Re: [PATCH 0/5] mm/kasan: advanced check

[Wengang <wen.gang.wang@oracle.com>]

On 11/20/2017 12:20 PM, Dmitry Vyukov wrote:
> On Mon, Nov 20, 2017 at 9:05 PM, Wengang <wen.gang.wang@oracle.com> wrote:
>>
>> On 11/20/2017 12:41 AM, Dmitry Vyukov wrote:
>>>
>>>> The reason I didn't submit the vchecker to mainline is that I didn't find
>>>> the case that this tool is useful in real life. Most of the system broken
>>>> case
>>>> can be debugged by other ways. Do you see the real case that this tool is
>>>> helpful?
>>> Hi,
>>>
>>> Yes, this is the main question here.
>>> How is it going to be used in real life? How widely?
>>>
>> I think the owner check can be enabled in the cases where KASAN is used. --
>> That is that we found there is memory issue, but don't know how it happened.
>
> But KASAN generally pinpoints the corruption as it happens. Why do we
> need something else?

Currently (without this patch set) kasan can't detect the overwritten 
issues that happen on allocated memory.

Say, A allocated a 128 bytes memory and B write to that memory at offset 
0 with length 100 unexpectedly.  Currently kasan won't report error for 
any writing to the offset 0 with len <= 128 including the B writting.  
This patch lets kasan report the B writing to offset 0 with length 100.

thanks,
wengang

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>