My question is somewhat different from others that have asked about fault addresses. I'm trying to implement a horrible hack to determine, from a signal handler, whether the signal interrupted a syscall or ordinary user code by inspecting the code at the saved instruction pointer and comparing it against the possible syscall entry instructions for the host architecture it's running on. This is part of implementing correct POSIX thread cancellation that does not suffer from the race condition and resource leak described in my old question:
How are POSIX cancellation points supposed to behave?
If this approach is unreliable or otherwise wrong, I'd also like to hear reasons.
I found this code on a LUG. Couldn't get to the page to point the URL here, so pasted the whole code. This code prints a small stack trace when SIGSEGV occurs. Not sure if there is some other way that does not use ucontext_t.