Skip to contentSkip to conversation

drm/i915: Name the anonymous per-engine context struct

PATCH Superseded Tvrtko Ursulin 4 participants 14 messages Mar 18, 2016 – Mar 22, 2016

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

PATCH 1/1 drm/i915: Name the anonymous per-engine context struct

From: Tvrtko Ursulin <[email protected]>

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

Name for the structure, intel_context_engine, and the local variable name were borrowed from a similar patch by Chris Wilson.

---
 drivers/gpu/drm/i915/i915_drv.h  |  2 +-
 drivers/gpu/drm/i915/intel_lrc.c | 96 +++++++++++++++++++++-------------------
 2 files changed, 51 insertions(+), 47 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +1 -1
index 00c41a4bde2a..480639c39543 100644
@@ -840,7 +840,7 @@ struct intel_context {
 	} legacy_hw_ctx;
 
 	/* Execlists */
-	struct {
+	struct intel_context_engine {
 		struct drm_i915_gem_object *state;
 		struct intel_ringbuffer *ringbuf;
 		int pin_count;
drivers/gpu/drm/i915/intel_lrc.c +50 -46

241 lines not on this page · raw message

Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by on v1 1/1
#1 lore raw
Quoted 12 lines · Tvrtko Ursulin

From: Tvrtko Ursulin <[email protected]>

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

Quoted 5 lines

Name for the structure, intel_context_engine, and the local variable name were borrowed from a similar patch by Chris Wilson.

Looks like a patch I've been maintaining for an eon, and I think Mika once did a similar thing.

If I'm far enough away from it to be impartial?
-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
#2 lore raw

PATCH v2 1/1 drm/i915: Name the anonymous per-engine context struct

In reply to Tvrtko Ursulin

From: Tvrtko Ursulin <[email protected]>

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

Name for the structure, intel_context_engine, and the local variable name were borrowed from a similar patch by Chris Wilson.

v2: Hate the engine->dev surprises, really do.

---
 drivers/gpu/drm/i915/i915_drv.h  |  2 +-
 drivers/gpu/drm/i915/intel_lrc.c | 94 +++++++++++++++++++++-------------------
 2 files changed, 50 insertions(+), 46 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +1 -1
index 00c41a4bde2a..480639c39543 100644
@@ -840,7 +840,7 @@ struct intel_context {
 	} legacy_hw_ctx;
 
 	/* Execlists */
-	struct {
+	struct intel_context_engine {
 		struct drm_i915_gem_object *state;
 		struct intel_ringbuffer *ringbuf;
 		int pin_count;
drivers/gpu/drm/i915/intel_lrc.c +49 -45

240 lines not on this page · raw message

Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw
Quoted 29 lines · Tvrtko Ursulin

From: Tvrtko Ursulin <[email protected]>

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

Name for the structure, intel_context_engine, and the local variable name were borrowed from a similar patch by Chris Wilson.

v2: Hate the engine->dev surprises, really do.

---
  drivers/gpu/drm/i915/i915_drv.h  |  2 +-
  drivers/gpu/drm/i915/intel_lrc.c | 94 +++++++++++++++++++++-------------------
  2 files changed, 50 insertions(+), 46 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +0 -0
index 00c41a4bde2a..480639c39543 100644

drivers/gpu/drm/i915/i915_drv.h struct intel_context { · line 843

Quoted 5 lines
drivers/gpu/drm/i915/i915_drv.h +1 -1
@@ -840,7 +840,7 @@ struct intel_context {
  	} legacy_hw_ctx;
 
  	/* Execlists */
-	struct {
+	struct intel_context_engine {

Good idea, I had a version of this too, derived from Chris' patch [157/190] drm/i915: Tidy execlists by using intel_context_engine locals.

The only thing to disagree with is the actual name; it should be "intel_engine_context" (or some abbreviation thereof), because in English (and German) the noun at the *end* of a compound noun-phrase is what it actually *is*, with all the others qualifying it. So a "railway bridge" is a type of bridge, not a type of railway, and
"eine Straßenbahnhaltestelle" (street-train-stopping-place => tram stop) is not a street.

[aside] My favourite in English is "Space Civilisation Power Struggle Game" (five nouns in a row!) describing a certain boxed game -- anyone recognise that? [/aside]

I'll post a version following that naming convention shortly ...

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

#4 lore raw
Quoted 39 lines · Dave Gordon
Quoted 29 lines · Tvrtko Ursulin

From: Tvrtko Ursulin <[email protected]>

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

Name for the structure, intel_context_engine, and the local variable name were borrowed from a similar patch by Chris Wilson.

v2: Hate the engine->dev surprises, really do.

---
 drivers/gpu/drm/i915/i915_drv.h  |  2 +-
 drivers/gpu/drm/i915/intel_lrc.c | 94 +++++++++++++++++++++-------------------
 2 files changed, 50 insertions(+), 46 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +0 -0
index 00c41a4bde2a..480639c39543 100644

drivers/gpu/drm/i915/i915_drv.h struct intel_context { · line 843

Quoted 5 lines
drivers/gpu/drm/i915/i915_drv.h +1 -1
@@ -840,7 +840,7 @@ struct intel_context {
 	} legacy_hw_ctx;
 
 	/* Execlists */
-	struct {
+	struct intel_context_engine {

Good idea, I had a version of this too, derived from Chris' patch [157/190] drm/i915: Tidy execlists by using intel_context_engine locals.

Quoted 2 lines

The only thing to disagree with is the actual name; it should be "intel_engine_context" (or some abbreviation thereof), because in

We have been using object_subobject.
-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
#5 lore raw
Quoted 40 lines · Chris Wilson
Quoted 39 lines · Dave Gordon
Quoted 29 lines · Tvrtko Ursulin

From: Tvrtko Ursulin <[email protected]>

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

Name for the structure, intel_context_engine, and the local variable name were borrowed from a similar patch by Chris Wilson.

v2: Hate the engine->dev surprises, really do.

---
  drivers/gpu/drm/i915/i915_drv.h  |  2 +-
  drivers/gpu/drm/i915/intel_lrc.c | 94 +++++++++++++++++++++-------------------
  2 files changed, 50 insertions(+), 46 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +0 -0
index 00c41a4bde2a..480639c39543 100644

drivers/gpu/drm/i915/i915_drv.h struct intel_context { · line 843

Quoted 5 lines
drivers/gpu/drm/i915/i915_drv.h +1 -1
@@ -840,7 +840,7 @@ struct intel_context {
  	} legacy_hw_ctx;
 
  	/* Execlists */
-	struct {
+	struct intel_context_engine {

Good idea, I had a version of this too, derived from Chris' patch [157/190] drm/i915: Tidy execlists by using intel_context_engine locals.

Quoted 5 lines
Quoted 2 lines

The only thing to disagree with is the actual name; it should be "intel_engine_context" (or some abbreviation thereof), because in

We have been using object_subobject.
-Chris

No we haven't. Examples of the above convention for structs include:

* "struct intel_device_info" - information about an intel device
* "struct i915_ctx_hang_stats" - statistics about hangs, per context
* "struct intel_uncore_funcs" - functions exported from uncore
* "struct i915_power_well_ops" - operations on power wells
* "struct i915_suspend_saved_registers" - registers saved at suspend

and for instance names:

* "struct list_head request_list" - a list of requests
* "struct i915_vma *lrc_vma" - the VMA for an LRC
* "uint32_t *lrc_reg_state" - the state of the registers in an LRC

Indeed it's quite difficult to find any compound names that do *not* follow this convention. Maybe your version would be more natural in a language where adjectives (or possessives) normally follow the noun they describe?

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

#6 lore raw
Quoted 59 lines · Dave Gordon
Quoted 42 lines · Chris Wilson
Quoted 41 lines · Dave Gordon
Quoted 31 lines · Tvrtko Ursulin

From: Tvrtko Ursulin <[email protected]>

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

Name for the structure, intel_context_engine, and the local variable name were borrowed from a similar patch by Chris Wilson.

v2: Hate the engine->dev surprises, really do.

---
  drivers/gpu/drm/i915/i915_drv.h  |  2 +-
  drivers/gpu/drm/i915/intel_lrc.c | 94
+++++++++++++++++++++-------------------
  2 files changed, 50 insertions(+), 46 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +0 -0
index 00c41a4bde2a..480639c39543 100644

drivers/gpu/drm/i915/i915_drv.h struct intel_context { · line 843

Quoted 5 lines
drivers/gpu/drm/i915/i915_drv.h +1 -1
@@ -840,7 +840,7 @@ struct intel_context {
      } legacy_hw_ctx;
 
      /* Execlists */
-    struct {
+    struct intel_context_engine {

Good idea, I had a version of this too, derived from Chris' patch [157/190] drm/i915: Tidy execlists by using intel_context_engine locals.

Quoted 5 lines
Quoted 2 lines

The only thing to disagree with is the actual name; it should be "intel_engine_context" (or some abbreviation thereof), because in

We have been using object_subobject.
-Chris

No we haven't. Examples of the above convention for structs include:

* "struct intel_device_info" - information about an intel device
* "struct i915_ctx_hang_stats" - statistics about hangs, per context
* "struct intel_uncore_funcs" - functions exported from uncore
* "struct i915_power_well_ops" - operations on power wells
* "struct i915_suspend_saved_registers" - registers saved at suspend

and for instance names:

Quoted 8 lines

* "struct list_head request_list" - a list of requests
* "struct i915_vma *lrc_vma" - the VMA for an LRC
* "uint32_t *lrc_reg_state" - the state of the registers in an LRC

Indeed it's quite difficult to find any compound names that do *not* follow this convention. Maybe your version would be more natural in a language where adjectives (or possessives) normally follow the noun they describe?

One example is maybe:

struct drm_i915_error_state {
struct drm_i915_error_ring

So per-ring state of the total error state, which sounds equivalent to per-engine state of a context.

Or:

struct intel_fbc {
	struct intel_fbc_state_cache {

More verbose version like "struct intel_context_engine_state"?

"struct intel_engine_context" reads like a context of an engine to me, which sounds opposite to what it should be - state of an engine for a context.

?

Regards,

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

#7 lore raw

PATCH v3 1/2 drm/i915: name the anonymous per-engine context struct

In reply to Dave Gordon

Introduce a name for the previously anonymous per-engine structure embedded inside the intel_context, and use local pointers to the relevant element of this array to simplify various execlist functions. This may improve the compiler's ability to avoid redundant dereference- and index operations.

This version is derived from an original by Chris Wilson, detached from the original set of patches in which it was posted and rebased to current nightly.

Then it was updated by Tvrtko Ursulin <[email protected]>, who noted:

This anonymous struct was causing a good amount of overly verbose code. Also, if we name it and cache the pointer locally when there are multiple accesses to it, not only the code is more readable, but the compiler manages to generate smaller binary.

Along the way I also shortened access to dev_priv and eliminated some unused variables and cached some where I spotted the opportunity.

This version uses the name "struct intel_engine_ctx" in accordance with the English (and German) convention that the concrete noun goes at the end of a noun-sequence, with earlier ones acting as adjectives i.e. an "engine-context" is the context of an engine, whereas a "context-engine" would be some sort of engine that processes contexts. Since object and structure names are noun-like, it's more consistent to name them this way.

---
 drivers/gpu/drm/i915/i915_drv.h  |   2 +-
 drivers/gpu/drm/i915/intel_lrc.c | 122 +++++++++++++++++++--------------------
 2 files changed, 62 insertions(+), 62 deletions(-)
drivers/gpu/drm/i915/i915_drv.h +1 -1
index 00c41a4..be3709f 100644
@@ -840,7 +840,7 @@ struct intel_context {
 	} legacy_hw_ctx;
 
 	/* Execlists */
-	struct {
+	struct intel_engine_ctx {
 		struct drm_i915_gem_object *state;
 		struct intel_ringbuffer *ringbuf;
 		int pin_count;
drivers/gpu/drm/i915/intel_lrc.c +61 -61

313 lines not on this page · raw message

Signature · 6 lines
1.9.1

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

PATCH v3 2/2 drm/i915: tidy up a few more references to engine[]

Now that we have a name for the previously anonymous per-engine structure embedded inside the intel_context for the benefit of execlist mode, we can optimise a few more places that access this array of structures. This may improve the compiler's ability to avoid redundant dereference and index operations.

---
 drivers/gpu/drm/i915/i915_debugfs.c        | 13 +++++--------
 drivers/gpu/drm/i915/i915_guc_submission.c | 17 +++++++++--------
 2 files changed, 14 insertions(+), 16 deletions(-)
drivers/gpu/drm/i915/i915_debugfs.c +5 -8
index ccdca2c..ce3b5e9 100644
@@ -1966,16 +1966,13 @@ static int i915_context_status(struct seq_file *m, void *unused)
 		if (i915.enable_execlists) {
 			seq_putc(m, '\n');
 			for_each_engine(engine, dev_priv, i) {
-				struct drm_i915_gem_object *ctx_obj =
-					ctx->engine[i].state;
-				struct intel_ringbuffer *ringbuf =
-					ctx->engine[i].ringbuf;
+				struct intel_engine_ctx *ectx = &ctx->engine[i];
 
 				seq_printf(m, "%s: ", engine->name);
-				if (ctx_obj)
-					describe_obj(m, ctx_obj);
-				if (ringbuf)
-					describe_ctx_ringbuf(m, ringbuf);
+				if (ectx->state)
+					describe_obj(m, ectx->state);
+				if (ectx->ringbuf)
+					describe_ctx_ringbuf(m, ectx->ringbuf);
 				seq_putc(m, '\n');
 			}
 		} else {
drivers/gpu/drm/i915/i915_guc_submission.c +9 -8
index ae1f58d..7352023 100644
@@ -392,6 +392,7 @@ static void guc_init_ctx_desc(struct intel_guc *guc,
 
 	for_each_engine(engine, dev_priv, i) {
 		struct guc_execlist_context *lrc = &desc.lrc[engine->guc_id];
+		struct intel_engine_ctx *ectx = &ctx->engine[i];
 		struct drm_i915_gem_object *obj;
 		uint64_t ctx_desc;
 
@@ -402,7 +403,7 @@ static void guc_init_ctx_desc(struct intel_guc *guc,
 		 * for now who owns a GuC client. But for future owner of GuC
 		 * client, need to make sure lrc is pinned prior to enter here.
 		 */
-		obj = ctx->engine[i].state;
+		obj = ectx->state;
 		if (!obj)
 			break;	/* XXX: continue? */
 
@@ -415,7 +416,7 @@ static void guc_init_ctx_desc(struct intel_guc *guc,
 		lrc->context_id = (client->ctx_index << GUC_ELC_CTXID_OFFSET) |
 				(engine->guc_id << GUC_ELC_ENGINE_OFFSET);
 
-		obj = ctx->engine[i].ringbuf->obj;
+		obj = ectx->ringbuf->obj;
 
 		lrc->ring_begin = i915_gem_obj_ggtt_offset(obj);
 		lrc->ring_end = lrc->ring_begin + obj->base.size - 1;
@@ -987,19 +988,19 @@ int intel_guc_suspend(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_guc *guc = &dev_priv->guc;
-	struct intel_context *ctx;
+	struct intel_engine_ctx *ectx;
 	u32 data[3];
 
 	if (!i915.enable_guc_submission)
 		return 0;
 
-	ctx = dev_priv->kernel_context;
+	ectx = &dev_priv->kernel_context->engine[RCS];
 
 	data[0] = HOST2GUC_ACTION_ENTER_S_STATE;
 	/* any value greater than GUC_POWER_D0 */
 	data[1] = GUC_POWER_D1;
 	/* first page is shared data with GuC */
-	data[2] = i915_gem_obj_ggtt_offset(ctx->engine[RCS].state);
+	data[2] = i915_gem_obj_ggtt_offset(ectx->state);
 
 	return host2guc_action(guc, data, ARRAY_SIZE(data));
 }
@@ -1013,18 +1014,18 @@ int intel_guc_resume(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_guc *guc = &dev_priv->guc;
-	struct intel_context *ctx;
+	struct intel_engine_ctx *ectx;
 	u32 data[3];
 
 	if (!i915.enable_guc_submission)
 		return 0;
 
-	ctx = dev_priv->kernel_context;
+;	ectx = &dev_priv->kernel_context->engine[RCS];
 
 	data[0] = HOST2GUC_ACTION_EXIT_S_STATE;
 	data[1] = GUC_POWER_D0;
 	/* first page is shared data with GuC */
-	data[2] = i915_gem_obj_ggtt_offset(ctx->engine[RCS].state);
+	data[2] = i915_gem_obj_ggtt_offset(ectx->state);
 
 	return host2guc_action(guc, data, ARRAY_SIZE(data));
 }
Signature · 6 lines
1.9.1

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

✗ Fi.CI.BAT: warning for drm/i915: Name the anonymous per-engine context struct (rev2)

In reply to Tvrtko Ursulin
== Series Details ==

== Series Details ==

Series: drm/i915: Name the anonymous per-engine context struct (rev2)
URL : https://patchwork.freedesktop.org/series/4635/
State : warning

== Summary ==

Series 4635v2 drm/i915: Name the anonymous per-engine context struct
http://patchwork.freedesktop.org/api/1.0/series/4635/revisions/2/mbox/

Test gem_ringfill:

        Subgroup basic-default-s3:
                dmesg-warn -> PASS       (bsw-nuc-2)
Test kms_flip:
        Subgroup basic-flip-vs-dpms:
                pass       -> DMESG-WARN (ilk-hp8440p) UNSTABLE
        Subgroup basic-flip-vs-wf_vblank:
                fail       -> PASS       (byt-nuc)
        Subgroup basic-plain-flip:
                pass       -> DMESG-WARN (hsw-brixbox)
Test kms_force_connector_basic:
        Subgroup force-connector-state:
                pass       -> SKIP       (ilk-hp8440p)
Test kms_pipe_crc_basic:
        Subgroup nonblocking-crc-pipe-a:
                dmesg-warn -> PASS       (snb-x220t)
        Subgroup nonblocking-crc-pipe-a-frame-sequence:
                pass       -> DMESG-WARN (snb-dellxps)
        Subgroup suspend-read-crc-pipe-c:
                pass       -> DMESG-WARN (bsw-nuc-2)
                incomplete -> PASS       (hsw-gt2)
Test pm_rpm:
        Subgroup basic-pci-d3-state:
                fail       -> DMESG-FAIL (snb-x220t)
                dmesg-warn -> PASS       (snb-dellxps)
bdw-nuci7        total:194  pass:182  dwarn:0   dfail:0   fail:0   skip:12 
bdw-ultra        total:194  pass:173  dwarn:0   dfail:0   fail:0   skip:21 
bsw-nuc-2        total:194  pass:156  dwarn:1   dfail:0   fail:0   skip:37 
byt-nuc          total:194  pass:159  dwarn:0   dfail:0   fail:0   skip:35 
hsw-brixbox      total:194  pass:171  dwarn:1   dfail:0   fail:0   skip:22 
hsw-gt2          total:194  pass:176  dwarn:1   dfail:0   fail:0   skip:17 
ilk-hp8440p      total:194  pass:129  dwarn:1   dfail:0   fail:0   skip:64 
ivb-t430s        total:194  pass:169  dwarn:0   dfail:0   fail:0   skip:25 
skl-i7k-2        total:194  pass:171  dwarn:0   dfail:0   fail:0   skip:23 
snb-dellxps      total:194  pass:159  dwarn:1   dfail:0   fail:0   skip:34 
snb-x220t        total:194  pass:160  dwarn:0   dfail:1   fail:0   skip:33 

Results at /archive/results/CI_IGT_test/Patchwork_1649/

e7a7673e9840fe8b50a5a2894c75565ec7858a00 drm-intel-nightly: 2016y-03m-19d-10h-09m-53s UTC integration manifest a1c5f9b1e8b9cbfdab0fb71ccf7a5a0838b56069 drm/i915: Name the anonymous per-engine context struct

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

#10 lore raw

Re: ✗ Fi.CI.BAT: warning for drm/i915: Name the anonymous per-engine context struct (rev2)

Quoted 11 lines · Patchwork

== Series Details ==

Series: drm/i915: Name the anonymous per-engine context struct (rev2)
URL : https://patchwork.freedesktop.org/series/4635/
State : warning

== Summary ==

Series 4635v2 drm/i915: Name the anonymous per-engine context struct
http://patchwork.freedesktop.org/api/1.0/series/4635/revisions/2/mbox/

Quoted 6 lines

Test gem_ringfill:

         Subgroup basic-default-s3:
                 dmesg-warn -> PASS       (bsw-nuc-2)
Test kms_flip:
         Subgroup basic-flip-vs-dpms:
                 pass       -> DMESG-WARN (ilk-hp8440p) UNSTABLE

Sporadic ILK pipe underruns: https://bugs.freedesktop.org/show_bug.cgi?id=93787

Quoted 4 lines
         Subgroup basic-flip-vs-wf_vblank:
                 fail       -> PASS       (byt-nuc)
         Subgroup basic-plain-flip:
                 pass       -> DMESG-WARN (hsw-brixbox)

Unrelated intermittent device suspended while HW access: https://bugs.freedesktop.org/show_bug.cgi?id=94349

Quoted 8 lines

Test kms_force_connector_basic:

         Subgroup force-connector-state:
                 pass       -> SKIP       (ilk-hp8440p)
Test kms_pipe_crc_basic:
         Subgroup nonblocking-crc-pipe-a:
                 dmesg-warn -> PASS       (snb-x220t)
         Subgroup nonblocking-crc-pipe-a-frame-sequence:
                 pass       -> DMESG-WARN (snb-dellxps)

Same.

Quoted 2 lines
         Subgroup suspend-read-crc-pipe-c:
                 pass       -> DMESG-WARN (bsw-nuc-2)

Unrelated lockdep splat: https://bugs.freedesktop.org/show_bug.cgi?id=94350

Quoted 4 lines
                 incomplete -> PASS       (hsw-gt2)
Test pm_rpm:
         Subgroup basic-pci-d3-state:
                 fail       -> DMESG-FAIL (snb-x220t)

Device suspended while HW access again.

Quoted 14 lines
                 dmesg-warn -> PASS       (snb-dellxps)
bdw-nuci7        total:194  pass:182  dwarn:0   dfail:0   fail:0   skip:12
bdw-ultra        total:194  pass:173  dwarn:0   dfail:0   fail:0   skip:21
bsw-nuc-2        total:194  pass:156  dwarn:1   dfail:0   fail:0   skip:37
byt-nuc          total:194  pass:159  dwarn:0   dfail:0   fail:0   skip:35
hsw-brixbox      total:194  pass:171  dwarn:1   dfail:0   fail:0   skip:22
hsw-gt2          total:194  pass:176  dwarn:1   dfail:0   fail:0   skip:17
ilk-hp8440p      total:194  pass:129  dwarn:1   dfail:0   fail:0   skip:64
ivb-t430s        total:194  pass:169  dwarn:0   dfail:0   fail:0   skip:25
skl-i7k-2        total:194  pass:171  dwarn:0   dfail:0   fail:0   skip:23
snb-dellxps      total:194  pass:159  dwarn:1   dfail:0   fail:0   skip:34
snb-x220t        total:194  pass:160  dwarn:0   dfail:1   fail:0   skip:33
Quoted 4 lines

Results at /archive/results/CI_IGT_test/Patchwork_1649/

e7a7673e9840fe8b50a5a2894c75565ec7858a00 drm-intel-nightly: 2016y-03m-19d-10h-09m-53s UTC integration manifest a1c5f9b1e8b9cbfdab0fb71ccf7a5a0838b56069 drm/i915: Name the anonymous per-engine context struct

So looking good.

Chris, r-b on v2? It was just a revert of a hunk which changed one instance of ctx->i915->dev->struct_mutex to engine->dev->struct_mutex which the CI reminded me is not allowed in some places.

Regards,

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

#11 lore raw

Re: ✗ Fi.CI.BAT: warning for drm/i915: Name the anonymous per-engine context struct (rev2)

Quoted 41 lines · Tvrtko Ursulin
Quoted 11 lines · Patchwork

== Series Details ==

Series: drm/i915: Name the anonymous per-engine context struct (rev2)
URL : https://patchwork.freedesktop.org/series/4635/
State : warning

== Summary ==

Series 4635v2 drm/i915: Name the anonymous per-engine context struct
http://patchwork.freedesktop.org/api/1.0/series/4635/revisions/2/mbox/

Quoted 6 lines

Test gem_ringfill:

        Subgroup basic-default-s3:
                dmesg-warn -> PASS       (bsw-nuc-2)
Test kms_flip:
        Subgroup basic-flip-vs-dpms:
                pass       -> DMESG-WARN (ilk-hp8440p) UNSTABLE

Sporadic ILK pipe underruns:
https://bugs.freedesktop.org/show_bug.cgi?id=93787

Quoted 4 lines
        Subgroup basic-flip-vs-wf_vblank:
                fail       -> PASS       (byt-nuc)
        Subgroup basic-plain-flip:
                pass       -> DMESG-WARN (hsw-brixbox)

Unrelated intermittent device suspended while HW access:
https://bugs.freedesktop.org/show_bug.cgi?id=94349

Quoted 8 lines

Test kms_force_connector_basic:

        Subgroup force-connector-state:
                pass       -> SKIP       (ilk-hp8440p)
Test kms_pipe_crc_basic:
        Subgroup nonblocking-crc-pipe-a:
                dmesg-warn -> PASS       (snb-x220t)
        Subgroup nonblocking-crc-pipe-a-frame-sequence:
                pass       -> DMESG-WARN (snb-dellxps)
Quoted 7 lines

Same.

Quoted 2 lines
        Subgroup suspend-read-crc-pipe-c:
                pass       -> DMESG-WARN (bsw-nuc-2)

Unrelated lockdep splat: https://bugs.freedesktop.org/show_bug.cgi?id=94350

Is it really the same one? There should be another lockdep chain that isn't in bugzilla...

Quoted 26 lines
Quoted 4 lines
                incomplete -> PASS       (hsw-gt2)
Test pm_rpm:
        Subgroup basic-pci-d3-state:
                fail       -> DMESG-FAIL (snb-x220t)

Device suspended while HW access again.

Quoted 14 lines
                dmesg-warn -> PASS       (snb-dellxps)
bdw-nuci7        total:194  pass:182  dwarn:0   dfail:0   fail:0   skip:12
bdw-ultra        total:194  pass:173  dwarn:0   dfail:0   fail:0   skip:21
bsw-nuc-2        total:194  pass:156  dwarn:1   dfail:0   fail:0   skip:37
byt-nuc          total:194  pass:159  dwarn:0   dfail:0   fail:0   skip:35
hsw-brixbox      total:194  pass:171  dwarn:1   dfail:0   fail:0   skip:22
hsw-gt2          total:194  pass:176  dwarn:1   dfail:0   fail:0   skip:17
ilk-hp8440p      total:194  pass:129  dwarn:1   dfail:0   fail:0   skip:64
ivb-t430s        total:194  pass:169  dwarn:0   dfail:0   fail:0   skip:25
skl-i7k-2        total:194  pass:171  dwarn:0   dfail:0   fail:0   skip:23
snb-dellxps      total:194  pass:159  dwarn:1   dfail:0   fail:0   skip:34
snb-x220t        total:194  pass:160  dwarn:0   dfail:1   fail:0   skip:33
Quoted 4 lines

Results at /archive/results/CI_IGT_test/Patchwork_1649/

e7a7673e9840fe8b50a5a2894c75565ec7858a00 drm-intel-nightly: 2016y-03m-19d-10h-09m-53s UTC integration manifest a1c5f9b1e8b9cbfdab0fb71ccf7a5a0838b56069 drm/i915: Name the anonymous per-engine context struct

Quoted 6 lines

So looking good.

Chris, r-b on v2? It was just a revert of a hunk which changed one instance of ctx->i915->dev->struct_mutex to engine->dev->struct_mutex which the CI reminded me is not allowed in some places.

That one again! One day we will get engine init/fini sorted. Yes, -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
#12 lore raw

Re: ✗ Fi.CI.BAT: warning for drm/i915: Name the anonymous per-engine context struct (rev2)

Quoted 44 lines · Chris Wilson
Quoted 43 lines · Tvrtko Ursulin
Quoted 11 lines · Patchwork

== Series Details ==

Series: drm/i915: Name the anonymous per-engine context struct (rev2)
URL : https://patchwork.freedesktop.org/series/4635/
State : warning

== Summary ==

Series 4635v2 drm/i915: Name the anonymous per-engine context struct
http://patchwork.freedesktop.org/api/1.0/series/4635/revisions/2/mbox/

Quoted 6 lines

Test gem_ringfill:

         Subgroup basic-default-s3:
                 dmesg-warn -> PASS       (bsw-nuc-2)
Test kms_flip:
         Subgroup basic-flip-vs-dpms:
                 pass       -> DMESG-WARN (ilk-hp8440p) UNSTABLE

Sporadic ILK pipe underruns:
https://bugs.freedesktop.org/show_bug.cgi?id=93787

Quoted 4 lines
         Subgroup basic-flip-vs-wf_vblank:
                 fail       -> PASS       (byt-nuc)
         Subgroup basic-plain-flip:
                 pass       -> DMESG-WARN (hsw-brixbox)

Unrelated intermittent device suspended while HW access:
https://bugs.freedesktop.org/show_bug.cgi?id=94349

Quoted 8 lines

Test kms_force_connector_basic:

         Subgroup force-connector-state:
                 pass       -> SKIP       (ilk-hp8440p)
Test kms_pipe_crc_basic:
         Subgroup nonblocking-crc-pipe-a:
                 dmesg-warn -> PASS       (snb-x220t)
         Subgroup nonblocking-crc-pipe-a-frame-sequence:
                 pass       -> DMESG-WARN (snb-dellxps)

Same.

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines
         Subgroup suspend-read-crc-pipe-c:
                 pass       -> DMESG-WARN (bsw-nuc-2)

Unrelated lockdep splat: https://bugs.freedesktop.org/show_bug.cgi?id=94350

Is it really the same one? There should be another lockdep chain that isn't in bugzilla...

Looks the same to me:

Bz:

[  179.762863] rtcwake/5995 is trying to acquire lock:
[  179.762877]  (s_active#6){++++.+}, at: [<ffffffff8124ec70>] kernfs_remove_by_name_ns+0x40/0xa0
[  179.762878]
but task is already holding lock:
[  179.762885]  (cpu_hotplug.lock){+.+.+.}, at: [<ffffffff81078c4d>] cpu_hotplug_begin+0x6d/0xc0
[  179.762886]

which lock already depends on the new lock.

This CI run:

[  127.210522] rtcwake/5947 is trying to acquire lock:
[  127.210539]  (s_active#6){++++.+}, at: [<ffffffff81250740>] kernfs_remove_by_name_ns+0x40/0xa0
[  127.210540]
but task is already holding lock:
[  127.210549]  (cpu_hotplug.lock){+.+.+.}, at: [<ffffffff81078f5d>] cpu_hotplug_begin+0x6d/0xc0
[  127.210550]

which lock already depends on the new lock.

Quoted 27 lines
Quoted 27 lines
Quoted 5 lines
                 incomplete -> PASS       (hsw-gt2)
Test pm_rpm:
         Subgroup basic-pci-d3-state:
                 fail       -> DMESG-FAIL (snb-x220t)

Device suspended while HW access again.

Quoted 18 lines
                 dmesg-warn -> PASS       (snb-dellxps)
bdw-nuci7        total:194  pass:182  dwarn:0   dfail:0   fail:0   skip:12
bdw-ultra        total:194  pass:173  dwarn:0   dfail:0   fail:0   skip:21
bsw-nuc-2        total:194  pass:156  dwarn:1   dfail:0   fail:0   skip:37
byt-nuc          total:194  pass:159  dwarn:0   dfail:0   fail:0   skip:35
hsw-brixbox      total:194  pass:171  dwarn:1   dfail:0   fail:0   skip:22
hsw-gt2          total:194  pass:176  dwarn:1   dfail:0   fail:0   skip:17
ilk-hp8440p      total:194  pass:129  dwarn:1   dfail:0   fail:0   skip:64
ivb-t430s        total:194  pass:169  dwarn:0   dfail:0   fail:0   skip:25
skl-i7k-2        total:194  pass:171  dwarn:0   dfail:0   fail:0   skip:23
snb-dellxps      total:194  pass:159  dwarn:1   dfail:0   fail:0   skip:34
snb-x220t        total:194  pass:160  dwarn:0   dfail:1   fail:0   skip:33

Results at /archive/results/CI_IGT_test/Patchwork_1649/

e7a7673e9840fe8b50a5a2894c75565ec7858a00 drm-intel-nightly: 2016y-03m-19d-10h-09m-53s UTC integration manifest a1c5f9b1e8b9cbfdab0fb71ccf7a5a0838b56069 drm/i915: Name the anonymous per-engine context struct

Quoted 8 lines
Quoted 6 lines

So looking good.

Chris, r-b on v2? It was just a revert of a hunk which changed one instance of ctx->i915->dev->struct_mutex to engine->dev->struct_mutex which the CI reminded me is not allowed in some places.

That one again! One day we will get engine init/fini sorted. Yes,

Yes r-b, just to be really sure?

Regards,

Tvrtko

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

Reviewed-by
#13 lore raw

Re: ✗ Fi.CI.BAT: warning for drm/i915: Name the anonymous per-engine context struct (rev2)

Quoted 29 lines · Tvrtko Ursulin
Quoted 2 lines · Chris Wilson

Is it really the same one? There should be another lockdep chain that isn't in bugzilla...

Looks the same to me:

Bz:

[  179.762863] rtcwake/5995 is trying to acquire lock:
[  179.762877]  (s_active#6){++++.+}, at: [<ffffffff8124ec70>]
kernfs_remove_by_name_ns+0x40/0xa0
[  179.762878]
but task is already holding lock:
[  179.762885]  (cpu_hotplug.lock){+.+.+.}, at: [<ffffffff81078c4d>]
cpu_hotplug_begin+0x6d/0xc0
[  179.762886]

which lock already depends on the new lock.

This CI run:

[  127.210522] rtcwake/5947 is trying to acquire lock:
[  127.210539]  (s_active#6){++++.+}, at: [<ffffffff81250740>]
kernfs_remove_by_name_ns+0x40/0xa0
[  127.210540]
but task is already holding lock:
[  127.210549]  (cpu_hotplug.lock){+.+.+.}, at: [<ffffffff81078f5d>]
cpu_hotplug_begin+0x6d/0xc0
[  127.210550]
Quoted 1 line

which lock already depends on the new lock.

Ok, the chain I'm looking for doesn't involve the kernfs link.

Quoted 29 lines
Quoted 29 lines
Quoted 29 lines
Quoted 5 lines
                incomplete -> PASS       (hsw-gt2)
Test pm_rpm:
        Subgroup basic-pci-d3-state:
                fail       -> DMESG-FAIL (snb-x220t)

Device suspended while HW access again.

Quoted 14 lines
                dmesg-warn -> PASS       (snb-dellxps)
bdw-nuci7        total:194  pass:182  dwarn:0   dfail:0   fail:0   skip:12
bdw-ultra        total:194  pass:173  dwarn:0   dfail:0   fail:0   skip:21
bsw-nuc-2        total:194  pass:156  dwarn:1   dfail:0   fail:0   skip:37
byt-nuc          total:194  pass:159  dwarn:0   dfail:0   fail:0   skip:35
hsw-brixbox      total:194  pass:171  dwarn:1   dfail:0   fail:0   skip:22
hsw-gt2          total:194  pass:176  dwarn:1   dfail:0   fail:0   skip:17
ilk-hp8440p      total:194  pass:129  dwarn:1   dfail:0   fail:0   skip:64
ivb-t430s        total:194  pass:169  dwarn:0   dfail:0   fail:0   skip:25
skl-i7k-2        total:194  pass:171  dwarn:0   dfail:0   fail:0   skip:23
snb-dellxps      total:194  pass:159  dwarn:1   dfail:0   fail:0   skip:34
snb-x220t        total:194  pass:160  dwarn:0   dfail:1   fail:0   skip:33
Quoted 4 lines

Results at /archive/results/CI_IGT_test/Patchwork_1649/

e7a7673e9840fe8b50a5a2894c75565ec7858a00 drm-intel-nightly: 2016y-03m-19d-10h-09m-53s UTC integration manifest a1c5f9b1e8b9cbfdab0fb71ccf7a5a0838b56069 drm/i915: Name the anonymous per-engine context struct

So looking good.

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

Chris, r-b on v2? It was just a revert of a hunk which changed one instance of ctx->i915->dev->struct_mutex to engine->dev->struct_mutex which the CI reminded me is not allowed in some places.

That one again! One day we will get engine init/fini sorted. Yes,

Yes r-b, just to be really sure?

-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