daniel229
a66423dfa3
Replace frame download in "Rewrite Portable"
2015-01-06 14:19:36 +08:00
Unknown W. Brackets
61d43f1616
Fix some glitches in the new AA3 handling.
...
Matches a test better.
2015-01-04 16:34:58 -08:00
Unknown W. Brackets
23893ffba8
x86jit: Fix crash in prefixes and simd enabled.
2015-01-04 15:50:12 -08:00
Henrik Rydgard
f188737a86
Just to investigate #7269 , skip the version check on ARMEABI (very old ARM)
2015-01-04 23:45:30 +01:00
Henrik Rydgard
90376267b1
More reg number asserts
2015-01-04 23:15:33 +01:00
Henrik Rydgard
fe63a15e22
x86jit fpu regcache: Add extra invariant check for too-large mips register numbers
2015-01-04 22:54:01 +01:00
Henrik Rydgård
82c6079f1f
Merge pull request #7275 from unknownbrackets/atrac-aa3
...
Correct sceAtracSetAA3*() handling
2015-01-04 22:21:28 +01:00
Unknown W. Brackets
585d0d3a56
Update armips.
2015-01-04 13:04:30 -08:00
Unknown W. Brackets
db5af4f838
Correct sceAtracSetAA3*() handling.
...
This is used to send OMA / AA3 files, which have a different format.
Couldn't have been working before, may not work correctly still.
2015-01-04 12:56:28 -08:00
Henrik Rydgård
7b2f05d0a0
Merge pull request #7270 from Bigpet/adhocthread
...
don't use Thread* to create thread
2015-01-04 20:34:19 +01:00
Peter Tissen
a7d9370a69
don't use Thread* to create thread
...
slightly less hacky, as in, a little bit closer to what the actual PSP does, although the timings are still complete guesses
also
* Protect the adhoc Events by a mutex lock
* don't use emplace_back since symbian doesn't compile with it
2015-01-04 18:58:42 +01:00
Henrik Rydgard
a5d961b8d7
Clean up some compiler warnings in adhoc code
2015-01-04 14:09:41 +01:00
Henrik Rydgard
d479fae481
Show some major networking errors on the OSD
2015-01-04 13:04:11 +01:00
Henrik Rydgard
101fe5c481
Don't launch the friend finder thread if the adhoc network init fails.
...
Adds a flag for network success but doesn't check it anywhere yet..
2015-01-04 13:04:10 +01:00
Henrik Rydgard
58238bc7ed
GCC buildfix
2015-01-04 13:04:10 +01:00
sum2012
9933ae0c6e
netAdhocctlInited should be true when return success
...
Update back new file
Windows (?) build fix
2015-01-04 13:04:09 +01:00
Bigpet
43b3707ee6
don't initialize adhoc connection when networking is disabled
2015-01-04 13:04:08 +01:00
sum2012
ee6f6d11fc
Move Kill AdhocServer Thread to __NetAdhocShutdown
2015-01-04 13:04:08 +01:00
sum2012
aa69489aaf
Move built-in AdhocServer Thread to __NetAdhocInit
...
There are games (ie. Naruto Kizuna Drive) who tried to initialize
network/connect to adhoc server early when the game started, if adhoc
server wasn't ready the player may need to restart the game in order to
connect successfully
2015-01-04 13:04:07 +01:00
sum2012
6ed15ab1f5
Add EnableWlan check
2015-01-04 13:04:06 +01:00
sum2012
21f3d0ecf2
Add built-in AdhocServer option
2015-01-04 13:04:06 +01:00
sum2012
589cf551a0
Revert "Fixed an issue when MipsCall executed on any random thread instead of it's own thread causing some registers to be corrupted and leading to invalid address access"
...
This reverts commit 8a65d73ad25ad0f35b02f0e3af84de37625550b0.
2015-01-04 13:04:05 +01:00
sum2012
accda25ce0
Travis CI Fix, typo
2015-01-04 13:04:05 +01:00
AdamN
1b8167f9df
Changing the socket order checking in AdhocPollSocket a little bit
...
PS: The way how ppsspp indexing the socket may need to be changed to prevent PDP and PTP socket having the possibility to use the same socket id
2015-01-04 13:04:04 +01:00
AdamN
1a61bbbdf2
Adding socket error checking on AdhocServer
2015-01-04 13:04:04 +01:00
sum2012
e0a6480223
Move database.db to Tools\PSP game data
2015-01-04 13:04:03 +01:00
sum2012
7c323f8571
Another build fix
2015-01-04 13:04:03 +01:00
AdamN
1c7152db3a
Adding Branch opcode to the MAKE_MIPS
...
Temp revert old version to fix conflict; Implementing built-in PRO Adhoc Server; Add back sceUsbGps
Implementing more Adhoc functions, and some bug fix (i don't remember them anymore)
2015-01-04 13:04:02 +01:00
AdamN
3dc6eed05b
Giving access to some Kernel functions in order to execute MipsCall on the right thread instead of any random thread (current thread)
2015-01-04 13:04:01 +01:00
AdamN
69fb339a83
Fixed an issue when MipsCall executed on any random thread instead of it's own thread causing some registers to be corrupted and leading to invalid address access
2015-01-04 13:04:01 +01:00
AdamN
2cc19f56b2
Rearrange Network Shutdown to be properly in order (NetInit->AdhocInit->network functions->AdhocTerm->NetTerm)
2015-01-04 13:04:00 +01:00
Henrik Rydgård
303b92d703
Merge pull request #7265 from unknownbrackets/mpeg-ringbuffer
...
Reverse packetsFree to packetsAvail
2015-01-04 09:55:13 +01:00
daniel229
6cebc3a0b0
Replace function with memmove in Jui Dr. Touma Jotarou
2015-01-04 14:07:17 +08:00
Unknown W. Brackets
ce33a9f128
Try to handle savestates better.
...
Just in case. This isn't perfect but it should catch common cases.
2015-01-03 21:46:40 -08:00
Unknown W. Brackets
67cbf2e904
Reverse packetsFree to packetsAvail.
...
This is the actual value stored in memory, which games may be accessing or
even writing to.
Makes the video/mpeg/ringbuffer/avail test pass.
2015-01-03 18:59:56 -08:00
Bovine
54c7a12306
Fix simd vmmul transpose optimizations.
...
Yep, gotta discard the S matrix when we're done with it.
2015-01-03 16:58:03 -07:00
Bovine
0fdebdc1ca
Fix discardVS.
...
It's not valid for non-away values to have a lane.
2015-01-03 16:54:31 -07:00
Henrik Rydgard
3cc633f091
x86jit: Minor logging improvements in FPU recache
2015-01-03 22:55:39 +01:00
Bovine
185d4db081
Fix simd vmmul transpose optimizations.
...
Need to ensure S has been written back before transposing it or we'll end
up writing back S'.
2015-01-03 14:48:54 -07:00
Unknown W. Brackets
ab978b1eb1
armjit: Fix vfad/vavg sign for -0.000.
...
But, NEON is still broken pretty bad, not sure why.
2015-01-03 11:14:52 -08:00
Henrik Rydgård
5fffe53671
Merge pull request #7258 from unknownbrackets/jit-minor
...
x86jit: Fix vavg, improve vfad
2015-01-03 19:44:22 +01:00
Unknown W. Brackets
e6a50a1371
Initialize vfpu regs are NAN, not 0.0.
...
This matches tests, the fpu, and seems logical.
2015-01-03 10:40:21 -08:00
Unknown W. Brackets
a1864b2c78
x86jit: Fix vavg scale.
...
Was completely wrong, darn it.
2015-01-03 10:39:58 -08:00
Unknown W. Brackets
8a089ee583
x86jit: Handle -0.0 in vavg/vfad.
...
Should result in a positive result.
2015-01-03 10:39:26 -08:00
Bovine
2ed4c1bd6b
Fix matrix disassembly notation
...
Changed disassembler output to match gas input for operations on
transposed matrices where row != col.
2015-01-03 02:04:08 -07:00
Henrik Rydgard
13c4a22967
Disable transpose optimization, seems to break games
2015-01-02 01:13:34 +01:00
Unknown W. Brackets
d0b6a3a339
Oops, fix a mistake in the screenshot func.
...
How did I lose the return?
2015-01-01 10:50:33 -08:00
Henrik Rydgard
6a7e5d1cc2
x86 Jit SIMD: Generate somewhat shorter code for handling transposed matrices in vmmul.
...
TODO: Build into regalloc instead, with a MapMatrix function?
2015-01-01 12:44:03 +01:00
Unknown W. Brackets
5f6f8ac0a2
x86jit: Fix another sequential detect problem.
...
Arg. Maybe should use voffset after all, but it won't fit in a u8
anyway...
2014-12-31 22:43:31 -08:00
Unknown W. Brackets
c5f7b6fa72
x86jit: Disable simd again for now.
2014-12-31 19:43:40 -08:00
