diff options
| author | Kenneth Graunke <[email protected]> | 2012-08-10 10:26:03 -0700 |
|---|---|---|
| committer | Kenneth Graunke <[email protected]> | 2012-08-12 20:12:13 -0700 |
| commit | 9da50667f490ba2c6240f4c91c9707e3f181adae (patch) | |
| tree | caae15ed5579ce228c0a5c8fc30c65738a1aee69 | |
| parent | 006c1a3c652803e2ff8d5f7ea55c9cb5d8353279 (diff) | |
intel: Move finish_batch() call before MI_BATCH_BUFFER_END and padding.
On Gen4+, brw_finish_batch() calls brw_emit_query_end(), which emits
some extra PIPE_CONTROLs to capture the current occlusion query data.
Unfortunately, it was being called *after* _intel_batchbuffer_flush
added the MI_BATCH_BUFFER_END, meaning those PIPE_CONTROLs didn't get
inside the batch.
Not only does this likely cause bogus occlusion query values, it can
also cause crashes: with the recent change to use 64-bit depth count
writes on Gen6+, we started emitting an odd-length PIPE_CONTROL, which
happened after the MI_NOOP padding. This resulted in an odd-length
batch buffer, which resulted in execbuf2 returning -EINVAL and the
application dying with an intel_do_flush_locked failure.
On older generations, finish_batch() doesn't emit any state, so this
change shouldn't have any effect.
Huge thanks to Chris Wilson for helping me figure this out.
NOTE: This is a candidate for stable release branches.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=53311
Signed-off-by: Kenneth Graunke <[email protected]>
Reviewed-by: Chris Wilson <[email protected]>
Reviewed-by: Eric Anholt <[email protected]>
| -rw-r--r-- | src/mesa/drivers/dri/intel/intel_batchbuffer.c | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/src/mesa/drivers/dri/intel/intel_batchbuffer.c b/src/mesa/drivers/dri/intel/intel_batchbuffer.c index ac133ee8e03..06cbaecc74d 100644 --- a/src/mesa/drivers/dri/intel/intel_batchbuffer.c +++ b/src/mesa/drivers/dri/intel/intel_batchbuffer.c @@ -230,6 +230,9 @@ _intel_batchbuffer_flush(struct intel_context *intel, intel->batch.reserved_space = 0; + if (intel->vtbl.finish_batch) + intel->vtbl.finish_batch(intel); + /* Mark the end of the buffer. */ intel_batchbuffer_emit_dword(intel, MI_BATCH_BUFFER_END); if (intel->batch.used & 1) { @@ -237,9 +240,6 @@ _intel_batchbuffer_flush(struct intel_context *intel, intel_batchbuffer_emit_dword(intel, MI_NOOP); } - if (intel->vtbl.finish_batch) - intel->vtbl.finish_batch(intel); - intel_upload_finish(intel); /* Check that we didn't just wrap our batchbuffer at a bad time. */ |