RE: Patches for piping to core pattern ?

From: Ravinandan Arakali (rarakali)
Date: Wed May 07 2008 - 16:48:30 EST


Upon adding some debugs, I find that it dumps core file upto
4096 bytes and after that, when pipe_write() notices that the
free memory is less than the requested size, it checks for
pending signals and exits from pipe_write().

But note that the user-level application on the read end of
the pipe does get to process the 4096 bytes since I get
a compressed core file. Problem is that the writer bails out
after 4K.

Any ideas ?

Thanks,
Ravi



-----Original Message-----
From: Ravinandan Arakali (rarakali)
Sent: Monday, May 05, 2008 6:55 PM
To: 'Randy Dunlap'
Cc: Linux Kernel
Subject: RE: Patches for piping to core pattern ?

Randy,
Thanks for the pointers to the patches.
I am now able to see the core getting streamed to a program (such as
gzip).

But I see that the core itself is quite small (truncated).
I see this behavior even when using the normal core_pattern (without the
"|" ). Wonder if it has something to do with the changes to binfmt_elf.c
? I'm not sure why these changes are required for pipe support.

Thanks,
Ravi


-----Original Message-----
From: Randy Dunlap [mailto:randy.dunlap@xxxxxxxxxx]
Sent: Monday, May 05, 2008 8:56 AM
To: Ravinandan Arakali (rarakali)
Cc: Linux Kernel
Subject: Re: Patches for piping to core pattern ?

On Sun, 4 May 2008 19:24:59 -0700 Ravinandan Arakali (rarakali) wrote:

> Hi,
> I am trying to port the "piping cores to an executable" to our kernel
> (2.6.10 based). I took Andi Kleen's patch (3/3) available at
> http://lwn.net/Articles/195310/ and some of the supporting functions
> from 2.6.19. But it's not working for me.
>
> Andi's patch mentions "Using the infrastructure created in previous
> patches implement support to pipe core dumps into programs."
>
> Can somebody point me to these previous patches ?

I don't know if these patches 1/3 and 2/3 represent the complete
infrastructure that Andi referred to, but you can find them here:

http://www.uwsg.indiana.edu/hypermail/linux/kernel/0608.1/2339.html
http://www.uwsg.indiana.edu/hypermail/linux/kernel/0608.1/2340.html


---
~Randy
--
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/