497386dc5c
wined3d: Remove the unused device type parameter from the fragment pipe's get_caps() function.
...
On the subject of actually supporting software devices, we should probably
implement those by creating a software or reference adapter with appropriate
gl_info and pipe implementations and creating a device based on that.
2010-02-04 13:39:26 +01:00
42c83f0260
wined3d: Dump ARB program shaders on compile failure.
...
Similar to the patch for GLSL.
2010-02-02 14:10:40 +01:00
705f96fa39
wined3d: Get rid of some unnecessary casts to specific shader types.
2010-01-26 12:08:13 +01:00
46cefd312c
wined3d: Recognize the SM4 cut opcode.
2010-01-07 09:26:54 +01:00
3687677427
wined3d: Recognize the SM4 lt opcode.
2010-01-07 09:26:30 +01:00
6d5aca61f3
wined3d: Recognize the SM4 iadd opcode.
2010-01-07 09:26:25 +01:00
22939f9d29
wined3d: Recognize the SM4 emit opcode.
2010-01-07 09:26:20 +01:00
c57c76b78e
wined3d: Recognize the SM4 ige opcode.
2010-01-07 09:26:08 +01:00
b548e387ae
wined3d: Move shader input/output signatures to IWineD3DBaseShaderClass.
2009-12-31 12:12:16 +01:00
b3defd4c80
wined3d: Rename "wineD3DDevice" to "device".
2009-12-10 11:54:47 +01:00
d2ad2df637
wined3d: Remove a few redundant assignments.
2009-12-09 13:50:48 +01:00
060934c557
wined3d: Don't define nameless structs or unions.
...
Fixes compilation on systems that don't support nameless unions or structs.
2009-12-07 14:18:25 +01:00
0e77a381ea
wined3d: Create clip coords only if they are used.
...
If GL vertex shader clipping is supported (GLSL or NVvp), we currently
always calculate clipping results in the shader. They're gracefully
discarded by opengl, but we need 6 constants / uniforms to calculate
unused values.
2009-11-12 15:18:00 +01:00
111307dfba
wined3d: Get rid of the GL_SUPPORT macro.
2009-10-29 13:09:33 +01:00
48e03bc189
wined3d: Introduce a separate structure for various OpenGL implementation limits.
2009-10-29 13:09:27 +01:00
eeb54b9922
wined3d: Keep better track of where we're using wined3d contexts.
...
The idea here is that we can restore the thread's current GL context on
context_release() if it doesn't correspond to the current wined3d context on
context_acquire().
2009-10-28 11:15:19 +01:00
0c72464abf
wined3d: Use the native constant limit for fitting private constants.
2009-10-26 11:42:36 +01:00
9ee6508132
wined3d: Remove dead code.
2009-10-26 11:42:29 +01:00
f15fc8f41c
wined3d: Take native constants into account when reporting ARB limits.
2009-10-26 11:42:20 +01:00
8d2512317f
wined3d: Free the blitter private data in arbfp_blit_free() (Valgrind).
2009-10-22 10:46:15 +02:00
6be41fd59a
wined3d: Get rid of GL_LIMITS.
2009-10-22 10:46:05 +02:00
55047f811a
wined3d: Fix some GLINFO_LOCATION abuse.
2009-10-22 10:45:57 +02:00
122ed40c2b
wined3d: Use ARB constant limits in ARB.
2009-10-20 14:46:16 +02:00
2d802af4d9
wined3d: Make shader specific constants accessible via GL_LIMITS.
2009-10-20 14:46:08 +02:00
8a68f418db
wined3d: Don't reserve clipping constants needlessly.
2009-10-08 12:03:50 +02:00
ed2860f789
wined3d: Split the clip_parameters[2] array.
2009-10-08 12:03:42 +02:00
7f6ae94c0a
wined3d: Don't write to result.color in arb_add_sRGB_correction.
...
add_sRGB_correction and the generate_pshader function generate the same MOV.
There's no need to do it in two places.
2009-10-08 12:03:30 +02:00
dc45657f2c
wined3d: Fix vs_2_0 sgn.
...
vs_2_0's sgn has two extra parameters - temporary registers for the
instruction. This patch fixes the crash of the d3d9:visual test on windows and
properly handles the vs_2_0 version of sgn.
2009-10-02 12:36:59 +02:00
404aecb8da
wined3d: Test if result.clip[] works correctly.
...
The Mac Nvidia driver doesn't implement the result.clip part of
GL_NV_vertex_program2_option correctly. It complains about a syntax
error if "result.clip[0]" is used. "result.clip" compiles correctly, but
supports only one clipplane, and running the shader results in severe
rendering bugs.
2009-09-28 17:33:07 +02:00
f88c6e5ce5
wined3d: Move common shader fields to IWineD3DBaseShaderClass.
2009-09-25 15:14:35 +02:00
9880cd7504
wined3d: Make some variables static.
2009-09-15 14:32:36 -05:00
f8a3a5fd49
wined3d: Check the correct debug channel in shader_arb_load_constantsF().
2009-09-08 13:06:07 +02:00
f928d6ec87
wined3d: Fix shader_arb_load_constantsF().
2009-09-08 13:06:01 +02:00
d6547c535b
wined3d: Only generate the clipplane emulation KIL if a clipplane is used.
...
The KIL is quite expensive because it forces drivers to disable early Z
discard. It is cheaper to generate and switch between two shaders.
2009-09-08 13:03:03 +02:00
4cdb1c9211
wined3d: Print a warning when an ARB program exceeds the native resource limits.
2009-09-02 11:25:58 +02:00
055906e2a9
wined3d: Properly check for ARB program compile errors.
2009-09-02 11:25:58 +02:00
560d63548d
wined3d: Remove trailing spaces.
2009-08-27 11:45:37 +02:00
3cd91a451d
wined3d: Use 50 chars for register names in texm3x3* functions.
2009-08-19 17:28:08 +02:00
d900491310
wined3d: Replace find_clip_texcoord() with a generic function for finding a free input register.
2009-08-17 13:57:50 +02:00
f7918b978f
wined3d: Make the "luminanceparams" shader_reg_maps member a bitmap.
2009-08-12 13:22:50 +02:00
95bb4c0486
wined3d: Make the "bumpmat" shader_reg_maps member a bitmap.
2009-08-12 13:22:44 +02:00
270f57e755
wined3d: Make the "address" shader_reg_maps member a bitmap.
2009-08-12 13:22:30 +02:00
6fa9fa15da
wined3d: Make the "temporary" shader_reg_maps member a bitmap.
2009-08-11 12:36:20 +02:00
50853e295b
wined3d: Make the "texcoord" shader_reg_maps member a bitmap.
2009-08-11 12:36:16 +02:00
e74efb5487
Assorted spelling fixes.
2009-08-11 12:35:38 +02:00
8d746c31ba
wined3d: Properly handle negative loop step in shader_glsl_loop().
2009-08-10 13:52:37 +02:00
14eedc5be7
wined3d: Watch out about higher constants when clamping ps 1.x consts.
...
ps 1.x supports only 8 constants, so the shader load code dirtifies the
first 8 on a shader switch. However, the constant load code reloaded all dirty
consts with clamping. That potentially left constants > 8 clamped.
2009-08-07 14:56:33 +02:00
3bf0ad4528
wined3d: Explicitly pass the context to the shader_select() handler.
2009-08-07 12:11:06 +02:00
4c1b3de936
wined3d: Call the shader_load_np2fixup_constants() handler directly from inside the shader backend.
2009-08-06 17:26:30 +02:00
03686cb6bd
wined3d: Explicitly pass the context to the shader_load_constants() handlers.
2009-08-06 17:26:30 +02:00
a7251f0f4e
wined3d: Track render_offscreen in the context.
...
This makes sure the relevant states in FindContext() are actually marked dirty
when needed.
2009-08-06 17:26:30 +02:00
2e9362d0fb
wined3d: ARB clipplane init needs the helper constant.
2009-08-05 11:19:14 +02:00
da1e5573da
wined3d: Rename WineD3DContext to struct wined3d_context.
2009-08-03 15:01:48 +02:00
be536ebe2f
wined3d: Get rid of activeContext.
2009-07-24 13:10:57 +02:00
ee4d18cc83
wined3d: Pass NULL to ActivateContext() when we don't need a specific target.
...
Once we keep track of the wined3d context for each thread,
lastActiveRenderTarget won't do what we want here.
2009-07-21 14:22:44 +02:00
cd7f43b00d
wined3d: Don't activate a context unless we need one.
2009-07-20 12:23:07 +02:00
43e6686a78
wined3d: Rename _WineD3D_GL_Info to struct wined3d_gl_info.
2009-07-17 11:22:07 +02:00
9d14dcab68
wined3d: Use the unmodified source in MOVA.
...
The code removed by this patch was a leftover from the time when we tried to
emulate MOVA with a plain ARL, which only supports one source in plain ARB.
This breaks the more sophisticated MOVA and 4 component ADDR register we have
now however.
2009-07-16 11:15:45 +02:00
387d337a70
wined3d: EXP and EXPP are scalar operations.
2009-07-15 13:17:02 +02:00
2bfdff198c
wined3d: Check the correct constant limit.
2009-07-10 13:07:33 +02:00
483d552e21
wined3d: Allow shader_buffer_init() to fail.
2009-07-09 11:38:02 +02:00
40b411958b
wined3d: Rename "SHADER_BUFFER" to something sane.
2009-07-09 11:37:58 +02:00
32fd8f2130
wined3d: checkGLcall() doesn't need a \n.
2009-07-07 11:50:22 +02:00
2ac34bf231
wined3d: Add some missing float suffixes.
...
The compiler should be smart enough to fix most of these, but it looks sloppy.
2009-07-07 11:50:05 +02:00
0caebe5f51
wined3d: Force NP2 constant reload in shader_arb_select.
2009-07-02 11:56:57 +02:00
ef280280c2
wined3d: Implement shader_arb_load_np2fixup_constants.
2009-07-02 11:56:51 +02:00
e29babbf6d
wined3d: Add NP2 fixup code to shader_hw_sample (ARB).
2009-07-02 11:56:45 +02:00
034fa4268c
wined3d: Add NP2 fixup code to shader_arb_generate_pshader.
2009-07-02 11:56:39 +02:00
ec30e9a85e
wined3d: Add arb_ps_np2fixup_info structure and improve structure packing for arb_ps_compiled_shader.
2009-07-02 11:56:33 +02:00
c82da7d975
wined3d: Don't dirtify too many shader constants.
...
If we have start = 4, count = 3, the highest dirty constant is the one with index 6. start + count gives 7,
so it already includes the zero-based array correction. Don't add an additional 1.
Additionally to the inefficiency of looking at one extra constant, this causes problems if the driver
rejects loading 257 constants on the initial load. In this case no constant is loaded if
GL_EXT_gpu_program_parameters is used.
2009-07-01 13:14:34 +02:00
40bc47901c
wined3d: Use flags for driver quirks.
2009-07-01 11:41:44 +02:00
d4312ebb16
wined3d: Add the vertex shader footer in the main function.
2009-06-29 12:28:40 +02:00
5befed21f6
wined3d: Write the vshader footer in a separate function.
2009-06-29 12:28:31 +02:00
5859d87754
wined3d: Implement function calls with NV exts.
2009-06-29 12:28:21 +02:00
031401a070
wined3d: ARB*p is already on after a depth blit.
2009-06-29 12:28:13 +02:00
9d63585352
wined3d: Don't disable ARBfp if the replacement pipeline is used.
...
If the replacement pipeline is used, ARBfp is always on. Disabling it
can break shaders or the replacement pipeline, because the shader and
ffp code assumes the extension is on.
2009-06-29 12:28:06 +02:00
7363b6e2e3
wined3d: Document GL context dependencies.
2009-06-26 12:05:07 +02:00
d1cd08b7b3
wined3d: Honor WINED3DSPSM_NOT in ARB.
2009-06-26 12:02:29 +02:00
8fb1e2acfb
wined3d: Jump to the else branch if cond is false, not endif.
2009-06-26 12:02:18 +02:00
a07b50c94c
wined3d: Support writemasks on texkill in ARB.
2009-06-26 12:02:12 +02:00
3a327908de
wined3d: Initialize the used clip planes even if no clip emulation is used.
2009-06-26 12:02:07 +02:00
4d86eac60a
wined3d: Unclamp vertex colors for 3.0 shaders in ARB if needed.
2009-06-26 12:02:01 +02:00
f69fb5c011
wined3d: POW and LOG operate on the absolute value.
2009-06-26 12:01:54 +02:00
bd14b6e2b4
wined3d: IFC requires GL_NV_fragment_program2.
2009-06-26 12:01:44 +02:00
d4854b636c
wined3d: Beware of double negations.
2009-06-26 12:01:36 +02:00
f98c6e3cff
wined3d: Allow the existence of other shader types in get_loop_control_const().
2009-06-24 11:28:50 +02:00
943ed4e1e9
wined3d: Eliminate a redundant local variable in get_loop_control_const().
2009-06-24 11:28:45 +02:00
aec6d319fa
wined3d: Avoid a redundant copy.
2009-06-24 11:28:38 +02:00
fd8ebebe56
wined3d: Use a proper structure for "loop_control".
2009-06-24 11:28:33 +02:00
4caf43ddb7
wined3d: Add some missing GL locking documentation to arb_program_shader.c.
2009-06-24 11:28:27 +02:00
a4f2d3d18e
wined3d: Fix a compiler warning.
2009-06-24 11:25:09 +02:00
e0ec956b3d
wined3d: Fix pixelshader ifc.
...
The NVfp2 control flow uses the D3D syntax, and doesn't need the inversed
branch condition we're using in the vertex shader.
2009-06-23 15:53:00 +02:00
444b7ac7e5
wined3d: Use a local parameter for the position fixup.
...
I don't like that I have to do this because the posFixup is in all
vertex programs, so its at the same position and could be loaded
globally. Unfortunately, there are only 256 env parameters usually,
which makes it impossible for any shader to use c256, even if it does
not use indirect addressing, and so we can't claim 256 constant
support.
2009-06-23 15:53:00 +02:00
7ffd10da28
wined3d: Find the clip texcoord before compiling.
...
The best is to put this together with the rest of the shader_arb_priv initialization.
2009-06-23 15:53:00 +02:00
e64d48371f
wined3d: Add a function to control use of NV_vp2 clipplanes.
...
I find it helpful for debugging to have this controlled at a central place,
without having to disable the entire GL extension or manually find all the
places where GL_SUPPORT(NV_VERTEX_PROGRAM2_OPTION) controls clipplane use. It
is useful for debugging the emulation code on NV cards and for debugging mac
driver issues.
2009-06-23 15:53:00 +02:00
85af0b2943
wined3d: Don't emulate clipplanes with ffp vp and fix a wrong if condition.
...
b2f09fd204
2009-06-23 15:53:00 +02:00
114b5bdefb
wined3d: Fix a comment.
...
This was actually broken by 122af07a30
2009-06-23 11:48:10 +02:00
569b448e13
wined3d: Fix logical condition in find_clip_texcoord().
2009-06-22 13:25:27 +02:00
3845b37a21
wined3d: Advertise SM 3.0 in ARB if the extensions are available.
2009-06-22 11:27:11 +02:00
e211a0b4a4
wined3d: Enable Shader Model 2.0 in ARB.
...
If the needed constants are available, we can support all vs_2_0 and ps_2_0
requirements with the plain ARB extensions. We cannot however, run SM 2.0a or
SM 2.0b.
2009-06-22 11:27:05 +02:00
e1c6b73e2d
wined3d: Make the ARB index offset work with emulated MOVA.
2009-06-22 11:26:57 +02:00
77820d5ea5
wined3d: XXXC CC doesn't work even with NV_FP2 on.
2009-06-22 11:26:50 +02:00
26d17fe04c
wined3d: Implement MRTs in ARB.
2009-06-22 11:26:43 +02:00
bb8bcf5878
wined3d: Reload the first 8 constants on a 1.x and != 1.x ps switch.
...
ps 1.x constants are clamped to [-1;1], constants in >= 2.0 pshaders
are not. This means we have to reload constants when switching between
those shader types in ARB. In GLSL this is not a concern because
constants are tied to program objects and are reloaded on a shader
change anyway.
2009-06-22 11:26:33 +02:00
2364ce08e5
d3d9: Add a SGN test.
2009-06-22 11:26:14 +02:00
b2f09fd204
wined3d: Emulate clipplanes in ARB if the NV extensions are not available.
...
This patch tries to find a free texture coordinate to load up to 4 clip
coordinates into the pixel shader, and uses KIL to throw away fragments
that are cut by a clipplane. If no free texture coordinate is found,
clipping is not done. If more than 4 clipplanes are used, only the first
4 are actually enabled. That should be pretty rare though.
2009-06-19 14:54:00 +02:00
7baf7c6518
wined3d: Implement vertex texture fetch in ARB.
2009-06-19 14:53:50 +02:00
f1e97cdfbc
wined3d: Don't run out of varyings because of clipplanes.
2009-06-19 14:53:43 +02:00
b53b5909ff
wined3d: Handle SM 3.0 varyings in ARB.
2009-06-19 14:53:20 +02:00
3aaa530ca9
wined3d: Properly load the y correction.
2009-06-18 15:17:26 +02:00
c9ef1ed28c
wined3d: Deal with instructions that do not support _SAT.
2009-06-17 12:24:43 +02:00
a7b605af64
wined3d: Implement D3DSPDM_PARTIALPRECISION support.
2009-06-17 12:24:35 +02:00
dfd338de77
wined3d: Add modifier support to tex and friends.
2009-06-17 12:20:54 +02:00
7856c1c26f
wined3d: Fix texm3x3pad.
...
The register number is already in the name we got from get_register_name,
don't add it twice.
2009-06-17 12:20:48 +02:00
58c3a0c832
wined3d: Implement vpos and vface.
2009-06-17 12:19:31 +02:00
ae600fe0c8
wined3d: Implement texldl in ARB.
2009-06-17 12:19:15 +02:00
4830a29101
wined3d: Implement texldd in ARB.
2009-06-16 13:57:34 +02:00
2f6a713805
wined3d: Use flags instead of bools to control shader_hw_sample.
2009-06-16 13:57:30 +02:00
e608ef856e
wined3d: Implement IFC in ARB.
2009-06-16 13:57:20 +02:00
dfefbcd00d
wined3d: Implement break and breakc.
2009-06-16 13:57:15 +02:00
c220bafa2c
wined3d: Unroll loops in ARB if needed.
...
This is needed to allow us to claim vs_2_0 support with plain
GL_ARB_vertex_program. vs_2_0 allows loops without breaks.
2009-06-16 13:57:04 +02:00
258d99d289
wined3d: Implement dsy in ARB.
2009-06-10 11:52:15 +02:00
51cdc9d361
wined3d: Implement loops with the NV extensions.
2009-06-10 11:52:01 +02:00
298bd3c50d
wined3d: Store the bumpmap ARB constants in ARB structures.
2009-06-10 11:51:49 +02:00
a44f84fdf7
wined3d: Add a proper enum value for "no extension".
...
This would allow an extension check for ARB_COLOR_BUFFER_FLOAT to work.
2009-06-04 12:19:02 +02:00
ad6279de9f
wined3d: Replace the wined3d hash table with the generic red-black tree.
2009-06-03 11:40:14 +02:00
2afcb794f2
wined3d: Implement clipplanes in the ARB backend.
...
Using GL_NV_vertex_program2_option so far. If we're really desparate we can
handle some cases without the extension by using a custom varying and texkill
in the fragment program.
2009-06-02 12:09:54 +02:00
dfc9141458
wined3d: Use condition code for sRGB write correction if available.
...
This gives a small performance improvement. Don't enable NVfp for it though,
because the NVfp penalty is bigger than the gain from this patch. But if NVfp
is enabled anyway, make use of it.
2009-06-02 12:09:38 +02:00
7e1100d0e4
wined3d: Handle LRP in vertex shaders.
...
The ARB "LRP" instruction is a fragment program only instruction. It is
however valid in vs_2_0+ vertex shaders, so we have to emulate it.
2009-06-02 12:09:26 +02:00
38076e08ca
wined3d: Add a comment about texbem and X2D.
2009-06-01 16:16:00 +02:00
de12f880ce
wined3d: Don't enable the NV frag extensions if we don't need them.
...
Enabling the NV extensions occupies a temp register for some reason. Avoid
needlessly enabling it.
2009-06-01 16:16:00 +02:00
5a07bacd36
wined3d: Write result.color in one mov.
...
Otherwise the Nvidia assembler needs one temp reg.
2009-06-01 16:16:00 +02:00
050735be79
wined3d: Revert "WineD3D: use CMP instead of SLT and SGE in sRGB correction".
...
This reverts patch ba35760f9f
2009-06-01 16:16:00 +02:00
8665ee5906
wined3d: Use Rx registers for sRGB correction if possible.
...
This removes another reason to declare TA, TB and TC.
2009-06-01 16:16:00 +02:00
7168b62323
wined3d: Use NRM from GL_NV_fragment_program2 if available.
2009-05-29 11:27:51 +02:00
d8e219be75
wined3d: Avoid the TMP_COLOR mov in some cases.
...
Many 2.0 and 3.0 shaders end with a "mov oC0, rx". If sRGB writing is enabled,
the ARB backend writes to a TMP_COLOR temporary, and at the end of the shader
writes the sRGB corrected color to result.color. If oC0 is not partially
rewritten after the mov, we can ignore the mov, not declare TMP_COLOR at all,
and just use the rx register as input for the sRGB correction code. This saves
a temporary and an instruction.
2009-05-29 11:27:44 +02:00
da7176be2a
wined3d: Use DP2A or X2D for dp2add if available.
2009-05-29 11:27:29 +02:00
e2be741bc0
wined3d: Remove a MESA hack.
2009-05-29 11:27:23 +02:00
5f723bc44a
wined3d: Allocate the proper size for the lconst map.
2009-05-29 11:27:14 +02:00
d37472f27b
wined3d: Add missing ActivateContext calls to shader_destroy() implementations.
2009-05-29 11:26:32 +02:00
e70c80add9
wined3d: Emulate if(bool) in ARB shaders.
2009-05-28 12:46:01 +02:00
e492dd858e
wined3d: Add a function around the handler table.
...
This reduces the number of methods in the shader backend(the instr
modifiers can be handled in that wrapper) and it will help flow
control emulation in the ARB backend.
2009-05-28 12:45:45 +02:00
aad92c0780
wined3d: Manage vs_compile_args in the backends.
2009-05-28 12:45:32 +02:00
814dd42f25
wined3d: Manage ps_compiled_shader in the backends.
2009-05-28 12:45:21 +02:00
8763f6eac5
wined3d: Make find_gl_vshader backend specific.
2009-05-28 12:45:10 +02:00
732a9a15d5
wined3d: Remove the forward declaration added in the last patch.
2009-05-28 12:44:59 +02:00
d0d681c81e
wined3d: Make find_gl_pshader backend private.
2009-05-28 12:44:43 +02:00
5865d9c01b
wined3d: Store the compile args in the compile context, not the shader.
2009-05-28 12:44:22 +02:00
6492622350
wined3d: sincos for vertex shaders.
...
SCS is unfortunately a fragment program only instruction. If we have the NV
extensions we can use SIN and COS. Otherwise we have to approximate sine and
cosine with a taylor series. Luckily we're provided with the necessary
constants by the application.
2009-05-27 11:56:53 +02:00
2f3faf4526
wined3d: Put some ARB declarations in the right place.
...
TMP_POS is only used in vertex shaders, declare it in the vshader
specific code. The sRGB constants are only used by pixel shaders, so
move them to the ps specific code, and avoid reading the stateblock.
2009-05-27 11:56:29 +02:00
f484cbbf91
wined3d: Get rid of Tx hardcoding in texbem.
2009-05-27 11:56:07 +02:00
da1b8d0c8e
wined3d: Replace a few hardcoded Tx registers with proper name getters.
2009-05-27 11:56:01 +02:00
7fc4279cb8
wined3d: Implement DSX in ARB.
...
DSY is not that easy because we have to deal with the inverted coordinate
system in onscreen(!) rendering - see GLSL dsy.
2009-05-27 11:55:52 +02:00
122af07a30
wined3d: Implement SGN in ARB.
2009-05-21 16:17:52 +02:00
8e45e48b0c
wined3d: Support vec4 A0 with NV_vertex_program2_option.
2009-05-21 16:17:52 +02:00
d0756ee537
wined3d: Support ABS and ABSNEG with NV extensions if available.
2009-05-21 16:17:52 +02:00
01ec5068e2
wined3d: Add NV asm extension support to the ARB backend.
2009-05-21 16:17:51 +02:00
60f20a2f39
wined3d: Support ABS and ABSNEG in ARB.
2009-05-21 16:17:51 +02:00
7b1d48721d
wined3d: Emulate the 4 component address register in ARB.
2009-05-20 10:55:17 +02:00
72e9e80f6e
wined3d: Prepare for vec4 address registers in ARB.
2009-05-20 10:55:09 +02:00
c7ca3793cc
wined3d: Implement mova rounding in arb.
2009-05-20 10:55:03 +02:00
bffb89101f
wined3d: Fix srgb correction.
2009-05-20 10:54:56 +02:00
e6cc81d287
wined3d: Get rid of the TMP register in fragment shaders.
...
To be able keep the temporary register in the type independent NRM
instruction, the vertex temporary register is renamed to TA to match
the name of a pixel shader register.
2009-05-20 10:54:50 +02:00
ba35760f9f
wined3d: Use CMP instead of SLT and SGE in sRGB correction.
...
This shortens the code and reduces the amount of temp regs used by 1.
2009-05-19 15:24:48 +02:00
ad217029b0
wined3d: Get rid of TMP accesses in texm3x3* instructions.
2009-05-19 15:24:39 +02:00
ced325f816
wined3d: Get rid of TMP access in texm3x2*.
...
texm3x2pad knows which register the following texm3x2depth or tex instruction
will use, and it knows that this register is uninitialized. So use it for
temporary storage instead of TMP.
2009-05-19 15:24:31 +02:00
17b001788f
wined3d: Get rid of TMP2 and some easy TMP register uses.
...
We have 5 temporary registers declared for private use. We can get rid of two
of them.
2009-05-19 15:24:21 +02:00
4bf980962a
wined3d: Only declare TMP_OUT in vertex shaders.
2009-05-19 15:24:13 +02:00
d7f282f241
WineD3D: Support more constants in ARB shaders.
...
GL_LIMIT - 1 for vertex programs and the GL limit in fragment programs. The
indirect addressing limitation in GLSL applies here as well.
2009-05-18 13:59:31 +02:00
c7839112c9
wined3d: Avoid declaring helper_const in vertex programs if possible.
2009-05-18 13:59:24 +02:00
2cb8f42168
wined3d: Support clipplanes with GLSL.
...
This is the Nth attemt to make clipping work with GLSL shaders. The patch now
uses the GLSL quirk table to handle cards that need a custom varying for
gl_ClipPos, and the code is adapted to the changed state table and shader
backend system.
2009-05-18 13:59:16 +02:00
45563979bd
wined3d: Update the bump constants after a shader change.
...
Since we're using local constants now, we have to update the constants after a
shader change.
2009-05-18 13:58:57 +02:00
3372846aa6
wined3d: Use local constants for bump parameters.
...
This simplifies the loading code a bit. The constants were never
designed to be at the same location in all shaders, so there's no
point in using program.env. This way we don't collide with the d3d
shader constants and its easier to work together with NP2 fixups and
other shaders.
2009-05-18 13:58:47 +02:00
6fe1a5daa5
wined3d: Don't use TMP_COLOR unless srgb is enabled.
...
This was needed unconditionally in the past to apply fog, but since we're
using the ARBfp fog defines it is only needed if an sRGB correction is done
at the end of the shader.
2009-05-15 11:45:18 +02:00
b35e469d9d
wined3d: Document functions that depend on the caller to do GL locking.
2009-05-15 10:25:41 +02:00
0123f81ce5
wined3d: Centralize WINED3DSPDM_SATURATE handling.
...
Besides the cleanup this patch will help to implement D3DSPDM_PARTIALPRECISION
with GL_NV_vertex_program2_option / GL_NV_fragment_program_option.
2009-05-14 14:27:33 +02:00
aeddd2da63
wined3d: Fix remaining incorrect dst register getters.
2009-05-14 14:27:25 +02:00
bcb20ae96c
wined3d: Fix a few more hardcoded pixel shader registers.
2009-05-14 14:27:19 +02:00
2327e316a2
wined3d: Tidy up pshader_hw_texreg2rgb and friends.
2009-05-14 14:27:14 +02:00
05cc344e7d
wined3d: Tidy up pshader_hw_dp2add.
2009-05-14 14:27:07 +02:00
ca1a979c69
wined3d: Tidy up pshader_hw_cmp.
2009-05-13 13:22:16 +02:00
c06aa361e0
wined3d: Tidy up pshader_hw_cnd.
2009-05-13 13:22:11 +02:00
90526ce761
wined3d: Separate Tx tempregs from Tx varyings in ARB.
...
ps_1_3 uses Tx to pass in texture coordinates, but also as temporary
registers. ps_1_4 and ps_2_0 only use them for texture coordinates. This patch
gets rid of the Tx = fragment.texcoord[x] assign in all shader versions, and
doesn't even declare Tx in ps_1_4 and ps_2_0.
The <=ps_1_3 instructions know which kind of input they expect from the Tx
register, so the instruction handlers now know if they have to read the
tempreg Tx or the varying fragment.texcoord[x].
2009-05-13 13:22:04 +02:00
fc5e511dda
wined3d: Fix 1.4 pshader projection in ARB.
...
shader_arb_add_src_param handled DW and TXP undid it again. Remove DZ DW from
the modifiers and handle it in the instruction. DZ cannot be handled by TXP as
is, so move the .z component to .w and make it DW-like. Using SZW+TXP is
likely more efficient than the RCP, MUL, TEX we'd get if we let
shader_arb_add_src_param do the job.
2009-05-13 13:21:43 +02:00
45799fdc4e
wined3d: Tidy up pshader_hw_tex.
...
Fix the indentation and replace the get_register_name with
shader_arb_get_dst_param.
2009-05-13 13:21:29 +02:00
59689262e7
wined3d: Add missing GL locking to shader_backend_t.shader_free_private() implementations.
2009-05-13 12:01:48 +02:00
b5aee04023
wined3d: Add missing GL locking to fragment_pipeline.enable_extension() implementations.
2009-05-13 12:01:43 +02:00
fef4253171
wined3d: Pass a struct wined3d_shader_register to shader_arb_get_register_name().
2009-05-12 11:17:45 +02:00
601030cfa9
wined3d: Standardize pshader_hw_texkill.
...
Use shader_arb_get_dst_param instead of get_register_name to find the register
name. Even though this adds support for modifiers(which aren't allowed by
native), this shouldn't hurt. If an app passes in an incorrect shader it
should be caught in the frontend.
2009-05-11 12:50:37 +02:00
5cfbfbc8b3
wined3d: Tidy up shader_hw_rsq_rcp.
2009-05-11 12:50:22 +02:00
da47f68522
wined3d: Simplify shader_hw_map2gl.
...
Most of the extra work done in the pixel shader code was not needed or
redundant. This patch unifies the two paths.
2009-05-11 12:50:01 +02:00
7a1f9c84f1
wined3d: Get rid of the strcat in shader_arb_add_dst_param and make it shader_arb_get_dst_param.
2009-05-11 12:49:54 +02:00
48c2b0d669
wined3d: Merge shader_arb_add_src_param and pshader_gen_input_modifier_line.
...
Mostly based on the code of pshader_gen_input_modifier_line. The space-adding
behavior of shader_arb_add_src_param was removed because the plurality of
instruction handlers passes an uninitialized buffer in and expects a register
name written to its start, and only map2gl and rcp_rsq use the space-adding
stuff. I'll change rcp_rsq in a later patch anyway. I changed the name to
shader_arb_get_src_param to reflect this behavior.
2009-05-11 12:48:35 +02:00
7a1f305d42
wined3d: Get rid of pshader_gen_output_modifier_line.
...
Use shader_arb_add_instruction_modifiers instead. This avoids calling the
fixup function from each single instruction handler to handle shifts. It does
not yet get rid of the modifier handler in each instruction because we don't
want a separate line if we can just append _SAT to the instruction name.
2009-05-08 12:20:37 +02:00
f025829449
wined3d: Pass the instr to pshader_gen_output_modifier_line.
2009-05-08 12:20:25 +02:00
2fd485ae31
winedd: Move shader_*_add_instruction_modifiers into the shader backend.
2009-05-08 12:20:18 +02:00
4daa3e5320
wined3d: Always declare single constants in ARB if rel addr is not used.
...
This is needed to raise the number of advertised constants to the GL
limit. The ARB assembler ususally does not optimize away unused
constants, so we have to do this.
2009-05-08 12:15:26 +02:00
5e473cb17f
wined3d: Group register information into struct wined3d_shader_register.
2009-05-08 12:13:17 +02:00
65622a052c
wined3d: Create a struct wined3d_shader_version to store version information.
...
The version tokens in SM1-3 and SM4 are similar, but not the same.
2009-05-07 14:44:21 +02:00
4c29105cb2
wined3d: Pack hardcoded local constants in ARB.
...
This makes the location of used program.local parameters more
predictable and eases sharing this space with other private constants.
2009-05-06 13:12:45 +02:00
df5d82dd5b
wined3d: Pack ARB srgb constants better.
...
There's no need to waste constants for repeatedly loading the same value.
2009-05-05 14:13:20 +02:00
2a5a6a3439
wined3d: Store the SM 1-3 opcode information tables in the SM 1-3 frontend.
2009-05-05 11:34:28 +02:00
2378108eb9
wined3d: Create a frontend for parsing shaders.
2009-05-04 12:53:08 +02:00
8d4c90496f
wined3d: Store the sampler type in struct shader_reg_maps.
...
Instead of the usage token.
2009-04-29 13:08:01 +02:00
6376941266
wined3d: Avoid using an implicit value to set fogcoord to 0.0.
2009-04-23 18:14:27 +02:00
fc60083ed2
wined3d: Handle the destination modifier shift in the frontend rather than the backend.
2009-04-23 14:22:06 +02:00
fec1fa93cf
wined3d: Handle the swizzle shift in the frontend rather than the backend.
2009-04-22 13:14:14 +02:00
629801bccb
wined3d: Get rid of the token field from struct wined3d_shader_src_param.
2009-04-22 13:14:03 +02:00
7a3d4ce76f
wined3d: Pass a struct wined3d_shader_src_param to shader_arb_get_swizzle().
2009-04-22 13:13:57 +02:00
db5ab97125
wined3d: Store the source swizzle in struct wined3d_shader_src_param.
2009-04-21 14:00:02 +02:00
644439ff0d
wined3d: Pass a struct wined3d_shader_src_param to pshader_gen_input_modifier_line().
2009-04-21 13:56:58 +02:00
ea0b3b42eb
wined3d: Pass a struct wined3d_shader_src_param to shader_arb_add_src_param().
2009-04-20 14:13:33 +02:00
1ae060be4a
wined3d: Simplify replicating the .x swizzle in shader_hw_mov().
...
It becomes even simpler once the swizzle shift and token field are eliminated.
2009-04-20 14:13:28 +02:00
915b91757f
wined3d: Store the destination relative addressing token as a struct wined3d_shader_src_param.
2009-04-20 14:13:08 +02:00
8ac4c98be2
wined3d: Store the source modifiers in struct wined3d_shader_src_param.
2009-04-15 15:57:35 -05:00
b6668b0296
wined3d: Store the register index in struct wined3d_shader_src_param.
2009-04-15 15:57:30 -05:00
ff62cab537
wined3d: Introduce struct wined3d_shader_src_param.
2009-04-15 15:57:19 -05:00
463de24e7b
wined3d: Group instruction context information together in struct wined3d_shader_context.
2009-04-15 15:57:13 -05:00
d3fc7dca99
wined3d: Get rid of the "token" field from struct wined3d_shader_dst_param.
2009-04-14 15:25:07 -05:00
18cdcc0b7e
wined3d: Pass a struct wined3d_shader_dst_param to shader_arb_get_write_mask().
2009-04-14 15:25:01 -05:00
0c2514bd8f
wined3d: Rename some NP2 fixup variables and update comments.
2009-04-13 17:32:11 -05:00
9657e81429
wined3d: Add newline to fixme.
2009-04-13 17:32:03 -05:00
6e970c5d95
wined3d: glsl: Enable load_np2fixup_constants.
2009-04-10 14:00:58 +02:00
9b067a6ca5
wined3d: Add load_np2fixup_constants function to shader backend.
2009-04-10 14:00:50 +02:00
754b5cf2ad
wined3d: Deal with reserved shader constants in the backend.
...
This moves the GLSL and ARB specific reserved constants out of directx.c into
the get_caps methods of the shader backends. That way the number of reserved
constants remains in the backend.
GL_LIMITS({v/p}shader_constantsF) now contains the real number of constants as
advertised by GL instead of some mixture of GL info and backend implementation
specifics. This makes it easier for backends to decide how many constants to
use.
2009-04-10 11:22:55 +02:00
3a5bbe6a0c
wined3d: Split vshader_program_add_param() in shader_arb_add_src_param() and shader_arb_add_dst_param().
2009-04-10 10:23:38 +02:00
e6031afdfa
wined3d: Merge pshader_get_register_name() and part of vshader_program_add_param().
2009-04-10 10:23:33 +02:00
9a28d273bc
wined3d: Explicitly pass the register type and index to pshader_get_register_name().
2009-04-10 10:23:24 +02:00
22e57d025c
wined3d: Store the destination shift in struct wined3d_shader_dst_param.
2009-04-10 10:23:15 +02:00
0f92bbc6bc
wined3d: Store the register type in struct wined3d_shader_dst_param.
2009-04-07 14:21:49 +02:00
6f66c1ddbd
wined3d: Store the write mask in struct wined3d_shader_dst_param.
2009-04-07 14:21:37 +02:00
f0de1622d0
wined3d: Store the destination modifiers in struct wined3d_shader_dst_param.
2009-04-06 12:57:52 +02:00
6660978b44
wined3d: Use register_idx in the ARB shader backend.
2009-04-03 13:01:50 +02:00
7b62165546
wined3d: Store the register index in struct wined3d_shader_dst_param.
2009-04-03 13:01:44 +02:00
7245cd2b81
wined3d: Add a struct wined3d_shader_dst_param for storing destination parameter information.
2009-04-03 13:01:38 +02:00
04b2e0b19b
wined3d: Remove the opcode field from struct wined3d_shader_instruction.
2009-04-03 13:01:30 +02:00
29c9dc989d
wined3d: Use the source and destination parameter count in the ARB shader backend.
2009-04-03 13:01:20 +02:00
7bde2792c2
wined3d: Store the source and destination parameter count in struct wined3d_shader_instruction.
2009-04-02 16:28:29 +02:00
d633b6d47e
wined3d: Avoid comparing against WINED3DSIO_* in the ARB shader backend.
2009-04-02 16:28:29 +02:00
44648b2b91
wined3d: Store the instruction handler index in struct wined3d_shader_instruction.
...
This allows us to avoid comparing against WINED3DSIO_*, which is specific to
the bytecode.
2009-04-02 16:28:29 +02:00
c3a01b315e
wined3d: Use a more reasonable name for "SHADER_OPCODE_ARG".
2009-04-01 14:24:19 +02:00
e44335585e
wined3d: Don't use the opcode_token field in the ARB backend.
2009-04-01 14:24:02 +02:00
3fbdb64270
wined3d: Update a comment.
...
shader_select takes care of compiling the shader since we have the shader
duplication infrastructure. However, there's still a motivation to skip the
shader_select call if the vertex shader is dirty.
2009-03-30 14:21:52 +02:00
703dcca61b
wined3d: Fix some whitespace.
2009-03-30 14:21:44 +02:00
afc5744e70
wined3d: Pass format_desc to set_shader().
...
This should be the last unnecessary getFormatDescEntry() call.
2009-03-24 12:57:45 +01:00
a56e3a102b
wined3d: Only return a pointer to struct GlPixelFormatDesc from getFormatDescEntry().
2009-03-13 11:29:10 +01:00
362bc0d694
wined3d: Don't compare const_num against -1.
2009-03-09 14:39:39 +01:00
5cf764aec2
wined3d: Put the color_correction calls into the instruction handlers.
...
The color correction cannot be done behind the back of the individual
instruction handlers because it might conflict with the instruction's
color modifications and the D3D provided writemask.
2009-02-27 13:14:23 +01:00
6012475e59
wined3d: Properly update last_was_pshader.
2009-02-24 16:57:30 +01:00
828f42cb1f
wined3d: Add some traces to state handlers.
...
Should make it a bit clearer what's happening in some cases.
2009-02-19 12:53:44 +01:00
Henri Verbeet