Skip to contentSkip to conversation

igt/gem_reset_stats: Fix pending batches status expectation

i-g-t Antonio Argenziano 3 participants 5 messages Jun 9, 2017 – Jun 20, 2017

lore
Share & tools
lore
#0 lore raw

Test expects pending batches to be discarded after a reset. That is no longer the case. Fixed to expect a normal execution.

Cc Michel Thierry <[email protected]>
---
 tests/gem_reset_stats.c | 24 +++++++-----------------
 1 file changed, 7 insertions(+), 17 deletions(-)
tests/gem_reset_stats.c +7 -17
index c4ce4ac2..73afeeb2 100644
@@ -239,7 +239,7 @@ static void test_rs(const struct intel_execution_engine *e,
 		if (i == hang_index)
 			assert_reset_status(i, fd[i], 0, RS_BATCH_ACTIVE);
 		if (i > hang_index)
-			assert_reset_status(i, fd[i], 0, RS_BATCH_PENDING);
+			assert_reset_status(i, fd[i], 0, RS_NO_ERROR);
 	}
 
 	igt_assert(igt_seconds_elapsed(&ts_injected) <= 30);
@@ -312,10 +312,10 @@ static void test_rs_ctx(const struct intel_execution_engine *e,
 						    RS_BATCH_ACTIVE);
 			if (i == hang_index && j > hang_context)
 				assert_reset_status(i, fd[i], ctx[i][j],
-						    RS_BATCH_PENDING);
+						    RS_NO_ERROR);
 			if (i > hang_index)
 				assert_reset_status(i, fd[i], ctx[i][j],
-						    RS_BATCH_PENDING);
+						    RS_NO_ERROR);
 		}
 	}
 
@@ -330,7 +330,7 @@ static void test_ban(const struct intel_execution_engine *e)
 	struct local_drm_i915_reset_stats rs_bad, rs_good;
 	int fd_bad, fd_good;
 	int ban, retry = 10;
-	int active_count = 0, pending_count = 0;
+	int active_count = 0;
 
 	fd_bad = drm_open_driver(DRIVER_INTEL);
 	fd_good = drm_open_driver(DRIVER_INTEL);
@@ -350,9 +350,6 @@ static void test_ban(const struct intel_execution_engine *e)
 	noop(fd_good, 0, e);
 	noop(fd_good, 0, e);
 
-	/* The second hang will count as pending and be discarded */
-	active_count--;
-	pending_count += 1; /* inject hang does 1 real exec + 1 dummy */
 	while (retry--) {
 		inject_hang(fd_bad, 0, e, BAN);
 		active_count++;
@@ -373,12 +370,10 @@ static void test_ban(const struct intel_execution_engine *e)
 	assert_reset_status(fd_bad, fd_bad, 0, RS_BATCH_ACTIVE);
 	igt_assert_eq(gem_reset_stats(fd_bad, 0, &rs_bad), 0);
 	igt_assert_eq(rs_bad.batch_active, active_count);
-	igt_assert_eq(rs_bad.batch_pending, pending_count);
 
-	assert_reset_status(fd_good, fd_good, 0, RS_BATCH_PENDING);
+	assert_reset_status(fd_good, fd_good, 0, RS_NO_ERROR);
 	igt_assert_eq(gem_reset_stats(fd_good, 0, &rs_good), 0);
 	igt_assert_eq(rs_good.batch_active, 0);
-	igt_assert_eq(rs_good.batch_pending, 2);
 
 	close(fd_bad);
 	close(fd_good);
@@ -389,7 +384,7 @@ static void test_ban_ctx(const struct intel_execution_engine *e)
 	struct local_drm_i915_reset_stats rs_bad, rs_good;
 	int fd, ban, retry = 10;
 	uint32_t ctx_good, ctx_bad;
-	int active_count = 0, pending_count = 0;
+	int active_count = 0;
 
 	fd = drm_open_driver(DRIVER_INTEL);
 
@@ -414,9 +409,6 @@ static void test_ban_ctx(const struct intel_execution_engine *e)
 	noop(fd, ctx_good, e);
 	noop(fd, ctx_good, e);
 
-	/* This second hang will count as pending and be discarded */
-	active_count--;
-	pending_count++;
 	while (retry--) {
 		inject_hang(fd, ctx_bad, e, BAN);
 		active_count++;
@@ -437,12 +429,10 @@ static void test_ban_ctx(const struct intel_execution_engine *e)
 	assert_reset_status(fd, fd, ctx_bad, RS_BATCH_ACTIVE);
 	igt_assert_eq(gem_reset_stats(fd, ctx_bad, &rs_bad), 0);
 	igt_assert_eq(rs_bad.batch_active, active_count);
-	igt_assert_eq(rs_bad.batch_pending, pending_count);
 
-	assert_reset_status(fd, fd, ctx_good, RS_BATCH_PENDING);
+	assert_reset_status(fd, fd, ctx_good, RS_NO_ERROR);
 	igt_assert_eq(gem_reset_stats(fd, ctx_good, &rs_good), 0);
 	igt_assert_eq(rs_good.batch_active, 0);
-	igt_assert_eq(rs_good.batch_pending, 2);
 
 	close(fd);
 }
Signature · 6 lines
2.11.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Reviewed-by
#1 lore raw
Quoted 2 lines · Antonio Argenziano

Test expects pending batches to be discarded after a reset. That is no longer the case. Fixed to expect a normal execution.

You could expand this to say:
after commit 821ed7df6e2a ("drm/i915: Update reset path to fix incomplete requests"), that is no longer the case.

Quoted 4 lines
Cc Michel Thierry <[email protected]>

If we want the test to pass, then it's ok. Someone else may say we need further subtests.

On the basis this brings existing tests to the current expectation,

(with the updated commit msg).

-Michel

Quoted 87 lines
---
  tests/gem_reset_stats.c | 24 +++++++-----------------
  1 file changed, 7 insertions(+), 17 deletions(-)
tests/gem_reset_stats.c +6 -14
index c4ce4ac2..73afeeb2 100644
@@ -239,7 +239,7 @@ static void test_rs(const struct intel_execution_engine *e,
  		if (i == hang_index)
  			assert_reset_status(i, fd[i], 0, RS_BATCH_ACTIVE);
  		if (i > hang_index)
-			assert_reset_status(i, fd[i], 0, RS_BATCH_PENDING);
+			assert_reset_status(i, fd[i], 0, RS_NO_ERROR);
  	}
 
  	igt_assert(igt_seconds_elapsed(&ts_injected) <= 30);
@@ -312,10 +312,10 @@ static void test_rs_ctx(const struct intel_execution_engine *e,
  						    RS_BATCH_ACTIVE);
  			if (i == hang_index && j > hang_context)
  				assert_reset_status(i, fd[i], ctx[i][j],
-						    RS_BATCH_PENDING);
+						    RS_NO_ERROR);
  			if (i > hang_index)
  				assert_reset_status(i, fd[i], ctx[i][j],
-						    RS_BATCH_PENDING);
+						    RS_NO_ERROR);
  		}
  	}
 
@@ -330,7 +330,7 @@ static void test_ban(const struct intel_execution_engine *e)
  	struct local_drm_i915_reset_stats rs_bad, rs_good;
  	int fd_bad, fd_good;
  	int ban, retry = 10;
-	int active_count = 0, pending_count = 0;
+	int active_count = 0;
 
  	fd_bad = drm_open_driver(DRIVER_INTEL);
  	fd_good = drm_open_driver(DRIVER_INTEL);
@@ -350,9 +350,6 @@ static void test_ban(const struct intel_execution_engine *e)
  	noop(fd_good, 0, e);
  	noop(fd_good, 0, e);
 
-	/* The second hang will count as pending and be discarded */
-	active_count--;
-	pending_count += 1; /* inject hang does 1 real exec + 1 dummy */
  	while (retry--) {
  		inject_hang(fd_bad, 0, e, BAN);
  		active_count++;
@@ -373,12 +370,10 @@ static void test_ban(const struct intel_execution_engine *e)
  	assert_reset_status(fd_bad, fd_bad, 0, RS_BATCH_ACTIVE);
  	igt_assert_eq(gem_reset_stats(fd_bad, 0, &rs_bad), 0);
  	igt_assert_eq(rs_bad.batch_active, active_count);
-	igt_assert_eq(rs_bad.batch_pending, pending_count);
 
-	assert_reset_status(fd_good, fd_good, 0, RS_BATCH_PENDING);
+	assert_reset_status(fd_good, fd_good, 0, RS_NO_ERROR);
  	igt_assert_eq(gem_reset_stats(fd_good, 0, &rs_good), 0);
  	igt_assert_eq(rs_good.batch_active, 0);
-	igt_assert_eq(rs_good.batch_pending, 2);
 
  	close(fd_bad);
  	close(fd_good);
@@ -389,7 +384,7 @@ static void test_ban_ctx(const struct intel_execution_engine *e)
  	struct local_drm_i915_reset_stats rs_bad, rs_good;
  	int fd, ban, retry = 10;
  	uint32_t ctx_good, ctx_bad;
-	int active_count = 0, pending_count = 0;
+	int active_count = 0;
 
  	fd = drm_open_driver(DRIVER_INTEL);
 
@@ -414,9 +409,6 @@ static void test_ban_ctx(const struct intel_execution_engine *e)
  	noop(fd, ctx_good, e);
  	noop(fd, ctx_good, e);
 
-	/* This second hang will count as pending and be discarded */
-	active_count--;
-	pending_count++;
  	while (retry--) {
  		inject_hang(fd, ctx_bad, e, BAN);
  		active_count++;
@@ -437,12 +429,10 @@ static void test_ban_ctx(const struct intel_execution_engine *e)
  	assert_reset_status(fd, fd, ctx_bad, RS_BATCH_ACTIVE);
  	igt_assert_eq(gem_reset_stats(fd, ctx_bad, &rs_bad), 0);
  	igt_assert_eq(rs_bad.batch_active, active_count);

tests/gem_reset_stats.c static void test_ban_ctx(const struct intel_execution_engine *e) · line 438

Quoted 10 lines
tests/gem_reset_stats.c +1 -3
-	igt_assert_eq(rs_bad.batch_pending, pending_count);
 
-	assert_reset_status(fd, fd, ctx_good, RS_BATCH_PENDING);
+	assert_reset_status(fd, fd, ctx_good, RS_NO_ERROR);
  	igt_assert_eq(gem_reset_stats(fd, ctx_good, &rs_good), 0);
  	igt_assert_eq(rs_good.batch_active, 0);
-	igt_assert_eq(rs_good.batch_pending, 2);
 
  	close(fd);
  }

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

#2 lore raw
Quoted 16 lines · Michel Thierry
Quoted 2 lines · Antonio Argenziano

Test expects pending batches to be discarded after a reset. That is no longer the case. Fixed to expect a normal execution.

You could expand this to say:
after commit 821ed7df6e2a ("drm/i915: Update reset path to fix incomplete requests"), that is no longer the case.

Quoted 4 lines
Cc Michel Thierry <[email protected]>

If we want the test to pass, then it's ok. Someone else may say we need further subtests.

Quoted 6 lines

On the basis this brings existing tests to the current expectation,

(with the updated commit msg).

-Michel

Hey,

What is the plan regarding this patch? It has been stale for a while.

We can push it with the updated commit message if Antonio is okay with the change.

On conditional reviews, it is a good practice, if the original author follows up with the "fixed" commit as a reply or gives us assent on IRC to do the change on our side.

Also, from CONTRIBUTING:
-----------------------------------------------------------------------------
Please use --subject-prefix="PATCH i-g-t" so that i-g-t patches are easily identified in the massive amount mails on intel-gfx. To ensure this is always done, autogen.sh will run:

git config format.subjectprefix "PATCH i-g-t"

on its first invocation.
-----------------------------------------------------------------------------

Lack of proper prefix breaks filtering / patchwork and makes changes harder to track.

The autogen.sh thing is a recent addition.

Thanks!

Signature · 6 lines
Cheers,
Arek
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw
Quoted 26 lines · Arkadiusz Hiler
Quoted 16 lines · Michel Thierry
Quoted 2 lines · Antonio Argenziano

Test expects pending batches to be discarded after a reset. That is no longer the case. Fixed to expect a normal execution.

You could expand this to say:
after commit 821ed7df6e2a ("drm/i915: Update reset path to fix incomplete requests"), that is no longer the case.

Quoted 4 lines
Cc Michel Thierry <[email protected]>

If we want the test to pass, then it's ok. Someone else may say we need further subtests.

Quoted 6 lines

On the basis this brings existing tests to the current expectation,

(with the updated commit msg).

-Michel

Hey,

Quoted 8 lines

What is the plan regarding this patch? It has been stale for a while.

We can push it with the updated commit message if Antonio is okay with the change.

On conditional reviews, it is a good practice, if the original author follows up with the "fixed" commit as a reply or gives us assent on IRC to do the change on our side.

Hi Arek, I have resent the patch with the corrected commit message. Michel RB'd it again. Just needs to be merged ;)

Quoted 9 lines

Also, from CONTRIBUTING:
-----------------------------------------------------------------------------
Please use --subject-prefix="PATCH i-g-t" so that i-g-t patches are easily identified in the massive amount mails on intel-gfx. To ensure this is always done, autogen.sh will run:

Quoted 4 lines

git config format.subjectprefix "PATCH i-g-t"

on its first invocation.
-----------------------------------------------------------------------------

Sorry I forgot the prefix, I agree it makes things more readable. Also the new patch doesn't have it. Do you want me to re-push it with the prefix?

Quoted 8 lines

Lack of proper prefix breaks filtering / patchwork and makes changes harder to track.

The autogen.sh thing is a recent addition.

Thanks!

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

#4 lore raw
Quoted 48 lines · Antonio Argenziano
Quoted 26 lines · Arkadiusz Hiler
Quoted 16 lines · Michel Thierry
Quoted 2 lines · Antonio Argenziano

Test expects pending batches to be discarded after a reset. That is no longer the case. Fixed to expect a normal execution.

You could expand this to say:
after commit 821ed7df6e2a ("drm/i915: Update reset path to fix incomplete requests"), that is no longer the case.

Quoted 4 lines
Cc Michel Thierry <[email protected]>

If we want the test to pass, then it's ok. Someone else may say we need further subtests.

Quoted 6 lines

On the basis this brings existing tests to the current expectation,

(with the updated commit msg).

-Michel

Hey,

Quoted 8 lines

What is the plan regarding this patch? It has been stale for a while.

We can push it with the updated commit message if Antonio is okay with the change.

On conditional reviews, it is a good practice, if the original author follows up with the "fixed" commit as a reply or gives us assent on IRC to do the change on our side.

Hi Arek, I have resent the patch with the corrected commit message. Michel RB'd it again. Just needs to be merged ;)

Quoted 6 lines

Also, from CONTRIBUTING:
-----------------------------------------------------------------------------
Please use --subject-prefix="PATCH i-g-t" so that i-g-t patches are easily identified in the massive amount mails on intel-gfx. To ensure this is always done, autogen.sh will run:

Quoted 7 lines
Quoted 4 lines

git config format.subjectprefix "PATCH i-g-t"

on its first invocation.
-----------------------------------------------------------------------------

Sorry I forgot the prefix, I agree it makes things more readable. Also the new patch doesn't have it. Do you want me to re-push it with the prefix?

That is why I have missed it. No need for resending, I'll push it in a minute - just try not forgetting it next time :-)

Thanks!

Signature · 2 lines
Cheers,
Arek
Quoted 9 lines
Quoted 9 lines

Lack of proper prefix breaks filtering / patchwork and makes changes harder to track.

The autogen.sh thing is a recent addition.

Thanks!

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