Retropex/dolphin
1
0
Fork 0
mirror of https://github.com/Retropex/dolphin.git synced 2025-05-18 14:20:42 +02:00
Code Issues Actions Packages Projects Releases Wiki Activity
20,510 Commits 11 Branches 32 Tags 698 MiB
28a3691e70
Commit Graph

1305 Commits

Author SHA1 Message Date
Matt Mastracci
1e08ad0d97 Add MOV optimization and MOV_sum 
Replaces incarnations of the A=B+C pattern throughout the
code so we can apply optimizations consistently.
 2016-06-27 15:43:52 -06:00
Matt Mastracci
b1296a7825 Refactor fastmem/trampoline code. 
Simplication to avoid reading back the generated instructions, allowing
us to handle all possible cases.
 2016-06-27 14:58:20 -06:00
Matt Mastracci
4aa5291f54 Refactor the paired load/store code 
Simplification/reduction of duplicated code. Detect other constant GQR values and inline loads (5-10% speedup) and do direct dispatch to AOT methods for stores.
 2016-06-26 19:49:02 -06:00
degasus
ca96302a36 PowerPC: Dedoublify CheckExternalExceptions. 2016-06-26 11:29:25 +02:00
Mat M
c896645313 Merge pull request #3947 from magumagu/interpreter-cleanup 
Misc cleanup in Interpreter.
 2016-06-26 02:11:09 -04:00
magumagu
1c28a27ee4 Delete dead variable named bFakeVMEM. 
(There's another bFakeVMEM in Memmap.h.)
 2016-06-25 23:05:55 -07:00
magumagu
b5ed27ffa8 Misc cleanup in Interpreter. 
Mostly comment changes; also gets rid of an unnecessary call to
CheckExceptions.
 2016-06-25 22:52:17 -07:00
Rohit Nirmal
84cd57f99b Fix building with PCH disabled. 2016-06-24 11:46:48 -05:00
Pierre Bourdon
5fcb4bb3ab Further fixes to the formatting change. WX sucks. 2016-06-24 12:16:10 +02:00
Pierre Bourdon
3570c7f03a Reformat all the things. Have fun with merge conflicts. 2016-06-24 10:43:46 +02:00
Pierre Bourdon
2115e8a4a6 Add annotations for code sections that need to remain manually formatted. 2016-06-24 10:41:46 +02:00
Mat M
08d45b9fea Merge pull request #3794 from EmptyChaos/frame-advance-race 
Core: Add synchronization to state changes (Fix Frame Step and FIFO Player -  Issue 8718)
 2016-05-22 15:19:16 -04:00
degasus
c5e8238cc7 JitArm64: Fix dispatcher with MMU. 2016-05-20 08:32:16 +02:00
degasus
8e902abca3 JitArm64: Implement DSI exception. 
Ok, this falls back on all memory instructions, but it's a way to start.
 2016-05-20 08:32:16 +02:00
degasus
5db61b4bbb JitArm64: Fix broken block handling. 2016-05-19 19:20:44 +02:00
degasus
37e2e92e33 JitArm64: Handle ISI excepction. 2016-05-19 19:20:44 +02:00
degasus
95ebb2991f JitArm64: Check Exceptions on mtmsr. 2016-05-19 19:20:44 +02:00
degasus
782c68cf00 JitArm64: Clean up LWZ idle skipping. 2016-05-19 19:20:44 +02:00
degasus
637851fce0 JitArm64: optimize bclrx 
the unconditional path is the most common, so no need to switch to far code here.
 2016-05-14 00:05:55 +02:00
degasus
b6f16352e4 JitArm64: Check downcount on block linking. 
This skips one B() call.
 2016-05-14 00:04:58 +02:00
degasus
10e716c623 JitCommon: Pass jit block instead of host pointer on linking. 
So the JIT may do more fancy stuff.
 2016-05-14 00:04:58 +02:00
degasus
cf3c65fbd0 JitArm64: Use B() instead of BR() to jumo to ASM. 
Avoid indirect jumps as good as possible. This is a noticeable speedup.
 2016-05-14 00:04:58 +02:00
degasus
b2be9bd7f7 JitArm64: Inline JitAsm in JitArm64. 
So they share the same emitter, and so they are in the same 128MB range.
This allows us to use B() to jump to the dispatcher.

However, so we have to regenerate them on every cache clear.
 2016-05-14 00:04:58 +02:00
EmptyChaos
c1922783f8 Core: Threadsafety Synchronization Fixes (Frame Advance / FifoPlayer) 
Fix Frame Advance and FifoPlayer pause/unpause/stop.

CPU::EnableStepping is not atomic but is called from multiple threads
which races and leaves the system in a random state; also instruction
stepping was unstable, m_StepEvent had an almost random value because
of the dual purpose it served which could cause races where CPU::Run
would SingleStep when it was supposed to be sleeping.

FifoPlayer never FinishStateMove()d which was causing it to deadlock.
Rather than partially reimplementing CPU::Run, just use CPUCoreBase
and then call CPU::Run(). More DRY and less likely to have weird bugs
specific to the player (i.e the previous freezing on pause/stop).

Refactor PowerPC::state into CPU since it manages the state of the
CPU Thread which is controlled by CPU, not PowerPC. This simplifies
the architecture somewhat and eliminates races that can be caused by
calling PowerPC state functions directly instead of using CPU's
(because they bypassed the EnableStepping lock).
 2016-05-13 09:23:44 +10:00
Matthew Parlane
ea82fcc24f Merge pull request #3658 from Tilka/helpers 
Interpreter: simplify some helpers
 2016-05-03 22:44:27 +12:00
Mathew Maidment
b99b685f22 Merge pull request #3800 from phire/unexceptional-exceptions 
Make exceptions consistent across all JITs/Interpreters (Fixes Pokemon Box)
 2016-04-29 19:48:55 -04:00
Mathew Maidment
b420abc908 Merge pull request #3769 from RisingFog/remove_framestep 
Remove Unused Framestop Functions
 2016-04-29 19:41:25 -04:00
Scott Mansell
eb5819f88a Make exceptions consistant across all JITs/Interpeters. 
They all handled it diffrently, so I've just moved it into Advance()

This fixes Pokemon Box booting in JIT/JITIL which shared a bug where
exceptions set in a scheduled event would be ignored untill the next
slice (upto 20,000 cycles).
 2016-04-28 17:22:52 +12:00
Chris Burgener
3140f9b226 Change GetStatePtr() Return to const 2016-04-05 12:45:48 -04:00
Pierre Bourdon
21eb1cd158 Merge pull request #3635 from phire/lazy_comex 
Implement BLR Overflow handling for Windows.
 2016-03-26 01:42:54 +01:00
Scott Mansell
d61baef2f6 Disable JitArm64's inline timebase implemenation, as it's incorrect. 
The interpeted version is correct.
 2016-03-24 05:17:10 +13:00
Scott Mansell
27beef1ff4 Store an inverted copy of lastOCfactor. 
The inverse operation is more common, especially when games check the
timer rapidly. So we do the division once and store the inverted copy.
 2016-03-24 05:17:10 +13:00
Scott Mansell
407f86e01a Mark global variables with g_ prefix 2016-03-24 04:32:12 +13:00
Scott Mansell
2ebbfd6f85 Adjust cycle counts so they are accurate to the jit block level 
Previously GlobalTimer was only updated at the end of each slice
when CoreTiming::Advance() was called, so it could be upto 20,000
cycles off.

This was causing huge problems with games which made heavy use of
the time base register, such as OoT (virtual console) and Pokemon
puzzle.

I've also made it so event scheduling will be accurate to the jit
block level, instead of accurate to the slice.
 2016-03-24 04:24:53 +13:00
Scott Mansell
c5b3a2efac Implement BLR Overflow handling for Windows. 2016-03-20 00:41:28 +13:00
degasus
54a4d68c9a JitArm64: implement dcbx 2016-03-05 00:29:38 +01:00
degasus
9bf50fb93d JitArm64: Fix mulli. 
mulli has no inst.OE
 2016-03-04 22:51:47 +01:00
degasus
9ed465f4ac JitArm64: Implement mulhwx 2016-03-04 22:51:46 +01:00
degasus
889a0d396d JitArm64: Align farcode entry points to 16 bytes 2016-03-04 22:49:02 +01:00
degasus
eaa0e275db JitArm64: Use immediate WriteExceptionExit more often. 2016-03-04 22:49:01 +01:00
degasus
689205f0a4 JitArm64: small cleanup, no changes 2016-03-04 22:49:01 +01:00
degasus
7b017c6a65 JitArm64: Use a register as PC argument for the dispatcher. 2016-03-04 19:50:12 +01:00
degasus
050932ed23 JitArm64: Rename global registers. 2016-03-04 19:50:12 +01:00
degasus
57d76cefb0 JitArm64: Merge WriteExceptionExit. 2016-03-04 19:50:12 +01:00
degasus
55b9ce9b5b JitArm64: Store memory pointer once, not per block. 2016-03-04 19:50:12 +01:00
degasus
56c8f65bc2 JitArm64: Precheck for exceptions before calling PowerPC::CheckExceptions. 2016-03-04 19:50:12 +01:00
degasus
aea48c4591 JitArm64: Drop unused WriteExceptionExit. 2016-03-04 19:50:11 +01:00
degasus
c5b0dc98c2 JitArm64: Fix LWZ idle skipping. 2016-03-04 19:50:11 +01:00
Ryan Houdek
36b476ed76 Merge pull request #3683 from Sonicadvance1/aarch64_fix_fcmp 
[AArch64] Fix fcmp.
 2016-03-01 09:29:56 -05:00
degasus
e2725792d4 JitArm64: Optimize addi 2016-02-29 08:39:24 +01:00