thirdparty/ppsspp
2
0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
3250 commits 39 branches 0 tags 279 MiB
Commit graph

60 commits

Author SHA1 Message Date
Unknown W. Brackets
e133d33167 Don't re-wake a thread already woken by delete. 
Arg, I thought I'd changed this already.
 2013-03-24 22:54:09 -07:00
Unknown W. Brackets
da3af79f3b Turn down logging for LwMutexes. 
No one seems to have run into issues with these for a while, and they are
quite noisy, since games use them as lightweight locks.
 2013-03-10 22:57:27 -07:00
Unknown W. Brackets
37dffb6210 Implement sceKernelReferLwMutexStatus() and ByID(). 2013-02-24 22:50:48 -08:00
Unknown W. Brackets
ad3152cb94 Fix sceKernelReferMutexStatus(). 2013-02-24 20:28:39 -08:00
Unknown W. Brackets
2a6457b6ab Cut down on h files including PointerWrap. 
This makes changes to it a bit faster to build.
 2013-02-04 08:26:59 -08:00
Unknown W. Brackets
1cb7a88137 Don't use a dynamic string for wait debugging. 
Happens in release, and shows up on the profiler.
Not huge, but should save ~0.5% of cpu time.
 2013-01-26 11:29:19 -08:00
Henrik Rydgard
8d845cc588 quick implementation of sceKernelReferMutexStatus 2013-01-20 23:11:48 +01:00
Unknown W. Brackets
43c8726554 Oops, 0 is a valid timer id. 2013-01-18 01:38:14 -08:00
Unknown W. Brackets
305c41c932 Prevent duplicate waiting threads in sema/mutex. 
If something did a tight loop of a short wait/timeout, it might end
up getting on the waitingThreads list multiple times.

Eventually this caused a massive performance problem.
 2012-12-31 23:23:52 -08:00
Unknown W. Brackets
bcecd881b1 Savestate cleanup + be more careful with vectors. 
For some reason I debugged it going into std::string so I just made
them all more explicit which fixed everything.
 2012-12-28 13:55:31 -08:00
Unknown W. Brackets
3ed2045f6b Add state saving to tertiary kernel modules. 2012-12-28 13:55:25 -08:00
Unknown W. Brackets
595759ef78 Implement a factory for kernel objects. 
Couldn't think of a better way to do this, maybe there's some fancy
one I don't know about.

Also finished up a couple left over kernel objects.

Maybe a quarter the way there?
 2012-12-28 13:55:22 -08:00
Unknown W. Brackets
991243fffd Add basic DoState() to tertiary kernel objects. 2012-12-28 13:52:52 -08:00
Henrik Rydgård
b2f29992a3 Wrap some functions, clean some stuff 2012-12-26 08:52:40 +01:00
Unknown W. Brackets
9ba0ed46fd Always create CoreTiming timers during kernel init. 2012-12-23 21:27:26 -08:00
Unknown W. Brackets
65cfc8de86 Cleanup and fix mutex init, no need for a flag now. 2012-12-23 18:27:07 -08:00
Henrik Rydgård
4a4d126dc7 Init/shutdown mutex 2012-12-23 11:19:55 +01:00
Unknown W. Brackets
87fcdc9080 Use KERNELOBJECT_MAX_NAME_LENGTH for name length. 
Just avoiding magic numbers, didn't know there was a constant.
 2012-12-16 21:36:52 -08:00
Unknown W. Brackets
bb10140728 Oops, event flags weren't always deleting properly. 
Added some comments to explain why removal from waitingThreads is
deferred.  Also had to move things around so the right outBitsPtr
is always written to on timeouts.
 2012-12-15 01:36:53 -08:00
Unknown W. Brackets
67f7f6b05d Oops, these flags can be or'd together of course. 2012-12-14 23:24:04 -08:00
Unknown W. Brackets
2ea992f2c4 Validate attr for mutexes, semas, and events. 2012-12-14 12:05:15 -08:00
Unknown W. Brackets
3d57a4c838 Cleanup some wrappers now that rescheds defer. 2012-12-09 01:04:54 -08:00
Unknown W. Brackets
5e8aa4c071 Implement deferred rescheduling/callbacks. 
This way most HLE functions can be wrapped normally.  Hurray, sanity.
 2012-12-09 01:04:53 -08:00
Unknown W. Brackets
fda41752e8 Simplify the API for HLE funcs to do callbacks. 
__KernelWaitCurThread() already (ultimately) calls __KernelCheckCallbacks.
 2012-12-09 01:04:52 -08:00
Unknown W. Brackets
ee1a729157 Turns out *CB() fires even when it succeeds. 2012-12-09 01:04:50 -08:00
Unknown W. Brackets
f9b561ee42 Implement priority unlocking for mutexes. 2012-12-09 01:04:49 -08:00
Unknown W. Brackets
a67505761e Mutexes and semaphores were rescheduling twice. 
This was causing it to "jump" the correct thread and go to the next.
 2012-12-07 00:46:18 -08:00
Unknown W. Brackets
00ca9d8dc7 Make the UMD wait funcs actually wait. 2012-12-01 23:29:24 -08:00
Unknown W. Brackets
998104e2eb Consider threads still waiting until they wake. 
This seems to match the actual behavior of timeouts much better.
 2012-11-30 23:04:24 -08:00
Unknown W. Brackets
48c5efd0c2 Match hw timing of mutex/sema timeouts better. 
Affects some tests passing/not passing due to minor timing issues, ugh.
 2012-11-30 22:58:42 -08:00
Unknown W. Brackets
3ab41e515d Mutexes and semaphores don't always reschedule. 
It's just the normal rescheduling happening.  Previous tests weren't
correct.

Also some minor cleanup.
 2012-11-30 22:58:41 -08:00
Henrik Rydgard
980d13fe50 Assorted GPU fixes: Advance vertex pointer, fixes missing triangles in SPT. More logging. Separate proj matrices for through and normal mode (through matrix don't need to get updated as often). Some cleanup. 2012-11-28 13:45:22 +01:00
Unknown W. Brackets
d3f2d9d10b Use u64 here to avoid warnings, better precision. 
The timeout value is u32 anyway, so an int was already a loss.
 2012-11-22 16:40:26 -08:00
Unknown W. Brackets
c2c2b77792 Fix threads locking more than one mutex at a time. 
Duh, of course they can.  Even my test did, I'm stupid.
 2012-11-22 16:33:16 -08:00
Unknown W. Brackets
1bf04790e9 Darn, buildfix for debug. 2012-11-21 07:28:54 -08:00
Unknown W. Brackets
879efb25cc Don't allow unlocking someone else's mutex. 2012-11-21 00:28:29 -08:00
Unknown W. Brackets
fa9b9840d0 Implement LwMutex locking and unlocking. 2012-11-21 00:21:25 -08:00
Unknown W. Brackets
db16f9309e Don't crash if something goes wrong with mutexes. 
Probably can't happen but being defensive is good.
 2012-11-20 22:57:14 -08:00
Unknown W. Brackets
b71d998a20 Broadcast a threadend event to wake mutexes. 2012-11-20 02:20:16 -08:00
Unknown W. Brackets
dd3db71b61 Forget mutex waiting threads when they timeout. 2012-11-20 02:20:16 -08:00
Unknown W. Brackets
8da2fb074c For mutexes/semaphores, register a global timer. 
Instead of leaking one per each.  Should be faster.
Since we can only wait on one thing per thread at a time, this is fine.
 2012-11-20 02:20:16 -08:00
Unknown W. Brackets
a81c138004 Create and Delete LwMutex funcs. 2012-11-20 02:20:15 -08:00
Unknown W. Brackets
fd41851503 Implement timeouts on semaphores too. 2012-11-20 02:20:14 -08:00
Unknown W. Brackets
4d71e32193 Update the value when mutexes timeout. 2012-11-20 02:20:14 -08:00
Unknown W. Brackets
aad7e48a1a Oops, turns out mutex timeouts are in micro. 
Found it documented as milli somewhere, but it was wrong.
 2012-11-20 02:20:14 -08:00
Unknown W. Brackets
d88fa153b7 Implement timeouts on mutexes. 
Still need to set timeoutPtr after they expire.
Pretty sure I did this the right way, seems CoreTiming does already
allow/use multiple events for the same type?
 2012-11-20 02:20:13 -08:00
Unknown W. Brackets
7b9a8bf766 Minor correctness on create mutex. 2012-11-20 02:20:12 -08:00
Unknown W. Brackets
7fa388952e Refactor mutex locking to simplify. 2012-11-20 02:20:12 -08:00
Unknown W. Brackets
0681c8e5c5 Add overflow error to mutex locking. 2012-11-20 02:20:12 -08:00
Unknown W. Brackets
2290397b94 sceKernelTryLockMutex() and fix other locking. 2012-11-20 02:19:21 -08:00