Due to how we emulate the guest signal handlers, we do the state setup
in the real host signal handler, then we jump out after state setup.
This was causing a situation where we were setting up the guest signal
handler state with the correct sa_mask.
Then after setting up the guest state we would leave the FEX signal
handler, restoring the signal mask to our original mask.
Instead now as we are setting up the guest state, we save our host
signal mask. Then on signal handler return we modify our host signal
mask to match what the guest wants.
Once we hit our sigreturn emulation we then restore the original signal
mask.
This looks to improve some stability problems regarding how wine uses
signals but it still doesn't fix the gvisor test sadly.
Now that we deparent FEXMountDaemon we can no longer use
PR_SET_PDEATHSIG.
Instead we rely on the ref counting and checking the pipe status to see
if the original parent has left us.
Further improvements that could be done in the future is that every user
of the mount point talks to the daemon to give it a pipe to check is
still live, since the ref counting sometimes is incorrect.
This flag is ignored with sigaltstack.
Fixes an early assert in:
- Splice
- No Time to Explain Remastered
- Ittledew
- Hyperdrive Massacre
- English Country Tune
We were using the glibc context structure layout which doesn't match
what the kernel is doing.
Switches over to allocating fpstate independentally of the ucontext_t,
then pointing to it from uc_mcontext how we're supposed to.
glibc uses the __fpregs_mem region for other purposes.
This also fixes xmm state being stored on the 32-bit side, and also
fixes SIGALRM and SIGVTALRM siginfo overwriting structure data.
On 32-bit we can't just memcpy the siginfo_t over because the sizes
don't match. Throw a message instead.
This allows FEXMountDaemon to remove FEXInterpreter as its parent.
Instead becoming the parent of whatever the current reaper process is.
Do it as early as possible this way FEXInterpreter won't get an
erroneous SIGCHLD.
There is no glibc wrapper for tkill and tgkill requires a tgid.
Kernel will reject us if we tried using -1 or 0 even though that is what
it does internally.
If it is going back to SIG_IGN or SIG_DFL then let them unregister.
This is useful for when they are wanting to ignore a signal after a
while.
Or only capture a fault during a time, then afterwards want the
application to crash on error.
Additionally clears up some other minor logic which wasn't being used
anymore
These syscall arguments are not laid out in a sane way.
Looks like they wanted to keep the interface the same for both x86-64
and 32-bit x86 so they split the offset argument in to two values.
Weirdly enough, even though these are 32-bit offsets on 32-bit x86; On
x86-64 these are still 64-bit. Which means the kernel weirdly allows you
to overlap the two 64-bit halves.
eg: `uint64_t Offset = (pos_high << 32) | pos_low;`
So you can have a 64-bit value that is the full range, but if you have
data in the lower 32-bits of pos_high then you corrupt the offset.
Additionally this allows you to interleave low and high if you want to
be obtuse.
Initially this was a small hack to work around sigqueueinfo sending over
siginfo_t. Now this is unnecessary and would crash if an application
used sigqueueinfo to a signal without the SIGINFO flag.
Ryan Houdek