* [PATCH i-g-t] tests/gem_cpu_reloc: Fix gem_cpu_reloc OOM failure
@ 2015-05-12 12:09 Derek Morton
2015-05-13 11:06 ` Gore, Tim
2015-05-13 16:41 ` Thomas Wood
0 siblings, 2 replies; 3+ messages in thread
From: Derek Morton @ 2015-05-12 12:09 UTC (permalink / raw)
To: intel-gfx; +Cc: thomas.wood
On android with small memory footprints gem_cpu_reloc can fail
due to OOM.
Refactor gem_cpu_reloc into 2 tests, a basic test which performs
10 relocations and a full test which skips if there is insufficient
memory.
Changed the memory required test to better estimate the actual
RAM used.
Signed-off-by: Derek Morton <derek.j.morton@intel.com>
---
tests/gem_cpu_reloc.c | 48 ++++++++++++++++++++++++++++++++++++------------
1 file changed, 36 insertions(+), 12 deletions(-)
diff --git a/tests/gem_cpu_reloc.c b/tests/gem_cpu_reloc.c
index 058089c..2f5fc60 100644
--- a/tests/gem_cpu_reloc.c
+++ b/tests/gem_cpu_reloc.c
@@ -159,16 +159,14 @@ uint32_t gen8_batch[] = {
uint32_t *batch = gen6_batch;
uint32_t batch_size = sizeof(gen6_batch);
-igt_simple_main
+void run_test(int fd, int count)
{
const uint32_t hang[] = {-1, -1, -1, -1};
const uint32_t end[] = {MI_BATCH_BUFFER_END, 0};
- uint64_t aper_size;
uint32_t noop;
uint32_t *handles;
- int fd, i, count;
+ int i;
- fd = drm_open_any();
noop = intel_get_drm_devid(fd);
use_blt = 0;
@@ -180,13 +178,6 @@ igt_simple_main
batch_size += 2 * 4;
}
- aper_size = gem_mappable_aperture_size();
- count = aper_size / 4096 * 2;
- if (igt_run_in_simulation())
- count = 10;
-
- intel_require_memory(1+count, 4096, CHECK_RAM);
-
handles = malloc (count * sizeof(uint32_t));
igt_assert(handles);
@@ -246,6 +237,39 @@ igt_simple_main
igt_progress("gem_cpu_reloc: ", 2*count+i, 3*count);
}
- igt_info("Test suceeded, cleanup up - this might take a while.\n");
+ igt_info("Subtest suceeded, cleanup up - this might take a while.\n");
+ for (i = 0; i < count; i++) {
+ gem_close(fd, handles[i]);
+ }
+ gem_close(fd, noop);
+ free(handles);
+}
+
+igt_main
+{
+ uint64_t aper_size;
+ int fd, count;
+
+ igt_fixture {
+ fd = drm_open_any();
+ }
+
+ igt_subtest("basic") {
+ run_test (fd, 10);
+ }
+
+ igt_skip_on_simulation();
+
+ igt_subtest("full") {
+ aper_size = gem_mappable_aperture_size();
+ count = aper_size / 4096 * 2;
+
+ /* count + 2 (noop & bad) buffers. A gem object appears to
+ require about 2kb + buffer + kernel overhead */
+ intel_require_memory(2+count, 2048+4096, CHECK_RAM);
+
+ run_test (fd, count);
+ }
+
close(fd);
}
--
1.9.1
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH i-g-t] tests/gem_cpu_reloc: Fix gem_cpu_reloc OOM failure
2015-05-12 12:09 [PATCH i-g-t] tests/gem_cpu_reloc: Fix gem_cpu_reloc OOM failure Derek Morton
@ 2015-05-13 11:06 ` Gore, Tim
2015-05-13 16:41 ` Thomas Wood
1 sibling, 0 replies; 3+ messages in thread
From: Gore, Tim @ 2015-05-13 11:06 UTC (permalink / raw)
To: Morton, Derek J, intel-gfx; +Cc: Wood, Thomas
> -----Original Message-----
> From: Morton, Derek J
> Sent: Tuesday, May 12, 2015 1:10 PM
> To: intel-gfx@lists.freedesktop.org
> Cc: Wood, Thomas; Gore, Tim; Morton, Derek J
> Subject: [PATCH i-g-t] tests/gem_cpu_reloc: Fix gem_cpu_reloc OOM failure
>
> On android with small memory footprints gem_cpu_reloc can fail due to
> OOM.
>
> Refactor gem_cpu_reloc into 2 tests, a basic test which performs
> 10 relocations and a full test which skips if there is insufficient memory.
>
> Changed the memory required test to better estimate the actual RAM used.
>
> Signed-off-by: Derek Morton <derek.j.morton@intel.com>
> ---
> tests/gem_cpu_reloc.c | 48 ++++++++++++++++++++++++++++++++++++-
> -----------
> 1 file changed, 36 insertions(+), 12 deletions(-)
>
> diff --git a/tests/gem_cpu_reloc.c b/tests/gem_cpu_reloc.c index
> 058089c..2f5fc60 100644
> --- a/tests/gem_cpu_reloc.c
> +++ b/tests/gem_cpu_reloc.c
> @@ -159,16 +159,14 @@ uint32_t gen8_batch[] = { uint32_t *batch =
> gen6_batch; uint32_t batch_size = sizeof(gen6_batch);
>
> -igt_simple_main
> +void run_test(int fd, int count)
> {
> const uint32_t hang[] = {-1, -1, -1, -1};
> const uint32_t end[] = {MI_BATCH_BUFFER_END, 0};
> - uint64_t aper_size;
> uint32_t noop;
> uint32_t *handles;
> - int fd, i, count;
> + int i;
>
> - fd = drm_open_any();
> noop = intel_get_drm_devid(fd);
>
> use_blt = 0;
> @@ -180,13 +178,6 @@ igt_simple_main
> batch_size += 2 * 4;
> }
>
> - aper_size = gem_mappable_aperture_size();
> - count = aper_size / 4096 * 2;
> - if (igt_run_in_simulation())
> - count = 10;
> -
> - intel_require_memory(1+count, 4096, CHECK_RAM);
> -
> handles = malloc (count * sizeof(uint32_t));
> igt_assert(handles);
>
> @@ -246,6 +237,39 @@ igt_simple_main
> igt_progress("gem_cpu_reloc: ", 2*count+i, 3*count);
> }
>
> - igt_info("Test suceeded, cleanup up - this might take a while.\n");
> + igt_info("Subtest suceeded, cleanup up - this might take a while.\n");
> + for (i = 0; i < count; i++) {
> + gem_close(fd, handles[i]);
> + }
> + gem_close(fd, noop);
> + free(handles);
> +}
> +
> +igt_main
> +{
> + uint64_t aper_size;
> + int fd, count;
> +
> + igt_fixture {
> + fd = drm_open_any();
> + }
> +
> + igt_subtest("basic") {
> + run_test (fd, 10);
> + }
> +
> + igt_skip_on_simulation();
> +
> + igt_subtest("full") {
> + aper_size = gem_mappable_aperture_size();
> + count = aper_size / 4096 * 2;
> +
> + /* count + 2 (noop & bad) buffers. A gem object appears to
> + require about 2kb + buffer + kernel overhead */
> + intel_require_memory(2+count, 2048+4096, CHECK_RAM);
> +
> + run_test (fd, count);
> + }
> +
> close(fd);
> }
> --
> 1.9.1
Works for me. This follows the pattern we have been using for other tests
that cannot run on low memory systems.
Reviewed-by: Tim Gore <tim.gore@intel.com>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH i-g-t] tests/gem_cpu_reloc: Fix gem_cpu_reloc OOM failure
2015-05-12 12:09 [PATCH i-g-t] tests/gem_cpu_reloc: Fix gem_cpu_reloc OOM failure Derek Morton
2015-05-13 11:06 ` Gore, Tim
@ 2015-05-13 16:41 ` Thomas Wood
1 sibling, 0 replies; 3+ messages in thread
From: Thomas Wood @ 2015-05-13 16:41 UTC (permalink / raw)
To: Derek Morton; +Cc: Intel Graphics Development
On 12 May 2015 at 13:09, Derek Morton <derek.j.morton@intel.com> wrote:
> On android with small memory footprints gem_cpu_reloc can fail
> due to OOM.
>
> Refactor gem_cpu_reloc into 2 tests, a basic test which performs
> 10 relocations and a full test which skips if there is insufficient
> memory.
>
> Changed the memory required test to better estimate the actual
> RAM used.
>
> Signed-off-by: Derek Morton <derek.j.morton@intel.com>
> ---
> tests/gem_cpu_reloc.c | 48 ++++++++++++++++++++++++++++++++++++------------
> 1 file changed, 36 insertions(+), 12 deletions(-)
>
> diff --git a/tests/gem_cpu_reloc.c b/tests/gem_cpu_reloc.c
> index 058089c..2f5fc60 100644
> --- a/tests/gem_cpu_reloc.c
> +++ b/tests/gem_cpu_reloc.c
> @@ -159,16 +159,14 @@ uint32_t gen8_batch[] = {
> uint32_t *batch = gen6_batch;
> uint32_t batch_size = sizeof(gen6_batch);
>
> -igt_simple_main
> +void run_test(int fd, int count)
This function should be marked static.
> {
> const uint32_t hang[] = {-1, -1, -1, -1};
> const uint32_t end[] = {MI_BATCH_BUFFER_END, 0};
> - uint64_t aper_size;
> uint32_t noop;
> uint32_t *handles;
> - int fd, i, count;
> + int i;
>
> - fd = drm_open_any();
> noop = intel_get_drm_devid(fd);
>
> use_blt = 0;
> @@ -180,13 +178,6 @@ igt_simple_main
> batch_size += 2 * 4;
> }
>
> - aper_size = gem_mappable_aperture_size();
> - count = aper_size / 4096 * 2;
> - if (igt_run_in_simulation())
> - count = 10;
> -
> - intel_require_memory(1+count, 4096, CHECK_RAM);
> -
> handles = malloc (count * sizeof(uint32_t));
> igt_assert(handles);
>
> @@ -246,6 +237,39 @@ igt_simple_main
> igt_progress("gem_cpu_reloc: ", 2*count+i, 3*count);
> }
>
> - igt_info("Test suceeded, cleanup up - this might take a while.\n");
> + igt_info("Subtest suceeded, cleanup up - this might take a while.\n");
> + for (i = 0; i < count; i++) {
> + gem_close(fd, handles[i]);
> + }
> + gem_close(fd, noop);
> + free(handles);
> +}
> +
> +igt_main
> +{
> + uint64_t aper_size;
> + int fd, count;
> +
> + igt_fixture {
> + fd = drm_open_any();
> + }
> +
> + igt_subtest("basic") {
> + run_test (fd, 10);
> + }
> +
> + igt_skip_on_simulation();
> +
> + igt_subtest("full") {
> + aper_size = gem_mappable_aperture_size();
> + count = aper_size / 4096 * 2;
> +
> + /* count + 2 (noop & bad) buffers. A gem object appears to
> + require about 2kb + buffer + kernel overhead */
> + intel_require_memory(2+count, 2048+4096, CHECK_RAM);
> +
> + run_test (fd, count);
> + }
> +
> close(fd);
This should be inside an igt_fixture block, otherwise fd will be
uninitialised when listing subtests.
> }
> --
> 1.9.1
>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2015-05-13 16:41 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-05-12 12:09 [PATCH i-g-t] tests/gem_cpu_reloc: Fix gem_cpu_reloc OOM failure Derek Morton
2015-05-13 11:06 ` Gore, Tim
2015-05-13 16:41 ` Thomas Wood
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.