Re: Detecting illegal memory access

[Jan Engelhardt]

>
>A user program can trap EFAULT.

There are no syscalls involved. It is a [arch-dependent] stack backtrace 
that walks up EBP. But object files compiled with -fomit-frame-pointer have 
random values in %ebp. So when doing

    ebp = *(void **)ebp;

it either succeeds or throws SIGSEGV. Catching signals is catchy because 
the backtrace might be done in a threaded application. Does *jmp() work 
with threads?


>However, you need to use siglongjmp()
>to get back into your program and continue, plus you need to do very
>careful coding to mask and unmask the appropriate signals though it
>all. Since protection is on a per-page basis, one could make a program
>that attempts to R/W from virtual page 0 on up, recoding success or
>failure for each page.


Jan Engelhardt
-- 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/