r600g: don't call buffer_wait in buffer_mmap_sync_with_rings
The winsys should do this, because it measures how much time we spend in buffer_map doing synchronization, which can be viewed with the gallium HUD. Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
parent
94d294137e
commit
862f69fbe1
|
@ -292,10 +292,9 @@ void *r600_buffer_mmap_sync_with_rings(struct r600_context *ctx,
|
|||
ctx->ws->cs_sync_flush(ctx->rings.dma.cs);
|
||||
}
|
||||
}
|
||||
ctx->ws->buffer_wait(resource->buf, rusage);
|
||||
|
||||
/* at this point everything is synchronized */
|
||||
return ctx->ws->buffer_map(resource->cs_buf, NULL, usage | PIPE_TRANSFER_UNSYNCHRONIZED);
|
||||
return ctx->ws->buffer_map(resource->cs_buf, NULL, usage);
|
||||
}
|
||||
|
||||
static void r600_flush_from_winsys(void *ctx, unsigned flags)
|
||||
|
|
Loading…
Reference in New Issue