[Click] Re: kernel level click

[Roman Chertov]

Juncheol Park wrote:
>  Hi,
> 
>  Finally, I could successfully install the kernel level click
>  and use it for my experiments by solving the unresolved symbols problem
>  with insmod proclikefs.o.
> 
>  As I mentioned in other question's reply,
>  I added more compiling option at Makefile.in under click/linuxmodule as follows.
> 
>  proclikefs.o: proclikefs.c Makefile
> -       $(COMPILE) -D__KERNEL__ -DMODULE -c $<
> +       $(COMPILE) -D__KERNEL__ -DMODULE -DMODVERSIONS -include /usr/src/linux-$(shell uname -r)/include/linux/modversions.h -c $<
> 
>  I am not sure if this could be a good way to avoid such problems or not
>  because this problem might come from my own specific problem related to my current kernel.

I have also seen this on a 2.4.26 kernel.  I had problems with sb_lock 
being unresolved.  The problem went away after I installed a fresh copy 
of the source and patched it, while already running in the patched 
2.4.26 kernel.  Obviously this was a lame solution, but I think it means 
that for whatever reason there is an occosional difficulty resolving 
symbols correctly.

Roman


> 
>  If someone can clarify this, it would be really appreciated.
> 
>  Cheers,
> 
>  Jun
> 
> 
> On Sat, 13 Aug 2005, Juncheol Park wrote:
> 
> 
>> Hi,
>>
>> I would like to use kernel-level click rather than userlevel.
>> Could any body tell me which version of click
>> is really working well at kernel-level and how it works?
>>
>> Any real experience of kernel level click will be
>> very helpful to my research.
>>
>> And it would be greatly appreciated if you can provide
>> all the details such as linux-kernel version, madwifi version,
>> and so on.
>>
>> Thanks in advance,
>>
>> Jun
>>
>>
> 
> 
> 
> _______________________________________________
> click mailing list
> click at amsterdam.lcs.mit.edu
> https://amsterdam.lcs.mit.edu/mailman/listinfo/click
>