* [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang()
@ 2019-10-26 19:34 Chris Wilson
2019-10-26 19:34 ` [Intel-gfx] " Chris Wilson
` (2 more replies)
0 siblings, 3 replies; 14+ messages in thread
From: Chris Wilson @ 2019-10-26 19:34 UTC (permalink / raw)
To: intel-gfx; +Cc: igt-dev
igt_allow_hang() checks that the GPU can be reset before allowing the
test to cause a GPU hang (which would need the reset to recover).
However, our checking for allowing a hang depends on i915.reset which we
later restore. Do that restoration before the check so that this test is
not affected by earlier fails.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
lib/igt_gt.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/lib/igt_gt.c b/lib/igt_gt.c
index 5ca77471c..256c7cbc0 100644
--- a/lib/igt_gt.c
+++ b/lib/igt_gt.c
@@ -175,8 +175,6 @@ igt_hang_t igt_allow_hang(int fd, unsigned ctx, unsigned flags)
if (!igt_check_boolean_env_var("IGT_HANG", true))
igt_skip("hang injection disabled by user");
gem_context_require_bannable(fd);
- if (!igt_check_boolean_env_var("IGT_HANG_WITHOUT_RESET", false))
- igt_require(has_gpu_reset(fd));
allow_reset = 1;
if ((flags & HANG_ALLOW_CAPTURE) == 0) {
@@ -191,6 +189,9 @@ igt_hang_t igt_allow_hang(int fd, unsigned ctx, unsigned flags)
}
igt_require(igt_sysfs_set_parameter(fd, "reset", "%d", allow_reset));
+ if (!igt_check_boolean_env_var("IGT_HANG_WITHOUT_RESET", false))
+ igt_require(has_gpu_reset(fd));
+
ban = context_get_ban(fd, ctx);
if ((flags & HANG_ALLOW_BAN) == 0)
context_set_ban(fd, ctx, 0);
--
2.24.0.rc1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Intel-gfx] [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang()
2019-10-26 19:34 [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang() Chris Wilson
@ 2019-10-26 19:34 ` Chris Wilson
2019-10-26 19:34 ` [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck Chris Wilson
2019-10-27 19:01 ` [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang() Andi Shyti
2 siblings, 0 replies; 14+ messages in thread
From: Chris Wilson @ 2019-10-26 19:34 UTC (permalink / raw)
To: intel-gfx; +Cc: igt-dev
igt_allow_hang() checks that the GPU can be reset before allowing the
test to cause a GPU hang (which would need the reset to recover).
However, our checking for allowing a hang depends on i915.reset which we
later restore. Do that restoration before the check so that this test is
not affected by earlier fails.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
lib/igt_gt.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/lib/igt_gt.c b/lib/igt_gt.c
index 5ca77471c..256c7cbc0 100644
--- a/lib/igt_gt.c
+++ b/lib/igt_gt.c
@@ -175,8 +175,6 @@ igt_hang_t igt_allow_hang(int fd, unsigned ctx, unsigned flags)
if (!igt_check_boolean_env_var("IGT_HANG", true))
igt_skip("hang injection disabled by user");
gem_context_require_bannable(fd);
- if (!igt_check_boolean_env_var("IGT_HANG_WITHOUT_RESET", false))
- igt_require(has_gpu_reset(fd));
allow_reset = 1;
if ((flags & HANG_ALLOW_CAPTURE) == 0) {
@@ -191,6 +189,9 @@ igt_hang_t igt_allow_hang(int fd, unsigned ctx, unsigned flags)
}
igt_require(igt_sysfs_set_parameter(fd, "reset", "%d", allow_reset));
+ if (!igt_check_boolean_env_var("IGT_HANG_WITHOUT_RESET", false))
+ igt_require(has_gpu_reset(fd));
+
ban = context_get_ban(fd, ctx);
if ((flags & HANG_ALLOW_BAN) == 0)
context_set_ban(fd, ctx, 0);
--
2.24.0.rc1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread* [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-26 19:34 [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang() Chris Wilson
2019-10-26 19:34 ` [Intel-gfx] " Chris Wilson
@ 2019-10-26 19:34 ` Chris Wilson
2019-10-26 19:34 ` [Intel-gfx] " Chris Wilson
2019-10-27 19:01 ` Andi Shyti
2019-10-27 19:01 ` [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang() Andi Shyti
2 siblings, 2 replies; 14+ messages in thread
From: Chris Wilson @ 2019-10-26 19:34 UTC (permalink / raw)
To: intel-gfx; +Cc: igt-dev
Wrap the test inside igt_allow_hang as we depend upon a fast reset to
cancel the hostile context.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
tests/i915/gem_ctx_exec.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/tests/i915/gem_ctx_exec.c b/tests/i915/gem_ctx_exec.c
index b552edfbd..6b0092250 100644
--- a/tests/i915/gem_ctx_exec.c
+++ b/tests/i915/gem_ctx_exec.c
@@ -207,6 +207,7 @@ static void nohangcheck_hostile(int i915)
{
int64_t timeout = NSEC_PER_SEC / 2;
igt_spin_t *spin;
+ igt_hang_t hang;
uint32_t ctx;
int err = 0;
int dir;
@@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
* we forcibly terminate that context.
*/
+ hang = igt_allow_hang(i915, 0, 0);
+
dir = igt_sysfs_open_parameters(i915);
igt_require(dir != -1);
@@ -235,6 +238,7 @@ static void nohangcheck_hostile(int i915)
igt_require(__enable_hangcheck(dir, true));
gem_quiescent_gpu(i915);
+ igt_disallow_hang(i915, hang);
igt_assert_f(err == 0,
"Hostile unpreemptable context was not cancelled immediately upon closure\n");
--
2.24.0.rc1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread* [Intel-gfx] [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-26 19:34 ` [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck Chris Wilson
@ 2019-10-26 19:34 ` Chris Wilson
2019-10-27 19:01 ` Andi Shyti
1 sibling, 0 replies; 14+ messages in thread
From: Chris Wilson @ 2019-10-26 19:34 UTC (permalink / raw)
To: intel-gfx; +Cc: igt-dev
Wrap the test inside igt_allow_hang as we depend upon a fast reset to
cancel the hostile context.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
tests/i915/gem_ctx_exec.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/tests/i915/gem_ctx_exec.c b/tests/i915/gem_ctx_exec.c
index b552edfbd..6b0092250 100644
--- a/tests/i915/gem_ctx_exec.c
+++ b/tests/i915/gem_ctx_exec.c
@@ -207,6 +207,7 @@ static void nohangcheck_hostile(int i915)
{
int64_t timeout = NSEC_PER_SEC / 2;
igt_spin_t *spin;
+ igt_hang_t hang;
uint32_t ctx;
int err = 0;
int dir;
@@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
* we forcibly terminate that context.
*/
+ hang = igt_allow_hang(i915, 0, 0);
+
dir = igt_sysfs_open_parameters(i915);
igt_require(dir != -1);
@@ -235,6 +238,7 @@ static void nohangcheck_hostile(int i915)
igt_require(__enable_hangcheck(dir, true));
gem_quiescent_gpu(i915);
+ igt_disallow_hang(i915, hang);
igt_assert_f(err == 0,
"Hostile unpreemptable context was not cancelled immediately upon closure\n");
--
2.24.0.rc1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 14+ messages in thread* Re: [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-26 19:34 ` [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck Chris Wilson
2019-10-26 19:34 ` [Intel-gfx] " Chris Wilson
@ 2019-10-27 19:01 ` Andi Shyti
2019-10-27 19:01 ` [Intel-gfx] " Andi Shyti
` (2 more replies)
1 sibling, 3 replies; 14+ messages in thread
From: Andi Shyti @ 2019-10-27 19:01 UTC (permalink / raw)
To: Chris Wilson; +Cc: igt-dev, intel-gfx
Hi Chris,
> @@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
> * we forcibly terminate that context.
> */
>
> + hang = igt_allow_hang(i915, 0, 0);
> +
why ctx = 0?
Andi
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [Intel-gfx] [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-27 19:01 ` Andi Shyti
@ 2019-10-27 19:01 ` Andi Shyti
2019-10-27 19:05 ` Chris Wilson
2019-10-27 19:06 ` Chris Wilson
2 siblings, 0 replies; 14+ messages in thread
From: Andi Shyti @ 2019-10-27 19:01 UTC (permalink / raw)
To: Chris Wilson; +Cc: igt-dev, intel-gfx
Hi Chris,
> @@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
> * we forcibly terminate that context.
> */
>
> + hang = igt_allow_hang(i915, 0, 0);
> +
why ctx = 0?
Andi
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-27 19:01 ` Andi Shyti
2019-10-27 19:01 ` [Intel-gfx] " Andi Shyti
@ 2019-10-27 19:05 ` Chris Wilson
2019-10-27 19:05 ` [Intel-gfx] " Chris Wilson
2019-10-27 19:33 ` Andi Shyti
2019-10-27 19:06 ` Chris Wilson
2 siblings, 2 replies; 14+ messages in thread
From: Chris Wilson @ 2019-10-27 19:05 UTC (permalink / raw)
To: Andi Shyti; +Cc: igt-dev, intel-gfx
Quoting Andi Shyti (2019-10-27 19:01:24)
> Hi Chris,
>
> > @@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
> > * we forcibly terminate that context.
> > */
> >
> > + hang = igt_allow_hang(i915, 0, 0);
> > +
>
> why ctx = 0?
Because I'm lazy and it didn't matter ;)
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [Intel-gfx] [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-27 19:05 ` Chris Wilson
@ 2019-10-27 19:05 ` Chris Wilson
2019-10-27 19:33 ` Andi Shyti
1 sibling, 0 replies; 14+ messages in thread
From: Chris Wilson @ 2019-10-27 19:05 UTC (permalink / raw)
To: Andi Shyti; +Cc: igt-dev, intel-gfx
Quoting Andi Shyti (2019-10-27 19:01:24)
> Hi Chris,
>
> > @@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
> > * we forcibly terminate that context.
> > */
> >
> > + hang = igt_allow_hang(i915, 0, 0);
> > +
>
> why ctx = 0?
Because I'm lazy and it didn't matter ;)
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-27 19:05 ` Chris Wilson
2019-10-27 19:05 ` [Intel-gfx] " Chris Wilson
@ 2019-10-27 19:33 ` Andi Shyti
2019-10-27 19:33 ` [Intel-gfx] " Andi Shyti
1 sibling, 1 reply; 14+ messages in thread
From: Andi Shyti @ 2019-10-27 19:33 UTC (permalink / raw)
To: Chris Wilson; +Cc: igt-dev, intel-gfx
> > > @@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
> > > * we forcibly terminate that context.
> > > */
> > >
> > > + hang = igt_allow_hang(i915, 0, 0);
> > > +
> >
> > why ctx = 0?
>
> Because I'm lazy and it didn't matter ;)
yes, sure :)
Acked-by: Andi Shyti <andi.shyti@intel.com>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-27 19:01 ` Andi Shyti
2019-10-27 19:01 ` [Intel-gfx] " Andi Shyti
2019-10-27 19:05 ` Chris Wilson
@ 2019-10-27 19:06 ` Chris Wilson
2019-10-27 19:06 ` [Intel-gfx] " Chris Wilson
2 siblings, 1 reply; 14+ messages in thread
From: Chris Wilson @ 2019-10-27 19:06 UTC (permalink / raw)
To: Andi Shyti; +Cc: igt-dev, intel-gfx
Quoting Andi Shyti (2019-10-27 19:01:24)
> Hi Chris,
>
> > @@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
> > * we forcibly terminate that context.
> > */
> >
> > + hang = igt_allow_hang(i915, 0, 0);
> > +
>
> why ctx = 0?
Imagine,
* we forcibly terminate that context.
*/
- hang = igt_allow_hang(i915, 0, 0);
-
dir = igt_sysfs_open_parameters(i915);
igt_require(dir != -1);
ctx = gem_context_create(i915);
+ hang = igt_allow_hang(i915, ctx, 0);
igt_require(__enable_hangcheck(dir, false));
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread* Re: [Intel-gfx] [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck
2019-10-27 19:06 ` Chris Wilson
@ 2019-10-27 19:06 ` Chris Wilson
0 siblings, 0 replies; 14+ messages in thread
From: Chris Wilson @ 2019-10-27 19:06 UTC (permalink / raw)
To: Andi Shyti; +Cc: igt-dev, intel-gfx
Quoting Andi Shyti (2019-10-27 19:01:24)
> Hi Chris,
>
> > @@ -216,6 +217,8 @@ static void nohangcheck_hostile(int i915)
> > * we forcibly terminate that context.
> > */
> >
> > + hang = igt_allow_hang(i915, 0, 0);
> > +
>
> why ctx = 0?
Imagine,
* we forcibly terminate that context.
*/
- hang = igt_allow_hang(i915, 0, 0);
-
dir = igt_sysfs_open_parameters(i915);
igt_require(dir != -1);
ctx = gem_context_create(i915);
+ hang = igt_allow_hang(i915, ctx, 0);
igt_require(__enable_hangcheck(dir, false));
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
* Re: [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang()
2019-10-26 19:34 [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang() Chris Wilson
2019-10-26 19:34 ` [Intel-gfx] " Chris Wilson
2019-10-26 19:34 ` [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck Chris Wilson
@ 2019-10-27 19:01 ` Andi Shyti
2019-10-27 19:01 ` [Intel-gfx] " Andi Shyti
2 siblings, 1 reply; 14+ messages in thread
From: Andi Shyti @ 2019-10-27 19:01 UTC (permalink / raw)
To: Chris Wilson; +Cc: igt-dev, intel-gfx
Hi Chris,
On Sat, Oct 26, 2019 at 08:34:55PM +0100, Chris Wilson wrote:
> igt_allow_hang() checks that the GPU can be reset before allowing the
> test to cause a GPU hang (which would need the reset to recover).
> However, our checking for allowing a hang depends on i915.reset which we
> later restore. Do that restoration before the check so that this test is
> not affected by earlier fails.
>
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Andi Shyti <andi.shyti@intel.com>
Andi
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 14+ messages in thread
end of thread, other threads:[~2019-10-27 19:33 UTC | newest]
Thread overview: 14+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-10-26 19:34 [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang() Chris Wilson
2019-10-26 19:34 ` [Intel-gfx] " Chris Wilson
2019-10-26 19:34 ` [PATCH i-g-t 2/2] i915/gem_ctx_exec: Enable resets for basic-nohangcheck Chris Wilson
2019-10-26 19:34 ` [Intel-gfx] " Chris Wilson
2019-10-27 19:01 ` Andi Shyti
2019-10-27 19:01 ` [Intel-gfx] " Andi Shyti
2019-10-27 19:05 ` Chris Wilson
2019-10-27 19:05 ` [Intel-gfx] " Chris Wilson
2019-10-27 19:33 ` Andi Shyti
2019-10-27 19:33 ` [Intel-gfx] " Andi Shyti
2019-10-27 19:06 ` Chris Wilson
2019-10-27 19:06 ` [Intel-gfx] " Chris Wilson
2019-10-27 19:01 ` [PATCH i-g-t 1/2] lib: Restore i915.reset before testing it in igt_allow_hang() Andi Shyti
2019-10-27 19:01 ` [Intel-gfx] " Andi Shyti
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox