but includes debugging to ensure that the reg being submitted is the
one that follows in the packet. Convert most uses of OUT_RING to it,
and convert a couple of OUT_REG sets to DMA_PACKET0/OUT_RING_REG. Also,
add checking to see if more registers are submitted to a DMA_PACKET0
than should be, to avoid hangs during stupid mistakes (checking for
less isn't done).
reading a register/attempting DMA. Now it'll wait for a certain number
of seconds rather than a certain number of times through the loop
before deciding that it's timed out and resetting the hardware. Also,
add more timeout handling, and reset the draw state after resetting the
engine.
are always emitted. Fix the real problem, which was not enough regs
being initialized in ati_draw.c. Fix a typo that was resulting in alpha
coming out as 0 * src or 0 * broken instead of src * 1 or src * mask.
Assign the blending results to R0, as appears to be necessary. Unbreak
the dst-alpha-blend-with-no-dst-alpha code. Yow. And set the right DMA
count for the r200 traps code.
non-mask rendering. Reenable it. Also, R200TexFormats was used instead
of R100 in one place. Harmless so far, because the formats were in the
same order.
troubles.
- Move the RemoveBlockAndWakeupHandlers to match
RegisterBlockAndWakeupHandlers.
- Enable R100 trapezoid "acceleration" when DRI is working, so that it can
be exposed and worked on.
instead of ErrorFs for things that are really bad, and put limits on
some loops. Now, sometimes instead of hanging the entire system, we
(mostly-) cleanly drop to console when the card has hung.
not doing an add of 1 per triangle like I hoped, and instead seems to
be saturating all the pixels or something. The R100 acceleration
renders pretty well, with some gaps. Note that both are slower than
software due to lack of DMA to submit vertices.
- Mostly fix R128 and Radeon transform support, including supporting
bilinear filtering on R128. Subpixel offsets are still probably an
issue (reported by rendercheck), but I want to make 100% sure about my
understanding of the protocol before changing everybody, including fb.
- Add support for dst formats without alpha to R128 Composite.
- Remove the R128 Blend code, which has long outlived its usefulness. (I
kept it around for one reason: It could be useful for the w/h > 1024
case with no mask and a non-src op. That seems pretty infrequent and
not worth the trouble).
might be able to do Precise in hardware, so leave it up to the driver.
- Add a helper function for computing a set of offsets for smooth trapezoid
rasterizing using many sharp trapezoids.
RasterizeTrapezoid screen function. These hooks will be called for
imprecise, non-sharp trapezoids with A8 destinations.
Note that the current main consumer of trapezoids, cairo, is requesting
precise, sharp trapezoids by not changing the default Picture
attributes, but gets non-sharp effects in software because fb bases its
choice of sharp/non-sharp on the mask format being A8 vs A1, and cairo
asks for A8. Follow fb's (poor?) example by ignoring the sharp setting
and basing the choice off of the mask being A8.
heuristic, delay the decision until the first
kaaPixmapUse{Screen|Memory}, and put it in framebuffer if UseScreen was
called. Provides a significant improvement in cairo speeds (100%
speedup in cairogears here) and is likely to improve text performance
as well.
rate selection necessary for Mac fbdev
Don't know about fb changes to pixmaps, so can't track dirt.
Add Mac specific 1280x854 mode. Warn when requested mode isn't found.
Add ability to soft-boot video cards.
Add region expand request. FIXME: need test cases
- On R128, don't refer to an old Composite's mask transform when the
current Composite doesn't have a mask.
- Staticize some global variables in r128_composite.c.
header include. I don't know how this happened, but I'm going to blame
a "few" fscks having happened between testing the code and committing
(due to other driver changes that made me not blame dirty for the
failures I saw) for disappearance of kasync.c changes. Fixes a lot of
corruption.
offscreen pixmap area, and set it when any rendering occurs. When
moving a pixmap out of offscreen, don't read data back if it wasn't
dirtied (compared to the system memory copy).
out when allocation can't find a free area of the requested size. When
offscreen pixmaps get used, the offscreen area's score is increased by
a constant value. Every certain number of increases, all offscreen area
scores get decreased by a fraction. When choosing a set of areas to
remove for a new allocation, the set of areas with the smallest total
score is chosen for removal. While this is not the smartest system, it
prevents things like always removing the first offscreen area in memory
(likely the most recent) to be kicked out when doing replacing.
solid-fill-based composite acceleration.
- Use a real pixmap when doing an UploadToScratch (For pDrawable->type ==
DRAWABLE_WINDOW, you need to get the backing pixmap).
- Pass back the x/y offsets from kaaGetOffscreenPixmap unconditionally,
because they'll be used in the scratch case.
- Turn on the Render acceleration for Rage 128 and Radeon 100-series at
last!
Fix a few allocation bugs with alternate visual ids Allow for non-8/8/8
alternate visuals
Turn off any existing shadow before enabling it again (avoids
re-registering existing damage)
Add some validation code to catch re-registered damages
addition to the ARGB32 one. This allows 'glitz' to run on top of any X
server using mesa.
Switch to using 32bpp for depth 24 pixmaps (even when the frame buffer is
not depth 24).
KdCheckSync -- the boolean used in the latter won't be set yet.
Oops. == instead of =.
Must sync hardware before rasterizing trapezoids in case the mask is in
off-screen memory and has just been erased. Yes, it is silly to place
masks in off-screen memory. That's a separate issue.
Off-screen reallocation could have used a stale pointer.
Separate framebuffer mapping computation from actual frame buffer mapping.
Now map the frame buffer from vesaEnable so that VT switch shares the
same mapping code. This makes sure any shadow framebuffer is allocated
again.
- Add monochrome hardware cursor support.
- Try to auto-detect AGP support for DRI on Radeons. And fail. Detect it
properly on R128.
- Set up card for pseudo-DMA if possible. Convert 2D rendering code to
prepare DMA packets only. Use generic code to decode DMA packets to
MMIO if PDMA is unavailable. Add WIP code to support "real" DMA without
DRM support.
- Dispatch pending DMA commands when the server sleeps. Otherwise some
things, such as typing in an xterm, wouldn't show up for a time.
- Fix Radeon Composite acceleration in many ways, and add Rage 128
Composite acceleration. Disable them both due to still-not-understood
issues they have. They fail with In, Out, AtopReverse, and Xor, and
text rendering is strange.
- Add textured XV support for R100 and Rage 128. No brightness/sat
controls, but it does support multiple ports, and cooperates with
Composite.
- Add WIP code for hostdata uploads.
- Many cleanups and fixes.
ati driver, cascading between VESA and FBDEV. We only have init
functions done so far; need to add all of the others. Fixed some
compiler warnings. Whitespace and formatting cleanups (using 4 spaces,
no tabs)
Get rid of all references to SCO.
Change SDK include dir to $(includedir)/xizzle. Add SDK libs where
necessary.
Reformat to be nicer and easier to shuffle around; also, fix lib ordering
so we get so much closer to the elusive final link. Shuffle
common/xf86Init.c into libxizzle.a.
Fix a couple of early snafus - s/BUILDXI/XINPUT/, et al; make the SDK stuff
conditional as needed; fix the SBus includes.
Name library os-support/libxizzleos.a, not os-support/foo/libxizzlefoo.a.
Clean up ARCH_SOURCES so it's always initialised to something.
Move linked libraries to _LIBADD, which somehow escaped my attention. Make
inclusion of drm/libxizzlelinuxdrm.a dependent on DRI.
s/VERSION/VBE_VERSION/;
Axe this redundant dir.
All Xizzle-specific: axe hw/xizzle/Xi, shuffle link order, fix list of
required modules to be vaguely sane, add some pertinent libs/incs.
anyone uses this in production, a big scary monster will eat them. hrm,
perhaps i should make it have a --i-know-what-i'm-doing param that it
doens't start without, heh
- Reserve areas for back/depth/span when USING_DRI && GLXEXT. This would be
better in a TransitionTo3d, but we'd need to work with the offscreen
memory manager for that.
- Misc. fixes to ati_dri.c for DRI+GLX. Needs more work still.
pixmap into temporary offscreen storage. Subsequent UploadToScratch may
clobber the data of previous ones. This allows hardware acceleration of
composite operations on glyphs.
- Add a new UploadToScreen kaa hook for doing the actual moving of data to
framebuffer. This would allow us to do things like hostdata blits or
memcpy to agp and then blit.
- Add an UploadToScreen on ATI which is just memcpy, but which will be
replaced with a hostdata blit soon.
- Add UploadToScratch on ATI and reserve 64k of scratch space. This
provided a 3x speedup of rgb24text on my Radeon.
emitting as a tri fan rather than a tri list. A rect list would save an
additional vertex (out of 4) per rectangle, but there's no measurable
speed difference and the tri fan may be useful when transforms come
into play.
accel. I don't 100% trust that the math works for Saturate, but I can't
tell from existing information.
- Fix texture pitch fallback checks.
- Fallback when src or mask have transforms.
- Disable Radeon Composite accel until the offset thing is fixed.
- Set offscreenPitch to 64 on Radeon thanks to new information and a kaa
fix. Fixes acceleration at width!=1024.
functions. Along with making things more readable, it fixes a problem
where the coordinates would get messed up if acceleration failed due to
things like pixmaps being in the wrong locations.
pixmaps are offscreen and don't have alpha maps. It is the last case
checked before going to software fallback
- Use the new Composite hook in the ati driver to implement acceleration of
most Composites that get done in an xcompmgr environment on r100 series
cards. It is only available when using the DRM. There are still some
corruption issues, but the DRI is still non-default and I need to get
this into version control.