Re: fs: proc: gpf in find_entry

[Andrey Ryabinin]

2014-12-22 23:39 GMT+03:00 Eric W. Biederman <ebiederm@xxxxxxxxxxxx>:
> Sasha Levin <sasha.levin@xxxxxxxxxx> writes:
>
>> On 12/22/2014 12:52 PM, Andrey Ryabinin wrote:
>>> 2014-12-22 18:51 GMT+03:00 Eric W. Biederman <ebiederm@xxxxxxxxxxxx>:
>>>> These two instructions:
>>>>>>   11:   4d 85 ff                test   %r15,%r15
>>>>>>   14:   0f 84 de 01 00 00       je     0x1f8
>>>>
>>>> Should prevent a NULL %r15 value from ever reaching the trapping
>>>> instruction.
>>>
>>> If they were executed, then yes. But I think there was jump from somewhere
>>> to the instructions below those two.
>>
>> There is indeed a jump direct to that point, which avoids the %r15
>> check.
>
> Where do you see that direct jump, that certainly has not been posted
> in this thread?
>
> There are certainly no such code paths I in the source code.  There is
> only one NULL pointer check in find_entry and it is executed every time
> the loop executes.
>
> So at this point all I know is some set of tools has totally destroyed
> the code and made what Sasha Levin's is testing so far from the source
> code that this is a useless bug report.
>
> I have no reason to even suspect this bug is actually in the upstream
> kernel.
>

Generated code looks correct to me (considering that it was built with
sanitizer tools).
So this looks like a real BUG to me.

AFAICT this is 'head' == NULL:

   head = ctl_node->header;
   entry = &head->ctl_table[ctl_node - head->node];


> This appears to be a kind of testing that slows development and wastes
> peoples time.  Can someone give me a patch that sets the TAINTED flag
> when KASAN is loaded?
>

Pay attention to the first line of report:
[ 2015.960381] general protection fault: 0000 [#1] PREEMPT SMP KASAN
--
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/