musings.txt

On Linx, SIGSEGV which are raised by the kernel (i.e. have si_code=128/SI_KERNEL) seem to always have si_addr=0.
This confuses implicit NULL checks because we can get NULL-pointer exception in places where we should actually crash!
It seems that SIGSEGVs raised by the kernel occur for very large addresses, which are not mapped into the user space (i.e. 0xffffffffffffffff, 0xbaadbabebaadbabe)

See:

os_linux_x86.cpp
----------------

#ifndef AMD64
    // Halt if SI_KERNEL before more crashes get misdiagnosed as Java bugs
    // This can happen in any running code (currently more frequently in
    // interpreter code but has been seen in compiled code)
    if (sig == SIGSEGV && info->si_addr == 0 && info->si_code == SI_KERNEL) {
      fatal("An irrecoverable SI_KERNEL SIGSEGV has occurred due "
            "to unstable signal handling in this distribution.");
    }
#endif // AMD64

http://unix.stackexchange.com/questions/71240/sigaction7-semantics-of-siginfo-ts-si-code-member