loading modules

loading modules

[stefania magistrali]

Hi all!
I am working with a PowerPC 860T based board.
I have a problem with loading modules: I made:

make dep make clean make zImage  make modules  and make modules_install

to create my kernel and modules  , but when I try to load the module
hdlc.o with insmod, I have a segmentation fault. I am using busybox
version 0.60.3  and linux version 2.4.19pre6. Here I send the last part
of the strace sequence of the command "insmod  hdlc"
Can anyone help me?
Thanks very much in advance!
Stefania

query_module(NULL, 0, NULL, 0)          = 0
query_module(NULL, QM_MODULES, { /* 0 entries */ }, 0) = 0
mmap(NULL, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
0) = 0x30033000
query_module(NULL, QM_SYMBOLS, 0x30032e08, 16384, 21866) = -1 ENOSPC (No
space left on device)
mmap(NULL, 24576, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
0) = 0x30037000
query_module(NULL, QM_SYMBOLS, { /* 961 entries */ }, 961) = 0
--- SIGSEGV (Segmentation fault) ---
+++ killed by SIGSEGV +++


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: loading modules

[Magnus Damm]

This might be related to a powerpc-specific bug in the busybox
insmod implementation. Try the 0.60.4 release or use this patch:

http://opensource.se/projects/busybox/busybox-0.60.3-insmod_rewrite.patch

/ magnus

stefania magistrali wrote:
>
> Hi all!
> I am working with a PowerPC 860T based board.
> I have a problem with loading modules: I made:
>
> make dep make clean make zImage  make modules  and make modules_install
>
> to create my kernel and modules  , but when I try to load the module
> hdlc.o with insmod, I have a segmentation fault. I am using busybox
> version 0.60.3  and linux version 2.4.19pre6. Here I send the last part
> of the strace sequence of the command "insmod  hdlc"
> Can anyone help me?
> Thanks very much in advance!
> Stefania
>
> query_module(NULL, 0, NULL, 0)          = 0
> query_module(NULL, QM_MODULES, { /* 0 entries */ }, 0) = 0
> mmap(NULL, 16384, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
> 0) = 0x30033000
> query_module(NULL, QM_SYMBOLS, 0x30032e08, 16384, 21866) = -1 ENOSPC (No
> space left on device)
> mmap(NULL, 24576, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
> 0) = 0x30037000
> query_module(NULL, QM_SYMBOLS, { /* 961 entries */ }, 961) = 0
> --- SIGSEGV (Segmentation fault) ---
> +++ killed by SIGSEGV +++
>

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

loading modules

[blesson paul]

Hi all
		I am a newbie  to Kernel world. When I looked into the file system files, 
I found that the initialization function ( where the file system is 
registered) is “init_filesystem”  where filesystem can be coda, vfat etc. As 
far as know, the initialization function is
	int init_module(void)
		Then how kernel takes different initialization functions. I want to know 
whether my know how is wrong or not

Thanking in advance
regards
Blesson Paul





_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com

Re: loading modules

[Alexandre P. Nunes]

blesson paul wrote:

> Hi all
> I am a newbie to Kernel world. When I looked into the file system 
> files, I found that the initialization function ( where the file 
> system is registered) is "init_filesystem" where filesystem can be 
> coda, vfat etc. As far as know, the initialization function is
> int init_module(void)
> Then how kernel takes different initialization functions. I want to 
> know whether my know how is wrong or not


you have something like (function parameters supressed for simplifcation):

/* Kernel calls this in your module */
init_module()
{

do_something();
init_filesystem(...);

return whatever;
}


So the entry point for every kind of module, including filesystem 
modules, is init_module(), you don't implement init_filesystem in the 
module, you call it to let the kernel know you want to register a 
filesystem.

>
> Thanking in advance
> regards
> Blesson Paul


Cheers,

Alex