In vm/compaction.cpp I rescoped some variables to lower the stack usage
from 592 to 560 bytes. I wasn't very successful with this. The stack
usage is larger than it looks because methods on the structures used
take an implicit this pointer and a reference to the data (so the data
has to live out it's full scope and can't be put in a register).
In vm/debug.cpp I made a large (1024 bytes) stack allocated buffer
simply dynamically allocated.
In vm/os-unix.cpp I rescoped signal handling structures to not coincide
with each other and reduced a very large (1024 bytes) amount of stack
usage to less than 500 bytes.
clang-format doesn't recognize casts to non-pointer/non-template types
so it winds up adding a space between the right paren and the expression
and then failing to recognize prefix operators in the process
(e.g. foo = (cell) & bar; should be foo = (cell)&bar;). This commit
manually fixes up the major cases (fixnum, cell, all types ending in _t).
Factor is finally a real C++ project and has a custom assert macro. Assertion failures were still getting caught as exceptions and causing failure loops. Write our own macro that calls factor::abort on failure.
* Clear faulting_p from a safepoint rather than inside general_error, because jumping into unwind-native-frames could blow up.
* Handle multiple faults from fatal_error by breakpointing. Is there anything else we can safely do at that point?
* Verify memory protection faults in the top half of the signal handlers because signal dispatch could fault. Treat memory faults during gc or fep as fatal errors.
* Add a function factor_vm::abort() that restores the default SIGABRT handler and ::abort()s. Use it from fatal_error() so we get useful context from gdb and so the user gets feedback from the system crash reporter that Factor blew up and didn't just disappear.
* In factorbug(), don't proceed with .s .r .c if it would be unsafe to do so.
* Don't pile on signals if we've already called fatal_error().
This gives us an escape hatch for when things go way south and ^C can't be handled. Also unmask SIGQUIT and SIGTERM from the stdin_loop thread so the thread doesn't keep the process alive after those signals.
The stdin_loop thread will keep trying to consume input unless we stop it by sending it a signal. Use SIGUSR2 to stop the read syscall and a mutex to hold up the loop while the fep is active.
Although SIGINT still can't interrupt the current Factor thread in a sane way, this will at least wake up the run loop when waiting for input and fix#348 when implemented at the application level.
On OS X it appears the mach exception thread and libdispatch queue threads occasionally get the SIGALRM from the itimer. Count those as foreign_thread_samples instead of letting the signal kill the process.
Set up the state necessary to collect samples. We still need to add GC support for walking the sample set, and the compiler needs to register GC roots before safepoints as well. We also need primitives to expose the data to Factor for reporting.
This will let us safely unguard the callstack overflow guard page as a grace page so the callstack overflow handler can do its business without discarding frames.
It's not robust currently to raise an exception because a lot of our code that isn't already written with exceptions in mind breaks. Also, a signal is likely to be received by an FFI callback installed on the IO multiplexer, which will cause Factor to die since the callback cannot handle the exception. We need a more robust solution to dealing with SIGINT.
Also lay some groundwork for counting profile samples and reporting non-interrupting asynchronous signals.
Remove the weird and broken functions fix_callstack_top and scrub_return_address. Instead, simply decrement the SP and store the PC from the front end of the signal handler so that the back end can return back into the original context normally. Currently aborts for leaf procedure frames pending a more robust solution.
Björn Lindqvist
Doug Coleman
Steven Stewart-Gallus
Erik Charlebois
Joe Groff
Philipp Brüschweiler