Re: [PATCH 6/6] kasan: improve double-free reports.

[Andrey Ryabinin]

On 08/02/2016 02:39 PM, Alexander Potapenko wrote:

>> +static void kasan_end_report(unsigned long *flags)
>> +{
>> +       pr_err("==================================================================\n");
>> +       add_taint(TAINT_BAD_PAGE, LOCKDEP_NOW_UNRELIABLE);
> Don't we want to add the taint as early as possible once we've
> detected the error?

What for?
It certainly shouldn't be before dump_stack(), otherwise on the first report the kernel will claimed as tainted.


>>
>> +void kasan_report_double_free(struct kmem_cache *cache, void *object,
>> +                       s8 shadow)
>> +{
>> +       unsigned long flags;
>> +
>> +       kasan_start_report(&flags);
>> +       pr_err("BUG: Double free or corrupt pointer\n");
> How about "Double free or freeing an invalid pointer\n"?
> I think "corrupt pointer" doesn't exactly reflect where the bug is.

Ok