Commit Graph
13 Commits
Author SHA1 Message Date
Stefanos Kornilios Mitsis Poiitidis 370f36c8f7 Merge pull request #1027 from Sonicadvance1/fix_ppoll
ppoll fixes
2021-05-18 09:20:32 +03:00
Ryan Houdek dfb15aaeb9 ppoll fixes
glibc implementation makes a copy of the timeout and the kernel is expected to update it
Can't use the glibc implementation because of this.
2021-05-11 18:33:24 -07:00
Ryan Houdek 565f0e6af3 Handle FD flag remapping correctly
FD flag remapping was broken. It would remap one flag on to another and then the next check would remap it back.

Instead keep a mask of the flags to be remapped then remap them all at the end.
Also goes through the ops and fixes a few cases where it was remapping wrong flags.
2021-05-11 18:29:21 -07:00
Ryan Houdek 83b4188569 Microoptimization for the DRM ioctls
For DRM applications have a three FD deep MRU cached for faster lookups of FD to DRM handlers.
In a completely DRM ioctl bound situation like es2gears or GL application without threaded context
Then this puts us /nearly/ at the performance of calling the ioctl32 handler directly.
Sadly there is overhead that can't be overcome so this is the best that can be done from userland
2021-05-06 02:55:22 -07:00
Ryan Houdek 45223cc7e7 Adds initial ioctl emulation code 2021-05-01 07:39:14 -07:00
Ryan Houdek 04e0baadc6 Changes LogMan assert to defines
C++ no-op functions can't optimize out the predicate arguments in all cases.
This was causing a problem where zero cost assertions weren't actually zero cost.

The only way to resolve this is to actually use macros sadly enough.

This will give a fairly hefty performance uplift with anything operating on IR.
2021-04-26 13:49:35 -07:00
Stefanos Kornilios Mitsis Poiitidis 942b8549c6 Docs: Add tags to the source code 2021-03-30 12:21:18 +03:00
Scott Mansell 6ed36e3304 Refactor syscalls to take Frame
The vast majoirty of syscalls don't need anything in thread or frame.
So lets save an indirection for all those syscalls.

Most of the syscalls which do need Thread (or CTX via
Thread are in Thread.cpp or Memory.cpp
These have all been modifiy to fetch Thread from Frame
2021-03-09 20:28:50 +13:00
Stefanos Kornilios Mitsis Poiitidis 568fb3643d Syscalls/x86-32: Fix selects to actually remarshal the sets after running the host syscall 2021-02-16 02:35:38 +02:00
Ryan Houdek 15df787710 Implements ioctl32 for x86-64
If you have the COMPAT feature enabled then this just works there.
AArch64 not implemented currently since changes aren't upstream
2021-01-27 08:02:49 -08:00
Ryan Houdek 433162004a Implements more syscalls 2021-01-12 14:50:17 -08:00
Ryan Houdek 66e98365eb Captures fsstatat in filesystem overlay
Fixes Vulkan loader searching for vulkan json ICD files
2021-01-10 02:13:48 -08:00
Ryan Houdek 37983006f0 Moves FEXCore's SignalDelegator and SyscallHandler to the frontend
Sadly these things can't be split without breaking functionality so it
turns in to a bit of a mess.

SyscallHandler is very much something that is a Linux only construct and
shouldn't be in FEXCore itself. Lets the frontend register a
Syscallhandler with FEXCore. FEXCore itself is then aware of the current
syscall ABI and handles the ABI in an optimal fashion.
So it is not a 100% clean break otherwise we would lose performance.

The SignalDelegator then needs to move to the frontend since the
SyscallHandler requires it for signal based syscalls.
The CPU backend signal handling still needs to happen in FEXCore because
it is a very tight coupling with the CPU backend.
Once we need to support more Signal handling we can give the backends
cleaner support to select which specific OS handler to handle.
2020-12-13 22:38:04 -08:00