| Message ID | 20191031152857.17143-4-janusz.krzysztofik@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Calculate softpin offsets from minimum GTT alignment | expand |
Quoting Janusz Krzysztofik (2019-10-31 15:28:56) > The basic-range subtest assumes 4kB GTT alignment while calculating > softpin offsets. On future backends with possibly larger minimum page > sizes the test will fail as a half of calculated offsets to be tested > will be incorrectly aligned. > > Replace hardcoded constants corresponding to the assumed 4kB GTT > alignment with variables initialized with actual minimum GTT alignment > size and order. > > v2: Simplify the code by reversing the size->order conversion, > - drop irrelevant modifications of requested object sizes. > v3: Reword commit message after removal of patch "Don't filter out > addresses on full PPGTT" from the series, > - initialize page size order with an actual minimum returned by a new > helper (inspired by Chris). > v4: Update the helper name, use the term 'minimum GTT alignment' across > the change, adjust variable names, > - refresh the commit message on top of the reintroduced patch that > fixes invalid offsets incorrectly assumed as occupied. > > Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com> > Cc: Katarzyna Dec <katarzyna.dec@intel.com> > Cc: Stuart Summers <stuart.summers@intel.com> > Cc: Chris Wilson <chris@chris-wilson.co.uk> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> -Chris
diff --git a/tests/i915/gem_exec_reloc.c b/tests/i915/gem_exec_reloc.c index 423fed8b..8050cd3e 100644 --- a/tests/i915/gem_exec_reloc.c +++ b/tests/i915/gem_exec_reloc.c @@ -520,14 +520,16 @@ static void basic_range(int fd, unsigned flags) uint64_t gtt_size = gem_aperture_size(fd); const uint32_t bbe = MI_BATCH_BUFFER_END; igt_spin_t *spin = NULL; + int alignment_order = gem_gtt_min_alignment_order(fd); + uint64_t alignment = 1ull << alignment_order; int count, n, err; igt_require(gem_has_softpin(fd)); - for (count = 12; gtt_size >> (count + 1); count++) + for (count = alignment_order; gtt_size >> (count + 1); count++) ; - count -= 12; + count -= alignment_order; memset(obj, 0, sizeof(obj)); memset(reloc, 0, sizeof(reloc)); @@ -536,7 +538,7 @@ static void basic_range(int fd, unsigned flags) n = 0; for (int i = 0; i <= count; i++) { obj[n].handle = gem_create(fd, 4096); - obj[n].offset = (1ull << (i + 12)) - 4096; + obj[n].offset = (1ull << (i + alignment_order)) - alignment; obj[n].offset = gen8_canonical_address(obj[n].offset); obj[n].flags = EXEC_OBJECT_PINNED | EXEC_OBJECT_SUPPORTS_48B_ADDRESS; gem_write(fd, obj[n].handle, 0, &bbe, sizeof(bbe)); @@ -559,7 +561,7 @@ static void basic_range(int fd, unsigned flags) } for (int i = 1; i < count; i++) { obj[n].handle = gem_create(fd, 4096); - obj[n].offset = 1ull << (i + 12); + obj[n].offset = 1ull << (i + alignment_order); obj[n].offset = gen8_canonical_address(obj[n].offset); obj[n].flags = EXEC_OBJECT_PINNED | EXEC_OBJECT_SUPPORTS_48B_ADDRESS; gem_write(fd, obj[n].handle, 0, &bbe, sizeof(bbe));
The basic-range subtest assumes 4kB GTT alignment while calculating softpin offsets. On future backends with possibly larger minimum page sizes the test will fail as a half of calculated offsets to be tested will be incorrectly aligned. Replace hardcoded constants corresponding to the assumed 4kB GTT alignment with variables initialized with actual minimum GTT alignment size and order. v2: Simplify the code by reversing the size->order conversion, - drop irrelevant modifications of requested object sizes. v3: Reword commit message after removal of patch "Don't filter out addresses on full PPGTT" from the series, - initialize page size order with an actual minimum returned by a new helper (inspired by Chris). v4: Update the helper name, use the term 'minimum GTT alignment' across the change, adjust variable names, - refresh the commit message on top of the reintroduced patch that fixes invalid offsets incorrectly assumed as occupied. Signed-off-by: Janusz Krzysztofik <janusz.krzysztofik@linux.intel.com> Cc: Katarzyna Dec <katarzyna.dec@intel.com> Cc: Stuart Summers <stuart.summers@intel.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> --- tests/i915/gem_exec_reloc.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)