Re: [RFC][PATCH] bitfields API

["Vegard Nossum" <vegard.nossum@gmail.com>]

On Thu, Aug 28, 2008 at 10:27 PM, Adrian Bunk <bunk@kernel.org> wrote:
> On Thu, Aug 28, 2008 at 09:40:47PM +0300, Pekka Enberg wrote:
>> Hi Alexey,
>>
>> Alexey Dobriyan wrote:
>>> On Thu, Aug 28, 2008 at 08:32:23PM +0200, Vegard Nossum wrote:
>>>> How do you feel about this patch? It's all about making kmemcheck more
>>>> useful... and not much else. Does it have any chance of entering the
>>>> kernel along with kmemcheck (when/if that happens)?
>>>
>>> DEFINE_BITFIELD is horrible.
>>
>> Heh, heh, one alternative is to have a kmemcheck_memset() thingy that
>> unconditionally zeroes bit fields and maybe is a no-op when kmemcheck is
>> disabled.
>
> This sounds as if this might cause bugs to disappear when debugging gets
> turned on?
>
> Or do I miss anything?

You are correct :-)

Almost all the possible solutions (at least the feasible ones) are
trade-offs between false-positives and false-negatives.

So here we are trading a bunch of false-positive errors (a couple of
thousand for transferring a 1M file over ssh :-)) for detecting any
code that uses an uninitialized flag in struct skbuff. So in this case
it is more useful to hide reports about this single bit-field.


Vegard

-- 
"The animistic metaphor of the bug that maliciously sneaked in while
the programmer was not looking is intellectually dishonest as it
disguises that the error is the programmer's own creation."
	-- E. W. Dijkstra, EWD1036