Skip to contentSkip to conversation

drm/i915: reset drm state backpointer in crtc_state

PATCH In review Chandra Konduru 4 participants 7 messages Apr 6, 2015 – Apr 8, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: reset drm state backpointer in crtc_state

At end of intel_crtc_set_config, reset crtc_state's drm_state back pointer to null.

---
 drivers/gpu/drm/i915/intel_display.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_display.c +3 -1
index c84926b..f9c2e4d 100644
@@ -12451,8 +12451,10 @@ fail:
 	}
 
 out_config:
-	if (state)
+	if (state) {
 		drm_atomic_state_free(state);
+		to_intel_crtc(set->crtc)->config->base.state = NULL;
+	}
 
 	intel_set_config_free(config);
 	return ret;
Signature · 6 lines
1.7.9.5

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Tested-by on v1 1/1
#1 lore raw

Task id: 6132 -------------------------------------Summary-------------------------------------
Platform Delta drm-intel-nightly Series Applied
PNV 272/272 272/272
ILK -1 302/302 301/302
SNB 303/303 303/303

IVB                 -1              338/338              337/338
BYT                 -1              287/287              286/287
HSW                                  361/361              361/361
BDW                                  308/308              308/308

-------------------------------------Detailed-------------------------------------
Platform Test drm-intel-nightly Series Applied
*ILK igt@kms_pipe_crc_basic@bad-nb-words-3 PASS(2) DMESG_WARN(1)PASS(1)
(dmesg patch applied)drm:drm_edid_block_valid[drm]]*ERROR*EDID_checksum_is_invalid,remainder_is@EDID checksum is .* remainder is
*IVB igt@gem_pwrite_pread@snooped-pwrite-blt-cpu_mmap-performance PASS(2) DMESG_WARN(1)PASS(1)
(dmesg patch applied)drm:i915_hangcheck_elapsed[i915]]*ERROR*Hangcheck_timer_elapsed...blitter_ring_idle@Hangcheck timer elapsed... blitter ring idle
*BYT igt@gem_exec_bad_domains@conflicting-write-domain PASS(19) FAIL(1)PASS(1)
Note: You need to pay more attention to line start with '*'
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

maintainer
#2 lore raw
In reply to Chandra Konduru
Quoted 2 lines · Chandra Konduru

At end of intel_crtc_set_config, reset crtc_state's drm_state back pointer to null.

This does not tell me anything that reading the patch already didn't. Please explain *why* this is needed in the commit message. What breaks without it? If this fixes a regression, please indicate which commit regressed.

BR,
Jani.

Quoted 29 lines
---
 drivers/gpu/drm/i915/intel_display.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_display.c +3 -1
index c84926b..f9c2e4d 100644
@@ -12451,8 +12451,10 @@ fail:
 	}
 
 out_config:
-	if (state)
+	if (state) {
 		drm_atomic_state_free(state);
+		to_intel_crtc(set->crtc)->config->base.state = NULL;
+	}
 
 	intel_set_config_free(config);
 	return ret;
Signature · 6 lines
1.7.9.5

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw
Quoted 8 lines

-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Tuesday, April 07, 2015 2:02 AM
To: Konduru, Chandra; [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: reset drm state backpointer in crtc_state

Quoted 7 lines
Quoted 2 lines · Chandra Konduru

At end of intel_crtc_set_config, reset crtc_state's drm_state back pointer to null.

This does not tell me anything that reading the patch already didn't. Please explain *why* this is needed in the commit message. What breaks without it? If this fixes a regression, please indicate which commit regressed.

Once atomic transaction is done, live crtc_state (i.e., intel_crtc->config) is carrying back pointer to drm_atomic_state which is freed. When a new non-atomic transaction is made (crtc_disable triggered off from set_mode), this stale pointer is carried into that transaction.
Depending on timing, this causes issue to scaler feature that I am working if panel fit to be disabled during crtc_disable.

It has potential to cause issues to wm work that Matt is doing, but not sure.

Quoted 30 lines

BR,
Jani.

Quoted 26 lines
---
 drivers/gpu/drm/i915/intel_display.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_display.c +3 -2
index c84926b..f9c2e4d 100644
@@ -12451,8 +12451,10 @@ fail:
 	}
 
 out_config:
-	if (state)
+	if (state) {
 		drm_atomic_state_free(state);
+		to_intel_crtc(set->crtc)->config->base.state = NULL;
+	}
 
 	intel_set_config_free(config);
 	return ret;
--

1.7.9.5

Quoted 7 lines
Quoted 4 lines

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

--
Jani Nikula, Intel Open Source Technology Center

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

#4 lore raw
Quoted 16 lines · Chandra
Quoted 8 lines

-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Tuesday, April 07, 2015 2:02 AM
To: Konduru, Chandra; [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: reset drm state backpointer in crtc_state

Quoted 7 lines
Quoted 2 lines · Chandra Konduru

At end of intel_crtc_set_config, reset crtc_state's drm_state back pointer to null.

This does not tell me anything that reading the patch already didn't. Please explain *why* this is needed in the commit message. What breaks without it? If this fixes a regression, please indicate which commit regressed.

Quoted 6 lines

Once atomic transaction is done, live crtc_state (i.e., intel_crtc->config) is carrying back pointer to drm_atomic_state which is freed. When a new non-atomic transaction is made (crtc_disable triggered off from set_mode), this stale pointer is carried into that transaction.
Depending on timing, this causes issue to scaler feature that I am working if panel fit to be disabled during crtc_disable.

You shouldn't ever be chasing the state backpointer of a committed state. It's ok to clean it up, but chasing that pointer itself is a bug. Why does the scaler code look at that pointer?
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#5 lore raw
Quoted 36 lines

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Wednesday, April 08, 2015 1:20 AM
To: Konduru, Chandra
Cc: Jani Nikula; [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: reset drm state backpointer in crtc_state

Quoted 9 lines · Chandra
Quoted 9 lines

-----Original Message-----
From: Jani Nikula [mailto:[email protected]]
Sent: Tuesday, April 07, 2015 2:02 AM
To: Konduru, Chandra; [email protected]
Subject: Re: [Intel-gfx] [PATCH] drm/i915: reset drm state backpointer in crtc_state

On Mon, 06 Apr 2015, Chandra Konduru <[email protected]>

Quoted 6 lines · wrote:
Quoted 6 lines
Quoted 2 lines

At end of intel_crtc_set_config, reset crtc_state's drm_state back pointer to null.

This does not tell me anything that reading the patch already didn't. Please explain *why* this is needed in the commit message. What breaks without it? If this fixes a regression, please indicate which

commit regressed.

Quoted 7 lines

Once atomic transaction is done, live crtc_state (i.e., intel_crtc->config) is carrying back pointer to drm_atomic_state which is freed. When a new non-atomic transaction is made (crtc_disable triggered off from set_mode), this stale pointer is carried into that transaction. Depending on timing, this causes issue to scaler feature that I am working if panel fit to be disabled during crtc_disable.

Quoted 3 lines

You shouldn't ever be chasing the state backpointer of a committed state. It's ok to clean it up, but chasing that pointer itself is a bug. Why does the scaler code look at that pointer?

While setting up scalers, it looks at that pointer to know other scaler users.

Quoted 5 lines

-Daniel
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

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

#6 lore raw
In reply to Chandra Konduru
Quoted 13 lines · Chandra Konduru

At end of intel_crtc_set_config, reset crtc_state's drm_state back pointer to null.

---
 drivers/gpu/drm/i915/intel_display.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_display.c +0 -0
index c84926b..f9c2e4d 100644

drivers/gpu/drm/i915/intel_display.c fail: · line 12456

Quoted 7 lines
drivers/gpu/drm/i915/intel_display.c +2 -1
@@ -12451,8 +12451,10 @@ fail:
 	}
 
 out_config:
-	if (state)
+	if (state) {
 		drm_atomic_state_free(state);
+		to_intel_crtc(set->crtc)->config->base.state = NULL;

This should be done in intel_crtc_set_state instead of just in this case here.
-Daniel

Quoted 11 lines
drivers/gpu/drm/i915/intel_display.c +1 -0
+	}
 
 	intel_set_config_free(config);
 	return ret;
Signature · 6 lines
1.7.9.5

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx