Re: [Intel-gfx] [PATCH i-g-t v2 2/2] tests/gem_exec_schedule: Add test for resetting preemptive batch

2017-12-08 Thread Chris Wilson
Quoting Antonio Argenziano (2017-12-04 23:45:41)
> This patch adds a test that will trigger a preemption of a low priority
> batch by a 'bad' batch buffer which will hang. The test aims at making
> sure that a hanging high priority batch will not disrupt the submission
> flow of low priority contexts.
> 
> -v2:
> - Rename subtest (Chris)
> - Use igt_hang_ctx to hang ring (Chris)
> - Add comment on execution order checks (Chris)
> 
> Cc: Chris Wilson 
> Cc: Michal Winiarski 
> Signed-off-by: Antonio Argenziano 
> ---
> +static void preemptive_hang(int fd, unsigned ring)
> +{
> +   igt_spin_t *spin[16];

Someone is going to shout one day, what is it with the magic number 16?

#define MAX_ELSP_QLEN 16 ? And should be set to some safe number times
the forseeable future.

> +   igt_hang_t hang;
> +   uint32_t ctx[2];
> +
> +   ctx[HI] = gem_context_create(fd);
> +   gem_context_set_priority(fd, ctx[HI], MAX_PRIO);
> +
> +   for (int n = 0; n < 16; n++) {
> +   ctx[LO] = gem_context_create(fd);
> +   gem_context_set_priority(fd, ctx[LO], MIN_PRIO);
> +
> +   spin[n] = __igt_spin_batch_new(fd, ctx[LO], ring, 0);
> +   igt_debug("spin[%d].handle=%d\n", n, spin[n]->handle);
> +
> +   gem_context_destroy(fd, ctx[LO]);
> +   }
> +
> +   hang = igt_hang_ctx(fd, ctx[HI], ring, 0, NULL);
> +   gem_wait(fd, hang.handle, NULL);

That should be igt_post_hang_ring(fd, hang);
-Chris
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx


Re: [Intel-gfx] [PATCH i-g-t v2 2/2] tests/gem_exec_schedule: Add test for resetting preemptive batch

2017-12-08 Thread Michał Winiarski
On Mon, Dec 04, 2017 at 03:45:41PM -0800, Antonio Argenziano wrote:
> This patch adds a test that will trigger a preemption of a low priority
> batch by a 'bad' batch buffer which will hang. The test aims at making
> sure that a hanging high priority batch will not disrupt the submission
> flow of low priority contexts.
> 
> -v2:
>   - Rename subtest (Chris)
>   - Use igt_hang_ctx to hang ring (Chris)
>   - Add comment on execution order checks (Chris)
> 
> Cc: Chris Wilson 
> Cc: Michal Winiarski 
> Signed-off-by: Antonio Argenziano 

Reviewed-by: Michał Winiarski 

> ---
>  tests/gem_exec_schedule.c | 37 +
>  1 file changed, 37 insertions(+)
___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx


[Intel-gfx] [PATCH i-g-t v2 2/2] tests/gem_exec_schedule: Add test for resetting preemptive batch

2017-12-04 Thread Antonio Argenziano
This patch adds a test that will trigger a preemption of a low priority
batch by a 'bad' batch buffer which will hang. The test aims at making
sure that a hanging high priority batch will not disrupt the submission
flow of low priority contexts.

-v2:
- Rename subtest (Chris)
- Use igt_hang_ctx to hang ring (Chris)
- Add comment on execution order checks (Chris)

Cc: Chris Wilson 
Cc: Michal Winiarski 
Signed-off-by: Antonio Argenziano 
---
 tests/gem_exec_schedule.c | 37 +
 1 file changed, 37 insertions(+)

diff --git a/tests/gem_exec_schedule.c b/tests/gem_exec_schedule.c
index d0f85b3e..0c2efaa2 100644
--- a/tests/gem_exec_schedule.c
+++ b/tests/gem_exec_schedule.c
@@ -515,6 +515,40 @@ static void preempt_self(int fd, unsigned ring)
gem_close(fd, result);
 }
 
+static void preemptive_hang(int fd, unsigned ring)
+{
+   igt_spin_t *spin[16];
+   igt_hang_t hang;
+   uint32_t ctx[2];
+
+   ctx[HI] = gem_context_create(fd);
+   gem_context_set_priority(fd, ctx[HI], MAX_PRIO);
+
+   for (int n = 0; n < 16; n++) {
+   ctx[LO] = gem_context_create(fd);
+   gem_context_set_priority(fd, ctx[LO], MIN_PRIO);
+
+   spin[n] = __igt_spin_batch_new(fd, ctx[LO], ring, 0);
+   igt_debug("spin[%d].handle=%d\n", n, spin[n]->handle);
+
+   gem_context_destroy(fd, ctx[LO]);
+   }
+
+   hang = igt_hang_ctx(fd, ctx[HI], ring, 0, NULL);
+   gem_wait(fd, hang.handle, NULL);
+
+   for (int n = 0; n < 16; n++) {
+   /* Current behavior is to execute requests in order of 
submission.
+* This is subject to change as the scheduler evolve. The test 
should
+* be updated to reflect such changes.
+*/
+   igt_assert(gem_bo_busy(fd, spin[n]->handle));
+   igt_spin_batch_free(fd, spin[n]);
+   }
+
+   gem_context_destroy(fd, ctx[HI]);
+}
+
 static void deep(int fd, unsigned ring)
 {
 #define XS 8
@@ -1044,6 +1078,9 @@ igt_main
preempt(fd, e->exec_id 
| e->flags, NEW_CTX | HANG_LP);
}
 
+   
igt_subtest_f("preemptive-hang-%s", e->name)
+   preemptive_hang(fd, 
e->exec_id | e->flags);
+
igt_fixture {
igt_disallow_hang(fd, 
hang);

igt_fork_hang_detector(fd);
-- 
2.14.2

___
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx