mirrors/dolphin
0
0
Fork 0
mirror of https://github.com/dolphin-emu/dolphin.git synced 2025-10-14 12:09:09 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
12020 commits 11 branches 17 tags 1.7 GiB
Commit graph

51 commits

Author SHA1 Message Date
Tillmann Karras
d05e205a24 FPURoundMode: revert use of enums in bit-fields 
The workaround of using fixed underlying types produces lots of warnings
in GCC because now the bit-fields are too small for the value range used
for conversion semantics.
 2014-03-09 15:24:35 +01:00
Tillmann Karras
aa7a2fc4bf Jit64::twx(): add assert message 
The implemented and the real control flow differ a bit.
 2014-03-09 15:24:33 +01:00
Matthew Parlane
d9ca5c83c9 Fix bug in PrintCallstack caused by trying to use MMIO 
Also tidy up methods in Debugger_SymbolMap.cpp
 2014-03-09 12:23:30 +13:00
Pierre Bourdon
8802770a85 JitUtil: Optimize MMIO loads by generating code to directly access a given MMIO. 2014-03-08 23:32:43 +01:00
Matthew Parlane
e5b250fa79 Merge pull request #122 from lioncash/rename-default 
Rename the JIT function Default() to FallBackToInterpreter(). Communicates intent way better in terms of telling the reader what's going on.
 2014-03-06 12:13:54 +13:00
Shawn Hoffman
7733463e65 commit 1a428de189 introduced a bug by using a signed enum in a bitfield, the value of which is then used in a ldmxcsr instruction. The sign-extension corrupts the value, causing an exception by attempting to load mxcsr with an invalid value. 2014-03-05 10:19:29 -08:00
Ryan Houdek
4f02132f93 Make our architecture defines less stupid. 
Our defines were never clear between what meant 64bit or x86_64
This makes a clear cut between bitness and architecture.
This commit also has the side effect of bringing up aarch64 compiling support.
 2014-03-04 09:36:59 -06:00
Pierre Bourdon
91676ad930 Merge pull request #128 from lioncash/normalize-debug-interface-func-names 
Change the DebugInterface, PPCDebugInterface, and DSPDebugInterface to use CamelCase function names.
 2014-03-03 18:58:36 +01:00
Tillmann Karras
891b334f52 Fix 32 bit non-Windows build 2014-03-03 10:46:08 +01:00
Lioncash
279a8c0148 Change the DebugInterface, PPCDebugInterface, and DSPDebugInterface to use CamelCase names. 
This is the standard coding convention in the codebase, so our interfaces should use it too.
 2014-03-03 00:39:08 -05:00
Lioncash
34b5a7871f Rename the JIT function Default() to FallBackToInterpreter(). Communicates intent way better in terms of telling the reader what's going on. 
Also did some minor cleanup in functions that used Default.
 2014-03-02 19:59:20 -05:00
Lioncash
13a007abed Remove another clamp function laying in the codebase and replace it with the one in MathUtil.h. 2014-03-02 13:57:27 -05:00
Tillmann Karras
315a8ba1c0 Various changes suggested by cppcheck 
- remove unused variables
- reduce the scope where it makes sense
- correct limits (did you know that strcat()'s last parameter does not
  include the \0 that is always added?)
- set some free()'d pointers to NULL
 2014-02-28 12:43:20 +01:00
Tillmann Karras
1a428de189 x64FPURoundMode: move things around a bit 2014-02-28 12:28:21 +01:00
Tillmann Karras
6914eca167 Fix various warnings reported by clang 
- mostly remove unused variables
- rename some generic JIT identifiers
 2014-02-28 12:28:19 +01:00
Pierre Bourdon
99db9f0fb1 Merge pull request #94 from Tilka/uninitialized 
Fix struct initialization
 2014-02-23 17:07:24 +01:00
Lioncash
e3e78128a2 isSameValue in IR.cpp/.h should be returning a boolean value. 2014-02-23 04:14:57 -05:00
Tillmann Karras
3038bca0a4 JitBackpatch: initialize InstructionInfo to 0 
DisassembleMov() doesn't always initialize all fields of 'info'.
 2014-02-23 04:36:26 +01:00
Pierre Bourdon
311caef094 Merge pull request #25 from Tilka/ppc_fp 
Fix non-IEEE mode
 2014-02-23 04:15:37 +01:00
Tillmann Karras
ee21cbe2d1 Add phire's more accurate DoubleToSingle version 
This method doesn't involve messing around with the quirks of the x87
FPU and should be reasonably fast. As a bonus, it does the correct thing
for out-of-range doubles.

However, it is also a little slower and only benefits programs that rely
on undefined behavior so it is disabled for now.
 2014-02-23 04:13:47 +01:00
Pierre Bourdon
83b7bb64aa Make Common/ mostly IWYU clean (and fix errors in rest of the project detected by this change). 2014-02-22 23:37:29 +01:00
Lioncash
146b301a91 Fix more header sorting issues in Core/ (now check-includes clean). 2014-02-20 01:01:11 +01:00
Lioncash
2afe215271 Convert all includes to relative paths. 2014-02-18 02:19:10 -05:00
Lioncash
3fd87a7636 Second and final pass of clearing out tabs. 2014-02-17 02:19:41 -05:00
Lioncash
6c4ee1753a Fix some vertical alignments 
ie. uses spaces for alignment.
 2014-02-16 20:12:05 -05:00
Ryan Houdek
9b57292395 Fix ARM build 2014-02-15 22:05:12 -06:00
Ryan Houdek
77851edc1b Fix the arbitrary exits. We have to make sure that the block links are cleared entirely. 2014-02-15 20:41:42 -06:00
Ryan Houdek
0a9fd93eda On block unlinking, wipe the address from the valid links. 2014-02-15 20:23:01 -06:00
Ryan Houdek
d2a90e6eab If block linking isn't enabled then make sure there isn't a performance hit from getting a block number. 2014-02-15 20:23:00 -06:00
Ryan Houdek
c5b8c65ddc Initialize linkStatus to false in the ARM JIT cores as well. 2014-02-15 20:23:00 -06:00
Ryan Houdek
22b86e64f5 redo 'jit: change our linking module to be able to handle arbitrary exit addresses' 2014-02-15 20:22:59 -06:00
Lioncash
867d0f01a2 Remove #if 0 statement in lfd() in Jit64/Jit_LoadStoreFloating.cpp 2014-02-14 23:46:09 -05:00
Lioncash
49d6be26cd Remove dead else branch in StoreFromRegister() in Jit64/JitRegCache.cpp 2014-02-14 23:43:51 -05:00
Lioncash
d42e8817a0 Clean out some old left-behind quantizer stuff in Jit64/Jit_LoadStorePaired.cpp and Jit64/Jit_SystemRegisters.cpp. 2014-02-14 23:32:37 -05:00
Tillmann Karras
404624bf0b Turn loops into range-based form 
and some things suggested by cppcheck and compiler warnings.
 2014-02-13 09:05:50 +01:00
Tillmann Karras
2ff794d299 Fix some warnings 2014-02-13 09:02:43 +01:00
Scott Mansell
7062cf8657 Interpeter: Fixed ConvertToDouble to match the manual. 
Also added some documntation comments.
 2014-02-12 23:12:17 +01:00
Tillmann Karras
f6897039c7 Interpreter: fix float conversions 
Can't use simple casting, otherwise we get the same problems as in Jit64.
 2014-02-12 23:12:15 +01:00
Tillmann Karras
db196d8c5b Jit64[IL]: fix float conversions 
Floating-point is complicated...

Some background: Denormals are floats that are too close to zero to be
stored in a normalized way (their exponent would need more bits). Since
they are stored unnormalized, they are hard to work with, even in
hardware.  That's why both PowerPC and SSE can be configured to operate
in faster but non-standard-conpliant modes in which these numbers are
simply rounded ('flushed') to zero.

Internally, we do the same as the PowerPC CPU and store all floats in
double format. This means that for loading and storing singles we need a
conversion. The PowerPC CPU does this in hardware. We previously did
this using CVTSS2SD/CVTSD2SS. Unfortunately, these instructions are
considered arithmetic and therefore flush denormals to zero if non-IEEE
mode is active. This normally wouldn't be a problem since the next
arithmetic floating-point instruction would do the same anyway but as it
turns out some games actually use floating-point instructions for
copying arbitrary data.

My idea for fixing this problem was to use x87 instructions since the
x87 FPU never supported flush-to-zero and thus doesn't mangle denormals.
However, there is one more problem to deal with: SNaNs are automatically
converted to QNaNs (by setting the most-significant bit of the
fraction). I opted to fix this by manually resetting the QNaN bit of all
values with all-1s exponent.
 2014-02-12 23:12:15 +01:00
lioncash
d2038049f5 Replace all include guard ifdefs with "#pragma once" 2014-02-10 18:07:16 -05:00
Lioncash
ebb48d019e Clean up some struct indentations 
Also cleaned up the indentations of some variable declarations.
 2014-02-09 19:40:11 -05:00
Lioncash
40182a48a5 Cleanup enum indentations. 2014-02-09 16:16:10 -05:00
Matthew Parlane
70d2592ffb Fix warnings found by StringFromFormat having printf style checking. 2014-02-07 01:38:08 +13:00
Ryan Houdek
00089433a6 Merge pull request #31 from Sonicadvance1/slowmemfix 
[JITARM] Fixes an issue in the slowmem path that causes grey THP movies,...
 2014-02-04 12:55:43 -08:00
Ryan Houdek
57f86f7f52 [JITARM] Fixes an issue in the slowmem path that causes grey THP movies, and issues of games not working like Mario Kart being incapable of getting ingame. This was due to me forgetting an instruction. 2014-02-04 13:44:26 -06:00
Tillmann Karras
442796b681 Jit64IL: nits 2014-02-04 09:09:06 +01:00
Lioncash
89001accd8 Move two macros that literally doesn't change across all JIT platforms (JITDISABLE and INSTRUCTION_START) and simply have it in the JitBase.h source file. This way the macro is just located in one spot rather than 6 places. 2014-01-30 22:11:39 -05:00
Tillmann Karras
e94a6f07f4 Fix warning 2014-01-25 21:42:44 +01:00
skidau
8b53385c54 Added a check for a redundant mov in the fastmem writes code. 
Fixes issue 6905.
 2014-01-13 22:18:19 +11:00
Pierre Bourdon
45f74840d6 Do not fastmem addresses that were constant propagated 2014-01-09 08:28:45 +01:00