Commit Graph

14468 Commits

Author SHA1 Message Date
Ryan Houdek 7cc586d615 Merge pull request #1100 from FioraAeterna/psq_insts
JIT: implement remaining psq_l/st instruction variants
2014-09-19 15:16:44 -05:00
Fiora 0f53bba45a JIT: change eieio and tlbsync to DoNothing
The interpreter functions for these are no-ops anyways.
Also add some missing DoNothings to the ARM64 JIT.
2014-09-19 13:14:49 -07:00
Ryan Houdek 522d7eb275 Merge pull request #1109 from FioraAeterna/ps_cmp
JIT: add ps_cmp0/ps_cmp1/ps_res/ps_rsqrte
2014-09-19 14:41:05 -05:00
Ryan Houdek 526e92464b Merge pull request #1103 from FioraAeterna/floatconstnaming
JIT: rename a few constants to be more accurate
2014-09-19 13:21:06 -05:00
Ryan Houdek af25b15348 Merge pull request #977 from Sonicadvance1/xfb-scaling-rebased
XFB Scaling
2014-09-19 12:54:19 -05:00
Ryan Houdek c5f9301e6e Add a comment to the software renderer that stride should be implemented 2014-09-19 12:33:15 -05:00
magumagu 32e5043b29 WIP XFB scaling.
Still an ugly mess.
2014-09-19 12:33:15 -05:00
Ryan Houdek 0d1c1e9477 Merge pull request #1122 from Sonicadvance1/Android-reduce-confusion
Show OpenGL or OpenGL ES on Android depending on what the device supports
2014-09-19 09:21:58 -05:00
Ryan Houdek e18860830a Show OpenGL or OpenGL ES on Android depending on what the device supports.
Reduces confusion for the users
2014-09-19 09:21:16 -05:00
Ryan Houdek d544c563ea Merge pull request #1112 from Sonicadvance1/AArch64-mov-aliases
Add AArch64 emitter aliases for MOV and MVN.
2014-09-19 09:19:33 -05:00
Ryan Houdek ec310811a8 Merge pull request #1114 from Sonicadvance1/AArch64-fix-build
Fix AArch64 JIT compiling.
2014-09-19 09:19:25 -05:00
Fiora c130a496f2 JIT: fix fsel/ps_sel implementations for NaN input
fselx was the main problem, but ps_sel was wrong too (even if there were no
known reported bugs with it).

This fixes Beyond Good and Evil (at the least).
2014-09-19 01:58:13 -07:00
Fiora d96016ed21 JIT: support byte-reversed stores/loads
4 more instructions down.

Store ones should be pretty well-tested; load ones seem to almost never be
used. I found them in Turok Evolution, so I was able to check code generation,
but the relevant code didn't seem to be called.
2014-09-18 20:38:44 -07:00
skidau bd740ae9c7 Merge pull request #1111 from Tilka/fix_gcpad_setmotor
Fix GCPad::SetMotor()
2014-09-19 13:26:05 +10:00
Fiora af8772700a JIT: add ps_res and ps_rsqrte using scalar implementations 2014-09-18 18:36:15 -07:00
Lioncash c33a9de6af FrameTools: Set focus on Windows via CFrame::SetFocus 2014-09-18 21:15:26 -04:00
Fiora 1b425dedd1 JIT: add ps_cmp0/ps_cmp1 implementations using current fcmp code 2014-09-18 17:57:28 -07:00
Fiora 3c49200b22 X64Emitter: add MOVHLPS/MOVLHPS 2014-09-18 17:57:27 -07:00
Ryan Houdek a8abbdae85 Merge pull request #1106 from FioraAeterna/fixdebug2
JIT: fix debug mode
2014-09-18 19:52:35 -05:00
Lioncash 68d62c3d32 Merge pull request #1110 from lioncash/apple-code
DolphinWX: Fix GC/Wiimote input windows randomly crashing on OSX
2014-09-18 20:17:16 -04:00
shuffle2 88196d373f Merge pull request #1117 from shuffle2/qt-plugins
windows/qt: add (hopefully) all plugins DolphinQt should possibly need.
2014-09-18 15:46:28 -07:00
shuffle2 fcf1016827 Merge pull request #1116 from comex/fix-blr-optimization-again-simpler
Reset RSP after calling Jit in case it cleared the code cache.
2014-09-18 15:46:10 -07:00
Shawn Hoffman 035a1c0ec2 windows/qt: add (hopefully) all plugins DolphinQt should possibly need.
For now they are all copied to binary directory, once the project is more mature
it can be stripped down.
2014-09-18 15:24:14 -07:00
comex 3df935b98e Reset RSP after calling Jit in case it cleared the code cache. 2014-09-18 18:23:36 -04:00
Ryan Houdek f709dda6aa Fix AArch64 JIT compiling. 2014-09-18 16:50:49 -05:00
Ryan Houdek 7608e3f11e Add AArch64 emitter aliases for MOV and MVN. 2014-09-18 16:30:40 -05:00
Tillmann Karras e35db54454 Fix and simplify GCPad::SetMotor()
abs() takes an int argument. Casting -0.5..0.5 to int always resulted in
zero.
2014-09-18 21:34:07 +02:00
Lioncash 207d7787a4 DolphinWX: Fix GC/Wiimote input windows randomly crashing on OSX 2014-09-18 14:52:12 -04:00
Ryan Houdek e9164247d6 Merge pull request #1108 from FioraAeterna/fixdebugscroll
Debugger: scroll by multiples of 4 bytes
2014-09-18 06:52:44 -05:00
Ryan Houdek d54c8eb6bf Merge pull request #1098 from FioraAeterna/cvtsi2ss
JIT: use cvtsi2ss in paired singles
2014-09-18 06:51:37 -05:00
Ryan Houdek 0294b344e2 Merge pull request #1086 from FioraAeterna/fixsrawint
Interpreter: fix carry calculation in srawx
2014-09-18 06:41:37 -05:00
Fiora 20c3a0f2d8 Debugger: scroll by multiples of 4 bytes
Avoids that weird effect where scrolling offsets code from 4-byte boundaries,
showing nonsense 75% of the time.
2014-09-18 03:54:57 -07:00
Fiora 8ce4676605 Debugger: make opcode search a bit better
Search a wider range (not all games fit in the originally searched range).
Print a notice if the opcode isn't found, instead of silently failing.
2014-09-18 03:47:46 -07:00
Fiora cd7853bd50 JIT: fix debug mode 2014-09-18 03:14:04 -07:00
skidau 74eee70fad Merge pull request #1105 from FioraAeterna/pssel
JIT: fix regression in ps_sel
2014-09-18 14:55:47 +10:00
Fiora 9ab816e6e9 JIT: fix regression in ps_sel
My code didn't maintain correct semantics with floating-point NaNs (a < b is
not the same as "not a >= b" in float), which seems to have broken FIFA 12.
2014-09-17 21:43:44 -07:00
shuffle2 8b151d6a61 Merge pull request #1104 from comex/its-like-the-bad-old-days-where-i-broke-master-over-and-over
Fix BLR optimization (sorry)
2014-09-17 19:38:28 -07:00
comex 217758b607 Correct inaccurate comment. 2014-09-17 22:30:33 -04:00
comex 6695b5acce Fix backwards #ifdef. 2014-09-17 22:30:20 -04:00
comex 97c9cb5882 Add missing push wrapper around UpdatePerformanceMonitor 2014-09-17 21:10:43 -04:00
comex 49a48a6057 Merge pull request #1025 from comex/blr-optimization
Opportunistically predict BLR destinations using RET.
2014-09-17 20:26:57 -04:00
Fiora 29fc151cc7 JIT: implement remaining psq_l/st instruction variants
These are used in at least a few popular games, including Super Smash Bros.
Brawl.
2014-09-17 17:10:27 -07:00
comex 7ad9027593 Be pedantic about stack overflow on Linux and OS X.
Add some magic to the fault handler to handle stack overflow due to BLR
optimization, and disable the optimization if fastmem is not enabled.
2014-09-17 20:08:09 -04:00
Fiora cfd65c2762 JIT: rename a few constants to be more accurate
Remove the '2'; it's not longer applicable.
2014-09-17 16:59:35 -07:00
comex 755bd2c445 Reorganize backpatching a bit. Untested on ARM.
Rather than *MemTools.cpp checking whether the address is in the
emulated range itself (which, as of the next commit, doesn't cover every
kind of access the JIT might want to intercept) and doing PC
replacement, they just pass the access address and context to
jit->HandleFault, which does the rest itself.

Because SContext is now in JitInterface, I wanted JitBackpatch.h (which
defines it) to be lightweight, so I moved TrampolineCache and associated
x64{Analyzer,Emitter} dependencies into its own file.  I hate adding new
files in three places, two of which are MSVC...

While I'm at it, edit a misleading comment.
2014-09-17 19:57:06 -04:00
comex 7b0fdb52cd Run exception handlers on an alternate stack on Linux.
*Completely untested.*  Someone please test.
2014-09-17 19:57:04 -04:00
comex bd4e75e69a Shorten the blr stub a bit. 2014-09-17 19:56:58 -04:00
comex b597ec3e08 Opportunistically predict BLR destinations using RET.
When executing a BL-type instruction, push the new LR onto the stack,
then CALL the dispatcher or linked block rather than JMPing to it.  When
executing BLR, compare [rsp+8] to LR, and RET if it's right, which it
usually will be unless the thread was switched out.  If it's not right,
reset RSP to avoid overflow.

This both saves a trip through the dispatcher and improves branch
prediction.

There is a small possibility of stack overflow anyway, which should
be handled... *yawn*
2014-09-17 19:56:09 -04:00
comex 558dee84ca Wrap some function calls in ABI_Push|PopRegistersAndAdjustStack(0, 0);
These calls are made outside of JIT blocks, and thus previously did not
read any protection - register use is taken into account and the outer
dispatcher stack frame is sufficient.  However, if data is to be stored
on the stack, these calls must reserve stack shadow space on Windows to
avoid clobbering it.
2014-09-17 19:56:09 -04:00
Fiora 2a0b06f08c JIT: simpler fallback conditions for load/store float
I wasn't really thinking when I wrote these; d/s are for float registers, not
gprs.
2014-09-17 10:27:33 -07:00