389e50a570
more hlsl
2019-12-18 20:12:03 +01:00
da6fc5f35c
TMP more stuff
2019-12-18 20:12:02 +01:00
62015b1499
initial
2019-12-18 20:12:02 +01:00
18a630bb78
Merge pull request #1800 from LukeUsher/half-life-2-hack
...
Fix when titles attempt to find the XBEH header in memory
2019-12-17 22:28:34 +01:00
be42ab7a0f
Nasty hack to fix hl2 crash during switch to ingame
2019-12-17 19:45:30 +00:00
8ebeeeb24c
Merge pull request #1799 from LukeUsher/fix-mechassault
...
kernel: Define and init IdexChannelObject
2019-12-16 23:47:29 +01:00
b62333551c
kernel: Define and init IdexChannelObject
2019-12-16 21:53:02 +00:00
d0da11cb1e
Merge pull request #1796 from LukeUsher/eeprom-serial-number
...
Eeprom Config: Correctly render Serial Number as Numeric Text, not HEX
2019-12-14 13:20:23 -06:00
23540f01b4
Eeprom Config: Correctly render Serial Number as Numeric Text, not HEX
2019-12-14 19:15:58 +00:00
7e46adb95e
Merge pull request #1797 from LukeUsher/minor-nic-fix
...
nvnet: Accept only broadcast and targeted traffic, not multicast
2019-12-14 13:13:19 -06:00
d1018cb773
nvnet: Accept only broadcast and targeted traffic, not multicast
...
Fixes XNET warnings when using Debug XDK libraries
2019-12-14 11:25:35 +00:00
63bc65ece2
Merge pull request #1791 from PatrickvL/hash_palete_in_texture_key
...
Speed up paletized texture
2019-12-02 08:04:41 +00:00
b76f4c5d5a
Fixed black texture regression in Bloodrayne (and possibly all other reported cases); Turned out to be an oversight (GetHostBaseTexture 2nd argument is not the TextureStage, but D3DUsage, dug!). I had to re-do each change line by line, testing as I went, to find this... Ah well, at least it didn't have anything to do with the new cache!
2019-12-01 17:16:57 +01:00
ce191a66e0
Fixed incorrect assumption : Xbox doesn't have a D3DCOMMON_VIDEOMEMORY flag at all, so we can't identify resources with it either!
2019-12-01 15:41:34 +01:00
8664e6cc3d
Potential fix for the "CreateCubeTexture Failed!" regression, by restoring GetHostBaseTexture to it's previous behaviour (if this works, I'll create a separate issue with all relevant detail so that we can follow-up on this later)
2019-11-28 17:00:50 +01:00
01bd2d026e
Lower impact of this PR by storing a few things that won't hurt much
2019-11-28 16:27:39 +01:00
de1b815542
Split up resource cache in two, one for paletized textures (which gets pruned when it overflows),
...
one for all other resources (which cache doesn't prune, like it did not do before, too).
This way, only paletized textures get pruned from their specialized cache, while other resources,
including render-targets (which can't be paletized) will stay resident in the general cache (which won't overflow).
Doing it this way, probably fixes the reported black texture regressions,
but it might also lead to less of a performance increase than before.
2019-11-28 15:56:31 +01:00
b03551535c
Changed our resource cache key into a struct, to greatly reduce the chance of hash collisions.
...
This can solve part of the reported regressions (but there's more, like not-pruneing render targets).
2019-11-26 15:22:43 +01:00
e281d1e490
Renamed EmuD3DActiveTexture into g_pXbox_SetTexture
2019-11-25 15:55:09 +01:00
db45305a5f
Call PruneResourceCache() once per draw (in CxbxUpdateNativeD3DResources instead of in EmuVerifyResourceIsRegistered), this might prevents black textures.
2019-11-25 15:55:08 +01:00
d09381c5a3
Prepare for setting surfaces as active SetTexture's (disabled for now).
...
Consequences of this change :
GetHostBaseTexture won't be called on X_D3DCOMMON_TYPE_SURFACE anymore.
GetHostTexture became obsolete because of this,
Also corected a few typo's & made bits of code simpler.
2019-11-25 15:55:08 +01:00
1b2b320b36
Avoid scaling factor rounding issues, by always applying g_RenderScaleFactor as a multiplier (never as a divisor).
2019-11-25 15:55:08 +01:00
1332ba3ef0
Revert unintentional change
2019-11-25 15:55:07 +01:00
f5bc947964
Minor code cleanups
2019-11-25 15:55:07 +01:00
3be6429300
Cater for actual palette size
2019-11-25 15:55:06 +01:00
5d3cae890c
Include a palette hash in the texture lookup key, instead of patching Palette_Lock functions and freeing textures when setting palette's.
...
To avoid overflowing the resource cache, PruneResourceCache now applies cache eviction as well.
2019-11-25 15:55:06 +01:00
67068303de
Merge pull request #1789 from PatrickvL/correct_d3d_enums
...
Corrected old ordering of D3DTOP enum elements.
2019-11-25 15:54:10 +01:00
fb00f60c27
Merge pull request #1785 from NZJenkins/log_handling
...
Implement extra log component calculations
2019-11-24 09:01:13 +01:00
100be789e2
Implement extra log component calculations
2019-11-22 19:18:29 +13:00
8f6bd2024c
Merge pull request #1767 from x1nixmzeng/cmake-debugger-fix
...
Fixed CMake designer settings for debugger gui
2019-11-21 14:42:32 -06:00
6ba97108c6
Corrected old ordering of D3DTOP enum elements.
...
This fixes alpha texture blending on titles that use D3D XDK 3948 and older
2019-11-21 11:43:12 +01:00
068cefe444
Merge pull request #1772 from LukeUsher/nic-debugging
...
Fix IP network traffic
2019-11-20 17:09:20 +01:00
aefb88eaf1
Merge pull request #1776 from NZJenkins/expp_handling
...
Update expp instruction handling
2019-11-20 17:03:29 +01:00
7cac893aa6
Merge branch 'develop' into expp_handling
2019-11-20 17:02:55 +01:00
8be7402227
Merge pull request #1783 from PatrickvL/impl
...
Introduce two CxbxImpl_*'s for as-yet unimplemented functions.
2019-11-20 15:56:40 +00:00
81bf7e06c7
Merge pull request #1771 from PatrickvL/correct_render_state_info
...
Corrections in render state information
2019-11-20 15:54:10 +00:00
8aef04f83a
Comment-only change, no function impact;
...
Re-orded render state verification comments incrementally, like this:
"Verified [absent in X, ]present in Y[ , removed in Z]" (whereby X < Y < Z).
This makes these comments easier to read and understand.
2019-11-20 16:38:29 +01:00
db90d6a8b8
Verified all render states; Updated D3DRS_MULTISAMPLEMODE to 4039 (might even be 4034)
2019-11-19 17:40:50 +01:00
ae05bdba58
Correct D3DRS_MULTISAMPLERENDERTARGETMODE from 4039 to 4034.
...
This has no functional impact anyway, since there's no host renderstate associated with this (and later) render states.
2019-11-19 15:52:34 +01:00
ae3fbc67c7
Rename g_SetStreamSources into g_Xbox_SetStreamSource (which better reflects it's contents)
2019-11-18 17:57:02 +01:00
ffa5feb709
Merge pull request #1781 from PatrickvL/VersionVariable
...
Turn Cxbx version define into a variable
2019-11-18 09:40:04 -06:00
ca1d2b2722
Reinstated version string in initial caption bar.
2019-11-18 14:44:47 +01:00
61b3743567
Addressed review remarks
2019-11-18 14:44:46 +01:00
0aac02c0c0
Fixed comment header
2019-11-18 14:44:46 +01:00
9cf3591393
Turn Cxbx version define into a variable
...
This is intended to reduce the number of files that need to be
recompiled when updating a repository.
2019-11-18 14:44:44 +01:00
fd7c9ccb47
Merge pull request #1782 from PatrickvL/even_more_cleanup
...
Some more cleanup
2019-11-18 13:20:07 +00:00
5814eaa7fd
Do the extended availability check of render states everywhere it might have an impact.
2019-11-18 12:06:48 +01:00
d5a5e9fc9a
Disable X_D3DRS_MULTISAMPLETYPE in a simpler way, which also allows for easier disabling of other render states as well (would the need ever arise).
2019-11-17 16:17:10 +01:00
2c46324320
Replaced two literal numbers by their corresponding symbols (no functional change, just an improvement that was due)
2019-11-16 22:32:49 +01:00
477ab534c5
This commit solved the fact that X_D3DRS_MULTISAMPLEMODE replaces the older X_D3DRS_MULTISAMPLETYPE (which was available below 4039).
...
This allows restoring X_D3DRS_MULTISAMPLEMODE back to 4361 where it belongs (as far as we know).
Adjusted X_D3DRS_MULTISAMPLERENDERTARGETMODE to XDK 4039 (was 4242)
Made comments on DxbxRenderStateInfo and _X_D3DRENDERSTATETYPE consistent again.
Also use X_D3DRS_* FIRST / LAST defines wherever applicable.
This is a subset of work done earlier in 67cfc36fca
2019-11-16 21:48:15 +01:00
Anthony Miles