On 10/10/13 09:33, Joe Perches wrote:
> On Thu, 2013-10-10 at 09:25 +1100, Ryan Mallon wrote:
>
>> if (kptr_restrict && (in_irq() || in_serving_softirq() ||
>> in_nmi())) {
>>
>> Is making sure that you don't have kernel code doing something like this:
>>
>> irqreturn_t some_irq_handler(int irq, void *data)
>> {
>> struct seq_file *seq = to_seq(data);
>>
>> seq_printf(seq, "value = %pK\n");
>> return IRQ_HANDLED;
>> }
>>
>> Because that obviously won't work when kptr_restrict=1 (because the
>> CAP_SYSLOG check is meaningless). However, the code is broken regardless
>> of the kptr_restrict value.
>
> The only brokenness I see here is that the code doesn't pass
> a pointer along with %pK
>
> seq_printf(seq, "value of seq: %pK\n", seq);
>
>> Since the default value of kptr_restrict is
>> 0, this kind of bug can go over-looked because the seq file will print
>> the pointer value correctly when kptr_restrict=0, and it will correctly
>> print 0's when kptr_restrict=2, but it will print 'pK-error' when
>> kptr_restrict=1. Doing the check in all cases makes it more likely that
>> bugs like this get found. In fact, doing something like:
>>
>> if (WARN_ON(in_irq() || in_serving_softirq() || in_nmi())) {
>>
>> Might be better, since that will print a stack-trace showing where the
>> offending vsprintf is.
>
> WARN_ON would be potentially _very_ noisy.
> Maybe a long period (once a day?) ratelimited dump_stack();
If it was noisy, it would indicate a bunch of broken kernel code which
needs fixing :-). Anyway, this is really a separate issue to what I am
trying to fix, which is why I left the original code intact. If you want
to change it, post a follow-up patch.
~Ryan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
