From: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
To: igt-dev@lists.freedesktop.org
Cc: Intel-gfx@lists.freedesktop.org,
Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Subject: [igt-dev] [PATCH i-g-t 1/2] tests/gem_exec_await: Relax the busy spinner
Date: Mon, 19 Nov 2018 15:22:28 +0000 [thread overview]
Message-ID: <20181119152229.8390-1-tvrtko.ursulin@linux.intel.com> (raw)
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Add some nop instructions between recursive batch buffer start calls to
give system some breathing room. Without these, especially when coupled
with memory pressure, false GPU hangs can be observed caused by the
inability of the chip to cope.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106680
---
tests/i915/gem_exec_await.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/tests/i915/gem_exec_await.c b/tests/i915/gem_exec_await.c
index 5cfeb8ec8bfa..3ea5b5903c6b 100644
--- a/tests/i915/gem_exec_await.c
+++ b/tests/i915/gem_exec_await.c
@@ -63,6 +63,7 @@ static void xchg_obj(void *array, unsigned i, unsigned j)
static void wide(int fd, int ring_size, int timeout, unsigned int flags)
{
const uint32_t bbe = MI_BATCH_BUFFER_END;
+ const unsigned int num_nops = 1000;
const int gen = intel_gen(intel_get_drm_devid(fd));
struct {
struct drm_i915_gem_exec_object2 *obj;
@@ -123,7 +124,7 @@ static void wide(int fd, int ring_size, int timeout, unsigned int flags)
exec[e].execbuf.buffer_count = 2;
exec[e].reloc.target_handle = 1; /* recurse */
- exec[e].reloc.offset = sizeof(uint32_t);
+ exec[e].reloc.offset = (1 + num_nops) * sizeof(uint32_t);
exec[e].reloc.read_domains = I915_GEM_DOMAIN_COMMAND;
if (gen < 4)
exec[e].reloc.delta = 1;
@@ -162,7 +163,7 @@ static void wide(int fd, int ring_size, int timeout, unsigned int flags)
gem_set_domain(fd, exec[e].exec[1].handle,
I915_GEM_DOMAIN_WC, I915_GEM_DOMAIN_WC);
- i = 0;
+ i = num_nops;
exec[e].cmd[i] = MI_BATCH_BUFFER_START;
if (gen >= 8) {
exec[e].cmd[i] |= 1 << 8 | 1;
@@ -200,7 +201,7 @@ static void wide(int fd, int ring_size, int timeout, unsigned int flags)
count += nengine;
for (unsigned e = 0; e < nengine; e++)
- exec[e].cmd[0] = MI_BATCH_BUFFER_END;
+ exec[e].cmd[num_nops] = MI_BATCH_BUFFER_END;
__sync_synchronize();
}
--
2.19.1
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev
next reply other threads:[~2018-11-19 15:22 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-11-19 15:22 Tvrtko Ursulin [this message]
2018-11-19 15:22 ` [igt-dev] [PATCH i-g-t 2/2] tests/gem_exec_await: Add a memory pressure subtest Tvrtko Ursulin
2018-11-19 15:36 ` Chris Wilson
2018-11-19 15:54 ` Tvrtko Ursulin
2018-11-19 17:07 ` [Intel-gfx] " Chris Wilson
2018-11-19 15:28 ` [igt-dev] [PATCH i-g-t 1/2] tests/gem_exec_await: Relax the busy spinner Chris Wilson
2018-11-19 15:33 ` Tvrtko Ursulin
2018-11-19 16:18 ` Chris Wilson
2018-11-19 19:18 ` Tvrtko Ursulin
2018-11-19 19:34 ` Chris Wilson
2018-11-19 16:44 ` [igt-dev] ✓ Fi.CI.BAT: success for series starting with [i-g-t,1/2] " Patchwork
2018-11-19 21:52 ` [igt-dev] ✓ Fi.CI.IGT: " Patchwork
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20181119152229.8390-1-tvrtko.ursulin@linux.intel.com \
--to=tvrtko.ursulin@linux.intel.com \
--cc=Intel-gfx@lists.freedesktop.org \
--cc=igt-dev@lists.freedesktop.org \
--cc=tvrtko.ursulin@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox