gmane.linux.kernel.kernelnewbies

Re: Communication between kernel and kernel module

Frank Ch. Eigler — 2012-05-23T11:48:16


One needs a proper hook mechanism such as kprobes or tracepoints.


The sched_switch tracepoint is probably the best spot for this.  A
systemtap script that listens for these events and simply traces them
to stdout looks like:

probe scheduler.ctxswitch { println(cpu(), ":", prev_tid, "->", next_tid) }

- FChE

A question about fs

王哲 — 2012-05-23T11:35:10

Hello Everyone:
      The first time to study fs ,and write a simple module to traverse the
double list with super_block.
and the output of module is different with the result from the command cat
/proc/filesystems.
 the output fs of module is 23,and the output fs of cat /proc/filesystems
is 26. why?

 My Environments:
# uname -r
2.6.39.1
# lsb_release -a
No LSB modules are available.
Distributor ID:    Ubuntu
Description:    Ubuntu 11.04
Release:    11.04
Codename:    natty
gcc version 4.5.2 (Ubuntu/Linaro 4.5.2-8ubuntu4)

my code is as followed:

#include<linux/module.h>
#include<linux/init.h>
#include<linux/sched.h>
#include <linux/fs.h>
#include <linux/list.h>
#include <linux/spinlock.h>

#define addr_super_blocks 0xc173d0a0
#define addr_sb_lock      0xc18f3ba0

#ifndef offsetof
#define offsetof(type, field)   ((long) &((type *)0)->field)
#endif   /* offsetof */

#ifndef container_of
#define container_of(ptr, type, member) ({                      \
        const typeof( ((type *)0)->member ) *__mptr = (p

Re: Incremental Linking

Sarbojit Ganguly — 2012-05-23T07:51:35

I think this might help (taking a cue from your last statement)

http://www.ibm.com/developerworks/linux/library/l-lkm/

On 23 May 2012 13:20, Sarbojit Ganguly <unixman.linuxboy< at >gmail.com> wrote:

Re: Incremental Linking

Sarbojit Ganguly — 2012-05-23T07:50:00

Can you please elaborate on that? What exactly you are trying to find
out and other details.


On 23 May 2012 12:51, somanath sahoo <bapi_mvit2004< at >yahoo.com> wrote:

Re: Incremental Linking

somanath sahoo — 2012-05-23T05:39:44

Hi Jonathan,
 
I came to know about "incremental linking" while reading "Embedded linux Primer".
 
 
Regards,
Somanath 


________________________________
From: Jonathan Neuschäfer <j.neuschaefer< at >gmx.net>
To: somanath sahoo <bapi_mvit2004< at >yahoo.com> 
Cc: "kernelnewbies< at >kernelnewbies.org" <kernelnewbies< at >kernelnewbies.org> 
Sent: Tuesday, May 22, 2012 8:47 PM
Subject: Re: Incremental Linking

On Tue, May 22, 2012 at 02:56:35AM -0700, somanath sahoo wrote:

Can you give us a pointer to where you read this?

Thanks,
    Jonathan Neuschäfer

Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Communication between kernel and kernel module

刘智猷 — 2012-05-23T01:07:31

Hi,
On 05/23/2012 02:18 AM, drewd77.gm wrote:
There are lots of ways to do it, maybe you don't need to write a module 
or hacking the kernel like that... Check perf and systemtap. They may help.
Regards,
LIU Zhiyou
_______________________________________________
Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Communication between kernel and kernel module

Vlad Dogaru — 2012-05-22T18:31:45

Kprobes [1] might be useful.  I for one haven't used Kprobes before,
but I'm sure others on this list can provide useful insight.

[1] http://lwn.net/Articles/132196/

Hope this helps,
Vlad

Communication between kernel and kernel module

drewd77.gm — 2012-05-22T18:18:32

Hi,

Is there any way to call a function defined in a kernel module from within
the kernel proper?

Here's what I'm trying to do:

I'm trying to keep track of when and for how long a process is being
executed on the processor, i.e. I need to know when context switches occur.

There is a function called context_switch() in linux/kernel/sched/core.c
that is called whenever a context switch happens. I would like to call a
function in a kernel module I've written from within context_switch() and
pass some data to it.

I've seen how to communicate between user space and kernel space, such as
using a file in /dev or using system calls, but not how to communicate
within the kernel itself.

Regards,
Drew
_______________________________________________
Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Incremental Linking

Mulyadi Santosa — 2012-05-22T17:40:28

perhaps you meant lazy binding? same thing like what glibc does in
user space I believe.

In short, symbols (functions etc) are not resolved right away, but
looked up and referenced when needed only.

Hopefully I point you the correct meaning. If not, feel free to CMIIW.

Re: [Q] USB urb completion handler

Ezequiel Garcia — 2012-05-22T17:37:36

Hi,


Actually, I was wrong because my driver (being a video capture driver)
submits several
isoc urbs at the same time.

I'm still not sure how this works regarding concurrency, but I guess
(or hope) the usb core calls
the completion handler one at a time OR the device sends interrupts
one at a time for each urb.

I know it's a terrible thing for me to guess (or hope), but these days
I've been testing
the driver under several conditions and it seems to work fine.
So I'm not worrying too much.

If someone more knowledgeable reads this and can put a light upon this issue,
I would appreciate it.

loody: Thanks for your answer regarding context ;)
Ezequiel.

ISO 8601 Timestamp <-> UNIX Timestamp

Dan Luedtke — 2012-05-22T17:12:33

Hi everyone,

are there any functions to convert from and to ISO 8601 formatted
timestamps? (Besides the mktime()-way)
Either ASCII/char* or maybe some kind of struct? Although I am trying
to avoid using them, I am just keen if there would be something like
this.

Regards.
Dan

Re: Incremental Linking

Jonathan Neuschäfer — 2012-05-22T15:17:52

Can you give us a pointer to where you read this?

Thanks,
Jonathan Neuschäfer


_______________________________________________
Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Why I can't compile a simple netfilter hook module?

Jonathan Neuschäfer — 2012-05-22T15:08:29

This symbol is declared in a section that's "only for userspace
compatibility" (netfilter_ipv4.h, line 10).

HTH,
Jonathan Neuschäfer

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Incremental Linking

somanath sahoo — 2012-05-22T09:56:35

Hi,
 
As newbie in linux kernel, I would like to understand the concept of  "incremental linking " w.r.t to linux kernel module. 
 
I have read that kernel loadable modules are being generated and inserted into a running kernel due to this "incremental linking " technology.
 
If anyone can provide me some pointers to understand the "incremental linking" concept, it will be needful.
 
 
Thanks,
Somanath_______________________________________________
Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: [Q] USB urb completion handler

loody — 2012-05-22T06:56:13

hi

2012/5/8 Ezequiel Garcia <elezegarcia< at >gmail.com>:

Both conditions will try to call complete functions.

Re: [Q] USB urb completion handler

loody — 2012-05-22T06:55:02

hi:

2012/5/8 Ezequiel Garcia <elezegarcia< at >gmail.com>:
how about multi-thread try to send urbs on the same pipe with the same
complete function?

回复: Why I can't compile a simple netfilter hook module?

c265n46 — 2012-05-22T04:38:00

what is your kernel version?
maybe you should use NF_INET_PRE_ROUTING instead.

2012-05-22



c265n46



发件人：Chir0n
发送时间：2012-05-08 12:14
主题：Why I can't compile a simple netfilter hook module?
收件人："kernelnewbies"<kernelnewbies< at >kernelnewbies.org>
抄送：

I'm using this Makefile:

obj-m += hello.o

all:
       make -C /lib/modules/$(shell uname -r)/build M=$(PWD) modules

clean:
       make -C /lib/modules/$(shell uname -r)/build M=$(PWD) clean


The hello.c is this:

#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/netfilter.h>
#include <linux/netfilter_ipv4.h>

static struct nf_hook_ops nfho;         //struct holding set of hook function options

//function to be called by hook
unsigned int hook_func(unsigned int hooknum, struct sk_buff **skb, const struct net_device *in, const struct net_device *out, int (*okfn)(struct sk_buff *))
{
 printk(KERN_INFO "packet dropped\n");

Re: Question about tlv message support

NorthPole — 2012-05-18T14:47:39

Hi all,
Now that I found some time to post

I just wanted to thank you all for the help, your posts helped a lot :-)

On Tue, Apr 17, 2012 at 12:34 PM, Javier Martinez Canillas
<martinez.javier< at >gmail.com> wrote:

Re: Can't find macro SAVE_ALL

王哲 — 2012-05-18T10:19:15

2012/5/18 HWSW Development <hwsw.development< at >googlemail.com>

         I have sloved the problem. this macro is written with assembly
，not c .
        No wonder that i don't search it out in the lxr.linux.no

  arch/x86/kernel/entry_32.s

 192 .macro SAVE_ALL
 193         cld
 194         PUSH_GS
 195         pushl_cfi %fs
 196         /*CFI_REL_OFFSET fs, 0;*/
 197         pushl_cfi %es
 198         /*CFI_REL_OFFSET es, 0;*/
 199         pushl_cfi %ds
 200         /*CFI_REL_OFFSET ds, 0;*/
 201         pushl_cfi %eax
 202         CFI_REL_OFFSET eax, 0
 203         pushl_cfi %ebp
 204         CFI_REL_OFFSET ebp, 0
 205         pushl_cfi %edi
 206         CFI_REL_OFFSET edi, 0
 207         pushl_cfi %esi
 208         CFI_REL_OFFSET esi, 0
 209         pushl_cfi %edx
 210         CFI_REL_OFFSET edx, 0
 211         pushl_cfi %ecx
 212         CFI_REL_OFFSET ecx, 0
 213         pushl_cfi %ebx
 214         CFI_REL_OFFSET ebx, 0
 215         movl $(__USER_DS), %edx
 216         movl %edx, %ds
 217         movl %

Re: Can't find macro SAVE_ALL

Bill Traynor — 2012-05-18T03:05:10


Couldn't you find it by doing something like this?  In your kernel source
tree:

$cd arch/x86
$grep -ircl ".macro SAVE_ALL" *




_______________________________________________
Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Can't find macro SAVE_ALL

王哲 — 2012-05-18T02:37:00

2012/5/17 Javier Martinez Canillas <martinez.javier< at >gmail.com>

   but the rearch result :
   Defined as a preprocessor macro in:

   - arch/cris/include/arch-v32/arch/irq.h, line
32<http://lxr.free-electrons.com/source/arch/cris/include/arch-v32/arch/irq.h#L32>
   - arch/cris/include/arch-v10/arch/irq.h, line
85<http://lxr.free-electrons.com/source/arch/cris/include/arch-v10/arch/irq.h#L85>
   - arch/sparc/include/asm/asmmacro.h, line
27<http://lxr.free-electrons.com/source/arch/sparc/include/asm/asmmacro.h#L27>

    is not in x86 Architecture.

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies< at >kernelnewbies.org
http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies