Ed Co | 1 Jun 08:03 2004
Picon

test message [please ignore]

This is a test message. please ignore!!!

	
		
__________________________________
Do you Yahoo!?
Friends.  Fun.  Try the all-new Yahoo! Messenger.
http://messenger.yahoo.com/ 

--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
FAQ:           http://kernelnewbies.org/faq/

Aniruddha Patwardhan | 1 Jun 08:11 2004

[COMMERCIAL] : Various openings at well funded Technology Startup...

Hi all,

Plz go through following mail by a friend of mine about the openings at
a well funded Technology startup.

Plz reply to mentioned address directly in case you are interested.

Thanks,

Aniruddha Patwardhan
http://aniruddha.talk.to

====================================================
Hi everyone,

Gridlogics Technologies is a fast growing product startup in Pune that
is into bleeding-edge storage networking technologies. We are developing
a high performance I/O platform for high performance clusters.

We are looking to form a skilled development team that will be involved
in the design and devlopment of the product. Innovation is the key and
those who join will be participating in devloping & filing patents too.

We are looking for Sr Dev./ Tech leads/Product Architects with following
qualifications:
- Anywhere between 2-9 yrs of experience in C programming.
- Atleast 1-2 yrs experience specifically in Linux Kernel Programming
and overall good familiarity with Linux kernel Code
- Should have good knowledge of local file systems such as ext3/reiserFS
and distributed file systems such as NFS and AFS. Should understand the
(Continue reading)

Franz Reinhardt | 1 Jun 08:16 2004
Picon

_alloc_pages: allocation failed

Hi,

on my linux system, I've got the following message:
_alloc_pages: 0-order allocation failed. Killing application foo

Can anybody explain me, what's happening ? Is it just a memory leak of the 
application ? Or has it to do with reusing swapped memory pages ?

Any hints are highly appreciated.

Kind regards,

Franz

--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
FAQ:           http://kernelnewbies.org/faq/

Jan Hudec | 1 Jun 09:12 2004
Picon

Re: _alloc_pages: allocation failed

On Tue, Jun 01, 2004 at 08:16:52 +0200, Franz Reinhardt wrote:
> Hi,
> 
> on my linux system, I've got the following message:
> _alloc_pages: 0-order allocation failed. Killing application foo
> 
> Can anybody explain me, what's happening ? Is it just a memory leak of the 
> application ? Or has it to do with reusing swapped memory pages ?

It's an out-of-memory kill. You were out of memory. Completely. So the
kernel decided that the only way to get the page was to kill some
process. And did so, using SIGKILL.

-------------------------------------------------------------------------------
						 Jan 'Bulb' Hudec <bulb <at> ucw.cz>
aksingh | 1 Jun 09:46 2004

EXPORT_SYMTAB


Hi,

      one of modules used a varaible dev_netloop[MAX_NET_DEVICE], I want to
use this variable in another module I have can i do it using the macro
EXPORT_SYMBOL(dev_netloop) in the first module. Even when i do this, my
second module doesnt compile it says, undeclared identifier !
Any clues ?

thanks
Amit

--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
FAQ:           http://kernelnewbies.org/faq/

mohanlal jangir | 1 Jun 11:10 2004

Re: ip_local_deliver


----- Original Message -----
From: <aksingh <at> hss.hns.com>
To: <kernelnewbies <at> nl.linux.org>
Sent: Monday, May 31, 2004 9:55 PM
Subject: ip_local_deliver
>
> HI,
>
>   can I call ip_local_deliver/ip_route_output from my kernel module,
though
> it is not exported in /usr/src/linux-2.4/netsyms.c, is there no way that i
> can use it directly ?. Also, I see a problem with the linux kernel, any
> packet that comes from user applications, even one whose dest_ip_addr ==
> that of the local machine(the machine on which that application is
> running), even then the packet goes out of the machine, i.e if i write one
> such packet to ip using a raw socket,ip_local_deliver is not called for
> that packet immediately,
Calling ip_local_deliver for loopback packets immediately is bad idea ( I
think so). Think of tcp 3 way handshake, where tcp output function will call
ip output function. If ip output function immediately calls ip input
(ip_local_deliver) , ip input shall call tcp input and that will call again
tcp output for sending syn-ack. This process will continute further for ack.

 it first goes out of the ethernet interface and
> then is caught again, why so ? why is it kept this way ?
>
> thanks
> Amit
>
(Continue reading)

Franz Reinhardt | 1 Jun 13:22 2004
Picon

Re: _alloc_pages: allocation failed

Am Dienstag, 1. Juni 2004 09:12 schrieben Sie:
> On Tue, Jun 01, 2004 at 08:16:52 +0200, Franz Reinhardt wrote:
> > Hi,
> >
> > on my linux system, I've got the following message:
> > _alloc_pages: 0-order allocation failed. Killing application foo
> >
> > Can anybody explain me, what's happening ? Is it just a memory leak of
>
> the
>
> > application ? Or has it to do with reusing swapped memory pages ?
>
> It's an out-of-memory kill. You were out of memory. Completely. So the
> kernel decided that the only way to get the page was to kill some
> process. And did so, using SIGKILL.
>
How can this happen ? I mean, you can't allocate new memory if it's not 
available, you'll get a NULL-pointer if you try to malloc(). And if you try 
to access to that memory, the application will end up with an segmentation 
fault.

Franz

--
Kernelnewbies: Help each other learn about the Linux kernel.
Archive:       http://mail.nl.linux.org/kernelnewbies/
FAQ:           http://kernelnewbies.org/faq/

(Continue reading)

Momchil Velikov | 1 Jun 13:30 2004
Picon

Re: _alloc_pages: allocation failed

>>>>> "Franz" == Franz Reinhardt <fre <at> wenglor.de> writes:

Franz> Am Dienstag, 1. Juni 2004 09:12 schrieben Sie:
>> On Tue, Jun 01, 2004 at 08:16:52 +0200, Franz Reinhardt wrote:
>> > Hi,
>> >
>> > on my linux system, I've got the following message:
>> > _alloc_pages: 0-order allocation failed. Killing application foo
>> >
>> > Can anybody explain me, what's happening ? Is it just a memory leak of
>> 
>> the
>> 
>> > application ? Or has it to do with reusing swapped memory pages ?
>> 
>> It's an out-of-memory kill. You were out of memory. Completely. So the
>> kernel decided that the only way to get the page was to kill some
>> process. And did so, using SIGKILL.
>> 

Franz> How can this happen ? I mean, you can't allocate new memory if
Franz> it's not available, you'll get a NULL-pointer if you try to
Franz> malloc(). And if you try to access to that memory, the
Franz> application will end up with an segmentation fault.

  The point is that you haven't got null pointer from malloc, but a
valid address, in a valid virtual page, it just happens that there's
no physical page to back up the virtual one.  Which happens because
kernel allocates more virtual pages than physical pages in RAM + swap.

(Continue reading)

Franz Reinhardt | 1 Jun 14:57 2004
Picon

Re: _alloc_pages: allocation failed

Am Dienstag, 1. Juni 2004 13:30 schrieb kernelnewbies-bounce <at> nl.linux.org:
> >>>>> "Franz" == Franz Reinhardt <fre <at> wenglor.de> writes:
>
> Franz> Am Dienstag, 1. Juni 2004 09:12 schrieben Sie:
> >> On Tue, Jun 01, 2004 at 08:16:52 +0200, Franz Reinhardt wrote:
> >> > Hi,
> >> >
> >> > on my linux system, I've got the following message:
> >> > _alloc_pages: 0-order allocation failed. Killing application foo
> >> >
> >> > Can anybody explain me, what's happening ? Is it just a memory leak of
> >>
> >> the
> >>
> >> > application ? Or has it to do with reusing swapped memory pages ?
> >>
> >> It's an out-of-memory kill. You were out of memory. Completely. So the
> >> kernel decided that the only way to get the page was to kill some
> >> process. And did so, using SIGKILL.
>
> Franz> How can this happen ? I mean, you can't allocate new memory if
> Franz> it's not available, you'll get a NULL-pointer if you try to
> Franz> malloc(). And if you try to access to that memory, the
> Franz> application will end up with an segmentation fault.
>
>   The point is that you haven't got null pointer from malloc, but a
> valid address, in a valid virtual page, it just happens that there's
> no physical page to back up the virtual one.  Which happens because
> kernel allocates more virtual pages than physical pages in RAM + swap.
>
(Continue reading)

Learner | 1 Jun 15:25 2004
Picon

Re: Finding size of a raw device

Hi ,

 Probably I found the answer to this . 

On a filesystem, by doing an lseek the OS can 
track the size of the file via the inodes .

Thus, an lseek with SEEK_END would give the 
file size , as the inode tables provide the 
needed support for the OS .

But since a raw device bound to a file is a 
direct interface , there is no way in which 
the OS can figure out its limits , and thus 
lseek with SEEK_END would yield a zero .

It would be nice if somebody could confirm 
the above explanation .

Cheers !

 --- Learner <ruxyz <at> yahoo.com> wrote: > Hi ,
> 
>   I was trying to find the size of a raw partition 
> using the below program , but the output is not 
> correct. Can anybody suggest how to rectify this
> program.
> 
>   Is it that "lseek" should be used only on regular 
> files and not special files ?Is there a proc entry 
(Continue reading)


Gmane