gibbed
d730784efb
Added XBDM stubs.
2017-01-09 04:51:55 -06:00
Dr. Chat
691c3d1021
EXTRACT_I32: Support constant src1
2016-11-11 16:50:42 -06:00
Dr. Chat
a148b965f1
KernelState should handle module launching
2016-10-24 11:01:10 -05:00
Dr. Chat
52c75c8dbc
x64 backend: Pass guest return address as first argument to guest functions.
2016-08-26 13:46:21 -05:00
Dr. Chat
7683069f6c
Set context/membase registers to rsi/rdi respectively.
...
Use offsetof for the backend thunks.
2016-08-26 11:00:42 -05:00
Dr. Chat
5f4416ee2f
x64 backend: Save nonvolatile XMM registers on host -> guest transitions
...
Define the context and membase registers in the x64 emitter.
2016-08-22 14:55:16 -05:00
Dr. Chat
e3fdb08ad7
x64 JIT: Allocate constant data outside of the guest heap.
2016-08-18 18:26:55 -05:00
Dr. Chat
22794902f3
CNTLZ: Invert the result using xor rather than subtraction
2016-08-01 13:25:44 -05:00
Dr. Chat
a72f1d949f
Constant propagation for vector SGT/UGE/UGT
...
Constant rounding (and fixed vec128 rounding only doing x)
2016-07-25 11:57:14 -05:00
Dr. Chat
8427acfada
Use vpminud to saturate rather than bitwise ops (shaves 6 instructions per 16 in 32 saturated pack)
2016-07-23 19:19:40 -05:00
Dr. Chat
2d55b12cc9
Fix a race condition in x64 code cache involving RtlGrowFunctionTable
2016-07-19 13:59:40 -05:00
Dr. Chat
18ff97e6b6
OPCODE_IS_NAN
2016-07-11 20:44:12 -05:00
Dr. Chat
3d1d4dea47
OPCODE_DOT_PRODUCT_4 constant propagation
2016-06-28 19:39:22 -05:00
Dr. Chat
f2ad6b8cb8
OPCODE_RECIP
2016-06-26 13:03:42 -05:00
gibbed
8965fda402
RSqrt for VEC128_TYPE.
2016-06-20 13:31:11 -05:00
gibbed
53e37c3167
Constant propogation for DOT_PRODUCT_3.
2016-06-20 13:30:19 -05:00
gibbed
450ca87120
Added comparison operator for --break_on_instruction.
2016-06-19 21:01:18 -05:00
gibbed
b5bef8c54d
Undoing the previous commit in favor of finding a better solution (xmm1/xmm2 can't safely be used here).
2016-06-19 02:33:52 -05:00
gibbed
d04a6cc983
Added support for when both src1 and src2 are constants in EmitCommutativeBinaryXmmOp and EmitAssociativeBinaryXmmOp.
2016-06-19 02:10:40 -05:00
Dr. Chat
d5865740e2
Add a few more vector tests
2016-06-12 21:45:51 -05:00
Dr. Chat
0ef16b10e2
Constant Propagation for OPCODE_VECTOR_ADD
2016-06-12 21:41:02 -05:00
Dr. Chat
841f6fe239
Merge branch 'vulkan'
2016-06-11 19:23:23 -05:00
Dr. Chat
32011a841e
Formatting.
2016-06-05 17:48:06 -05:00
Haydn Trigg
ff5529714c
clang-format to Google Standards
...
Added formatting to code
2016-05-29 13:22:14 +09:30
Haydn Trigg
f2ca689ea0
Expanded vec128 and VectorSub
...
Added double precision (f64) values to the vec128 structure
Added cases for all formats of data for vector subtraction in the
Value::VectorSub function
NOTE: Unsure what the saturate function is for however maintained
original functionality
2016-05-17 05:34:09 +09:30
Dr. Chat
0e3c113375
Physical write watches -> access watches (read and/or write watching)
2016-03-17 21:55:16 -05:00
Ben Vanik
06ba273492
Warn and be ok with symbol services failing.
2016-02-21 14:26:16 -08:00
Ben Vanik
05107d2d3e
Skeleton xenia::gpu::vulkan implementation, enough to start trace viewer.
2016-02-18 16:43:41 -08:00
sephiroth99
5d9a4b6830
Handle all other cases in switch statements
2016-01-20 11:00:37 -05:00
sephiroth99
4c0f9e6342
Add missing includes
2016-01-20 10:36:01 -05:00
Ben Vanik
6777ce6668
Reconcile debugger and save state stuff into a single implementation.
...
Fixes #497 and fixes #496 .
Still rough edges, but at least less duplication.
2016-01-18 11:48:21 -08:00
Ben Vanik
9473f20c7f
Fixing travis lint diff.
2016-01-13 22:55:28 -08:00
Ben Vanik
2e826f9321
Supporting constants in Haswell codepaths for mul_add/sub.
2016-01-13 22:29:39 -08:00
Dr. Chat
0771eea0b3
Change comments to comply with local regulations.
2016-01-13 19:41:15 -06:00
Dr. Chat
bd6bf16bd1
New flag: disable_global_lock - Disables global lock usage in guest code.
2016-01-13 19:40:00 -06:00
Dr. Chat
14c5e48120
Enable speculative PACK_TYPE_FLOAT16_3
2016-01-13 01:48:18 -06:00
Dr. Chat
36372d1eae
Enforce xex memory protections
2016-01-13 01:47:06 -06:00
Dr. Chat
f0802d75fa
Use atomic compare exchange in stdcx/stwcx (still under the global lock)
2016-01-13 01:45:31 -06:00
Dr. Chat
51dde62f09
x64 backend vector optimizations
2016-01-13 01:43:23 -06:00
Dr. Chat
7b6f1388d6
OPCODE_ATOMIC_COMPARE_EXCHANGE
2016-01-13 01:41:58 -06:00
Dr. Chat
3ccb2a978d
Propagate vector constants.
2016-01-13 01:40:44 -06:00
Dr. Chat
6d2f083b09
Up the maximum function count to 50,000
2016-01-13 01:30:15 -06:00
Ben Vanik
86b706d87c
Merge pull request #509 from x1nixmzeng/xex-resources
...
Rebrand the Xenia window with module metadata
2016-01-10 08:55:47 -08:00
x1nixmzeng
3eb602c93a
Assert library name index is in range of the string table
...
As suggested by @DrChat
2016-01-09 17:55:57 +00:00
x1nixmzeng
6530b9dc49
Added callback when new module is launched
...
This avoids having to guess the display window title format
Also manually fixed the linting errors picked out by travis which do not
get picked up using xb format locally
2016-01-09 17:36:46 +00:00
x1nixmzeng
06f259c87d
Mask the name index instead
...
The maximum number of import libraries (32) is already asserted
2016-01-09 02:43:29 +00:00
x1nixmzeng
7e4fdf7669
Fixed rare crash when accessing library import by name
...
The library name index may only be 8-bits.
This bug was not present in the previous implementation due to a bitmask
(0xFF) - see
ea99ba8e3b/src/xenia/kernel/util/xex2.cc (L272)
2016-01-09 01:19:05 +00:00
Ben Vanik
a95be25e2f
Merge branch 'linuxfix' of https://github.com/sephiroth99/xenia into sephiroth99-linuxfix
2015-12-31 14:39:23 -08:00
Ben Vanik
952d35911c
Folding build_tools back into the main repo for simplicity.
2015-12-30 16:53:14 -08:00
Ben Vanik
214532a3e8
Fiddling with FPSCR and such. Still not implemented, but wiring done.
2015-12-30 16:53:14 -08:00
