984bb31709
Fix another race condition with multithreading.
...
It's possible that after it's marked completed, the CPU will grab it to
enqueue. This actually happened to mean with decent reproducibility.
2014-05-03 18:40:27 -07:00
99e4300f95
Eat a few more cycles on END cmds.
...
Fixes #3049 , which apparently expects an END/FINISH to not immediately
trigger (it drawsyncs soon after, and completely expects that to
reschedule.)
Either way, it seems plausible an END would take a few more cycles than
most instructions, since it generates an interrupt.
2014-04-29 00:08:43 -07:00
97c18e7f0e
Comment out a few unsafe replacement funcs.
2014-04-22 08:07:10 -07:00
256e98b841
Add the common funcs to the direct exec calls.
2014-04-17 01:18:52 -07:00
b2f12daba9
Don't reset OFFSETADDR, changes offsetAddr.
2014-04-17 01:18:51 -07:00
4c1c694d4b
Use flags to avoid hashing textures when unchanged.
...
If only parameters change (like wrapping or clut, etc.) we don't need to
rehash the data - we know it hasn't changed.
Should reduce the distance between lazy texture hashing on and off.
2014-04-13 14:22:39 -07:00
e93407f33e
Oops, downgrade from ERROR to DEBUG, and -report.
2014-04-12 00:04:02 -07:00
26933384a7
Don't mark a list complete too early.
...
This would matter probably mostly if we implemented CONTINUE properly.
2014-04-11 23:50:20 -07:00
4561440df7
Handle GE pause signals per tests.
...
They really should pause, but we were resetting them incorrectly.
And most importantly, sceGeContinue() inside the signal handler absolutely
must work. Games use this a lot.
2014-04-11 23:50:20 -07:00
702294fe60
Don't trigger pause GE signals either, comments.
...
This better explains what each signal does.
2014-04-10 00:23:36 -07:00
9c75c1e6fd
Avoid triggering SYNC ge signals.
...
Fixes #5806 , bcause it avoids rescheduling at the signal, which it should
not do. Possibly this is incorrect for all GE signals, but we don't need
to trigger any behavior on a sync anyway.
2014-04-08 22:20:33 -07:00
8a440f93b1
Super stupid typo.
2014-04-05 14:08:44 -07:00
82df1b1fa8
Simplify UpdatePC() a bit more.
2014-04-05 12:04:10 -07:00
cbb6c01d4c
Avoid some debug stats when disabled.
...
This function can get called a lot sometimes, and an if helps on ARM a
little.
2014-04-05 01:41:54 -07:00
7cde2311e0
Include sceKernelThread.h in less headers.
2014-03-29 17:02:41 -07:00
a4327702f1
Reduce some includes under GPU/.
2014-03-29 16:51:38 -07:00
4fbb245382
Avoid leaving the fast runloop on jumps.
...
Jumps are actually very common in some games, like FF4 and Crisis Core,
and tons more. They are used to jump around vertex data.
Improves performance by a few percent in FF4.
2014-03-05 23:24:18 -08:00
e11e4cfff2
GCC buildfix
2014-03-04 11:38:33 +01:00
0a880b09ad
Allow fast bone mtx loads w/o software skinning.
2014-03-02 18:12:40 -08:00
db59559a92
Reduce the "Ret: Stack Empty" log level to DEBUG.
...
Still reporting. Avoids logspam in buggy games.
2014-02-25 21:23:21 +07:00
183b4cb45c
Add a static method for creating PSPPointers.
...
This is still kinda ugly, but the assignment thing was just too ugly, I
couldn't do it anymore.
2013-12-16 23:50:52 -08:00
435115441d
Reinitialize the GE on sceKernelLoadExec().
...
Mostly, reset the lists and draw completion, etc.
May be required for Capcom Classic Collection Reloaded to work.
2013-11-28 15:35:12 -08:00
dc03796ccb
The bone matrix optimization is only safe when software skinning
2013-11-14 15:02:22 +01:00
8c562a615f
Bone matrix loading speedup. Bit of a hack but seems reliable enough.
2013-11-14 14:02:31 +01:00
1633aa689c
Remove the extra process queues hack.
...
It seems like it's not helping anymore so it could be hurting.
2013-11-09 23:08:44 -08:00
2c74693f6c
VSync codepath only works on _WIN32 so only show it and check it there.
2013-10-27 23:09:46 +10:00
4972e7b680
EnqueueList: Relocate code that depends on compatibility check to simplify the code
2013-10-27 03:49:46 +10:00
afb578a325
Use the newer interface for Qt's gpu debugging.
2013-10-20 00:14:39 -07:00
acb9c49ea2
Try to use a smaller value for cyclesExecuted
2013-10-12 16:20:27 +08:00
2cdb2d1e1f
Extra cycles execute .Fix #3979 & #3883
2013-10-09 19:45:28 +08:00
fc8aa7bd8f
Crashfix dumping display lists that start at the start of VRAM (seems to hit this in GTA).
...
Add a couple of comments.
2013-10-07 22:58:46 +02:00
8f7a9656c8
Merge pull request #4062 from unknownbrackets/debugger
...
Show tabs for depth/stencil in ge debugger, other improvements
2013-10-06 23:59:54 -07:00
ff0e9cf32d
Make it possible to flip flags in the ge debugger.
2013-10-06 22:17:56 -07:00
0c03dc2dec
Handle it better when viewing an invalid addr.
2013-10-06 22:17:52 -07:00
a09119ba04
Correct savestates when dls had context saving.
...
It was wrong, and also they were different between 64/32 bit.
2013-10-06 22:10:44 -07:00
74908f6b38
Add errors and rescheduling to sceDmac*.
...
The functionality seems pretty straight-forward.
2013-09-29 16:51:49 -07:00
424b4cbd5f
Move GE_CMD_BJUMP to GPUCommon. Add the flag. (no actual bbox impl yet)
2013-09-24 14:14:04 +02:00
bdfc8bf4b3
Mark texture as maybe changed when running lists.
2013-09-23 23:17:56 -07:00
4c3b3484d7
Actually apply the restored gstate.
2013-09-23 23:17:36 -07:00
3787471791
Add some initial methods to GPUDebugInterface.
2013-09-22 10:22:33 -07:00
eabd8b5302
Add GPU debugging hooks to Host, not yet in use.
2013-09-22 08:14:55 -07:00
2e2fa53201
Preseve offsetAddr on stall, instead of reset to 0.
...
Only for new lists is it reset to 0, afaict.
2013-09-21 21:01:41 -07:00
881cefbc83
A paused list will allow a context save.
...
Just not stall, drawing, etc.
2013-09-21 21:01:41 -07:00
e10ae1530b
Error and reporting for sceGeBreak's second param.
2013-09-21 21:01:40 -07:00
e9db63f61d
Refuse to dequeue a running/completed list.
2013-09-21 15:03:30 -07:00
c5a709d039
Don't update stall address on completed lists.
2013-09-21 15:03:30 -07:00
e7bd716c71
Allocate ge list ids using round robin.
...
Even if you enqueue and then dequeue 0, you don't get it again right away.
2013-09-21 15:03:29 -07:00
863eb83e4c
Add support for sceGeGetStack() for debugging.
2013-09-21 10:04:31 -07:00
3c954ea0e4
Save the base address in signal calls.
...
It's not saved in regular calls but it is in signal ones.
2013-09-21 00:07:49 -07:00
9116b879a4
Restore context even without an interrupt.
2013-09-20 21:03:39 -07:00
Unknown W. Brackets