12444f6305
Compiles, runs, and hangs in normal wait.
2013-05-23 18:46:09 -07:00
1b833b6789
Removing libjit backend.
2013-05-23 18:05:03 -07:00
bedf86d259
Adding 'x64' backend.
...
Does not compile.
2013-05-23 18:04:12 -07:00
86002b2f44
Removing old llvm cruft.
2013-05-23 17:51:03 -07:00
63cee87420
Removing library loader, as it's not used.
2013-05-23 17:50:05 -07:00
57b486c4ad
Removing libjit, because it is pretty terrible.
2013-05-23 17:48:01 -07:00
7f4f448369
Adding beaengine.
2013-05-23 17:39:03 -07:00
4eafb6b070
Adding asmjit.
2013-05-23 17:30:13 -07:00
eee856be04
Tweaking things again. Now not getting very far. Sigh.
2013-05-23 15:52:57 -07:00
f08e80efbc
Fixing calling into code. Interpreter can now be used.
2013-05-23 15:44:45 -07:00
9487fa027c
libjit doesn't encode mov with imm64 offsets, fixing.
2013-05-23 13:55:09 -07:00
70a9a1053e
The previous truncation was useless.
2013-05-23 13:16:39 -07:00
89672e111d
Optimizing memory accesses a bit by letting libjit use relative addressing.
2013-05-23 11:46:26 -07:00
fc8eb3d40a
Adding disassembly to traces and bumping libjit.
2013-05-23 11:39:29 -07:00
730dba4aea
Adding addzex.
2013-05-23 02:13:04 -07:00
7d95098792
Fix for bswap16
2013-05-23 02:05:08 -07:00
d6f2a0b3f0
Adding --trace_branches.
...
Simple output right now, can always be enhanced.
2013-05-23 01:31:41 -07:00
4495637616
Making sure shim calls are cdecl.
2013-05-23 01:04:14 -07:00
426fccdbb4
Bumping libjit to fix the tail call bug.
2013-05-23 00:55:16 -07:00
6bde04de26
Flushing logging.
2013-05-23 00:26:55 -07:00
1a01d365b5
Making code dumps conditional on --log_codegen
2013-05-22 22:01:34 -07:00
fcb66660c0
(partial) addx and logging compiled addresses.
2013-05-22 21:53:21 -07:00
5cb8cf2630
Hacky ALU ops. Need overflow checks.
2013-05-22 21:14:16 -07:00
9ed911cdad
Fixing comparisons and disabling OutputDebugString logging.
2013-05-22 21:11:27 -07:00
2d01dced5a
Adding back invalid instruction logs.
2013-05-22 20:58:48 -07:00
f2c16c24f3
Basic control instructions (no indirection branches yet).
2013-05-22 20:15:04 -07:00
1bcb3e8f92
Many more ALU instructions.
2013-05-22 18:51:06 -07:00
64b619023d
Adding back in quite a few instructions.
2013-05-22 18:03:43 -07:00
c0dd60bde8
Adding memory access/etc.
...
Needs byte swapping.
2013-05-22 15:02:26 -07:00
7227ba2693
Setting up register fill/spill.
2013-05-22 13:37:44 -07:00
16d2b73a73
Main codegen loop now running through instructions.
2013-05-22 13:10:07 -07:00
dad99769cb
Prepping for basic block generation.
2013-05-22 11:53:52 -07:00
33e2bc01ee
Kernel call thunks and cleanup on exports.
2013-05-22 11:30:53 -07:00
37f93d2974
Calling to/from native works!
2013-05-22 11:13:27 -07:00
0c127828af
Disabling libjit warnings. Now down to 12.
2013-05-21 17:11:48 -07:00
65a8d3c039
libjit now uses beaengine for disassembly.
...
Fixes #20 .
2013-05-21 17:06:30 -07:00
81c76a854a
Fixing some libjit warnings.
2013-05-21 16:24:49 -07:00
a064e68621
Fixing more warnings.
2013-05-21 16:19:11 -07:00
720651ba9e
Removing wslay socket support in prep for switch to mongoose.
...
Work on #21 .
2013-05-21 16:16:33 -07:00
1ec06bd7d0
Tweaking compile options and switching back to VS2010.
2013-05-21 16:08:56 -07:00
85804f7415
Initial libjit skeleton.
...
Not currently generating instructions, but most of the flow is right up to
that point. A lot of work required to refactor the emitter, but wanted to
get this checked in.
2013-05-21 15:36:58 -07:00
94e3e592dd
Making disassembly much faster by moving string processing to only the
...
dump path.
2013-05-19 19:32:55 -07:00
d78f045123
libjit update.
2013-05-19 17:23:24 -07:00
7d93795c10
Fixing Windows compilation errors.
2013-05-19 09:29:31 -07:00
1ce68ca7fe
Removing LLVM submodule and build information.
...
Code is still present until transitioned over to libjit.
2013-05-19 09:29:22 -07:00
0aa7d1ec05
Adding libjit.
...
Removing some old xethunk cruft.
2013-05-19 09:23:50 -07:00
b018b6fe56
Broken, incomplete, but need to move forward with rewrite.
2013-04-21 12:34:20 -07:00
4987147055
Moving disasm functions to ppc/.
2013-03-30 02:01:33 -07:00
85bdbd24d1
In-progress work: refactoring PAL not to be instanced.
...
This removes a lot of useless passing around of the PAL object.
2013-03-29 05:07:32 -07:00
c46093266e
Tracking code low/high addresses per module.
2013-02-24 11:45:16 -08:00
Ben Vanik