| Message ID | 1415260951-27700-1-git-send-email-yakui.zhao@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Thu, Nov 06, 2014 at 04:02:30PM +0800, Zhao Yakui wrote: > After applying the commit(982f7eb238a0898c456e0574dee7c4507738d75f), the OUT_RELOC is > updated on Broadwell and later, which is to handle the 64-bit field of gfx address > internally. In such case some commands should be fixed, otherwise GPU hang will > be triggered when running rendercopy. > (It is already fixed on Broadwell) > > Signed-off-by: Zhao Yakui <yakui.zhao@intel.com> Thanks a lot for taking care of this. Taking about 982f7e, this is a good reason to post the patches on the ml so we have a chance to catch "silent" bugs. Pushed both to master (with reducing the commit message to reasonable line lengths).
diff --git a/lib/rendercopy_gen9.c b/lib/rendercopy_gen9.c index 9ff4b3a..e20a84f 100644 --- a/lib/rendercopy_gen9.c +++ b/lib/rendercopy_gen9.c @@ -397,7 +397,6 @@ static void gen7_emit_vertex_buffer(struct intel_batchbuffer *batch, GEN7_VB0_BUFFER_ADDR_MOD_EN | /* Address Modify Enable */ VERTEX_SIZE << VB0_BUFFER_PITCH_SHIFT); OUT_RELOC(batch->bo, I915_GEM_DOMAIN_VERTEX, 0, offset); - OUT_BATCH(0); OUT_BATCH(3 * VERTEX_SIZE); } @@ -523,12 +522,10 @@ gen9_emit_state_base_address(struct intel_batchbuffer *batch) { /* surface */ OUT_RELOC(batch->bo, I915_GEM_DOMAIN_SAMPLER, 0, BASE_ADDRESS_MODIFY); - OUT_BATCH(0); /* dynamic */ OUT_RELOC(batch->bo, I915_GEM_DOMAIN_RENDER | I915_GEM_DOMAIN_INSTRUCTION, 0, BASE_ADDRESS_MODIFY); - OUT_BATCH(0); /* indirect */ OUT_BATCH(0); @@ -536,7 +533,6 @@ gen9_emit_state_base_address(struct intel_batchbuffer *batch) { /* instruction */ OUT_RELOC(batch->bo, I915_GEM_DOMAIN_INSTRUCTION, 0, BASE_ADDRESS_MODIFY); - OUT_BATCH(0); /* general state buffer size */ OUT_BATCH(0xfffff000 | 1);
After applying the commit(982f7eb238a0898c456e0574dee7c4507738d75f), the OUT_RELOC is updated on Broadwell and later, which is to handle the 64-bit field of gfx address internally. In such case some commands should be fixed, otherwise GPU hang will be triggered when running rendercopy. (It is already fixed on Broadwell) Signed-off-by: Zhao Yakui <yakui.zhao@intel.com> --- lib/rendercopy_gen9.c | 4 ---- 1 file changed, 4 deletions(-)