| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Makefile.template
|
|
|
|
Instead of passing in a fixed struct, the loader now passes in a list
of __DRIextension structs, to advertise the functionality it can provide
to the driver. Each extension is individually versioned and can be
extended or phased out as the interface develops.
|
|
* Fix crash at context creation in most drivers supporting vblank.
* Don't pass vblank sequence or flags to functions that get passed the drawable
private already.
* Attempt to initialize vblank related drawable private fields just once
per drawable. May need more work in some drivers.
|
|
Consolidate support for synchronizing to and retrieving vblank counters. Also
fix the core vblank code to return monotonic MSC counters, which are required
by some GLX extensions. Adding support for multiple pipes to a low level
driver is fairly easy, the Intel 965 driver provides simple example code (see
intel_buffers.c:intelWindowMoved()).
The new code bumps the media stream counter extension version to 2 and adds a
new getDrawableMSC callback. This callback takes a drawablePrivate pointer,
which is used to calculate the MSC value seen by clients based on the actual
vblank counter(s) returned from the kernel. The new drawable private fields
are as follows:
- vblSeq - used for tracking vblank counts for buffer swapping
- vblFlags - flags (e.g. current pipe), updated by low level driver
- msc_base - MSC counter from the last time the current pipe changed
- vblank_base - kernel DRM vblank counter from the last time the pipe changed
Using the above variables, the core vblank code (in vblank.c) can calculate a
monotonic MSC value. The low level DRI drivers are responsible for updating
the current pipe (by setting VBLANK_FLAG_SECONDARY for example in vblFlags)
along with msc_base and vblank_base whenever the pipe associated with a given
drawable changes (again, see intelWindowMoved for an example of this).
Drivers should fill in the GetDrawableMSC DriverAPIRec field to point to
driDrawableGetMSC32 and add code for pipe switching as outlined above to fully
support the new scheme.
|
|
Many DRI entry points took a __DRInativeDisplay pointer and a screen
index as arguments. The only use for the native display pointer was to
pass it back to the loader when looking up the __DRIscreen for the given
screen index.
Instead, let's just pass in the __DRIscreen pointer directly, which
let's drop the __DRInativeDisplay type and the getScreen function.
The assumption is now that the loader will be able to retrieve context
from the __DRIscreen pointer when necessary.
|
|
|
|
|
|
The coordinates need to be computed after we've got the hw lock.
Code updated to:
1. Ignore all/x/y/width/height/ params passed to Clear func.
2. Pass 0,0,0,0,0 to _swrast_Clear() until they're totally removed.
|
|
request at least drmMinor 6 anyway.
|
|
intention as drm seems to never return EAGAIN) in all radeon drivers.
|
|
|
|
- Makes all three drivers use the same screen structure and setup code, with a
few ifdefs for the separate compilation to deal with symbols not being
available to all drivers and the fact that we have no mechanism for dealing
with different config options for different chip families in the same driver.
These issues should be dealt with later.
- Introduces IS_R[123]00_CLASS(radeonScreenPtr) macro for code for taking
different paths depending on the general class of chipset.
- Adds many new R300-class PCI IDs, though not all those listed in
radeon_driver.c.
|
|
radeon/r200AllocDmaRegionVerts function. No longer compile the radeon_compat stuff which was for drm 1.1 compatibility presumably but unused since about forever into the driver.
|
|
|
|
Use driRenderbuffer's offset, pitch fields in the span routines.
Remove the SetBuffer driver function.
Consolidate the code for setting CTX_RB3D_COLOROFFSET and CTX_RB3D_COLORPITCH
state in new radeonUpdateDrawBuffer() function.
Old code is surrounded by #if 000 / #endif, temporarily.
|
|
that are currently obtained via glXGetProcAddress and all of the XF86DRI
functions are replaced with a funciton table. This table will be passed to
__driCreateNewScreen.
One of the functions in the table is getProcAddress. This allows some
loaders to expose functionality not in all loaders. This will be immediatly
used for glxEnableExtension (formerly known to drivers as
__glXScrEnableExtension). libGL (and in the future libglx) expose this
function so that drivers can enable GLX extensions. libEGL should exposed
eglEnableExtension to enable EGL extensions. The same function cannot be
used for both because the extensions have different names and (possibly)
different semantics. Drivers can optionally use one, both, or neither.
The key parts are in the __DRIinterfaceMethodsRec structure in
dri_interface.h. A pointer to one of these structures is passed into
__driCreateNewScreen. Because of this, the version of the API is bumped to
20050725. Since the previous version(s) were never in a release, their
existance is erased.
I was actually a little surprised by how much code this cuts from the
drivers. A lot of glXGetProcAddress calls disappear, and a lot of
version checks go with them. Nice.
The one thing I'm not sure of is removing __glXInitialize. For some
reason that function was in the glXGetProcAddress table, but *nothing*
in the Mesa tree used it. Did something with DRI conf. use this
function? It seems odd...
|
|
Main driver impacts:
- new code for creating the Mesa GLframebuffer
- new span/pixel read/write code
Some drivers not yet updated/tested.
|
|
|
|
radeon and r200 driver
|
|
|
|
fast z clear and z buffer compression are supported for now, hierarchical-z is not. Still problems with multiple apps and z/stencil readback, which is why hyperz is disabled per default. Also add the new point sprite packet drm 1.13 accepts to the sanity code.
|
|
happen before any state had been set, causing a hang later on. Fix this by
calling r200Flush instead of FIREVERTICES (which checks if any state has been
emitted but not flushed, before calling Flush) in r200Clear. While here, add
some more debugging info which was useful, and remove an unnecessary
save/restore in BackUpAndEmit.
|
|
between pointers in appropriate types.
Submitted by: Ronny V. Vindenes <s864@ii.uib.no>
|
|
state in a ready-to-emit cmdbuf, which avoids the issue Nicolai Haehnle reported
where the check() could return differently during backup-and-emit than it should
have if it were called at the right time. Move the lit emission before most of
the TCL state emission on r200, which fixes neverball issues.
Tested with: r100/r200 with neverball, tuxracer, chromium, quake3, ipers
|
|
|
|
a new cmdbuf, to ensure that state wasn't lost across UNLOCK/LOCK pairs (in the
case of context switching). This was rather inefficient. Instead, after
flushing a cmdbuf, mark the state as needing to be saved on UNLOCK. Then, at
the beginning of flushing a cmdbuf, if we actually have lost the context, go
back and emit a new cmdbuf with the full set of state, before continuing with
the cmdbuf flush. Also, remove the dirty/clean atom lists, since atoms are
emitted in a fixed order these days, and go with a simpler single list.
Provides a 14% improvement in ipers performance in my tests, along with other
apps.
|
|
last commit. I've been convinced by keithw that it's sufficient, and put a note
in the code about it.
Close another race for state in the Clear functions. I made the situation worse
in my last commit, but this should fix things. Might be a slight performance
hit, which could be regained by splitting the R*_FIREVERTICES calls in r*Clear
up so that the EmitState doesn't happen in a separate new cmdbuf.
|
|
errors on r100 and rendering errors and hangs on r200 (same for R100 without
OLD_PACKETS).
If a command buffer filled after some state (EmitState or a VBPNTR write) was
emitted, the lock was grabbed, the buffer flushed, a new buffer prepared, and
the lock dropped. Another client could come in, set its own state as part of
rendering, and when the first client flushed the rendering commands depending
on the previous state, it got the 2nd client's state. This is fixed by checking
for enough space before beginning a set of state emits and rendering, and
flushing the buffer first if so. This guarantees that the buffer won't wrap.
Also, move the "lost_context = 1" from the end of cmdbuf flushing to
UNLOCK_HARDWARE for clarity (at a minimum) that any time the lock is dropped,
state may get overwritten. We don't have enough information at the point of the
LOCK_HARDWARE to reset our state to the last UNLOCK_HARDWARE point in the case
that we did lose our context, but saving the information to rebuild that state
may be a useful optimization (ipers data suggests up to 5%).
|
|
uses the NEW_INTERFACE now so _SOLO isn't necessary anymore.
Tested with the hardware that I own.
|
|
|
|
|
|
|
|
|
|
|
|
of glx/mini. removes glx/mini/drm.h glx/mini/sarea.h
|
|
by calling _mesa_init_driver_functions() and then plugging in the driver-
specific functions.
In particular, make sure ctx->Driver.NewTextureObject points to the
appropriate driver function so that _all_ texture objects are augmented
with the driver-specific data.
Put in a bunch of assertions in the texture-related driver functions that
texObj->DriverData is valid. Remove old dead code in near future.
|
|
|
|
- Allocate vertices explicitly, rather than trying to talk
about dma buffers.
- Clean up the various Flush() operations.
- Don't allow fallbacks any longer. Provide a support function
to detect them ahead o ftime
Updates to tnl_dd_vbtmp.h
- Get rid of power-of-two vertex strides. Pack all vertices tightly.
- Get texunit 2,3 emit working coorrectly.
Other stuff:
- Get rid of lingering Ubyte color support.
- Fix a few compiler warnings.
|
|
|
|
|
|
|
|
|
