summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2008-01-15 i965: fix an assert fail in brw_new_batchZou Nan hai
2008-01-14[i915] Fix recursive lock hang in intelContendedLock handling.Eric Anholt
2008-01-14[intel] Delay unmapping of driver-private buffers to reduce map/unmap thrashing.Eric Anholt
Good for a 10-15% improvement to OA.
2008-01-14[965] Force a new vertex upload buffer at new batch time.Eric Anholt
Otherwise, we could choose to upload into the temporary VBO that we just fired off to the hardware. Good for a 60% OA performance improvement.
2008-01-11Allow osmesa to be enabled or disabledDan Nicholson
The conditional in src/mesa/Makefile currently hardcodes the cases where libOSMesa can be built on libGL. Likewise, the xlib case always includes libOSMesa in the stand-alone target. This changes the conditional to a loop over the DRIVER_DIRS variable. This means that any driver configuration can enable or disable osmesa. The current "stand-alone" rule is changed so that DRIVER_DIRS=x11 and DRIVER_DIRS="x11 osmesa" are both respected. The configure option is changed to --enable-gl-osmesa as this change allows libOSMesa to be built upon any of the libGL-enabling drivers.
2008-01-10[965] Remove an open-coded memcpy.Eric Anholt
2008-01-10[intel] Remove gratuitous (batchbuffer) flush before doing buffer clears.Eric Anholt
Increases OA performance by about 3%.
2008-01-10[intel] Report buffer size when allocating with INTEL_DEBUG=bufmgr and TTM.Eric Anholt
2008-01-10[965] Improve performance by allocating CURBE buffers a page at a time.Eric Anholt
Since each one is only 64b, and kernel allocations are a page anyway, this lets us reduce buffer allocation by packing many CURBEs into one buffer, for each batchbuffer submitted. Improves openarena performance by around 10%.
2008-01-10[intel] Remove a gratuitous flush at the end of ClearWithTris.Eric Anholt
2008-01-10[intel] Only flush batch when changing draw buffers, not every cliprect update.Eric Anholt
The previous code would reference freed memory on window moves.
2008-01-10[intel] Add more cliprect modes to cover other meanings for batch emits.Eric Anholt
The previous change gave us only two modes, one which looped over the batch per cliprect (3d drawing) and one that didn't (state updeast). However, we really want 4: - Batch doesn't care about cliprects (state updates) - Batch needs DRAWING_RECTANGLE looping per cliprect (3d drawing) - Batch needs to be executed just once (region fills, copies, etc.) - Batch already includes cliprect handling, and must be flushed by unlock time (copybuffers, clears). All callers should now be fixed to use one of these states for any batchbuffer emits. Thanks to Keith Whitwell for pointing out the failure.
2008-01-10i965: fix bug #14002Xiang, Haihao
2008-01-10i965: fix segfault caused by commit e131c46b20241737ceba4856dbe01dcca6dd2c03.Xiang, Haihao
2008-01-09[intel] Simplify intelCreateBuffer() a bit.Kristian Høgsberg
Drop a bunch of unused arguments from intel_create_renderbuffer() and introduce intel_renderbuffer_set_region() to set the region for a renderbuffer.
2008-01-09[intel] Prepare intelCopyBuffer() for private back buffers.Kristian Høgsberg
2008-01-09Factor out code to do the DRM_I830_GETPARAM ioctl.Kristian Høgsberg
2008-01-09Fold DoBindContext() helper function into driBindContext().Kristian Høgsberg
2008-01-09[intel] Re-allocate backing buffer objects on BufferData.Eric Anholt
This may allow better concurrency (noop in openarena performance now), but is also important for the previous commit -- otherwise, we may end up with BufferData, draw_prims, BufferData and the draw_prims would use the new VBO data instead of old. This could still occur with user-supplied VBOs and poor use of MapBuffer without BufferData.
2008-01-09[965] Allow more than one draw_prims per batchbuffer.Eric Anholt
The comment about (vbo)_exec_api.c appeared to be stale, as the VBO code seems to only use non-named VBOs (not actual VBOs) or freshly-allocated VBO data. This brings a 2x speedup to openarena, because we can submit nearly-full batchbuffers instead of many 450-byte ones.
2008-01-09[965] Replace the always_update dirty flag with BRW_NEW_BATCH.Eric Anholt
This allows us to avoid re-emitting some state when validate_state happens multiple times per batchbuffer. Even though we flush batch per primitive currently, that may still happen already if the primitive changed (this should probably be fixed as well).
2008-01-09[intel] Rename lost_hardware vtbl entry to new_batch.Eric Anholt
Both drivers have ended up relying on lost_hardware being called after each batch buffer, so update the name. This removes one of the calls on 965 whic h was outside of the batchbuffer handling code and just duplicating what had already happened through batchbuffer handling.
2008-01-09[965] Remove drawing rect upload, which is handled (better) by the kernel.Eric Anholt
2008-01-09[intel] Clean up cliprect handling in intel drivers.Eric Anholt
In particular, batch buffers are no longer flushed when switching from CLIPRECTS to NO_CLIPRECTS or vice versa, and 965 just uses DRM cliprect handling for primitives instead of trying to sneak in its own to avoid the DRM stuff. The disadvantage is that we will re-execute state updates per cliprect, but the advantage is that we will be able to accumulate larger batch buffers, which were proving to be a major overhead.
2008-01-09[intel] Remove the dead intel->need_flush member.Eric Anholt
2008-01-09[965] Clarify a bit of index buffer upload code.Eric Anholt
2008-01-09fix bug 13978: glDrawBuffersARB() didn't set all necessary stateBrian
2008-01-08[intel] Add TTM bufmgr debug for type or offset migration of buffers.Eric Anholt
2008-01-07i915: Keith Whitwell's swizzling TEX patch. fix #8283Xiang, Haihao
2008-01-06Replace gl_framebuffer's _ColorDrawBufferMask with _ColorDrawBufferIndexesBrian
Each array element is now a BUFFER_x token rather than a BUFFER_BIT_x bitmask. The number of active color buffers is specified by _NumColorDrawBuffers. This builds on the previous DrawBuffer changes and will help with drivers implementing GL_ARB_draw_buffers.
2008-01-06Simplify ctx->_NumColorDrawBuffers, _ColorDrawBuffers and fix bug 13835.Brian
These fields are no longer indexed by shader output. Now, we just have a simple array of renderbuffer pointers. If the shader writes to gl_FragData[i], send those colors to the N _ColorDrawBuffers. Otherwise, replicate the single gl_FragColor (or the fixed-function color) to the N _ColorDrawBuffers. A few more changes and simplifications can follow from this...
2008-01-05fix depth/1 typo in glTexImage3D proxy codeBrian
2008-01-05check if span length > 0 before calling _swrast_write_rgba_span()Brian
2008-01-04i915: don't validate PS program when falling back to softwareXiang, Haihao
rendering. fix #12786
2008-01-04intel: some initialization for dri_bufmgr_ttmXiang, Haihao
2008-01-03[intel] Add a single-entry relocation buffer cache.Eric Anholt
By avoiding the repeated relocation buffer creation/map/unmap/destroy for each new batch buffer, this improves OpenArena framerates by 30%. Caching batch buffers themselves doesn't appear to be a significant performance win over this change.
2008-01-03[intel] Convert relocations to not be cleared out on buffer submit.Eric Anholt
We have two consumers of relocations. One is static state buffers, which want the same relocation every time. The other is the batchbuffer, which gets thrown out immediately after submit. This lets us reduce repeated computation for static state buffers, and clean up the code by moving relocations nearer to where the state buffer is computed.
2008-01-03[965] Fix some missing initialization in WM keys.Eric Anholt
2008-01-03 fix fd.o bug #13761Zou Nan hai
MRD computation is now changed in mesa core
2008-01-02[965] Convert WM unit to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02[965] Convert VS unit to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02[965] Convert SF unit to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02[965] Convert GS unit to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02[965] Convert clip unit to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02[965] Convert CC unit to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02[965] Convert surface state to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02[965] Convert sampler state to use a cache key instead of brw_cache_data.Eric Anholt
2008-01-02Revert "[intel] Use the memory type mask containing the caching flags."Eric Anholt
This reverts commit 8bb9ae3693362a302206255c61f512d942df9bbf. Validating our kernel buffers with the caching off in flags but on in mask means that the kernel migrates the buffer to be uncached, which is undesired.
2008-01-02[intel] Use the memory type mask containing the caching flags.Eric Anholt
2008-01-02Set correct flags mask when validating buffers.Keith Packard
The 'mask' value used in the validation operation specifies which of the 'flags' bits are being modified. Buffer validation wants to pass the memory type and access mode (rwx) to the kernel so that the buffer will be placed correctly, and so that the right kind of fence will be created (read vs write). That means we actually want a constant mask for these operations, and not something computed from the bits coming in. The constant we want is DRM_BO_MASK_MEM | DRM_BO_FLAG_READ | DRM_BO_FLAG_WRITE | DRM_BO_FLAG_EXE.