This doesn't currently install thunks which can come a bit later.
We require a postinst and prerm step for importing and unimporting the binfmt_misc files.
Easy enough
The allocator functions don't return -1 on error, instead return the actual error.
This was setup to match the syscall behaviour rather than the glibc/errno behaviour
In the case of a kernel older than 4.17 then MAP_FIXED_NOREPLACE doesn't exist.
This adds a check in to see if the pointer returned was valid but not what we asked for
Instead of forcing a fixed offset for the stack. Allow it to get placed automatically.
64-bit was already doing this; Now we can also do it in 32-bit.
This is possible because the mapper that is mapping the code will always map in the 32-bit space
for a 32-bit guest.
Currently our 32-bit code only gave one TLS slot and crashed
if you needed more.
First switch over to the same initial slot as the Linux kernel.
Then allow searching the slots for a free spot.
This allows us to more closely match the behaviour of the Linux kernel just in case
anything has hardcoded the TLS slots
This makes it so Saints Row: The Third stops crashing at boot, plays a few intro videos, then hangs instead.
These 32bit syscalls use a compat statfs which is only 64bytes in size.
This was overwriting data on the guest stack and causing crashes.
Describe a 32-bit statfs struct and ensure with struct verifier that it matches.
Fixes a crash in KOTOR2 that would happen just before the main menu.
This was altered in kernel 2.3.23 to include a mem_unit variable.
mem_unit states what each of the memory units is scaled by.
The kernel will always have this be 1 or page size on x86, so follow that behaviour.
Instead of maxing out the values to what uint32_t can handle, scale all of these by page size instead.
This fixes a crash in KOTOR2's boot sequence where it would divide by zero if everything was maxed out.
Test sleeps for 1 second and expects to come back within 10ms of the time.
When the CPU is doing other things then it can end up missing that timeframe. Thus flake.
Just disable it
Places them in RO where they can't be modified.
While we're in the area, we can use an alias to prevent duplicated array
types, and also add a static assert to ensure the arrays are always the
same size.
This allows us to avoid needing to bounds check several accesses in a
row that we know will always be successful.
This function does everything required for setting up a squashfs as a rootfs.
- Checks if the file is a valid squashfs
- Executes the FEXMountDaemon
- Error checks to ensure it was mounted correctly
- Updates CONFIG_ROOTFS to point to the mounted location
- Takes 200-400ms more startup time
This is a tool that communicates with FEXLoader/FEXInterpreter to automatically mount
squashfs based rootfs files on execve.
This tool will launch automatically if you have a squashfs based rootfs selected.
Once the rootfs is mounted, it will watch for the parent FEX processe to completely exit.
Once the parent FEX exits it will unmount and cleanup after itself.
This tool has a dependency on your host having FUSE, fusermount, and squashfuse applications.
If anything goes wrong in the bringup process then it propagates the erro up the chain and will
let FEX know that it couldn't mount.
We already have a Config.h which conflicts with this file.
In issue cropped up if you need to include Config.h (the generated one)
from inside Common/ then it would only pull the Common/Config.h file.
Just change the name to not be confusing
Ryan Houdek