quoted 20 lines
 drivers/gpu/drm/i915/intel_lrc.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_lrc.c +2 -1
index f50feaa..b2c0d50 100644
@@ -499,7 +499,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 			 * context (even though a different request) to
 			 * the second port.
 			 */
-			if (ctx_single_port_submission(cursor->ctx))
+			if (ctx_single_port_submission(last->ctx) ||
+			    ctx_single_port_submission(cursor->ctx))
 				break;
 
 			GEM_BUG_ON(last->ctx == cursor->ctx);
Signature · 1 line
1.9.1
Signature · 3 lines
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#3 lore raw
In reply to Min He
Quoted 4 lines · Min He

For a singl_port_submission context, it can only be submitted to port 0, and there shouldn't be any other context in port 1 at the same time. This is required by GVT-g context to have an opportunity to save/restore some non-hw context render registers.

This statement is not true. It has the opportunity to modify the GVT context if a non-GVT context was in port 0 or port 1.

Quoted 1 line

This patch is to implement the correct logic in execlists_dequeue.

I object. This is not the correct logic here, but to workaround a failure in GVT.

Any way pushed until GVT is fixed.
-Chris

Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
author
#4 lore raw
Quoted 10 lines

-----Original Message-----
From: Chris Wilson [mailto:[email protected]]
Sent: Thursday, November 17, 2016 4:01 PM
To: He, Min <[email protected]>
Cc: [email protected]
Subject: Re: [Intel-gfx] [PATCH v3] drm/i915: fix the dequeue logic for single_port_submission context

Quoted 8 lines
Quoted 4 lines · Min He

For a singl_port_submission context, it can only be submitted to port 0, and there shouldn't be any other context in port 1 at the same time. This is required by GVT-g context to have an opportunity to save/restore some non-hw context render registers.

This statement is not true. It has the opportunity to modify the GVT context if a non-GVT context was in port 0 or port 1.

If a non-GVT context is in port 0, and a GVT context in port 1, after non-GVT context completes, GPU will switch to GVT context directly, and we will lose the opportunity to save the non-GVT context registers and restore GVT context registers. So non-GVT context's register will impact GVT context. And vice versa.

Quoted 11 lines
Quoted 2 lines

This patch is to implement the correct logic in execlists_dequeue.

I object. This is not the correct logic here, but to workaround a failure in GVT.

Any way pushed until GVT is fixed.
-Chris

--
Chris Wilson, Intel Open Source Technology Centre

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx