>:( i had put those warnings to prevent the crash itself and what's causing
>it.
> may be i had missed that pos can be < 0 in addition to (pos > 37).
> i am just looking at those areas of code for some other issue. hope i
> can a find a fix
> for both of them
I am building kernel with pos<0 check added...
I hoped for some ad-hoc stress-test recommendation, but I guess this will have
to wait.
On Fri, Jan 6, 2012 at 8:05 PM, Mohammed Shafi <[email protected]> wrote:
> 2012/1/6 MR <[email protected]>:
>> ?>:( ?i had put those warnings to prevent the crash itself and what's causing
>> ?>it.
>> ?> may be i had missed that pos can be < 0 in addition to (pos > 37).
>> ?> i am just looking at those areas of code for some other issue. hope i
>> ?> can a find a fix
>> ?> for both of them
>>
>> I am building kernel with pos<0 check added...
>
> thanks! ?if it does not works, i have only one more idea(a patch
> which removes some suspicious code which abuses a pointer any way that
> has to be fixed properly ). otherwise i need to recreate the issue and
> capture the stack trace completely, put debug prints. i ran an
> overnight test but was unable to recreate the issue. later today AN ?i
> got a crash accidentally but only once, after that never.
attaching the patch for your reference, but this is not the fix. it
has to be fixed properly.
if it does not helps, then the issue itself is something else i had assumed.
>
>>
>> I hoped for some ad-hoc stress-test recommendation, but I guess this will have
>> to wait.
any test if we can recreate the panic ASAP will be highly helpful. it
will help us narrow down the issue quickly and put more debug
prints/ideas
>>
>>
>
>
>
> --
> shafi
--
shafi
2012/1/6 MR <[email protected]>:
> ?>:( ?i had put those warnings to prevent the crash itself and what's causing
> ?>it.
> ?> may be i had missed that pos can be < 0 in addition to (pos > 37).
> ?> i am just looking at those areas of code for some other issue. hope i
> ?> can a find a fix
> ?> for both of them
>
> I am building kernel with pos<0 check added...
thanks! if it does not works, i have only one more idea(a patch
which removes some suspicious code which abuses a pointer any way that
has to be fixed properly ). otherwise i need to recreate the issue and
capture the stack trace completely, put debug prints. i ran an
overnight test but was unable to recreate the issue. later today AN i
got a crash accidentally but only once, after that never.
>
> I hoped for some ad-hoc stress-test recommendation, but I guess this will have
> to wait.
>
>
--
shafi