Upgrading the Engine discovery
v2 cover Upgrading the Engine discovery
Replaces the old static engine identification with new uAPI for engine discovery.
Marked as RFC, because this is not tested on HW yet. And we need initial review along with CI reaction.
Ramalingam C (43):
test/i915: gem_bad_reloc: use the gem_engine_topology library test/i915: gem_exec_async: use the gem_engine_topology library test/i915: gem_exec_capture: use the gem_engine_topology library test/i915: gem_exec_params: use the gem_engine_topology library lib/i915: Helper func for engine class to exec_flag test/prime_vgem: use the gem_engine_topology library tests/i915/gem_exec_reuse: use the gem_engine_topology library tests/i915/gem_sync: use the gem_engine_topology library tests/i915/gem_busy: use the gem_engine_topology library tests/i915/gem_concurrent_all: use the gem_engine_topology library tests/i915/gem_cs_prefetch: use the gem_engine_topology library tests/i915/gem_ctx_create: use the gem_engine_topology library tests/i915/gem_ctx_isolation: use the gem_engine_topology library tests/i915/gem_ctx_switch: use the gem_engine_topology library tests/i915/gem_ctx_thrash: use the gem_engine_topology library tests/i915/gem_eio: use the gem_engine_topology library tests/i915/gem_exec_await: use the gem_engine_topology library tests/i915/gem_exec_create: use the gem_engine_topology library tests/i915/gem_exec_fence: use the gem_engine_topology library tests/i915/gem_exec_flush: use the gem_engine_topology library tests/i915/gem_exec_gttfill: use the gem_engine_topology library tests/i915/gem_exec_latency: use the gem_engine_topology library tests/i915/gem_exec_nop: use the gem_engine_topology library tests/i915/gem_exec_reloc: use the gem_engine_topology library tests/i915/gem_exec_schedule: use the gem_engine_topology library tests/i915/gem_exec_suspend: use the gem_engine_topology library tests/i915/gem_exec_whisper: use the gem_engine_topology library tests/i915/gem_mocs_settings: use the gem_engine_topology library tests/i915/gem_reset_stats: use the gem_engine_topology library tests/i915/gem_ring_sync_loop: use the gem_engine_topology library tests/i915/gem_ringfill: use the gem_engine_topology library tests/i915/gem_shrink: use the gem_engine_topology library tests/i915/gem_spin_batch: use the gem_engine_topology library tests/i915/gem_storedw_loop: use the gem_engine_topology library tests/i915/gem_userptr_blits: use the gem_engine_topology library tests/i915/i915_hangman: use the gem_engine_topology library tests/i915/i915_module_load: use the gem_engine_topology library tests/kms_busy: use the gem_engine_topology library tests/prime_busy: use the gem_engine_topology library tests/amdgpu/amd_prime: use the gem_engine_topology library tests/i915/gem_ctx_shared: use the gem_engine_topology library
lib: replace the legacy funcs with gem_engine_topology libraries lib/gem_engine_topology: rename the macros
benchmarks/gem_syslatency.c | 7 +- lib/i915/gem_engine_topology.c | 26 ++- lib/i915/gem_engine_topology.h | 6 +- lib/i915/gem_ring.c | 7 +- lib/i915/gem_submission.c | 6 +- lib/igt_gt.c | 11 -- lib/igt_gt.h | 21 -- tests/amdgpu/amd_prime.c | 5 +- tests/i915/gem_bad_reloc.c | 14 +- tests/i915/gem_busy.c | 23 +-- tests/i915/gem_concurrent_all.c | 8 +- tests/i915/gem_cs_prefetch.c | 18 +- tests/i915/gem_cs_tlb.c | 2 +- tests/i915/gem_ctx_create.c | 34 ++-- tests/i915/gem_ctx_engines.c | 10 +- tests/i915/gem_ctx_exec.c | 2 +- tests/i915/gem_ctx_isolation.c | 19 +- tests/i915/gem_ctx_shared.c | 99 +++++----- tests/i915/gem_ctx_switch.c | 26 +-- tests/i915/gem_ctx_thrash.c | 15 +- tests/i915/gem_eio.c | 63 +++--- tests/i915/gem_exec_async.c | 34 ++-- tests/i915/gem_exec_await.c | 5 +- tests/i915/gem_exec_basic.c | 8 +- tests/i915/gem_exec_capture.c | 24 +-- tests/i915/gem_exec_create.c | 6 +- tests/i915/gem_exec_fence.c | 112 +++++------ tests/i915/gem_exec_flush.c | 72 ++++--- tests/i915/gem_exec_gttfill.c | 30 +-- tests/i915/gem_exec_latency.c | 101 ++++------ tests/i915/gem_exec_nop.c | 119 ++++++------ tests/i915/gem_exec_parallel.c | 4 +- tests/i915/gem_exec_params.c | 41 +--- tests/i915/gem_exec_reloc.c | 27 +-- tests/i915/gem_exec_reuse.c | 12 +- tests/i915/gem_exec_schedule.c | 332 ++++++++++++++++---------------- tests/i915/gem_exec_store.c | 4 +- tests/i915/gem_exec_suspend.c | 52 ++--- tests/i915/gem_exec_whisper.c | 27 +-- tests/i915/gem_mocs_settings.c | 54 +++--- tests/i915/gem_reset_stats.c | 60 +++--- tests/i915/gem_ring_sync_loop.c | 4 +- tests/i915/gem_ringfill.c | 32 ++- tests/i915/gem_shrink.c | 22 ++- tests/i915/gem_spin_batch.c | 36 ++-- tests/i915/gem_storedw_loop.c | 39 ++-- tests/i915/gem_sync.c | 232 +++++++++++----------- tests/i915/gem_userptr_blits.c | 8 +- tests/i915/gem_wait.c | 4 +- tests/i915/i915_hangman.c | 8 +- tests/i915/i915_module_load.c | 29 +-- tests/kms_busy.c | 62 +++--- tests/perf_pmu.c | 12 +- tests/prime_busy.c | 24 +-- tests/prime_vgem.c | 72 ++++--- 55 files changed, 1047 insertions(+), 1083 deletions(-)
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 1/43 test/i915: gem_bad_reloc: use the gem_engine_topology library
Replace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_bad_reloc.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)
tests/i915/gem_bad_reloc.c +7 -7
@@ -50,7 +50,8 @@ IGT_TEST_DESCRIPTION("Simulates SNA behaviour using negative self-relocations"* than the total size of the GTT), the GPU will hang. * See https://bugs.freedesktop.org/show_bug.cgi?id=78533 */ -static void negative_reloc (int fd,unsigned engine,unsigned flags) +static void negative_reloc (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj;
@@ -60,7 +61,6 @@ static void negative_reloc(int fd, unsigned engine, unsigned flags)uint64_t *offsets;int i; -gem_require_ring (fd, engine);igt_require (intel_gen (intel_get_drm_devid (fd)) >=7 );memset (&obj,0 ,sizeof (obj));
@@ -70,7 +70,7 @@ static void negative_reloc(int fd, unsigned engine, unsigned flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )&obj; execbuf.buffer_count =1 ; - execbuf.flags = engine | (flags &USE_LUT ); + execbuf.flags = e->flags | (flags &USE_LUT );igt_require (__gem_execbuf (fd, &execbuf) ==0 );igt_info ("Found offset %lld for 4k batch\n" , (long long )obj.offset);
@@ -185,7 +185,7 @@ static void negative_reloc_blt(int fd) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd =-1 ; igt_fixture {
@@ -193,12 +193,12 @@ igt_mainigt_require_gem (fd); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) { igt_subtest_f("negative-reloc-%s", e->name) -negative_reloc (fd, e->exec_id | e->flags,0 ); +negative_reloc (fd, e,0 ); igt_subtest_f("negative-reloc-lut-%s", e->name) -negative_reloc (fd, e->exec_id | e->flags,USE_LUT ); +negative_reloc (fd, e,USE_LUT ); }igt_subtest ("negative-reloc-bltcopy" )
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 01/43] test/i915: gem_bad_reloc: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:03)
Quoted 49 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.Signed-off-by Ramalingam C <[email protected]>--- tests/i915/gem_bad_reloc.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)tests/i915/gem_bad_reloc.c +4
-4@@ -50,7 +50,8 @@ IGT_TEST_DESCRIPTION("Simulates SNA behaviour using negative self-relocations"* than the total size of the GTT), the GPU will hang. * See https://bugs.freedesktop.org/show_bug.cgi?id=78533 */ -static void negative_reloc (int fd,unsigned engine,unsigned flags) +static void negative_reloc (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj;@@ -60,7 +61,6 @@ static void negative_reloc(int fd, unsigned engine, unsigned flags)uint64_t *offsets;int i; -gem_require_ring (fd, engine);igt_require (intel_gen (intel_get_drm_devid (fd)) >=7 );memset (&obj,0 ,sizeof (obj));@@ -70,7 +70,7 @@ static void negative_reloc(int fd, unsigned engine, unsigned flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )&obj; execbuf.buffer_count =1 ; - execbuf.flags = engine | (flags &USE_LUT ); + execbuf.flags = e->flags | (flags &USE_LUT );igt_require (__gem_execbuf (fd, &execbuf) ==0 );igt_info ("Found offset %lld for 4k batch\n" , (long long )obj.offset);@@ -185,7 +185,7 @@ static void negative_reloc_blt(int fd) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd =-1 ; igt_fixture {
tests/i915/gem_bad_reloc.c igt_main · line 196
Quoted 5 lines
tests/i915/gem_bad_reloc.c +1
-1@@ -193,12 +193,12 @@ igt_mainigt_require_gem (fd); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) {
uABI testing not physical engines.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 01/43] test/i915: gem_bad_reloc: use the gem_engine_topology library
Quoted 53 lines · Chris Wilson
Quoting Ramalingam C (2019-06-21 11:03:03)
Quoted 52 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.Signed-off-by Ramalingam C <[email protected]>--- tests/i915/gem_bad_reloc.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-)tests/i915/gem_bad_reloc.c +4
-4@@ -50,7 +50,8 @@ IGT_TEST_DESCRIPTION("Simulates SNA behaviour using negative self-relocations"* than the total size of the GTT), the GPU will hang. * See https://bugs.freedesktop.org/show_bug.cgi?id=78533 */ -static void negative_reloc (int fd,unsigned engine,unsigned flags) +static void negative_reloc (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj;@@ -60,7 +61,6 @@ static void negative_reloc(int fd, unsigned engine, unsigned flags)uint64_t *offsets;int i; -gem_require_ring (fd, engine);igt_require (intel_gen (intel_get_drm_devid (fd)) >=7 );memset (&obj,0 ,sizeof (obj));@@ -70,7 +70,7 @@ static void negative_reloc(int fd, unsigned engine, unsigned flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )&obj; execbuf.buffer_count =1 ; - execbuf.flags = engine | (flags &USE_LUT ); + execbuf.flags = e->flags | (flags &USE_LUT );igt_require (__gem_execbuf (fd, &execbuf) ==0 );igt_info ("Found offset %lld for 4k batch\n" , (long long )obj.offset);@@ -185,7 +185,7 @@ static void negative_reloc_blt(int fd) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd =-1 ; igt_fixture {@@ -193,12 +193,12 @@ igt_mainigt_require_gem (fd); }
Quoted 4 lines
tests/i915/gem_bad_reloc.c igt_main · line 196
Quoted 2 lines
tests/i915/gem_bad_reloc.c +1
-1-for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) {uABI testing not physical engines.
Should we add something like for_each_legacy_abi_engine to use in these tests so it is more self-documenting? (And get rid of raw access to intel_execution_engines.)
Regards,
Tvrtko
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 01/43] test/i915: gem_bad_reloc: use the gem_engine_topology library
Quoting Tvrtko Ursulin (2019-06-21 11:05:46)
Quoted 7 lines · Chris Wilson
Quoted 5 lines
Quoting Ramalingam C (2019-06-21 11:03:03)
Quoted 4 lines
tests/i915/gem_bad_reloc.c +0
-0@@ -193,12 +193,12 @@ igt_mainigt_require_gem (fd); }
Quoted 8 lines
Quoted 4 lines
tests/i915/gem_bad_reloc.c igt_main · line 196
Quoted 2 lines
tests/i915/gem_bad_reloc.c +1
-1-for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) {uABI testing not physical engines.
Should we add something like
for_each_legacy_abi_engineto use in these tests so it is more self-documenting? (And get rid of raw access tointel_execution_engines.)
Yes, a for_each_legacy_abi_engine would be most welcome. The basic parameter testing should exist for both . Above that I would like to keep some basic behavioural checking sufficient so that we can be confident that yesterday's mesa/etc keeps on working on tomorrow's kernel. After that, so long as we can say "we can tested this behaviour with the new ABI and have proven the equivalence via this and this test, ergo we can be confident the legacy ABI will work (if applicable)". -Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 2/43 test/i915: gem_exec_async: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
v2:
class and instance are compared for engine identification [Andi]
--- tests/i915/gem_exec_async.c | 34 +++++++++++++++------------------- 1 file changed, 15 insertions(+), 19 deletions(-)
tests/i915/gem_exec_async.c +15 -19
@@ -28,7 +28,7 @@IGT_TEST_DESCRIPTION ("Check that we can issue concurrent writes across the engines." ); -static void store_dword (int fd,unsigned ring, +static void store_dword (int fd,const struct intel_execution_engine2 *e,uint32_t target,uint32_t offset,uint32_t value) {const int gen =intel_gen (intel_get_drm_devid (fd));
@@ -41,7 +41,7 @@ static void store_dword(int fd, unsigned ring,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = ring; + execbuf.flags = e->flags;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -80,15 +80,15 @@ static void store_dword(int fd, unsigned ring,gem_close (fd, obj[1 ].handle); } -static void one (int fd,unsigned ring,uint32_t flags) +static void one (int fd,const struct intel_execution_engine2 *e) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_exec_object2 obj[2 ]; +struct intel_execution_engine2 *engine;#define SCRATCH 0#define BATCH 1struct drm_i915_gem_relocation_entry reloc;struct drm_i915_gem_execbuffer2 execbuf; -unsigned int other;uint32_t *batch;int i;
@@ -138,19 +138,21 @@ static void one(int fd, unsigned ring, uint32_t flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = ring | flags; + execbuf.flags = e->flags;igt_require (__gem_execbuf (fd, &execbuf) ==0 );gem_close (fd, obj[BATCH ].handle); i =0 ; -for_each_physical_engine (fd, other) { -if (other == ring) + +__for_each_physical_engine (fd, engine) { +if (engine->class == e->class && + engine->instance == e->instance)continue ; -if (!gem_can_store_dword (fd, other)) +if (!gem_class_can_store_dword (fd, engine->class))continue ; -store_dword (fd, other, obj[SCRATCH ].handle,4 *i, i); +store_dword (fd, engine, obj[SCRATCH ].handle,4 *i, i); i++; }
@@ -185,7 +187,7 @@ static bool has_async_execbuf(int fd) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd =-1 ;igt_skip_on_simulation ();
@@ -199,17 +201,11 @@ igt_mainigt_fork_hang_detector (fd); } -for (e = intel_execution_engines; e->name; e++) { -/* default exec-id is purely symbolic */ -if (e->exec_id ==0 ) -continue ; - +__for_each_physical_engine (fd, e) igt_subtest_f("concurrent-writes-%s", e->name) { -igt_require (gem_ring_has_physical_engine (fd, e->exec_id | e->flags)); -igt_require (gem_can_store_dword (fd, e->exec_id | e->flags)); -one (fd, e->exec_id, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); +one (fd, e); } - } igt_fixture {igt_stop_hang_detector ();
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 3/43 test/i915: gem_exec_capture: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_capture.c | 24 ++++++++++-------------- 1 file changed, 10 insertions(+), 14 deletions(-)
tests/i915/gem_exec_capture.c +10 -14
@@ -61,7 +61,8 @@ static void check_error_state(int dir, struct drm_i915_gem_exec_object2 *obj)igt_assert (found); } -static void __capture1 (int fd,int dir,unsigned ring,uint32_t target) +static void __capture1 (int fd,int dir,const struct intel_execution_engine2 *e, +uint32_t target) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_exec_object2 obj[4 ];
@@ -145,7 +146,7 @@ static void __capture1(int fd, int dir, unsigned ring, uint32_t target)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )obj; execbuf.buffer_count =ARRAY_SIZE (obj); - execbuf.flags = ring; + execbuf.flags = e ? e->flags :0 ;if (gen >3 && gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -168,12 +169,12 @@ static void __capture1(int fd, int dir, unsigned ring, uint32_t target)gem_close (fd, obj[SCRATCH ].handle); } -static void capture (int fd,int dir,unsigned ring) +static void capture (int fd,int dir,const struct intel_execution_engine2 *e) {uint32_t handle; handle =gem_create (fd,4096 ); -__capture1 (fd, dir, ring, handle); +__capture1 (fd, dir, e, handle);gem_close (fd, handle); }
@@ -501,7 +502,7 @@ static void userptr(int fd, int dir)igt_assert (posix_memalign (&ptr,4096 ,4096 ) ==0 );igt_require (__gem_userptr (fd, ptr,4096 ,0 ,0 , &handle) ==0 ); -__capture1 (fd, dir,0 , handle); +__capture1 (fd, dir,NULL , handle);gem_close (fd, handle);free (ptr);
@@ -526,7 +527,7 @@ static size_t safer_strlen(const char *s) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;igt_hang_t hang;int fd =-1 ;int dir =-1 ;
@@ -552,15 +553,10 @@ igt_mainigt_require (safer_strlen (igt_sysfs_get (dir,"error" )) >0 ); } -for (e = intel_execution_engines; e->name; e++) { -/* default exec-id is purely symbolic */ -if (e->exec_id ==0 ) -continue ; - +__for_each_physical_engine (fd, e) { igt_subtest_f("capture-%s", e->name) { -igt_require (gem_ring_has_physical_engine (fd, e->exec_id | e->flags)); -igt_require (gem_can_store_dword (fd, e->exec_id | e->flags)); -capture (fd, dir, e->exec_id | e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); +capture (fd, dir, e); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 4/43 test/i915: gem_exec_params: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
v2:
has_ring() check is removed [tvertko] Engine type is identified through CLASS [tvrtko]
--- tests/i915/gem_exec_params.c | 41 ++++++------------------------------ 1 file changed, 7 insertions(+), 34 deletions(-)
tests/i915/gem_exec_params.c +7 -34
@@ -52,30 +52,6 @@#define LOCAL_I915_EXEC_BATCH_FIRST (1 << 18)#define LOCAL_I915_EXEC_FENCE_ARRAY (1 << 19) -static bool has_ring (int fd,unsigned ring_exec_flags) -{ -switch (ring_exec_flags &I915_EXEC_RING_MASK ) { -case 0 : -case I915_EXEC_RENDER : -return true; - -case I915_EXEC_BSD : -if (ring_exec_flags &LOCAL_I915_EXEC_BSD_MASK ) -return gem_has_bsd2 (fd); -else -return gem_has_bsd (fd); - -case I915_EXEC_BLT : -return gem_has_blt (fd); - -case I915_EXEC_VEBOX : -return gem_has_vebox (fd); - } - -igt_assert_f (0 ,"invalid exec flag 0x%x\n" , ring_exec_flags); -return false; -} -static bool has_exec_batch_first (int fd) {int val =-1 ;
@@ -201,7 +177,7 @@ int fd; igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e; igt_fixture { fd =drm_open_driver (DRIVER_INTEL );
@@ -235,11 +211,9 @@ igt_main } igt_subtest("control") { -for (e = intel_execution_engines; e->name; e++) { -if (has_ring (fd, e->exec_id | e->flags)) { - execbuf.flags = e->exec_id | e->flags; -gem_execbuf (fd, &execbuf); - } +__for_each_physical_engine (fd, e) { + execbuf.flags = e->flags; +gem_execbuf (fd, &execbuf); } }
@@ -387,14 +361,13 @@ igt_main igt_subtest("rs-invalid") {bool has_rs =has_resource_streamer (fd); -unsigned int engine; -for_each_engine (fd, engine) { +__for_each_physical_engine (fd, e) {int expect = -EINVAL ; -if (has_rs && (engine ==0 || engine ==I915_EXEC_RENDER )) +if (has_rs && (e->class ==I915_ENGINE_CLASS_RENDER )) expect =0 ; - execbuf.flags = engine |LOCAL_I915_EXEC_RESOURCE_STREAMER ; + execbuf.flags = e->flags |LOCAL_I915_EXEC_RESOURCE_STREAMER ;igt_assert_eq (__gem_execbuf (fd, &execbuf), expect); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 5/43 lib/i915: Helper func for engine class to exec_flag
In reply to Ramalingam CHelper function is defined to get the execbuf flag for a engine of a class mapped for a ctx.
--- lib/i915/gem_engine_topology.c | 22 ++++++++++++++++++++++ lib/i915/gem_engine_topology.h | 2 ++ 2 files changed, 24 insertions(+)
lib/i915/gem_engine_topology.c +22 -0
@@ -289,3 +289,25 @@ bool gem_has_engine_topology(int fd)return !__gem_context_get_param (fd, ¶m); } + +/* Return -EINVAL when the Engine with mentioned class is not found */ +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class) +{ +const struct intel_execution_engine2 *e; +DEFINE_CONTEXT_ENGINES_PARAM (engines, param, ctx_id,GEM_MAX_ENGINES ); +int i; + +if (gem_topology_get_param (fd, ¶m)) { +__for_each_static_engine (e) { +if (e->class == class) +return e->flags; + } +return -EINVAL ; + } + +for (i =0 ; i < param.size; i++) +if (engines.engines[i].engine_class == class) +return i; + +return -EINVAL ; +}
lib/i915/gem_engine_topology.h +2 -0
@@ -53,6 +53,8 @@ int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id,void gem_context_set_all_engines (int fd,uint32_t ctx); +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class); +#define __for_each_static_engine (e__) \ for ((e__) = intel_execution_engines2; (e__)->name; (e__)++)
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 05/43] lib/i915: Helper func for engine class to exec_flag
Quoted 2 lines · Ramalingam C
Helper function is defined to get the execbuf flag for a engine of a class mapped for a ctx.
We need matching on both class and instance. You can take https://patchwork.freedesktop.org/patch/312751/?series=62454 from Andi who wrote it in the meantime.
Regards,
Tvrtko
Quoted 42 lines
Signed-off-by Ramalingam C <[email protected]>--- lib/i915/gem_engine_topology.c | 22 ++++++++++++++++++++++ lib/i915/gem_engine_topology.h | 2 ++ 2 files changed, 24 insertions(+)lib/i915/gem_engine_topology.c +22
-0@@ -289,3 +289,25 @@ bool gem_has_engine_topology(int fd)return !__gem_context_get_param (fd, ¶m); } + +/* Return -EINVAL when the Engine with mentioned class is not found */ +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class) +{ +const struct intel_execution_engine2 *e; +DEFINE_CONTEXT_ENGINES_PARAM (engines, param, ctx_id,GEM_MAX_ENGINES ); +int i; + +if (gem_topology_get_param (fd, ¶m)) { +__for_each_static_engine (e) { +if (e->class == class) +return e->flags; + } +return -EINVAL ; + } + +for (i =0 ; i < param.size; i++) +if (engines.engines[i].engine_class == class) +return i; + +return -EINVAL ; +}lib/i915/gem_engine_topology.h +0
-0
lib/i915/gem_engine_topology.h int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id, · line 60
Quoted 8 lines
lib/i915/gem_engine_topology.h +2
-0@@ -53,6 +53,8 @@ int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id,void gem_context_set_all_engines (int fd,uint32_t ctx); +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class); +#define __for_each_static_engine (e__) \ for ((e__) = intel_execution_engines2; (e__)->name; (e__)++)
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 05/43] lib/i915: Helper func for engine class to exec_flag
Quoted 8 lines · Tvrtko Ursulin
Quoted 2 lines · Ramalingam C
Helper function is defined to get the execbuf flag for a engine of a class mapped for a ctx.
We need matching on both class and instance. You can take
https://patchwork.freedesktop.org/patch/312751/?series=62454 from Andi who wrote it in the meantime.
Ok Sure.
-Ram
Quoted 43 lines
Regards,
Tvrtko
Quoted 38 lines
Signed-off-by Ramalingam C <[email protected]>--- lib/i915/gem_engine_topology.c | 22 ++++++++++++++++++++++ lib/i915/gem_engine_topology.h | 2 ++ 2 files changed, 24 insertions(+)lib/i915/gem_engine_topology.c +22
-0@@ -289,3 +289,25 @@ bool gem_has_engine_topology(int fd)return !__gem_context_get_param (fd, ¶m); } + +/* Return -EINVAL when the Engine with mentioned class is not found */ +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class) +{ +const struct intel_execution_engine2 *e; +DEFINE_CONTEXT_ENGINES_PARAM (engines, param, ctx_id,GEM_MAX_ENGINES ); +int i; + +if (gem_topology_get_param (fd, ¶m)) { +__for_each_static_engine (e) { +if (e->class == class) +return e->flags; + } +return -EINVAL ; + } + +for (i =0 ; i < param.size; i++) +if (engines.engines[i].engine_class == class) +return i; + +return -EINVAL ; +}lib/i915/gem_engine_topology.h +0
-1--- a/lib/i915/gem_engine_topology.h
Quoted 8 lines
lib/i915/gem_engine_topology.h int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id, · line 58
Quoted 8 lines
+++
b/lib/i915/gem_engine_topology.hlib/i915/gem_engine_topology.h +2
-0@@ -53,6 +53,8 @@ int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id,void gem_context_set_all_engines (int fd,uint32_t ctx); +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class); +#define __for_each_static_engine (e__) \ for ((e__) = intel_execution_engines2; (e__)->name; (e__)++)
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 05/43] lib/i915: Helper func for engine class to exec_flag
In reply to Ramalingam CQuoted 14 lines · Ramalingam C
Helper function is defined to get the execbuf flag for a engine of a class mapped for a ctx.
Signed-off-by Ramalingam C <[email protected]>--- lib/i915/gem_engine_topology.c | 22 ++++++++++++++++++++++ lib/i915/gem_engine_topology.h | 2 ++ 2 files changed, 24 insertions(+)lib/i915/gem_engine_topology.c +0
-0
lib/i915/gem_engine_topology.c bool gem_has_engine_topology(int fd) · line 298
Quoted 10 lines
lib/i915/gem_engine_topology.c +7
-0@@ -289,3 +289,25 @@ bool gem_has_engine_topology(int fd)return !__gem_context_get_param (fd, ¶m); } + +/* Return -EINVAL when the Engine with mentioned class is not found */ +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class) +{ +const struct intel_execution_engine2 *e; +DEFINE_CONTEXT_ENGINES_PARAM (engines, param, ctx_id,GEM_MAX_ENGINES ); +int i;
Here Chris would complain that variable declaration should be declared in length order.
lib/i915/gem_engine_topology.c bool gem_has_engine_topology(int fd) · line 306
Quoted 7 lines
lib/i915/gem_engine_topology.c +7
-0+if (gem_topology_get_param (fd, ¶m)) { +__for_each_static_engine (e) { +if (e->class == class) +return e->flags; + } +return -EINVAL ; + }
Here the case where length is '0' is missing.
Quoted 11 lines
lib/i915/gem_engine_topology.c +6
-0+for (i =0 ; i < param.size; i++) +if (engines.engines[i].engine_class == class) +return i; + +return -EINVAL ; +}lib/i915/gem_engine_topology.h +0
-0
lib/i915/gem_engine_topology.h int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id, · line 59
Quoted 7 lines
lib/i915/gem_engine_topology.h +2
-0@@ -53,6 +53,8 @@ int gem_context_lookup_engine(int fd, uint64_t engine, uint32_t ctx_id,void gem_context_set_all_engines (int fd,uint32_t ctx); +int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class); +#define __for_each_static_engine (e__) \ for ((e__) = intel_execution_engines2; (e__)->name; (e__)++)
oh... looks familiar :) perhaps you can put my patch (or ours, they are basically the same) in this series and add sob by you, as well... there would be three sob this way (tvrtko, me, you).
Andi
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 6/43 test/prime_vgem: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
v2:
helper for getting the execbuf_flag from class is used.
--- tests/prime_vgem.c | 72 ++++++++++++++++++++++------------------------ 1 file changed, 34 insertions(+), 38 deletions(-)
tests/prime_vgem.c +34 -38
@@ -325,7 +325,7 @@ static bool prime_busy(int fd, bool excl)return poll (&pfd,1 ,0 ) ==0 ; } -static void work (int i915,int dmabuf,unsigned ring,uint32_t flags) +static void work (int i915,int dmabuf,const struct intel_execution_engine2 *e) {const int SCRATCH =0 ;const int BATCH =1 ;
@@ -341,7 +341,9 @@ static void work(int i915, int dmabuf, unsigned ring, uint32_t flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )obj; execbuf.buffer_count =2 ; - execbuf.flags = ring | flags; + execbuf.flags = e ? e->flags : +gem_engine_class_to_execbuf_flags (i915,0 , +I915_ENGINE_CLASS_RENDER );if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -420,7 +422,7 @@ static void work(int i915, int dmabuf, unsigned ring, uint32_t flags)igt_assert (read_busy && write_busy); } -static void test_busy (int i915,int vgem,unsigned ring,uint32_t flags) +static void test_busy (int i915,int vgem,const struct intel_execution_engine2 *e) {struct vgem_bo scratch;struct timespec tv;
@@ -434,7 +436,7 @@ static void test_busy(int i915, int vgem, unsigned ring, uint32_t flags)vgem_create (vgem, &scratch); dmabuf =prime_handle_to_fd (vgem, scratch.handle); -work (i915, dmabuf, ring, flags); +work (i915, dmabuf, e);/* Calling busy in a loop should be enough to flush the rendering */ memset (&tv,0 ,sizeof (tv));
@@ -450,7 +452,8 @@ static void test_busy(int i915, int vgem, unsigned ring, uint32_t flags)close (dmabuf); } -static void test_wait (int i915,int vgem,unsigned ring,uint32_t flags) +static void test_wait (int i915,int vgem, +const struct intel_execution_engine2 *e) {struct vgem_bo scratch;struct pollfd pfd;
@@ -463,7 +466,7 @@ static void test_wait(int i915, int vgem, unsigned ring, uint32_t flags)vgem_create (vgem, &scratch); pfd.fd =prime_handle_to_fd (vgem, scratch.handle); -work (i915, pfd.fd, ring, flags); +work (i915, pfd.fd, e); pfd.events =POLLIN ;igt_assert_eq (poll (&pfd,1 ,10000 ),1 );
@@ -477,7 +480,7 @@ static void test_wait(int i915, int vgem, unsigned ring, uint32_t flags)close (pfd.fd); } -static void test_sync (int i915,int vgem,unsigned ring,uint32_t flags) +static void test_sync (int i915,int vgem,const struct intel_execution_engine2 *e) {struct vgem_bo scratch;uint32_t *ptr;
@@ -494,7 +497,7 @@ static void test_sync(int i915, int vgem, unsigned ring, uint32_t flags)igt_assert (ptr !=MAP_FAILED );gem_close (vgem, scratch.handle); -work (i915, dmabuf, ring, flags); +work (i915, dmabuf, e);prime_sync_start (dmabuf, false);for (i =0 ; i <1024 ; i++)
@@ -506,7 +509,8 @@ static void test_sync(int i915, int vgem, unsigned ring, uint32_t flags)munmap (ptr, scratch.size); } -static void test_fence_wait (int i915,int vgem,unsigned ring,unsigned flags) +static void test_fence_wait (int i915,int vgem, +const struct intel_execution_engine2 *e) {struct vgem_bo scratch;uint32_t fence;
@@ -527,7 +531,7 @@ static void test_fence_wait(int i915, int vgem, unsigned ring, unsigned flags)igt_assert (ptr !=MAP_FAILED ); igt_fork(child, 1) -work (i915, dmabuf, ring, flags); +work (i915, dmabuf, e);sleep (1 );
@@ -568,7 +572,7 @@ static void test_fence_hang(int i915, int vgem, unsigned flags)igt_assert (ptr !=MAP_FAILED );gem_close (vgem, scratch.handle); -work (i915, dmabuf,I915_EXEC_DEFAULT ,0 ); +work (i915, dmabuf,NULL );/* The work should have been cancelled */
@@ -811,7 +815,7 @@ static void test_flip(int i915, int vgem, unsigned hang) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int i915 =-1 ;int vgem =-1 ;
@@ -840,39 +844,32 @@ igt_main igt_subtest("coherency-gtt")test_gtt_interleaved (vgem, i915); -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (i915, e) {igt_subtest_f ("%ssync-%s" , - e->exec_id ==0 ?"basic-" :"" , - e->name) { -gem_require_ring (i915, e->exec_id | e->flags); -igt_require (gem_can_store_dword (i915, e->exec_id | e->flags)); + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", e->name) { +igt_require (gem_class_can_store_dword (i915, e->class));gem_quiescent_gpu (i915); -test_sync (i915, vgem, e->exec_id, e->flags); +test_sync (i915, vgem, e); } - } -for (e = intel_execution_engines; e->name; e++) {igt_subtest_f ("%sbusy-%s" , - e->exec_id ==0 ?"basic-" :"" , - e->name) { -gem_require_ring (i915, e->exec_id | e->flags); -igt_require (gem_can_store_dword (i915, e->exec_id | e->flags)); + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", e->name) { +igt_require (gem_can_store_dword (i915, e->class));gem_quiescent_gpu (i915); -test_busy (i915, vgem, e->exec_id, e->flags); +test_busy (i915, vgem, e); } - } -for (e = intel_execution_engines; e->name; e++) {igt_subtest_f ("%swait-%s" , - e->exec_id ==0 ?"basic-" :"" , - e->name) { -gem_require_ring (i915, e->exec_id | e->flags); -igt_require (gem_can_store_dword (i915, e->exec_id | e->flags)); + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", e->name) { +igt_require (gem_can_store_dword (i915, e->class));gem_quiescent_gpu (i915); -test_wait (i915, vgem, e->exec_id, e->flags); +test_wait (i915, vgem, e); } }
@@ -887,15 +884,14 @@ igt_main igt_subtest("basic-fence-mmap")test_fence_mmap (i915, vgem); -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (i915, e) {igt_subtest_f ("%sfence-wait-%s" , - e->exec_id ==0 ?"basic-" :"" , - e->name) { -gem_require_ring (i915, e->exec_id | e->flags); -igt_require (gem_can_store_dword (i915, e->exec_id | e->flags)); + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", e->name) { +igt_require (gem_can_store_dword (i915, e->class));gem_quiescent_gpu (i915); -test_fence_wait (i915, vgem, e->exec_id, e->flags); +test_fence_wait (i915, vgem, e); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 06/43] test/prime_vgem: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:08)
Quoted 130 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.v2:
helper for getting theexecbuf_flagfrom class is used.Signed-off-by Ramalingam C <[email protected]>--- tests/prime_vgem.c | 72 ++++++++++++++++++++++------------------------ 1 file changed, 34 insertions(+), 38 deletions(-)tests/prime_vgem.c +16
-12@@ -325,7 +325,7 @@ static bool prime_busy(int fd, bool excl)return poll (&pfd,1 ,0 ) ==0 ; } -static void work (int i915,int dmabuf,unsigned ring,uint32_t flags) +static void work (int i915,int dmabuf,const struct intel_execution_engine2 *e) {const int SCRATCH =0 ;const int BATCH =1 ;@@ -341,7 +341,9 @@ static void work(int i915, int dmabuf, unsigned ring, uint32_t flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )obj; execbuf.buffer_count =2 ; - execbuf.flags = ring | flags; + execbuf.flags = e ? e->flags : +gem_engine_class_to_execbuf_flags (i915,0 , +I915_ENGINE_CLASS_RENDER );if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;@@ -420,7 +422,7 @@ static void work(int i915, int dmabuf, unsigned ring, uint32_t flags)igt_assert (read_busy && write_busy); } -static void test_busy (int i915,int vgem,unsigned ring,uint32_t flags) +static void test_busy (int i915,int vgem,const struct intel_execution_engine2 *e) {struct vgem_bo scratch;struct timespec tv;@@ -434,7 +436,7 @@ static void test_busy(int i915, int vgem, unsigned ring, uint32_t flags)vgem_create (vgem, &scratch); dmabuf =prime_handle_to_fd (vgem, scratch.handle); -work (i915, dmabuf, ring, flags); +work (i915, dmabuf, e);/* Calling busy in a loop should be enough to flush the rendering */ memset (&tv,0 ,sizeof (tv));@@ -450,7 +452,8 @@ static void test_busy(int i915, int vgem, unsigned ring, uint32_t flags)close (dmabuf); } -static void test_wait (int i915,int vgem,unsigned ring,uint32_t flags) +static void test_wait (int i915,int vgem, +const struct intel_execution_engine2 *e) {struct vgem_bo scratch;struct pollfd pfd;@@ -463,7 +466,7 @@ static void test_wait(int i915, int vgem, unsigned ring, uint32_t flags)vgem_create (vgem, &scratch); pfd.fd =prime_handle_to_fd (vgem, scratch.handle); -work (i915, pfd.fd, ring, flags); +work (i915, pfd.fd, e); pfd.events =POLLIN ;igt_assert_eq (poll (&pfd,1 ,10000 ),1 );@@ -477,7 +480,7 @@ static void test_wait(int i915, int vgem, unsigned ring, uint32_t flags)close (pfd.fd); } -static void test_sync (int i915,int vgem,unsigned ring,uint32_t flags) +static void test_sync (int i915,int vgem,const struct intel_execution_engine2 *e) {struct vgem_bo scratch;uint32_t *ptr;@@ -494,7 +497,7 @@ static void test_sync(int i915, int vgem, unsigned ring, uint32_t flags)igt_assert (ptr !=MAP_FAILED );gem_close (vgem, scratch.handle); -work (i915, dmabuf, ring, flags); +work (i915, dmabuf, e);prime_sync_start (dmabuf, false);for (i =0 ; i <1024 ; i++)@@ -506,7 +509,8 @@ static void test_sync(int i915, int vgem, unsigned ring, uint32_t flags)munmap (ptr, scratch.size); } -static void test_fence_wait (int i915,int vgem,unsigned ring,unsigned flags) +static void test_fence_wait (int i915,int vgem, +const struct intel_execution_engine2 *e) {struct vgem_bo scratch;uint32_t fence;@@ -527,7 +531,7 @@ static void test_fence_wait(int i915, int vgem, unsigned ring, unsigned flags)igt_assert (ptr !=MAP_FAILED ); igt_fork(child, 1) -work (i915, dmabuf, ring, flags); +work (i915, dmabuf, e);sleep (1 );@@ -568,7 +572,7 @@ static void test_fence_hang(int i915, int vgem, unsigned flags)igt_assert (ptr !=MAP_FAILED );gem_close (vgem, scratch.handle); -work (i915, dmabuf,I915_EXEC_DEFAULT ,0 ); +work (i915, dmabuf,NULL );/* The work should have been cancelled */ @@ -811,7 +815,7 @@ static void test_flip(int i915, int vgem, unsigned hang) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int i915 =-1 ;int vgem =-1 ;@@ -840,39 +844,32 @@ igt_main igt_subtest("coherency-gtt")test_gtt_interleaved (vgem, i915);
tests/prime_vgem.c igt_main · line 850
Quoted 10 lines
tests/prime_vgem.c +3
-5-for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (i915, e) {igt_subtest_f ("%ssync-%s" , - e->exec_id ==0 ?"basic-" :"" , - e->name) { -gem_require_ring (i915, e->exec_id | e->flags); -igt_require (gem_can_store_dword (i915, e->exec_id | e->flags)); + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", e->name) {
Nak. Please restore this to using EXEC_DEFAULT explicitly. -Chris
_______________________________________________
igt-dev mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 7/43 tests/i915/gem_exec_reuse: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_reuse.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)
tests/i915/gem_exec_reuse.c +6 -6
@@ -39,7 +39,7 @@ struct noop { };static void noop (struct noop *n, -unsigned ring,unsigned ctx, +unsigned exec_flags,unsigned ctx,unsigned int count,unsigned int offset) {struct drm_i915_gem_execbuffer2 execbuf;
@@ -52,7 +52,7 @@ static void noop(struct noop *n,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (n->obj); execbuf.buffer_count = count +1 ; - execbuf.flags = ring |1 <<12 ; + execbuf.flags = exec_flags |1 <<12 ; execbuf.rsvd1 = ctx;gem_execbuf (n->fd, &execbuf); }
@@ -89,7 +89,7 @@ igt_main igt_fixture {uint64_t gtt_size, max;uint32_t bbe =MI_BATCH_BUFFER_END ; -unsigned engine; +const struct intel_execution_engine2 *e;igt_allow_unlimited_files ();
@@ -118,9 +118,9 @@ igt_main no.obj =malloc (sizeof (struct drm_i915_gem_exec_object2 ) * (no.max_age +1 )); nengine =0 ; -for_each_engine (no.fd, engine) -if (engine) - engines[nengine++] = engine; +__for_each_static_engine (e) +if (e) + engines[nengine++] = e->flags;igt_require (nengine); no.batch =gem_create (no.fd,4096 );
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 8/43 tests/i915/gem_sync: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_sync.c | 234 +++++++++++++++++++++--------------------- 1 file changed, 118 insertions(+), 116 deletions(-)
tests/i915/gem_sync.c +118 -116
@@ -79,25 +79,26 @@ out: }static void -sync_ring (int fd,unsigned ring,int num_children,int timeout) +sync_ring (int fd,const struct intel_execution_engine2 *e,int num_children, +int timeout) {unsigned engines[16 ];const char *names[16 ];int num_engines =0 ; -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { - names[num_engines] = e__->name; - engines[num_engines++] = ring; +if (!e) { +__for_each_physical_engine (fd, e) { + names[num_engines] = e->name; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; } num_children *= num_engines; }else { -gem_require_ring (fd, ring); +gem_require_ring (fd, e->flags); names[num_engines] =NULL ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags; }intel_detect_and_clear_missed_interrupts (fd);
@@ -139,7 +140,7 @@ sync_ring(int fd, unsigned ring, int num_children, int timeout) }static void -idle_ring (int fd,unsigned ring,int timeout) +idle_ring (int fd,const struct intel_execution_engine2 *e,int timeout) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_exec_object2 object;
@@ -147,7 +148,7 @@ idle_ring(int fd, unsigned ring, int timeout)double start, elapsed;unsigned long cycles; -gem_require_ring (fd, ring); +gem_require_ring (fd, e->flags);memset (&object,0 ,sizeof (object)); object.handle =gem_create (fd,4096 );
@@ -156,7 +157,7 @@ idle_ring(int fd, unsigned ring, int timeout)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (&object); execbuf.buffer_count =1 ; - execbuf.flags = ring; + execbuf.flags = e->flags;gem_execbuf (fd, &execbuf);gem_sync (fd, object.handle);
@@ -178,28 +179,29 @@ idle_ring(int fd, unsigned ring, int timeout) }static void -wakeup_ring (int fd,unsigned ring,int timeout,int wlen) +wakeup_ring (int fd,const struct intel_execution_engine2 *e,int timeout, +int wlen) {unsigned engines[16 ];const char *names[16 ];int num_engines =0 ; -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { -if (!gem_can_store_dword (fd, ring)) +if (!e) { +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - names[num_engines] = e__->name; - engines[num_engines++] = ring; + names[num_engines] = e->name; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; }igt_require (num_engines); }else { -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); names[num_engines] =NULL ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags; }intel_detect_and_clear_missed_interrupts (fd);
@@ -221,7 +223,7 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen) execbuf.flags = engines[child % num_engines]; spin = __igt_spin_new(fd, - .engine = execbuf.flags, + .engine = engines[child % num_engines], .flags = (IGT_SPIN_POLL_RUN |IGT_SPIN_FAST ));igt_assert (igt_spin_has_poll (spin));
@@ -290,28 +292,28 @@ wakeup_ring(int fd, unsigned ring, int timeout, int wlen)igt_assert_eq (intel_detect_and_clear_missed_interrupts (fd),0 ); } -static void active_ring (int fd,unsigned ring,int timeout) +static void active_ring (int fd,const struct intel_execution_engine2 *e,int timeout) {unsigned engines[16 ];const char *names[16 ];int num_engines =0 ; -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { -if (!gem_can_store_dword (fd, ring)) +if (!e) { +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - names[num_engines] = e__->name; - engines[num_engines++] = ring; + names[num_engines] = e->name; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; }igt_require (num_engines); }else { -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); names[num_engines] =NULL ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags; }intel_detect_and_clear_missed_interrupts (fd);
@@ -321,11 +323,11 @@ static void active_ring(int fd, unsigned ring, int timeout)igt_spin_t *spin[2 ]; spin[0] = __igt_spin_new(fd, - .engine = ring, + .engine = e->flags, .flags =IGT_SPIN_FAST ); spin[1] = __igt_spin_new(fd, - .engine = ring, + .engine = e->flags, .flags =IGT_SPIN_FAST ); start =gettime ();
@@ -357,28 +359,29 @@ static void active_ring(int fd, unsigned ring, int timeout) }static void -active_wakeup_ring (int fd,unsigned ring,int timeout,int wlen) +active_wakeup_ring (int fd,const struct intel_execution_engine2 *e,int timeout, +int wlen) {unsigned engines[16 ];const char *names[16 ];int num_engines =0 ; -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { -if (!gem_can_store_dword (fd, ring)) +if (!e) { +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - names[num_engines] = e__->name; - engines[num_engines++] = ring; + names[num_engines] = e->name; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; }igt_require (num_engines); }else { -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); names[num_engines] =NULL ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags; }intel_detect_and_clear_missed_interrupts (fd);
@@ -400,13 +403,13 @@ active_wakeup_ring(int fd, unsigned ring, int timeout, int wlen) execbuf.flags = engines[child % num_engines]; spin[0] = __igt_spin_new(fd, - .engine = execbuf.flags, + .engine = engines[child % num_engines], .flags = (IGT_SPIN_POLL_RUN |IGT_SPIN_FAST ));igt_assert (igt_spin_has_poll (spin[0 ])); spin[1] = __igt_spin_new(fd, - .engine = execbuf.flags, + .engine = engines[child % num_engines], .flags = (IGT_SPIN_POLL_RUN |IGT_SPIN_FAST ));
@@ -491,30 +494,30 @@ active_wakeup_ring(int fd, unsigned ring, int timeout, int wlen) }static void -store_ring (int fd,unsigned ring,int num_children,int timeout) +store_ring (int fd,const struct intel_execution_engine2 *e,int num_children,int timeout) {const int gen =intel_gen (intel_get_drm_devid (fd));unsigned engines[16 ];const char *names[16 ];int num_engines =0 ; -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { -if (!gem_can_store_dword (fd, ring)) +if (!e) { +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - names[num_engines] = e__->name; - engines[num_engines++] = ring; + names[num_engines] = e->name; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; } num_children *= num_engines; }else { -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); names[num_engines] =NULL ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags; }intel_detect_and_clear_missed_interrupts (fd);
@@ -606,7 +609,7 @@ store_ring(int fd, unsigned ring, int num_children, int timeout) }static void -switch_ring (int fd,unsigned ring,int num_children,int timeout) +switch_ring (int fd,const struct intel_execution_engine2 *e,int num_children,int timeout) {const int gen =intel_gen (intel_get_drm_devid (fd));unsigned engines[16 ];
@@ -615,23 +618,23 @@ switch_ring(int fd, unsigned ring, int num_children, int timeout)gem_require_contexts (fd); -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { -if (!gem_can_store_dword (fd, ring)) +if (!e) { +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - names[num_engines] = e__->name; - engines[num_engines++] = ring; + names[num_engines] = e->name; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; } num_children *= num_engines; }else { -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); names[num_engines] =NULL ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags; }intel_detect_and_clear_missed_interrupts (fd);
@@ -803,7 +806,7 @@ static void *waiter(void *arg) }static void -__store_many (int fd,unsigned ring,int timeout,unsigned long *cycles) +__store_many (int fd,const struct intel_execution_engine2 *e,int timeout,unsigned long *cycles) {const int gen =intel_gen (intel_get_drm_devid (fd));const uint32_t bbe =MI_BATCH_BUFFER_END ;
@@ -817,7 +820,7 @@ __store_many(int fd, unsigned ring, int timeout, unsigned long *cycles)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (object); - execbuf.flags = ring; + execbuf.flags = e->flags; execbuf.flags |=LOCAL_I915_EXEC_NO_RELOC ; execbuf.flags |=LOCAL_I915_EXEC_HANDLE_LUT ;if (gen <6 )
@@ -931,7 +934,7 @@ __store_many(int fd, unsigned ring, int timeout, unsigned long *cycles) }static void -store_many (int fd,unsigned ring,int timeout) +store_many (int fd,const struct intel_execution_engine2 *e,int timeout) {unsigned long *shared;const char *names[16 ];
@@ -942,24 +945,22 @@ store_many(int fd, unsigned ring, int timeout)intel_detect_and_clear_missed_interrupts (fd); -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { -if (!gem_can_store_dword (fd, ring)) +if (!e) { +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; igt_fork(child, 1) -__store_many (fd, - ring, - timeout, +__store_many (fd, e, timeout, &shared[n]); - names[n++] = e__->name; + names[n++] = e->name; }igt_waitchildren (); }else { -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); -__store_many (fd, ring, timeout, &shared[n]); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); +__store_many (fd, e, timeout, &shared[n]); names[n++] =NULL ; }
@@ -974,11 +975,12 @@ store_many(int fd, unsigned ring, int timeout)static void sync_all (int fd,int num_children,int timeout) { -unsigned engines[16 ], engine; +unsigned engines[16 ]; +const struct intel_execution_engine2 *engine;int num_engines =0 ; -for_each_physical_engine (fd, engine) { - engines[num_engines++] = engine; +__for_each_physical_engine (fd, engine) { + engines[num_engines++] = engine->flags;if (num_engines ==ARRAY_SIZE (engines))break ; }
@@ -1028,13 +1030,13 @@ store_all(int fd, int num_children, int timeout)const int gen =intel_gen (intel_get_drm_devid (fd));unsigned engines[16 ];int num_engines =0 ; -unsigned int ring; +const struct intel_execution_engine2 *e; -for_each_physical_engine (fd, ring) { -if (!gem_can_store_dword (fd, ring)) +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; }
@@ -1131,26 +1133,26 @@ store_all(int fd, int num_children, int timeout) }static void -preempt (int fd,unsigned ring,int num_children,int timeout) +preempt (int fd,const struct intel_execution_engine2 *e,int num_children,int timeout) {unsigned engines[16 ];const char *names[16 ];int num_engines =0 ;uint32_t ctx[2 ]; -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, ring) { - names[num_engines] = e__->name; - engines[num_engines++] = ring; +if (!e) { +__for_each_physical_engine (fd, e) { + names[num_engines] = e->name; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; } num_children *= num_engines; }else { -gem_require_ring (fd, ring); +gem_require_ring (fd, e->flags); names[num_engines] =NULL ; - engines[num_engines++] = ring; + engines[num_engines++] = e->flags; } ctx[0 ] =gem_context_create (fd);
@@ -1185,7 +1187,7 @@ preempt(int fd, unsigned ring, int num_children, int timeout)igt_spin_t *spin = __igt_spin_new(fd, .ctx = ctx[0], - .engine = execbuf.flags); + .engine = engines[child % num_engines]); do { gem_execbuf(fd, &execbuf);
@@ -1210,7 +1212,7 @@ preempt(int fd, unsigned ring, int num_children, int timeout) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;const int ncpus =sysconf (_SC_NPROCESSORS_ONLN);int fd =-1 ;
@@ -1225,55 +1227,55 @@ igt_mainigt_fork_hang_detector (fd); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) { igt_subtest_f("%s", e->name) -sync_ring (fd, e->exec_id | e->flags,1 ,150 ); +sync_ring (fd, e,1 ,150 ); igt_subtest_f("idle-%s", e->name) -idle_ring (fd, e->exec_id | e->flags,150 ); +idle_ring (fd, e,150 ); igt_subtest_f("active-%s", e->name) -active_ring (fd, e->exec_id | e->flags,150 ); +active_ring (fd, e,150 ); igt_subtest_f("wakeup-%s", e->name) -wakeup_ring (fd, e->exec_id | e->flags,150 ,1 ); +wakeup_ring (fd, e,150 ,1 ); igt_subtest_f("active-wakeup-%s", e->name) -active_wakeup_ring (fd, e->exec_id | e->flags,150 ,1 ); +active_wakeup_ring (fd, e,150 ,1 ); igt_subtest_f("double-wakeup-%s", e->name) -wakeup_ring (fd, e->exec_id | e->flags,150 ,2 ); +wakeup_ring (fd, e,150 ,2 ); igt_subtest_f("store-%s", e->name) -store_ring (fd, e->exec_id | e->flags,1 ,150 ); +store_ring (fd, e,1 ,150 ); igt_subtest_f("switch-%s", e->name) -switch_ring (fd, e->exec_id | e->flags,1 ,150 ); +switch_ring (fd, e,1 ,150 ); igt_subtest_f("forked-switch-%s", e->name) -switch_ring (fd, e->exec_id | e->flags, ncpus,150 ); +switch_ring (fd, e, ncpus,150 ); igt_subtest_f("many-%s", e->name) -store_many (fd, e->exec_id | e->flags,150 ); +store_many (fd, e,150 ); igt_subtest_f("forked-%s", e->name) -sync_ring (fd, e->exec_id | e->flags, ncpus,150 ); +sync_ring (fd, e, ncpus,150 ); igt_subtest_f("forked-store-%s", e->name) -store_ring (fd, e->exec_id | e->flags, ncpus,150 ); +store_ring (fd, e, ncpus,150 ); } igt_subtest("basic-each") -sync_ring (fd,ALL_ENGINES ,1 ,5 ); +sync_ring (fd,NULL ,1 ,5 ); igt_subtest("basic-store-each") -store_ring (fd,ALL_ENGINES ,1 ,5 ); +store_ring (fd,NULL ,1 ,5 ); igt_subtest("basic-many-each") -store_many (fd,ALL_ENGINES ,5 ); +store_many (fd,NULL ,5 ); igt_subtest("switch-each") -switch_ring (fd,ALL_ENGINES ,1 ,150 ); +switch_ring (fd,NULL ,1 ,150 ); igt_subtest("forked-switch-each") -switch_ring (fd,ALL_ENGINES , ncpus,150 ); +switch_ring (fd,NULL , ncpus,150 ); igt_subtest("forked-each") -sync_ring (fd,ALL_ENGINES , ncpus,150 ); +sync_ring (fd,NULL , ncpus,150 ); igt_subtest("forked-store-each") -store_ring (fd,ALL_ENGINES , ncpus,150 ); +store_ring (fd,NULL , ncpus,150 ); igt_subtest("active-each") -active_ring (fd,ALL_ENGINES ,150 ); +active_ring (fd,NULL ,150 ); igt_subtest("wakeup-each") -wakeup_ring (fd,ALL_ENGINES ,150 ,1 ); +wakeup_ring (fd,NULL ,150 ,1 ); igt_subtest("active-wakeup-each") -active_wakeup_ring (fd,ALL_ENGINES ,150 ,1 ); +active_wakeup_ring (fd,NULL ,150 ,1 ); igt_subtest("double-wakeup-each") -wakeup_ring (fd,ALL_ENGINES ,150 ,2 ); +wakeup_ring (fd,NULL ,150 ,2 ); igt_subtest("basic-all")sync_all (fd,1 ,5 );
@@ -1297,11 +1299,11 @@ igt_main } igt_subtest("preempt-all") -preempt (fd,ALL_ENGINES ,1 ,20 ); +preempt (fd,NULL ,1 ,20 ); -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) { igt_subtest_f("preempt-%s", e->name) -preempt (fd, e->exec_id | e->flags, ncpus,150 ); +preempt (fd, e, ncpus,150 ); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 9/43 tests/i915/gem_busy: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_busy.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
tests/i915/gem_busy.c +3 -2
@@ -305,6 +305,7 @@ static void close_race(int fd) {const unsigned int ncpus =sysconf (_SC_NPROCESSORS_ONLN);const unsigned int nhandles =gem_measure_ring_inflight (fd,ALL_ENGINES ,0 ) /2 ; +const struct intel_execution_engine2 *e;unsigned int engines[16 ], nengine;unsigned long *control;uint32_t *handles;
@@ -319,8 +320,8 @@ static void close_race(int fd)*/ nengine =0 ; -for_each_engine (fd, i) - engines[nengine++] = i; +__for_each_static_engine (e) + engines[nengine++] = e->flags;igt_require (nengine); control =mmap (NULL ,4096 ,PROT_WRITE ,MAP_SHARED |MAP_ANON ,-1 ,0 );
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 10/43 tests/i915/gem_concurrent_all: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_concurrent_all.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
tests/i915/gem_concurrent_all.c +4 -4
@@ -949,15 +949,15 @@ static igt_hang_t rcs_hang(void)static igt_hang_t all_hang (void ) {igt_hang_t hang =igt_hang_ring (fd,I915_EXEC_RENDER ); -unsigned engine; +const struct intel_execution_engine2 *e; -for_each_physical_engine (fd, engine) { +__for_each_physical_engine (fd, e) {struct drm_i915_gem_execbuffer2 eb = hang.spin->execbuf; -if (engine ==I915_EXEC_RENDER ) +if (e->class ==I915_ENGINE_CLASS_RENDER )continue ; - eb.flags = engine; + eb.flags = e->flags;__gem_execbuf (fd, &eb); }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 11/43 tests/i915/gem_cs_prefetch: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_cs_prefetch.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-)
tests/i915/gem_cs_prefetch.c +9 -9
@@ -79,7 +79,7 @@ static void setup(int fd, int gen, struct shadow *shadow) shadow->reloc.write_domain =I915_GEM_DOMAIN_INSTRUCTION ; } -static void can_test_ring (unsigned ring) +static void can_test_ring (const struct intel_execution_engine2 *e) {int master =drm_open_driver_master (DRIVER_INTEL );int fd =drm_open_driver (DRIVER_INTEL );
@@ -87,12 +87,12 @@ static void can_test_ring(unsigned ring)/* Dance to avoid dying with master open */ close (master);igt_require_gem (fd); -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class));close (fd); } -static void test_ring (unsigned ring) +static void test_ring (const struct intel_execution_engine2 *e) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj[2 ];
@@ -100,7 +100,7 @@ static void test_ring(unsigned ring)uint64_t i, count;int fd, gen; -can_test_ring (ring); +can_test_ring (e); fd =drm_open_driver_master (DRIVER_INTEL ); gen =intel_gen (intel_get_drm_devid (fd));
@@ -116,7 +116,7 @@ static void test_ring(unsigned ring)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); - execbuf.flags = ring; + execbuf.flags = e->flags;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -139,11 +139,11 @@ static void test_ring(unsigned ring) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;igt_skip_on_simulation (); -for (e = intel_execution_engines; e->name; e++) +__for_each_static_engine (e) igt_subtest_f("%s", e->name) -test_ring (e->exec_id | e->flags); +test_ring (e); }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 11/43] tests/i915/gem_cs_prefetch: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:13)
Quoted 56 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.Signed-off-by Ramalingam C <[email protected]>--- tests/i915/gem_cs_prefetch.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-)tests/i915/gem_cs_prefetch.c +6
-6@@ -79,7 +79,7 @@ static void setup(int fd, int gen, struct shadow *shadow) shadow->reloc.write_domain =I915_GEM_DOMAIN_INSTRUCTION ; } -static void can_test_ring (unsigned ring) +static void can_test_ring (const struct intel_execution_engine2 *e) {int master =drm_open_driver_master (DRIVER_INTEL );int fd =drm_open_driver (DRIVER_INTEL );@@ -87,12 +87,12 @@ static void can_test_ring(unsigned ring)/* Dance to avoid dying with master open */ close (master);igt_require_gem (fd); -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class));close (fd); } -static void test_ring (unsigned ring) +static void test_ring (const struct intel_execution_engine2 *e) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj[2 ];@@ -100,7 +100,7 @@ static void test_ring(unsigned ring)uint64_t i, count;int fd, gen; -can_test_ring (ring); +can_test_ring (e); fd =drm_open_driver_master (DRIVER_INTEL ); gen =intel_gen (intel_get_drm_devid (fd));@@ -116,7 +116,7 @@ static void test_ring(unsigned ring)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); - execbuf.flags = ring; + execbuf.flags = e->flags;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
tests/i915/gem_cs_prefetch.c static void test_ring(unsigned ring) · line 146
Quoted 10 lines
tests/i915/gem_cs_prefetch.c +2
-2@@ -139,11 +139,11 @@ static void test_ring(unsigned ring) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;igt_skip_on_simulation (); -for (e = intel_execution_engines; e->name; e++) +__for_each_static_engine (e)
Now this is exercising a bit of HW lore, so could be phys. -Chris
_______________________________________________
igt-dev mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 12/43 tests/i915/gem_ctx_create: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_ctx_create.c | 34 +++++++++++++++------------------- 1 file changed, 15 insertions(+), 19 deletions(-)
tests/i915/gem_ctx_create.c +15 -19
@@ -124,7 +124,8 @@ static void files(int core, int timeout, const int ncpus)gem_close (core, batch); } -static void active (int fd,unsigned engine,int timeout,int ncpus) +static void active (int fd,const struct intel_execution_engine2 *e, +int timeout,int ncpus) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_execbuffer2 execbuf;
@@ -132,14 +133,13 @@ static void active(int fd, unsigned engine, int timeout, int ncpus)unsigned int nengine, engines[16 ];unsigned *shared; -if (engine ==ALL_ENGINES ) { +if (!e) {igt_require (all_nengine); nengine = all_nengine;memcpy (engines, all_engines,sizeof (engines[0 ])*nengine); }else { -gem_require_ring (fd, engine); nengine =1 ; - engines[0 ] = engine; + engines[0 ] = e->flags; } shared =mmap (NULL ,4096 ,PROT_WRITE ,MAP_SHARED |MAP_ANON ,-1 ,0 );
@@ -157,7 +157,7 @@ static void active(int fd, unsigned engine, int timeout, int ncpus) igt_fork(child, ppgtt_nengine) {unsigned long count =0 ; -if (ppgtt_engines[child] == engine) +if (ppgtt_engines[child] == e->flags)continue ; execbuf.flags = ppgtt_engines[child];
@@ -524,17 +524,16 @@ igt_main {const int ncpus =sysconf (_SC_NPROCESSORS_ONLN);struct drm_i915_gem_context_create create; +const struct intel_execution_engine2 *e;int fd =-1 ; igt_fixture { -unsigned engine; - fd =drm_open_driver (DRIVER_INTEL );igt_require_gem (fd);gem_require_contexts (fd); -for_each_physical_engine (fd, engine) - all_engines[all_nengine++] = engine; +__for_each_physical_engine (fd, e) + all_engines[all_nengine++] = e->flags;igt_require (all_nengine);if (gem_uses_full_ppgtt (fd)) {
@@ -575,20 +574,17 @@ igt_mainfiles (fd,150 , ncpus); igt_subtest("active-all") -active (fd,ALL_ENGINES ,120 ,1 ); +active (fd,NULL ,120 ,1 ); igt_subtest("forked-active-all") -active (fd,ALL_ENGINES ,120 , ncpus); +active (fd,NULL ,120 , ncpus); -for (const struct intel_execution_engine *e = intel_execution_engines; - e->name; e++) { +__for_each_physical_engine (fd, e) { igt_subtest_f("active-%s", e->name) -active (fd, e->exec_id | e->flags,20 ,1 ); +active (fd, e,20 ,1 ); igt_subtest_f("forked-active-%s", e->name) -active (fd, e->exec_id | e->flags,20 , ncpus); -if (e->exec_id) { - igt_subtest_f("hog-%s", e->name) -active (fd, e->exec_id | e->flags,20 ,-1 ); - } +active (fd, e,20 , ncpus); + igt_subtest_f("hog-%s", e->name) +active (fd, e,20 ,-1 ); } igt_fixture {
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 13/43 tests/i915/gem_ctx_isolation: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_ctx_isolation.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-)
tests/i915/gem_ctx_isolation.c +8 -11
@@ -556,7 +556,6 @@ static void nonpriv(int fd,0x0505c0c0 , 0xdeadbeef }; -unsigned int engine = e->flags;unsigned int num_values =ARRAY_SIZE (values);/* Sigh -- hsw: we need cmdparser access to our own registers! */
@@ -574,7 +573,7 @@ static void nonpriv(int fd,tmpl_regs (fd, ctx, e, tmpl, values[v]); - spin = igt_spin_new(fd, .ctx = ctx, .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx, .engine = e->flags);igt_debug ("%s[%d]: Setting all registers to 0x%08x\n" , __func__, v, values[v]);
@@ -612,7 +611,6 @@ static void isolation(int fd,0xaaaaaaaa , 0xdeadbeef }; -unsigned int engine = e->flags;unsigned int num_values = flags & (DIRTY1 |DIRTY2 ) ?ARRAY_SIZE (values) :1 ;
@@ -625,7 +623,7 @@ static void isolation(int fd, ctx[0 ] =gem_context_create (fd); regs[0 ] =read_regs (fd, ctx[0 ], e, flags); - spin = igt_spin_new(fd, .ctx = ctx[0], .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx[0], .engine = e->flags);if (flags &DIRTY1 ) { igt_debug("%s[%d]: Setting all registers of ctx 0 to 0x%08x\n",
@@ -678,11 +676,11 @@ static void isolation(int fd,#define S4 (4 << 8)#define SLEEP_MASK (0xf << 8) -static void inject_reset_context (int fd,unsigned int engine) +static void inject_reset_context (int fd,const struct intel_execution_engine2 *e) {struct igt_spin_factory opts = { .ctx =gem_context_create (fd), - .engine = engine, + .engine = e->flags, .flags =IGT_SPIN_FAST , };igt_spin_t *spin;
@@ -693,7 +691,7 @@ static void inject_reset_context(int fd, unsigned int engine)* HW for screwing up if the context was already broken. */ -if (gem_can_store_dword (fd, engine)) +if (gem_class_can_store_dword (fd, e->class)) opts.flags |=IGT_SPIN_POLL_RUN ; spin =__igt_spin_factory (fd, &opts);
@@ -723,7 +721,6 @@ static void preservation(int fd, 0xdeadbeef };const unsigned int num_values =ARRAY_SIZE (values); -unsigned int engine = e->flags;uint32_t ctx[num_values +1 ];uint32_t regs[num_values +1 ][2 ];igt_spin_t *spin;
@@ -731,7 +728,7 @@ static void preservation(int fd,gem_quiescent_gpu (fd); ctx[num_values] =gem_context_create (fd); - spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = e->flags); regs[num_values][0 ] =read_regs (fd, ctx[num_values], e, flags);for (int v =0 ; v < num_values; v++) { ctx[v] =gem_context_create (fd);
@@ -744,7 +741,7 @@ static void preservation(int fd,igt_spin_free (fd, spin);if (flags &RESET ) -inject_reset_context (fd, engine); +inject_reset_context (fd, e);switch (flags &SLEEP_MASK ) {case NOSLEEP :
@@ -771,7 +768,7 @@ static void preservation(int fd,break ; } - spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = e->flags);for (int v =0 ; v < num_values; v++) regs[v][1 ] =read_regs (fd, ctx[v], e, flags); regs[num_values][1 ] =read_regs (fd, ctx[num_values], e, flags);
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 13/43] tests/i915/gem_ctx_isolation: use the gem_engine_topology library
For tests/i915/gem_ctx_isolation the changes you have made look good, but you will also need something like the following:
----------------------------------------------------------------
tests/i915/gem_ctx_isolation.c +2 -2
@@ -812,6 +812,7 @@ static unsigned int __has_context_isolation(int fd) igt_main {unsigned int has_context_isolation =0 ; +const struct intel_execution_engine2 *e;int fd =-1 ;uint16_t devid;
@@ -833,8 +834,7 @@ igt_mainigt_skip_on (gen >LAST_KNOWN_GEN ); } -for (const struct intel_execution_engine2 *e = intel_execution_engines2; - e->name; e++) { +__for_each_physical_engine (fd, e) { igt_subtest_group { igt_fixture {igt_require (has_context_isolation & (1 << e->class));
----------------------------------------------------------------
Quoted 119 lines · Ramalingam C
Date: Fri, 21 Jun 2019 15:33:15 +0530
From: Ramalingam C <[email protected]>
To: [email protected], [email protected],
[email protected]
Subject: [igt-dev] [RFC v2 13/43] tests/i915/gem_ctx_isolation: use the
gem_engine_topologylibrary
X-Mailer: git-send-email 2.19.1Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.Signed-off-by Ramalingam C <[email protected]>--- tests/i915/gem_ctx_isolation.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-)tests/i915/gem_ctx_isolation.c +8
-11@@ -556,7 +556,6 @@ static void nonpriv(int fd,0x0505c0c0 , 0xdeadbeef }; -unsigned int engine = e->flags;unsigned int num_values =ARRAY_SIZE (values);/* Sigh -- hsw: we need cmdparser access to our own registers! */ @@ -574,7 +573,7 @@ static void nonpriv(int fd,tmpl_regs (fd, ctx, e, tmpl, values[v]); - spin = igt_spin_new(fd, .ctx = ctx, .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx, .engine = e->flags);igt_debug ("%s[%d]: Setting all registers to 0x%08x\n" , __func__, v, values[v]);@@ -612,7 +611,6 @@ static void isolation(int fd,0xaaaaaaaa , 0xdeadbeef }; -unsigned int engine = e->flags;unsigned int num_values = flags & (DIRTY1 |DIRTY2 ) ?ARRAY_SIZE (values) :1 ;@@ -625,7 +623,7 @@ static void isolation(int fd, ctx[0 ] =gem_context_create (fd); regs[0 ] =read_regs (fd, ctx[0 ], e, flags); - spin = igt_spin_new(fd, .ctx = ctx[0], .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx[0], .engine = e->flags);if (flags &DIRTY1 ) { igt_debug("%s[%d]: Setting all registers of ctx 0 to 0x%08x\n",@@ -678,11 +676,11 @@ static void isolation(int fd,#define S4 (4 << 8)#define SLEEP_MASK (0xf << 8) -static void inject_reset_context (int fd,unsigned int engine) +static void inject_reset_context (int fd,const struct intel_execution_engine2 *e) {struct igt_spin_factory opts = { .ctx =gem_context_create (fd), - .engine = engine, + .engine = e->flags, .flags =IGT_SPIN_FAST , };igt_spin_t *spin;@@ -693,7 +691,7 @@ static void inject_reset_context(int fd, unsigned int engine)* HW for screwing up if the context was already broken. */ -if (gem_can_store_dword (fd, engine)) +if (gem_class_can_store_dword (fd, e->class)) opts.flags |=IGT_SPIN_POLL_RUN ; spin =__igt_spin_factory (fd, &opts);@@ -723,7 +721,6 @@ static void preservation(int fd, 0xdeadbeef };const unsigned int num_values =ARRAY_SIZE (values); -unsigned int engine = e->flags;uint32_t ctx[num_values +1 ];uint32_t regs[num_values +1 ][2 ];igt_spin_t *spin;@@ -731,7 +728,7 @@ static void preservation(int fd,gem_quiescent_gpu (fd); ctx[num_values] =gem_context_create (fd); - spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = e->flags); regs[num_values][0 ] =read_regs (fd, ctx[num_values], e, flags);for (int v =0 ; v < num_values; v++) { ctx[v] =gem_context_create (fd);@@ -744,7 +741,7 @@ static void preservation(int fd,igt_spin_free (fd, spin);if (flags &RESET ) -inject_reset_context (fd, engine); +inject_reset_context (fd, e);switch (flags &SLEEP_MASK ) {case NOSLEEP :@@ -771,7 +768,7 @@ static void preservation(int fd,break ; } - spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = engine); + spin = igt_spin_new(fd, .ctx = ctx[num_values], .engine = e->flags);for (int v =0 ; v < num_values; v++) regs[v][1 ] =read_regs (fd, ctx[v], e, flags); regs[num_values][1 ] =read_regs (fd, ctx[num_values], e, flags);Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 14/43 tests/i915/gem_ctx_switch: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_ctx_switch.c | 28 +++++++++++++++------------- 1 file changed, 15 insertions(+), 13 deletions(-)
tests/i915/gem_ctx_switch.c +15 -13
@@ -107,7 +107,7 @@ static int measure_qlen(int fd, }static void single (int fd,uint32_t handle, -const struct intel_execution_engine *e, +const struct intel_execution_engine2 *e,unsigned flags,const int ncpus,int timeout)
@@ -125,8 +125,6 @@ static void single(int fd, uint32_t handle, shared =mmap (NULL ,4096 ,PROT_WRITE ,MAP_SHARED |MAP_ANON ,-1 ,0 );igt_assert (shared !=MAP_FAILED ); -gem_require_ring (fd, e->exec_id | e->flags); -for (n =0 ; n <64 ; n++) {if (flags &QUEUE ) contexts[n] =gem_queue_create (fd);
@@ -152,12 +150,12 @@ static void single(int fd, uint32_t handle, execbuf.buffers_ptr =to_user_pointer (&obj); execbuf.buffer_count =1 ; execbuf.rsvd1 = contexts[0 ]; - execbuf.flags = e->exec_id | e->flags; + execbuf.flags = e->flags; execbuf.flags |=LOCAL_I915_EXEC_HANDLE_LUT ; execbuf.flags |=LOCAL_I915_EXEC_NO_RELOC ;igt_require (__gem_execbuf (fd, &execbuf) ==0 );if (__gem_execbuf (fd, &execbuf)) { - execbuf.flags = e->exec_id | e->flags; + execbuf.flags = e->flags; reloc.target_handle = obj.handle;gem_execbuf (fd, &execbuf); }
@@ -223,16 +221,17 @@ static void all(int fd, uint32_t handle, unsigned flags, int timeout) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj[2 ]; -unsigned int engine[16 ], e; +const struct intel_execution_engine2 *e; +unsigned int engine[16 ];const char *name[16 ];uint32_t contexts[65 ];unsigned int nengine;int n, qlen; nengine =0 ; -for_each_physical_engine (fd, e) { - engine[nengine] = e; - name[nengine] = e__->name; +__for_each_physical_engine (fd, e) { + engine[nengine] = e->flags; + name[nengine] = e->name; nengine++; }igt_require (nengine);
@@ -306,7 +305,7 @@ static void all(int fd, uint32_t handle, unsigned flags, int timeout) igt_main {const int ncpus =sysconf (_SC_NPROCESSORS_ONLN); -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e; static const struct { const char *name; unsigned int flags;
@@ -338,7 +337,7 @@ igt_mainigt_fork_hang_detector (fd); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) {for (typeof (*phases) *p = phases; p->name; p++) { igt_subtest_group { igt_fixture {
@@ -346,12 +345,15 @@ igt_mainigt_require (p->require (fd)); } - igt_subtest_f("%s%s%s", e->exec_id == 0 ? "basic-" : "", e->name, p->name) +igt_subtest_f ("%s%s%s" , e->class == +I915_ENGINE_CLASS_RENDER ? + "basic-" : "", e->name, p->name)single (fd, light, e, p->flags,1 ,5 );igt_skip_on_simulation (); - igt_subtest_f("%s%s-heavy%s", e->exec_id == 0 ? "basic-" : "", e->name, p->name) +igt_subtest_f ("%s%s-heavy%s" , e->class == + I915_ENGINE_CLASS_RENDER ? "basic-" : "", e->name, p->name)single (fd, heavy, e, p->flags,1 ,5 ); igt_subtest_f("forked-%s%s", e->name, p->name)single (fd, light, e, p->flags, ncpus,150 );
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 14/43] tests/i915/gem_ctx_switch: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:16)
tests/i915/gem_ctx_switch.c igt_main · line 340
Quoted 6 lines
tests/i915/gem_ctx_switch.c +1
-1@@ -338,7 +337,7 @@ igt_mainigt_fork_hang_detector (fd); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) {
~o~ I'm leaning towards saying that we do need to exercise uABI here, since this is the only spot where we explicitly try and check that contexts exist across the uABI.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 15/43 tests/i915/gem_ctx_thrash: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_ctx_thrash.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-)
tests/i915/gem_ctx_thrash.c +8 -7
@@ -103,13 +103,13 @@ static void single(const char *name, bool all_engines) num_engines =0 ;if (all_engines) { -unsigned engine; +const struct intel_execution_engine2 *e; -for_each_physical_engine (fd, engine) { -if (!gem_can_store_dword (fd, engine)) +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - engines[num_engines++] = engine; + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; }
@@ -220,7 +220,8 @@ static void single(const char *name, bool all_engines)static void processes (void ) { -unsigned engines[16 ], engine; +const struct intel_execution_engine2 *e; +unsigned engines[16 ];int num_engines;struct rlimit rlim;unsigned num_ctx;
@@ -230,8 +231,8 @@ static void processes(void) fd =drm_open_driver (DRIVER_INTEL ); num_engines =0 ; -for_each_physical_engine (fd, engine) { - engines[num_engines++] = engine; +__for_each_physical_engine (fd, e) { + engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ; }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 16/43 tests/i915/gem_eio: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_eio.c | 63 +++++++++++++++++++++++++++----------------- 1 file changed, 39 insertions(+), 24 deletions(-)
tests/i915/gem_eio.c +39 -24
@@ -170,15 +170,15 @@ static int __gem_wait(int fd, uint32_t handle, int64_t timeout)return err; } -static igt_spin_t *__spin_poll (int fd,uint32_t ctx,unsigned long flags) +static igt_spin_t *__spin_poll (int fd,uint32_t ctx,const struct intel_execution_engine2 *e) {struct igt_spin_factory opts = { .ctx = ctx, - .engine = flags, + .engine = e->flags, .flags =IGT_SPIN_FAST , }; -if (gem_can_store_dword (fd, opts.engine)) +if (gem_class_can_store_dword (fd, e->class)) opts.flags |=IGT_SPIN_POLL_RUN ;return __igt_spin_factory (fd, &opts);
@@ -194,9 +194,9 @@ static void __spin_wait(int fd, igt_spin_t *spin) } } -static igt_spin_t *spin_sync (int fd,uint32_t ctx,unsigned long flags) +static igt_spin_t *spin_sync (int fd,uint32_t ctx,const struct intel_execution_engine2 *e) { -igt_spin_t *spin =__spin_poll (fd, ctx, flags); +igt_spin_t *spin =__spin_poll (fd, ctx, e);__spin_wait (fd, spin);
@@ -315,6 +315,7 @@ static void __test_banned(int fd) .buffer_count = 1, };const uint32_t bbe =MI_BATCH_BUFFER_END ; +const struct intel_execution_engine2 *e;unsigned long count =0 ;gem_write (fd, obj.handle,0 , &bbe,sizeof (bbe));
@@ -344,7 +345,12 @@ static void __test_banned(int fd) }/* Trigger a reset, making sure we are detected as guilty */ - hang =spin_sync (fd,0 ,0 ); +__for_each_physical_engine (fd, e) +if (e->class ==I915_ENGINE_CLASS_RENDER ) { + hang =spin_sync (fd,0 , e); +break ; + } +trigger_reset (fd);igt_spin_free (fd, hang);
@@ -368,6 +374,7 @@ static void test_banned(int fd)static void test_wait (int fd,unsigned int flags,unsigned int wait) {igt_spin_t *hang; +const struct intel_execution_engine2 *e; fd =gem_reopen_driver (fd);igt_require_gem (fd);
@@ -382,7 +389,11 @@ static void test_wait(int fd, unsigned int flags, unsigned int wait)else igt_require (i915_reset_control (true)); - hang =spin_sync (fd,0 ,I915_EXEC_DEFAULT ); +__for_each_physical_engine (fd, e) +if (e->class ==I915_ENGINE_CLASS_RENDER ){ + hang =spin_sync (fd,0 , e); +break ; + }check_wait (fd, hang->handle, wait,NULL );
@@ -416,18 +427,18 @@ static void test_suspend(int fd, int state)static void test_inflight (int fd,unsigned int wait) {int parent_fd = fd; -unsigned int engine; +const struct intel_execution_engine2 *engine;int fence[64 ];/* mostly conservative estimate of ring size */ int max;igt_require_gem (fd);igt_require (gem_has_exec_fence (fd)); - max =gem_measure_ring_inflight (fd,-1 ,0 ); + max =gem_measure_ring_inflight (fd,ALL_ENGINES ,0 );igt_require (max >1 ); max =min (max -1 ,ARRAY_SIZE (fence)); -for_each_engine (parent_fd, engine) { +__for_each_physical_engine (parent_fd, engine) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_execbuffer2 execbuf;
@@ -442,7 +453,8 @@ static void test_inflight(int fd, unsigned int wait)gem_write (fd, obj[1 ].handle,0 , &bbe,sizeof (bbe));gem_quiescent_gpu (fd); -igt_debug ("Starting %s on engine '%s'\n" , __func__, e__->name); +igt_debug ("Starting %s on engine '%s'\n" , + __func__, engine->name);igt_require (i915_reset_control (false)); hang =spin_sync (fd,0 , engine);
@@ -451,7 +463,7 @@ static void test_inflight(int fd, unsigned int wait)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = engine |I915_EXEC_FENCE_OUT ; + execbuf.flags = engine->flags |I915_EXEC_FENCE_OUT ;for (unsigned int n =0 ; n < max; n++) {gem_execbuf_wr (fd, &execbuf);
@@ -484,7 +496,7 @@ static void test_inflight_suspend(int fd)igt_spin_t *hang;int max; - max =gem_measure_ring_inflight (fd,-1 ,0 ); + max =gem_measure_ring_inflight (fd,ALL_ENGINES ,0 );igt_require (max >1 ); max =min (max -1 ,ARRAY_SIZE (fence));
@@ -548,13 +560,13 @@ static uint32_t context_create_safe(int i915)static void test_inflight_contexts (int fd,unsigned int wait) {int parent_fd = fd; -unsigned int engine; +const struct intel_execution_engine2 *engine;igt_require_gem (fd);igt_require (gem_has_exec_fence (fd));gem_require_contexts (fd); -for_each_engine (parent_fd, engine) { +__for_each_physical_engine (parent_fd, engine) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_execbuffer2 execbuf;
@@ -570,7 +582,8 @@ static void test_inflight_contexts(int fd, unsigned int wait)gem_quiescent_gpu (fd); -igt_debug ("Starting %s on engine '%s'\n" , __func__, e__->name); +igt_debug ("Starting %s on engine '%s'\n" , + __func__, engine->name);igt_require (i915_reset_control (false));memset (obj,0 ,sizeof (obj));
@@ -584,7 +597,7 @@ static void test_inflight_contexts(int fd, unsigned int wait)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = engine |I915_EXEC_FENCE_OUT ; + execbuf.flags = engine->flags |I915_EXEC_FENCE_OUT ;for (unsigned int n =0 ; n <ARRAY_SIZE (fence); n++) { execbuf.rsvd1 = ctx[n];
@@ -671,7 +684,8 @@ static void test_inflight_internal(int fd, unsigned int wait)struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj[2 ];uint32_t bbe =MI_BATCH_BUFFER_END ; -unsigned engine, nfence =0 ; +unsigned nfence =0 ; +const struct intel_execution_engine2 *engine;int fences[16 ];igt_spin_t *hang;
@@ -692,8 +706,8 @@ static void test_inflight_internal(int fd, unsigned int wait)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; -for_each_engine (fd, engine) { - execbuf.flags = engine |I915_EXEC_FENCE_OUT ; +__for_each_physical_engine (fd, engine) { + execbuf.flags = engine->flags |I915_EXEC_FENCE_OUT ;gem_execbuf_wr (fd, &execbuf);
@@ -716,7 +730,8 @@ static void test_inflight_internal(int fd, unsigned int wait) }static void reset_stress (int fd, -uint32_t ctx0,unsigned int engine, +uint32_t ctx0, +const struct intel_execution_engine2 *engine,unsigned int flags) {const uint32_t bbe =MI_BATCH_BUFFER_END ;
@@ -726,7 +741,7 @@ static void reset_stress(int fd,struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr =to_user_pointer (&obj), .buffer_count =1 , - .flags = engine, + .flags = engine->flags, };igt_stats_t stats;
@@ -794,9 +809,9 @@ static void reset_stress(int fd,static void test_reset_stress (int fd,unsigned int flags) {uint32_t ctx0 =context_create_safe (fd); -unsigned int engine; +const struct intel_execution_engine2 *engine; -for_each_engine (fd, engine) +__for_each_physical_engine (fd, engine)reset_stress (fd, ctx0, engine, flags);gem_context_destroy (fd, ctx0);
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 16/43] tests/i915/gem_eio: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:18)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
gem_eio is first and foremost a uABI test, we cannot simply discard exercising the legacy interface.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 17/43 tests/i915/gem_exec_await: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_await.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
tests/i915/gem_exec_await.c +4 -3
@@ -73,14 +73,15 @@ static void wide(int fd, int ring_size, int timeout, unsigned int flags) } *exec;struct drm_i915_gem_exec_object2 *obj;struct drm_i915_gem_execbuffer2 execbuf; +const struct intel_execution_engine2 *engine;unsigned engines[16 ]; -unsigned nengine, engine; +unsigned nengine;unsigned long count;double time; nengine =0 ; -for_each_physical_engine (fd, engine) - engines[nengine++] = engine; +__for_each_physical_engine (fd, engine) + engines[nengine++] = engine->flags;igt_require (nengine); exec =calloc (nengine,sizeof (*exec));
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 18/43 tests/i915/gem_exec_create: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_create.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
tests/i915/gem_exec_create.c +3 -3
@@ -63,11 +63,11 @@ static void all(int fd, unsigned flags, int timeout, int ncpus)struct drm_i915_gem_exec_object2 obj;unsigned engines[16 ];unsigned nengine; -unsigned engine; +const struct intel_execution_engine2 *e; nengine =0 ; -for_each_physical_engine (fd, engine) - engines[nengine++] = engine; +__for_each_physical_engine (fd, e) + engines[nengine++] = e->flags;igt_require (nengine);memset (&obj,0 ,sizeof (obj));
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 18/43] tests/i915/gem_exec_create: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:20)
Quoted 15 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.Signed-off-by Ramalingam C <[email protected]>--- tests/i915/gem_exec_create.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)tests/i915/gem_exec_create.c +0
-0@@ -63,11 +63,11 @@ static void all(int fd, unsigned flags, int timeout, int ncpus)struct drm_i915_gem_exec_object2 obj;unsigned engines[16 ];
tests/i915/gem_exec_create.c static void all(int fd, unsigned flags, int timeout, int ncpus) · line 71
Quoted 10 lines
tests/i915/gem_exec_create.c +3
-3unsigned nengine; -unsigned engine; +const struct intel_execution_engine2 *e; nengine =0 ; -for_each_physical_engine (fd, engine) - engines[nengine++] = engine; +__for_each_physical_engine (fd, e) + engines[nengine++] = e->flags;igt_require (nengine);
You will have noticed this pattern quite frequently, and possibly considered a helper to simply return the engine map? -Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 18/43] tests/i915/gem_exec_create: use the gem_engine_topology library
Quoted 13 lines
tests/i915/gem_exec_create.c static void all(int fd, unsigned flags, int timeout, int ncpus) · line 71
Quoted 12 lines
struct drm_i915_gem_exec_object2 obj; unsigned engines[16]; unsigned nengine;tests/i915/gem_exec_create.c +3
-3-unsigned engine; +const struct intel_execution_engine2 *e; nengine =0 ; -for_each_physical_engine (fd, engine) - engines[nengine++] = engine; +__for_each_physical_engine (fd, e) + engines[nengine++] = e->flags;igt_require (nengine);
Quoted 2 lines
You will have noticed this pattern quite frequently, and possibly considered a helper to simply return the engine map?
I think that just calling "intel_init_engine_list()" and retrieve a "struct inte_engine_data" does it without making any other helper.
you would get the list of engines/flags in:
(struct intel_engine_data *)->engines[n].flags
where we also have the flags.
Andi
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 19/43 tests/i915/gem_exec_fence: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_fence.c | 112 +++++++++++++++++++----------------- 1 file changed, 58 insertions(+), 54 deletions(-)
tests/i915/gem_exec_fence.c +58 -54
@@ -57,7 +57,8 @@ struct sync_merge_data {#define SYNC_IOC_MERGE _IOWR(SYNC_IOC_MAGIC, 3, struct sync_merge_data)#endif -static void store (int fd,unsigned ring,int fence,uint32_t target,unsigned offset_value) +static void store (int fd,const struct intel_execution_engine2 *e,int fence, +uint32_t target,unsigned offset_value) {const int SCRATCH =0 ;const int BATCH =1 ;
@@ -71,7 +72,7 @@ static void store(int fd, unsigned ring, int fence, uint32_t target, unsigned ofmemset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = ring |LOCAL_EXEC_FENCE_IN ; + execbuf.flags = e->flags |LOCAL_EXEC_FENCE_IN ; execbuf.rsvd2 = fence;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -119,7 +120,8 @@ static bool fence_busy(int fence)#define NONBLOCK 0x2#define WAIT 0x4 -static void test_fence_busy (int fd,unsigned ring,unsigned flags) +static void test_fence_busy (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_exec_object2 obj;
@@ -134,7 +136,7 @@ static void test_fence_busy(int fd, unsigned ring, unsigned flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (&obj); execbuf.buffer_count =1 ; - execbuf.flags = ring |LOCAL_EXEC_FENCE_OUT ; + execbuf.flags = e->flags |LOCAL_EXEC_FENCE_OUT ;memset (&obj,0 ,sizeof (obj)); obj.handle =gem_create (fd,4096 );
@@ -216,7 +218,7 @@ static void test_fence_busy_all(int fd, unsigned flags)struct drm_i915_gem_execbuffer2 execbuf;struct timespec tv;uint32_t *batch; -unsigned int engine; +const struct intel_execution_engine2 *e;int all, i, timeout;gem_quiescent_gpu (fd);
@@ -263,10 +265,10 @@ static void test_fence_busy_all(int fd, unsigned flags) i++; all =-1 ; -for_each_engine (fd, engine) { +__for_each_static_engine (e) {int fence, new; - execbuf.flags = engine |LOCAL_EXEC_FENCE_OUT ; + execbuf.flags = e->flags |LOCAL_EXEC_FENCE_OUT ; execbuf.rsvd2 =-1 ;gem_execbuf_wr (fd, &execbuf); fence = execbuf.rsvd2 >>32 ;
@@ -315,11 +317,12 @@ static void test_fence_busy_all(int fd, unsigned flags)gem_quiescent_gpu (fd); } -static void test_fence_await (int fd,unsigned ring,unsigned flags) +static void test_fence_await (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {uint32_t scratch =gem_create (fd,4096 );igt_spin_t *spin; -unsigned engine; +const struct intel_execution_engine2 *other;uint32_t *out;int i;
@@ -330,20 +333,20 @@ static void test_fence_await(int fd, unsigned ring, unsigned flags) I915_GEM_DOMAIN_GTT, I915_GEM_DOMAIN_GTT); spin = igt_spin_new(fd, - .engine = ring, + .engine = e->flags, .flags =IGT_SPIN_FENCE_OUT );igt_assert (spin->out_fence !=-1 ); i =0 ; -for_each_physical_engine (fd, engine) { -if (!gem_can_store_dword (fd, engine)) +__for_each_physical_engine (fd, other) { +if (!gem_class_can_store_dword (fd, other->class))continue ;if (flags &NONBLOCK ) { -store (fd, engine, spin->out_fence, scratch, i); +store (fd, other, spin->out_fence, scratch, i); }else { igt_fork(child, 1) -store (fd, engine, spin->out_fence, scratch, i); +store (fd, other, spin->out_fence, scratch, i); } i++;
@@ -370,13 +373,14 @@ static void test_fence_await(int fd, unsigned ring, unsigned flags)gem_close (fd, scratch); } -static void resubmit (int fd,uint32_t handle,unsigned int ring,int count) +static void resubmit (int fd,uint32_t handle, +const struct intel_execution_engine2 *e,int count) {struct drm_i915_gem_exec_object2 obj = { .handle = handle };struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr =to_user_pointer (&obj), .buffer_count =1 , - .flags = ring, + .flags = e->flags, };while (count--)gem_execbuf (fd, &execbuf);
@@ -398,7 +402,7 @@ static int __execbuf(int fd, struct drm_i915_gem_execbuffer2 *execbuf)return err; } -static void test_parallel (int fd,unsigned int master) +static void test_parallel (int fd,const struct intel_execution_engine2 *e) {const int SCRATCH =0 ;const int BATCH =1 ;
@@ -411,7 +415,7 @@ static void test_parallel(int fd, unsigned int master)uint32_t handle[16 ];uint32_t batch[16 ];igt_spin_t *spin; -unsigned engine; +const struct intel_execution_engine2 *other;IGT_CORK_HANDLE (c);uint32_t plug;int i, x =0 ;
@@ -421,14 +425,14 @@ static void test_parallel(int fd, unsigned int master)/* Fill the queue with many requests so that the next one has to * wait before it can be executed by the hardware. */ - spin = igt_spin_new(fd, .engine = master, .dependency = plug); -resubmit (fd, spin->handle, master,16 ); + spin = igt_spin_new(fd, .engine = e->flags, .dependency = plug); +resubmit (fd, spin->handle, e,16 );/* Now queue the master request and its secondaries */ memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = master |LOCAL_EXEC_FENCE_OUT ; + execbuf.flags = e->flags |LOCAL_EXEC_FENCE_OUT ;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -497,11 +501,11 @@ static void test_parallel(int fd, unsigned int master) obj[BATCH ].relocation_count =1 ;/* Queue all secondaries */ -for_each_physical_engine (fd, engine) { -if (engine == master) +__for_each_physical_engine (fd, other) { +if (other == e)continue ; - execbuf.flags = engine |LOCAL_EXEC_FENCE_SUBMIT ; + execbuf.flags = other->flags |LOCAL_EXEC_FENCE_SUBMIT ;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -588,7 +592,8 @@ static inline uint32_t upper_32_bits(uint64_t x)return x >>32 ; } -static void test_keep_in_fence (int fd,unsigned int engine,unsigned int flags) +static void test_keep_in_fence (int fd,const struct intel_execution_engine2 *e, +unsigned int flags) {struct sigaction sa = { .sa_handler = alarm_handler };struct drm_i915_gem_exec_object2 obj = {
@@ -597,14 +602,14 @@ static void test_keep_in_fence(int fd, unsigned int engine, unsigned int flags)struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr =to_user_pointer (&obj), .buffer_count =1 , - .flags = engine |LOCAL_EXEC_FENCE_OUT , + .flags = e->flags |LOCAL_EXEC_FENCE_OUT , };unsigned long count, last;struct itimerval itv;igt_spin_t *spin;int fence; - spin = igt_spin_new(fd, .engine = engine); + spin = igt_spin_new(fd, .engine = e->flags);gem_execbuf_wr (fd, &execbuf); fence =upper_32_bits (execbuf.rsvd2);
@@ -662,7 +667,8 @@ static void test_long_history(int fd, long ring_size, unsigned flags)const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_execbuffer2 execbuf; -unsigned int engines[16 ], engine; +unsigned int engines[16 ]; +const struct intel_execution_engine2 *e;unsigned int nengine, n, s;unsigned long limit;int all_fences;
@@ -673,8 +679,8 @@ static void test_long_history(int fd, long ring_size, unsigned flags) limit = ring_size /3 ; nengine =0 ; -for_each_physical_engine (fd, engine) - engines[nengine++] = engine; +__for_each_physical_engine (fd, e) + engines[nengine++] = e->flags;igt_require (nengine);gem_quiescent_gpu (fd);
@@ -1155,7 +1161,7 @@ static void test_syncobj_wait(int fd) .handle = syncobj_create(fd), };igt_spin_t *spin; -unsigned engine; +const struct intel_execution_engine2 *e;unsigned handle[16 ];int n;
@@ -1188,13 +1194,13 @@ static void test_syncobj_wait(int fd)gem_write (fd, obj.handle,0 , &bbe,sizeof (bbe)); n =0 ; -for_each_engine (fd, engine) { +__for_each_static_engine (e) { obj.handle =gem_create (fd,4096 );gem_write (fd, obj.handle,0 , &bbe,sizeof (bbe));/* No inter-engine synchronisation, will complete */ -if (engine ==I915_EXEC_BLT ) { - execbuf.flags = engine; +if (e->class ==I915_ENGINE_CLASS_COPY ) { + execbuf.flags = e->flags; execbuf.cliprects_ptr =0 ; execbuf.num_cliprects =0 ;gem_execbuf (fd, &execbuf);
@@ -1204,7 +1210,7 @@ static void test_syncobj_wait(int fd)igt_assert (gem_bo_busy (fd, spin->handle));/* Now wait upon the blocked engine */ - execbuf.flags =LOCAL_EXEC_FENCE_ARRAY | engine; + execbuf.flags =LOCAL_EXEC_FENCE_ARRAY | e->flags; execbuf.cliprects_ptr =to_user_pointer (&fence); execbuf.num_cliprects =1 ; fence.flags =LOCAL_EXEC_FENCE_WAIT ;
@@ -1513,7 +1519,7 @@ static void test_syncobj_channel(int fd) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int i915 =-1 ; igt_fixture {
@@ -1545,11 +1551,10 @@ igt_maintest_fence_busy_all (i915,WAIT |HANG ); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (i915, e) { igt_subtest_group { igt_fixture { -igt_require (gem_has_ring (i915, e->exec_id | e->flags)); -igt_require (gem_can_store_dword (i915, e->exec_id | e->flags)); +igt_require (gem_class_can_store_dword (i915, e->class)); } igt_subtest_group {
@@ -1558,29 +1563,28 @@ igt_main }igt_subtest_f ("%sbusy-%s" , - e->exec_id ==0 ?"basic-" :"" , + e->class ==I915_ENGINE_CLASS_RENDER ?"basic-" :"" , e->name) -test_fence_busy (i915, e->exec_id | e->flags,0 ); +test_fence_busy (i915, e,0 );igt_subtest_f ("%swait-%s" , - e->exec_id ==0 ?"basic-" :"" , + e->class ==I915_ENGINE_CLASS_RENDER ?"basic-" :"" , e->name) -test_fence_busy (i915, e->exec_id | e->flags,WAIT ); +test_fence_busy (i915, e,WAIT );igt_subtest_f ("%sawait-%s" , - e->exec_id ==0 ?"basic-" :"" , + e->class ==I915_ENGINE_CLASS_RENDER ?"basic-" :"" , e->name) -test_fence_await (i915, e->exec_id | e->flags,0 ); +test_fence_await (i915, e,0 ); igt_subtest_f("nb-await-%s", e->name) -test_fence_await (i915, e->exec_id | e->flags,NONBLOCK ); +test_fence_await (i915, e,NONBLOCK ); igt_subtest_f("keep-in-fence-%s", e->name) -test_keep_in_fence (i915, e->exec_id | e->flags,0 ); +test_keep_in_fence (i915, e,0 ); -if (e->exec_id && - !(e->exec_id ==I915_EXEC_BSD && !e->flags)) { +if (!(e->class ==I915_ENGINE_CLASS_VIDEO && !e->flags)) { igt_subtest_f("parallel-%s", e->name) {igt_require (has_submit_fence (i915)); igt_until_timeout(2) -test_parallel (i915, e->exec_id | e->flags); +test_parallel (i915, e); } }
@@ -1599,13 +1603,13 @@ igt_main } igt_subtest_f("busy-hang-%s", e->name) -test_fence_busy (i915, e->exec_id | e->flags,HANG ); +test_fence_busy (i915, e,HANG ); igt_subtest_f("wait-hang-%s", e->name) -test_fence_busy (i915, e->exec_id | e->flags,HANG |WAIT ); +test_fence_busy (i915, e,HANG |WAIT ); igt_subtest_f("await-hang-%s", e->name) -test_fence_await (i915, e->exec_id | e->flags,HANG ); +test_fence_await (i915, e,HANG ); igt_subtest_f("nb-await-hang-%s", e->name) -test_fence_await (i915, e->exec_id | e->flags,NONBLOCK |HANG ); +test_fence_await (i915, e,NONBLOCK |HANG ); igt_fixture {igt_disallow_hang (i915, hang); }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 19/43] tests/i915/gem_exec_fence: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:21)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
Also a uABI test.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 20/43 tests/i915/gem_exec_flush: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_flush.c | 72 ++++++++++++++++++------------------- 1 file changed, 35 insertions(+), 37 deletions(-)
tests/i915/gem_exec_flush.c +35 -37
@@ -74,8 +74,8 @@ static uint32_t movnt(uint32_t *map, int i) }#endif -static void run (int fd,unsigned ring,int nchild,int timeout, -unsigned flags) +static void run (int fd,const struct intel_execution_engine2 *e, +int nchild,int timeout,unsigned flags) {const int gen =intel_gen (intel_get_drm_devid (fd));
@@ -162,7 +162,7 @@ static void run(int fd, unsigned ring, int nchild, int timeout,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =3 ; - execbuf.flags = ring | (1 <<11 ) | (1 <<12 ); + execbuf.flags = e->flags | (1 <<11 ) | (1 <<12 );if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -352,7 +352,8 @@ enum batch_mode {BATCH_GTT , BATCH_WC, }; -static void batch (int fd,unsigned ring,int nchild,int timeout, +static void batch (int fd,const struct intel_execution_engine2 *e, +int nchild,int timeout,enum batch_mode mode,unsigned flags) {const int gen =intel_gen (intel_get_drm_devid (fd));
@@ -394,7 +395,7 @@ static void batch(int fd, unsigned ring, int nchild, int timeout,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = ring | (1 <<11 ) | (1 <<12 ); + execbuf.flags = e->flags | (1 <<11 ) | (1 <<12 );if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -552,7 +553,7 @@ static const char *yesno(bool x) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;const int ncpus =sysconf (_SC_NPROCESSORS_ONLN); const struct batch { const char *name;
@@ -599,78 +600,75 @@ igt_mainigt_fork_hang_detector (fd); } - for (e = intel_execution_engines; e->name; e++) igt_subtest_group { -unsigned ring = e->exec_id | e->flags; -unsigned timeout =5 +120 *!!e->exec_id; + __for_each_physical_engine(fd, e) igt_subtest_group { +unsigned timeout =5 +120 ; igt_fixture { -gem_require_ring (fd ,ring ); -igt_require (gem_can_store_dword (fd, ring)); +igt_require (gem_class_can_store_dword (fd, e->class)); }for (const struct batch *b = batches; b->name; b++) {igt_subtest_f ("%sbatch-%s-%s-uc" , - b == batches && e->exec_id ==0 ?"basic-" :"" , - b->name, - e->name) -batch (fd, ring, ncpus, timeout, b->mode,0 ); + b == batches && + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", b->name, e->name) +batch (fd, e, ncpus, timeout, b->mode,0 );igt_subtest_f ("%sbatch-%s-%s-wb" , - b == batches && e->exec_id ==0 ?"basic-" :"" , - b->name, - e->name) -batch (fd, ring, ncpus, timeout, b->mode,COHERENT ); + b == batches && + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", b->name, e->name) +batch (fd, e, ncpus, timeout, b->mode,COHERENT );igt_subtest_f ("%sbatch-%s-%s-cmd" , - b == batches && e->exec_id ==0 ?"basic-" :"" , - b->name, - e->name) -batch (fd, ring, ncpus, timeout, b->mode, + b == batches && + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", b->name, e->name) +batch (fd, e, ncpus, timeout, b->mode,COHERENT |CMDPARSER ); }for (const struct mode *m = modes; m->name; m++) {igt_subtest_f ("%suc-%s-%s" , - (m->flags &BASIC && e->exec_id ==0 ) ?"basic-" :"" , - m->name, - e->name) -run (fd, ring, ncpus, timeout, + (m->flags &BASIC && + e->class ==I915_ENGINE_CLASS_RENDER ) ? + "basic-" : "", m->name, e->name) +run (fd, e, ncpus, timeout,UNCACHED | m->flags);igt_subtest_f ("uc-%s-%s-interruptible" , m->name, e->name) -run (fd, ring, ncpus, timeout, +run (fd, e, ncpus, timeout,UNCACHED | m->flags |INTERRUPTIBLE );igt_subtest_f ("%swb-%s-%s" , - e->exec_id ==0 ?"basic-" :"" , - m->name, - e->name) -run (fd, ring, ncpus, timeout, + e->class ==I915_ENGINE_CLASS_RENDER ? + "basic-" : "", m->name, e->name) +run (fd, e, ncpus, timeout,COHERENT | m->flags);igt_subtest_f ("wb-%s-%s-interruptible" , m->name, e->name) -run (fd, ring, ncpus, timeout, +run (fd, e, ncpus, timeout,COHERENT | m->flags |INTERRUPTIBLE );igt_subtest_f ("wc-%s-%s" , m->name, e->name) -run (fd, ring, ncpus, timeout, +run (fd, e, ncpus, timeout,COHERENT |WC | m->flags);igt_subtest_f ("wc-%s-%s-interruptible" , m->name, e->name) -run (fd, ring, ncpus, timeout, +run (fd, e, ncpus, timeout,COHERENT |WC | m->flags |INTERRUPTIBLE );igt_subtest_f ("stream-%s-%s" , m->name, e->name) {igt_require (cpu &SSE4_1 ); -run (fd, ring, ncpus, timeout, +run (fd, e, ncpus, timeout,MOVNT |COHERENT |WC | m->flags); }
@@ -678,7 +676,7 @@ igt_main m->name, e->name) {igt_require (cpu &SSE4_1 ); -run (fd, ring, ncpus, timeout, +run (fd, e, ncpus, timeout,MOVNT |COHERENT |WC | m->flags |INTERRUPTIBLE ); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 21/43 tests/i915/gem_exec_gttfill: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_gttfill.c | 30 +++++++++++++++--------------- 1 file changed, 15 insertions(+), 15 deletions(-)
tests/i915/gem_exec_gttfill.c +15 -15
@@ -104,16 +104,16 @@ static void submit(int fd, int gen,gem_sync (fd, obj.handle); } -static void fillgtt (int fd,unsigned ring,int timeout) +static void fillgtt (int fd,const struct intel_execution_engine2 *e, +int timeout) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_relocation_entry reloc[2 ];volatile uint64_t *shared;struct batch *batches; -unsigned engines[16 ]; -unsigned nengine; -unsigned engine; +unsigned nengine, engines[16 ]; +const struct intel_execution_engine2 *other;uint64_t size;unsigned count;
@@ -121,18 +121,18 @@ static void fillgtt(int fd, unsigned ring, int timeout)igt_assert (shared !=MAP_FAILED ); nengine =0 ; -if (ring ==0 ) { -for_each_physical_engine (fd, engine) { -if (!gem_can_store_dword (fd, engine)) +if (!e) { +__for_each_physical_engine (fd, other) { +if (!gem_class_can_store_dword (fd, other->class))continue ; - engines[nengine++] = engine; + engines[nengine++] = other->flags; } }else { -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); - engines[nengine++] = ring; + engines[nengine++] = e->flags; }igt_require (nengine);
@@ -207,7 +207,7 @@ static void fillgtt(int fd, unsigned ring, int timeout) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int device =-1 ;igt_skip_on_simulation ();
@@ -222,12 +222,12 @@ igt_main igt_subtest("basic")fillgtt (device,0 ,1 );/* just enough to run a single pass */ -for (e = intel_execution_engines; e->name; e++) +__for_each_physical_engine (device, e) igt_subtest_f("%s", e->name) -fillgtt (device, e->exec_id | e->flags,20 ); +fillgtt (device, e,20 ); igt_subtest("all") -fillgtt (device,0 ,150 ); +fillgtt (device,NULL ,150 ); igt_fixture {igt_stop_hang_detector ();
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 22/43 tests/i915/gem_exec_latency: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_latency.c | 103 ++++++++++++++-------------------- 1 file changed, 42 insertions(+), 61 deletions(-)
tests/i915/gem_exec_latency.c +42 -61
@@ -63,10 +63,10 @@ static unsigned int ring_size;static double rcs_clock;static void -poll_ring (int fd,unsigned ring,const char *name) +poll_ring (int fd,const struct intel_execution_engine2 *e,const char *name) {const struct igt_spin_factory opts = { - .engine = ring, + .engine = e->flags, .flags =IGT_SPIN_POLL_RUN |IGT_SPIN_FAST , };struct timespec tv = {};
@@ -74,8 +74,8 @@ poll_ring(int fd, unsigned ring, const char *name)igt_spin_t *spin[2 ];uint64_t elapsed; -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); spin[0 ] =__igt_spin_factory (fd, &opts);igt_assert (igt_spin_has_poll (spin[0 ]));
@@ -109,8 +109,8 @@ poll_ring(int fd, unsigned ring, const char *name)#define RCS_TIMESTAMP (0x2000 + 0x358)static void latency_on_ring (int fd, -unsigned ring,const char *name, -unsigned flags) +const struct intel_execution_engine2 *e, +const char *name,unsigned flags) {const int gen =intel_gen (intel_get_drm_devid (fd));const int has_64bit_reloc = gen >=8 ;
@@ -131,7 +131,7 @@ static void latency_on_ring(int fd,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (&obj[1 ]); execbuf.buffer_count =2 ; - execbuf.flags = ring; + execbuf.flags = e->flags; execbuf.flags |=LOCAL_I915_EXEC_NO_RELOC |LOCAL_I915_EXEC_HANDLE_LUT ;memset (obj,0 ,sizeof (obj));
@@ -188,7 +188,7 @@ static void latency_on_ring(int fd, }if (flags &LIVE ) - spin = igt_spin_new(fd, .engine = ring); + spin = igt_spin_new(fd, .engine = e->flags); start = *reg;for (j =0 ; j < repeats; j++) {
@@ -256,8 +256,8 @@ static void latency_on_ring(int fd, }static void latency_from_ring (int fd, -unsigned ring,const char *name, -unsigned flags) +const struct intel_execution_engine2 *e, +const char *name,unsigned flags) {const int gen =intel_gen (intel_get_drm_devid (fd));const int has_64bit_reloc = gen >=8 ;
@@ -265,7 +265,7 @@ static void latency_from_ring(int fd,struct drm_i915_gem_relocation_entry reloc;struct drm_i915_gem_execbuffer2 execbuf;const unsigned int repeats = ring_size /2 ; -unsigned int other; +const struct intel_execution_engine2 *other;uint32_t *map, *results;uint32_t ctx[2 ] = {};int i, j;
@@ -281,7 +281,7 @@ static void latency_from_ring(int fd,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (&obj[1 ]); execbuf.buffer_count =2 ; - execbuf.flags = ring; + execbuf.flags = e->flags; execbuf.flags |=LOCAL_I915_EXEC_NO_RELOC |LOCAL_I915_EXEC_HANDLE_LUT ; execbuf.rsvd1 = ctx[1 ];
@@ -311,7 +311,7 @@ static void latency_from_ring(int fd, reloc.presumed_offset = obj[1 ].offset; reloc.target_handle = flags &CORK ?1 :0 ; -for_each_physical_engine (fd, other) { +__for_each_physical_engine (fd, other) {igt_spin_t *spin =NULL ;IGT_CORK_HANDLE (c);
@@ -322,7 +322,7 @@ static void latency_from_ring(int fd,if (flags &PREEMPT ) spin = __igt_spin_new(fd, .ctx = ctx[0], - .engine = ring); + .engine = e->flags);if (flags &CORK ) { obj[0 ].handle =igt_cork_plug (&c, fd);
@@ -334,7 +334,7 @@ static void latency_from_ring(int fd,uint64_t offset; execbuf.flags &= ~ENGINE_FLAGS ; - execbuf.flags |= ring; + execbuf.flags |= e->flags; execbuf.batch_start_offset =64 * j; reloc.offset =
@@ -359,7 +359,7 @@ static void latency_from_ring(int fd,gem_execbuf (fd, &execbuf); execbuf.flags &= ~ENGINE_FLAGS ; - execbuf.flags |= other; + execbuf.flags |= other->flags; execbuf.batch_start_offset =64 * (j + repeats); reloc.offset =
@@ -392,7 +392,7 @@ static void latency_from_ring(int fd,igt_spin_free (fd, spin);igt_info ("%s-%s delay: %.2fns\n" , - name, e__->name, + name, e->name, (results[2 *repeats-1 ] - results[0 ]) / (double )repeats * rcs_clock); }
@@ -441,7 +441,8 @@ static bool __spin_wait(int fd, igt_spin_t *spin)* reasonable latency. */ static void -rthog_latency_on_ring(int fd, unsigned int engine, const char *name, unsigned int flags) +rthog_latency_on_ring (int fd,const struct intel_execution_engine2 *e, + const char *name, unsigned int flags)#define RTIDLE 0x1 {const char *passname[] = {
@@ -458,7 +459,7 @@ rthog_latency_on_ring(int fd, unsigned int engine, const char *name, unsigned in#define NPASS ARRAY_SIZE(passname)#define MMAP_SZ (64 << 10)const struct igt_spin_factory opts = { - .engine = engine, + .engine = e->flags, .flags =IGT_SPIN_POLL_RUN |IGT_SPIN_FAST , };struct rt_pkt *results;
@@ -472,19 +473,19 @@ rthog_latency_on_ring(int fd, unsigned int engine, const char *name, unsigned inigt_assert (results !=MAP_FAILED ); nengine =0 ; -if (engine ==ALL_ENGINES ) { -for_each_physical_engine (fd, engine) { -if (!gem_can_store_dword (fd, engine)) +if (!e) { +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - engines[nengine] = engine; - names[nengine] = e__->name; + engines[nengine] = e->flags; + names[nengine] = e->name; nengine++; }igt_require (nengine >1 ); }else { -igt_require (gem_can_store_dword (fd, engine)); - engines[nengine] = engine; +igt_require (gem_class_can_store_dword (fd, e->class)); + engines[nengine] = e->flags; names[nengine] = name; nengine++; }
@@ -494,7 +495,6 @@ rthog_latency_on_ring(int fd, unsigned int engine, const char *name, unsigned in igt_fork(child, nengine) {unsigned int pass =0 ;/* Three phases: warmup, normal, rt. */ - engine = engines[child]; do { struct igt_mean mean; double min = HUGE_VAL;
@@ -547,7 +547,7 @@ rthog_latency_on_ring(int fd, unsigned int engine, const char *name, unsigned inigt_spin_reset (spin);igt_nsec_elapsed (&ts); -__submit_spin (fd, spin, engine); +__submit_spin (fd, spin, engines[child]);if (!__spin_wait (fd, spin)) {igt_warn ("Wait timeout! (%s)\n" , passname[pass]);
@@ -650,7 +650,7 @@ static double clockrate(int i915, int reg) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int device =-1 ; igt_fixture {
@@ -675,71 +675,53 @@ igt_main } igt_subtest("all-rtidle-submit") -rthog_latency_on_ring (device,ALL_ENGINES ,"all" ,RTIDLE ); +rthog_latency_on_ring (device,NULL ,"all" ,RTIDLE ); igt_subtest("all-rthog-submit") -rthog_latency_on_ring (device,ALL_ENGINES ,"all" ,0 ); +rthog_latency_on_ring (device,NULL ,"all" ,0 ); igt_subtest_group {igt_fixture igt_require(intel_gen(intel_get_drm_devid(device)) >= 7); -for (e = intel_execution_engines; e->name; e++) { -if (e->exec_id ==0 ) -continue ; - +__for_each_physical_engine (device, e) { igt_subtest_group { - igt_fixture { -igt_require (gem_ring_has_physical_engine (device, e->exec_id | e->flags)); - } - igt_subtest_f("%s-dispatch", e->name)latency_on_ring (device, - e->exec_id | e->flags, - e->name,0 ); + e, e->name,0 ); igt_subtest_f("%s-live-dispatch", e->name)latency_on_ring (device, - e->exec_id | e->flags, - e->name,LIVE ); + e, e->name,LIVE ); igt_subtest_f("%s-poll", e->name)poll_ring (device, - e->exec_id | e->flags, - e->name); + e, e->name); igt_subtest_f("%s-rtidle-submit", e->name)rthog_latency_on_ring (device, - e->exec_id | - e->flags, - e->name, + e, e->name,RTIDLE ); igt_subtest_f("%s-rthog-submit", e->name)rthog_latency_on_ring (device, - e->exec_id | - e->flags, - e->name, + e, e->name,0 ); igt_subtest_f("%s-live-dispatch-queued", e->name)latency_on_ring (device, - e->exec_id | e->flags, - e->name,LIVE |CORK ); + e, e->name,LIVE |CORK ); igt_subtest_f("%s-dispatch-queued", e->name)latency_on_ring (device, - e->exec_id | e->flags, - e->name,CORK ); + e, e->name,CORK ); igt_subtest_f("%s-synchronisation", e->name)latency_from_ring (device, - e->exec_id | e->flags, - e->name,0 ); + e, e->name,0 ); igt_subtest_f("%s-synchronisation-queued", e->name)latency_from_ring (device, - e->exec_id | e->flags, - e->name,CORK ); + e, e->name,CORK ); igt_subtest_group { igt_fixture {
@@ -749,8 +731,7 @@ igt_main igt_subtest_f("%s-preemption", e->name)latency_from_ring (device, - e->exec_id | e->flags, - e->name,PREEMPT ); + e, e->name,PREEMPT ); } } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 23/43 tests/i915/gem_exec_nop: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_nop.c | 119 ++++++++++++++++++++------------------ 1 file changed, 63 insertions(+), 56 deletions(-)
tests/i915/gem_exec_nop.c +63 -56
@@ -66,7 +66,8 @@ static double elapsed(const struct timespec *start, const struct timespec *end) (end->tv_nsec - start->tv_nsec)*1e-9); } -static double nop_on_ring (int fd,uint32_t handle,unsigned ring_id, +static double nop_on_ring (int fd,uint32_t handle, +const struct intel_execution_engine2 *e,int timeout,unsigned long *out) {struct drm_i915_gem_execbuffer2 execbuf;
@@ -80,11 +81,11 @@ static double nop_on_ring(int fd, uint32_t handle, unsigned ring_id,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (&obj); execbuf.buffer_count =1 ; - execbuf.flags = ring_id; + execbuf.flags = e->flags; execbuf.flags |=LOCAL_I915_EXEC_HANDLE_LUT ; execbuf.flags |=LOCAL_I915_EXEC_NO_RELOC ;if (__gem_execbuf (fd, &execbuf)) { - execbuf.flags = ring_id; + execbuf.flags = e->flags;gem_execbuf (fd, &execbuf); }intel_detect_and_clear_missed_interrupts (fd);
@@ -104,7 +105,8 @@ static double nop_on_ring(int fd, uint32_t handle, unsigned ring_id,return elapsed (&start, &now); } -static void poll_ring (int fd,unsigned engine,const char *name,int timeout) +static void poll_ring (int fd,const struct intel_execution_engine2 *e, +const char *name,int timeout) {const int gen =intel_gen (intel_get_drm_devid (fd));const uint32_t MI_ARB_CHK =0x5 <<23 ;
@@ -121,8 +123,8 @@ static void poll_ring(int fd, unsigned engine, const char *name, int timeout)if (gen ==4 || gen ==5 ) flags |=I915_EXEC_SECURE ; -gem_require_ring (fd, engine); -igt_require (gem_can_store_dword (fd, engine)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class));memset (&obj,0 ,sizeof (obj)); obj.handle =gem_create (fd,4096 );
@@ -186,7 +188,7 @@ static void poll_ring(int fd, unsigned engine, const char *name, int timeout)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (&obj); execbuf.buffer_count =1 ; - execbuf.flags = engine | flags; + execbuf.flags = e->flags | flags; cycles =0 ; do {
@@ -222,7 +224,8 @@ static void poll_sequential(int fd, const char *name, int timeout)struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_relocation_entry reloc[4 ], *r;uint32_t *bbe[2 ], *state, *batch; -unsigned engines[16 ], nengine, engine, flags; +unsigned engines[16 ], nengine, flags; +const struct intel_execution_engine2 *e;struct timespec tv = {};unsigned long cycles;uint64_t elapsed;
@@ -233,11 +236,11 @@ static void poll_sequential(int fd, const char *name, int timeout) flags |=I915_EXEC_SECURE ; nengine =0 ; -for_each_physical_engine (fd, engine) { -if (!gem_can_store_dword (fd, engine)) +__for_each_physical_engine (fd, e) { +if (!gem_class_can_store_dword (fd, e->class))continue ; - engines[nengine++] = engine; + engines[nengine++] = e->flags; }igt_require (nengine);
@@ -342,20 +345,22 @@ static void poll_sequential(int fd, const char *name, int timeout) }static void single (int fd,uint32_t handle, -unsigned ring_id,const char *ring_name) +const struct intel_execution_engine2 *e, +const char *ring_name) {double time;unsigned long count; -gem_require_ring (fd, ring_id); +gem_require_ring (fd, e->flags); - time =nop_on_ring (fd, handle, ring_id,20 , &count); + time =nop_on_ring (fd, handle, e,20 , &count);igt_info ("%s: %'lu cycles: %.3fus\n" , ring_name, count, time*1e6 / count); }static double -stable_nop_on_ring (int fd,uint32_t handle,unsigned int engine, +stable_nop_on_ring (int fd,uint32_t handle, +const struct intel_execution_engine2 *e,int timeout,int reps) {igt_stats_t s;
@@ -370,7 +375,7 @@ stable_nop_on_ring(int fd, uint32_t handle, unsigned int engine,unsigned long count;double time; - time =nop_on_ring (fd, handle, engine, timeout, &count); + time =nop_on_ring (fd, handle, e, timeout, &count);igt_stats_push_float (&s, time / count); }
@@ -432,20 +437,20 @@ static void parallel(int fd, uint32_t handle, int timeout)unsigned engines[16 ];const char *names[16 ];unsigned nengine; -unsigned engine; +const struct intel_execution_engine2 *e;unsigned long count;double time, sum; sum =0 ; nengine =0 ; -for_each_physical_engine (fd, engine) { - engines[nengine] = engine; - names[nengine] = e__->name; +__for_each_physical_engine (fd, e) { + engines[nengine] = e->flags; + names[nengine] = e->name; nengine++; - time =nop_on_ring (fd, handle, engine,1 , &count) / count; + time =nop_on_ring (fd, handle, e,1 , &count) / count; sum += time; -igt_debug ("%s: %.3fus\n" , e__->name,1e6 *time); +igt_debug ("%s: %.3fus\n" , e->name,1e6 *time); }igt_require (nengine);igt_info ("average (individually): %.3fus\n" , sum/nengine*1e6 );
@@ -491,25 +496,25 @@ static void series(int fd, uint32_t handle, int timeout) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj; +const struct intel_execution_engine2 *e;struct timespec start, now, sync;unsigned engines[16 ];unsigned nengine; -unsigned engine;unsigned long count;double time, max =0 , min =HUGE_VAL , sum =0 ;const char *name; nengine =0 ; -for_each_physical_engine (fd, engine) { - time =nop_on_ring (fd, handle, engine,1 , &count) / count; +__for_each_physical_engine (fd, e) { + time =nop_on_ring (fd, handle, e,1 , &count) / count;if (time > max) { - name = e__->name; + name = e->name; max = time; }if (time < min) min = time; sum += time; - engines[nengine++] = engine; + engines[nengine++] = e->flags; }igt_require (nengine); igt_info("Maximum execution latency on %s, %.3fus, min %.3fus, total %.3fus per cycle, average %.3fus\n",
@@ -582,11 +587,12 @@ static void sequential(int fd, uint32_t handle, unsigned flags, int timeout)const int ncpus = flags &FORKED ?sysconf (_SC_NPROCESSORS_ONLN) :1 ;struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj[2 ]; +const struct intel_execution_engine2 *e;unsigned engines[16 ];unsigned nengine;double *results;double time, sum; -unsigned n; +int n;gem_require_contexts (fd);
@@ -595,14 +601,14 @@ static void sequential(int fd, uint32_t handle, unsigned flags, int timeout) nengine =0 ; sum =0 ; -for_each_physical_engine (fd, n) { +__for_each_physical_engine (fd, e) {unsigned long count; - time =nop_on_ring (fd, handle, n,1 , &count) / count; + time =nop_on_ring (fd, handle, e,1 , &count) / count; sum += time; -igt_debug ("%s: %.3fus\n" , e__->name,1e6 *time); +igt_debug ("%s: %.3fus\n" , e->name,1e6 *time); - engines[nengine++] = n; + engines[nengine++] = e->flags; }igt_require (nengine); igt_info("Total (individual) execution latency %.3fus per cycle\n",
@@ -710,10 +716,11 @@ static bool fence_wait(int fence) }static void fence_signal (int fd,uint32_t handle, -unsigned ring_id,const char *ring_name, -int timeout) +const struct intel_execution_engine2 *e, +const char *ring_name,int timeout) {#define NFENCES 512 +const struct intel_execution_engine2 *other;struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj;struct timespec start, now;
@@ -725,12 +732,12 @@ static void fence_signal(int fd, uint32_t handle,igt_require (gem_has_exec_fence (fd)); nengine =0 ; -if (ring_id ==ALL_ENGINES ) { -for_each_physical_engine (fd, n) - engines[nengine++] = n; +if (!e) { +__for_each_physical_engine (fd, other) + engines[nengine++] = other->flags; }else { -gem_require_ring (fd, ring_id); - engines[nengine++] = ring_id; +gem_require_ring (fd, e->flags); + engines[nengine++] = e->flags; }igt_require (nengine);
@@ -754,14 +761,14 @@ static void fence_signal(int fd, uint32_t handle,clock_gettime (CLOCK_MONOTONIC , &start); do { for (int loop = 0; loop < 1024; loop++) { - for (int e = 0; e < nengine; e++) { + for (int i = 0; i < nengine; i++) { if (fences[n] != -1) { igt_assert(fence_wait(fences[n])); close(fences[n]); } execbuf.flags &= ~ENGINE_FLAGS; - execbuf.flags |= engines[e]; + execbuf.flags |= engines[i]; gem_execbuf_wr(fd, &execbuf); /* Enable signaling by doing a poll() */
@@ -787,7 +794,8 @@ static void fence_signal(int fd, uint32_t handle, }static void preempt (int fd,uint32_t handle, -unsigned ring_id,const char *ring_name) +const struct intel_execution_engine2 *e, +const char *ring_name) {struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj;
@@ -795,7 +803,7 @@ static void preempt(int fd, uint32_t handle,unsigned long count;uint32_t ctx[2 ]; -gem_require_ring (fd, ring_id); +gem_require_ring (fd, e->flags); ctx[0 ] =gem_context_create (fd);gem_context_set_priority (fd, ctx[0 ],MIN_PRIO );
@@ -809,11 +817,11 @@ static void preempt(int fd, uint32_t handle,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (&obj); execbuf.buffer_count =1 ; - execbuf.flags = ring_id; + execbuf.flags = e->flags; execbuf.flags |=LOCAL_I915_EXEC_HANDLE_LUT ; execbuf.flags |=LOCAL_I915_EXEC_NO_RELOC ;if (__gem_execbuf (fd, &execbuf)) { - execbuf.flags = ring_id; + execbuf.flags = e->flags;gem_execbuf (fd, &execbuf); } execbuf.rsvd1 = ctx[1 ];
@@ -825,7 +833,7 @@ static void preempt(int fd, uint32_t handle,igt_spin_t *spin = __igt_spin_new(fd, .ctx = ctx[0], - .engine = ring_id); + .engine = e->flags);for (int loop =0 ; loop <1024 ; loop++)gem_execbuf (fd, &execbuf);
@@ -846,7 +854,7 @@ static void preempt(int fd, uint32_t handle, igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;uint32_t handle =0 ;int device =-1 ;
@@ -873,15 +881,15 @@ igt_main igt_subtest("basic-sequential")sequential (device, handle,0 ,5 ); -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (device, e) { igt_subtest_f("%s", e->name) -single (device, handle, e->exec_id | e->flags, e->name); +single (device, handle, e, e->name); igt_subtest_f("signal-%s", e->name) -fence_signal (device, handle, e->exec_id | e->flags, e->name,5 ); +fence_signal (device, handle, e, e->name,5 ); } igt_subtest("signal-all") -fence_signal (device, handle,ALL_ENGINES ,"all" ,150 ); +fence_signal (device, handle,NULL ,"all" ,150 ); igt_subtest("series")series (device, handle,150 );
@@ -908,9 +916,9 @@ igt_mainigt_require (gem_scheduler_has_preemption (device)); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (device, e) { igt_subtest_f("preempt-%s", e->name) -preempt (device, handle, e->exec_id | e->flags, e->name); +preempt (device, handle, e, e->name); } }
@@ -919,11 +927,10 @@ igt_mainigt_device_set_master (device); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (device, e) {/* Requires master for STORE_DWORD on gen4/5 */ igt_subtest_f("poll-%s", e->name) -poll_ring (device, - e->exec_id | e->flags, e->name,20 ); +poll_ring (device, e, e->name,20 ); }igt_subtest ("poll-sequential" )
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 23/43] tests/i915/gem_exec_nop: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:25)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
Also a uABI exercise.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 24/43 tests/i915/gem_exec_reloc: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_reloc.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-)
tests/i915/gem_exec_reloc.c +14 -13
@@ -246,7 +246,7 @@ static void check_bo(int fd, uint32_t handle)munmap (map,4096 ); } -static void active (int fd,unsigned engine) +static void active (int fd,const struct intel_execution_engine2 *e) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_relocation_entry reloc;
@@ -257,15 +257,15 @@ static void active(int fd, unsigned engine)int pass; nengine =0 ; -if (engine ==ALL_ENGINES ) { -for_each_physical_engine (fd, engine) { -if (gem_can_store_dword (fd, engine)) - engines[nengine++] = engine; +if (!e) { +__for_each_physical_engine (fd, e) { +if (gem_class_can_store_dword (fd, e->class)) + engines[nengine++] = e->flags; } }else { -igt_require (gem_has_ring (fd, engine)); -igt_require (gem_can_store_dword (fd, engine)); - engines[nengine++] = engine; +igt_require (gem_has_ring (fd, e->flags)); +igt_require (gem_class_can_store_dword (fd, e->class)); + engines[nengine++] = e->flags; }igt_require (nengine);
@@ -654,6 +654,7 @@ static void basic_softpin(int fd) igt_main { +const struct intel_execution_engine2 *e; const struct mode { const char *name; unsigned before, after;
@@ -750,12 +751,12 @@ igt_mainfrom_gpu (fd); igt_subtest("active") -active (fd,ALL_ENGINES ); -for (const struct intel_execution_engine *e = intel_execution_engines; - e->name; e++) { +active (fd,NULL ); + +__for_each_physical_engine (fd, e) igt_subtest_f("active-%s", e->name) -active (fd, e->exec_id | e->flags); - } +active (fd, e); +igt_fixture close (fd ); }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 24/43] tests/i915/gem_exec_reloc: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:26)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
Also uABI.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 25/43 tests/i915/gem_exec_schedule: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_schedule.c | 354 ++++++++++++++++----------------- 1 file changed, 177 insertions(+), 177 deletions(-)
tests/i915/gem_exec_schedule.c +177 -177
@@ -83,7 +83,8 @@ void __sync_read_u32_count(int fd, uint32_t handle, uint32_t *dst, uint64_t sizegem_read (fd, handle,0 , dst, size); } -static uint32_t __store_dword (int fd,uint32_t ctx,unsigned ring, +static uint32_t __store_dword (int fd,uint32_t ctx, +const struct intel_execution_engine2 *e,uint32_t target,uint32_t offset,uint32_t value,uint32_t cork,unsigned write_domain) {
@@ -97,7 +98,7 @@ static uint32_t __store_dword(int fd, uint32_t ctx, unsigned ring,memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj + !cork); execbuf.buffer_count =2 + !!cork; - execbuf.flags = ring; + execbuf.flags = e->flags;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ; execbuf.rsvd1 = ctx;
@@ -138,11 +139,12 @@ static uint32_t __store_dword(int fd, uint32_t ctx, unsigned ring,return obj[2 ].handle; } -static void store_dword (int fd,uint32_t ctx,unsigned ring, +static void store_dword (int fd,uint32_t ctx, +const struct intel_execution_engine2 *e,uint32_t target,uint32_t offset,uint32_t value,uint32_t cork,unsigned write_domain) { -gem_close (fd,__store_dword (fd, ctx, ring, +gem_close (fd,__store_dword (fd, ctx, e, target, offset, value, cork, write_domain)); }
@@ -161,14 +163,15 @@ static uint32_t create_highest_priority(int fd)return ctx; } -static void unplug_show_queue (int fd,struct igt_cork *c,unsigned int engine) +static void unplug_show_queue (int fd,struct igt_cork *c, +const struct intel_execution_engine2 *e) {igt_spin_t *spin[MAX_ELSP_QLEN ];for (int n =0 ; n <ARRAY_SIZE (spin); n++) {const struct igt_spin_factory opts = { .ctx =create_highest_priority (fd), - .engine = engine, + .engine = e->flags, }; spin[n] =__igt_spin_factory (fd, &opts);gem_context_destroy (fd, opts.ctx);
@@ -182,7 +185,7 @@ static void unplug_show_queue(int fd, struct igt_cork *c, unsigned int engine) } -static void fifo (int fd,unsigned ring) +static void fifo (int fd,const struct intel_execution_engine2 *e) {IGT_CORK_HANDLE (cork);uint32_t scratch, plug;
@@ -193,10 +196,10 @@ static void fifo(int fd, unsigned ring) plug =igt_cork_plug (&cork, fd);/* Same priority, same timeline, final result will be the second eb */ -store_dword (fd,0 , ring, scratch,0 ,1 , plug,0 ); -store_dword (fd,0 , ring, scratch,0 ,2 , plug,0 ); +store_dword (fd,0 , e, scratch,0 ,1 , plug,0 ); +store_dword (fd,0 , e, scratch,0 ,2 , plug,0 ); -unplug_show_queue (fd, &cork, ring); +unplug_show_queue (fd, &cork, e);gem_close (fd, plug); result =__sync_read_u32 (fd, scratch,0 );
@@ -205,15 +208,15 @@ static void fifo(int fd, unsigned ring)igt_assert_eq_u32 (result,2 ); } -static void independent (int fd,unsigned int engine) +static void independent (int fd,const struct intel_execution_engine2 *e) {IGT_CORK_HANDLE (cork);uint32_t scratch, plug, batch;igt_spin_t *spin =NULL ; -unsigned int other; +const struct intel_execution_engine2 *other;uint32_t *ptr; -igt_require (engine !=0 ); +igt_require (e); scratch =gem_create (fd,4096 ); ptr =gem_mmap__gtt (fd, scratch,4096 ,PROT_READ );
@@ -222,32 +225,32 @@ static void independent(int fd, unsigned int engine) plug =igt_cork_plug (&cork, fd);/* Check that we can submit to engine while all others are blocked */ -for_each_physical_engine (fd, other) { -if (other == engine) +__for_each_physical_engine (fd, other) { +if (other == e)continue ; -if (!gem_can_store_dword (fd, other)) +if (!gem_class_can_store_dword (fd, other->class))continue ;if (spin ==NULL ) { - spin = __igt_spin_new(fd, .engine = other); + spin = __igt_spin_new(fd, .engine = other->flags); }else {struct drm_i915_gem_execbuffer2 eb = { .buffer_count =1 , .buffers_ptr =to_user_pointer (&spin->obj[IGT_SPIN_BATCH ]), - .flags = other, + .flags = other->flags, };gem_execbuf (fd, &eb); } -store_dword (fd,0 , other, scratch,0 , other, plug,0 ); +store_dword (fd,0 , other, scratch,0 , other->flags, plug,0 ); }igt_require (spin);/* Same priority, but different timeline (as different engine) */ - batch =__store_dword (fd,0 , engine, scratch,0 , engine, plug,0 ); + batch =__store_dword (fd,0 , e, scratch,0 , e->flags, plug,0 ); -unplug_show_queue (fd, &cork, engine); +unplug_show_queue (fd, &cork, e);gem_close (fd, plug);gem_sync (fd, batch);
@@ -257,34 +260,35 @@ static void independent(int fd, unsigned int engine)/* Only the local engine should be free to complete. */ igt_assert (gem_bo_busy (fd, scratch)); -igt_assert_eq (ptr[0 ], engine); +igt_assert_eq (ptr[0 ], e->flags);igt_spin_free (fd, spin);gem_quiescent_gpu (fd);/* And we expect the others to have overwritten us, order unspecified */ igt_assert (!gem_bo_busy (fd, scratch)); -igt_assert_neq (ptr[0 ], engine); +igt_assert_neq (ptr[0 ], e->flags);munmap (ptr,4096 );gem_close (fd, scratch); } -static void smoketest (int fd,unsigned ring,unsigned timeout) +static void smoketest (int fd,const struct intel_execution_engine2 *e, +unsigned timeout) {const int ncpus =sysconf (_SC_NPROCESSORS_ONLN); -unsigned engines[MAX_ENGINES ]; +const struct intel_execution_engine2 *engines[MAX_ENGINES ];unsigned nengine; -unsigned engine; +const struct intel_execution_engine2 *other;uint32_t scratch;uint32_t result[2 * ncpus]; nengine =0 ; -if (ring ==ALL_ENGINES ) { -for_each_physical_engine (fd, engine) - engines[nengine++] = engine; +if (!e) { +__for_each_physical_engine (fd, other) + engines[nengine++] = other; }else { - engines[nengine++] = ring; + engines[nengine++] = e; }igt_require (nengine);
@@ -302,12 +306,12 @@ static void smoketest(int fd, unsigned ring, unsigned timeout) prio =hars_petruska_f54_1_random_unsafe_max (MAX_PRIO -MIN_PRIO ) +MIN_PRIO ;gem_context_set_priority (fd, ctx, prio); - engine = engines[hars_petruska_f54_1_random_unsafe_max (nengine)]; -store_dword (fd, ctx, engine, scratch, + other = engines[hars_petruska_f54_1_random_unsafe_max (nengine)]; +store_dword (fd, ctx, other, scratch,8 *child +0 , ~child,0 ,0 );for (unsigned int step =0 ; step <8 ; step++) -store_dword (fd, ctx, engine, scratch, +store_dword (fd, ctx, other, scratch,8 *child +4 , count++,0 ,0 ); }
@@ -352,7 +356,7 @@ static void semaphore_userlock(int i915) .handle = batch_create(i915), };igt_spin_t *spin =NULL ; -unsigned int engine; +const struct intel_execution_engine2 *e;uint32_t scratch;igt_require (gem_scheduler_has_semaphores (i915));
@@ -365,16 +369,16 @@ static void semaphore_userlock(int i915)*/ scratch =gem_create (i915,4096 ); -for_each_physical_engine (i915, engine) { +__for_each_physical_engine (i915, e) {if (!spin) { spin = igt_spin_new(i915, .dependency = scratch, - .engine = engine); + .engine = e->flags); }else {uint64_t saved = spin->execbuf.flags; spin->execbuf.flags &= ~ENGINE_MASK ; - spin->execbuf.flags |= engine; + spin->execbuf.flags |= e->flags;gem_execbuf (i915, &spin->execbuf);
@@ -390,15 +394,15 @@ static void semaphore_userlock(int i915)* taking precedence. */ scratch =gem_context_create (i915); -for_each_physical_engine (i915, engine) { +__for_each_physical_engine (i915, e) {struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr =to_user_pointer (&obj), .buffer_count =1 , - .flags = engine, + .flags = e->flags, .rsvd1 = scratch, }; -if (engine == (spin->execbuf.flags &ENGINE_MASK )) +if (e->flags == (spin->execbuf.flags &ENGINE_MASK ))continue ;gem_execbuf (i915, &execbuf);
@@ -415,7 +419,7 @@ static void semaphore_codependency(int i915)struct {igt_spin_t *xcs, *rcs; } task[2 ]; -unsigned int engine; +const struct intel_execution_engine2 *e, *other;int i;/*
@@ -429,13 +433,13 @@ static void semaphore_codependency(int i915)*/ i =0 ; -for_each_physical_engine (i915, engine) { +__for_each_physical_engine (i915, e) {uint32_t ctx; -if (engine ==I915_EXEC_RENDER ) +if (e->class ==I915_ENGINE_CLASS_RENDER )continue ; -if (!gem_can_store_dword (i915, engine)) +if (!gem_class_can_store_dword (i915, e->class))continue ; ctx =gem_context_create (i915);
@@ -443,16 +447,19 @@ static void semaphore_codependency(int i915) task[i].xcs = __igt_spin_new(i915, .ctx = ctx, - .engine = engine, + .engine = e->flags, .flags =IGT_SPIN_POLL_RUN );igt_spin_busywait_until_started (task[i].xcs); -/* Common rcs tasks will be queued in FIFO */ - task[i].rcs = - __igt_spin_new(i915, - .ctx = ctx, - .engine = I915_EXEC_RENDER, - .dependency = task[i].xcs->handle); +__for_each_physical_engine (i915, other) { +if (other->class ==I915_ENGINE_CLASS_RENDER ) +/* Common rcs tasks will be queued in FIFO */ + task[i].rcs = + __igt_spin_new(i915, + .ctx = ctx, + .engine = other->flags, + .dependency = task[i].xcs->handle); + }gem_context_destroy (i915, ctx);
@@ -481,7 +488,7 @@ static void semaphore_resolve(int i915) {const uint32_t SEMAPHORE_ADDR =64 <<10 ;uint32_t semaphore, outer, inner, *sema; -unsigned int engine; +const struct intel_execution_engine2 *e;/* * Userspace may submit batches that wait upon unresolved
@@ -501,7 +508,7 @@ static void semaphore_resolve(int i915) semaphore =gem_create (i915,4096 ); sema =gem_mmap__wc (i915, semaphore,0 ,4096 ,PROT_WRITE ); -for_each_physical_engine (i915, engine) { +__for_each_physical_engine (i915, e) {struct drm_i915_gem_exec_object2 obj[3 ];struct drm_i915_gem_execbuffer2 eb;uint32_t handle, cancel;
@@ -509,10 +516,10 @@ static void semaphore_resolve(int i915)igt_spin_t *spin;int64_t poke =1 ; -if (!gem_can_store_dword (i915, engine)) +if (!gem_class_can_store_dword (i915, e->class))continue ; - spin = __igt_spin_new(i915, .engine = engine); + spin = __igt_spin_new(i915, .engine = e->flags);igt_spin_end (spin);/* we just want its address for later */ gem_sync (i915, spin->handle);igt_spin_reset (spin);
@@ -609,26 +616,26 @@ static void semaphore_resolve(int i915)static void semaphore_noskip (int i915) {const int gen =intel_gen (intel_get_drm_devid (i915)); -unsigned int engine, other; +const struct intel_execution_engine2 *engine, *other;uint32_t ctx;igt_require (gen >=6 );/* MI_STORE_DWORD_IMM convenience */ ctx =gem_context_create (i915); -for_each_physical_engine (i915, engine) { -for_each_physical_engine (i915, other) { +__for_each_physical_engine (i915, engine) { +__for_each_physical_engine (i915, other) {struct drm_i915_gem_exec_object2 obj[3 ];struct drm_i915_gem_execbuffer2 eb;uint32_t handle, *cs, *map;igt_spin_t *chain, *spin; -if (other == engine || !gem_can_store_dword (i915, other)) +if (other == engine || !gem_class_can_store_dword (i915, other->class))continue ; - chain = __igt_spin_new(i915, .engine = engine); + chain = __igt_spin_new(i915, .engine = engine->flags); - spin = __igt_spin_new(i915, .engine = other); + spin = __igt_spin_new(i915, .engine = other->flags);igt_spin_end (spin);/* we just want its address for later */ gem_sync (i915, spin->handle);igt_spin_reset (spin);
@@ -662,7 +669,7 @@ static void semaphore_noskip(int i915) eb.buffer_count =3 ; eb.buffers_ptr =to_user_pointer (obj); eb.rsvd1 = ctx; - eb.flags = other; + eb.flags = other->flags;gem_execbuf (i915, &eb);/* port1: dependency chain from port0 */
@@ -673,7 +680,7 @@ static void semaphore_noskip(int i915)memset (&eb,0 ,sizeof (eb)); eb.buffer_count =2 ; eb.buffers_ptr =to_user_pointer (obj); - eb.flags = other; + eb.flags = other->flags;gem_execbuf (i915, &eb);igt_spin_set_timeout (chain,NSEC_PER_SEC /100 );
@@ -688,7 +695,8 @@ static void semaphore_noskip(int i915)gem_context_destroy (i915, ctx); } -static void reorder(int fd, unsigned ring, unsigned flags) +static void reorder (int fd,const struct intel_execution_engine2 *e, + unsigned flags)#define EQUAL 1 {IGT_CORK_HANDLE (cork);
@@ -708,10 +716,10 @@ static void reorder(int fd, unsigned ring, unsigned flags)/* We expect the high priority context to be executed first, and * so the final result will be value from the low priority context. */ -store_dword (fd, ctx[LO ], ring, scratch,0 , ctx[LO ], plug,0 ); -store_dword (fd, ctx[HI ], ring, scratch,0 , ctx[HI ], plug,0 ); +store_dword (fd, ctx[LO ], e, scratch,0 , ctx[LO ], plug,0 ); +store_dword (fd, ctx[HI ], e, scratch,0 , ctx[HI ], plug,0 ); -unplug_show_queue (fd, &cork, ring); +unplug_show_queue (fd, &cork, e);gem_close (fd, plug);gem_context_destroy (fd, ctx[LO ]);
@@ -726,7 +734,7 @@ static void reorder(int fd, unsigned ring, unsigned flags)igt_assert_eq_u32 (result, ctx[LO ]); } -static void promotion (int fd,unsigned ring) +static void promotion (int fd,const struct intel_execution_engine2 *e) {IGT_CORK_HANDLE (cork);uint32_t result, dep;
@@ -753,16 +761,16 @@ static void promotion(int fd, unsigned ring)* fifo would be NOISE, LO, HI. * strict priority would be HI, NOISE, LO */ -store_dword (fd, ctx[NOISE ], ring, result,0 , ctx[NOISE ], plug,0 ); -store_dword (fd, ctx[LO ], ring, result,0 , ctx[LO ], plug,0 ); +store_dword (fd, ctx[NOISE ], e, result,0 , ctx[NOISE ], plug,0 ); +store_dword (fd, ctx[LO ], e, result,0 , ctx[LO ], plug,0 );/* link LO <-> HI via a dependency on another buffer */ -store_dword (fd, ctx[LO ], ring, dep,0 , ctx[LO ],0 ,I915_GEM_DOMAIN_INSTRUCTION ); -store_dword (fd, ctx[HI ], ring, dep,0 , ctx[HI ],0 ,0 ); +store_dword (fd, ctx[LO ], e, dep,0 , ctx[LO ],0 ,I915_GEM_DOMAIN_INSTRUCTION ); +store_dword (fd, ctx[HI ], e, dep,0 , ctx[HI ],0 ,0 ); -store_dword (fd, ctx[HI ], ring, result,0 , ctx[HI ],0 ,0 ); +store_dword (fd, ctx[HI ], e, result,0 , ctx[HI ],0 ,0 ); -unplug_show_queue (fd, &cork, ring); +unplug_show_queue (fd, &cork, e);gem_close (fd, plug);gem_context_destroy (fd, ctx[NOISE ]);
@@ -781,7 +789,8 @@ static void promotion(int fd, unsigned ring)#define NEW_CTX (0x1 << 0)#define HANG_LP (0x1 << 1) -static void preempt (int fd,unsigned ring,unsigned flags) +static void preempt (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {uint32_t result =gem_create (fd,4096 );uint32_t result_read;
@@ -796,7 +805,7 @@ static void preempt(int fd, unsigned ring, unsigned flags)gem_context_set_priority (fd, ctx[HI ],MAX_PRIO );if (flags &HANG_LP ) - hang =igt_hang_ctx (fd, ctx[LO ], ring,0 ); + hang =igt_hang_ctx (fd, ctx[LO ], e->flags,0 );for (int n =0 ; n <ARRAY_SIZE (spin); n++) {if (flags &NEW_CTX ) {
@@ -806,10 +815,10 @@ static void preempt(int fd, unsigned ring, unsigned flags) } spin[n] = __igt_spin_new(fd, .ctx = ctx[LO], - .engine = ring); + .engine = e->flags);igt_debug ("spin[%d].handle=%d\n" , n, spin[n]->handle); -store_dword (fd, ctx[HI ], ring, result,0 , n +1 ,0 ,I915_GEM_DOMAIN_RENDER ); +store_dword (fd, ctx[HI ], e, result,0 , n +1 ,0 ,I915_GEM_DOMAIN_RENDER ); result_read =__sync_read_u32 (fd, result,0 );igt_assert_eq_u32 (result_read, n +1 );
@@ -833,21 +842,21 @@ static void preempt(int fd, unsigned ring, unsigned flags)static igt_spin_t *__noise (int fd,uint32_t ctx,int prio,igt_spin_t *spin) { -unsigned other; +const struct intel_execution_engine2 *other;gem_context_set_priority (fd, ctx, prio); -for_each_physical_engine (fd, other) { +__for_each_physical_engine (fd, other) {if (spin ==NULL ) { spin = __igt_spin_new(fd, .ctx = ctx, - .engine = other); + .engine = other->flags); }else {struct drm_i915_gem_execbuffer2 eb = { .buffer_count =1 , .buffers_ptr =to_user_pointer (&spin->obj[IGT_SPIN_BATCH ]), .rsvd1 = ctx, - .flags = other, + .flags = other->flags, };gem_execbuf (fd, &eb); }
@@ -858,21 +867,23 @@ static igt_spin_t *__noise(int fd, uint32_t ctx, int prio, igt_spin_t *spin)static void __preempt_other (int fd,uint32_t *ctx, -unsigned int target,unsigned int primary, +const struct intel_execution_engine2 *target_e, +const struct intel_execution_engine2 *primary_e,unsigned flags) {uint32_t result =gem_create (fd,4096 );uint32_t result_read[4096 /sizeof (uint32_t )]; -unsigned int n, i, other; +unsigned int n, i; +const struct intel_execution_engine2 *other; n =0 ; -store_dword (fd, ctx[LO ], primary, +store_dword (fd, ctx[LO ], primary_e, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER ); n++;if (flags &CHAIN ) { -for_each_physical_engine (fd, other) { +__for_each_physical_engine (fd, other) {store_dword (fd, ctx[LO ], other, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER );
@@ -880,7 +891,7 @@ static void __preempt_other(int fd, } } -store_dword (fd, ctx[HI ], target, +store_dword (fd, ctx[HI ], target_e, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER );
@@ -896,9 +907,10 @@ static void __preempt_other(int fd,gem_close (fd, result); } -static void preempt_other (int fd,unsigned ring,unsigned int flags) +static void preempt_other (int fd,const struct intel_execution_engine2 *e, +unsigned int flags) { -unsigned int primary; +const struct intel_execution_engine2 *primary;igt_spin_t *spin =NULL ;uint32_t ctx[3 ];
@@ -921,10 +933,9 @@ static void preempt_other(int fd, unsigned ring, unsigned int flags) ctx[HI ] =gem_context_create (fd);gem_context_set_priority (fd, ctx[HI ],MAX_PRIO ); -for_each_physical_engine (fd, primary) { -igt_debug ("Primary engine: %s\n" , e__->name); -__preempt_other (fd, ctx, ring, primary, flags); - +__for_each_physical_engine (fd, primary) { +igt_debug ("Primary engine: %s\n" , e->name); +__preempt_other (fd, ctx, e, primary, flags); }igt_assert (gem_bo_busy (fd, spin->handle));
@@ -936,13 +947,15 @@ static void preempt_other(int fd, unsigned ring, unsigned int flags) }static void __preempt_queue (int fd, -unsigned target,unsigned primary, +const struct intel_execution_engine2 *target_e, +const struct intel_execution_engine2 *primary_e,unsigned depth,unsigned flags) {uint32_t result =gem_create (fd,4096 );uint32_t result_read[4096 /sizeof (uint32_t )];igt_spin_t *above =NULL , *below =NULL ; -unsigned int other, n, i; +unsigned int n, i; +const struct intel_execution_engine2 *other;int prio =MAX_PRIO ;uint32_t ctx[3 ] = {gem_context_create (fd),
@@ -971,13 +984,13 @@ static void __preempt_queue(int fd,gem_context_set_priority (fd, ctx[LO ], prio--); n =0 ; -store_dword (fd, ctx[LO ], primary, +store_dword (fd, ctx[LO ], primary_e, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER ); n++;if (flags &CHAIN ) { -for_each_physical_engine (fd, other) { +__for_each_physical_engine (fd, other) {store_dword (fd, ctx[LO ], other, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER );
@@ -985,7 +998,7 @@ static void __preempt_queue(int fd, } } -store_dword (fd, ctx[HI ], target, +store_dword (fd, ctx[HI ], target_e, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER );
@@ -1016,22 +1029,23 @@ static void __preempt_queue(int fd,gem_close (fd, result); } -static void preempt_queue (int fd,unsigned ring,unsigned int flags) +static void preempt_queue (int fd,const struct intel_execution_engine2 *e, +unsigned int flags) { -unsigned other; +const struct intel_execution_engine2 *other; -for_each_physical_engine (fd, other) { +__for_each_physical_engine (fd, other) {for (unsigned depth =0 ; depth <=MAX_ELSP_QLEN ; depth++) -__preempt_queue (fd, ring, other, depth, flags); +__preempt_queue (fd, e, other, depth, flags); } } -static void preempt_self (int fd,unsigned ring) +static void preempt_self (int fd,const struct intel_execution_engine2 *e) {uint32_t result =gem_create (fd,4096 );uint32_t result_read[4096 /sizeof (uint32_t )];igt_spin_t *spin[MAX_ELSP_QLEN ]; -unsigned int other; +const struct intel_execution_engine2 *other;unsigned int n, i;uint32_t ctx[3 ];
@@ -1049,17 +1063,17 @@ static void preempt_self(int fd, unsigned ring) n =0 ;gem_context_set_priority (fd, ctx[HI ],MIN_PRIO ); -for_each_physical_engine (fd, other) { +__for_each_physical_engine (fd, other) { spin[n] = __igt_spin_new(fd, .ctx = ctx[NOISE], - .engine = other); + .engine = other->flags);store_dword (fd, ctx[HI ], other, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER ); n++; }gem_context_set_priority (fd, ctx[HI ],MAX_PRIO ); -store_dword (fd, ctx[HI ], ring, +store_dword (fd, ctx[HI ], e, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER );
@@ -1082,7 +1096,7 @@ static void preempt_self(int fd, unsigned ring)gem_close (fd, result); } -static void preemptive_hang (int fd,unsigned ring) +static void preemptive_hang (int fd,const struct intel_execution_engine2 *e) {igt_spin_t *spin[MAX_ELSP_QLEN ];igt_hang_t hang;
@@ -1097,12 +1111,12 @@ static void preemptive_hang(int fd, unsigned ring) spin[n] = __igt_spin_new(fd, .ctx = ctx[LO], - .engine = ring); + .engine = e->flags);gem_context_destroy (fd, ctx[LO ]); } - hang =igt_hang_ctx (fd, ctx[HI ], ring,0 ); + hang =igt_hang_ctx (fd, ctx[HI ], e->flags,0 );igt_post_hang_ring (fd, hang);for (int n =0 ; n <ARRAY_SIZE (spin); n++) {
@@ -1117,7 +1131,7 @@ static void preemptive_hang(int fd, unsigned ring)gem_context_destroy (fd, ctx[HI ]); } -static void deep (int fd,unsigned ring) +static void deep (int fd,const struct intel_execution_engine2 *e) {#define XS 8const unsigned int max_req =MAX_PRIO -MIN_PRIO ;
@@ -1138,10 +1152,10 @@ static void deep(int fd, unsigned ring) ctx[n] =gem_context_create (fd); } - nreq =gem_measure_ring_inflight (fd, ring,0 ) / (4 *XS ) *MAX_CONTEXTS ; + nreq =gem_measure_ring_inflight (fd, e->flags,0 ) / (4 *XS ) *MAX_CONTEXTS ;if (nreq > max_req) nreq = max_req; -igt_info ("Using %d requests (prio range %d)\n" , nreq, max_req); +igt_info ("Using %d requests (prio range %d)\n" , nreq, max_req); result =gem_create (fd, size);for (int m =0 ; m <XS ; m ++)
@@ -1163,7 +1177,7 @@ static void deep(int fd, unsigned ring)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =XS +2 ; - execbuf.flags = ring; + execbuf.flags = e->flags;for (n =0 ; n <MAX_CONTEXTS ; n++) { execbuf.rsvd1 = ctx[n];gem_execbuf (fd, &execbuf);
@@ -1182,7 +1196,7 @@ static void deep(int fd, unsigned ring)struct drm_i915_gem_execbuffer2 eb = { .buffers_ptr =to_user_pointer (obj), .buffer_count =3 , - .flags = ring | (gen <6 ?I915_EXEC_SECURE :0 ), + .flags = e->flags | (gen <6 ?I915_EXEC_SECURE :0 ), .rsvd1 = ctx[n %MAX_CONTEXTS ], };uint32_t batch[16 ];
@@ -1235,15 +1249,15 @@ static void deep(int fd, unsigned ring)gem_context_set_priority (fd, context,MAX_PRIO - nreq + n);for (int m =0 ; m <XS ; m++) { -store_dword (fd, context, ring, result,4 *n, context, dep[m],0 ); -store_dword (fd, context, ring, result,4 *m, context,0 ,I915_GEM_DOMAIN_INSTRUCTION ); +store_dword (fd, context, e, result,4 *n, context, dep[m],0 ); +store_dword (fd, context, e, result,4 *m, context,0 ,I915_GEM_DOMAIN_INSTRUCTION ); } expected = context; }igt_info ("Second deptree: %d requests [%.3fs]\n" , n *XS ,1e-9 *igt_nsec_elapsed (&tv)); -unplug_show_queue (fd, &cork, ring); +unplug_show_queue (fd, &cork, e);gem_close (fd, plug);igt_require (expected);/* too slow */
@@ -1281,10 +1295,10 @@ static int __execbuf(int fd, struct drm_i915_gem_execbuffer2 *execbuf)return err; } -static void wide (int fd,unsigned ring) +static void wide (int fd,const struct intel_execution_engine2 *e) {struct timespec tv = {}; -unsigned int ring_size =gem_measure_ring_inflight (fd, ring,MEASURE_RING_NEW_CTX ); +unsigned int ring_size =gem_measure_ring_inflight (fd, e->flags,MEASURE_RING_NEW_CTX );IGT_CORK_HANDLE (cork);uint32_t plug;
@@ -1306,13 +1320,13 @@ static void wide(int fd, unsigned ring)igt_seconds_elapsed (&tv) <5 && count < ring_size; count++) {for (int n =0 ; n <MAX_CONTEXTS ; n++) { -store_dword (fd, ctx[n], ring, result,4 *n, ctx[n], plug,I915_GEM_DOMAIN_INSTRUCTION ); +store_dword (fd, ctx[n], e, result,4 *n, ctx[n], plug,I915_GEM_DOMAIN_INSTRUCTION ); } }igt_info ("Submitted %d requests over %d contexts in %.1fms\n" , count,MAX_CONTEXTS ,igt_nsec_elapsed (&tv) *1e-6 ); -unplug_show_queue (fd, &cork, ring); +unplug_show_queue (fd, &cork, e);gem_close (fd, plug);for (int n =0 ; n <MAX_CONTEXTS ; n++)
@@ -1326,14 +1340,14 @@ static void wide(int fd, unsigned ring)free (ctx); } -static void reorder_wide (int fd,unsigned ring) +static void reorder_wide (int fd,const struct intel_execution_engine2 *e) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_relocation_entry reloc;struct drm_i915_gem_exec_object2 obj[3 ];struct drm_i915_gem_execbuffer2 execbuf;struct timespec tv = {}; -unsigned int ring_size =gem_measure_ring_inflight (fd, ring,MEASURE_RING_NEW_CTX ); +unsigned int ring_size =gem_measure_ring_inflight (fd, e->flags,MEASURE_RING_NEW_CTX );IGT_CORK_HANDLE (cork);uint32_t result, target, plug;uint32_t result_read[1024 ];
@@ -1360,7 +1374,7 @@ static void reorder_wide(int fd, unsigned ring)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =3 ; - execbuf.flags = ring; + execbuf.flags = e->flags;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -1414,7 +1428,7 @@ static void reorder_wide(int fd, unsigned ring)gem_context_destroy (fd, execbuf.rsvd1); } -unplug_show_queue (fd, &cork, ring); +unplug_show_queue (fd, &cork, e);gem_close (fd, plug);__sync_read_u32_count (fd, result, result_read,sizeof (result_read));
@@ -1440,7 +1454,7 @@ static void bind_to_cpu(int cpu)igt_assert (sched_setaffinity (getpid (),sizeof (cpu_set_t), &allowed) ==0 ); } -static void test_pi_ringfull (int fd,unsigned int engine) +static void test_pi_ringfull (int fd,const struct intel_execution_engine2 *e) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct sigaction sa = { .sa_handler = alarm_handler };
@@ -1463,7 +1477,7 @@ static void test_pi_ringfull(int fd, unsigned int engine) execbuf.buffers_ptr =to_user_pointer (&obj[1 ]); execbuf.buffer_count =1 ; - execbuf.flags = engine; + execbuf.flags = e->flags;/* Warm up both (hi/lo) contexts */ execbuf.rsvd1 =gem_context_create (fd);
@@ -1571,11 +1585,11 @@ static void test_pi_ringfull(int fd, unsigned int engine)static void measure_semaphore_power (int i915) {struct gpu_power power; -unsigned int engine, signaler; +const struct intel_execution_engine2 *engine, *signaler;igt_require (gpu_power_open (&power) ==0 ); -for_each_physical_engine (i915, signaler) { +__for_each_physical_engine (i915, signaler) {struct gpu_power_sample s_spin[2 ];struct gpu_power_sample s_sema[2 ];double baseline, total;
@@ -1583,7 +1597,7 @@ static void measure_semaphore_power(int i915)igt_spin_t *spin; spin = __igt_spin_new(i915, - .engine = signaler, + .engine = signaler->flags, .flags =IGT_SPIN_POLL_RUN );gem_wait (i915, spin->handle, &jiffie);/* waitboost */ igt_spin_busywait_until_started (spin);
@@ -1593,14 +1607,14 @@ static void measure_semaphore_power(int i915)gpu_power_read (&power, &s_spin[1 ]);/* Add a waiter to each engine */ -for_each_physical_engine (i915, engine) { +__for_each_physical_engine (i915, engine) {igt_spin_t *sema;if (engine == signaler)continue ; sema = __igt_spin_new(i915, - .engine = engine, + .engine = engine->flags, .dependency = spin->handle);igt_spin_free (i915, sema);
@@ -1617,7 +1631,7 @@ static void measure_semaphore_power(int i915) total =gpu_power_W (&power, &s_sema[0 ], &s_sema[1 ]);igt_info ("%s: %.1fmW + %.1fmW (total %1.fmW)\n" , - e__->name, + engine->name,1e3 * baseline,1e3 * (total - baseline),1e3 * total);
@@ -1628,7 +1642,7 @@ static void measure_semaphore_power(int i915) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd =-1 ;igt_skip_on_simulation ();
@@ -1646,21 +1660,15 @@ igt_main } igt_subtest_group { -for (e = intel_execution_engines; e->name; e++) { -/* default exec-id is purely symbolic */ -if (e->exec_id ==0 ) -continue ; - +__for_each_physical_engine (fd, e) { igt_subtest_f("fifo-%s", e->name) { -igt_require (gem_ring_has_physical_engine (fd, e->exec_id | e->flags)); -igt_require (gem_can_store_dword (fd, e->exec_id | e->flags)); -fifo (fd, e->exec_id | e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); +fifo (fd, e); } igt_subtest_f("independent-%s", e->name) { -igt_require (gem_ring_has_physical_engine (fd, e->exec_id | e->flags)); -igt_require (gem_can_store_dword (fd, e->exec_id | e->flags)); -independent (fd, e->exec_id | e->flags); +igt_require (gem_class_can_store_dword (fd, e->class)); +independent (fd, e); } } }
@@ -1681,26 +1689,22 @@ igt_mainsemaphore_noskip (fd); igt_subtest("smoketest-all") -smoketest (fd,ALL_ENGINES ,30 ); - -for (e = intel_execution_engines; e->name; e++) { -if (e->exec_id ==0 ) -continue ; +smoketest (fd,NULL ,30 ); +__for_each_physical_engine (fd, e) { igt_subtest_group { igt_fixture { -igt_require (gem_ring_has_physical_engine (fd, e->exec_id | e->flags)); -igt_require (gem_can_store_dword (fd, e->exec_id | e->flags)); +igt_require (gem_class_can_store_dword (fd, e->class)); } igt_subtest_f("in-order-%s", e->name) -reorder (fd, e->exec_id | e->flags,EQUAL ); +reorder (fd, e,EQUAL ); igt_subtest_f("out-order-%s", e->name) -reorder (fd, e->exec_id | e->flags,0 ); +reorder (fd, e,0 ); igt_subtest_f("promotion-%s", e->name) -promotion (fd, e->exec_id | e->flags); +promotion (fd, e); igt_subtest_group { igt_fixture {
@@ -1708,30 +1712,30 @@ igt_main } igt_subtest_f("preempt-%s", e->name) -preempt (fd, e->exec_id | e->flags,0 ); +preempt (fd, e,0 ); igt_subtest_f("preempt-contexts-%s", e->name) -preempt (fd, e->exec_id | e->flags,NEW_CTX ); +preempt (fd, e,NEW_CTX ); igt_subtest_f("preempt-self-%s", e->name) -preempt_self (fd, e->exec_id | e->flags); +preempt_self (fd, e); igt_subtest_f("preempt-other-%s", e->name) -preempt_other (fd, e->exec_id | e->flags,0 ); +preempt_other (fd, e,0 ); igt_subtest_f("preempt-other-chain-%s", e->name) -preempt_other (fd, e->exec_id | e->flags,CHAIN ); +preempt_other (fd, e,CHAIN ); igt_subtest_f("preempt-queue-%s", e->name) -preempt_queue (fd, e->exec_id | e->flags,0 ); +preempt_queue (fd, e,0 ); igt_subtest_f("preempt-queue-chain-%s", e->name) -preempt_queue (fd, e->exec_id | e->flags,CHAIN ); +preempt_queue (fd, e,CHAIN ); igt_subtest_f("preempt-queue-contexts-%s", e->name) -preempt_queue (fd, e->exec_id | e->flags,CONTEXTS ); +preempt_queue (fd, e,CONTEXTS ); igt_subtest_f("preempt-queue-contexts-chain-%s", e->name) -preempt_queue (fd, e->exec_id | e->flags,CONTEXTS |CHAIN ); +preempt_queue (fd, e,CONTEXTS |CHAIN ); igt_subtest_group {igt_hang_t hang;
@@ -1742,11 +1746,11 @@ igt_main } igt_subtest_f("preempt-hang-%s", e->name) { -preempt (fd, e->exec_id | e->flags,NEW_CTX |HANG_LP ); +preempt (fd, e,NEW_CTX |HANG_LP ); } igt_subtest_f("preemptive-hang-%s", e->name) -preemptive_hang (fd, e->exec_id | e->flags); +preemptive_hang (fd, e); igt_fixture {igt_disallow_hang (fd, hang);
@@ -1756,16 +1760,16 @@ igt_main } igt_subtest_f("deep-%s", e->name) -deep (fd, e->exec_id | e->flags); +deep (fd, e); igt_subtest_f("wide-%s", e->name) -wide (fd, e->exec_id | e->flags); +wide (fd, e); igt_subtest_f("reorder-wide-%s", e->name) -reorder_wide (fd, e->exec_id | e->flags); +reorder_wide (fd, e); igt_subtest_f("smoketest-%s", e->name) -smoketest (fd, e->exec_id | e->flags,5 ); +smoketest (fd, e,5 ); } } }
@@ -1779,18 +1783,14 @@ igt_mainigt_require (gem_has_execlists (fd)); } -for (e = intel_execution_engines; e->name; e++) { -if (e->exec_id ==0 ) -continue ; - +__for_each_physical_engine (fd, e) { igt_subtest_group { igt_fixture { -igt_require (gem_ring_has_physical_engine (fd, e->exec_id | e->flags));igt_require (gem_scheduler_has_preemption (fd)); } igt_subtest_f("pi-ringfull-%s", e->name) -test_pi_ringfull (fd, e->exec_id | e->flags); +test_pi_ringfull (fd, e); } } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 25/43] tests/i915/gem_exec_schedule: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:27)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
We still need to verify the legacy uABI.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 26/43 tests/i915/gem_exec_suspend: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_suspend.c | 54 +++++++++++++++++------------------ 1 file changed, 27 insertions(+), 27 deletions(-)
tests/i915/gem_exec_suspend.c +27 -27
@@ -47,7 +47,8 @@#define CACHED (1<<8)#define HANG (2<<8) -static void run_test (int fd,unsigned ring,unsigned flags); +static void run_test (int fd,const struct intel_execution_engine2 *e, +unsigned flags);static void check_bo (int fd,uint32_t handle) {
@@ -64,10 +65,10 @@ static void check_bo(int fd, uint32_t handle)static void test_all (int fd,unsigned flags) { -unsigned engine; +const struct intel_execution_engine2 *engine; -for_each_physical_engine (fd, engine) -if (gem_can_store_dword (fd, engine)) +__for_each_physical_engine (fd, engine) +if (gem_class_can_store_dword (fd, engine->class))run_test (fd, engine, flags & ~0xff ); }
@@ -86,7 +87,8 @@ static bool has_semaphores(int fd)return val >0 ; } -static void run_test (int fd,unsigned engine,unsigned flags) +static void run_test (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {const int gen =intel_gen (intel_get_drm_devid (fd));const uint32_t bbe =MI_BATCH_BUFFER_END ;
@@ -98,7 +100,7 @@ static void run_test(int fd, unsigned engine, unsigned flags)igt_spin_t *spin =NULL ; nengine =0 ; -if (engine ==ALL_ENGINES ) { +if (!e) {/* If we don't have semaphores, then every ring switch * will result in a CPU stall until the previous write * has finished. This is likely to hide any issue with
@@ -106,19 +108,19 @@ static void run_test(int fd, unsigned engine, unsigned flags)* GPU is then unlikely to be active!) */ if (has_semaphores (fd)) { -for_each_physical_engine (fd, engine) { -if (gem_can_store_dword (fd, engine)) - engines[nengine++] = engine; +__for_each_physical_engine (fd, e) { +if (gem_class_can_store_dword (fd, e->class)) + engines[nengine++] = e->flags; } }else {igt_require (gem_has_ring (fd,0 )); -igt_require (gem_can_store_dword (fd,0 )); +igt_require (gem_class_can_store_dword (fd,I915_ENGINE_CLASS_RENDER )); engines[nengine++] =0 ; } }else { -igt_require (gem_has_ring (fd, engine)); -igt_require (gem_can_store_dword (fd, engine)); - engines[nengine++] = engine; +igt_require (gem_has_ring (fd, e->flags)); +igt_require (gem_class_can_store_dword (fd, e->class)); + engines[nengine++] = e->flags; }igt_require (nengine);
@@ -189,7 +191,7 @@ static void run_test(int fd, unsigned engine, unsigned flags) }if (flags &HANG ) - spin = igt_spin_new(fd, .engine = engine); + spin = igt_spin_new(fd, .engine = e->flags);switch (mode (flags)) {case NOSLEEP :
@@ -239,7 +241,7 @@ igt_main { "-S4", HIBERNATE }, { NULL, 0 } }, *m; -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;igt_hang_t hang;int fd;
@@ -254,24 +256,22 @@ igt_main } igt_subtest("basic") -run_test (fd,ALL_ENGINES ,NOSLEEP ); +run_test (fd,NULL ,NOSLEEP ); igt_subtest("basic-S3-devices") -run_test (fd,ALL_ENGINES ,SUSPEND_DEVICES ); +run_test (fd,NULL ,SUSPEND_DEVICES ); igt_subtest("basic-S3") -run_test (fd,ALL_ENGINES ,SUSPEND ); +run_test (fd,NULL ,SUSPEND ); igt_subtest("basic-S4-devices") -run_test (fd,ALL_ENGINES ,HIBERNATE_DEVICES ); +run_test (fd,NULL ,HIBERNATE_DEVICES ); igt_subtest("basic-S4") -run_test (fd,ALL_ENGINES ,HIBERNATE ); +run_test (fd,NULL ,HIBERNATE ); -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) {for (m = modes; m->suffix; m++) { igt_subtest_f("%s-uncached%s", e->name, m->suffix) -run_test (fd, e->exec_id | e->flags, - m->mode |UNCACHED ); +run_test (fd, e, m->mode |UNCACHED ); igt_subtest_f("%s-cached%s", e->name, m->suffix) -run_test (fd, e->exec_id | e->flags, - m->mode |CACHED ); +run_test (fd, e, m->mode |CACHED ); } }
@@ -281,9 +281,9 @@ igt_main } igt_subtest("hang-S3") -run_test (fd,0 ,SUSPEND |HANG ); +run_test (fd,NULL ,SUSPEND |HANG ); igt_subtest("hang-S4") -run_test (fd,0 ,HIBERNATE |HANG ); +run_test (fd,NULL ,HIBERNATE |HANG ); igt_fixture {igt_disallow_hang (fd, hang);
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 27/43 tests/i915/gem_exec_whisper: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_exec_whisper.c | 29 +++++++++++++++-------------- 1 file changed, 15 insertions(+), 14 deletions(-)
tests/i915/gem_exec_whisper.c +15 -14
@@ -166,7 +166,8 @@ static void ctx_set_random_priority(int fd, uint32_t ctx)gem_context_set_priority (fd, ctx, prio); } -static void whisper (int fd,unsigned engine,unsigned flags) +static void whisper (int fd,const struct intel_execution_engine2 *engine, +unsigned flags) {const uint32_t bbe =MI_BATCH_BUFFER_END ;const int gen =intel_gen (intel_get_drm_devid (fd));
@@ -202,16 +203,16 @@ static void whisper(int fd, unsigned engine, unsigned flags)gpu_power_open (&power); nengine =0 ; -if (engine ==ALL_ENGINES ) { -for_each_physical_engine (fd, engine) { -if (gem_can_store_dword (fd, engine)) - engines[nengine++] = engine; +if (!engine) { +__for_each_physical_engine (fd, engine) { +if (gem_class_can_store_dword (fd, engine->class)) + engines[nengine++] = engine->flags; } }else {igt_assert (!(flags &ALL )); -igt_require (gem_has_ring (fd, engine)); -igt_require (gem_can_store_dword (fd, engine)); - engines[nengine++] = engine; +igt_require (gem_has_ring (fd, engine->flags)); +igt_require (gem_class_can_store_dword (fd, engine->class)); + engines[nengine++] = engine->flags; }igt_require (nengine);
@@ -520,6 +521,7 @@ static void whisper(int fd, unsigned engine, unsigned flags) igt_main { +const struct intel_execution_engine2 *e; const struct mode { const char *name; unsigned flags;
@@ -565,19 +567,18 @@ igt_mainfor (const struct mode *m = modes; m->name; m++) { igt_subtest_f("%s", m->name) -whisper (fd,ALL_ENGINES , m->flags); +whisper (fd,NULL , m->flags); igt_subtest_f("%s-all", m->name) -whisper (fd,ALL_ENGINES , m->flags |ALL ); +whisper (fd,NULL , m->flags |ALL ); } -for (const struct intel_execution_engine *e = intel_execution_engines; - e->name; e++) { +__for_each_physical_engine (fd, e) {for (const struct mode *m = modes; m->name; m++) {if (m->flags &CHAIN )continue ; igt_subtest_f("%s-%s", e->name, m->name) -whisper (fd, e->exec_id | e->flags, m->flags); +whisper (fd, e, m->flags); } }
@@ -590,7 +591,7 @@ igt_mainif (m->flags &INTERRUPTIBLE )continue ; igt_subtest_f("hang-%s", m->name) -whisper (fd,ALL_ENGINES , m->flags |HANG ); +whisper (fd,NULL , m->flags |HANG ); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 27/43] tests/i915/gem_exec_whisper: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:29)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
Also elements of uABI testing.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 28/43 tests/i915/gem_mocs_settings: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_mocs_settings.c | 54 +++++++++++++++------------------- 1 file changed, 23 insertions(+), 31 deletions(-)
tests/i915/gem_mocs_settings.c +23 -31
@@ -213,7 +213,7 @@ static void do_read_registers(int fd,uint32_t dst_handle, uint32_t reg_base, uint32_t size, - uint32_t engine_id) + const struct intel_execution_engine2 *engine) { struct drm_i915_gem_execbuffer2 execbuf;struct drm_i915_gem_exec_object2 obj[2 ];
@@ -236,7 +236,7 @@ static void do_read_registers(int fd, execbuf.batch_len =create_read_batch (reloc, batch, dst_handle, size, reg_base);i915_execbuffer2_set_context_id (execbuf, ctx_id); - execbuf.flags =I915_EXEC_SECURE | engine_id; + execbuf.flags =I915_EXEC_SECURE | engine->flags;gem_write (fd, handle,0 , batch, execbuf.batch_len);gem_execbuf (fd, &execbuf);
@@ -270,7 +270,7 @@ static void write_registers(int fd,uint32_t reg_base, const uint32_t *values, uint32_t size, - uint32_t engine_id, + const struct intel_execution_engine2 *engine, bool privileged) { struct drm_i915_gem_exec_object2 obj;
@@ -288,9 +288,9 @@ static void write_registers(int fd, execbuf.batch_len =create_write_batch (batch, values, size, reg_base);i915_execbuffer2_set_context_id (execbuf, ctx_id);if (privileged) - execbuf.flags =I915_EXEC_SECURE | engine_id; + execbuf.flags =I915_EXEC_SECURE | engine->flags;else - execbuf.flags = engine_id; + execbuf.flags = engine->flags;gem_write (fd, handle,0 , batch, execbuf.batch_len);gem_execbuf (fd, &execbuf);
@@ -298,11 +298,11 @@ static void write_registers(int fd, }static void check_control_registers (int fd, -unsigned engine, +const struct intel_execution_engine2 *engine,uint32_t ctx_id,bool dirty) { -const uint32_t reg_base =get_engine_base (engine); +const uint32_t reg_base =get_engine_base (engine->flags);uint32_t dst_handle =gem_create (fd,4096 );uint32_t *read_regs;struct mocs_table table;
@@ -331,7 +331,7 @@ static void check_control_registers(int fd, }static void check_l3cc_registers (int fd, -unsigned engine, +const struct intel_execution_engine2 *engine,uint32_t ctx_id,bool dirty) {
@@ -408,17 +408,17 @@ static void rc6_wait(int i915) }static void check_mocs_values (int fd, -unsigned engine,uint32_t ctx_id, +const struct intel_execution_engine2 *engine,uint32_t ctx_id,bool dirty) {check_control_registers (fd, engine, ctx_id, dirty); -if (engine ==I915_EXEC_RENDER ) +if (engine->class ==I915_ENGINE_CLASS_RENDER )check_l3cc_registers (fd, engine, ctx_id, dirty); }static void write_dirty_mocs (int fd, -unsigned engine,uint32_t ctx_id, +const struct intel_execution_engine2 *engine,uint32_t ctx_id,bool privileged) {int num_of_mocs_entries;
@@ -428,17 +428,17 @@ static void write_dirty_mocs(int fd,else num_of_mocs_entries =GEN9_NUM_MOCS_ENTRIES ; -write_registers (fd, ctx_id,get_engine_base (engine), +write_registers (fd, ctx_id,get_engine_base (engine->flags), write_values, num_of_mocs_entries, engine, privileged); -if (engine ==I915_EXEC_RENDER ) +if (engine->class ==I915_ENGINE_CLASS_RENDER )write_registers (fd, ctx_id,GEN9_LNCFCMOCS0 , write_values, num_of_mocs_entries/2 , engine, privileged); } -static void run_test (int fd,unsigned engine,unsigned flags,unsigned mode) +static void run_test (int fd,const struct intel_execution_engine2 *engine,unsigned flags,unsigned mode) {uint32_t ctx_id =0 ;uint32_t ctx_clean_id;
@@ -451,10 +451,10 @@ static void run_test(int fd, unsigned engine, unsigned flags, unsigned mode)if (flags &MOCS_DIRTY_VALUES )igt_skip_on (intel_gen (intel_get_drm_devid (fd)) >=11 ); -gem_require_ring (fd, engine); +gem_require_ring (fd, engine->flags);/* Skip if we don't know where the registers are for this engine */ -igt_require (get_engine_base (engine)); +igt_require (get_engine_base (engine->flags));if (flags &MOCS_NON_DEFAULT_CTX ) ctx_id =gem_context_create (fd);
@@ -491,7 +491,7 @@ static void run_test(int fd, unsigned engine, unsigned flags, unsigned mode)gem_context_destroy (fd, ctx_id); } -static void isolation_test (int fd,unsigned engine) +static void isolation_test (int fd,const struct intel_execution_engine2 *engine) {uint32_t ctx[2 ] = {gem_context_create (fd),gem_context_create (fd) };
@@ -505,7 +505,7 @@ static void isolation_test(int fd, unsigned engine) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;struct mocs_table table;int fd =-1 ;
@@ -516,15 +516,7 @@ igt_mainigt_require (get_mocs_settings (fd, &table, false)); } -for (e = intel_execution_engines; e->name; e++) { -/* We don't know which engine will be assigned to us if we're -* using plain I915_EXEC_BSD, I915_EXEC_DEFAULT is just -* duplicating render -*/ -if ((e->exec_id ==I915_EXEC_BSD && !e->flags) || - e->exec_id ==I915_EXEC_DEFAULT ) -continue ; - +__for_each_static_engine (e) {for (unsigned mode =NONE ; mode <MAX_MOCS_TEST_MODES ; mode++) { igt_subtest_group {igt_hang_t hang = {};
@@ -542,7 +534,7 @@ igt_main* Let's also limit testing values on non-default * contexts to render-only. */ -if (flags && e->exec_id !=I915_EXEC_RENDER ) +if (flags && e->class !=I915_ENGINE_CLASS_RENDER )continue ; igt_subtest_f("mocs-%s%s%s-%s",
@@ -553,7 +545,7 @@ igt_mainif (flags & (MOCS_NON_DEFAULT_CTX |MOCS_DIRTY_VALUES ))gem_require_contexts (fd); -run_test (fd, e->exec_id | e->flags, flags, mode); +run_test (fd, e, flags, mode); } }
@@ -565,10 +557,10 @@ igt_main } igt_subtest_f("mocs-isolation-%s", e->name) { -gem_require_ring (fd, e->exec_id | e->flags); +gem_require_ring (fd, e->flags);gem_require_contexts (fd); -isolation_test (fd, e->exec_id | e->flags); +isolation_test (fd, e); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 29/43 tests/i915/gem_reset_stats: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_reset_stats.c | 60 +++++++++++++++++------------------- 1 file changed, 29 insertions(+), 31 deletions(-)
tests/i915/gem_reset_stats.c +29 -31
@@ -73,7 +73,7 @@ static void sync_gpu(void)close (fd); } -static int noop (int fd,uint32_t ctx,const struct intel_execution_engine *e) +static int noop (int fd,uint32_t ctx,const struct intel_execution_engine2 *e) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_execbuffer2 eb;
@@ -88,7 +88,7 @@ static int noop(int fd, uint32_t ctx, const struct intel_execution_engine *e)memset (&eb,0 ,sizeof (eb)); eb.buffers_ptr =to_user_pointer (&exec); eb.buffer_count =1 ; - eb.flags = e->exec_id | e->flags; + eb.flags = e->flags;i915_execbuffer2_set_context_id (eb, ctx); ret =__gem_execbuf (fd, &eb);
@@ -102,7 +102,7 @@ static int noop(int fd, uint32_t ctx, const struct intel_execution_engine *e)static int has_engine (int fd,uint32_t ctx, -const struct intel_execution_engine *e) +const struct intel_execution_engine2 *e) {int handle =noop (fd, ctx, e);if (handle <0 )
@@ -111,7 +111,7 @@ static int has_engine(int fd,return 1 ; } -static void check_context (const struct intel_execution_engine *e) +static void check_context (const struct intel_execution_engine2 *e) {int fd =drm_open_driver (DRIVER_INTEL );
@@ -157,14 +157,14 @@ static struct timespec ts_injected;#define BAN HANG_ALLOW_BAN#define ASYNC 2static void inject_hang (int fd,uint32_t ctx, -const struct intel_execution_engine *e, +const struct intel_execution_engine2 *e,unsigned flags) {igt_hang_t hang;clock_gettime (CLOCK_MONOTONIC , &ts_injected); - hang =igt_hang_ctx (fd, ctx, e->exec_id | e->flags, flags &BAN ); + hang =igt_hang_ctx (fd, ctx, e->flags, flags &BAN );if ((flags &ASYNC ) ==0 )igt_post_hang_ring (fd, hang); }
@@ -207,7 +207,7 @@ static int _assert_reset_status(int idx, int fd, int ctx, int status)#define assert_reset_status (idx, fd, ctx, status) \ igt_assert(_assert_reset_status(idx, fd, ctx, status) == 0) -static void test_rs (const struct intel_execution_engine *e, +static void test_rs (const struct intel_execution_engine2 *e,int num_fds,int hang_index,int rs_assumed_no_hang) {int fd[MAX_FD ];
@@ -253,7 +253,7 @@ static void test_rs(const struct intel_execution_engine *e, }#define MAX_CTX 100 -static void test_rs_ctx (const struct intel_execution_engine *e, +static void test_rs_ctx (const struct intel_execution_engine2 *e,int num_fds,int num_ctx,int hang_index,int hang_context) {
@@ -329,7 +329,7 @@ static void test_rs_ctx(const struct intel_execution_engine *e, } } -static void test_ban (const struct intel_execution_engine *e) +static void test_ban (const struct intel_execution_engine2 *e) {struct local_drm_i915_reset_stats rs_bad, rs_good;int fd_bad, fd_good;
@@ -383,7 +383,7 @@ static void test_ban(const struct intel_execution_engine *e)close (fd_good); } -static void test_ban_ctx (const struct intel_execution_engine *e) +static void test_ban_ctx (const struct intel_execution_engine2 *e) {struct local_drm_i915_reset_stats rs_bad, rs_good;int fd, ban, retry =10 ;
@@ -441,7 +441,7 @@ static void test_ban_ctx(const struct intel_execution_engine *e)close (fd); } -static void test_unrelated_ctx (const struct intel_execution_engine *e) +static void test_unrelated_ctx (const struct intel_execution_engine2 *e) {int fd1,fd2;int ctx_guilty, ctx_unrelated;
@@ -480,7 +480,7 @@ static int get_reset_count(int fd, int ctx)return rs.reset_count; } -static void test_close_pending_ctx (const struct intel_execution_engine *e) +static void test_close_pending_ctx (const struct intel_execution_engine2 *e) {int fd =drm_open_driver (DRIVER_INTEL );uint32_t ctx =gem_context_create (fd);
@@ -494,7 +494,7 @@ static void test_close_pending_ctx(const struct intel_execution_engine *e)close (fd); } -static void test_close_pending (const struct intel_execution_engine *e) +static void test_close_pending (const struct intel_execution_engine2 *e) {int fd =drm_open_driver (DRIVER_INTEL );
@@ -509,7 +509,7 @@ static void noop_on_each_ring(int fd, const bool reverse)const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_execbuffer2 eb;struct drm_i915_gem_exec_object2 obj; -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;memset (&obj,0 ,sizeof (obj)); obj.handle =gem_create (fd,4096 );
@@ -520,15 +520,15 @@ static void noop_on_each_ring(int fd, const bool reverse) eb.buffer_count =1 ;if (reverse) { -for (e = intel_execution_engines; e->name; e++) +__for_each_static_engine (e) ; -while (--e >= intel_execution_engines) { - eb.flags = e->exec_id | e->flags; +while (--e >= intel_execution_engines2) { + eb.flags = e->flags;__gem_execbuf (fd, &eb); } }else { -for (e = intel_execution_engines; e->name; e++) { - eb.flags = e->exec_id | e->flags; +__for_each_static_engine (e) { + eb.flags = e->flags;__gem_execbuf (fd, &eb); } }
@@ -537,7 +537,7 @@ static void noop_on_each_ring(int fd, const bool reverse)gem_close (fd, obj.handle); } -static void test_close_pending_fork (const struct intel_execution_engine *e, +static void test_close_pending_fork (const struct intel_execution_engine2 *e,const bool reverse) {int fd =drm_open_driver (DRIVER_INTEL );
@@ -546,7 +546,7 @@ static void test_close_pending_fork(const struct intel_execution_engine *e,assert_reset_status (fd, fd,0 ,RS_NO_ERROR ); - hang =igt_hang_ctx (fd,0 , e->exec_id | e->flags,0 ); + hang =igt_hang_ctx (fd,0 , e->flags,0 );sleep (1 );/* Avoid helpers as we need to kill the child
@@ -579,7 +579,7 @@ static void test_close_pending_fork(const struct intel_execution_engine *e,close (fd); } -static void test_reset_count (const struct intel_execution_engine *e, +static void test_reset_count (const struct intel_execution_engine2 *e,const bool create_ctx) {int fd =drm_open_driver (DRIVER_INTEL );
@@ -699,23 +699,21 @@ static void test_params(void)close (fd); } -static const struct intel_execution_engine * -next_engine (int fd,const struct intel_execution_engine *e) +static const struct intel_execution_engine2 * +next_engine (int fd,const struct intel_execution_engine2 *e) {do { e++;if (e->name ==NULL ) - e = intel_execution_engines; -if (e->exec_id ==0 ) - e++; + e = intel_execution_engines2; }while (!has_engine (fd,0 , e));return e; } -static void defer_hangcheck (const struct intel_execution_engine *engine) +static void defer_hangcheck (const struct intel_execution_engine2 *engine) { -const struct intel_execution_engine *next; +const struct intel_execution_engine2 *next;int fd, count_start, count_end;int seconds =30 ;
@@ -772,7 +770,7 @@ static bool gem_has_reset_stats(int fd) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;igt_skip_on_simulation (); igt_fixture {
@@ -805,7 +803,7 @@ igt_main igt_subtest_f("params-ctx")RUN_TEST (test_params_ctx ()); -for (e = intel_execution_engines; e->name; e++) { +__for_each_static_engine (e) { igt_subtest_f("reset-stats-%s", e->name)RUN_TEST (test_rs (e,4 ,1 ,0 ));
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 29/43] tests/i915/gem_reset_stats: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:31)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
I would err on the side of saying that the legacy uABI should be verified as well. We definitely need to preserve the existing use-cases of mesa with the legacy uABI.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 30/43 tests/i915/gem_ring_sync_loop: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_ring_sync_loop.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
tests/i915/gem_ring_sync_loop.c +3 -3
@@ -44,12 +44,12 @@ sync_loop(int fd)struct drm_i915_gem_relocation_entry reloc[1 ];unsigned engines[16 ];unsigned nengine; -unsigned engine; +const struct intel_execution_engine2 *engine;int i; nengine =0 ; -for_each_physical_engine (fd, engine) - engines[nengine++] = engine; +__for_each_physical_engine (fd, engine) + engines[nengine++] = engine->flags;igt_require (nengine);memset (object,0 ,sizeof (object));
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 31/43 tests/i915/gem_ringfill: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_ringfill.c | 32 ++++++++++++++------------------ 1 file changed, 14 insertions(+), 18 deletions(-)
tests/i915/gem_ringfill.c +14 -18
@@ -96,7 +96,7 @@ static int setup_execbuf(int fd,struct drm_i915_gem_execbuffer2 *execbuf, struct drm_i915_gem_exec_object2 *obj, struct drm_i915_gem_relocation_entry *reloc, - unsigned int ring) + unsigned int flags) { const int gen = intel_gen(intel_get_drm_devid(fd));const uint32_t bbe =MI_BATCH_BUFFER_END ;
@@ -109,7 +109,7 @@ static int setup_execbuf(int fd,memset (reloc,0 ,1024 *sizeof (*reloc)); execbuf->buffers_ptr =to_user_pointer (obj); - execbuf->flags = ring | (1 <<11 ) | (1 <<12 ); + execbuf->flags = flags | (1 <<11 ) | (1 <<12 );if (gen >3 && gen <6 ) execbuf->flags |=I915_EXEC_SECURE ;
@@ -170,25 +170,25 @@ static int setup_execbuf(int fd,return 0 ; } -static void run_test (int fd,unsigned ring,unsigned flags,unsigned timeout) +static void run_test (int fd,const struct intel_execution_engine2 *e, +unsigned flags,unsigned timeout) {struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_relocation_entry reloc[1024 ];struct drm_i915_gem_execbuffer2 execbuf;igt_hang_t hang; -gem_require_ring (fd, ring); -igt_require (gem_can_store_dword (fd, ring)); +igt_require (gem_class_can_store_dword (fd, e->class));if (flags & (SUSPEND |HIBERNATE )) -run_test (fd, ring,0 ,0 ); +run_test (fd, e,0 ,0 );gem_quiescent_gpu (fd); -igt_require (setup_execbuf (fd, &execbuf, obj, reloc, ring) ==0 ); +igt_require (setup_execbuf (fd, &execbuf, obj, reloc, e->flags) ==0 );memset (&hang,0 ,sizeof (hang));if (flags &HANG ) - hang =igt_hang_ring (fd, ring & ~(3 <<13 )); + hang =igt_hang_ring (fd, e->flags & ~(3 <<13 ));if (flags & (CHILD |FORKED |BOMB )) {int nchild;
@@ -204,7 +204,7 @@ static void run_test(int fd, unsigned ring, unsigned flags, unsigned timeout) igt_fork(child, nchild) {if (flags &NEWFD ) { fd =drm_open_driver (DRIVER_INTEL ); -setup_execbuf (fd, &execbuf, obj, reloc, ring); +setup_execbuf (fd, &execbuf, obj, reloc, e->flags); }fill_ring (fd, &execbuf, flags, timeout); }
@@ -235,7 +235,7 @@ static void run_test(int fd, unsigned ring, unsigned flags, unsigned timeout)gem_quiescent_gpu (fd);if (flags & (SUSPEND |HIBERNATE )) -run_test (fd, ring,0 ,0 ); +run_test (fd, e,0 ,0 ); } igt_main
@@ -278,18 +278,14 @@ igt_main }for (m = modes; m->suffix; m++) { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e; -for (e = intel_execution_engines; e->name; e++) { -igt_subtest_f ("%s%s%s" , - m->basic && !e->exec_id ?"basic-" :"" , - e->name, - m->suffix) { +__for_each_physical_engine (fd, e) { + igt_subtest_f("%s%s", e->name, m->suffix) {igt_skip_on (m->flags &NEWFD && master);if (m->flags & (HANG |SUSPEND |HIBERNATE ))igt_skip_on_simulation (); -run_test (fd, e->exec_id | e->flags, - m->flags, m->timeout); +run_test (fd, e, m->flags, m->timeout); } } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 31/43] tests/i915/gem_ringfill: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:33)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
uABI exercise.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 32/43 tests/i915/gem_shrink: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_shrink.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-)
tests/i915/gem_shrink.c +14 -8
@@ -328,7 +328,7 @@ static void run_test(int nchildren, uint64_t alloc,igt_waitchildren (); } -static void reclaim (unsigned engine,int timeout) +static void reclaim (const struct intel_execution_engine2 *engine,int timeout) {const uint64_t timeout_100ms =100000000LL ;int fd =drm_open_driver (DRIVER_INTEL );
@@ -346,9 +346,9 @@ static void reclaim(unsigned engine, int timeout) }while (!*shared); } - spin = igt_spin_new(fd, .engine = engine); + spin = igt_spin_new(fd, .engine = engine->flags);igt_until_timeout (timeout) { - igt_spin_t *next = __igt_spin_new(fd, .engine = engine); + igt_spin_t *next = __igt_spin_new(fd, .engine = engine->flags);igt_spin_set_timeout (spin, timeout_100ms);gem_sync (fd, spin->handle);
@@ -395,12 +395,12 @@ igt_main };uint64_t alloc_size =0 ;int num_processes =0 ; +const struct intel_execution_engine2 *engine;igt_skip_on_simulation (); igt_fixture {uint64_t mem_size =intel_get_total_ram_mb (); -unsigned int engine;int fd; fd =drm_open_driver (DRIVER_INTEL );
@@ -423,15 +423,21 @@ igt_main CHECK_SWAP | CHECK_RAM); nengine =0 ; -for_each_engine (fd, engine) - engines[nengine++] = engine; +__for_each_static_engine (engine) + engines[nengine++] = engine->flags;igt_require (nengine);close (fd); } - igt_subtest("reclaim") -reclaim (I915_EXEC_DEFAULT ,2 ); + igt_subtest("reclaim") { +__for_each_static_engine (engine) { +if (engine->class ==I915_ENGINE_CLASS_RENDER ) +break ; + } + +reclaim (engine,2 ); + }for (const struct test *t = tests; t->name; t++) {for (const struct mode *m = modes; m->suffix; m++) {
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 32/43] tests/i915/gem_shrink: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:34)
tests/i915/gem_shrink.c igt_main · line 433
Quoted 3 lines
tests/i915/gem_shrink.c +1
-2- igt_subtest("reclaim") -reclaim (I915_EXEC_DEFAULT ,2 ); + igt_subtest("reclaim") {
Why change?
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 33/43 tests/i915/gem_spin_batch: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_spin_batch.c | 40 +++++++++++++++++++------------------ 1 file changed, 21 insertions(+), 19 deletions(-)
tests/i915/gem_spin_batch.c +21 -19
@@ -32,7 +32,8 @@ "'%s' != '%s' (%lld not within %d%% tolerance of %lld)\n",\#x , #ref, (long long)x, tolerance, (long long)ref) -static void spin (int fd,unsigned int engine,unsigned int timeout_sec) +static void spin (int fd,const struct intel_execution_engine2 *e, +unsigned int timeout_sec) {const uint64_t timeout_100ms =100000000LL ;unsigned long loops =0 ;
@@ -41,9 +42,9 @@ static void spin(int fd, unsigned int engine, unsigned int timeout_sec)struct timespec itv = { };uint64_t elapsed; - spin = __igt_spin_new(fd, .engine = engine); + spin = __igt_spin_new(fd, .engine = e->flags);while ((elapsed =igt_nsec_elapsed (&tv)) >>30 < timeout_sec) { - igt_spin_t *next = __igt_spin_new(fd, .engine = engine); + igt_spin_t *next = __igt_spin_new(fd, .engine = e->flags);igt_spin_set_timeout (spin, timeout_100ms -igt_nsec_elapsed (&itv));
@@ -69,13 +70,14 @@ static void spin(int fd, unsigned int engine, unsigned int timeout_sec)#define RESUBMIT_NEW_CTX (1 << 0)#define RESUBMIT_ALL_ENGINES (1 << 1) -static void spin_resubmit (int fd,unsigned int engine,unsigned int flags) +static void spin_resubmit (int fd,const struct intel_execution_engine2 *e, +unsigned int flags) {const uint32_t ctx0 =gem_context_create (fd);const uint32_t ctx1 = (flags &RESUBMIT_NEW_CTX ) ?gem_context_create (fd) : ctx0; - igt_spin_t *spin = __igt_spin_new(fd, .ctx = ctx0, .engine = engine); -unsigned int other; + igt_spin_t *spin = __igt_spin_new(fd, .ctx = ctx0, .engine = e->flags); +const struct intel_execution_engine2 *other;struct drm_i915_gem_execbuffer2 eb = { .buffer_count =1 ,
@@ -84,15 +86,15 @@ static void spin_resubmit(int fd, unsigned int engine, unsigned int flags) };if (flags &RESUBMIT_ALL_ENGINES ) { -for_each_physical_engine (fd, other) { -if (other == engine) +__for_each_physical_engine (fd, other) { +if (other == e)continue ; - eb.flags = other; + eb.flags = other->flags;gem_execbuf (fd, &eb); } }else { - eb.flags = engine; + eb.flags = e->flags;gem_execbuf (fd, &eb); }
@@ -115,9 +117,9 @@ static void spin_exit_handler(int sig)static void spin_on_all_engines (int fd,unsigned int timeout_sec) { -unsigned engine; +const struct intel_execution_engine2 *engine; -for_each_physical_engine (fd, engine) { +__for_each_physical_engine (fd, engine) { igt_fork(child, 1) {igt_install_exit_handler (spin_exit_handler);spin (fd, engine, timeout_sec);
@@ -129,7 +131,7 @@ static void spin_on_all_engines(int fd, unsigned int timeout_sec) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd =-1 ;igt_skip_on_simulation ();
@@ -140,21 +142,21 @@ igt_mainigt_fork_hang_detector (fd); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) { igt_subtest_f("basic-%s", e->name) -spin (fd, e->exec_id,3 ); +spin (fd, e,3 ); igt_subtest_f("resubmit-%s", e->name) -spin_resubmit (fd, e->exec_id,0 ); +spin_resubmit (fd, e,0 ); igt_subtest_f("resubmit-new-%s", e->name) -spin_resubmit (fd, e->exec_id,RESUBMIT_NEW_CTX ); +spin_resubmit (fd, e,RESUBMIT_NEW_CTX ); igt_subtest_f("resubmit-all-%s", e->name) -spin_resubmit (fd, e->exec_id,RESUBMIT_ALL_ENGINES ); +spin_resubmit (fd, e,RESUBMIT_ALL_ENGINES ); igt_subtest_f("resubmit-new-all-%s", e->name) -spin_resubmit (fd, e->exec_id, +spin_resubmit (fd, e,RESUBMIT_NEW_CTX |RESUBMIT_ALL_ENGINES ); }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 34/43 tests/i915/gem_storedw_loop: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_storedw_loop.c | 39 +++++++++++++++++------------------ 1 file changed, 19 insertions(+), 20 deletions(-)
tests/i915/gem_storedw_loop.c +19 -20
@@ -65,7 +65,7 @@ mmap_coherent(int fd, uint32_t handle, int size) }static void -store_dword_loop (int fd,int ring,int divider) +store_dword_loop (int fd,const struct intel_execution_engine2 *e,int divider) {int i, val =0 ;struct drm_i915_gem_execbuffer2 execbuf;
@@ -107,7 +107,7 @@ store_dword_loop(int fd, int ring, int divider)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; - execbuf.flags = ring; + execbuf.flags = e->flags;igt_info ("running storedw loop on render with stall every %i batch\n" , divider);
@@ -143,32 +143,31 @@ store_dword_loop(int fd, int ring, int divider) }static void -store_test (int fd,int ring) +store_test (int fd,const struct intel_execution_engine2 *e) { -gem_require_ring (fd, ring); -store_dword_loop (fd, ring,1 ); -store_dword_loop (fd, ring,2 ); +gem_require_ring (fd, e->flags); +store_dword_loop (fd, e,1 ); +store_dword_loop (fd, e,2 );if (!igt_run_in_simulation ()) { -store_dword_loop (fd, ring,3 ); -store_dword_loop (fd, ring,5 ); -store_dword_loop (fd, ring,7 ); -store_dword_loop (fd, ring,11 ); -store_dword_loop (fd, ring,13 ); -store_dword_loop (fd, ring,17 ); -store_dword_loop (fd, ring,19 ); +store_dword_loop (fd, e,3 ); +store_dword_loop (fd, e,5 ); +store_dword_loop (fd, e,7 ); +store_dword_loop (fd, e,11 ); +store_dword_loop (fd, e,13 ); +store_dword_loop (fd, e,17 ); +store_dword_loop (fd, e,19 ); } }static void -check_test_requirements (int fd,int ringid) +check_test_requirements (int fd,const struct intel_execution_engine2 *e) { -gem_require_ring (fd, ringid); -igt_require (gem_can_store_dword (fd, ringid)); +igt_require (gem_class_can_store_dword (fd, e->class)); } igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd; igt_fixture {
@@ -183,10 +182,10 @@ igt_mainigt_require (gem_uses_ppgtt (fd)); } -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (fd, e) { igt_subtest_f("store-%s", e->name) { -check_test_requirements (fd, e->exec_id); -store_test (fd, e->exec_id | e->flags); +check_test_requirements (fd, e); +store_test (fd, e); } }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 35/43 tests/i915/gem_userptr_blits: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_userptr_blits.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
tests/i915/gem_userptr_blits.c +4 -4
@@ -996,7 +996,7 @@ static int test_dmabuf(void)return 0 ; } -static void store_dword_rand (int i915,unsigned int engine, +static void store_dword_rand (int i915,const struct intel_execution_engine2 *engine,uint32_t target,uint64_t sz,int count) {
@@ -1025,7 +1025,7 @@ static void store_dword_rand(int i915, unsigned int engine,memset (&exec,0 ,sizeof (exec)); exec.buffer_count =2 ; exec.buffers_ptr =to_user_pointer (obj); - exec.flags = engine; + exec.flags = engine->flags;if (gen <6 ) exec.flags |=I915_EXEC_SECURE ;
@@ -1140,14 +1140,14 @@ static void test_readonly(int i915)igt_assert (mprotect (space, total,PROT_READ ) ==0 ); igt_fork(child, 1) { -unsigned int engine; +const struct intel_execution_engine2 *engine;char *orig; orig =g_compute_checksum_for_data (G_CHECKSUM_SHA1 , pages, sz);gem_userptr (i915, space, total, true, userptr_flags, &rhandle); -for_each_engine (i915, engine) { +__for_each_physical_engine (i915, engine) {char *ref, *result;/* First tweak the backing store through the write */
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 35/43] tests/i915/gem_userptr_blits: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:37)
Quoted 35 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.Signed-off-by Ramalingam C <[email protected]>--- tests/i915/gem_userptr_blits.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)tests/i915/gem_userptr_blits.c +2
-3@@ -996,7 +996,7 @@ static int test_dmabuf(void)return 0 ; } -static void store_dword_rand (int i915,unsigned int engine, +static void store_dword_rand (int i915,const struct intel_execution_engine2 *engine,uint32_t target,uint64_t sz,int count) {@@ -1025,7 +1025,7 @@ static void store_dword_rand(int i915, unsigned int engine,memset (&exec,0 ,sizeof (exec)); exec.buffer_count =2 ; exec.buffers_ptr =to_user_pointer (obj); - exec.flags = engine; + exec.flags = engine->flags;if (gen <6 ) exec.flags |=I915_EXEC_SECURE ;@@ -1140,14 +1140,14 @@ static void test_readonly(int i915)igt_assert (mprotect (space, total,PROT_READ ) ==0 ); igt_fork(child, 1) { -unsigned int engine;
tests/i915/gem_userptr_blits.c static void test_readonly(int i915) · line 1151
Quoted 10 lines
tests/i915/gem_userptr_blits.c +2
-1+const struct intel_execution_engine2 *engine;char *orig; orig =g_compute_checksum_for_data (G_CHECKSUM_SHA1 , pages, sz);gem_userptr (i915, space, total, true, userptr_flags, &rhandle); -for_each_engine (i915, engine) { +__for_each_physical_engine (i915, engine) {char *ref, *result;
As written this is testing uABI, so you need to explain why it is not. -Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 36/43 tests/i915/i915_hangman: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/i915_hangman.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
tests/i915/i915_hangman.c +3 -3
@@ -199,7 +199,7 @@ static void check_error_state(const char *expected_ring_name,igt_assert (found); } -static void test_error_state_capture (unsigned ring_id, +static void test_error_state_capture (const struct intel_execution_engine2 *e,const char *ring_name) {uint32_t *batch;
@@ -208,7 +208,7 @@ static void test_error_state_capture(unsigned ring_id,clear_error_state (); - hang =igt_hang_ctx (device,0 , ring_id,HANG_ALLOW_CAPTURE ); + hang =igt_hang_ctx (device,0 , e->flags,HANG_ALLOW_CAPTURE ); offset = hang.spin->obj[IGT_SPIN_BATCH ].offset; batch =gem_mmap__cpu (device, hang.spin->handle,0 ,4096 ,PROT_READ );
@@ -278,7 +278,7 @@ igt_main__for_each_physical_engine (device, e) igt_subtest_f("error-state-capture-%s", e->name) -test_error_state_capture (e->flags, e->name); +test_error_state_capture (e, e->name); igt_subtest("hangcheck-unterminated")hangcheck_unterminated ();
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 37/43 tests/i915/i915_module_load: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/i915_module_load.c | 29 +++++++++++++++-------------- 1 file changed, 15 insertions(+), 14 deletions(-)
tests/i915/i915_module_load.c +15 -14
@@ -42,7 +42,7 @@#define ENGINE_MASK (I915_EXEC_RING_MASK | LOCAL_I915_EXEC_BSD_MASK) -static void store_dword (int fd,unsigned ring) +static void store_dword (int fd,const struct intel_execution_engine2 *e) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_exec_object2 obj[2 ];
@@ -51,17 +51,17 @@ static void store_dword(int fd, unsigned ring)uint32_t batch[16 ];int i; -if (!gem_can_store_dword (fd, ring)) +if (!gem_class_can_store_dword (fd, e->class))return ; -if (!gem_has_ring (fd, ring)) +if (!gem_has_ring (fd, e->flags))return ;intel_detect_and_clear_missed_interrupts (fd);memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )obj; execbuf.buffer_count =2 ; - execbuf.flags = ring; + execbuf.flags = e->flags;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -113,11 +113,12 @@ static void store_all(int fd)unsigned engines[16 ], permuted[16 ];uint32_t batch[16 ];uint64_t offset; -unsigned engine, nengine; +unsigned nengine; +const struct intel_execution_engine2 *engine;int value;int i, j; -if (!gem_can_store_dword (fd,0 )) +if (!gem_class_can_store_dword (fd,I915_ENGINE_CLASS_RENDER ))return ;memset (&execbuf,0 ,sizeof (execbuf));
@@ -151,14 +152,14 @@ static void store_all(int fd) nengine =0 ;intel_detect_and_clear_missed_interrupts (fd); -for_each_engine (fd, engine) { -if (!gem_can_store_dword (fd, engine)) +__for_each_physical_engine (fd, engine) { +if (!gem_class_can_store_dword (fd, engine->class))continue ;igt_assert (2 *(nengine+1 )*sizeof (batch) <=4096 ); execbuf.flags &= ~ENGINE_MASK ; - execbuf.flags |= engine; + execbuf.flags |= engine->flags; j =2 *nengine; reloc[j].target_handle = obj[0 ].handle;
@@ -190,7 +191,7 @@ static void store_all(int fd) execbuf.batch_start_offset = j*sizeof (batch);gem_execbuf (fd, &execbuf); - engines[nengine++] = engine; + engines[nengine++] = engine->flags; }gem_sync (fd, obj[1 ].handle);
@@ -281,14 +282,14 @@ static voidgem_exec_store (void) {int fd; -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e; fd =__drm_open_driver (DRIVER_INTEL );igt_fork_hang_detector (fd); -for (e = intel_execution_engines; e->name; e++) { -if (gem_can_store_dword (fd, e->exec_id | e->flags)) -store_dword (fd, e->exec_id | e->flags); +__for_each_static_engine (e) { +if (gem_class_can_store_dword (fd, e->class)) +store_dword (fd, e); }store_all (fd);
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 38/43 tests/kms_busy: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/kms_busy.c | 62 ++++++++++++++++++++++++------------------------ 1 file changed, 31 insertions(+), 31 deletions(-)
tests/kms_busy.c +31 -31
@@ -69,7 +69,7 @@ static void do_cleanup_display(igt_display_t *dpy)static void flip_to_fb (igt_display_t *dpy,int pipe,igt_output_t *output, -struct igt_fb *fb,unsigned ring, +struct igt_fb *fb,const struct intel_execution_engine2 *e,const char *name,bool modeset) {struct pollfd pfd = { .fd = dpy->drm_fd, .events =POLLIN };
@@ -77,7 +77,7 @@ static void flip_to_fb(igt_display_t *dpy, int pipe,struct drm_event_vblank ev; igt_spin_t *t = igt_spin_new(dpy->drm_fd, - .engine = ring, + .engine = e->flags, .dependency = fb->gem_handle);if (modeset) {
@@ -134,7 +134,9 @@ static void flip_to_fb(igt_display_t *dpy, int pipe,igt_spin_free (dpy->drm_fd, t); } -static void test_flip (igt_display_t *dpy,unsigned ring,int pipe,bool modeset) +static void test_flip (igt_display_t *dpy, +const struct intel_execution_engine2 *e,int pipe, +bool modeset) {struct igt_fb fb[2 ];int warmup[] = {0 ,1 ,0 ,-1 };
@@ -167,10 +169,10 @@ static void test_flip(igt_display_t *dpy, unsigned ring, int pipe, bool modeset) }/* Make the frontbuffer busy and try to flip to itself */ -flip_to_fb (dpy, pipe, output, &fb[0 ], ring,"fb[0]" , modeset); +flip_to_fb (dpy, pipe, output, &fb[0 ], e,"fb[0]" , modeset);/* Repeat for flip to second buffer */ -flip_to_fb (dpy, pipe, output, &fb[1 ], ring,"fb[1]" , modeset); +flip_to_fb (dpy, pipe, output, &fb[1 ], e,"fb[1]" , modeset);do_cleanup_display (dpy);igt_remove_fb (dpy->drm_fd, &fb[1 ]);
@@ -178,10 +180,11 @@ static void test_flip(igt_display_t *dpy, unsigned ring, int pipe, bool modeset) }static void test_atomic_commit_hang (igt_display_t *dpy,igt_plane_t *primary, -struct igt_fb *busy_fb,unsigned ring) +struct igt_fb *busy_fb, +const struct intel_execution_engine2 *e) { igt_spin_t *t = igt_spin_new(dpy->drm_fd, - .engine = ring, + .engine = e->flags, .dependency = busy_fb->gem_handle);struct pollfd pfd = { .fd = dpy->drm_fd, .events =POLLIN };unsigned flags =0 ;
@@ -213,7 +216,7 @@ static void test_atomic_commit_hang(igt_display_t *dpy, igt_plane_t *primary,igt_spin_end (t); } -static void test_hang (igt_display_t *dpy,unsigned ring, +static void test_hang (igt_display_t *dpy,const struct intel_execution_engine2 *e,enum pipe pipe,bool modeset,bool hang_newfb) {struct igt_fb fb[2 ];
@@ -234,12 +237,12 @@ static void test_hang(igt_display_t *dpy, unsigned ring,/* Test modeset disable with hang */ igt_output_set_pipe (output,PIPE_NONE );igt_plane_set_fb (primary, &fb[1 ]); -test_atomic_commit_hang (dpy, primary, &fb[hang_newfb], ring); +test_atomic_commit_hang (dpy, primary, &fb[hang_newfb], e);/* Test modeset enable with hang */ igt_plane_set_fb (primary, &fb[0 ]);igt_output_set_pipe (output, pipe); -test_atomic_commit_hang (dpy, primary, &fb[!hang_newfb], ring); +test_atomic_commit_hang (dpy, primary, &fb[!hang_newfb], e); }else {/* * Test what happens with a single hanging pageflip.
@@ -247,7 +250,7 @@ static void test_hang(igt_display_t *dpy, unsigned ring,* timeouts taking care of it. */ igt_plane_set_fb (primary, &fb[1 ]); -test_atomic_commit_hang (dpy, primary, &fb[hang_newfb], ring); +test_atomic_commit_hang (dpy, primary, &fb[hang_newfb], e); }do_cleanup_display (dpy);
@@ -256,7 +259,8 @@ static void test_hang(igt_display_t *dpy, unsigned ring, }static void test_pageflip_modeset_hang (igt_display_t *dpy, -unsigned ring,enum pipe pipe) +const struct intel_execution_engine2 *e, +enum pipe pipe) {struct igt_fb fb;struct drm_event_vblank ev;
@@ -270,7 +274,7 @@ static void test_pageflip_modeset_hang(igt_display_t *dpy,igt_display_commit2 (dpy, dpy->is_atomic ?COMMIT_ATOMIC :COMMIT_LEGACY ); t = igt_spin_new(dpy->drm_fd, - .engine = ring, + .engine = e->flags, .dependency = fb.gem_handle);do_or_die (drmModePageFlip (dpy->drm_fd, dpy->pipes[pipe].crtc_id, fb.fb_id,DRM_MODE_PAGE_FLIP_EVENT , &fb));
@@ -291,7 +295,7 @@ igt_main {igt_display_t display = { .drm_fd =-1 , .n_pipes =IGT_MAX_PIPES };/* we only test on render */ -const struct intel_execution_engine *e = &intel_execution_engines[1 ]; +const struct intel_execution_engine2 *e = &intel_execution_engines2[0 ];enum pipe n;igt_skip_on_simulation ();
@@ -319,32 +323,28 @@ igt_mainigt_subtest_f ("basic-flip-%s" , kmstest_pipe_name(n)) { -igt_require (gem_has_ring (display.drm_fd, - e->exec_id | e->flags)); +igt_require (gem_has_ring (display.drm_fd, e->flags)); -test_flip (&display, e->exec_id | e->flags, n, false); +test_flip (&display, e, n, false); }igt_subtest_f ("basic-modeset-%s" , kmstest_pipe_name(n)) { -igt_require (gem_has_ring (display.drm_fd, - e->exec_id | e->flags)); +igt_require (gem_has_ring (display.drm_fd, e->flags)); -test_flip (&display, e->exec_id | e->flags, n, true); +test_flip (&display, e, n, true); } igt_fixture { -igt_require (gem_has_ring (display.drm_fd, - e->exec_id | e->flags)); +igt_require (gem_has_ring (display.drm_fd, e->flags)); hang =igt_allow_hang (display.drm_fd,0 ,0 ); }igt_subtest_f ("extended-pageflip-modeset-hang-oldfb-%s-%s" , e->name, kmstest_pipe_name(n)) { -igt_require (gem_has_ring (display.drm_fd, - e->exec_id | e->flags)); +igt_require (gem_has_ring (display.drm_fd, e->flags)); -test_pageflip_modeset_hang (&display, e->exec_id | e->flags, n); +test_pageflip_modeset_hang (&display, e, n); } igt_fixture
@@ -352,25 +352,25 @@ igt_mainigt_subtest_f ("extended-pageflip-hang-oldfb-%s-%s" , e->name, kmstest_pipe_name(n)) -test_hang (&display, e->exec_id | e->flags, n, false, false); +test_hang (&display, e, n, false, false);igt_subtest_f ("extended-pageflip-hang-newfb-%s-%s" , e->name, kmstest_pipe_name(n)) -test_hang (&display, e->exec_id | e->flags, n, false, true); +test_hang (&display, e, n, false, true);igt_subtest_f ("extended-modeset-hang-oldfb-%s-%s" , e->name, kmstest_pipe_name(n)) -test_hang (&display, e->exec_id | e->flags, n, true, false); +test_hang (&display, e, n, true, false);igt_subtest_f ("extended-modeset-hang-newfb-%s-%s" , e->name, kmstest_pipe_name(n)) -test_hang (&display, e->exec_id | e->flags, n, true, true); +test_hang (&display, e, n, true, true);igt_subtest_f ("extended-modeset-hang-oldfb-with-reset-%s-%s" , e->name, kmstest_pipe_name(n)) {igt_set_module_param_int ("force_reset_modeset_test" ,1 ); -test_hang (&display, e->exec_id | e->flags, n, true, false); +test_hang (&display, e, n, true, false);igt_set_module_param_int ("force_reset_modeset_test" ,0 ); }
@@ -379,7 +379,7 @@ igt_main e->name, kmstest_pipe_name(n)) {igt_set_module_param_int ("force_reset_modeset_test" ,1 ); -test_hang (&display, e->exec_id | e->flags, n, true, true); +test_hang (&display, e, n, true, true);igt_set_module_param_int ("force_reset_modeset_test" ,0 ); }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 38/43] tests/kms_busy: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:40)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
So why isn't this exercising legacy execbuffer uABI? -Chris
_______________________________________________
igt-dev mailing list [email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 39/43 tests/prime_busy: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/prime_busy.c | 24 +++++++++++++----------- 1 file changed, 13 insertions(+), 11 deletions(-)
tests/prime_busy.c +13 -11
@@ -24,6 +24,7 @@#include "igt.h" #include <sys/poll.h> +#include <i915/gem_engine_topology.h> IGT_TEST_DESCRIPTION ("Basic check of polling for prime fences." );
@@ -38,7 +39,8 @@ static bool prime_busy(struct pollfd *pfd, bool excl)#define HANG 0x4#define POLL 0x8 -static void busy (int fd,unsigned ring,unsigned flags) +static void busy (int fd,const struct intel_execution_engine2 *e, +unsigned flags) {const int gen =intel_gen (intel_get_drm_devid (fd));struct drm_i915_gem_exec_object2 obj[2 ];
@@ -57,7 +59,7 @@ static void busy(int fd, unsigned ring, unsigned flags)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr = (uintptr_t )obj; execbuf.buffer_count =2 ; - execbuf.flags = ring; + execbuf.flags = e->flags;if (gen <6 ) execbuf.flags |=I915_EXEC_SECURE ;
@@ -175,15 +177,15 @@ static void busy(int fd, unsigned ring, unsigned flags) }static void test_engine_mode (int fd, -const struct intel_execution_engine *e, +const struct intel_execution_engine2 *e,const char *name,unsigned int flags) {igt_hang_t hang = {}; igt_subtest_group { igt_fixture { -gem_require_ring (fd, e->exec_id | e->flags); -igt_require (gem_can_store_dword (fd, e->exec_id | e->flags)); +gem_require_ring (fd, e->flags); +igt_require (gem_class_can_store_dword (fd, e->class));if ((flags &HANG ) ==0 ) {
@@ -197,14 +199,14 @@ static void test_engine_mode(int fd, }igt_subtest_f ("%s%s-%s" , - !e->exec_id && !(flags &HANG ) ?"basic-" :"" , + !(flags &HANG ) ?"basic-" :"" , name, e->name) -busy (fd, e->exec_id | e->flags, flags); +busy (fd, e, flags);igt_subtest_f ("%swait-%s-%s" , - !e->exec_id && !(flags &HANG ) ?"basic-" :"" , + !(flags &HANG ) ?"basic-" :"" , name, e->name) -busy (fd, e->exec_id | e->flags, flags |POLL ); +busy (fd, e, flags |POLL ); igt_fixture {if ((flags &HANG ) ==0 )
@@ -217,7 +219,7 @@ static void test_engine_mode(int fd, igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int fd =-1 ; igt_fixture {
@@ -239,7 +241,7 @@ igt_mainigt_fixture gem_require_mmap_wc (fd ); -for (e = intel_execution_engines; e->name; e++) { +__for_each_static_engine (e) {for (const struct mode *m = modes; m->name; m++)test_engine_mode (fd, e, m->name, m->flags); }
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 39/43] tests/prime_busy: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:41)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
This is a uABI test, so I need an explanation for why the reduction is justified.
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 40/43 tests/amdgpu/amd_prime: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/amdgpu/amd_prime.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
tests/amdgpu/amd_prime.c +4 -3
@@ -173,14 +173,15 @@ static void i915_to_amd(int i915, int amd, amdgpu_device_handle device)const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_execbuffer2 execbuf; -unsigned int engines[16 ], engine; +unsigned int engines[16 ]; +const struct intel_execution_engine2 *engine;unsigned int nengine;unsigned long count;struct cork c; nengine =0 ; -for_each_physical_engine (i915, engine) - engines[nengine++] = engine; +__for_each_physical_engine (i915, engine) + engines[nengine++] = engine->flags;igt_require (nengine);memset (obj,0 ,sizeof (obj));
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 41/43 tests/i915/gem_ctx_shared: use the gem_engine_topology library
In reply to Ramalingam CReplace the legacy for_each_engine* defines with the ones implemented in the gem_engine_topology library.
--- tests/i915/gem_ctx_shared.c | 105 ++++++++++++++++++------------------ 1 file changed, 53 insertions(+), 52 deletions(-)
tests/i915/gem_ctx_shared.c +53 -52
@@ -181,7 +181,7 @@ static void exhaust_shared_gtt(int i915, unsigned int flags)igt_waitchildren (); } -static void exec_shared_gtt (int i915,unsigned int ring) +static void exec_shared_gtt (int i915,const struct intel_execution_engine2 *e) {const int gen =intel_gen (intel_get_drm_devid (i915));const uint32_t bbe =MI_BATCH_BUFFER_END ;
@@ -189,14 +189,13 @@ static void exec_shared_gtt(int i915, unsigned int ring)struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr =to_user_pointer (&obj), .buffer_count =1 , - .flags = ring, + .flags = e->flags, };uint32_t scratch, *s;uint32_t batch[16 ];int i; -gem_require_ring (i915, ring); -igt_require (gem_can_store_dword (i915, ring)); +igt_require (gem_class_can_store_dword (i915, e->class));/* Find a hole big enough for both objects later */ scratch =gem_create (i915,16384 );
@@ -310,7 +309,7 @@ static void single_timeline(int i915)struct sync_file_info sync_file_info = { .num_fences =1 , }; -unsigned int engine; +const struct intel_execution_engine2 *engine;int n;igt_require (has_single_timeline (i915));
@@ -331,7 +330,7 @@ static void single_timeline(int i915) I915_CONTEXT_CREATE_FLAGS_SINGLE_TIMELINE); execbuf.flags =I915_EXEC_FENCE_OUT ; n =0 ; -for_each_engine (i915, engine) { +for_each_context_engine (i915,0 , engine) {gem_execbuf_wr (i915, &execbuf); sync_file_info.sync_fence_info =to_user_pointer (&rings[n]);do_ioctl (execbuf.rsvd2 >>32 ,SYNC_IOC_FILE_INFO , &sync_file_info);
@@ -350,13 +349,14 @@ static void single_timeline(int i915) } } -static void exec_single_timeline (int i915,unsigned int engine) +static void exec_single_timeline (int i915, +const struct intel_execution_engine2 *engine) { -unsigned int other; +const struct intel_execution_engine2 *other;igt_spin_t *spin;uint32_t ctx; -igt_require (gem_ring_has_physical_engine (i915, engine)); +igt_require (gem_ring_has_physical_engine (i915, engine->flags));igt_require (has_single_timeline (i915));/*
@@ -365,24 +365,24 @@ static void exec_single_timeline(int i915, unsigned int engine)*/ ctx =0 ; spin =NULL ; -for_each_physical_engine (i915, other) { +__for_each_physical_engine (i915, other) {if (other == engine)continue ;if (spin ==NULL ) { - spin = __igt_spin_new(i915, .ctx = ctx, .engine = other); + spin = __igt_spin_new(i915, .ctx = ctx, .engine = other->flags); }else {struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr = spin->execbuf.buffers_ptr, .buffer_count = spin->execbuf.buffer_count, - .flags = other, + .flags = other->flags, .rsvd1 = ctx, };gem_execbuf (i915, &execbuf); } }igt_require (spin); -igt_assert_eq (nop_sync (i915, ctx, engine,NSEC_PER_SEC ),0 ); +igt_assert_eq (nop_sync (i915, ctx, engine->flags,NSEC_PER_SEC ),0 );igt_spin_free (i915, spin);/*
@@ -393,24 +393,24 @@ static void exec_single_timeline(int i915, unsigned int engine) ctx =gem_context_clone (i915,0 ,0 ,I915_CONTEXT_CREATE_FLAGS_SINGLE_TIMELINE ); spin =NULL ; -for_each_physical_engine (i915, other) { +__for_each_physical_engine (i915, other) {if (other == engine)continue ;if (spin ==NULL ) { - spin = __igt_spin_new(i915, .ctx = ctx, .engine = other); + spin = __igt_spin_new(i915, .ctx = ctx, .engine = other->flags); }else {struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr = spin->execbuf.buffers_ptr, .buffer_count = spin->execbuf.buffer_count, - .flags = other, + .flags = other->flags, .rsvd1 = ctx, };gem_execbuf (i915, &execbuf); } }igt_assert (spin); -igt_assert_eq (nop_sync (i915, ctx, engine,NSEC_PER_SEC ), -ETIME ); +igt_assert_eq (nop_sync (i915, ctx, engine->flags,NSEC_PER_SEC ), -ETIME );igt_spin_free (i915, spin); }
@@ -542,7 +542,8 @@ static uint32_t store_timestamp(int i915,return obj.handle; } -static void independent (int i915,unsigned ring,unsigned flags) +static void independent (int i915,const struct intel_execution_engine2 *e, +unsigned flags) {const int TIMESTAMP =1023 ;uint32_t handle[ARRAY_SIZE (priorities)];
@@ -550,7 +551,7 @@ static void independent(int i915, unsigned ring, unsigned flags)unsigned int mmio_base;/* XXX i915_query()! */ -switch (ring) { +switch (e->flags) {case I915_EXEC_DEFAULT :case I915_EXEC_RENDER : mmio_base =0x2000 ;
@@ -580,7 +581,7 @@ static void independent(int i915, unsigned ring, unsigned flags)for (int n =0 ; n <ARRAY_SIZE (spin); n++) {const struct igt_spin_factory opts = { .ctx =create_highest_priority (i915), - .engine = ring, + .engine = e->flags, }; spin[n] =__igt_spin_factory (i915, &opts);gem_context_destroy (i915, opts.ctx);
@@ -589,7 +590,7 @@ static void independent(int i915, unsigned ring, unsigned flags)for (int i =0 ; i <ARRAY_SIZE (priorities); i++) {uint32_t ctx =gem_queue_create (i915);gem_context_set_priority (i915, ctx, priorities[i]); - handle[i] =store_timestamp (i915, ctx, ring, mmio_base,TIMESTAMP ); + handle[i] =store_timestamp (i915, ctx, e->flags, mmio_base,TIMESTAMP );gem_context_destroy (i915, ctx); }
@@ -614,7 +615,8 @@ static void independent(int i915, unsigned ring, unsigned flags)igt_assert ((int32_t )(handle[HI ] - handle[LO ]) <0 ); } -static void reorder(int i915, unsigned ring, unsigned flags) +static void reorder (int i915,const struct intel_execution_engine2 *e, + unsigned flags)#define EQUAL 1 {IGT_CORK_HANDLE (cork);
@@ -635,10 +637,10 @@ static void reorder(int i915, unsigned ring, unsigned flags)/* We expect the high priority context to be executed first, and * so the final result will be value from the low priority context. */ -store_dword (i915, ctx[LO ], ring, scratch,0 , ctx[LO ], plug,0 ); -store_dword (i915, ctx[HI ], ring, scratch,0 , ctx[HI ], plug,0 ); +store_dword (i915, ctx[LO ], e->flags, scratch,0 , ctx[LO ], plug,0 ); +store_dword (i915, ctx[HI ], e->flags, scratch,0 , ctx[HI ], plug,0 ); -unplug_show_queue (i915, &cork, ring); +unplug_show_queue (i915, &cork, e->flags);gem_close (i915, plug);gem_context_destroy (i915, ctx[LO ]);
@@ -656,7 +658,7 @@ static void reorder(int i915, unsigned ring, unsigned flags)munmap (ptr,4096 ); } -static void promotion (int i915,unsigned ring) +static void promotion (int i915,const struct intel_execution_engine2 *e) {IGT_CORK_HANDLE (cork);uint32_t result, dep;
@@ -683,16 +685,16 @@ static void promotion(int i915, unsigned ring)* fifo would be NOISE, LO, HI. * strict priority would be HI, NOISE, LO */ -store_dword (i915, ctx[NOISE ], ring, result,0 , ctx[NOISE ], plug,0 ); -store_dword (i915, ctx[LO ], ring, result,0 , ctx[LO ], plug,0 ); +store_dword (i915, ctx[NOISE ], e->flags, result,0 , ctx[NOISE ], plug,0 ); +store_dword (i915, ctx[LO ], e->flags, result,0 , ctx[LO ], plug,0 );/* link LO <-> HI via a dependency on another buffer */ -store_dword (i915, ctx[LO ], ring, dep,0 , ctx[LO ],0 ,I915_GEM_DOMAIN_INSTRUCTION ); -store_dword (i915, ctx[HI ], ring, dep,0 , ctx[HI ],0 ,0 ); +store_dword (i915, ctx[LO ], e->flags, dep,0 , ctx[LO ],0 ,I915_GEM_DOMAIN_INSTRUCTION ); +store_dword (i915, ctx[HI ], e->flags, dep,0 , ctx[HI ],0 ,0 ); -store_dword (i915, ctx[HI ], ring, result,0 , ctx[HI ],0 ,0 ); +store_dword (i915, ctx[HI ], e->flags, result,0 , ctx[HI ],0 ,0 ); -unplug_show_queue (i915, &cork, ring); +unplug_show_queue (i915, &cork, e->flags);gem_close (i915, plug);gem_context_destroy (i915, ctx[NOISE ]);
@@ -716,18 +718,18 @@ static void promotion(int i915, unsigned ring)munmap (ptr,4096 ); } -static void smoketest (int i915,unsigned ring,unsigned timeout) +static void smoketest (int i915,unsigned timeout) {const int ncpus =sysconf (_SC_NPROCESSORS_ONLN);unsigned engines[16 ];unsigned nengine; -unsigned engine; +const struct intel_execution_engine2 *engine;uint32_t scratch;uint32_t *ptr; nengine =0 ; -for_each_physical_engine (i915, engine) - engines[nengine++] = engine; +__for_each_physical_engine (i915, engine) + engines[nengine++] = engine->flags;igt_require (nengine); scratch =gem_create (i915,4096 );
@@ -740,16 +742,17 @@ static void smoketest(int i915, unsigned ring, unsigned timeout) ctx =gem_queue_create (i915);igt_until_timeout (timeout) {int prio; +unsigned flags; prio =hars_petruska_f54_1_random_unsafe_max (MAX_PRIO -MIN_PRIO ) +MIN_PRIO ;gem_context_set_priority (i915, ctx, prio); - engine = engines[hars_petruska_f54_1_random_unsafe_max (nengine)]; -store_dword (i915, ctx, engine, scratch, + flags = engines[hars_petruska_f54_1_random_unsafe_max (nengine)]; +store_dword (i915, ctx, flags, scratch,8 *child +0 , ~child,0 ,0 );for (unsigned int step =0 ; step <8 ; step++) -store_dword (i915, ctx, engine, scratch, +store_dword (i915, ctx, flags, scratch,8 *child +4 , count++,0 ,0 ); }
@@ -777,7 +780,7 @@ static void smoketest(int i915, unsigned ring, unsigned timeout) igt_main { -const struct intel_execution_engine *e; +const struct intel_execution_engine2 *e;int i915 =-1 ; igt_fixture {
@@ -809,13 +812,12 @@ igt_main igt_subtest("exhaust-shared-gtt-lrc")exhaust_shared_gtt (i915,EXHAUST_LRC ); -for (e = intel_execution_engines; e->name; e++) { +__for_each_physical_engine (i915, e) { igt_subtest_f("exec-shared-gtt-%s", e->name) -exec_shared_gtt (i915, e->exec_id | e->flags); +exec_shared_gtt (i915, e); igt_subtest_f("exec-single-timeline-%s", e->name) -exec_single_timeline (i915, - e->exec_id | e->flags); +exec_single_timeline (i915, e);/* * Check that the shared contexts operate independently,
@@ -826,33 +828,32 @@ igt_main*/ igt_subtest_group { igt_fixture { -gem_require_ring (i915, e->exec_id | e->flags); -igt_require (gem_can_store_dword (i915, e->exec_id | e->flags)); +igt_require (gem_class_can_store_dword (i915, e->class));igt_require (gem_scheduler_enabled (i915));igt_require (gem_scheduler_has_ctx_priority (i915)); } igt_subtest_f("Q-independent-%s", e->name) -independent (i915, e->exec_id | e->flags,0 ); +independent (i915, e,0 ); igt_subtest_f("Q-in-order-%s", e->name) -reorder (i915, e->exec_id | e->flags,EQUAL ); +reorder (i915, e,EQUAL ); igt_subtest_f("Q-out-order-%s", e->name) -reorder (i915, e->exec_id | e->flags,0 ); +reorder (i915, e,0 ); igt_subtest_f("Q-promotion-%s", e->name) -promotion (i915, e->exec_id | e->flags); +promotion (i915, e); igt_subtest_f("Q-smoketest-%s", e->name) -smoketest (i915, e->exec_id | e->flags,5 ); +smoketest (i915,5 ); } } igt_subtest("Q-smoketest-all") {igt_require (gem_scheduler_enabled (i915));igt_require (gem_scheduler_has_ctx_priority (i915)); -smoketest (i915,-1 ,30 ); +smoketest (i915,30 ); } igt_fixture {
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Re: [igt-dev] [RFC v2 41/43] tests/i915/gem_ctx_shared: use the gem_engine_topology library
Quoting Ramalingam C (2019-06-21 11:03:43)
Quoted 2 lines
Replace the legacy for_each_engine* defines with the ones implemented in the
gem_engine_topologylibrary.
Where's my legacy testing at?
-Chris
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 42/43 lib: replace the legacy funcs with gem_engine_topology libraries
In reply to Ramalingam CAll legacy functions are replaced with the functions from the gem_engine_topology library.
--- benchmarks/gem_syslatency.c | 7 ++++--- lib/i915/gem_ring.c | 7 +++++-- lib/i915/gem_submission.c | 6 ++++-- lib/igt_gt.c | 11 ----------- lib/igt_gt.h | 21 --------------------- 5 files changed, 13 insertions(+), 39 deletions(-)
benchmarks/gem_syslatency.c +4 -3
@@ -100,14 +100,15 @@ static void *gem_busyspin(void *arg) bs->sz ? bs->sz +sizeof (bbe) : bs->leak ?16 <<20 :4 <<10 ;unsigned engines[16 ];unsigned nengine; -unsigned engine; +const struct intel_execution_engine2 *engine;int fd; fd =drm_open_driver (DRIVER_INTEL ); nengine =0 ; -for_each_engine (fd, engine) -if (!ignore_engine (fd, engine)) engines[nengine++] = engine; +for_each_context_engine (fd,0 , engine) +if (!ignore_engine (fd, engine->flags)) + engines[nengine++] = engine->flags;memset (obj,0 ,sizeof (obj)); obj[0 ].handle =gem_create (fd,4096 );
lib/i915/gem_ring.c +5 -2
@@ -33,6 +33,8 @@#include "igt_dummyload.h" #include "igt_gt.h" +#include <i915/gem_engine_topology.h> +static int __execbuf (int fd,struct drm_i915_gem_execbuffer2 *execbuf) {int err;
@@ -141,9 +143,10 @@ gem_measure_ring_inflight(int fd, unsigned int engine, enum measure_ring_flags f {if (engine ==ALL_ENGINES ) {unsigned int global_min = ~0u ; +const struct intel_execution_engine2 *e; -for_each_physical_engine (fd, engine) { -unsigned int engine_min =__gem_measure_ring_inflight (fd, engine, flags); +for_each_context_engine (fd,0 , e) { +unsigned int engine_min =__gem_measure_ring_inflight (fd, e->flags, flags);if (engine_min < global_min) global_min = engine_min;
lib/i915/gem_submission.c +4 -2
@@ -36,6 +36,7 @@#include "ioctl_wrappers.h" #include "i915/gem_submission.h" +#include "i915/gem_engine_topology.h" /** * SECTION:gem_submission
@@ -207,6 +208,7 @@ void gem_test_engine(int i915, unsigned int engine) .buffers_ptr = to_user_pointer(&obj), .buffer_count = 1, }; +const struct intel_execution_engine2 *e; i915 =gem_reopen_driver (i915);igt_assert (!is_wedged (i915));
@@ -215,8 +217,8 @@ void gem_test_engine(int i915, unsigned int engine)gem_write (i915, obj.handle,0 , &bbe,sizeof (bbe));if (engine ==ALL_ENGINES ) { -for_each_physical_engine (i915, engine) { - execbuf.flags = engine; +__for_each_physical_engine (i915, e) { + execbuf.flags = e->flags;gem_execbuf (i915, &execbuf); } }else {
lib/igt_gt.c +0 -11
@@ -546,17 +546,6 @@ unsigned intel_detect_and_clear_missed_interrupts(int fd)return missed; } -const struct intel_execution_engine intel_execution_engines[] = { - {"default" ,NULL ,0 ,0 }, - {"render" ,"rcs0" ,I915_EXEC_RENDER ,0 }, - {"bsd" ,"vcs0" ,I915_EXEC_BSD ,0 }, - {"bsd1" ,"vcs0" ,I915_EXEC_BSD ,1 <<13 /*I915_EXEC_BSD_RING1*/ }, - {"bsd2" ,"vcs1" ,I915_EXEC_BSD ,2 <<13 /*I915_EXEC_BSD_RING2*/ }, - {"blt" ,"bcs0" ,I915_EXEC_BLT ,0 }, - {"vebox" ,"vecs0" ,I915_EXEC_VEBOX ,0 }, - {NULL ,0 ,0 } -}; -bool gem_class_can_store_dword (int fd,int class) {uint16_t devid =intel_get_drm_devid (fd);
lib/igt_gt.h +0 -21
@@ -65,27 +65,6 @@ unsigned intel_detect_and_clear_missed_interrupts(int fd);#define ALL_ENGINES ~0u/* Use in interfaces to iterate all engines */ -extern const struct intel_execution_engine { -const char *name; -const char *full_name; -unsigned exec_id; -unsigned flags; -} intel_execution_engines[]; - -#define for_if (expr__) if (!(expr__)) {} else - -#define for_each_engine (fd__, flags__) \ - for (const struct intel_execution_engine *e__ = intel_execution_engines;\ - e__->name; \ - e__++) \ - for_if (gem_has_ring(fd__, flags__ = e__->exec_id | e__->flags)) - -#define for_each_physical_engine (fd__, flags__) \ - for (const struct intel_execution_engine *e__ = intel_execution_engines;\ - e__->name; \ - e__++) \ - for_if (gem_ring_has_physical_engine(fd__, flags__ = e__->exec_id | e__->flags)) -bool gem_ring_is_physical_engine (int fd,unsigned int ring);bool gem_ring_has_physical_engine (int fd,unsigned int ring);
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
RFC v2 43/43 lib/gem_engine_topology: rename the macros
In reply to Ramalingam C%s/__for_each_physical_engine/for_each_physical_engine %s/__for_each_static_engine/for_each_static_engine
--- lib/i915/gem_engine_topology.c | 6 ++--- lib/i915/gem_engine_topology.h | 4 ++-- lib/i915/gem_submission.c | 2 +- tests/amdgpu/amd_prime.c | 2 +- tests/i915/gem_bad_reloc.c | 2 +- tests/i915/gem_busy.c | 20 ++++++++--------- tests/i915/gem_concurrent_all.c | 2 +- tests/i915/gem_cs_prefetch.c | 2 +- tests/i915/gem_cs_tlb.c | 2 +- tests/i915/gem_ctx_create.c | 4 ++-- tests/i915/gem_ctx_engines.c | 10 ++++----- tests/i915/gem_ctx_exec.c | 2 +- tests/i915/gem_ctx_shared.c | 8 +++---- tests/i915/gem_ctx_switch.c | 4 ++-- tests/i915/gem_ctx_thrash.c | 4 ++-- tests/i915/gem_eio.c | 12 +++++----- tests/i915/gem_exec_async.c | 4 ++-- tests/i915/gem_exec_await.c | 2 +- tests/i915/gem_exec_basic.c | 8 +++---- tests/i915/gem_exec_capture.c | 2 +- tests/i915/gem_exec_create.c | 2 +- tests/i915/gem_exec_fence.c | 12 +++++----- tests/i915/gem_exec_flush.c | 2 +- tests/i915/gem_exec_gttfill.c | 4 ++-- tests/i915/gem_exec_latency.c | 6 ++--- tests/i915/gem_exec_nop.c | 16 ++++++------- tests/i915/gem_exec_parallel.c | 4 ++-- tests/i915/gem_exec_params.c | 4 ++-- tests/i915/gem_exec_reloc.c | 4 ++-- tests/i915/gem_exec_reuse.c | 2 +- tests/i915/gem_exec_schedule.c | 40 ++++++++++++++++----------------- tests/i915/gem_exec_store.c | 4 ++-- tests/i915/gem_exec_suspend.c | 6 ++--- tests/i915/gem_exec_whisper.c | 4 ++-- tests/i915/gem_mocs_settings.c | 2 +- tests/i915/gem_reset_stats.c | 6 ++--- tests/i915/gem_ring_sync_loop.c | 2 +- tests/i915/gem_ringfill.c | 2 +- tests/i915/gem_shrink.c | 4 ++-- tests/i915/gem_spin_batch.c | 6 ++--- tests/i915/gem_storedw_loop.c | 2 +- tests/i915/gem_sync.c | 24 ++++++++++---------- tests/i915/gem_userptr_blits.c | 2 +- tests/i915/gem_wait.c | 4 ++-- tests/i915/i915_hangman.c | 2 +- tests/i915/i915_module_load.c | 4 ++-- tests/perf_pmu.c | 12 +++++----- tests/prime_busy.c | 2 +- tests/prime_vgem.c | 4 ++-- 49 files changed, 145 insertions(+), 145 deletions(-)
lib/i915/gem_engine_topology.c +3 -3
@@ -116,7 +116,7 @@ static void init_engine(struct intel_execution_engine2 *e2,return ; } -__for_each_static_engine (__e2) +for_each_static_engine (__e2)if (__e2->class == class && __e2->instance == instance)break ;
@@ -219,7 +219,7 @@ struct intel_engine_data intel_init_engine_list(int fd, uint32_t ctx_id)igt_debug ("using pre-allocated engine list\n" ); -__for_each_static_engine (e2) { +for_each_static_engine (e2) {struct intel_execution_engine2 *__e2 = &engine_data.engines[engine_data.nengines];
@@ -298,7 +298,7 @@ int gem_engine_class_to_execbuf_flags(int fd, uint32_t ctx_id, int class)int i;if (gem_topology_get_param (fd, ¶m)) { -__for_each_static_engine (e) { +for_each_static_engine (e) {if (e->class == class)return e->flags; }
lib/i915/gem_engine_topology.h +2 -2
@@ -55,7 +55,7 @@ void gem_context_set_all_engines(int fd, uint32_t ctx);int gem_engine_class_to_execbuf_flags (int fd,uint32_t ctx_id,int class); -#define __for_each_static_engine (e__) \ +#define for_each_static_engine (e__) \ for ((e__) = intel_execution_engines2; (e__)->name; (e__)++)#define for_each_context_engine (fd__, ctx__, e__) \
@@ -64,7 +64,7 @@ int gem_engine_class_to_execbuf_flags(int fd, uint32_t ctx_id, int class); intel_next_engine(&i__))/* needs to replace "for_each_physical_engine" when conflicts are fixed */ -#define __for_each_physical_engine (fd__, e__) \ +#define for_each_physical_engine (fd__, e__) \ for (struct intel_engine_data i__ = intel_init_engine_list(fd__, 0); \ ((e__) = intel_get_current_physical_engine(&i__)); \ intel_next_engine(&i__))
lib/i915/gem_submission.c +1 -1
@@ -217,7 +217,7 @@ void gem_test_engine(int i915, unsigned int engine)gem_write (i915, obj.handle,0 , &bbe,sizeof (bbe));if (engine ==ALL_ENGINES ) { -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) { execbuf.flags = e->flags;gem_execbuf (i915, &execbuf); }
tests/amdgpu/amd_prime.c +1 -1
@@ -180,7 +180,7 @@ static void i915_to_amd(int i915, int amd, amdgpu_device_handle device)struct cork c; nengine =0 ; -__for_each_physical_engine (i915, engine) +for_each_physical_engine (i915, engine) engines[nengine++] = engine->flags;igt_require (nengine);
tests/i915/gem_bad_reloc.c +1 -1
@@ -193,7 +193,7 @@ igt_mainigt_require_gem (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("negative-reloc-%s", e->name)negative_reloc (fd, e,0 );
tests/i915/gem_busy.c +10 -10
@@ -134,7 +134,7 @@ static void semaphore(int fd, const struct intel_execution_engine2 *e)/* Now queue it for a read across all available rings */ active =0 ; -__for_each_physical_engine (fd, __e) { +for_each_physical_engine (fd, __e) {if (exec_noop (fd, handle, __e->flags, false)) active |=1 << __e->class; }
@@ -251,7 +251,7 @@ static void one(int fd, const struct intel_execution_engine2 *e, unsigned test_fif (test_flags &PARALLEL ) {struct intel_execution_engine2 *e2; -__for_each_physical_engine (fd, e2) { +for_each_physical_engine (fd, e2) {if (e2->class == e->class && e2->instance == e->instance)continue ;
@@ -320,7 +320,7 @@ static void close_race(int fd)*/ nengine =0 ; -__for_each_static_engine (e) +for_each_static_engine (e) engines[nengine++] = e->flags;igt_require (nengine);
@@ -465,7 +465,7 @@ static void all(int i915) {const struct intel_execution_engine2 *e; -__for_each_physical_engine (i915, e) +for_each_physical_engine (i915, e)basic (i915, e,0 ); }
@@ -491,7 +491,7 @@ igt_mainall (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_group { igt_subtest_f("busy-%s", e->name) {gem_quiescent_gpu (fd);
@@ -506,7 +506,7 @@ igt_maingem_require_mmap_wc (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("extended-%s", e->name) {igt_require (gem_class_can_store_dword (fd, e->class));
@@ -516,7 +516,7 @@ igt_main } } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("extended-parallel-%s", e->name) {igt_require (gem_class_can_store_dword (fd, e->class));
@@ -533,7 +533,7 @@ igt_mainigt_require (has_semaphores (fd)); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("extended-semaphore-%s", e->name)semaphore (fd, e); }
@@ -554,7 +554,7 @@ igt_main hang =igt_allow_hang (fd,0 ,0 ); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {igt_subtest_f ("%shang-%s" , e->class ==I915_ENGINE_CLASS_RENDER ? "basic-" : "", e->name) {
@@ -570,7 +570,7 @@ igt_maingem_require_mmap_wc (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("extended-hang-%s", e->name) {igt_skip_on_simulation ();igt_require (gem_class_can_store_dword (fd, e->class));
tests/i915/gem_concurrent_all.c +1 -1
@@ -951,7 +951,7 @@ static igt_hang_t all_hang(void)igt_hang_t hang =igt_hang_ring (fd,I915_EXEC_RENDER );const struct intel_execution_engine2 *e; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {struct drm_i915_gem_execbuffer2 eb = hang.spin->execbuf;if (e->class ==I915_ENGINE_CLASS_RENDER )
tests/i915/gem_cs_prefetch.c +1 -1
@@ -143,7 +143,7 @@ igt_mainigt_skip_on_simulation (); -__for_each_static_engine (e) +for_each_static_engine (e) igt_subtest_f("%s", e->name)test_ring (e); }
tests/i915/gem_cs_tlb.c +1 -1
@@ -150,7 +150,7 @@ igt_mainigt_require_gem (fd); } -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) igt_subtest_f("%s", e->name)run_on_ring (fd, e->flags, e->name);
tests/i915/gem_ctx_create.c +2 -2
@@ -532,7 +532,7 @@ igt_mainigt_require_gem (fd);gem_require_contexts (fd); -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) all_engines[all_nengine++] = e->flags;igt_require (all_nengine);
@@ -578,7 +578,7 @@ igt_main igt_subtest("forked-active-all")active (fd,NULL ,120 , ncpus); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("active-%s", e->name)active (fd, e,20 ,1 );igt_subtest_f ("forked-active-%s" , e->name)
tests/i915/gem_ctx_engines.c +5 -5
@@ -193,7 +193,7 @@ static void idempotent(int i915) idx =0 ;memset (&engines,0 ,sizeof (engines)); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) { engines.engines[idx].engine_class = e->class; engines.engines[idx].engine_instance = e->instance; idx++;
@@ -269,7 +269,7 @@ static void execute_one(int i915)gem_execbuf (i915, &execbuf);gem_sync (i915, obj.handle); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {struct drm_i915_gem_busy busy = { .handle = obj.handle };for (int i =-1 ; i <=I915_EXEC_RING_MASK ; i++) {
@@ -338,7 +338,7 @@ static void execute_oneforall(int i915) };const struct intel_execution_engine2 *e; -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {memset (&engines,0 ,sizeof (engines));for (int i =0 ; i <=I915_EXEC_RING_MASK ; i++) {engine_class (&engines, i) = e->class;
@@ -378,7 +378,7 @@ static void execute_allforone(int i915) i =0 ;memset (&engines,0 ,sizeof (engines)); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {engine_class (&engines, i) = e->class;engine_instance (&engines, i) = e->instance; i++;
@@ -387,7 +387,7 @@ static void execute_allforone(int i915)gem_context_set_param (i915, ¶m); i =0 ; -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {struct drm_i915_gem_busy busy = {};igt_spin_t *spin;
tests/i915/gem_ctx_exec.c +1 -1
@@ -229,7 +229,7 @@ igt_maingem_sync (fd, handle); } -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) igt_subtest_f("basic-invalid-context-%s", e->name)invalid_context (fd, e, handle);
tests/i915/gem_ctx_shared.c +4 -4
@@ -365,7 +365,7 @@ static void exec_single_timeline(int i915,*/ ctx =0 ; spin =NULL ; -__for_each_physical_engine (i915, other) { +for_each_physical_engine (i915, other) {if (other == engine)continue ;
@@ -393,7 +393,7 @@ static void exec_single_timeline(int i915, ctx =gem_context_clone (i915,0 ,0 ,I915_CONTEXT_CREATE_FLAGS_SINGLE_TIMELINE ); spin =NULL ; -__for_each_physical_engine (i915, other) { +for_each_physical_engine (i915, other) {if (other == engine)continue ;
@@ -728,7 +728,7 @@ static void smoketest(int i915, unsigned timeout)uint32_t *ptr; nengine =0 ; -__for_each_physical_engine (i915, engine) +for_each_physical_engine (i915, engine) engines[nengine++] = engine->flags;igt_require (nengine);
@@ -812,7 +812,7 @@ igt_main igt_subtest("exhaust-shared-gtt-lrc")exhaust_shared_gtt (i915,EXHAUST_LRC ); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) { igt_subtest_f("exec-shared-gtt-%s", e->name)exec_shared_gtt (i915, e);
tests/i915/gem_ctx_switch.c +2 -2
@@ -229,7 +229,7 @@ static void all(int fd, uint32_t handle, unsigned flags, int timeout)int n, qlen; nengine =0 ; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { engine[nengine] = e->flags; name[nengine] = e->name; nengine++;
@@ -337,7 +337,7 @@ igt_mainigt_fork_hang_detector (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {for (typeof (*phases) *p = phases; p->name; p++) { igt_subtest_group { igt_fixture {
tests/i915/gem_ctx_thrash.c +2 -2
@@ -105,7 +105,7 @@ static void single(const char *name, bool all_engines)if (all_engines) {const struct intel_execution_engine2 *e; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -231,7 +231,7 @@ static void processes(void) fd =drm_open_driver (DRIVER_INTEL ); num_engines =0 ; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))break ;
tests/i915/gem_eio.c +6 -6
@@ -345,7 +345,7 @@ static void __test_banned(int fd) }/* Trigger a reset, making sure we are detected as guilty */ -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e)if (e->class ==I915_ENGINE_CLASS_RENDER ) { hang =spin_sync (fd,0 , e);break ;
@@ -389,7 +389,7 @@ static void test_wait(int fd, unsigned int flags, unsigned int wait)else igt_require (i915_reset_control (true)); -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e)if (e->class ==I915_ENGINE_CLASS_RENDER ){ hang =spin_sync (fd,0 , e);break ;
@@ -438,7 +438,7 @@ static void test_inflight(int fd, unsigned int wait)igt_require (max >1 ); max =min (max -1 ,ARRAY_SIZE (fence)); -__for_each_physical_engine (parent_fd, engine) { +for_each_physical_engine (parent_fd, engine) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_execbuffer2 execbuf;
@@ -566,7 +566,7 @@ static void test_inflight_contexts(int fd, unsigned int wait)igt_require (gem_has_exec_fence (fd));gem_require_contexts (fd); -__for_each_physical_engine (parent_fd, engine) { +for_each_physical_engine (parent_fd, engine) {const uint32_t bbe =MI_BATCH_BUFFER_END ;struct drm_i915_gem_exec_object2 obj[2 ];struct drm_i915_gem_execbuffer2 execbuf;
@@ -706,7 +706,7 @@ static void test_inflight_internal(int fd, unsigned int wait)memset (&execbuf,0 ,sizeof (execbuf)); execbuf.buffers_ptr =to_user_pointer (obj); execbuf.buffer_count =2 ; -__for_each_physical_engine (fd, engine) { +for_each_physical_engine (fd, engine) { execbuf.flags = engine->flags |I915_EXEC_FENCE_OUT ;gem_execbuf_wr (fd, &execbuf);
@@ -811,7 +811,7 @@ static void test_reset_stress(int fd, unsigned int flags)uint32_t ctx0 =context_create_safe (fd);const struct intel_execution_engine2 *engine; -__for_each_physical_engine (fd, engine) +for_each_physical_engine (fd, engine)reset_stress (fd, ctx0, engine, flags);gem_context_destroy (fd, ctx0);
tests/i915/gem_exec_async.c +2 -2
@@ -144,7 +144,7 @@ static void one(int fd, const struct intel_execution_engine2 *e) i =0 ; -__for_each_physical_engine (fd, engine) { +for_each_physical_engine (fd, engine) {if (engine->class == e->class && engine->instance == e->instance)continue ;
@@ -201,7 +201,7 @@ igt_mainigt_fork_hang_detector (fd); } -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) igt_subtest_f("concurrent-writes-%s", e->name) {igt_require (gem_class_can_store_dword (fd, e->class));one (fd, e);
tests/i915/gem_exec_await.c +1 -1
@@ -80,7 +80,7 @@ static void wide(int fd, int ring_size, int timeout, unsigned int flags)double time; nengine =0 ; -__for_each_physical_engine (fd, engine) +for_each_physical_engine (fd, engine) engines[nengine++] = engine->flags;igt_require (nengine);
tests/i915/gem_exec_basic.c +4 -4
@@ -118,7 +118,7 @@ static void all(int i915) {const struct intel_execution_engine2 *e; -__for_each_physical_engine (i915, e) +for_each_physical_engine (i915, e)noop (i915, e->flags); }
@@ -126,7 +126,7 @@ static void readonly_all(int i915) {const struct intel_execution_engine2 *e; -__for_each_physical_engine (i915, e) +for_each_physical_engine (i915, e)readonly (i915, e->flags); }
@@ -134,7 +134,7 @@ static void gtt_all(int i915) {const struct intel_execution_engine2 *e; -__for_each_physical_engine (i915, e) +for_each_physical_engine (i915, e)gtt (i915, e->flags); }
@@ -159,7 +159,7 @@ igt_main igt_subtest("gtt-all")gtt_all (fd); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("basic-%s", e->name)noop (fd, e->flags);igt_subtest_f ("readonly-%s" , e->name)
tests/i915/gem_exec_capture.c +1 -1
@@ -553,7 +553,7 @@ igt_mainigt_require (safer_strlen (igt_sysfs_get (dir,"error" )) >0 ); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("capture-%s", e->name) {igt_require (gem_class_can_store_dword (fd, e->class));capture (fd, dir, e);
tests/i915/gem_exec_create.c +1 -1
@@ -66,7 +66,7 @@ static void all(int fd, unsigned flags, int timeout, int ncpus)const struct intel_execution_engine2 *e; nengine =0 ; -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) engines[nengine++] = e->flags;igt_require (nengine);
tests/i915/gem_exec_fence.c +6 -6
@@ -265,7 +265,7 @@ static void test_fence_busy_all(int fd, unsigned flags) i++; all =-1 ; -__for_each_static_engine (e) { +for_each_static_engine (e) {int fence, new; execbuf.flags = e->flags |LOCAL_EXEC_FENCE_OUT ;
@@ -338,7 +338,7 @@ static void test_fence_await(int fd, const struct intel_execution_engine2 *e,igt_assert (spin->out_fence !=-1 ); i =0 ; -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {if (!gem_class_can_store_dword (fd, other->class))continue ;
@@ -501,7 +501,7 @@ static void test_parallel(int fd, const struct intel_execution_engine2 *e) obj[BATCH ].relocation_count =1 ;/* Queue all secondaries */ -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {if (other == e)continue ;
@@ -679,7 +679,7 @@ static void test_long_history(int fd, long ring_size, unsigned flags) limit = ring_size /3 ; nengine =0 ; -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) engines[nengine++] = e->flags;igt_require (nengine);
@@ -1194,7 +1194,7 @@ static void test_syncobj_wait(int fd)gem_write (fd, obj.handle,0 , &bbe,sizeof (bbe)); n =0 ; -__for_each_static_engine (e) { +for_each_static_engine (e) { obj.handle =gem_create (fd,4096 );gem_write (fd, obj.handle,0 , &bbe,sizeof (bbe));
@@ -1551,7 +1551,7 @@ igt_maintest_fence_busy_all (i915,WAIT |HANG ); } -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) { igt_subtest_group { igt_fixture {igt_require (gem_class_can_store_dword (i915, e->class));
tests/i915/gem_exec_flush.c +1 -1
@@ -600,7 +600,7 @@ igt_mainigt_fork_hang_detector (fd); } - __for_each_physical_engine(fd, e) igt_subtest_group { + for_each_physical_engine(fd, e) igt_subtest_group {unsigned timeout =5 +120 ; igt_fixture {
tests/i915/gem_exec_gttfill.c +2 -2
@@ -122,7 +122,7 @@ static void fillgtt(int fd, const struct intel_execution_engine2 *e, nengine =0 ;if (!e) { -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {if (!gem_class_can_store_dword (fd, other->class))continue ;
@@ -222,7 +222,7 @@ igt_main igt_subtest("basic")fillgtt (device,0 ,1 );/* just enough to run a single pass */ -__for_each_physical_engine (device, e) +for_each_physical_engine (device, e) igt_subtest_f("%s", e->name)fillgtt (device, e,20 );
tests/i915/gem_exec_latency.c +3 -3
@@ -311,7 +311,7 @@ static void latency_from_ring(int fd, reloc.presumed_offset = obj[1 ].offset; reloc.target_handle = flags &CORK ?1 :0 ; -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {igt_spin_t *spin =NULL ;IGT_CORK_HANDLE (c);
@@ -474,7 +474,7 @@ rthog_latency_on_ring(int fd, const struct intel_execution_engine2 *e, nengine =0 ;if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -684,7 +684,7 @@ igt_mainigt_fixture igt_require(intel_gen(intel_get_drm_devid(device)) >= 7); -__for_each_physical_engine (device, e) { +for_each_physical_engine (device, e) { igt_subtest_group { igt_subtest_f("%s-dispatch", e->name) latency_on_ring(device,
tests/i915/gem_exec_nop.c +8 -8
@@ -236,7 +236,7 @@ static void poll_sequential(int fd, const char *name, int timeout) flags |=I915_EXEC_SECURE ; nengine =0 ; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -443,7 +443,7 @@ static void parallel(int fd, uint32_t handle, int timeout) sum =0 ; nengine =0 ; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { engines[nengine] = e->flags; names[nengine] = e->name; nengine++;
@@ -505,7 +505,7 @@ static void series(int fd, uint32_t handle, int timeout)const char *name; nengine =0 ; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { time =nop_on_ring (fd, handle, e,1 , &count) / count;if (time > max) { name = e->name;
@@ -601,7 +601,7 @@ static void sequential(int fd, uint32_t handle, unsigned flags, int timeout) nengine =0 ; sum =0 ; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {unsigned long count; time =nop_on_ring (fd, handle, e,1 , &count) / count;
@@ -733,7 +733,7 @@ static void fence_signal(int fd, uint32_t handle, nengine =0 ;if (!e) { -__for_each_physical_engine (fd, other) +for_each_physical_engine (fd, other) engines[nengine++] = other->flags; }else {gem_require_ring (fd, e->flags);
@@ -881,7 +881,7 @@ igt_main igt_subtest("basic-sequential")sequential (device, handle,0 ,5 ); -__for_each_physical_engine (device, e) { +for_each_physical_engine (device, e) { igt_subtest_f("%s", e->name)single (device, handle, e, e->name);igt_subtest_f ("signal-%s" , e->name)
@@ -916,7 +916,7 @@ igt_mainigt_require (gem_scheduler_has_preemption (device)); } -__for_each_physical_engine (device, e) { +for_each_physical_engine (device, e) { igt_subtest_f("preempt-%s", e->name)preempt (device, handle, e, e->name); }
@@ -927,7 +927,7 @@ igt_mainigt_device_set_master (device); } -__for_each_physical_engine (device, e) { +for_each_physical_engine (device, e) {/* Requires master for STORE_DWORD on gen4/5 */ igt_subtest_f("poll-%s", e->name)poll_ring (device, e, e->name,20 );
tests/i915/gem_exec_parallel.c +2 -2
@@ -172,7 +172,7 @@ static void all(int fd, struct intel_execution_engine2 *engine, unsigned flags) nengine =0 ;if (!engine) {struct intel_execution_engine2 *e; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (gem_class_can_store_dword (fd, e->class)) engines[nengine++] = e->flags; }
@@ -254,7 +254,7 @@ igt_main/* NULL value means all engines */ all (fd,NULL , m->flags); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {for (const struct mode *m = modes; m->name; m++) igt_subtest_f("%s%s%s", e->name,
tests/i915/gem_exec_params.c +2 -2
@@ -211,7 +211,7 @@ igt_main } igt_subtest("control") { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { execbuf.flags = e->flags;gem_execbuf (fd, &execbuf); }
@@ -362,7 +362,7 @@ igt_main igt_subtest("rs-invalid") {bool has_rs =has_resource_streamer (fd); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {int expect = -EINVAL ;if (has_rs && (e->class ==I915_ENGINE_CLASS_RENDER )) expect =0 ;
tests/i915/gem_exec_reloc.c +2 -2
@@ -258,7 +258,7 @@ static void active(int fd, const struct intel_execution_engine2 *e) nengine =0 ;if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (gem_class_can_store_dword (fd, e->class)) engines[nengine++] = e->flags; }
@@ -753,7 +753,7 @@ igt_main igt_subtest("active")active (fd,NULL ); -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) igt_subtest_f("active-%s", e->name)active (fd, e);
tests/i915/gem_exec_reuse.c +1 -1
@@ -118,7 +118,7 @@ igt_main no.obj =malloc (sizeof (struct drm_i915_gem_exec_object2 ) * (no.max_age +1 )); nengine =0 ; -__for_each_static_engine (e) +for_each_static_engine (e)if (e) engines[nengine++] = e->flags;igt_require (nengine);
tests/i915/gem_exec_schedule.c +20 -20
@@ -225,7 +225,7 @@ static void independent(int fd, const struct intel_execution_engine2 *e) plug =igt_cork_plug (&cork, fd);/* Check that we can submit to engine while all others are blocked */ -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {if (other == e)continue ;
@@ -285,7 +285,7 @@ static void smoketest(int fd, const struct intel_execution_engine2 *e, nengine =0 ;if (!e) { -__for_each_physical_engine (fd, other) +for_each_physical_engine (fd, other) engines[nengine++] = other; }else { engines[nengine++] = e;
@@ -369,7 +369,7 @@ static void semaphore_userlock(int i915)*/ scratch =gem_create (i915,4096 ); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {if (!spin) { spin = igt_spin_new(i915, .dependency = scratch,
@@ -394,7 +394,7 @@ static void semaphore_userlock(int i915)* taking precedence. */ scratch =gem_context_create (i915); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {struct drm_i915_gem_execbuffer2 execbuf = { .buffers_ptr =to_user_pointer (&obj), .buffer_count =1 ,
@@ -433,7 +433,7 @@ static void semaphore_codependency(int i915)*/ i =0 ; -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {uint32_t ctx;if (e->class ==I915_ENGINE_CLASS_RENDER )
@@ -451,7 +451,7 @@ static void semaphore_codependency(int i915) .flags = IGT_SPIN_POLL_RUN);igt_spin_busywait_until_started (task[i].xcs); -__for_each_physical_engine (i915, other) { +for_each_physical_engine (i915, other) {if (other->class ==I915_ENGINE_CLASS_RENDER )/* Common rcs tasks will be queued in FIFO */ task[i].rcs =
@@ -508,7 +508,7 @@ static void semaphore_resolve(int i915) semaphore =gem_create (i915,4096 ); sema =gem_mmap__wc (i915, semaphore,0 ,4096 ,PROT_WRITE ); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {struct drm_i915_gem_exec_object2 obj[3 ];struct drm_i915_gem_execbuffer2 eb;uint32_t handle, cancel;
@@ -623,8 +623,8 @@ static void semaphore_noskip(int i915) ctx =gem_context_create (i915); -__for_each_physical_engine (i915, engine) { -__for_each_physical_engine (i915, other) { +for_each_physical_engine (i915, engine) { +for_each_physical_engine (i915, other) {struct drm_i915_gem_exec_object2 obj[3 ];struct drm_i915_gem_execbuffer2 eb;uint32_t handle, *cs, *map;
@@ -846,7 +846,7 @@ static igt_spin_t *__noise(int fd, uint32_t ctx, int prio, igt_spin_t *spin)gem_context_set_priority (fd, ctx, prio); -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {if (spin ==NULL ) { spin = __igt_spin_new(fd, .ctx = ctx,
@@ -883,7 +883,7 @@ static void __preempt_other(int fd, n++;if (flags &CHAIN ) { -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {store_dword (fd, ctx[LO ], other, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER );
@@ -933,7 +933,7 @@ static void preempt_other(int fd, const struct intel_execution_engine2 *e, ctx[HI ] =gem_context_create (fd);gem_context_set_priority (fd, ctx[HI ],MAX_PRIO ); -__for_each_physical_engine (fd, primary) { +for_each_physical_engine (fd, primary) {igt_debug ("Primary engine: %s\n" , e->name);__preempt_other (fd, ctx, e, primary, flags); }
@@ -990,7 +990,7 @@ static void __preempt_queue(int fd, n++;if (flags &CHAIN ) { -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {store_dword (fd, ctx[LO ], other, result, (n +1 )*sizeof (uint32_t ), n +1 ,0 ,I915_GEM_DOMAIN_RENDER );
@@ -1034,7 +1034,7 @@ static void preempt_queue(int fd, const struct intel_execution_engine2 *e, {const struct intel_execution_engine2 *other; -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {for (unsigned depth =0 ; depth <=MAX_ELSP_QLEN ; depth++)__preempt_queue (fd, e, other, depth, flags); }
@@ -1063,7 +1063,7 @@ static void preempt_self(int fd, const struct intel_execution_engine2 *e) n =0 ;gem_context_set_priority (fd, ctx[HI ],MIN_PRIO ); -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) { spin[n] = __igt_spin_new(fd, .ctx = ctx[NOISE], .engine = other->flags);
@@ -1589,7 +1589,7 @@ static void measure_semaphore_power(int i915)igt_require (gpu_power_open (&power) ==0 ); -__for_each_physical_engine (i915, signaler) { +for_each_physical_engine (i915, signaler) {struct gpu_power_sample s_spin[2 ];struct gpu_power_sample s_sema[2 ];double baseline, total;
@@ -1607,7 +1607,7 @@ static void measure_semaphore_power(int i915)gpu_power_read (&power, &s_spin[1 ]);/* Add a waiter to each engine */ -__for_each_physical_engine (i915, engine) { +for_each_physical_engine (i915, engine) {igt_spin_t *sema;if (engine == signaler)
@@ -1660,7 +1660,7 @@ igt_main } igt_subtest_group { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("fifo-%s", e->name) {igt_require (gem_class_can_store_dword (fd, e->class));fifo (fd, e);
@@ -1691,7 +1691,7 @@ igt_main igt_subtest("smoketest-all")smoketest (fd,NULL ,30 ); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_group { igt_fixture {igt_require (gem_class_can_store_dword (fd, e->class));
@@ -1783,7 +1783,7 @@ igt_mainigt_require (gem_has_execlists (fd)); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_group { igt_fixture {igt_require (gem_scheduler_has_preemption (fd));
tests/i915/gem_exec_store.c +2 -2
@@ -220,7 +220,7 @@ static void store_all(int fd) nengine =0 ;intel_detect_and_clear_missed_interrupts (fd); -__for_each_physical_engine (fd, engine) { +for_each_physical_engine (fd, engine) {if (!gem_class_can_store_dword (fd, engine->class))continue ;
@@ -329,7 +329,7 @@ igt_mainigt_fork_hang_detector (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("basic-%s", e->name)store_dword (fd, e);
tests/i915/gem_exec_suspend.c +3 -3
@@ -67,7 +67,7 @@ static void test_all(int fd, unsigned flags) {const struct intel_execution_engine2 *engine; -__for_each_physical_engine (fd, engine) +for_each_physical_engine (fd, engine)if (gem_class_can_store_dword (fd, engine->class))run_test (fd, engine, flags & ~0xff ); }
@@ -108,7 +108,7 @@ static void run_test(int fd, const struct intel_execution_engine2 *e,* GPU is then unlikely to be active!) */ if (has_semaphores (fd)) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (gem_class_can_store_dword (fd, e->class)) engines[nengine++] = e->flags; }
@@ -266,7 +266,7 @@ igt_main igt_subtest("basic-S4")run_test (fd,NULL ,HIBERNATE ); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {for (m = modes; m->suffix; m++) { igt_subtest_f("%s-uncached%s", e->name, m->suffix)run_test (fd, e, m->mode |UNCACHED );
tests/i915/gem_exec_whisper.c +2 -2
@@ -204,7 +204,7 @@ static void whisper(int fd, const struct intel_execution_engine2 *engine, nengine =0 ;if (!engine) { -__for_each_physical_engine (fd, engine) { +for_each_physical_engine (fd, engine) {if (gem_class_can_store_dword (fd, engine->class)) engines[nengine++] = engine->flags; }
@@ -572,7 +572,7 @@ igt_mainwhisper (fd,NULL , m->flags |ALL ); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {for (const struct mode *m = modes; m->name; m++) {if (m->flags &CHAIN )continue ;
tests/i915/gem_mocs_settings.c +1 -1
@@ -516,7 +516,7 @@ igt_mainigt_require (get_mocs_settings (fd, &table, false)); } -__for_each_static_engine (e) { +for_each_static_engine (e) {for (unsigned mode =NONE ; mode <MAX_MOCS_TEST_MODES ; mode++) { igt_subtest_group {igt_hang_t hang = {};
tests/i915/gem_reset_stats.c +3 -3
@@ -520,14 +520,14 @@ static void noop_on_each_ring(int fd, const bool reverse) eb.buffer_count =1 ;if (reverse) { -__for_each_static_engine (e) +for_each_static_engine (e) ;while (--e >= intel_execution_engines2) { eb.flags = e->flags;__gem_execbuf (fd, &eb); } }else { -__for_each_static_engine (e) { +for_each_static_engine (e) { eb.flags = e->flags;__gem_execbuf (fd, &eb); }
@@ -803,7 +803,7 @@ igt_main igt_subtest_f("params-ctx")RUN_TEST (test_params_ctx ()); -__for_each_static_engine (e) { +for_each_static_engine (e) { igt_subtest_f("reset-stats-%s", e->name)RUN_TEST (test_rs (e,4 ,1 ,0 ));
tests/i915/gem_ring_sync_loop.c +1 -1
@@ -48,7 +48,7 @@ sync_loop(int fd)int i; nengine =0 ; -__for_each_physical_engine (fd, engine) +for_each_physical_engine (fd, engine) engines[nengine++] = engine->flags;igt_require (nengine);
tests/i915/gem_ringfill.c +1 -1
@@ -280,7 +280,7 @@ igt_mainfor (m = modes; m->suffix; m++) {const struct intel_execution_engine2 *e; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("%s%s", e->name, m->suffix) {igt_skip_on (m->flags &NEWFD && master);if (m->flags & (HANG |SUSPEND |HIBERNATE ))
tests/i915/gem_shrink.c +2 -2
@@ -423,7 +423,7 @@ igt_main CHECK_SWAP | CHECK_RAM); nengine =0 ; -__for_each_static_engine (engine) +for_each_static_engine (engine) engines[nengine++] = engine->flags;igt_require (nengine);
@@ -431,7 +431,7 @@ igt_main } igt_subtest("reclaim") { -__for_each_static_engine (engine) { +for_each_static_engine (engine) {if (engine->class ==I915_ENGINE_CLASS_RENDER )break ; }
tests/i915/gem_spin_batch.c +3 -3
@@ -86,7 +86,7 @@ static void spin_resubmit(int fd, const struct intel_execution_engine2 *e, };if (flags &RESUBMIT_ALL_ENGINES ) { -__for_each_physical_engine (fd, other) { +for_each_physical_engine (fd, other) {if (other == e)continue ;
@@ -119,7 +119,7 @@ static void spin_on_all_engines(int fd, unsigned int timeout_sec) {const struct intel_execution_engine2 *engine; -__for_each_physical_engine (fd, engine) { +for_each_physical_engine (fd, engine) { igt_fork(child, 1) {igt_install_exit_handler (spin_exit_handler);spin (fd, engine, timeout_sec);
@@ -142,7 +142,7 @@ igt_mainigt_fork_hang_detector (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("basic-%s", e->name)spin (fd, e,3 );
tests/i915/gem_storedw_loop.c +1 -1
@@ -182,7 +182,7 @@ igt_mainigt_require (gem_uses_ppgtt (fd)); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("store-%s", e->name) {check_test_requirements (fd, e);store_test (fd, e);
tests/i915/gem_sync.c +12 -12
@@ -87,7 +87,7 @@ sync_ring(int fd, const struct intel_execution_engine2 *e, int num_children,int num_engines =0 ;if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { names[num_engines] = e->name; engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))
@@ -187,7 +187,7 @@ wakeup_ring(int fd, const struct intel_execution_engine2 *e, int timeout,int num_engines =0 ;if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -299,7 +299,7 @@ static void active_ring(int fd, const struct intel_execution_engine2 *e, int timint num_engines =0 ;if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -367,7 +367,7 @@ active_wakeup_ring(int fd, const struct intel_execution_engine2 *e, int timeout,int num_engines =0 ;if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -502,7 +502,7 @@ store_ring(int fd, const struct intel_execution_engine2 *e, int num_children, inint num_engines =0 ;if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -619,7 +619,7 @@ switch_ring(int fd, const struct intel_execution_engine2 *e, int num_children, igem_require_contexts (fd);if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -946,7 +946,7 @@ store_many(int fd, const struct intel_execution_engine2 *e, int timeout)intel_detect_and_clear_missed_interrupts (fd);if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -979,7 +979,7 @@ sync_all(int fd, int num_children, int timeout)const struct intel_execution_engine2 *engine;int num_engines =0 ; -__for_each_physical_engine (fd, engine) { +for_each_physical_engine (fd, engine) { engines[num_engines++] = engine->flags;if (num_engines ==ARRAY_SIZE (engines))break ;
@@ -1032,7 +1032,7 @@ store_all(int fd, int num_children, int timeout)int num_engines =0 ;const struct intel_execution_engine2 *e; -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {if (!gem_class_can_store_dword (fd, e->class))continue ;
@@ -1141,7 +1141,7 @@ preempt(int fd, const struct intel_execution_engine2 *e, int num_children, int tuint32_t ctx[2 ];if (!e) { -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { names[num_engines] = e->name; engines[num_engines++] = e->flags;if (num_engines ==ARRAY_SIZE (engines))
@@ -1227,7 +1227,7 @@ igt_mainigt_fork_hang_detector (fd); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("%s", e->name)sync_ring (fd, e,1 ,150 );igt_subtest_f ("idle-%s" , e->name)
@@ -1301,7 +1301,7 @@ igt_main igt_subtest("preempt-all")preempt (fd,NULL ,1 ,20 ); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("preempt-%s", e->name)preempt (fd, e, ncpus,150 ); }
tests/i915/gem_userptr_blits.c +1 -1
@@ -1147,7 +1147,7 @@ static void test_readonly(int i915)gem_userptr (i915, space, total, true, userptr_flags, &rhandle); -__for_each_physical_engine (i915, engine) { +for_each_physical_engine (i915, engine) {char *ref, *result;/* First tweak the backing store through the write */
tests/i915/gem_wait.c +2 -2
@@ -177,7 +177,7 @@ igt_mainbasic (fd,ALL_ENGINES ,WRITE ); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_group { igt_subtest_f("busy-%s", e->name) {gem_quiescent_gpu (fd);
@@ -234,7 +234,7 @@ igt_mainbasic (fd,ALL_ENGINES ,WRITE |HANG ); } -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) { igt_subtest_f("hang-busy-%s", e->name) {gem_quiescent_gpu (fd);basic (fd, e->flags,HANG |BUSY );
tests/i915/i915_hangman.c +1 -1
@@ -276,7 +276,7 @@ igt_main igt_subtest("error-state-basic")test_error_state_basic (); -__for_each_physical_engine (device, e) +for_each_physical_engine (device, e) igt_subtest_f("error-state-capture-%s", e->name)test_error_state_capture (e, e->name);
tests/i915/i915_module_load.c +2 -2
@@ -152,7 +152,7 @@ static void store_all(int fd) nengine =0 ;intel_detect_and_clear_missed_interrupts (fd); -__for_each_physical_engine (fd, engine) { +for_each_physical_engine (fd, engine) {if (!gem_class_can_store_dword (fd, engine->class))continue ;
@@ -287,7 +287,7 @@ gem_exec_store(void) fd =__drm_open_driver (DRIVER_INTEL );igt_fork_hang_detector (fd); -__for_each_static_engine (e) { +for_each_static_engine (e) {if (gem_class_can_store_dword (fd, e->class))store_dword (fd, e); }
tests/perf_pmu.c +6 -6
@@ -445,7 +445,7 @@ busy_check_all(int gem_fd, const struct intel_execution_engine2 *e, i =0 ; fd[0 ] =-1 ; -__for_each_physical_engine (gem_fd, e_) { +for_each_physical_engine (gem_fd, e_) {if (e->class == e_->class && e->instance == e_->instance) busy_idx = i;
@@ -508,7 +508,7 @@ most_busy_check_all(int gem_fd, const struct intel_execution_engine2 *e,unsigned int idle_idx, i; i =0 ; -__for_each_physical_engine (gem_fd, e_) { +for_each_physical_engine (gem_fd, e_) {if (e->class == e_->class && e->instance == e_->instance) idle_idx = i;else if (spin)
@@ -565,7 +565,7 @@ all_busy_check_all(int gem_fd, const unsigned int num_engines,unsigned int i; i =0 ; -__for_each_physical_engine (gem_fd, e) { +for_each_physical_engine (gem_fd, e) {if (spin)__submit_spin (gem_fd, spin, e,64 );else
@@ -1686,7 +1686,7 @@ igt_mainigt_require_gem (fd);igt_require (i915_type_id () >0 ); -__for_each_physical_engine (fd, e) +for_each_physical_engine (fd, e) num_engines++; }
@@ -1696,7 +1696,7 @@ igt_main igt_subtest("invalid-init")invalid_init (); -__for_each_physical_engine (fd, e) { +for_each_physical_engine (fd, e) {const unsigned int pct[] = {2 ,50 ,98 };/**
@@ -1900,7 +1900,7 @@ igt_maingem_quiescent_gpu (fd); } -__for_each_physical_engine (render_fd, e) { +for_each_physical_engine (render_fd, e) { igt_subtest_group { igt_fixture { gem_context_has_engine(render_fd,
tests/prime_busy.c +1 -1
@@ -241,7 +241,7 @@ igt_mainigt_fixture gem_require_mmap_wc (fd ); -__for_each_static_engine (e) { +for_each_static_engine (e) {for (const struct mode *m = modes; m->name; m++)test_engine_mode (fd, e, m->name, m->flags); }
tests/prime_vgem.c +2 -2
@@ -844,7 +844,7 @@ igt_main igt_subtest("coherency-gtt")test_gtt_interleaved (vgem, i915); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {igt_subtest_f ("%ssync-%s" , e->class ==I915_ENGINE_CLASS_RENDER ? "basic-" : "", e->name) {
@@ -884,7 +884,7 @@ igt_main igt_subtest("basic-fence-mmap")test_fence_mmap (i915, vgem); -__for_each_physical_engine (i915, e) { +for_each_physical_engine (i915, e) {igt_subtest_f ("%sfence-wait-%s" , e->class ==I915_ENGINE_CLASS_RENDER ? "basic-" : "", e->name) {
Signature · 6 lines
2.19.1 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✓ Fi.CI.BAT: success for Upgrading the Engine discovery
In reply to Ramalingam C== Series Details ==
== Series Details ==
Series: Upgrading the Engine discovery
URL : https://patchwork.freedesktop.org/series/62512/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_6322 -> IGTPW_3186
====================================================
Summary
-------
**SUCCESS**
No regressions found.
External URL: https://patchwork.freedesktop.org/api/1.0/series/62512/revisions/1/mbox/
Known issues
------------
Here are the changes found in IGTPW_3186 that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@gem_flink_basic@double-flink:
- fi-icl-u3: [PASS][1] -> [DMESG-WARN][2] ([fdo#107724])
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/fi-icl-u3/igt@[email protected]
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/fi-icl-u3/igt@[email protected]
* igt@i915_selftest@live_evict:
- fi-icl-dsi: [PASS][3] -> [INCOMPLETE][4] ([fdo#107713])
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/fi-icl-dsi/igt@i915_selftest@live_evict.html
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/fi-icl-dsi/igt@i915_selftest@live_evict.html
#### Possible fixes ####
* igt@kms_frontbuffer_tracking@basic:
- fi-icl-u2: [FAIL][5] ([fdo#103167]) -> [PASS][6]
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/fi-icl-u2/igt@[email protected]
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/fi-icl-u2/igt@[email protected]
[fdo#103167]: https://bugs.freedesktop.org/show_bug.cgi?id=103167
[fdo#107713]: https://bugs.freedesktop.org/show_bug.cgi?id=107713
[fdo#107724]: https://bugs.freedesktop.org/show_bug.cgi?id=107724
Participating hosts (53 -> 44)
------------------------------
Missing (9): fi-kbl-soraka fi-ilk-m540 fi-hsw-4200u fi-skl-guc fi-byt-squawks fi-bsw-cyan fi-icl-y fi-byt-clapper fi-bdw-samus
Build changes
-------------
* IGT: IGT_5064 -> IGTPW_3186
CI_DRM_6322: 07ceacd2be6fc33beff4c551d57c573ce5741d3d @ git://anongit.freedesktop.org/gfx-ci/linux
IGTPW_3186: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/
IGT_5064: 22850c1906550fb97b405c019275dcfb34be8cf7 @ git://anongit.freedesktop.org/xorg/app/intel-gpu-tools
== Testlist changes ==
+++ 1437 lines
--- 1619 lines
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✗ Fi.CI.IGT: failure for Upgrading the Engine discovery
In reply to Ramalingam C== Series Details ==
== Series Details ==
Series: Upgrading the Engine discovery
URL : https://patchwork.freedesktop.org/series/62512/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_6322_full -> IGTPW_3186_full
====================================================
Summary
-------
**FAILURE**
Serious unknown changes coming with IGTPW_3186_full absolutely need to be verified manually.
If you think the reported changes have nothing to do with the changes introduced in IGTPW_3186_full, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI.
External URL: https://patchwork.freedesktop.org/api/1.0/series/62512/revisions/1/mbox/
Possible new issues
-------------------
Here are the unknown changes that may have been introduced in IGTPW_3186_full:
### IGT changes ###
#### Possible regressions ####
* igt@gem_busy@close-race:
- shard-glk: [PASS][1] -> [FAIL][2] +4 similar issues
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-glk3/igt@[email protected]
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-glk6/igt@[email protected]
* {igt@gem_ctx_shared@exec-single-timeline-bcs0} (NEW):
- shard-iclb: NOTRUN -> [FAIL][3] +2 similar issues
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-iclb2/igt@[email protected]
* {igt@gem_ctx_shared@exec-single-timeline-vcs0} (NEW):
- shard-glk: NOTRUN -> [FAIL][4] +2 similar issues
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-glk7/igt@[email protected]
* {igt@gem_ctx_shared@exec-single-timeline-vcs1} (NEW):
- shard-kbl: NOTRUN -> [FAIL][5] +2 similar issues
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-kbl7/igt@[email protected]
* {igt@gem_ctx_shared@q-independent-vcs0} (NEW):
- shard-iclb: NOTRUN -> [SKIP][6] +8 similar issues
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-iclb1/igt@[email protected]
* igt@gem_eio@in-flight-internal-10ms:
- shard-hsw: [PASS][7] -> [CRASH][8] +4 similar issues
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-hsw1/igt@[email protected]
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-hsw1/igt@[email protected]
- shard-apl: NOTRUN -> [CRASH][9] +2 similar issues
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-apl7/igt@[email protected]
* igt@gem_eio@in-flight-internal-1us:
- shard-iclb: NOTRUN -> [CRASH][10]
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-iclb1/igt@[email protected]
* igt@gem_eio@in-flight-internal-immediate:
- shard-glk: [PASS][11] -> [CRASH][12] +4 similar issues
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-glk4/igt@[email protected]
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-glk1/igt@[email protected]
- shard-iclb: [PASS][13] -> [CRASH][14] +4 similar issues
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-iclb7/igt@[email protected]
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-iclb6/igt@[email protected]
* igt@gem_eio@in-flight-suspend:
- shard-kbl: [PASS][15] -> [CRASH][16] +5 similar issues
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-kbl7/igt@[email protected]
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-kbl1/igt@[email protected]
- shard-apl: [PASS][17] -> [CRASH][18] +2 similar issues
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-apl1/igt@[email protected]
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-apl1/igt@[email protected]
- shard-snb: [PASS][19] -> [CRASH][20] +4 similar issues
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-snb4/igt@[email protected]
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-snb4/igt@[email protected]
* igt@gem_exec_reuse@baggage:
- shard-snb: [PASS][21] -> [FAIL][22] +4 similar issues
[21]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-snb6/igt@[email protected]
[22]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-snb6/igt@[email protected]
* igt@gem_exec_reuse@contexts:
- shard-hsw: [PASS][23] -> [FAIL][24] +4 similar issues
[23]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-hsw8/igt@[email protected]
[24]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-hsw8/igt@[email protected]
- shard-kbl: [PASS][25] -> [FAIL][26] +3 similar issues
[25]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-kbl6/igt@[email protected]
[26]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-kbl3/igt@[email protected]
- shard-apl: [PASS][27] -> [FAIL][28] +4 similar issues
[27]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-apl5/igt@[email protected]
[28]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-apl8/igt@[email protected]
* igt@gem_exec_reuse@single:
- shard-iclb: [PASS][29] -> [FAIL][30] +3 similar issues
[29]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-iclb1/igt@[email protected]
[30]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-iclb6/igt@[email protected]
* {igt@gem_exec_schedule@independent-bcs0} (NEW):
- shard-snb: NOTRUN -> [FAIL][31] +1 similar issue
[31]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-snb4/igt@[email protected]
* {igt@gem_exec_schedule@independent-rcs0} (NEW):
- shard-apl: NOTRUN -> [FAIL][32] +6 similar issues
[32]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-apl2/igt@[email protected]
* {igt@gem_exec_schedule@independent-vcs0} (NEW):
- shard-hsw: NOTRUN -> [FAIL][33] +3 similar issues
[33]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-hsw1/igt@[email protected]
* igt@gem_exec_schedule@semaphore-noskip:
- shard-apl: [PASS][34] -> [DMESG-FAIL][35]
[34]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-apl7/igt@[email protected]
[35]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-apl5/igt@[email protected]
#### Warnings ####
* igt@gem_eio@in-flight-suspend:
- shard-glk: [FAIL][36] ([fdo#110667]) -> [CRASH][37]
[36]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_6322/shard-glk6/igt@[email protected]
[37]: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/shard-glk4/igt@[email protected]
New tests
---------
New tests have been introduced between CI_DRM_6322_full and IGTPW_3186_full:
### New IGT tests (353) ###
* igt@gem_bad_reloc@negative-reloc-bcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_bad_reloc@negative-reloc-lut-bcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_bad_reloc@negative-reloc-lut-rcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_bad_reloc@negative-reloc-lut-vcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_bad_reloc@negative-reloc-lut-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.0] s
* igt@gem_bad_reloc@negative-reloc-lut-vcs2:
- Statuses : 1 pass(s)
- Exec time: [0.00] s
* igt@gem_bad_reloc@negative-reloc-lut-vecs0:
- Statuses : 5 pass(s)
- Exec time: [0.0, 0.00] s
* igt@gem_bad_reloc@negative-reloc-rcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_bad_reloc@negative-reloc-vcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_bad_reloc@negative-reloc-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.00] s
* igt@gem_bad_reloc@negative-reloc-vcs2:
- Statuses : 1 pass(s)
- Exec time: [0.0] s
* igt@gem_bad_reloc@negative-reloc-vecs0:
- Statuses : 5 pass(s)
- Exec time: [0.0, 0.00] s
* igt@gem_ctx_shared@exec-shared-gtt-bcs0:
- Statuses : 6 pass(s)
- Exec time: [0.00] s
* igt@gem_ctx_shared@exec-shared-gtt-rcs0:
- Statuses : 6 pass(s)
- Exec time: [0.00] s
* igt@gem_ctx_shared@exec-shared-gtt-vcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_ctx_shared@exec-shared-gtt-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.00] s
* igt@gem_ctx_shared@exec-shared-gtt-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_ctx_shared@exec-shared-gtt-vecs0:
- Statuses : 5 pass(s)
- Exec time: [0.0, 0.00] s
* igt@gem_ctx_shared@exec-single-timeline-bcs0:
- Statuses : 4 fail(s) 2 skip(s)
- Exec time: [0.0, 1.06] s
* igt@gem_ctx_shared@exec-single-timeline-rcs0:
- Statuses : 6 skip(s)
- Exec time: [0.0] s
* igt@gem_ctx_shared@exec-single-timeline-vcs0:
- Statuses : 3 fail(s) 3 skip(s)
- Exec time: [0.0, 1.48] s
* igt@gem_ctx_shared@exec-single-timeline-vcs1:
- Statuses : 1 fail(s)
- Exec time: [1.06] s
* igt@gem_ctx_shared@exec-single-timeline-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_ctx_shared@exec-single-timeline-vecs0:
- Statuses : 4 fail(s) 1 skip(s)
- Exec time: [0.0, 1.07] s
* igt@gem_ctx_shared@q-in-order-bcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_ctx_shared@q-in-order-rcs0:
- Statuses : 3 pass(s) 2 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_ctx_shared@q-in-order-vcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.02] s
* igt@gem_ctx_shared@q-in-order-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_ctx_shared@q-in-order-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_ctx_shared@q-in-order-vecs0:
- Statuses : 4 pass(s) 1 skip(s)
- Exec time: [0.0, 0.01] s
* igt@gem_ctx_shared@q-independent-bcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.02] s
* igt@gem_ctx_shared@q-independent-rcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.02] s
* igt@gem_ctx_shared@q-independent-vcs0:
- Statuses : 6 skip(s)
- Exec time: [0.0] s
* igt@gem_ctx_shared@q-independent-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_ctx_shared@q-independent-vcs2:
- Statuses : 1 pass(s)
- Exec time: [0.00] s
* igt@gem_ctx_shared@q-independent-vecs0:
- Statuses : 4 pass(s)
- Exec time: [0.00, 0.02] s
* igt@gem_ctx_shared@q-out-order-bcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_ctx_shared@q-out-order-rcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_ctx_shared@q-out-order-vcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_ctx_shared@q-out-order-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_ctx_shared@q-out-order-vcs2:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_ctx_shared@q-out-order-vecs0:
- Statuses : 4 pass(s) 1 skip(s)
- Exec time: [0.0, 0.01] s
* igt@gem_ctx_shared@q-promotion-bcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.02] s
* igt@gem_ctx_shared@q-promotion-rcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.04] s
* igt@gem_ctx_shared@q-promotion-vcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.02] s
* igt@gem_ctx_shared@q-promotion-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_ctx_shared@q-promotion-vcs2:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_ctx_shared@q-promotion-vecs0:
- Statuses : 4 pass(s) 1 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_ctx_shared@q-smoketest-bcs0:
- Statuses : 3 pass(s) 2 skip(s)
- Exec time: [0.0, 5.42] s
* igt@gem_ctx_shared@q-smoketest-rcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 5.42] s
* igt@gem_ctx_shared@q-smoketest-vcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 5.42] s
* igt@gem_ctx_shared@q-smoketest-vcs1:
- Statuses : 1 pass(s)
- Exec time: [5.39] s
* igt@gem_ctx_shared@q-smoketest-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_ctx_shared@q-smoketest-vecs0:
- Statuses : 4 pass(s) 1 skip(s)
- Exec time: [0.0, 5.42] s
* igt@gem_ctx_switch@basic-rcs0:
- Statuses : 6 pass(s)
- Exec time: [5.01, 5.09] s
* igt@gem_ctx_switch@basic-rcs0-heavy:
- Statuses : 6 pass(s)
- Exec time: [5.26, 14.07] s
* igt@gem_ctx_switch@basic-rcs0-heavy-interruptible:
- Statuses : 6 pass(s)
- Exec time: [5.43, 17.42] s
* igt@gem_ctx_switch@basic-rcs0-heavy-queue:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 14.14] s
* igt@gem_ctx_switch@basic-rcs0-heavy-queue-interruptible:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 17.09] s
* igt@gem_ctx_switch@basic-rcs0-interruptible:
- Statuses : 6 pass(s)
- Exec time: [5.01, 5.06] s
* igt@gem_ctx_switch@basic-rcs0-queue:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 5.09] s
* igt@gem_ctx_switch@basic-rcs0-queue-interruptible:
- Statuses : 4 pass(s) 1 skip(s)
- Exec time: [0.0, 5.07] s
* igt@gem_exec_async@concurrent-writes-bcs0:
- Statuses : 6 pass(s)
- Exec time: [0.00] s
* igt@gem_exec_async@concurrent-writes-rcs0:
- Statuses : 6 pass(s)
- Exec time: [0.00] s
* igt@gem_exec_async@concurrent-writes-vcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.00] s
* igt@gem_exec_async@concurrent-writes-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.00] s
* igt@gem_exec_async@concurrent-writes-vcs2:
- Statuses : 1 pass(s)
- Exec time: [0.00] s
* igt@gem_exec_async@concurrent-writes-vecs0:
- Statuses : 5 pass(s)
- Exec time: [0.00, 0.01] s
* igt@gem_exec_capture@capture-bcs0:
- Statuses : 6 pass(s)
- Exec time: [0.00, 0.02] s
* igt@gem_exec_capture@capture-rcs0:
- Statuses : 6 pass(s)
- Exec time: [0.00, 0.01] s
* igt@gem_exec_capture@capture-vcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.01] s
* igt@gem_exec_capture@capture-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_exec_capture@capture-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_exec_capture@capture-vecs0:
- Statuses : 5 pass(s)
- Exec time: [0.00, 0.01] s
* igt@gem_exec_fence@basic-await-rcs0:
- Statuses : 6 pass(s)
- Exec time: [1.01, 1.04] s
* igt@gem_exec_fence@basic-busy-rcs0:
- Statuses : 6 pass(s)
- Exec time: [0.01, 0.03] s
* igt@gem_exec_fence@basic-wait-rcs0:
- Statuses : 5 pass(s)
- Exec time: [0.01, 0.03] s
* igt@gem_exec_flush@basic-batch-kernel-rcs0-cmd:
- Statuses : 1 pass(s) 5 skip(s)
- Exec time: [0.0, 134.38] s
* igt@gem_exec_flush@basic-batch-kernel-rcs0-uc:
- Statuses : 6 pass(s)
- Exec time: [134.33, 134.64] s
* igt@gem_exec_flush@basic-batch-kernel-rcs0-wb:
- Statuses : 6 pass(s)
- Exec time: [134.34, 134.78] s
* igt@gem_exec_flush@basic-uc-pro-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.24, 134.27] s
* igt@gem_exec_flush@basic-uc-prw-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.24, 134.32] s
* igt@gem_exec_flush@basic-uc-ro-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.23, 134.31] s
* igt@gem_exec_flush@basic-uc-rw-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.23, 134.25] s
* igt@gem_exec_flush@basic-uc-set-rcs0:
- Statuses : 5 pass(s)
- Exec time: [134.23, 134.27] s
* igt@gem_exec_flush@basic-wb-pro-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.23, 134.26] s
* igt@gem_exec_flush@basic-wb-prw-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.24, 134.27] s
* igt@gem_exec_flush@basic-wb-ro-before-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.23, 134.26] s
* igt@gem_exec_flush@basic-wb-ro-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.24, 134.27] s
* igt@gem_exec_flush@basic-wb-rw-before-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.24, 134.31] s
* igt@gem_exec_flush@basic-wb-rw-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.23, 134.25] s
* igt@gem_exec_flush@basic-wb-set-rcs0:
- Statuses : 6 pass(s)
- Exec time: [134.24, 134.26] s
* igt@gem_exec_schedule@deep-bcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 5.99] s
* igt@gem_exec_schedule@deep-rcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 5.99] s
* igt@gem_exec_schedule@deep-vcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 6.42] s
* igt@gem_exec_schedule@deep-vcs1:
- Statuses : 1 pass(s)
- Exec time: [4.91] s
* igt@gem_exec_schedule@deep-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_exec_schedule@deep-vecs0:
- Statuses : 4 pass(s) 1 skip(s)
- Exec time: [0.0, 5.87] s
* igt@gem_exec_schedule@fifo-bcs0:
- Statuses : 6 pass(s)
- Exec time: [0.00, 0.03] s
* igt@gem_exec_schedule@fifo-rcs0:
- Statuses : 6 pass(s)
- Exec time: [0.01, 0.03] s
* igt@gem_exec_schedule@fifo-vcs0:
- Statuses : 5 pass(s) 1 skip(s)
- Exec time: [0.0, 0.16] s
* igt@gem_exec_schedule@fifo-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.02] s
* igt@gem_exec_schedule@fifo-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_exec_schedule@fifo-vecs0:
- Statuses : 5 pass(s)
- Exec time: [0.00, 0.01] s
* igt@gem_exec_schedule@in-order-bcs0:
- Statuses : 3 pass(s) 2 skip(s)
- Exec time: [0.0, 0.06] s
* igt@gem_exec_schedule@in-order-rcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_exec_schedule@in-order-vcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.05] s
* igt@gem_exec_schedule@in-order-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.01] s
* igt@gem_exec_schedule@in-order-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_exec_schedule@in-order-vecs0:
- Statuses : 4 pass(s) 1 skip(s)
- Exec time: [0.0, 0.02] s
* igt@gem_exec_schedule@independent-bcs0:
- Statuses : 6 fail(s)
- Exec time: [12.01, 16.17] s
* igt@gem_exec_schedule@independent-rcs0:
- Statuses : 6 fail(s)
- Exec time: [8.57, 17.41] s
* igt@gem_exec_schedule@independent-vcs0:
- Statuses : 5 fail(s) 1 skip(s)
- Exec time: [0.0, 17.19] s
* igt@gem_exec_schedule@independent-vcs1:
- Statuses : 1 fail(s)
- Exec time: [4.42] s
* igt@gem_exec_schedule@independent-vcs2:
- Statuses : 1 fail(s)
- Exec time: [12.73] s
* igt@gem_exec_schedule@independent-vecs0:
- Statuses : 5 fail(s)
- Exec time: [11.76, 15.99] s
* igt@gem_exec_schedule@out-order-bcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.03] s
* igt@gem_exec_schedule@out-order-rcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.02] s
* igt@gem_exec_schedule@out-order-vcs0:
- Statuses : 4 pass(s) 2 skip(s)
- Exec time: [0.0, 0.01] s
* igt@gem_exec_schedule@out-order-vcs1:
- Statuses : 1 pass(s)
- Exec time: [0.02] s
* igt@gem_exec_schedule@out-order-vcs2:
- Statuses :
- Exec time: [None] s
* igt@gem_exec_schedule@out-order-vecs0:
- Statuses : 4 pass(s) 1 skip(
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_3186/
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev
Patchwork
CI[igt-dev] ✗ Fi.CI.BAT: failure for Upgrading the Engine discovery (rev2)
In reply to Ramalingam C== Series Details ==
== Series Details ==
Series: Upgrading the Engine discovery (rev2)
URL : https://patchwork.freedesktop.org/series/62512/
State : failure
== Summary ==
Applying: test/i915: gem_bad_reloc: use the gem_engine_topology library
Applying: test/i915: gem_exec_async: use the gem_engine_topology library
Applying: test/i915: gem_exec_capture: use the gem_engine_topology library
Applying: test/i915: gem_exec_params: use the gem_engine_topology library
Applying: lib/i915: Helper func for engine class to exec_flag Using index info to reconstruct a base tree...
M lib/i915/gem_engine_topology.c
M lib/i915/gem_engine_topology.h
Falling back to patching base and 3-way merge...
Auto-merging lib/i915/gem_engine_topology.h
CONFLICT (content): Merge conflict in lib/i915/gem_engine_topology.h Auto-merging lib/i915/gem_engine_topology.c
CONFLICT (content): Merge conflict in lib/i915/gem_engine_topology.c Patch failed at 0005 lib/i915: Helper func for engine class to exec_flag When you have resolved this problem, run "git am --continue". If you prefer to skip this patch, run "git am --skip" instead. To restore the original branch and stop patching, run "git am --abort".
_______________________________________________
igt-dev mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/igt-dev