| Age | Commit message (Collapse) | Author |
|---|
|
This is currently believed to work but be a significant performance loss.
Performance recovery should be soon to follow.
The dri_bo_fake_disable_backing_store() call was added to allow backing store
disable like bufmgr_fake.c did, which is a significant performance win (though
it's missing the no-fence-subdata part).
This commit is a squash merge of the 965-ttm branch, which had some history
I wanted to avoid pulling due to noisiness and brokenness at many points
for git-bisecting.
|
|
|
|
In the process, fix some alignment issues:
- Scratch space allocation was aligned into units of 1KB, while the allocation
wanted units of bytes, so we never allocated enough space for scratch.
- GRF register count was programmed as ALIGN(val - 1, 16) / 16 instead of
ALIGN(val, 16) / 16 - 1, which overcounted for val != 16n+1.
|
|
This reverts commit b2f1aa2389473ed09170713301b042661d70a48e.
Somehow I ended up with my branch's save-this-while-I-work-on-master commit
actually on master.
|
|
|
|
This code existed to dump logs of hardware access to be replayed in simulation.
Since we have real hardware now, it's not really needed.
|
|
vertex/fragment programs provided as const.
bmSetFenceLock should return bmSetFence value.
|
|
Signed-off-by: Keith Packard <keithp@neko.keithp.com>
|
|
passing them to the kernel. This works because all drawing commands
in the 965 driver are emitted with the lock held and the batchbuffer
is always flushed prior to releasing the lock. This allows multiple
cliprects to be dealt with, without replaying entire batchbuffers and
redundantly re-emitting state.
|
|
identify context switches between members of a share group -
ie. multiple contexts in a single application, possibly on different
threads. In this case the contexts share a bufmgr instance and there
is no need to evict textures - so don't.
2) Use a new flag 'need_flush' to ensure hardware rendering is flushed
prior to starting a software fallback.
|
|
|
|
hack to the drm to disable command verification on the cmd_buffer
ioctl. Doesn't exactly replay as commands are normally delivered as
batchbuffers but are captured and replayed as commands on the ring.
|
|
Get aubfile generation working again.
|
|
rendering operations to take place after evicting all resident
buffers.
Cope better with memory allocation failures throughout the driver and
improve tracking of failures.
|
|
|
|
a multiple of alignment to avoid accumulating unusable free blocks.
|
|
This driver comes from Tungsten Graphics, with a few further modifications by
Intel.
|
