summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorKenneth Graunke <[email protected]>2013-07-11 10:24:15 -0700
committerKenneth Graunke <[email protected]>2013-07-18 16:57:21 -0700
commit7ce5c6b2147f9af524f4073b6ee3b29c0a5a1522 (patch)
treee97358dbc8cb24baa1fdb2e29feb2e70d6e79274
parent30f33deccb7d481f638e34f2d9d9c107df871eab (diff)
i965: Combine URB code emission into a single group.
All four URB packets need to be programmed together in order for the GPU state to be valid. Putting them in separate BEGIN..ADVANCE blocks is risky: if we're nearing the end of a batch, the batch could be flushed inbetween two of the commands, causing the URB programming to be split into two batchbuffers. This -might- be okay with hardware contexts, but it offers no advantages over keeping them together, and has a potential for hangs. Putting them into a single BEGIN..ADVANCE block ensures they'll be kept in the same batch, which seems wise. Signed-off-by: Kenneth Graunke <[email protected]>
-rw-r--r--src/mesa/drivers/dri/i965/gen7_urb.c12
1 files changed, 2 insertions, 10 deletions
diff --git a/src/mesa/drivers/dri/i965/gen7_urb.c b/src/mesa/drivers/dri/i965/gen7_urb.c
index ce01ddab66a..350f644dfcd 100644
--- a/src/mesa/drivers/dri/i965/gen7_urb.c
+++ b/src/mesa/drivers/dri/i965/gen7_urb.c
@@ -60,12 +60,10 @@ gen7_allocate_push_constants(struct brw_context *brw)
if (brw->is_haswell && brw->gt == 3)
size = 16;
- BEGIN_BATCH(2);
+ BEGIN_BATCH(4);
OUT_BATCH(_3DSTATE_PUSH_CONSTANT_ALLOC_VS << 16 | (2 - 2));
OUT_BATCH(size);
- ADVANCE_BATCH();
- BEGIN_BATCH(2);
OUT_BATCH(_3DSTATE_PUSH_CONSTANT_ALLOC_PS << 16 | (2 - 2));
OUT_BATCH(size | size << GEN7_PUSH_CONSTANT_BUFFER_OFFSET_SHIFT);
ADVANCE_BATCH();
@@ -105,27 +103,21 @@ void
gen7_emit_urb_state(struct brw_context *brw, GLuint nr_vs_entries,
GLuint vs_size, GLuint vs_start)
{
- BEGIN_BATCH(2);
+ BEGIN_BATCH(8);
OUT_BATCH(_3DSTATE_URB_VS << 16 | (2 - 2));
OUT_BATCH(nr_vs_entries |
((vs_size - 1) << GEN7_URB_ENTRY_SIZE_SHIFT) |
(vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));
- ADVANCE_BATCH();
/* Allocate the GS, HS, and DS zero space - we don't use them. */
- BEGIN_BATCH(2);
OUT_BATCH(_3DSTATE_URB_GS << 16 | (2 - 2));
OUT_BATCH((0 << GEN7_URB_ENTRY_SIZE_SHIFT) |
(vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));
- ADVANCE_BATCH();
- BEGIN_BATCH(2);
OUT_BATCH(_3DSTATE_URB_HS << 16 | (2 - 2));
OUT_BATCH((0 << GEN7_URB_ENTRY_SIZE_SHIFT) |
(vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));
- ADVANCE_BATCH();
- BEGIN_BATCH(2);
OUT_BATCH(_3DSTATE_URB_DS << 16 | (2 - 2));
OUT_BATCH((0 << GEN7_URB_ENTRY_SIZE_SHIFT) |
(vs_start << GEN7_URB_STARTING_ADDRESS_SHIFT));