Char Limit issue on 2.6.29 kernels

Char Limit issue on 2.6.29 kernels

[naresh kamboju]

Hi,

I have come across with char Limit issue on 2.6.29 kernels.


SIGFAULT:

#define SIZE            8053759
char buf1[SIZE] ;

NO SIGFAULT:

#define SIZE            8053758
char buf1[SIZE] ;

This SIGFAULT not observed on 2.6.23 kernels.

What could be the changes that effecting from 2.6.23 to 2.6.29 kernels.
Please share me the information.


Best regards,
Naresh Kamboju

Re: [LTP] Char Limit issue on 2.6.29 kernels

[Mike Frysinger]

[-- Attachment #1: Type: text/plain, Size: 647 bytes --]

On Friday 24 July 2009 11:31:59 naresh kamboju wrote:
> I have come across with char Limit issue on 2.6.29 kernels.
>
>
> SIGFAULT:
>
> #define SIZE            8053759
> char buf1[SIZE] ;
>
> NO SIGFAULT:
>
> #define SIZE            8053758
> char buf1[SIZE] ;
>
> This SIGFAULT not observed on 2.6.23 kernels.

-ENOINFO

post some actual code rather than saying "i declared a large buffer and things 
crashed".  is there a reason you cc-ed LTP ?  is this a crash i a LTP test ?

> What could be the changes that effecting from 2.6.23 to 2.6.29 kernels.
> Please share me the information.

git log v2.6.23..v2.6.29
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [LTP] Char Limit issue on 2.6.29 kernels

[naresh kamboju]

Hi Mike,

On Fri, Jul 24, 2009 at 11:19 PM, Mike Frysinger<vapier@gentoo.org> wrote:
> On Friday 24 July 2009 11:31:59 naresh kamboju wrote:
>> I have come across with char Limit issue on 2.6.29 kernels.
>>
>>
>> SIGFAULT:
>>
>> #define SIZE            8053759
>> char buf1[SIZE] ;
>>
>> NO SIGFAULT:
>>
>> #define SIZE            8053758
>> char buf1[SIZE] ;
>>
>> This SIGFAULT not observed on 2.6.23 kernels.
>
> -ENOINFO
>
> post some actual code rather than saying "i declared a large buffer and things
> crashed".

currently i am debugging this issue.
i'll come up with more details.

>is there a reason you cc-ed LTP ?  is this a crash i >a LTP test ?
No.
it is not a test cases of LTP.
i am a member of LTP. want to share this information to LTP community.

-naresh

>
>> What could be the changes that effecting from 2.6.23 to 2.6.29 kernels.
>> Please share me the information.
>
> git log v2.6.23..v2.6.29
> -mike
>

Re: [LTP] Char Limit issue on 2.6.29 kernels

[Tim Bird]

naresh kamboju wrote:
> Hi Mike,
> 
> On Fri, Jul 24, 2009 at 11:19 PM, Mike Frysinger<vapier@gentoo.org> wrote:
>> On Friday 24 July 2009 11:31:59 naresh kamboju wrote:
>>> I have come across with char Limit issue on 2.6.29 kernels.
>>>
>>>
>>> SIGFAULT:
>>>
>>> #define SIZE � � � � � �8053759
>>> char buf1[SIZE] ;
>>>
>>> NO SIGFAULT:
>>>
>>> #define SIZE � � � � � �8053758
>>> char buf1[SIZE] ;
>>>
>>> This SIGFAULT not observed on 2.6.23 kernels.
>> -ENOINFO
>>
>> post some actual code rather than saying "i declared a large buffer and things
>> crashed".
> 
> currently i am debugging this issue.
> i'll come up with more details.

Some more feedback...

It's SEGFAULT - short for "Segmentation Fault", not SIGFAULT.

Please include in your details at least the following:
 * what software this is in (is this declaration in the kernel?, a library, a
 user space program, etc.)
 * what operation is being performed when the SEGFAULT occurs.
 * what compiler you are using.
 * the code surrounding the operation when the fault occurs

If you can, make the smallest possible program that reproduces this error,
and send the complete source for that.
 -- Tim

=============================
Tim Bird
Architecture Group Chair, CE Linux Forum
Senior Staff Engineer, Sony Corporation of America
=============================

Re: Char Limit issue on 2.6.29 kernels

[Ray Lee]

On Fri, Jul 24, 2009 at 8:31 AM, naresh kamboju <naresh.kernel@gmail.com> wrote:
>
> Hi,
>
> I have come across with char Limit issue on 2.6.29 kernels.
>
>
> SIGFAULT:
>
> #define SIZE            8053759
> char buf1[SIZE] ;
>
> NO SIGFAULT:
>
> #define SIZE            8053758
> char buf1[SIZE] ;
>
> This SIGFAULT not observed on 2.6.23 kernels.
>
> What could be the changes that effecting from 2.6.23 to 2.6.29 kernels.
> Please share me the information.

is "ulimit -a" the same on both kernels?

Can you provide a small program that shows the issue?