tests/gem_tiled_fence_blits: split into subtests
PATCH 1/1 tests/gem_tiled_fence_blits: split into subtests
From: Tim Gore <[email protected]>
The gem_tiled_fence_blits test tends to get oom killed on low memory (< 4GB) Android systems. This is because the test tries to allocate (sysinfo.totalram * 9 / 10) in buffer objects and the remaining 10% of memory is not always enough for the Android system.
As with a similar issue with gem_render_linear_blits, this problem is resolved by splitting into subtests. A "basic" subtest uses minimal memory buffers to test the basic operation, and a "normal" subtest which is skipped if there is insufficient memory.
I also took the opportunity to cull some numeric literals.
--- tests/gem_tiled_fence_blits.c | 51 ++++++++++++++++++++++++++++--------------- 1 file changed, 33 insertions(+), 18 deletions(-)
tests/gem_tiled_fence_blits.c +33 -18
@@ -62,8 +62,9 @@static drm_intel_bufmgr *bufmgr;struct intel_batchbuffer *batch; -static int width =512 , height =512 ; -static uint32_t linear[1024 *1024 /4 ]; +enum {width=512 , height=512 }; +static const int bo_size = width * height *4 ; +static uint32_t linear[width * height];static drm_intel_bo *create_bo (int fd,uint32_t start_val)
@@ -72,13 +73,13 @@ create_bo(int fd, uint32_t start_val)uint32_t tiling =I915_TILING_X ;int ret, i; - bo =drm_intel_bo_alloc (bufmgr,"tiled bo" ,1024 *1024 ,4096 ); + bo =drm_intel_bo_alloc (bufmgr,"tiled bo" , bo_size,4096 ); ret =drm_intel_bo_set_tiling (bo, &tiling, width *4 );igt_assert_eq (ret,0 );igt_assert (tiling ==I915_TILING_X );/* Fill the BO with dwords starting at start_val */ -for (i =0 ; i <1024 *1024 /4 ; i++) +for (i =0 ; i < width * height; i++) linear[i] = start_val++;gem_write (fd, bo->handle,0 , linear,sizeof (linear));
@@ -93,7 +94,7 @@ check_bo(int fd, drm_intel_bo *bo, uint32_t start_val)gem_read (fd, bo->handle,0 , linear,sizeof (linear)); -for (i =0 ; i <1024 *1024 /4 ; i++) { +for (i =0 ; i < width * height; i++) { igt_assert_f(linear[i] == start_val, "Expected 0x%08x, found 0x%08x " "at offset 0x%08x\n",
@@ -102,21 +103,13 @@ check_bo(int fd, drm_intel_bo *bo, uint32_t start_val) } } -igt_simple_main +static void run_test (int fd,int count) {drm_intel_bo *bo[4096 ];uint32_t bo_start_val[4096 ];uint32_t start =0 ; -int fd, i, count; - -igt_skip_on_simulation (); +int i; - fd =drm_open_any (); - count =3 *gem_aperture_size (fd) / (1024 *1024 ) /2 ; -if (count >intel_get_total_ram_mb () *9 /10 ) { - count =intel_get_total_ram_mb () *9 /10 ; -igt_info ("not enough RAM to run test, reducing buffer count\n" ); - } count |=1 ;igt_info ("Using %d 1MiB buffers\n" , count);
@@ -133,12 +126,12 @@ igt_simple_maincheck_bo(bo[i], bo_start_val[i]); */ - start +=1024 *1024 /4 ; + start += width * height; }for (i =0 ; i < count; i++) {int src = count - i -1 ; -intel_copy_bo (batch, bo[i], bo[src], width*height*4 ); +intel_copy_bo (batch, bo[i], bo[src], bo_size); bo_start_val[i] = bo_start_val[src]; }
@@ -149,7 +142,7 @@ igt_simple_mainif (src == dst)continue ; -intel_copy_bo (batch, bo[dst], bo[src], width*height*4 ); +intel_copy_bo (batch, bo[dst], bo[src], bo_size); bo_start_val[dst] = bo_start_val[src];/*
@@ -170,6 +163,28 @@ igt_simple_mainintel_batchbuffer_free (batch);drm_intel_bufmgr_destroy (bufmgr); +} + +igt_main +{ +int fd, count; + + igt_fixture { + fd =drm_open_any (); + } + + igt_subtest("basic") { +run_test (fd,2 ); + } + +/* the rest of the tests are too long for simulation */ +igt_skip_on_simulation (); + + igt_subtest("normal") { + count =3 *gem_aperture_size (fd) / (bo_size) /2 ; +intel_require_memory (count, bo_size,CHECK_RAM ); +run_test (fd, count); + }close (fd); }
Signature · 6 lines
2.3.0 _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Quoted 13 lines · On Tue, Mar 24, 2015 at 02:16:04PM +0000, [email protected] wrote:
From: Tim Gore <[email protected]>
The
gem_tiled_fence_blitstest tends to get oom killed on low memory (< 4GB) Android systems. This is because the test tries to allocate (sysinfo.totalram * 9 / 10) in buffer objects and the remaining 10% of memory is not always enough for the Android system.
As with a similar issue withgem_render_linear_blits, this problem is resolved by splitting into subtests. A "basic" subtest uses minimal memory buffers to test the basic operation, and a "normal" subtest which is skipped if there is insufficient memory.
Quoted 3 lines
I also took the opportunity to cull some numeric literals.
Signed-off-by Tim Gore <[email protected]>
Looks good, thanks.
-Chris
Signature · 5 lines
Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/intel-gfx