[igt-dev] [PATCH i-g-t] test/gem_exec_suspend: Enable the test for discrete GPUs
Thomas Hellström
thomas.hellstrom at linux.intel.com
Mon Sep 6 09:31:53 UTC 2021
On 9/6/21 8:25 AM, Petri Latvala wrote:
> On Fri, Sep 03, 2021 at 04:03:48PM +0200, Thomas Hellström wrote:
>> Discrete GPUs don't support gem_set_caching() and uses a single,
>> kernel provided mapping mode, FIXED. Adjust the test accordingly.
>>
>> This makes the test run on DG1.
>>
>> Signed-off-by: Thomas Hellström <thomas.hellstrom at linux.intel.com>
>> ---
>> tests/i915/gem_exec_suspend.c | 17 ++++++++++++++++-
>> 1 file changed, 16 insertions(+), 1 deletion(-)
>>
>> diff --git a/tests/i915/gem_exec_suspend.c b/tests/i915/gem_exec_suspend.c
>> index dbe0c8a7..5f17e600 100644
>> --- a/tests/i915/gem_exec_suspend.c
>> +++ b/tests/i915/gem_exec_suspend.c
>> @@ -114,7 +114,8 @@ static void run_test(int fd, const intel_ctx_t *ctx,
>>
>> memset(obj, 0, sizeof(obj));
>> obj[0].handle = gem_create(fd, 4096);
>> - gem_set_caching(fd, obj[0].handle, !!(flags & CACHED));
>> + if (!gem_has_lmem(fd))
>> + gem_set_caching(fd, obj[0].handle, !!(flags & CACHED));
>> obj[0].flags |= EXEC_OBJECT_WRITE;
>> obj[1].handle = gem_create(fd, 4096);
>> gem_write(fd, obj[1].handle, 0, &bbe, sizeof(bbe));
>> @@ -293,6 +294,7 @@ igt_main
>> const struct intel_execution_engine2 *e;
>> igt_hang_t hang;
>> const intel_ctx_t *ctx;
>> + bool has_lmem;
>> int fd;
>>
>> igt_fixture {
>> @@ -302,6 +304,7 @@ igt_main
>> ctx = intel_ctx_create_all_physical(fd);
>>
>> igt_fork_hang_detector(fd);
>> + has_lmem = gem_has_lmem(fd);
>> }
>>
>> igt_subtest("basic")
>> @@ -318,6 +321,18 @@ igt_main
>> run_test(fd, ctx, ALL_ENGINES, HIBERNATE);
>>
>> for (m = modes; m->suffix; m++) {
>> + if (has_lmem) {
>> + igt_subtest_with_dynamic_f("fixed%s", m->suffix) {
> You can't do
>
> if (something_that_depends_on_the_running_system)
> igt_subtest_of_any_kind(...)
>
>
> Instead you need to do
>
> igt_subtest_of_any_kind(...) {
> igt_require(something_that_depends_on_the_running_system);
> ...
> }
>
>
Thanks. I'll respin.
/Thomas
More information about the igt-dev
mailing list