Commit graph

1398 commits

Author SHA1 Message Date
Léo Lam
39fd6dcd5b Fix missing includes
Aren't indirect includes great?
2016-10-07 23:46:41 +02:00
Shawn Hoffman
fead9a05bc Fix static_assert raised by commit 789975e350 2016-10-04 00:56:02 -07:00
shuffle2
ab01dec178 Merge pull request #3806 from degasus/idle
ConfigManager: Drop SkipIdle option.
2016-10-03 23:59:38 -07:00
Markus Wick
cf3513f7fc Merge pull request #4123 from hthh/improve-const-stores
Jit: FIFO optimization improvements
2016-10-04 08:32:48 +02:00
degasus
40aa54447d Jit64: Drop very outdated comment about idle skipping. 2016-10-04 08:31:43 +02:00
degasus
0c67297ea0 ConfigManager: Drop SkipIdle option.
This option is safe + deterministic, so let's always enable it.
2016-10-04 08:31:43 +02:00
Shawn Hoffman
f000bb7590 Jit64: use domain-appropriate instructions
Original commit:
commit 62d3423ccf
Author: Tillmann Karras <tilkax@gmail.com>
Date:   Wed Nov 18 20:32:20 2015 +0100
2016-10-03 07:15:26 -07:00
degasus
f7e07f8f3b MMU: Don't generate virtual memory area on 32bit builds.
There is no JIT, so no need for fastmem. This memory won't be used at all.
2016-10-03 10:26:23 +02:00
degasus
0de1c6c99c MMU: Sort physical access by common access pattern. 2016-10-03 10:26:23 +02:00
degasus
d1488d0dc4 MMU: Drop force inline usage.
This is not needed, the compiler will do it on it's own if required.
2016-10-03 10:26:19 +02:00
degasus
4b3e388bd7 Jit64: Fix merge failure of dynamic_bat. 2016-10-03 10:15:52 +02:00
degasus
ebc47a4b36 Memmap: Drop redundant global bFakeVMEM flag. 2016-10-03 10:15:52 +02:00
shuffle2
fba6801851 Merge pull request #4216 from EmptyChaos/geckocodes-cleanup
GeckoCodes: Don't run PPC code in CoreTiming callbacks
2016-10-02 22:37:51 -07:00
EmptyChaos
83407263e5 HLE/GeckoCode: Add new HLE hook exit trampoline
Dolphin emulates GeckoCodes by fiddling with the CPU state when a
VI Interrupt occurs. The problem with this is that we don't know
where the PC is so it's non-deterministic and not necessarily
suitable for use with the codehandler.

There are two options: Patch the game like Gecko OS either directly
or using HLE::Patch, or use a trampoline so we can branch from any
PC even if it would otherwise not be valid. The problem with Gecko OS
patches is there are 10 of them and they have to be configured
manually (i.e. Game INIs to would need to have a [Core]GeckoHookType
property).

HLE_Misc::GeckoReturnTrampoline enables the Code Handler to be
entered from anywhere, the trampoline restores all the registers that
had to be secretly saved to the stack.
2016-10-03 16:27:42 +11:00
shuffle2
bd1218a3c4 Merge pull request #4250 from leoetlino/hle-patch-fix
HLE_OS: Minor fixes (function patching, output encoding)
2016-10-02 22:13:15 -07:00
shuffle2
456d869791 Merge pull request #4200 from EmptyChaos/jit64-dispatch-bug
Jit64: Fix Dispatcher ABI error
2016-10-02 20:41:39 -07:00
Mat M
ccfc081697 Merge pull request #4245 from aldelaro5/logs-levels-changes
Lots of Logs levels changes (also enable INFO level in every build)
2016-10-02 16:51:44 -04:00
aldelaro5
f0aa9b3751 Reorganise a ton of logs level
Most of this commits changes performance decreasing logs from info to debug and also cleans up innacurate levels.
2016-10-01 15:50:28 -04:00
Lioncash
91bbe56f59 SystemRegisters: Get rid of pointer casting 2016-09-30 08:28:07 -04:00
Lioncash
6b58ebfadc MMU: Remove useless casts 2016-09-30 06:48:13 -04:00
Lioncash
e8b5e38d98 MMU: Get rid of type punning in FP read/write functions
The previous code is actually considered undefined behavior.
2016-09-30 03:50:08 -04:00
Lioncash
e1705ba6c5 PPCAnalyst: Get rid of two casts
HostRead_Instruction does the same thing behind the scenes without casts.
2016-09-29 18:17:56 -04:00
Lioncash
ba628b3cab PPCAnalyst: Compress loop into std::any_of 2016-09-28 17:12:21 -04:00
EmptyChaos
f9a88adddc PowerPC: Fix Dynamic BAT savestates 2016-09-28 14:26:26 +10:00
Markus Wick
3696c2b022 Merge pull request #4210 from degasus/arm
JitArm64: Small cleanup + speedups.
2016-09-27 18:45:14 +02:00
Lioncash
082275d785 PPCAnalyst: Convert #define into a constant 2016-09-26 20:04:17 -04:00
Lioncash
a947391556 PPCAnalyst: Make local constants constexpr 2016-09-26 19:51:48 -04:00
Léo Lam
cb73bcc72e PPCSymbolDB: Drop useless const qualifier 2016-09-26 22:57:40 +02:00
Léo Lam
303325768b SymbolDB: Only match against the function name
This changes GetSymbolFromName to not require the passed name to
completely match with the symbol name. Instead, we now match
against the stripped symbol name (i.e. only the function name).

This fixes a regression introduced by #4160, which prevented
HLE::PatchFunctions() from working properly.
2016-09-26 22:56:42 +02:00
degasus
11bfc7fe77 JitArm64: Move memcheck check into fallback code.
So now each implemented slowmem instruction should check for DSI on each own.
2016-09-26 22:17:25 +02:00
degasus
39c08694a0 JitArm64: Optimize addic.
If simm < 0, it's faster to call MOVI2R(-simm) as only the lower bits are set.
2016-09-26 22:17:25 +02:00
degasus
1f94abea18 JitArm64: Optimize logic immediate instructions.
Try to use also the immediate instruction on ARM.
2016-09-26 22:17:25 +02:00
Mat M
9e1cfbee0e Merge pull request #4160 from sepalani/symbolmap
Map: Prevents symbols from being stripped
2016-09-18 22:52:12 -04:00
Scott Mansell
514ce3c6ed Merge pull request #4201 from EmptyChaos/interpreter-coretiming
Interpreter/CachedInterpreter/JitArm64: Fix CoreTiming::Advance usage
2016-09-15 09:57:00 +12:00
Sepalani
78d50c9a31 PPCSymbolDB: Rename unknown variable 2016-09-14 20:00:15 +02:00
Scott Mansell
ea7e734496 Merge pull request #4186 from aldelaro5/registerView-update-fix
Fix registerView updates issues when stepping over or stepping out
2016-09-14 18:10:41 +12:00
EmptyChaos
070fe7f499 Jit64: Fix Dispatcher ABI (downcount) violation
The dispatcher requires that the flags contain the downcount state.
2016-09-14 15:46:54 +10:00
EmptyChaos
55a7f576aa JitArm64: Fix CoreTiming contract
Call Advance at the start of each timing slice instead of the end.

Possibly fixed a bug where a slice would randomly branch straight
back to Advance() because the flags were not set to the right
values before branching to the dispatcher entrypoint.
2016-09-14 15:44:51 +10:00
EmptyChaos
f5bfce657c CachedInterpreter: Fix CoreTiming contract
Call CoreTiming::Advance() before executing the slice, not after.
2016-09-14 15:44:51 +10:00
EmptyChaos
1bcd129683 Interpreter: Fix CoreTiming contract
The interpreter does not use CoreTiming correctly. Calls to Advance
must be made in advance of executing the associated slice, not
afterwards.
2016-09-14 15:44:51 +10:00
Lioncash
00ddbee786 SymbolDB: Change Symbol's 'analyzed' member into a boolean
It's only ever used as one
2016-09-13 21:23:28 -04:00
Lioncash
0ac77b0288 SymbolDB: Use an enum class for representing symbol type 2016-09-13 21:21:03 -04:00
Markus Wick
bdcee1c585 Merge pull request #4205 from degasus/arm
JitArm64: Cleanup + small fix.
2016-09-12 11:47:51 +02:00
degasus
a0dcebba37 JitArm64: replace MOVI2R with MOVP2R when possible.
This skips the (u64) cast.
2016-09-10 11:01:18 +02:00
degasus
789e776a47 JitArm64: Fix block linking.
Double negation fail, sorry.
2016-09-10 11:01:18 +02:00
degasus
be763949bc JitArm64: Drop 32bit alignment restriction of the memarea.
ADD() is almost as fast as MOVK(), and we already have a register for the memory location. Let's just use it.
2016-09-10 11:01:18 +02:00
aldelaro5
09df343247 Move Memchecks support out of debug only builds
It wouldn't impact performance until at least one memcheck is enabled.  Because of this, it can be used in release builds without much impact, the only thing that woudl change is the use of HasAny method instead of preprocessor conditionals.  Since the perforamnce decrease comes right when the first memcheck is added and restored when the last is removed, it basically is all beneficial and works the same way.
2016-09-09 15:05:54 -04:00
Scott Mansell
fe41dff228 Fix the master build. 2016-09-07 20:10:03 +12:00
Scott Mansell
d951f9e765 Merge pull request #3771 from phire/pr3771
Remove instant ARAM DMA hack (Fixes Sarge's War)
2016-09-07 19:09:39 +12:00
Scott Mansell
7311cd8b6e Merge pull request #4137 from aldelaro5/memCheck-crash-fix
Fix a crash when adding the first memory check while the game is running
2016-09-07 14:19:19 +12:00