preserve-emulation/ppsspp
1
0
Fork 0
mirror of https://github.com/hrydgard/ppsspp.git synced 2026-07-27 01:05:07 +02:00
Code Issues Packages Projects Releases Wiki Activity
45,601 Commits 79 Branches 131 Tags
https-linux-poc
Commit Graph
6340 Commits
Author SHA1 Message Date
Unknown W. Brackets fd546ff21d Fix some type comparison warnings. 2016-08-05 10:47:45 -07:00
Unknown W. Brackets e55198f7e7 Correct some initialization order warnings. 
Also, another missing init with IRBlocks.
 2016-08-05 10:46:11 -07:00
sum2012 390ecb3184 Better support zh_gb.pgf 2016-08-05 19:58:31 +08:00
Unknown W. Brackets db12337b39 Add basic kernel mode file func entries. 2016-08-04 09:46:29 -07:00
Unknown W. Brackets e78a389ecc Move sceIo kernel tables to sceIo. 2016-08-04 09:46:29 -07:00
Unknown W. Brackets 0b43e28595 Move SysMemForKernel to sceKernelMemory. 2016-08-04 09:46:27 -07:00
Unknown W. Brackets 4b4770bbc5 Log kernel syscalls with a K. 
This way they can be properly differentiated.
 2016-08-04 09:46:26 -07:00
sum2012 1cc645b5e4 Revert "Temporarily disable PMP video for non-windows" 
This revert
https://github.com/hrydgard/ppsspp/pull/8693/commits/5e20ba18114c4c37c15e144598952d370e81edfe
 2016-07-25 20:21:53 +08:00
Unknown W. Brackets 558b4620e8 Mpeg: Parse video streams from PSMF header. 
Without doing this, FFmpeg will try to probe the streams to detect them
instead.  When it does this, sometimes it tries to read beyond the data
that's available - and then gets confused by EOFs.

Parsing this way allows us to control the situation.

An example is Valkyrie Profile, corruption in the first frames of the
second video during the intro.  Thi doesn't fix it yet, but now it's just
a matter of buffering.
 2016-07-24 15:58:06 -07:00
Henrik RydgårdandGitHub a1e815e356 Merge pull request #8651 from unknownbrackets/android-clang 
Switch Android build to using clang (needs buildbot update)
 2016-07-24 22:14:58 +02:00
Henrik Rydgard 2251c9c8f7 Help #8816 by partially reverting #8753. Does not match the tests though. Firmware difference? 2016-07-24 22:12:17 +02:00
Unknown W. Brackets 266ee632e8 Cleanup FFmpeg funcs deprecated in 3.x. 2016-07-24 12:53:54 -07:00
Henrik Rydgard 40549b14fd Error code cleanup 2016-07-24 21:22:26 +02:00
Henrik Rydgard 745081cc4e Return errors when trying to close file with pending operations. May help #6357. 2016-07-24 16:03:34 +02:00
Unknown W. Brackets 5009dcab7a Add reporting for sceKernelDevkitVersion() calls. 
We currently default to 1.50 as the reported version, but we more or less
emulate a 6.60 system.  It may be better to return 6.60, but let's see
what would actually be impacted by that first.
 2016-06-26 20:52:48 -07:00
Unknown W. Brackets 80d72acaec Psmf: Correct default stream. 
Fixes Jeanne d'Arc, others.
 2016-06-07 20:41:13 -07:00
Henrik Rydgård ce8d2bef57 Merge pull request #8803 from unknownbrackets/mpeg 
Reject invalid MPEG puts for certain lib versions
 2016-06-05 13:44:38 +02:00
Unknown W. Brackets bda1956f90 Psmf: Enable error check on player create. 
Reporting makes it look like this should be safe.
 2016-06-04 18:17:38 -07:00
Unknown W. Brackets cf2796de41 Psmf: Allow each struct to have its own stream. 
Some games will create a clone, and then set a different stream on it,
e.g. one for the video, another for the audio stream.
 2016-06-04 18:17:37 -07:00
Unknown W. Brackets b13c68a373 Psmf: Correct get video/audio info errors. 
This actually grabs from the currently specified stream, and fails if it's
the wrong type.  Additionally, return stream-specific info... unlikely to
matter, hard to imagine a video with a mono and stereo stream.

Also reduces log verbosity.
 2016-06-04 18:17:37 -07:00
Unknown W. Brackets 185ad781f3 Psmf: Add better error handling for SetPsmf. 
Checked on different devkit versions in case of changes.
 2016-06-04 18:17:36 -07:00
Unknown W. Brackets 04dcaeae1d Mpeg: Move, rather than resetting, packetsAvail. 
When we've got garbage data, this has to stay incorrect.  Without this,
Megaman X gets confused when playing its video (because it enqueues
garbage packets.)

See #3318.
 2016-06-04 18:16:23 -07:00
Unknown W. Brackets 4ee39f82e4 Mpeg: Validate added packets. 
Older libraries only, but this will cause it to reject packets that don't
make sense.  So far, this seems to mirror the behavior of various garbage
packets sent to the real firmware.
 2016-06-04 18:16:22 -07:00
Unknown W. Brackets 41f2999a2b Mpeg: Write into the ringbuffer using actual ring. 
This mimics behavior with clamped sizes and wrap around, and also makes it
easier to implement the garbage data handling the PSP has.
 2016-06-04 18:16:21 -07:00
Unknown W. Brackets 6d14c17f26 Mpeg: Treat negative available packets as zero. 2016-06-04 18:16:19 -07:00
Unknown W. Brackets deffb246ae Mpeg: Require less memory for older lib versions. 
It seems it started requiring 64k at 1.05.
 2016-06-04 18:16:18 -07:00
Henrik Rydgård 606856c9ad Merge pull request #8791 from unknownbrackets/delay 
Minor sysclock delay accuracy improvements
 2016-05-31 21:08:00 +02:00
Unknown W. Brackets 91887ad85c Psmf: Correct errors in current stream num. 2016-05-31 09:55:51 -07:00
Unknown W. Brackets 0a00777403 Psmf: Only write current type/channel if valid. 
And improve error handling in general.
 2016-05-31 09:55:51 -07:00
Unknown W. Brackets 7ad07ca82e Psmf: Handle stream specification per tests. 
These seem buggy, but generally sane.
 2016-05-31 09:55:50 -07:00
Unknown W. Brackets cca9a73105 Psmf: Start at stream 0 when not specified. 2016-05-31 09:55:49 -07:00
Unknown W. Brackets 510db2051d Psmf: Detect PCM streams from private stream info. 
This logic comes from Jpcsp.  Let's report though and see if there's more.
 2016-05-31 09:55:49 -07:00
Unknown W. Brackets 74239c199d Psmf: Allow AUDIO type when specifying streams. 2016-05-31 09:55:48 -07:00
Unknown W. Brackets 5f7c95f249 Psmf: Count PCM as an AUDIO stream as well. 2016-05-31 09:55:47 -07:00
Unknown W. Brackets 9df727f94f Psmf: Stop pretending streamMap is a vector. 2016-05-31 09:55:47 -07:00
Unknown W. Brackets 8df298f622 Thread: Auto-resume dispatch for worse threads. 
Seems this doesn't happen when you start a better priority thread, pretty
reproducibly.
 2016-05-31 09:55:20 -07:00
Unknown W. Brackets 197f3393df Thread: Better/minor accuracy for sysclock delay. 
Certain negative values shouldn't cause a delay, could help if a game has
a bug.
 2016-05-31 09:55:19 -07:00
Unknown W. Brackets baac12b327 Log args for dispatch/interrupt early-fail. 2016-05-31 09:55:18 -07:00
Unknown W. Brackets d0d570c6ac ThreadEvent: Delete threads after handler runs. 
It should actually run on the thread itself, it seems, but that's probably
not as important as the thread still existing.  This allows the handler to
get the thread name or etc.
 2016-05-30 13:00:23 -07:00
Unknown W. Brackets 21ba67e1d7 Module: Add legacy save state stub fixup. 
Forgot this, when we don't have the stub info.  Stub info is better, since
we might handle things differently.
 2016-05-30 02:11:28 -07:00
Unknown W. Brackets 65820d6184 Thread: Use a short delay for 0us delays after all. 
It seems the behavior varies.  Should help #6924 and #6981.
 2016-05-29 22:04:47 -07:00
Unknown W. Brackets 1a3fc9cdeb Module: Avoid writing missing stubs on reimport. 
In case they were manually linked, or something.
 2016-05-29 08:29:51 -07:00
Unknown W. Brackets 9b706975ad Module: Reimport stubs on load state. 2016-05-29 08:27:19 -07:00
Unknown W. Brackets 47f16ed97b Module: Refactor stub imports to dedicated func. 2016-05-29 08:07:01 -07:00
Unknown W. Brackets 12b907ed60 Put syscalls back where they were supposed to be. 
Oops, didn't mean to move this.  Might've broken save states.
 2016-05-29 07:31:22 -07:00
Unknown W. Brackets 4fc12cdd29 ThreadEvent: Handle getting scheduled off a thread. 2016-05-29 00:59:52 -07:00
Unknown W. Brackets 97cce6ccfe ThreadEvent: Trigger handlers for start/exit too. 
Start isn't in the right place or on the right thread, and exit isn't on
the right thread either.  But these don't seem very particular about what
thread they run on.
 2016-05-28 21:19:28 -07:00
Unknown W. Brackets c6cd614f2b Thread: Allow mipscalls to be scheduled in a row. 
Should be fine to just use the stack to save these things, likely that's
what happens in real firmware.  This fixes issues when a second mipscall
is scheduled in the same HLE syscall.
 2016-05-28 21:18:03 -07:00
Unknown W. Brackets ceb36fc81d ThreadEvent: Allow an exit event handler on self. 2016-05-28 21:17:54 -07:00
Unknown W. Brackets be1cde15ba ThreadEvent: Initial support for create/delete. 2016-05-28 21:17:36 -07:00