Re: [patch 1/3] vdso: print fatal signals

[Zachary Amsden <zach@vmware.com>]

Ingo Molnar wrote:
> Index: linux-vdso-rand.q/kernel/signal.c
> ===================================================================
> --- linux-vdso-rand.q.orig/kernel/signal.c
> +++ linux-vdso-rand.q/kernel/signal.c
> @@ -763,6 +763,37 @@ out_set:
>  #define LEGACY_QUEUE(sigptr, sig) \
>  	(((sig) < SIGRTMIN) && sigismember(&(sigptr)->signal, (sig)))
>  
> +int print_fatal_signals = 0;
> +
> +static void print_fatal_signal(struct pt_regs *regs, int signr)
> +{
> +	printk("%s/%d: potentially unexpected fatal signal %d.\n",
> +		current->comm, current->pid, signr);
> +
> +#ifdef __i386__
> +	printk("code at %08lx: ", regs->eip);
> +	{
> +		int i;
> +		for (i = 0; i < 16; i++) {
> +			unsigned char insn;
> +
> +			__get_user(insn, (unsigned char *)(regs->eip + i));
> +			printk("%02x ", insn);
> +		}
> +	}
> +#endif
>   


This looks ok for debugging boot problems.  Perhaps you could print the 
registers too?  The instruction dump won't help much for indirect access.

The get_user of eip+i is ok, but doesn't account for segment offsets.  
Not that I think it needs to here.  But it is one of a many growing 
number of places that now try to inspect or modify a potentially 
segmented area of memory (page fault handler must inspect for prefetch 
instructions, kprobes reads and patches code, FPU emulation).  Perhaps a 
common interface would be a nice thing at some point in time.

Zach