drm/i915/gt: Make sure that errors are propagated through request chains

PATCH In review Andi Shyti 3 participants 4 messages Feb 10, 2023 – Feb 24, 2023

Subsystems (3)

Stable branch DRM drivers Intel i915

lore

Share & tools

lore

Andi Shyti

author

2023-02-10 14:06 UTC #0 lore raw

PATCH 1/1 drm/i915/gt: Make sure that errors are propagated through request chains

Currently, for operations like memory clear or copy for big chunks of memory, we generate multiple requests executed in a chain.

But if one of the requests generated fails we would not know it to unless it happens to the last request, because errors are not properly propagated.

For this we need to keep propagating the chain of fence notification in order to always reach the final fence associated to the final request.

This way we would know that the memory operation has failed and whether the memory is still invalid.

On copy and clear migration signal fences upon completion.

Fixes cf586021642d80 ("drm/i915/gt: Pipelined page migration")

Reported-by Matthew Auld <[email protected]>

Signed-off-by Andi Shyti <[email protected]>

Cc [email protected]

---
 drivers/gpu/drm/i915/gt/intel_migrate.c | 31 +++++++++++++++++--------
 1 file changed, 21 insertions(+), 10 deletions(-)

drivers/gpu/drm/i915/gt/intel_migrate.c +21 ~~-10~~

index 3f638f1987968..8a293045a7b96 100644
@@ -748,7 +748,7 @@ intel_context_migrate_copy(struct intel_context *ce,
 		rq = i915_request_create(ce);
 		if (IS_ERR(rq)) {
 			err = PTR_ERR(rq);
-			goto out_ce;
+			break;
 		}
 
 		if (deps) {

@@ -878,10 +878,14 @@ intel_context_migrate_copy(struct intel_context *ce,
 
 		/* Arbitration is re-enabled between requests. */
 out_rq:
-		if (*out)
-			i915_request_put(*out);
-		*out = i915_request_get(rq);
+		i915_sw_fence_await(&rq->submit);
+		i915_request_get(rq);
 		i915_request_add(rq);
+		if (*out) {
+			i915_sw_fence_complete(&(*out)->submit);
+			i915_request_put(*out);
+		}
+		*out = rq;
 
 		if (err)
 			break;

@@ -905,7 +909,8 @@ intel_context_migrate_copy(struct intel_context *ce,
 		cond_resched();
 	} while (1);
 
-out_ce:
+	if (*out)
+		i915_sw_fence_complete(&(*out)->submit);
 	return err;
 }

@@ -1005,7 +1010,7 @@ intel_context_migrate_clear(struct intel_context *ce,
 		rq = i915_request_create(ce);
 		if (IS_ERR(rq)) {
 			err = PTR_ERR(rq);
-			goto out_ce;
+			break;
 		}
 
 		if (deps) {

@@ -1056,17 +1061,23 @@ intel_context_migrate_clear(struct intel_context *ce,
 
 		/* Arbitration is re-enabled between requests. */
 out_rq:
-		if (*out)
-			i915_request_put(*out);
-		*out = i915_request_get(rq);
+		i915_sw_fence_await(&rq->submit);
+		i915_request_get(rq);
 		i915_request_add(rq);
+		if (*out) {
+			i915_sw_fence_complete(&(*out)->submit);
+			i915_request_put(*out);
+		}
+		*out = rq;
+
 		if (err || !it.sg || !sg_dma_len(it.sg))
 			break;
 
 		cond_resched();
 	} while (1);
 
-out_ce:
+	if (*out)
+		i915_sw_fence_complete(&(*out)->submit);
 	return err;
 }

Signature · 1 line

2.39.1

Patchwork

2023-02-10 17:53 UTC #1 lore raw

[Intel-gfx] ✓ Fi.CI.BAT: success for drm/i915/gt: Make sure that errors are propagated through request chains

== Series Details ==

Series: drm/i915/gt: Make sure that errors are propagated through request chains
URL : https://patchwork.freedesktop.org/series/113892/
State : success

== Summary ==

CI Bug Log - changes from CI_DRM_12727 -> Patchwork_113892v1
====================================================

Summary
-------

**SUCCESS**

No regressions found.

External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/index.html

Participating hosts (40 -> 39)
------------------------------

Missing (1): fi-snb-2520m

Possible new issues
-------------------

Here are the unknown changes that may have been introduced in Patchwork_113892v1:

### IGT changes ###

#### Suppressed ####

The following results come from untrusted machines, tests, or statuses. They do not affect the overall result.

* igt@i915_selftest@live@migrate:
- {bat-adlp-9}: [PASS][1] -> [ABORT][2]
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/bat-adlp-9/igt@i915_selftest@[email protected]
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/bat-adlp-9/igt@i915_selftest@[email protected]
- {bat-rpls-2}: [PASS][3] -> [ABORT][4]
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/bat-rpls-2/igt@i915_selftest@[email protected]
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/bat-rpls-2/igt@i915_selftest@[email protected]
- {bat-rpls-1}: [PASS][5] -> [ABORT][6]
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/bat-rpls-1/igt@i915_selftest@[email protected]
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/bat-rpls-1/igt@i915_selftest@[email protected]

Known issues
------------

Here are the changes found in Patchwork_113892v1 that come from known issues:

### IGT changes ###

#### Issues hit ####

* igt@kms_chamelium_hpd@common-hpd-after-suspend:
- fi-bsw-nick: NOTRUN -> [SKIP][7] ([fdo#109271]) +1 similar issue
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/fi-bsw-nick/igt@[email protected]

#### Possible fixes ####

* igt@gem_exec_gttfill@basic:
- fi-pnv-d510: [FAIL][8] ([i915#7229]) -> [PASS][9]
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/fi-pnv-d510/igt@[email protected]
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/fi-pnv-d510/igt@[email protected]

* igt@i915_pm_rpm@module-reload:
- {bat-adln-1}: [FAIL][10] -> [PASS][11]
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/bat-adln-1/igt@[email protected]
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/bat-adln-1/igt@[email protected]

* igt@i915_selftest@live@execlists:
- fi-bsw-nick: [ABORT][12] ([i915#7911] / [i915#7913]) -> [PASS][13]
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/fi-bsw-nick/igt@i915_selftest@[email protected]
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/fi-bsw-nick/igt@i915_selftest@[email protected]

* igt@i915_selftest@live@gt_heartbeat:
- {bat-adln-1}: [DMESG-WARN][14] -> [PASS][15] +10 similar issues
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/bat-adln-1/igt@i915_selftest@live@gt_heartbeat.html
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/bat-adln-1/igt@i915_selftest@live@gt_heartbeat.html

* igt@i915_selftest@live@perf:
- {bat-adln-1}: [DMESG-WARN][16] ([i915#2867]) -> [PASS][17] +7 similar issues
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/bat-adln-1/igt@i915_selftest@[email protected]
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/bat-adln-1/igt@i915_selftest@[email protected]

{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).

Build changes
-------------

* Linux: CI_DRM_12727 -> Patchwork_113892v1

CI-20190529: 20190529
CI_DRM_12727: 571111c8567d76c5a5ec4641ba771bb1bdd278fc @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_7156: d10786443fef434534ec4d08e93db92dfad84c7b @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
Patchwork_113892v1: 571111c8567d76c5a5ec4641ba771bb1bdd278fc @ git://anongit.freedesktop.org/gfx-ci/linux

### Linux commits

3718f65264e4 drm/i915/gt: Make sure that errors are propagated through request chains

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/index.html

Patchwork

2023-02-11 08:34 UTC #2 lore raw

[Intel-gfx] ✓ Fi.CI.IGT: success for drm/i915/gt: Make sure that errors are propagated through request chains

In reply to Andi Shyti

== Series Details ==

Series: drm/i915/gt: Make sure that errors are propagated through request chains
URL : https://patchwork.freedesktop.org/series/113892/
State : success

== Summary ==

CI Bug Log - changes from CI_DRM_12727_full -> Patchwork_113892v1_full
====================================================

Summary
-------

**SUCCESS**

No regressions found.

Participating hosts (9 -> 10)
------------------------------

Additional (1): shard-rkl0

Possible new issues
-------------------

Here are the unknown changes that may have been introduced in Patchwork_113892v1_full:

### IGT changes ###

#### Suppressed ####

The following results come from untrusted machines, tests, or statuses. They do not affect the overall result.

* igt@kms_frontbuffer_tracking@fbc-rgb565-draw-mmap-cpu:
- {shard-dg1}: NOTRUN -> [DMESG-WARN][1]
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-dg1-17/igt@[email protected]

Known issues
------------

Here are the changes found in Patchwork_113892v1_full that come from known issues:

### IGT changes ###

#### Issues hit ####

* igt@kms_ccs@pipe-a-missing-ccs-buffer-y_tiled_gen12_rc_ccs_cc:
- shard-snb: NOTRUN -> [SKIP][2] ([fdo#109271]) +10 similar issues
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-snb5/igt@kms_ccs@pipe-a-missing-ccs-buffer-y_tiled_gen12_rc_ccs_cc.html

* igt@perf@stress-open-close:
- shard-glk: [PASS][3] -> [ABORT][4] ([i915#5213])
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-glk7/igt@[email protected]
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-glk1/igt@[email protected]

#### Possible fixes ####

* igt@drm_fdinfo@virtual-idle:
- {shard-rkl}: [FAIL][5] ([i915#7742]) -> [PASS][6] +1 similar issue
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-2/igt@[email protected]
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-5/igt@[email protected]

* igt@fbdev@pan:
- {shard-rkl}: [SKIP][7] ([i915#2582]) -> [PASS][8] +1 similar issue
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@[email protected]
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@[email protected]

* igt@gem_exec_fair@basic-pace-share@rcs0:
- shard-glk: [FAIL][9] ([i915#2842]) -> [PASS][10]
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-glk1/igt@gem_exec_fair@[email protected]
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-glk8/igt@gem_exec_fair@[email protected]

* igt@gem_exec_reloc@basic-cpu-read:
- {shard-rkl}: [SKIP][11] ([i915#3281]) -> [PASS][12]
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-2/igt@[email protected]
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-5/igt@[email protected]

* igt@gem_partial_pwrite_pread@write-uncached:
- {shard-rkl}: [SKIP][13] ([i915#3282]) -> [PASS][14] +1 similar issue
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-2/igt@[email protected]
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-5/igt@[email protected]

* igt@gen9_exec_parse@allowed-all:
- {shard-rkl}: [SKIP][15] ([i915#2527]) -> [PASS][16]
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-2/igt@[email protected]
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-5/igt@[email protected]

* igt@i915_module_load@reload:
- shard-snb: [ABORT][17] -> [PASS][18]
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-snb5/igt@[email protected]
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-snb5/igt@[email protected]

* igt@i915_pipe_stress@stress-xrgb8888-ytiled:
- {shard-rkl}: [SKIP][19] ([i915#4098]) -> [PASS][20] +1 similar issue
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@[email protected]
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@[email protected]

* igt@i915_pm_rpm@modeset-lpsp-stress-no-wait:
- {shard-rkl}: [SKIP][21] ([i915#1397]) -> [PASS][22]
[21]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@[email protected]
[22]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@[email protected]

* igt@kms_frontbuffer_tracking@psr-1p-offscren-pri-shrfb-draw-blt:
- {shard-rkl}: [SKIP][23] ([i915#1849] / [i915#4098]) -> [PASS][24] +12 similar issues
[23]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@[email protected]
[24]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@[email protected]

* igt@kms_plane@plane-panning-bottom-right@pipe-a-planes:
- {shard-rkl}: [SKIP][25] ([i915#1849]) -> [PASS][26] +2 similar issues
[25]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@kms_plane@[email protected]
[26]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@kms_plane@[email protected]

* igt@kms_psr@cursor_plane_onoff:
- {shard-rkl}: [SKIP][27] ([i915#1072]) -> [PASS][28] +1 similar issue
[27]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@kms_psr@cursor_plane_onoff.html
[28]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@kms_psr@cursor_plane_onoff.html

* igt@kms_universal_plane@cursor-fb-leak-pipe-b:
- {shard-rkl}: [SKIP][29] ([i915#1845] / [i915#4070] / [i915#4098]) -> [PASS][30] +1 similar issue
[29]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@[email protected]
[30]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@[email protected]

* igt@kms_vblank@pipe-b-wait-busy-hang:
- {shard-rkl}: [SKIP][31] ([i915#1845] / [i915#4098]) -> [PASS][32] +13 similar issues
[31]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-1/igt@[email protected]
[32]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-6/igt@[email protected]

* igt@perf@polling-small-buf:
- {shard-rkl}: [FAIL][33] ([i915#1722]) -> [PASS][34]
[33]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12727/shard-rkl-2/igt@[email protected]
[34]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/shard-rkl-5/igt@[email protected]

{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).

Build changes
-------------

* Linux: CI_DRM_12727 -> Patchwork_113892v1

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_113892v1/index.html

Matthew Auld

Reviewed-by on v1 1/1

2023-02-24 11:25 UTC #3 lore raw

In reply to Andi Shyti

Quoted 16 lines · Andi Shyti

Currently, for operations like memory clear or copy for big chunks of memory, we generate multiple requests executed in a chain.
But if one of the requests generated fails we would not know it to unless it happens to the last request, because errors are not properly propagated.
For this we need to keep propagating the chain of fence notification in order to always reach the final fence associated to the final request.
This way we would know that the memory operation has failed and whether the memory is still invalid.

Quoted 7 lines

On copy and clear migration signal fences upon completion.
Fixes cf586021642d80 ("drm/i915/gt: Pipelined page migration")
Reported-by Matthew Auld <[email protected]>
Suggested-by Chris Wilson <[email protected]>
Signed-off-by Andi Shyti <[email protected]>
Cc [email protected]

Reviewed-by Matthew Auld <[email protected]>