In the case that there is an immediate configuration failure. Use stderr specifically for outputting.
These errors won't be output typically because silent logging is enabled by default.
In the case of executable missing or rootfs configuration failure, print directly to stderr.
Previously it looked like FEX just exited for no reason.
We had multiple users encounter this and be confused
We already support receiving user signals in our signal handlers for this.
Just need to push through syscalls.
This fixes Dead Island Definitive Edition so it runs.
This is a base implementation of signalfd.
Signalfd allows the application to receive siginfo_t information through an FD.
The FD is either provided by the application or created by the kernel depending.
This specifically doesn't pick up *true* synchronous signals. tgkill of the number
should theoretically go through this interface.
This very specifically skips our internal required signals for now.
This means it won't pick up SIGILL, SIGBUS, or SIG63.
This is enough to capture an application that just wants to poll for SIGCHLD.
Anything more complex has the same problems of the guest handling a siginfo_t.
gzip ships a PT_LOAD program section without a file size and only a memory size.
In the case of a zero file size PT_LOAD then return success immediately in this section loader.
This fixes Steam using gzip to package up crash logs
Don't print how many instructions are installed in the tables.
This isn't useful anymore
Not installing signal 32 and 33 are something we don't support right now. Stop complaining in that case.
Stop printing when a thread is starting up and shutting down. If you want to see this then gdb shows it well.
Don't print clone flags unless we are hitting a case where we are printing another log message.
Only the frontends need to deal with ELF files specifically.
The backend doesn't need to be aware of them at all.
Since the ELF handling is the frontend's responsibility, move all the code to the frontend.
In a newer version of the kernel there was a feature to disallow cpuid.
We can emulate this by saying it is always enabled and disallow the ability to disable it.
This was overwriting the cmd argument and then being checked in the switch statement
after the call.
Since it was overwritten, it wasn't falling down the correct path, returning a flock_32 instead of a flock64_32
Since the vector was being initialized with braces, this caused the
arrays to only ever have one element within it instead of the size of
the array.
While we're at it, we can also construct a vector in place with the
necessary size.
We were sending these non-ipc variants of these syscalls through
the generic path. This is broken but luckily glibc from Ubuntu isn't shipping
a version using these yet.
Now that are aren't consuming all signals we need to more aggressively handle the host signal mask.
Now more signals are getting masked and blocked how they should be.
I saw a red herring that I thought the high cpu usage in steamwebhelper could come from signal handlers.
This turned out to not be the case, but now I've got this implemented.
Installs the few signal handlers that we need upfront but for everything that isn't a mandatory signal
we instead now wait until the guest also installs that signal handler.
This fixes#1107
Ryan Houdek