Skip to contentSkip to conversation

tests/gem_eio: Skip in-flight-suspend on snb

PATCH In review Daniel Vetter 7 participants 12 messages Oct 19, 2017 – Oct 26, 2017

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 tests/gem_eio: Skip in-flight-suspend on snb

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]>
References: https://intel-gfx-ci.01.org/tree/drm-tip/igt@[email protected]
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

---
 tests/gem_eio.c | 3 +++
 1 file changed, 3 insertions(+)
tests/gem_eio.c +3 -0
index 899cb62728e3..28375e208232 100644
@@ -218,6 +218,9 @@ static void test_inflight_suspend(int fd)
 	igt_require(gem_has_exec_fence(fd));
 	igt_require(i915_reset_control(false));
 
+	igt_skip_on_f(IS_SANDYBRIDGE(intel_get_drm_devid(fd)),
+		      "random incompletes in CI with this test\n");
+
 	memset(obj, 0, sizeof(obj));
 	obj[0].flags = EXEC_OBJECT_WRITE;
 	obj[1].handle = gem_create(fd, 4096);
Signature · 6 lines
2.15.0.rc1

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw
Quoted 5 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Quoted 7 lines

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]>
References: https://intel-gfx-ci.01.org/tree/drm-tip/igt@[email protected]
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

For more context, since I forgot to add: I'm definitely not advertising for abusing igt_skip to handle problematic testcases in general. What makes this special here is the combo of
- new testcase
- old machine where we don't have priority to fix things

Hence why I think it'll make sense to treat this as a feature-like thing, where we simply skip if stuff doesn't work/isn't exposed on older platforms and shrug it off. And once someone does a free time project to fix it up, we can then remove the skip.

I hope that explains a bit the reasoning from my behind using skip here.

The other bit is that if/once Maarten figured out what's wrong with watermarks, we should be able to enable shard-snb reporting in CI results, which would be really great. Except we really can't have tests that incomplete, because they victimize too much else and so would need to blacklist until fixed one way or the other anyways.
-Daniel

Quoted 21 lines
---
 tests/gem_eio.c | 3 +++
 1 file changed, 3 insertions(+)
tests/gem_eio.c +3 -0
index 899cb62728e3..28375e208232 100644
@@ -218,6 +218,9 @@ static void test_inflight_suspend(int fd)
 	igt_require(gem_has_exec_fence(fd));
 	igt_require(i915_reset_control(false));
 
+	igt_skip_on_f(IS_SANDYBRIDGE(intel_get_drm_devid(fd)),
+		      "random incompletes in CI with this test\n");
+
 	memset(obj, 0, sizeof(obj));
 	obj[0].flags = EXEC_OBJECT_WRITE;
 	obj[1].handle = gem_create(fd, 4096);
Signature · 1 line
2.15.0.rc1
Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
Quoted 35 lines

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Thursday, October 19, 2017 12:57 PM
To: Intel Graphics Development <[email protected]>
Cc: Daniel Vetter <[email protected]>; Joonas Lahtinen <[email protected]>; Chris Wilson <[email protected]>; Lofstedt, Marta <[email protected]>; Martin Peres <[email protected]>
Subject: Re: [PATCH i-g-t] tests/gem_eio: Skip in-flight-suspend on snb

Quoted 13 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]> References:
https://intel-gfx-ci.01.org/tree/drm-tip/igt@gem_eio@in-flight-suspend
.html
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

Quoted 1 line

For more context, since I forgot to add: I'm definitely not advertising for abusing igt_skip to handle problematic testcases in general. What makes this special here is the combo of
- new testcase
- old machine where we don't have priority to fix things

Quoted 6 lines

Hence why I think it'll make sense to treat this as a feature-like thing, where we simply skip if stuff doesn't work/isn't exposed on older platforms and shrug it off. And once someone does a free time project to fix it up, we can then remove the skip.

I hope that explains a bit the reasoning from my behind using skip here.

I am not buying this.
Could you define which old machines that we are not going to care about to find out that we are having this real issue? I also don't understand why new test-cases should be treated differently compare to the old bad behaving ones we already have.

/Marta

Quoted 20 lines

The other bit is that if/once Maarten figured out what's wrong with watermarks, we should be able to enable shard-snb reporting in CI results, which would be really great. Except we really can't have tests that incomplete, because they victimize too much else and so would need to blacklist until fixed one way or the other anyways.
-Daniel

Quoted 12 lines
---
 tests/gem_eio.c | 3 +++
 1 file changed, 3 insertions(+)
tests/gem_eio.c index renamed from tests/gem_eio.c +0 -0

899cb62728e3..28375e208232 100644

tests/gem_eio.c +0 -0
@@ -218,6 +218,9 @@ static void test_inflight_suspend(int fd)
 	igt_require(gem_has_exec_fence(fd));
 	igt_require(i915_reset_control(false));
 
Quoted 14 lines
Quoted 9 lines
tests/gem_eio.c +3 -1
+	igt_skip_on_f(IS_SANDYBRIDGE(intel_get_drm_devid(fd)),
+		      "random incompletes in CI with this test\n");
+
 	memset(obj, 0, sizeof(obj));
 	obj[0].flags = EXEC_OBJECT_WRITE;
 	obj[1].handle = gem_create(fd, 4096);
--

2.15.0.rc1

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

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

#3 lore raw
Quoted 40 lines · Marta
Quoted 40 lines

-----Original Message-----
From: Daniel Vetter [mailto:[email protected]] On Behalf Of Daniel Vetter
Sent: Thursday, October 19, 2017 12:57 PM
To: Intel Graphics Development <[email protected]>
Cc: Daniel Vetter <[email protected]>; Joonas Lahtinen <[email protected]>; Chris Wilson <[email protected]>; Lofstedt, Marta <[email protected]>; Martin Peres <[email protected]>
Subject: Re: [PATCH i-g-t] tests/gem_eio: Skip in-flight-suspend on snb

Quoted 13 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]> References:
https://intel-gfx-ci.01.org/tree/drm-tip/igt@gem_eio@in-flight-suspend
.html
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

Quoted 1 line

For more context, since I forgot to add: I'm definitely not advertising for abusing igt_skip to handle problematic testcases in general. What makes this special here is the combo of
- new testcase
- old machine where we don't have priority to fix things

Hence why I think it'll make sense to treat this as a feature-like thing, where we simply skip if stuff doesn't work/isn't exposed on older platforms and shrug it off. And once someone does a free time project to fix it up, we can then remove the skip.

Quoted 5 lines
Quoted 1 line

I hope that explains a bit the reasoning from my behind using skip here.

I am not buying this.
Could you define which old machines that we are not going to care about to find out that we are having this real issue? I also don't understand why new test-cases should be treated differently compare to the old bad behaving ones we already have.

What other old bad behaving ones do we have? The only other incomplete I'm seeing on older boxes is gem_exec_suspend@basice-s3, and I'm semi-tempted to do the same for that box too. But at least with fast-feedback the run order is fixed, so the tests which are not run are always the same ones.

There's one more cibuglog entry for pre-gen9 machines:

https://bugs.freedesktop.org/show_bug.cgi?id=102890

But cibuglog stats say reproduction rate is just 1%. I think that's ok, even for an incomplete.

I don't see anything else. There's tons of issues on newer machines where we still care, and where we still are working on stabilizing them. But nothing else causing incompletes on gen8 or older afaict. -Daniel

Quoted 36 lines

/Marta

Quoted 34 lines

The other bit is that if/once Maarten figured out what's wrong with watermarks, we should be able to enable shard-snb reporting in CI results, which would be really great. Except we really can't have tests that incomplete, because they victimize too much else and so would need to blacklist until fixed one way or the other anyways.
-Daniel

Quoted 19 lines
---
 tests/gem_eio.c | 3 +++
 1 file changed, 3 insertions(+)
tests/gem_eio.c index renamed from tests/gem_eio.c +0 -0

899cb62728e3..28375e208232 100644

tests/gem_eio.c +3 -1
@@ -218,6 +218,9 @@ static void test_inflight_suspend(int fd)
    igt_require(gem_has_exec_fence(fd));
    igt_require(i915_reset_control(false));
 
+   igt_skip_on_f(IS_SANDYBRIDGE(intel_get_drm_devid(fd)),
+                 "random incompletes in CI with this test\n");
+
    memset(obj, 0, sizeof(obj));
    obj[0].flags = EXEC_OBJECT_WRITE;
    obj[1].handle = gem_create(fd, 4096);
--
Quoted 2 lines

2.15.0.rc1

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

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#4 lore raw

✗ Fi.CI.BAT: failure for tests/gem_eio: Skip in-flight-suspend on snb

In reply to Daniel Vetter
== Series Details ==

== Series Details ==

Series: tests/gem_eio: Skip in-flight-suspend on snb
URL : https://patchwork.freedesktop.org/series/32280/
State : failure

== Summary ==

IGT patchset tested on top of latest successful build abc08cba366a64a07f7f4deb167ae7d6ae059958 lib: Free all internal buffers before measuring available memory

with latest DRM-Tip kernel build CI_DRM_3261
399dd92e2b42 drm-tip: 2017y-10m-18d-21h-54m-46s UTC integration manifest

No testlist changes.

Test kms_pipe_crc_basic:

        Subgroup read-crc-pipe-a-frame-sequence:
                pass       -> INCOMPLETE (fi-skl-6700hq)
fi-bdw-5557u     total:289  pass:268  dwarn:0   dfail:0   fail:0   skip:21  time:441s
fi-bdw-gvtdvm    total:289  pass:265  dwarn:0   dfail:0   fail:0   skip:24  time:455s
fi-blb-e6850     total:289  pass:223  dwarn:1   dfail:0   fail:0   skip:65  time:376s
fi-bsw-n3050     total:289  pass:243  dwarn:0   dfail:0   fail:0   skip:46  time:529s
fi-bwr-2160      total:289  pass:183  dwarn:0   dfail:0   fail:0   skip:106 time:264s
fi-bxt-dsi       total:289  pass:259  dwarn:0   dfail:0   fail:0   skip:30  time:493s
fi-bxt-j4205     total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:496s
fi-byt-j1900     total:289  pass:253  dwarn:1   dfail:0   fail:0   skip:35  time:499s
fi-byt-n2820     total:289  pass:249  dwarn:1   dfail:0   fail:0   skip:39  time:480s
fi-cfl-s         total:289  pass:253  dwarn:4   dfail:0   fail:0   skip:32  time:564s
fi-elk-e7500     total:289  pass:229  dwarn:0   dfail:0   fail:0   skip:60  time:413s
fi-gdg-551       total:289  pass:178  dwarn:1   dfail:0   fail:1   skip:109 time:248s
fi-glk-1         total:289  pass:261  dwarn:0   dfail:0   fail:0   skip:28  time:586s
fi-hsw-4770      total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:455s
fi-hsw-4770r     total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:431s
fi-ilk-650       total:289  pass:228  dwarn:0   dfail:0   fail:0   skip:61  time:433s
fi-ivb-3520m     total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:486s
fi-ivb-3770      total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:463s
fi-kbl-7500u     total:289  pass:264  dwarn:1   dfail:0   fail:0   skip:24  time:488s
fi-kbl-7560u     total:289  pass:270  dwarn:0   dfail:0   fail:0   skip:19  time:572s
fi-kbl-7567u     total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:478s
fi-kbl-r         total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:584s
fi-pnv-d510      total:289  pass:222  dwarn:1   dfail:0   fail:0   skip:66  time:552s
fi-skl-6260u     total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:450s
fi-skl-6700hq    total:240  pass:215  dwarn:0   dfail:0   fail:0   skip:24 
fi-skl-6700k     total:289  pass:265  dwarn:0   dfail:0   fail:0   skip:24  time:524s
fi-skl-6770hq    total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:510s
fi-skl-gvtdvm    total:289  pass:266  dwarn:0   dfail:0   fail:0   skip:23  time:456s
fi-snb-2520m     total:289  pass:250  dwarn:0   dfail:0   fail:0   skip:39  time:571s
fi-snb-2600      total:289  pass:249  dwarn:0   dfail:0   fail:0   skip:40  time:430s

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_384/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#5 lore raw

Re: ✗ Fi.CI.BAT: failure for tests/gem_eio: Skip in-flight-suspend on snb

Ah, domains.

It is a real bug, but the quickest way to paper over it is to s/TEST_DEVICES/TEST_NONE/ (that way we still have the test in arguably a more realistic mode), we have other reset tests that also trigger the death. The bug tells us reset on snb is unreliable (writing to GDRST causes a machine hang a short time later before the reset is complete). Exactly what engine setup prior to the GPU hang triggers the death and how to workaround the death are currently eluding me.
-Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#6 lore raw

✗ Fi.CI.BAT: warning for tests/gem_eio: Skip in-flight-suspend on snb

In reply to Daniel Vetter
== Series Details ==

== Series Details ==

Series: tests/gem_eio: Skip in-flight-suspend on snb
URL : https://patchwork.freedesktop.org/series/32280/
State : warning

== Summary ==

IGT patchset tested on top of latest successful build abc08cba366a64a07f7f4deb167ae7d6ae059958 lib: Free all internal buffers before measuring available memory

with latest DRM-Tip kernel build CI_DRM_3265
93f001963c99 drm-tip: 2017y-10m-19d-10h-52m-17s UTC integration manifest

No testlist changes.

Test drv_module_reload:

        Subgroup basic-no-display:
                pass       -> DMESG-WARN (fi-skl-6770hq)
fi-bdw-5557u     total:289  pass:268  dwarn:0   dfail:0   fail:0   skip:21  time:445s
fi-bdw-gvtdvm    total:289  pass:265  dwarn:0   dfail:0   fail:0   skip:24  time:453s
fi-blb-e6850     total:289  pass:223  dwarn:1   dfail:0   fail:0   skip:65  time:373s
fi-bsw-n3050     total:289  pass:243  dwarn:0   dfail:0   fail:0   skip:46  time:525s
fi-bwr-2160      total:289  pass:183  dwarn:0   dfail:0   fail:0   skip:106 time:264s
fi-bxt-dsi       total:289  pass:259  dwarn:0   dfail:0   fail:0   skip:30  time:504s
fi-bxt-j4205     total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:499s
fi-byt-j1900     total:289  pass:253  dwarn:1   dfail:0   fail:0   skip:35  time:493s
fi-byt-n2820     total:289  pass:249  dwarn:1   dfail:0   fail:0   skip:39  time:489s
fi-cfl-s         total:289  pass:253  dwarn:4   dfail:0   fail:0   skip:32  time:561s
fi-elk-e7500     total:289  pass:229  dwarn:0   dfail:0   fail:0   skip:60  time:427s
fi-gdg-551       total:289  pass:178  dwarn:1   dfail:0   fail:1   skip:109 time:250s
fi-glk-1         total:289  pass:261  dwarn:0   dfail:0   fail:0   skip:28  time:585s
fi-hsw-4770      total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:452s
fi-hsw-4770r     total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:427s
fi-ilk-650       total:289  pass:228  dwarn:0   dfail:0   fail:0   skip:61  time:438s
fi-ivb-3520m     total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:494s
fi-ivb-3770      total:289  pass:260  dwarn:0   dfail:0   fail:0   skip:29  time:460s
fi-kbl-7500u     total:289  pass:264  dwarn:1   dfail:0   fail:0   skip:24  time:493s
fi-kbl-7560u     total:289  pass:270  dwarn:0   dfail:0   fail:0   skip:19  time:571s
fi-kbl-7567u     total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:474s
fi-kbl-r         total:289  pass:262  dwarn:0   dfail:0   fail:0   skip:27  time:584s
fi-pnv-d510      total:289  pass:222  dwarn:1   dfail:0   fail:0   skip:66  time:548s
fi-skl-6260u     total:289  pass:269  dwarn:0   dfail:0   fail:0   skip:20  time:454s
fi-skl-6700hq    total:289  pass:263  dwarn:0   dfail:0   fail:0   skip:26  time:648s
fi-skl-6700k     total:289  pass:265  dwarn:0   dfail:0   fail:0   skip:24  time:522s
fi-skl-6770hq    total:289  pass:268  dwarn:1   dfail:0   fail:0   skip:20  time:504s
fi-skl-gvtdvm    total:289  pass:266  dwarn:0   dfail:0   fail:0   skip:23  time:458s
fi-snb-2520m     total:289  pass:250  dwarn:0   dfail:0   fail:0   skip:39  time:568s
fi-snb-2600      total:289  pass:249  dwarn:0   dfail:0   fail:0   skip:40  time:420s

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_386/
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#7 lore raw
In reply to Daniel Vetter
Quoted 21 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]>
References: https://intel-gfx-ci.01.org/tree/drm-tip/igt@[email protected]
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

---
  tests/gem_eio.c | 3 +++
  1 file changed, 3 insertions(+)
tests/gem_eio.c +0 -0
index 899cb62728e3..28375e208232 100644

tests/gem_eio.c static void test_inflight_suspend(int fd) · line 223

Quoted 6 lines
tests/gem_eio.c +3 -0
@@ -218,6 +218,9 @@ static void test_inflight_suspend(int fd)
  	igt_require(gem_has_exec_fence(fd));
  	igt_require(i915_reset_control(false));
 
+	igt_skip_on_f(IS_SANDYBRIDGE(intel_get_drm_devid(fd)),
+		      "random incompletes in CI with this test\n");
+

So, let's recap the problem here:
- Any incomplete in sharded runs mean that the platform is unfit for
pre-merge (because any other test after will go from pass to notrun)
- We can't fix issues immediately, especially for old platforms

This patch is sweeping the test under the rug by using the skip output, which is not only hard to track, it is also misleading.

After discussing with Marta, Arek and Petri, we found some consensus on the following proposal (terminology is up for debate):

- Introduce igt_dodge_on(cond, label): Report a pre-emptive 'fail' when the condition is true. Make sure this is over-ridable with IGT_DODGE=0 so as we can easily run these tests without recompiling them.

- Introduce a new piglit result (dodged), so as we can more easily keep track of the issue (no need to open the piglit results).

Any thoughts?

tests/gem_eio.c static void test_inflight_suspend(int fd) · line 226

Quoted 4 lines
tests/gem_eio.c +0 -0
  	memset(obj, 0, sizeof(obj));
  	obj[0].flags = EXEC_OBJECT_WRITE;
  	obj[1].handle = gem_create(fd, 4096);

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

#8 lore raw

+ Petri

Quoted 13 lines · Martin Peres
Quoted 12 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]>
References: https://intel-gfx-ci.01.org/tree/drm-tip/igt@[email protected]
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

<SNIP>

Quoted 8 lines

So, let's recap the problem here:
- Any incomplete in sharded runs mean that the platform is unfit for
pre-merge (because any other test after will go from pass to notrun)
- We can't fix issues immediately, especially for old platforms

This patch is sweeping the test under the rug by using the skip output, which is not only hard to track, it is also misleading.

Quoted 6 lines

After discussing with Marta, Arek and Petri, we found some consensus on the following proposal (terminology is up for debate):

- Introduce igt_dodge_on(cond, label): Report a pre-emptive 'fail' when the condition is true. Make sure this is over-ridable with IGT_DODGE=0 so as we can easily run these tests without recompiling them.

Make this igt_skip_on_ci(cond) and require IGT_CI=1 to activate them. Much like with simulation.

Still, a BIOS update to one of the CI machines might mean (if it's not now the case, not very far fetched for the future) that we go churn in the IGT codebase to drop bunch of these. That's not the optimal workflow I can think of when we're discussing a separate mailing list for IGT discussion and patches to make it more self-contained. Then we bind that new mailing list to our CI farm contents, and bind making fixes to the CI farm operation directly to the IGT reviewing bandwidth?

I'm still thinking best way would be that CI would mask the known problematic ones from the failure/pass criteria, and then somebody actually looks at the masked on after their testing coverage is prioritized. I think IGT should try to provide a wide range of tests that are supposed to work on any certain hardware. If they don't, it's not a reason to change the tests itself.

With the filter, we can grow the testing coverage for the new platforms, even if CI happens to have odd machines that may not pass those tests (and we may not have the resources to immediately fix those). All this without churning on the IGT codebase.

But if this is the only technically viable solution in short-term, then so be it. I just see better options too.

Regards, Joonas

Signature · 7 lines
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#9 lore raw
Quoted 30 lines · Joonas Lahtinen

+ Petri

Quoted 13 lines · Martin Peres
Quoted 12 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]>
References: https://intel-gfx-ci.01.org/tree/drm-tip/igt@[email protected]
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

<SNIP>

Quoted 11 lines

So, let's recap the problem here:
- Any incomplete in sharded runs mean that the platform is unfit for
pre-merge (because any other test after will go from pass to notrun)
- We can't fix issues immediately, especially for old platforms

This patch is sweeping the test under the rug by using the skip output, which is not only hard to track, it is also misleading.

After discussing with Marta, Arek and Petri, we found some consensus on the following proposal (terminology is up for debate):

Quoted 6 lines
Quoted 3 lines

- Introduce igt_dodge_on(cond, label): Report a pre-emptive 'fail' when the condition is true. Make sure this is over-ridable with IGT_DODGE=0 so as we can easily run these tests without recompiling them.

Make this igt_skip_on_ci(cond) and require IGT_CI=1 to activate them. Much like with simulation.

replace skip with fail, and we agree. Skips are too easy to ignore!

Quoted 8 lines

Still, a BIOS update to one of the CI machines might mean (if it's not now the case, not very far fetched for the future) that we go churn in the IGT codebase to drop bunch of these. That's not the optimal workflow I can think of when we're discussing a separate mailing list for IGT discussion and patches to make it more self-contained. Then we bind that new mailing list to our CI farm contents, and bind making fixes to the CI farm operation directly to the IGT reviewing bandwidth?

Isn't what we are proposing doing exactly this? By changing the source code of IGT, we allow people to send patches to remove some workarounds and see if they pass or fail in the same way they would propose any change to IGT. Moreover, we make the running of IGT in our farm as transparent as possible.

Quoted 7 lines

I'm still thinking best way would be that CI would mask the known problematic ones from the failure/pass criteria, and then somebody actually looks at the masked on after their testing coverage is prioritized. I think IGT should try to provide a wide range of tests that are supposed to work on any certain hardware. If they don't, it's not a reason to change the tests itself.

This is true that some skips will be highly-machine specific, but isn't our role as developers to know what and what doesn't work? By pushing this to an external whitelist only for CI, we miss an opportunity to improve this CI blacklist.

Now, let me remind you that this blacklist is *only* for tests that hang the machine or leave it in an inconsistant state which will lead to a crash later.

Quoted 5 lines

With the filter, we can grow the testing coverage for the new platforms, even if CI happens to have odd machines that may not pass those tests (and we may not have the resources to immediately fix those). All this without churning on the IGT codebase.

You are describing what cibuglog is already doing. Failing tests cases are suppressed in pre-merge, and associated bugs[1].

See above for what this proposal is about.

[1] https://intel-gfx-ci.01.org/cibuglog/

Quoted 3 lines

But if this is the only technically viable solution in short-term, then so be it. I just see better options too.

Maybe we need a write up our workflow. This time, a public one!

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

#10 lore raw

Since the discussion on this died and I believe that everyone are scared that the dodge suggestion would require someone to do some work. I could Ack the patch if:

+ igt_skip_on_f(IS_SANDYBRIDGE(intel_get_drm_devid(fd)),
+ "random incompletes in CI with this test\n");
+

Was replaced with an igt_warn

/Marta

Quoted 98 lines

-----Original Message-----
From: Martin Peres [mailto:[email protected]]
Sent: Monday, October 23, 2017 12:29 PM
To: Joonas Lahtinen <[email protected]>; Daniel Vetter:
<[email protected]>; Intel Graphics Development <intel- [email protected]>; Latvala, Petri <[email protected]>
Cc: Chris Wilson <[email protected]>; Lofstedt, Marta <[email protected]>
Subject: Re: [PATCH i-g-t] tests/gem_eio: Skip in-flight-suspend on snb

Quoted 19 lines · Joonas Lahtinen

+ Petri

Quoted 6 lines · Martin Peres
Quoted 5 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Quoted 9 lines
Quoted 9 lines

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]> References:
https://intel-gfx-ci.01.org/tree/drm-tip/igt@gem_eio@in-flight-suspe
nd.html
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

Quoted 16 lines

<SNIP>

Quoted 14 lines

So, let's recap the problem here:
- Any incomplete in sharded runs mean that the platform is unfit
for pre-merge (because any other test after will go from pass to notrun)
- We can't fix issues immediately, especially for old platforms

This patch is sweeping the test under the rug by using the skip output, which is not only hard to track, it is also misleading.

After discussing with Marta, Arek and Petri, we found some consensus on the following proposal (terminology is up for debate):

- Introduce igt_dodge_on(cond, label): Report a pre-emptive 'fail' when the condition is true. Make sure this is over-ridable with IGT_DODGE=0 so as we can easily run these tests without recompiling

them.

Quoted 3 lines

Make this igt_skip_on_ci(cond) and require IGT_CI=1 to activate them. Much like with simulation.

replace skip with fail, and we agree. Skips are too easy to ignore!

Quoted 8 lines

Still, a BIOS update to one of the CI machines might mean (if it's not now the case, not very far fetched for the future) that we go churn in the IGT codebase to drop bunch of these. That's not the optimal workflow I can think of when we're discussing a separate mailing list for IGT discussion and patches to make it more self-contained. Then we bind that new mailing list to our CI farm contents, and bind making fixes to the CI farm operation directly to the IGT reviewing bandwidth?

Isn't what we are proposing doing exactly this? By changing the source code of IGT, we allow people to send patches to remove some workarounds and see if they pass or fail in the same way they would propose any change to IGT. Moreover, we make the running of IGT in our farm as transparent as possible.

Quoted 7 lines

I'm still thinking best way would be that CI would mask the known problematic ones from the failure/pass criteria, and then somebody actually looks at the masked on after their testing coverage is prioritized. I think IGT should try to provide a wide range of tests that are supposed to work on any certain hardware. If they don't, it's not a reason to change the tests itself.

This is true that some skips will be highly-machine specific, but isn't our role as developers to know what and what doesn't work? By pushing this to an external whitelist only for CI, we miss an opportunity to improve this CI blacklist.

Now, let me remind you that this blacklist is *only* for tests that hang the machine or leave it in an inconsistant state which will lead to a crash later.

Quoted 5 lines

With the filter, we can grow the testing coverage for the new platforms, even if CI happens to have odd machines that may not pass those tests (and we may not have the resources to immediately fix those). All this without churning on the IGT codebase.

You are describing what cibuglog is already doing. Failing tests cases are suppressed in pre-merge, and associated bugs[1].

See above for what this proposal is about.

[1] https://intel-gfx-ci.01.org/cibuglog/

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

But if this is the only technically viable solution in short-term, then so be it. I just see better options too.

Maybe we need a write up our workflow. This time, a public one!

I hope

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

#11 lore raw
In reply to Martin Peres
Quoted 20 lines

-----Original Message-----
From: Lofstedt, Marta
Sent: Thursday, October 26, 2017 1:57 PM
To: 'Martin Peres' <[email protected]>; Joonas Lahtinen <[email protected]>; Daniel Vetter <[email protected]>; Intel Graphics Development <[email protected]>; Latvala, Petri <[email protected]>
Cc: Chris Wilson <[email protected]>
Subject: RE: [PATCH i-g-t] tests/gem_eio: Skip in-flight-suspend on snb

Since the discussion on this died and I believe that everyone are scared that the dodge suggestion would require someone to do some work.
I could Ack the patch if:

tests/gem_eio.c +3 -0
+	igt_skip_on_f(IS_SANDYBRIDGE(intel_get_drm_devid(fd)),
+		      "random incompletes in CI with this test\n");
+
Quoted 2 lines

Was replaced with an igt_warn

Forgot to write, it should be igt_warn paired with success on the test. This will produce the WARN result, note this is not the same as dmesg-warn. This is quite rare and it will be noticed.

Quoted 81 lines

/Marta

Quoted 48 lines

-----Original Message-----
From: Martin Peres [mailto:[email protected]]
Sent: Monday, October 23, 2017 12:29 PM
To: Joonas Lahtinen <[email protected]>; Daniel Vetter:
<[email protected]>; Intel Graphics Development <intel- [email protected]>; Latvala, Petri <[email protected]>
Cc: Chris Wilson <[email protected]>; Lofstedt, Marta <[email protected]>
Subject: Re: [PATCH i-g-t] tests/gem_eio: Skip in-flight-suspend on snb

Quoted 20 lines · Joonas Lahtinen

+ Petri

Quoted 6 lines · Martin Peres
Quoted 5 lines · Daniel Vetter

CI gets upset about it resulting in an incomplete, let's skip it until that's fixed to avoid havoc in the CI farm. Of course this should/will be reverted as soon as we have a fix (similar to how we dealt with the snb-dies-in-blt-hangs issue).

Quoted 10 lines
Quoted 10 lines

Cc: Joonas Lahtinen <[email protected]>
Cc: Chris Wilson <[email protected]>
Cc: "Lofstedt, Marta" <[email protected]>
Cc: Martin Peres <[email protected]> References:
https://intel-gfx-ci.01.org/tree/drm-tip/igt@gem_eio@in-flight-sus
pe
nd.html
References: https://bugs.freedesktop.org/show_bug.cgi?id=103289

Quoted 16 lines

<SNIP>

Quoted 14 lines

So, let's recap the problem here:
- Any incomplete in sharded runs mean that the platform is unfit
for pre-merge (because any other test after will go from pass to notrun)
- We can't fix issues immediately, especially for old platforms

This patch is sweeping the test under the rug by using the skip output, which is not only hard to track, it is also misleading.

After discussing with Marta, Arek and Petri, we found some consensus on the following proposal (terminology is up for debate):

- Introduce igt_dodge_on(cond, label): Report a pre-emptive 'fail' when the condition is true. Make sure this is over-ridable with IGT_DODGE=0 so as we can easily run these tests without recompiling

Quoted 15 lines

them.

Quoted 3 lines

Make this igt_skip_on_ci(cond) and require IGT_CI=1 to activate them. Much like with simulation.

replace skip with fail, and we agree. Skips are too easy to ignore!

Quoted 8 lines

Still, a BIOS update to one of the CI machines might mean (if it's not now the case, not very far fetched for the future) that we go churn in the IGT codebase to drop bunch of these. That's not the optimal workflow I can think of when we're discussing a separate mailing list for IGT discussion and patches to make it more self-contained. Then we bind that new mailing list to our CI farm contents, and bind making fixes to the CI farm operation directly to the

IGT reviewing bandwidth?

Quoted 15 lines

Isn't what we are proposing doing exactly this? By changing the source code of IGT, we allow people to send patches to remove some workarounds and see if they pass or fail in the same way they would propose any change to IGT. Moreover, we make the running of IGT in our farm as transparent as possible.

Quoted 7 lines

I'm still thinking best way would be that CI would mask the known problematic ones from the failure/pass criteria, and then somebody actually looks at the masked on after their testing coverage is prioritized. I think IGT should try to provide a wide range of tests that are supposed to work on any certain hardware. If they don't, it's not a reason to change the tests itself.

Quoted 29 lines
Quoted 7 lines

This is true that some skips will be highly-machine specific, but isn't our role as developers to know what and what doesn't work? By pushing this to an external whitelist only for CI, we miss an opportunity to improve this CI blacklist.

Now, let me remind you that this blacklist is *only* for tests that hang the machine or leave it in an inconsistant state which will lead to a

crash later.

Quoted 21 lines
Quoted 6 lines

With the filter, we can grow the testing coverage for the new platforms, even if CI happens to have odd machines that may not pass those tests (and we may not have the resources to immediately fix those). All this without churning on the IGT codebase.

You are describing what cibuglog is already doing. Failing tests cases are suppressed in pre-merge, and associated bugs[1].

See above for what this proposal is about.

[1] https://intel-gfx-ci.01.org/cibuglog/

Quoted 3 lines

But if this is the only technically viable solution in short-term, then so be it. I just see better options too.

Maybe we need a write up our workflow. This time, a public one!

I hope

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