summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2007-08-21i915tex: Document zone rendering related definitions.i915tex-zone-renderingKeith Whitwell1-1/+9
2007-07-13fix light_twosideBrian1-7/+4
2007-07-11Fix a few polygon offset errors.Brian3-23/+15
1. edge vectors were miscomputed in do_offset_tri() 2. units and scale assignments were transposed in update_clip_state(). 3. intel->polygon_offset_scale is the MRD, apply it to offset_units, not the scale factor.
2007-07-09fix stage, stage->next mix-upBrian1-1/+1
2007-06-01Move scissor state to dynamic indirect.Keith Whitwell5-56/+89
2007-06-01Larger allocation for indirect state when not using state pool.Keith Whitwell1-3/+13
Work around unnecessary flushes when we run out of indirect state space.
2007-05-30Slight tweak for relocations, search most to least recent.Keith Whitwell1-2/+5
2007-05-30Prelocate the batch and state buffers, avoid most SWZ relocs.Keith Whitwell4-1/+48
2007-05-30Revive pool_validateBuffer for batch pool.Keith Whitwell1-0/+29
Actually hook it up this time.
2007-05-23i915tex: Set frame mode to flushed before flushing old mode.Michel Dänzer1-5/+6
This avoids a recursive flush leading to a crash with several glxgears instances. Hopefully this will also fix multi-context apps.
2007-05-22i915tex: Clear pointers to unreferenced texture buffers.Michel Dänzer1-1/+3
Fixes a crash with fgl_glxgears.
2007-05-22Revert "Reduce size of the state pool allocation"Michel Dänzer1-1/+1
This reverts commit aa85f0fea0add5468ebc4d670e8938f2004a92a1.
2007-05-22Only call driBOOffset() once per validated bufferKeith Whitwell2-8/+16
2007-05-22Add local list of buffers on the dri validation list.Keith Whitwell1-5/+47
Don't repeatedly add buffers to the dri list as there is a cost for doing so in terms of pthread mutex lock/unlock calls.
2007-05-22Rebalance batch buffer and pool sizes.Keith Whitwell2-2/+12
Avoid waiting on batchbuffers during aristocrat benchmark.
2007-05-22Reduce size of the state pool allocationKeith Whitwell1-1/+1
It doesn't need to be so large as the buffers we take from it are much smaller than batchbuffers.
2007-05-22Add driBOAddListItemAndReference()Keith Whitwell2-0/+30
Increment refcount at the same time a buffer is added to a list. Reduces pthread mutex lock/unlock overhead.
2007-05-22Reduce driBO Ref/Unref callsKeith Whitwell1-9/+26
2007-05-14i915tex: Remove some dead code.Michel Dänzer2-34/+0
This eliminates the last compile warnings in the driver.
2007-05-14i915tex: Adapt to HWZ userspace interface cleanup.Michel Dänzer2-32/+2
Also drop private copy of its definitions.
2007-05-14Add new directories for tarballs target.Michel Dänzer1-0/+4
2007-05-14revert one copyright headerKeith Whitwell1-29/+26
2007-05-14more todo updatesKeith Whitwell1-5/+1
2007-05-14update copyright headersKeith Whitwell3-79/+95
2007-05-14TODO updates.Michel Dänzer1-18/+4
2007-05-14i915tex: Disable WindowMoved debugging output by default.Michel Dänzer1-1/+3
2007-05-14increase size of zone arraysKeith Whitwell1-1/+1
2007-05-14update copyright headersKeith Whitwell7-86/+107
2007-05-14Tweak resize predictorKeith Whitwell1-1/+5
2007-05-13i915tex: Fix SWZ FBO coordinate issues.Michel Dänzer2-6/+24
Flush batchbuffer after changing target colour renderbuffer but before setting new cliprects, and only apply drawing offsets when rendering to a window.
2007-05-13Rename some state atomsKeith Whitwell6-12/+11
2007-05-13Improvements to the frame tracker.Keith Whitwell9-166/+88
2007-05-13Reset the contended_lock flag.Keith Whitwell1-0/+2
Wasn't happening previously...
2007-05-13i915tex: Add WAIT_VBLANK call for classic.Michel Dänzer1-4/+9
Also silence classic debugging output by default.
2007-05-13i915tex: Window vs. user created FBO cleanups.Michel Dänzer8-19/+15
2007-05-13Try to make sure dynamic state is reemitted after a flush.Keith Whitwell1-1/+2
2007-05-13Add assertions for indirect state validity.Keith Whitwell1-2/+8
2007-05-13i915tex: Silence remaining HWZ debugging output by default.Michel Dänzer2-4/+6
2007-05-13Make sure MI_WAIT packets are emitted for SWZKeith Whitwell6-36/+75
Untested...
2007-05-13Hack-fix BUFFER_BITS check for clearing FBO'sKeith Whitwell3-4/+8
In many places we check BUFFER_BITS_FRONT_LEFT / BUFFER_BITS_BACK_LEFT, but this is incorrect for FBO rendering which sets BUFFER_BITS_COLOR0, etc. This changes enough code to get clears working on fbos, but it is not a complete fix.
2007-05-13i915tex: Fix and re-enable check_hwz().Michel Dänzer1-3/+1
2007-05-13Merge remote branch 'tg/i915tex-zone-rendering' into i915tex-zone-renderingMichel Dänzer16-113/+211
2007-05-13i915tex: SWZ cliprect handling improvements.Michel Dänzer3-7/+28
Before firing the SWZ batchbuffer, check that the cliprects still match the binning parameters and drop the batchbuffer if not. Also only allow ZONE_INIT to be emitted when there's at least one cliprect.
2007-05-13i915tex: Consolidate code for unmapping batchbuffer data.Michel Dänzer3-38/+18
2007-05-13i915tex: Make sure buffer state gets re-evaluated after cliprect update.Michel Dänzer1-0/+2
2007-05-13i915tex: Protect intel_bufferobj_data against nested locking.Michel Dänzer1-2/+5
2007-05-13i915tex: Make nested locking assertion more useful.Michel Dänzer1-1/+1
We would previously deadlock on the locking mutex before detecting nested locking.
2007-05-13Disable check_hwz()Keith Whitwell1-0/+2
I seem to be ending up in the hwz path even though INTEL_HWZ isn't set?!?
2007-05-13initialize state to swapped buffersKeith Whitwell1-0/+1
2007-05-13improve debug for render trackingKeith Whitwell9-7/+54
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-18 01:42:55 +0000