cai567890 / pcsx2

Automatically exported from code.google.com/p/pcsx2
1 stars 0 forks source link

Mana Khemia regression (weird scereen) in r989 #174

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago
Well, in r989 when hero goto Students Affairs for his first assignment
(it's the second time he go to this location) picture become inadequate:
swizzled and inappropriate. The reason of trouble is line 130 in file GS.cpp:

u32 qwc = ((psHu32(GIF_MODE) & 0x4) && (vif1Regs->mskpath3)) ? min(8,
(int)gif->qwc) : min( gifsplit, (int)gif->qwc );

I check this issue in Windows later, if nobody do it before.

p.S. Somewhere near r984 Mana Khemia backgrounds become correctly aligned,
nice!

Original issue reported on code.google.com by Zeydl...@gmail.com on 22 Apr 2009 at 7:15

GoogleCodeExporter commented 9 years ago
get a newer checkout of gsdx off the forum, that bug has been fixed by gabest 
(was a
gsdx bug)

Original comment by refraction on 22 Apr 2009 at 8:37

GoogleCodeExporter commented 9 years ago
If Zeydlitz is going to check in Windows later, I don't think he's using gsdx...

Original comment by arcum42@gmail.com on 22 Apr 2009 at 8:40

GoogleCodeExporter commented 9 years ago
could be zerogs ogl has a similar bug, i dont know, i only know the dx version 
doesnt
do it :p

Original comment by refraction on 22 Apr 2009 at 8:58

GoogleCodeExporter commented 9 years ago
ah youre zzogl dude, check r1005 for gsdx9, that fixes the problem in having 
split
path 3 transfers

Original comment by refraction on 22 Apr 2009 at 9:03

GoogleCodeExporter commented 9 years ago
So as I see in Windows, ZeroGS DX, OpenGL and ZZogl are affected. Poor me.

Original comment by Zeydl...@gmail.com on 22 Apr 2009 at 9:18

GoogleCodeExporter commented 9 years ago
And GSDX 846 does not have this issue, so it's Zero specific issue.

Original comment by Zeydl...@gmail.com on 22 Apr 2009 at 9:25

GoogleCodeExporter commented 9 years ago
As we dont really bother with ZeroGS, im marking this as a wontfix

Original comment by refraction on 22 Apr 2009 at 9:31

GoogleCodeExporter commented 9 years ago
Question: how did plugun knows is that 128-bit package is final package of the 
longer
one (in older versions?) O'k, if older one was not divided by 128, I could check
divisor by 128 and find final one by this. But what shod I do if I need skip to 
the
end of 1024 package? Could I?

Original comment by Zeydl...@gmail.com on 22 Apr 2009 at 5:09

GoogleCodeExporter commented 9 years ago
O'k. I found a good solution for all ZeroGS plugins. Maybe it also resolve some 
gust
trouble with them. Need to made a thoughfull tests.

Original comment by Zeydl...@gmail.com on 22 Apr 2009 at 6:04

GoogleCodeExporter commented 9 years ago
Glad to hear it. After you've tested it a bit, if you've got a patch that would 
work
on ZeroGS,let me know, and I'll be willing to apply the patch in svn.

Though I usually just use ZZOgl these days. :)

Original comment by arcum42@gmail.com on 22 Apr 2009 at 8:29

GoogleCodeExporter commented 9 years ago
that could be nice, as the GIF has an "intermediate" transfer mode (aka 
transfer 8qw
at a time) this could solve a few issues with that too.  After i got gabest 
fixing it
in gsdx, Need for Speed Most Wanted looked tons better for it!

Original comment by refraction on 22 Apr 2009 at 10:42

GoogleCodeExporter commented 9 years ago
O'k. I check all my game list. Some gust games have an improvements (at least I 
was
anle to run Atelier Iris 1). And the soultion is trivial: in gsmain.cpp:984, in
function _GSgifTransfer there is following code:

            if(path->tag.nloop == 0 ) 
            {
                if( path == &gs.path1 ) 
                {
                    // ffx hack
                    if( path->tag.eop )
                        return;
                    continue;                   
                }

                if( !path->tag.eop ) 
                {
                    //DEBUG_LOG("continuing from eop\n");
                    continue;
                }

                break;
            }

And this last break should be continue, not break.

Original comment by Zeydl...@gmail.com on 23 Apr 2009 at 4:51

GoogleCodeExporter commented 9 years ago
Zeydlitz
How to applies your source? For ZeroGS OGL or ZeroGS DX.

Original comment by kenzozus...@hotmail.com on 23 Apr 2009 at 6:21

GoogleCodeExporter commented 9 years ago
Found this code in GSmain.cpp and change it manually.

Original comment by Zeydl...@gmail.com on 23 Apr 2009 at 6:35

GoogleCodeExporter commented 9 years ago
Made the change. I'll play with it a little and then commit. Atelier Iris 1's my
favorite of the series, so anything that improves it is welcome...

Original comment by arcum42@gmail.com on 23 Apr 2009 at 9:48

GoogleCodeExporter commented 9 years ago
Applied. And, since it's silly to have something committed that marked "Won't 
Fix",
lets change that.

Original comment by arcum42@gmail.com on 23 Apr 2009 at 10:19