Ben Skeggs [Wed, 16 Mar 2005 00:54:09 +0000 (00:54 +0000)]
Fixed some strangeness when mixing textured/flat primitives. Skipping texture units should render correctly also. Aapo, I dont think I broke glxgears again on you...
Roland Scheidegger [Tue, 15 Mar 2005 22:23:29 +0000 (22:23 +0000)]
add support for user-configurable brilinear filtering on r200
Aapo Tahkola [Tue, 15 Mar 2005 20:55:29 +0000 (20:55 +0000)]
Fix to get glxgears display anything.
Ben Skeggs [Tue, 15 Mar 2005 17:09:35 +0000 (17:09 +0000)]
For some reason the Makefile didn't commit..
Ben Skeggs [Tue, 15 Mar 2005 17:00:44 +0000 (17:00 +0000)]
Initial multitexturing support. Old behaviour can be re-enabled by changing ifdefs in r300_setup_rs_unit and r300SetupPixelShader in r300_state.c\nRequires DRM update.
Ian Romanick [Mon, 14 Mar 2005 22:09:31 +0000 (22:09 +0000)]
If setting GL_PACK_INVERT_MESA results in a GL error, log it.
Felix Kuehling [Mon, 14 Mar 2005 21:52:46 +0000 (21:52 +0000)]
When timestamping textures, need to emit the event with SAVAGE_WAIT_3D
set. Otherwise the 3D engine may still be using a texture that is
being overwritten. In order to minimize the cost of waiting, timestamp
textures only when needed: when a texture image changes, when a
different texture is bound to a texture unit or when a texture unit is
disabled. (Used to be after flushing every command buffer.)
Peter Zubaj [Mon, 14 Mar 2005 20:35:00 +0000 (20:35 +0000)]
Stencil support
Reflex from mesa demos doesn't work
TODO - double side stencil
I hope that I didn't break anything
Michal Krol [Mon, 14 Mar 2005 12:43:59 +0000 (12:43 +0000)]
re-enable slang sources
Michal Krol [Mon, 14 Mar 2005 12:24:32 +0000 (12:24 +0000)]
make it compile under linux
Michal Krol [Mon, 14 Mar 2005 12:23:14 +0000 (12:23 +0000)]
make it compile under gcc 3.3.2
Michal Krol [Mon, 14 Mar 2005 12:19:36 +0000 (12:19 +0000)]
fix enum comma
Aapo Tahkola [Sun, 13 Mar 2005 12:23:08 +0000 (12:23 +0000)]
Replacing alpha test bits with values that give desired results as compaired to software rendering. Alpha test in textures is still completely broken as can be seen in progs/demos/fire and various games.
Felix Kuehling [Sun, 13 Mar 2005 00:20:03 +0000 (00:20 +0000)]
Updated savage_dri.h from Xorg CVS. This removes definitions that are
not part of the DDX->DRI interface and should never have been in this
file in the first place.
Felix Kuehling [Sat, 12 Mar 2005 23:00:53 +0000 (23:00 +0000)]
Removed savagedma.[ch]. They have been unused for a while and were only
kept around for reference. Now that command DMA is implemented in the
DRM module they are no longer needed.
Aapo Tahkola [Sat, 12 Mar 2005 10:14:09 +0000 (10:14 +0000)]
Unlike radeon and r200 drivers r300AllocCmdBuf allocs dwords not bytes.
Ian Romanick [Fri, 11 Mar 2005 20:55:03 +0000 (20:55 +0000)]
Add support for ARB_draw_buffers and ARB_occlusion_query. The GLX protocol
for these extensions (as well as ARB_vertex_program and ARB_matrix_palette)
was just approved by the ARB on 8-Mar-2005. Now the only extension missing
for 1.5 support is ARB_vertex_buffer_object.
The opcodes for ARB_matrix_palette were also added to gl_API.xml. Since
this extension isn't supported by Mesa, no code is generated for it. Some
tabs were also converted to spaces in the comment for
GetCompressedTexImageARB.
Brian Paul [Fri, 11 Mar 2005 14:54:51 +0000 (14:54 +0000)]
add MAT_DIRTY_FLAGS in _math_matrix_mul_floats(), (bug 2696)
Paul Mackerras [Fri, 11 Mar 2005 11:38:23 +0000 (11:38 +0000)]
Add suitable definitions for the _REV formats, plus a comment
explaining why they are the same as the non-REV formats now that
textures don't get byte-swapped on their way to the chip.
Fix the ARGB4444 and L8 formats.
Add entries for the A8 and RGB332 formats.
Felix Kuehling [Thu, 10 Mar 2005 22:45:00 +0000 (22:45 +0000)]
- Flush before updating the timestamp of a texture heap from a bound texture.
- Flush before using a new texture or before disabling a texture unit,
because savageFlushCmdBuf can only update the timestamp of the last
used texture. This fixes corruption in quake2 with single-textured
lighting.
Felix Kuehling [Wed, 9 Mar 2005 15:28:47 +0000 (15:28 +0000)]
If the texture image changed force emitting the texture address, even
if it didn't change. This seems to flush texture cashes. Fixes
multi-textured lighting in quake2.
Vladimir Dergachev [Wed, 9 Mar 2005 05:31:31 +0000 (05:31 +0000)]
Don't use R200 texture constants - R300 fields are all different.
Vladimir Dergachev [Wed, 9 Mar 2005 00:40:55 +0000 (00:40 +0000)]
Deal with texture formats in a more robust way.
Aapo Tahkola [Tue, 8 Mar 2005 22:29:45 +0000 (22:29 +0000)]
Replacing some RADEON cmd types with r300 counterparts.
Aapo Tahkola [Tue, 8 Mar 2005 03:29:46 +0000 (03:29 +0000)]
Improving Vladimirs alpha test fix a bit as it turns out r300Enable didnt correctly disable alpha test. Vertex programs with high vertex count now run a lot faster as software fallback is no longer on. Also disabling 'temp to result'-instructions as they probably violate the spec and dont seem to be something thats needed.
Ian Romanick [Sun, 6 Mar 2005 08:55:39 +0000 (08:55 +0000)]
Actually *use* the count_scale value associated with a parameter.
Felix Kuehling [Sun, 6 Mar 2005 03:52:01 +0000 (03:52 +0000)]
Don't synchronize after each frame. Allow the GPU to be one frame
ahead of the CPU for more parallelism of CPU and GPU. Unfortunately
there seems to be some broken hardware (like my ProSavageDDR) on which
status register updates are delayed sometimes. This leads to very
jerky animation if the hardware can buffer more than the current
frame. A new option "sync_frames" can be used as a remedy on such
broken hardware.
Felix Kuehling [Sun, 6 Mar 2005 01:28:30 +0000 (01:28 +0000)]
- Fixed two bugs related to DMA buffer handling that showed up with the
new DRM version 2.4 and command DMA.
- Flush less.
- Bumped the driver date.
Vladimir Dergachev [Sat, 5 Mar 2005 22:20:57 +0000 (22:20 +0000)]
Add format 15. It turned out to be X,X,X,X,X8.
Dave Airlie [Sat, 5 Mar 2005 06:38:59 +0000 (06:38 +0000)]
Fixing memset on ia64 & other archs
From: Stephane Marchesin
Vladimir Dergachev [Sat, 5 Mar 2005 06:31:36 +0000 (06:31 +0000)]
Enable VB mode rendering by default. If anyone thinks this is wrong, *please* feel free to change it back.
My reasons:
* immediate mode is broken
* vb mode is faster
* vb mode is the right way to do it.
Vladimir Dergachev [Sat, 5 Mar 2005 06:10:09 +0000 (06:10 +0000)]
Do properly initialize ALPHA_TEST register instead of carefully preserving bogus value. This fixes many, many rendering issues which I wrongly blamed on texture programming or performance..
Dave Airlie [Sat, 5 Mar 2005 04:00:09 +0000 (04:00 +0000)]
fix for solo to set the mode, and make solo gcc2 friendly...
From: Stephane Marchesin
Roland Scheidegger [Sat, 5 Mar 2005 00:04:20 +0000 (00:04 +0000)]
remove one of the two _tnl_isolate_materials function calls
Adam Jackson [Fri, 4 Mar 2005 20:18:55 +0000 (20:18 +0000)]
Add preliminary EGL headers.
Ian Romanick [Fri, 4 Mar 2005 17:53:24 +0000 (17:53 +0000)]
Import fixes from X.org tree (Søren Sandmann <sandmann@redhat.com>).
Rune Petersen [Fri, 4 Mar 2005 15:48:40 +0000 (15:48 +0000)]
Set pipes for R420 cards. Gives a nice speedup. This needs to be tested on r420 cards with less than 16 pipes.
Vladimir Dergachev [Thu, 3 Mar 2005 23:18:19 +0000 (23:18 +0000)]
Add a couple of helper functions for completeness.
Ian Romanick [Thu, 3 Mar 2005 21:22:46 +0000 (21:22 +0000)]
Fix some errors in the parameter descriptions in serveral vertex / fragment
program related functions.
Ian Romanick [Thu, 3 Mar 2005 21:21:59 +0000 (21:21 +0000)]
Check for some common function parameter description errors in the
endElement handler for <function>. This catches the errors as early as
possible and makes debugging other code easier.
Ian Romanick [Thu, 3 Mar 2005 19:56:11 +0000 (19:56 +0000)]
When building in the X-Server, attr.depth doesn't exist. Instead, use
(XMesaDrawable)w)->depth in the error message.
Brian Paul [Thu, 3 Mar 2005 02:05:33 +0000 (02:05 +0000)]
change gl_buffer_object's Size field to GLsizeiptrARB type
Brian Paul [Thu, 3 Mar 2005 01:44:42 +0000 (01:44 +0000)]
Added linux-x86-32 config to force building 32-bit objects in a 64-bit environment.
Brian Paul [Thu, 3 Mar 2005 01:38:13 +0000 (01:38 +0000)]
if building 32-bit objects in 64-bit environment, use -m32 flag
Alan Hourihane [Wed, 2 Mar 2005 18:57:01 +0000 (18:57 +0000)]
use COPY_CLEAN_4V macro to replace using both ASSIGN_4V & COPY_SZ_4V
Brian Paul [Wed, 2 Mar 2005 16:37:24 +0000 (16:37 +0000)]
Added a bunch of new comments, minor code clean-ups.
Brian Paul [Wed, 2 Mar 2005 16:31:58 +0000 (16:31 +0000)]
Using glColor3 commands to update materials could result in an undefined
alpha value. Fixed.
Brian Paul [Tue, 1 Mar 2005 02:51:07 +0000 (02:51 +0000)]
added stub functions for GLX_MESA_allocate_memory
Ian Romanick [Mon, 28 Feb 2005 19:37:10 +0000 (19:37 +0000)]
Clean-up some warnings caused by the addition of vertex / fragment program
protocol support (reported by Adam Jackson). Added code to validate the
'type' parameter to the draw element functions.
Brian Paul [Mon, 28 Feb 2005 17:45:11 +0000 (17:45 +0000)]
additional parenthesis in TNL_CONTEXT macro (Nicolai Haehnle)
Aapo Tahkola [Mon, 28 Feb 2005 00:27:02 +0000 (00:27 +0000)]
fix for 'nasty bug' and some sanity checks to avoid buffer overruns. Bumping VSF_MAX_FRAGMENT_LENGTH as it seems a bit low otherwise.
Brian Paul [Sun, 27 Feb 2005 16:23:41 +0000 (16:23 +0000)]
silence warnings
hmarson [Sun, 27 Feb 2005 12:12:25 +0000 (12:12 +0000)]
Added 0x4e51 - Saphire 9600 256MB
0x4e71 - Saphire 9600 256MB - Second Head...
Dave Airlie [Sun, 27 Feb 2005 07:43:18 +0000 (07:43 +0000)]
precedence precedence precedence.. add some brackets
Brian Paul [Sat, 26 Feb 2005 17:16:12 +0000 (17:16 +0000)]
Rename _mesa_update_buffers() to _mesa_update_draw_buffer_bounds() and do
additional checks.
Replace _mesa_init_buffers() with _mesa_init_scissor() and _mesa_init_multisample().
Dave Airlie [Sat, 26 Feb 2005 05:24:04 +0000 (05:24 +0000)]
Add a dri config option to enable the max texture level hack
make ycbcr depend on a CHIPSET define .. needs to be filled in though
Dave Airlie [Sat, 26 Feb 2005 04:56:25 +0000 (04:56 +0000)]
use girl2.rgb by default
Dave Airlie [Sat, 26 Feb 2005 04:51:47 +0000 (04:51 +0000)]
Add glXAllocateMemoryMESA demo app using ycbcr...
add girl2.rgb which is a slightly resized girl.rgb so client texturing works
Dave Airlie [Sat, 26 Feb 2005 03:55:44 +0000 (03:55 +0000)]
bring over structs from Xorg glx.h to make Mesa glx.h the one true glx.h
Dave Airlie [Sat, 26 Feb 2005 03:51:17 +0000 (03:51 +0000)]
Add GLX_MESA_allocate_memory from Xorg glx.h
Ian Romanick [Sat, 26 Feb 2005 01:09:35 +0000 (01:09 +0000)]
Fairly significant changes to enums.c and the way it is generated. enums.c
now contains 3 static tables. The first table is a single, large string of
all the enum names. The second table is an array, sorted by enum name, of
indexes to the string table and the matching enum value. The extra string
table is used to eliminate relocs (and save space) in the compiled file.
The third table is an array, sorted by enum value, of indexes into the
second table.
The [name, enum] table contains all of the enums, but the table sorted by
enum-value does not. This table contains one entry per enum value. For
enum values that have multiple names (e.g., 0x84C0 has GL_TEXTURE0_ARB and
GL_TEXTURE0), only an index to the "best" name will appear in the table.
gl_enums.py gives precedence to "core" GL versions of names, followed by ARB
versions, followed by EXT versions, followed, finally, by vendor versions
(i.e., anything that doesn't fall into one of the previous categories). By
filtering the unneeded elements from this table, not only can we guarantee
determinism in the generated tables, but we save 364 elements in the table.
The optimizations outlined above reduced the size of the stripped enums.o
(on x86) from ~80KB to ~53KB.
The internal organization of gl_enums.py was also heavily modified.
Previously enums were stored in an unsorted list as [value, name] tuples
(basically). This list was then sorted, using a user-specified compare
function (i.e., VERY slow in most Python implementations) to generate a
table sorted by enum value. It was then sorted again, using another
user-specified compare function, to generate a table sorted by name.
Enums are now stored in a dictionary, called enum_table, with the enum value
as the key. Each dictionary element is a list of [name, priority] pairs.
The priority is determined as described above. The table sorted by enum
value is generated by sorting the keys of enum_table (i.e., very fast). The
tables sorted by name are generated by creating a list, called name_table,
of [name, enum value] pairs. This table can then be sorted by doing
name_table.sort() (i.e., very fast).
The result is a fair amount more Python code, but execution time was reduced
from ~14 seconds to ~2 seconds.
Ian Romanick [Fri, 25 Feb 2005 23:02:33 +0000 (23:02 +0000)]
Accidentally committed wrong version with the previous commit set.
Ian Romanick [Fri, 25 Feb 2005 22:46:30 +0000 (22:46 +0000)]
Add GLX protocol support for ARB_fragement_program,
ARB_fragment_program_shadow, ARB_vertex_program, NV_fragment_program,
NV_fragment_program_option, NV_fragment_program2, NV_vertex_program,
NV_vertex_program1_1, NV_vertex_program2, NV_vertex_program2_option,
NV_vertex_program3, and ATI_text_fragment_shader.
Ian Romanick [Fri, 25 Feb 2005 01:48:25 +0000 (01:48 +0000)]
Convert the static functions read_pixel_reply, read_reply,
setup_single_request, and setup_vendor_request to the global functions
__glXReadPixelReply, __glXReadReply, __glXSetupSingleRequest, and
__glXSetupVendorRequest. This will make it easier to add handcoded Single /
VendorPrivate / VendorPrivteWithReply functions.
Aapo Tahkola [Thu, 24 Feb 2005 13:59:56 +0000 (13:59 +0000)]
Another nasty struct size bug found.
Brian Paul [Thu, 24 Feb 2005 05:47:06 +0000 (05:47 +0000)]
More GL_EXT_framebuffer_object: rename some things, added device driver hooks.
Felix Kuehling [Wed, 23 Feb 2005 20:51:27 +0000 (20:51 +0000)]
Don't forget to update a heap's timestamp when a texture image is
changed. Other drivers don't need to do this because they're swapping
modified textures out of texture memory, which implies a timestamp
update.
Ian Romanick [Wed, 23 Feb 2005 20:06:52 +0000 (20:06 +0000)]
Missed the commit of glxext.c yesterday.
Fixed a flow control problem in glGet*v that could result in the display
not being unlocked. This also resulted in refactoring a lot more code
out of the glGet*v routines into get_array_data, which was renamed to
get_client_data.
Ian Romanick [Wed, 23 Feb 2005 19:24:30 +0000 (19:24 +0000)]
Put quotes around $(CC) and $(CXX) so that 'CC=cache gcc' will work.
Brian Paul [Wed, 23 Feb 2005 16:54:18 +0000 (16:54 +0000)]
minor changes to aid debugging, and a new comment
Felix Kuehling [Wed, 23 Feb 2005 16:37:27 +0000 (16:37 +0000)]
With Savage DRM version 2.3.x or later use event counters for texture
heap aging, similar to the way it's done in the i810 and i855 drivers.
This avoids idling the engine on every texture upload.
Brian Paul [Wed, 23 Feb 2005 16:36:17 +0000 (16:36 +0000)]
Sort the enums in the Python code, instead of at runtime. (Zack Rusin)
Ian Romanick [Wed, 23 Feb 2005 00:47:11 +0000 (00:47 +0000)]
Corrected the value for X_GLrop_SecondaryColor3svEXT.
Ian Romanick [Wed, 23 Feb 2005 00:24:18 +0000 (00:24 +0000)]
Missed the Makefile with the previous commit.
Ian Romanick [Tue, 22 Feb 2005 22:36:31 +0000 (22:36 +0000)]
Added __glExtensionBiIsEnabled and __GLXcontext::gl_extension_bits. This
enables libGL to query which extension are exported to applications.
Refactored array-query functionality (from glGet*v) in src/glx/x11/single2.c.
Massive re-write of indirect vertex array support. The most noticable
effect is that glDrawElements now generates DrawArrays protocol. The
side-effects (and the main reasons for the re-work) are that it is much
easier to add support for new arrays (e.g., GL_VERTEX_ATTRIB_ARRAY,
GL_WEIGHT_ARRAY_ARB, etc.) and it is much easier to add support for the new
DrawArrays protocol (required to support ARB_vertex_buffer_object).
These changes were primarilly tested with progs/demos/isosurf.
Brian Paul [Tue, 22 Feb 2005 15:42:30 +0000 (15:42 +0000)]
now generated with Python script
Brian Paul [Tue, 22 Feb 2005 15:39:46 +0000 (15:39 +0000)]
generate enums.c file (Zack Rusin)
Aapo Tahkola [Tue, 22 Feb 2005 05:16:42 +0000 (05:16 +0000)]
New elt buffer code should be fairly stable.
Vladimir Dergachev [Mon, 21 Feb 2005 06:09:30 +0000 (06:09 +0000)]
Add code to autogenerate vertex shaders.
This is needed for multitexturing to work properly.
Dave Airlie [Mon, 21 Feb 2005 00:20:19 +0000 (00:20 +0000)]
Add DESTDIR install target
Dave Airlie [Sun, 20 Feb 2005 10:46:34 +0000 (10:46 +0000)]
add more functionality to miniglx for glitz
Aapo Tahkola [Sun, 20 Feb 2005 03:01:09 +0000 (03:01 +0000)]
Adding support for front buffer mode, removing some code we dont really need and some new debugging stuff.
Aapo Tahkola [Sat, 19 Feb 2005 16:30:56 +0000 (16:30 +0000)]
Minor fix to border and blend color.
Aapo Tahkola [Sat, 19 Feb 2005 03:04:08 +0000 (03:04 +0000)]
Disabling some fallbacks as they cause misc programs not to start and adding some comments.
Aapo Tahkola [Sat, 19 Feb 2005 02:25:53 +0000 (02:25 +0000)]
Fix for buffer overrun caused by ALLOC_STATE not having args surrounded by parenthesis. Can you see it? HINT: Anything texture related should now work slightly better. And yes it took me several hours to find it.
Nicolai Haehnle [Fri, 18 Feb 2005 22:07:20 +0000 (22:07 +0000)]
Print an error if idling the engine before the buffer copy fails.
Be a bit more useful about the sync message after flushing command buffers.
Add an "allmsg" debug name that enables all log messages but does not
enable syncing.
Michel Dänzer [Fri, 18 Feb 2005 20:03:33 +0000 (20:03 +0000)]
fd.o bug #2576: Add support for ATI RN50/ES1000. (ATI Technologies Inc.)
Nicolai Haehnle [Fri, 18 Feb 2005 18:57:38 +0000 (18:57 +0000)]
r300RefillCurrentDmaRegion: Retry buffer acquisition only once, after
waiting for the engine to idle. There's no way for another buffer to
become free anyway once the engine is idle.
Nicolai Haehnle [Fri, 18 Feb 2005 18:28:52 +0000 (18:28 +0000)]
Always submit command buffers, even when there are no cliprects,
so that we no longer leak DMA buffers (plus, this just might fix some
state-setting related problems, if there were any - but that's unlikely).
Update the DRM to cope with cmdbuf->nbox == 0.
Aapo Tahkola [Fri, 18 Feb 2005 16:58:19 +0000 (16:58 +0000)]
We dont need to call r300EmitAOS if we arent going to draw anything(r300_get_num_verts returns 0).
Nicolai Haehnle [Fri, 18 Feb 2005 16:34:14 +0000 (16:34 +0000)]
Reset nr_released_bufs to 0 after the command buffer has been submitted.
Added a more verbose comment about nr_released_bufs in r300_context.h
Brian Paul [Fri, 18 Feb 2005 14:49:27 +0000 (14:49 +0000)]
XMesaGetDepthBuffer() returned wrong bytesPerValue (Karl Schultz)
Brian Paul [Fri, 18 Feb 2005 14:44:00 +0000 (14:44 +0000)]
s/0/NULL/ (Jeff Muizelaar)
Nicolai Haehnle [Fri, 18 Feb 2005 11:27:24 +0000 (11:27 +0000)]
Fix: RADEON_DEBUG is a bitfield, so use & to check for flags, not ==
Nicolai Haehnle [Thu, 17 Feb 2005 22:07:32 +0000 (22:07 +0000)]
Fixing *lots* of warning messages, especially concerning unportable
whitespace before preprocessor commands.
Please, can you try to keep the warnings down? Try running make with
make -s sometime to see just how bad an offender the current code is.
Nicolai Haehnle [Thu, 17 Feb 2005 21:45:20 +0000 (21:45 +0000)]
RADEON_DEBUG is a bitfield, so use & instead of ==.
Also, put the hash in preprocessor directives at the beginning of the line
to fix error messages.
Aapo Tahkola [Thu, 17 Feb 2005 18:16:46 +0000 (18:16 +0000)]
Switching to start_index32_packet.
Aapo Tahkola [Thu, 17 Feb 2005 18:02:28 +0000 (18:02 +0000)]
magic_1 handling...
Ben Skeggs [Thu, 17 Feb 2005 14:34:51 +0000 (14:34 +0000)]
oops.
Brian Paul [Thu, 17 Feb 2005 14:12:50 +0000 (14:12 +0000)]
s/0/NULL/ (Jeff Muizelaar)