steam-frame/FEX-Emu--FEX
1
0 Fork 0
mirror of https://github.com/FEX-Emu/FEX.git synced 2026-10-26 22:01:47 +01:00
Code Issues Packages Projects Releases 67 Wiki Activity
4,879 Commits 11 Branches 71 Tags
ffb4de9fd9320f428e3ff5283f5fa830ee23ffb5
Commit Graph
4879 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Ryan Houdek ffb4de9fd9 Merge pull request #2079 from Sonicadvance1/ensure_armemitter_uses_allocator 
Ensure Arm64Emitter uses FEX allocator
 2022-10-15 15:37:23 -07:00
Ryan Houdek 6b3d8886e5 Merge pull request #2077 from Sonicadvance1/fix_thunks_with_lots_args 
Thunks: Fixes indirect thunks with 8+ arguments
 2022-10-15 15:37:09 -07:00
Ryan Houdek ddc10272a0 Merge pull request #2076 from Sonicadvance1/update_vulkan 
Thunks: Update Vulkan thunk to v1.3.231
 2022-10-15 15:14:30 -07:00
Ryan Houdek 0b5ef00165 Thunks: Fixes indirect thunks with 8+ arguments 
Due to how we use a modified ABI for these indirect functions, we don't
have a clean way to say that the host_addr lives in a side-argument.

The previous inline asm that moved the value from r11 in to a variable
worked up until you hit functions with 8 or more arguments. At that
point the compiler was generating code before our inline assembly and
using r11 as a temporary, thus destroying our value.
Then a crash would occur and it was very hard to determine why. It would
end up calling some random function (0x1 in this case) from an indirect
call.

This made it /look/ like it was calling an invalid function returned
from the loader but in reality it was a corrupt register loading bad
data.

To work around this case, we can use an inline asm register variable and
a volatile asm block that "sets" the variable. In this case GCC and
Clang both seem to extend the live range of the register from the start
of the function to the use of the variable.

This resolves the issue for now, and I tested quite a large number of
function signatures to see if it would break in the future.

Theoretically our functional testing should catch this, but we don't
currently have something that abuses all the functions like this
currently.
 2022-10-15 15:13:40 -07:00
Ryan Houdek abb44d3327 Merge pull request #2069 from wannacu/main 
Flags: Refine _Bfe's shift
 2022-10-14 22:53:05 -07:00
Ryan Houdek 11eaf3d48a Ensure Arm64Emitter uses FEX allocator 
Otherwise we will end up allocating code buffers in the lower 32-bits,
consuming precious virtual address space.
 2022-10-14 22:04:07 -07:00
Ryan Houdek 0e6c8bd12e Thunks: Update Vulkan thunk to v1.3.231 
Only missing a few function definitions, resorted to match order of
definitions in the headers so future changes don't mix up as much
 2022-10-14 01:51:35 -07:00
Ryan Houdek a9fb008317 External: Update Vulkan-Headers to v1.3.231 2022-10-14 01:50:49 -07:00
Ryan Houdek e9f3a5b3e4 Merge pull request #2074 from lioncash/vuxtl 
IR: Handle 256-bit VUXTL/VUXTL2
 2022-10-13 13:15:47 -07:00
Ryan Houdek ebc45dff45 Merge pull request #2070 from lioncash/vuabdl 
IR: Handle 256-bit VUABDL
 2022-10-13 12:17:47 -07:00
lioncash fc4a5ebfd3 IR: Handle 256-bit VUXTL2 
Extends VUXTL2 to handle 256-bit vectors.
 2022-10-13 19:15:05 +00:00
lioncash 1d7b688c55 IR: Handle 256-bit VUXTL 
Extends VUXTL to handle 256-bit values.
 2022-10-13 19:07:57 +00:00
Ryan Houdek f14a5ffbbf Merge pull request #2073 from lioncash/vsxtl 
IR: Handle 256-bit VSXTL/VSXTL2
 2022-10-13 11:53:12 -07:00
Ryan Houdek cada0d593c Merge pull request #2072 from lioncash/test 
unittests: Amend mm register usage in H0F38/66_04.asm test
 2022-10-13 11:28:27 -07:00
lioncash 0436540791 IR: Handle 256-bit VSXTL2 
Extends VSXTL2 to handle 256-bit vectors.
 2022-10-13 18:21:25 +00:00
lioncash a87ac86e18 IR: Handle 256-bit VSXTL 
Extends VSXTL to handle 256-bit vectors.
 2022-10-13 18:21:22 +00:00
lioncash 1278b23150 unittests: Amend mm register usage in H0F38/66_04.asm test 
This should be using xmm2 rather than mm2.
 2022-10-13 17:09:40 +00:00
lioncash 02f5ea4b9d IR: Handle 256-bit VUABDL 
Extends VUABDL to handle 256-bit vectors.
 2022-10-13 15:37:10 +00:00
wannacu 2e14e613d0 Flags: Refine _Bfe's shift 2022-10-13 16:41:04 +08:00
Ryan Houdek 85c2889652 Docs: Update for release FEX-2210 FEX-2210 2022-10-13 00:46:59 -07:00
Ryan Houdek 23dd056b60 Merge pull request #2067 from lioncash/vmul 
IR: Handle 256-bit VSMul/VUMul
 2022-10-12 15:59:03 -07:00
lioncash 71043e372a IR: Handle 256-bit VSMul/VUMul 
Extends VSMul and VUMul to handle 256-bit vectors.
 2022-10-12 00:00:15 +00:00
Ryan Houdek c412d073b9 Merge pull request #2066 from lioncash/vrev64 
IR: Handle 256-bit VRev64
 2022-10-11 14:21:45 -07:00
lioncash 8fb03ff1b9 IR: Handle 256-bit VRev64 
Extends VRev64 to handle 256-bit vectors.
 2022-10-11 20:17:04 +00:00
Ryan Houdek c2b6aef6f4 Merge pull request #2065 from lioncash/shift-imm 
IR: Handle 256-bit VShlI/VUShlI/VUShrI
 2022-10-11 12:16:44 -07:00
lioncash 24547318c6 IR: Handle 256-bit VShlI 
Extends VShlI to handle 256-bit vectors.
 2022-10-11 18:19:23 +00:00
Ryan Houdek b693112c80 Merge pull request #2058 from Sonicadvance1/add_opencl_thunk_db 
Add opencl thunk db
 2022-10-11 11:06:20 -07:00
lioncash 48d1184066 IR: Handle 256-bit VSShrI 
Extends VSShrI to handle 256-bit vectors.
 2022-10-11 18:04:13 +00:00
Ryan Houdek 8da9ebc2e0 Merge pull request #2062 from wannacu/main 
SMC: Fix possible deadlock
 2022-10-11 10:13:02 -07:00
lioncash 5ba510474b IR: Handle 256-bit VUShrI 
Extends VUShrI to handle 256-bit vectors.
 2022-10-11 17:08:35 +00:00
Ryan Houdek 51214d1be1 Merge pull request #2064 from lioncash/vushls 
IR: Handle 256-bit VSShrS/VUShlS/VUShrS
 2022-10-11 09:59:03 -07:00
Ryan Houdek 4d6e15d7af Merge pull request #2063 from lioncash/interp-shift 
Interpreter: Handle 256-bit VSShr/VUShl/VUShr
 2022-10-11 09:22:11 -07:00
lioncash 4721894427 IR: Handle 256-bit VSShrS 
Extends VSShrS to handle 256-bit vectors.
 2022-10-11 16:21:50 +00:00
lioncash d429865b6e IR: Handle 256-bit VUShrS 
Extends VUShrS to handle 256-bit vectors.
 2022-10-11 16:08:09 +00:00
lioncash ca5881a72c IR: Handle 256-bit VUShlS 
Extends VUShlS to handle 256-bit vectors.
 2022-10-11 15:45:39 +00:00
lioncash 7151b9daff Interpreter/VectorOps: Remove lingering magic 32 constants 
Makes these functions consistent with the rest that explicitly test for
256 bit width.
 2022-10-11 15:02:41 +00:00
lioncash d9b5e28b22 Interpreter: Handle 256-bit VSShr 
This is only implemented in the interpreter, so this is trivial.
 2022-10-11 14:58:40 +00:00
lioncash aa7954a7d6 Interpreter: Handle 256-bit VUShr 
This is only implemented in the interpreter, so this is trivial.
 2022-10-11 14:56:19 +00:00
lioncash 802c70d1ab Interpreter: Handle 256-bit VUShl 
This is only implemented in the interpreter, so this is trivial.
 2022-10-11 14:55:17 +00:00
wannacu 8f905988e9 Use compatible syscall helpers 2022-10-11 16:35:17 +08:00
wannacu 478c5595ad SMC: Fix possible deadlock 2022-10-11 15:31:27 +08:00
Ryan Houdek 3977e1f29e Merge pull request #2055 from Sonicadvance1/update_description_ripping_script 
Scripts: Updates DefinitionExtract
 2022-10-10 10:11:16 -07:00
Ryan Houdek 2b1ef97354 Merge pull request #2060 from Sonicadvance1/clang_thunks 
Thunks: Add support for building with clang
 2022-10-10 09:44:58 -07:00
Mai eaddf7f1a5 Merge pull request #2061 from Sonicadvance1/fix_linker_script_depends 
Thunks: Adds dependency on linker script
 2022-10-10 12:17:20 -04:00
Ryan Houdek 3237de3085 Merge pull request #2056 from Sonicadvance1/guest_function_bool 
Thunks/Host: Adds bool operator to fex_guest_function_ptr
 2022-10-10 09:10:34 -07:00
Ryan Houdek df3d398d31 Thunks/Host: Adds bool operator to fex_guest_function_ptr 
Lets us check if nullptr was passed in
 2022-10-10 08:52:50 -07:00
Ryan Houdek b44b3401b7 Merge pull request #2015 from Sonicadvance1/map_regular_offset 
ELFCodeLoader: Map primary ELF more like the kernel
 2022-10-10 08:49:49 -07:00
Ryan Houdek c28ca0fac9 ELFCodeLoader: Map primary ELF more like the kernel 
The kernel maps the primary ELF with a hint to some place *near* the
middle of the virtual address space. While the interpreter stays at the
top of the address space.

This also adds ASLR to the ELF loading, with a define for debugging and
testing purposes.

Requires both #2013 and #2014 merged first.
 2022-10-10 08:38:15 -07:00
Ryan Houdek 235e2b6c2c FEXCore/Allocator: Store what the host VA is 
Calling this function multiple times without this will change the
result.

Necessary so we can determine what the host VA is  from multiple
locations.
 2022-10-10 08:35:40 -07:00
Mai d68b84bc27 Merge pull request #2013 from Sonicadvance1/fix_mapper 
ELFCodeloader: Map once and then use MAP_FIXED to overwrite
 2022-10-10 10:05:53 -04:00