Commit Graph

147 Commits

Author SHA1 Message Date
Gregory Hainaut dc00af8829 psx|spu2x|gsdx: shift negative value is undefined in C standard
Instead does the negation after the shift

v2: assert shift <= 15 in clamp_mix
2017-04-30 21:20:23 +02:00
Gregory Hainaut 1a768ca438 gsdx sw: sync thread if current target pages are already used by input texture
Otherwise you have a write before read typical race condition. It works
most of the time because textures are stored in temporary buffers (aka
texture cache). So the race condition requires texture invalidation in the mix.

I hope the perf impact will be small enough.

Fix #1691
Blood Will Tell: gray scale effect description

Frame is renderer in 0x700
Sync 0x700 (RT will be used as input)

Foreach page of frame
    // The missing Sync was this one. You can't copy new data to 0x2800
    // until you finish the rendering that use 0x2800 as input texture
    // (AKA end of this foreach loop)
    Sync 0x2800 (not the first iteration, texture will be used as a RT)
    Copy page from 0x700+offset to 0x2800

    Sync 0x2800 (RT will be used as input)
    Render Effect line1 from 0x2800 to 0x700
2017-03-15 18:30:46 +01:00
Gregory Hainaut 6d6ed1a205 gsdx sw: do const q division in ConvertVertexBuffer
It allow to do the division before the size multiplication
It avoid a float overflow if T is too big.
Old behavior: (T * size) / Q
New behavior: (T / Q) * size

Performance Note:
* Rcp was replaced by a slow division (more accurate)
* At least we avoid a 2nd loop on the vertex buffer

It helps on Pro Soccer Club and Galerians Ash rendering

Tric Note:
SPRITE must be handled differently because the 'q' of first vertex could
be invalid
2017-03-10 20:45:14 +01:00
Gregory Hainaut 0a5f796c33 gsdx sw: missing texture due to wrong optimization fix
Typical bug, missing/wrong texture on the SW renderer but working fine on the HW renderer

Debugged on ATV Quad Power Racing 2 but I suspect couple of game are impacted

Bug description:
GSdx flatten the Q value of sprite. So m_vt.m_eq.q is true when Q(2N+1) are the same.
Q(2N) values could be random. The fix replaces Q0 by Q1 for the uniform Q value.
2017-03-05 20:27:19 +01:00
Akash 09c72375ab GSDX: Pass total height of framebuffer on GetOutput()
Some PSX games seem to store image data of the drawing results in an undeterminate area out of range from the current context buffer. At such cases, calculate the height of both the frame memory rectangles combined.

What happens on "Crash bash" -

* At first draw, scissoring is limited to SCAY0- 0 & SCAY1- 255
* At second draw, scissoring is limited to SCAY0- 255 & SCAY0-511

Previously, we limited the height to the value of one single output texture, so instead of that let's calculate the total height of both the two buffers combined to prevent such issues.
2017-01-04 22:17:26 +05:30
Akash 07d7905896 GSDX: Fix output texture height calculation
Previously, the height of the frame offset was also considered for the total height of the texture which was obviously wrong as the portion before the offset value isn't part of the frame memory.
2016-12-08 22:14:05 +01:00
Gregory Hainaut 51c64fcbe6 gsdx sw: trick GetFeedbackOutput to handle basic merge loopback
It is enough for Xenosaga.
2016-10-14 18:13:20 +02:00
Gregory Hainaut 22ebdcf1cf gsdx debug: create a nice dump function for kernel gs registers
Add missing register
2016-10-08 23:42:39 +02:00
Gregory Hainaut 79666b26c0 gsdx sw: restore the const reference
It was removed for PSX hack which was removed a couple of commits ago
2016-10-06 20:02:22 +02:00
Pseudonym a0629587be gsdx memory: implement read texture of PS GPU24
convert the swizzled block as tightly packed 24 bits RGB

Then convert scanline to standard 32 bits RGBA

The HW renderer requires the preload data hack
2016-10-05 21:59:42 +02:00
Gregory Hainaut 3653a7746a gsdx psxmode: small hack to "see" video until we found a better solution 2016-10-05 09:43:24 +02:00
Gregory Hainaut 97172f5e51 gsdx debug: count draw call in a different way
* Always do +1 before the draw call
* Prefix texture name with i (as input) to keep them before the FB

Goal is to ensure that all renderers share the same draw call value.
2016-09-28 19:41:11 +02:00
Gregory Hainaut 78b2848622 gsdx: factorize the TEX0 generation for mipmap 2016-09-25 11:51:48 +02:00
Gregory Hainaut e38aba0dff gsdx: rename SW texture dump filename
* Use texraw for the unconverted texture (keep index fmt)
  => avoid bad filename order with the multiple texture layers
* add the real mipmap address
* Use a nice string format
2016-09-24 09:15:40 +02:00
Gregory Hainaut ebdf54a60b gsdx sw: don't enable auto flush hack in replayer
Otherwise we don't have same number of draw call in HW/SW which is
awful to compare
2016-09-17 12:05:05 +02:00
Gregory Hainaut 3994141d40 gsdx sw: always enable the auto flush hack on the SW renderer
There are enough option. Will remove it if too slow.

Positively impact Jak, GTA, Battlefield 2
2016-09-15 17:24:40 +02:00
Gregory Hainaut c638e5ec87 gsdx debug: use a pretty format name when dumping texture
Much faster to read the format on the file name :)

+ remove a couple of useless ifdef
2016-09-07 08:34:24 +02:00
Gregory Hainaut a123e65185 gsdx sw: handle the invalidation of the palette when written by the GS
Manual gives all setup to upload a palette from the host. But nothing forbid to render
directly in the palette buffer. (GS rule nb 1, there is no rule ^^)

Fix Virtua Fighter 2 dark colors

However I'm not sure we can fix HW renderer. Rendering is done on the GPU but palette
handling is done on the CPU... So we need to read back data (ouch, and slow). A quick
test didn't get the expected results. Potentially there are others bugs (aka not gonna
happen on the HW renderer)
2016-09-03 00:21:56 +02:00
Gregory Hainaut d01194da0b gsdx sw: help debug
* quick way to disable verbose fixed tex0
* use palette address/format when the palette is dumped :)
2016-09-03 00:10:23 +02:00
Gregory Hainaut 15db7eeb81 gsdx: use correct format in printf
u => unsigned
d => signed

zu => size_t
2016-08-13 12:48:35 +02:00
Gregory Hainaut 41afd85a5d gsdx sw: fix GCC warning enumeral and non-enumeral type in conditional expression 2016-08-12 19:30:14 +02:00
Gregory Hainaut 0dcce91a6e Merge pull request #1504 from ssakash/RemoveFunction
GSDX: Remove GetDeviceSize()
2016-08-08 16:15:07 +02:00
Akash bc24e90070 GSDX: Remove GetDeviceSize()
It's not used anywhere and it doesn't seem needed.
2016-08-07 12:43:03 +05:30
Jonathan Li f978f9a07d Merge pull request #1477 from turtleli/gsdx-defer-init
gsdx: Avoid illegal instruction crash on older CPUs
2016-08-02 23:00:19 +01:00
Gregory Hainaut 92ae8a5be7 gsdx sw: print current draw call of GetSizeFixedTEX0
To easily detect compare with hardware renderer
2016-07-31 13:19:38 +02:00
Jonathan Li e5ea4cc5d4 gsdx: Defer vector initialisation to GS/GPUinit
It can cause PCSX2 to crash if the instructions aren't supported.
2016-07-27 23:34:14 +01:00
Gregory Hainaut 25bc2dff07 gsdx: dump and log EE texture read
It gives a visual opportunity to detect a bad read of the texture cache
2016-07-14 19:45:06 +02:00
Gregory Hainaut 063d2e568a gsdx tc: re-implement frame lookup
Try to avoid random black screen frame

v2: don't force the preload hack on the frame
It creates a ghost image over FMV

v3: support offset within a frame
2016-04-04 21:30:37 +02:00
Gregory Hainaut f904cd6c4a gsdx: add atomic for SW Renderer
V2: fix assertion
2016-02-22 19:12:37 +01:00
Gregory Hainaut 19c9a0b441 gsdx: remove aggressive threading
http://wiki.pcsx2.net/index.php/PCSX2_Documentation/Threading_Basics
2015-11-20 17:36:03 +01:00
Gregory Hainaut c12958bf10 gsdx: sign-compare
Need review
2015-11-12 12:11:44 +01:00
Akash 04b765a674 gsdx: check for null deference.
CID 146839 (#1 of 1): Explicit null dereferenced (FORWARD_NULL)11. var_deref_model: Passing null pointer fb_pages to UsePages, which dereferences it.

CID 146840 (#1 of 1): Explicit null dereferenced (FORWARD_NULL)11. var_deref_model: Passing null pointer zb_pages to UsePages, which dereferences it.

* Prevent a potential null pointer deference in ```void GSRendererSW::UsePages()```
2015-10-17 09:14:07 +05:30
Gregory Hainaut 5fb8c7e65c gsdx: initialize members in constructor of objects
A couple of useless members were removed too.

Also fix wnd initialization

Coverity:
CID 146955 (#1 of 1): Uninitialized pointer read (UNINIT)
18. uninit_use: Using uninitialized value wnd[i].
2015-09-23 09:46:53 +02:00
Gregory Hainaut 4eed4ca3a1 gsdx-debug: restore 16 bit dump format in the SW renderer
Actually dump both 32/16 bits (former for texture shuffle and later
for real 16 bits RT)
2015-09-08 12:41:05 +02:00
gabest11 d826d925db gsdx: eliminated a few bugs in the texture size changer algorithm 2015-08-07 02:08:29 +02:00
gabest11 49b3acea72 gsdx: texture size reduction in sw mode, fixes Stolen, less memory usage in general. 2015-08-05 19:11:41 +02:00
gabest11 42f51591df gsdx: lupin 3 fix, texture addressing outside the limits, only for sw and opencl yet 2015-08-04 13:27:08 +02:00
gabest11 d631030608 gsdx: fixed some of the renames where it made no sense, also added a commend about lupin 3rd. 2015-08-04 03:36:42 +02:00
Gregory Hainaut 3b127f663b gsdx-tc: trace the texture format to detect texture shuffling
It fixes games that uses 16 bits RT (like snow engine games)
2015-07-01 09:36:53 +02:00
Gregory Hainaut b62a2d6a3c gsdx-debug: dump texture in 32 bits when a shuffing effect is detected
Otherwise it is unreadable
2015-07-01 09:30:20 +02:00
Gregory Hainaut 5daf7aa98f gsdx-debug: keep s_n/pop in sync 2015-06-07 12:39:48 +02:00
Gregory Hainaut 9d69ff1f91 gsdx-debug: create the savef option
The idea is to easily extract bad frame from gs dump and to compare them
with the sw renderer.

Here the summary of the options:
savet => dump all input textures
save  => dump the render target
savez => dump the depth buffer
savef => dump the frame
2015-05-31 17:38:52 +02:00
Gregory Hainaut b884e0c0c0 gsdx-debug: improve debug experience
* Dump context before the increase of s_n
  => aligned with the global call number
* Don't print colclip not supported when it is optimized away
* dettach the input texture when it is useless
  => avoid to show a wrong texture in the debugger
2015-05-23 12:23:05 +02:00
Gregory Hainaut 28bb64aae8 gsdx: sed/dr/dRect/ 2015-05-15 20:49:32 +02:00
Gregory Hainaut d870188d21 gsdx: sed/o/off/ 2015-05-15 20:40:09 +02:00
Gregory Hainaut 8e1db43431 gsdx-debug: debug stuff 2015-05-08 19:28:17 +02:00
Gregory Hainaut cc4713d379 gsdx-debug: extend ogl debug capabilities
Group opengl calls into a nice name.

Apitrace shows them in a tree format that support folding. Previously it
was a long flat list (10K-40K of lines by frame)

I align the call number with the internal s_n variable. This way it is
easy to map GSdx dump output with the GL debugger :)
2015-05-06 19:09:13 +02:00
Gregory Hainaut 6fca3b421d gsdx-debug: dump register context in a .txt file
+ Realign s_n number if frame is skipped

This way it is easier to understand current draw. It also eases
comparaison between SW/HW.
2015-05-06 19:09:12 +02:00
Gregory Hainaut 7bfee7e377 gsdx-debug: stop the dump after a limited number of draw call
Otherwise it fills your disk in a short time with useless data.
2015-05-03 16:46:48 +02:00
Gregory Hainaut c8a3db114c gsdx debug: create savet option
0: don't dump input texture
1: dump input texture

Now, you can do a first pass with only RT. When you find the wrong call, you can redump the input texture of the bad draw.
2015-05-01 13:35:21 +02:00