Gurol,

On Mon, 01 Sep 2003, Gurol Akman wrote:

> my attempts to use 'insmod' i.s.o. 'modprobe' also fail. 'insmod -f' doesn't
> help either... i know that i update my kernel image from 2.2.20 to 2.4.20

Stop right there.  You upgraded a kernel on a 2.2 system to 2.4.20?  You
need gcc 2.95.3, make 3.77, binutils 2.9.1.0.25, util-linux 2.10o, and
-- perhaps most important to your situation -- modutils 2.4.2.  An old
modutils would never load the newer module.

Did you upgraed your tool chain as described in linux-2.4/Documentation/Changes?

--brian


> properly and boot from the new image but perhaps i need to do additional
> things re: header files, etc. i'll check the debian forum re: kernel
> upgrades. so let's put this problem on hold for the moment.
> 
> to be able to continue my work for now, i switched from 'loadable' to
> 'linked in' drivers and was able to install my streams-xyz driver w/o any
> problems. but then i hit on another problem as follows...
> 
> i have an internal library (e.g., libcmn.a) which contains a set of
> kernel-space functions. since these functions are used by many of my
> drivers, i link all my drivers with this library at compile/link time. if i
> try to configure 2 of my drivers (i.e., streams-abc, streams-xyz) as linked
> in drivers (i.e., by creating the appropriate /usr/src/LiS/pkg
> sub-directories and Config, Makefile, driver.o files), and do a 'make
> install' i run into multiple definition problems re: some functions listed
> in the libcmn.a library. i also get additional complaints re: multiple
> definition of vsprintf and sprintf functions. as a result, i'm unable to get
> my 2 drivers installed at the same time. 
> 
> 1. how can resolve multiple definition problems caused by libcmn.a routines?
> 
> 2. what causes sprintf/vsprintf multiple definition problems & how can i
> alleviate them?
> 
> thanks for any suggestions,
> 
> gurol
> 
> -----Original Message-----
> From: Brian F. G. Bidulock [mailto:[EMAIL PROTECTED]
> Sent: Sunday, August 31, 2003 12:39 AM
> To: Gurol Akman
> Cc: '[EMAIL PROTECTED]'
> Subject: Re: [Linux-streams] dynamically loadable drivers
> 
> 
> Gurol,
> 
> Several things to check.  Make sure that you have installed you kernel
> and kernel modules properly.  modprobe looks for modules by kernel version.
> You can try 'insmod streams-xyz' instead of modprobe to see if it is
> a kernel installation problem or not.  If you get a version error from
> 'insmod' try 'insmod -f streams-xyz'.
> 
> --brian
> 
> On Sat, 30 Aug 2003, Gurol Akman wrote:
> 
> > 1. disabled the version checks and rebuilt the 2.4.20 kernel.
> > 2. rebooted my machine with the new kernel.
> > 3. compiled LiS 2.16.12 from scratch.
> > 4. removed -DMODVERSIONS flag and recompiled my driver.
> > 5. did 'streams start'
> > 6. did 'modprobe streams-xyz' 
> > 
> > I still get the same complaint re: kernel version not found. This doesn't
> > make sense!!! I was expecting modprobe to skip any version checks;
> > apparently it doesn't. I noted that LiS detects the kernel change I made
> and
> > does not use the -DMODVERSIONS flag when compiling its sub components. It
> > looks like 'modprobe' didn't take the change I made into account however.
> > 
> > I looks like there's no way for me to load my driver dynamically --
> > regardless of version checks being enabled or disabled in the kernel. Any
> > other suggestions to get this functionality to work? Thanks,
> > 
> > Gurol
> > 
> > -----Original Message-----
> > From: Brian F. G. Bidulock [mailto:[EMAIL PROTECTED]
> > Sent: Saturday, August 30, 2003 9:24 PM
> > To: Gurol Akman
> > Cc: '[EMAIL PROTECTED]'
> > Subject: Re: [Linux-streams] dynamically loadable drivers
> > 
> > 
> > Gurol,
> > 
> > Do you have the following at the top of your driver?
> > 
> > #include <linux/config.h>
> > #include <linux/version.h>
> > #ifdef MODVERSIONS
> > #include <linux/modversions.h>
> > #endif
> > #include <linux/module.h>
> > 
> > that might help.  I find that kernel symbol and module versions cause more
> > grief than good during developement and I usually turn them off in my
> > development kernels.
> > 
> > --brian
> > 
> > On Sat, 30 Aug 2003, Gurol Akman wrote:
> > 
> > > i believe so brian...
> > > 
> > > debian 3.0 comes with 2.2.20 kernel; i then installed & compiled 2.4.20
> > > kernel under /usr/src/linux and upgraded my system from 2.2.20 to
> 2.4.20.
> > i
> > > then compiled & installed LiS 2.16.12. finally, i compiled my driver on
> > this
> > > system and try to load it. this is when i get the errors listed. do you
> > see
> > > anything incorrect in this sequence of events? thanks,
> > > 
> > > gurol
> > >  
> > > -----Original Message-----
> > > From: Brian F. G. Bidulock [mailto:[EMAIL PROTECTED]
> > > Sent: Saturday, August 30, 2003 12:37 AM
> > > To: Gurol Akman
> > > Cc: '[EMAIL PROTECTED]'
> > > Subject: Re: [Linux-streams] dynamically loadable drivers
> > > 
> > > 
> > > Gurol,
> > > 
> > > Did you compile your streams module with the same kernel header files as
> > > you compiled the kernel?
> > > 
> > > --brian
> > > 
> > > On Fri, 29 Aug 2003, Gurol Akman wrote:
> > > 
> > > > 
> > > >    When  I  try  to use 'modprobe' to load my driver, i get the
> > following
> > > >    errors:
> > > > 
> > > > 
> > > > 
> > > >        couldn't find the kernel version the module was compiled for
> > > > 
> > > >        insmod /lib/modules/2.4.20/misc/streams-xyz.o failed
> > > > 
> > > > 
> > > > 
> > > >    1. I've compiled my driver with the -DMODULE and -DMODVERSIONS
> flags.
> > > > 
> > > >    2.  I  have the init_module() and cleanup_module() calls defined in
> > my
> > > >    driver code.
> > > > 
> > > >    3.  I'm using LiS 2.16.2 under Debian 3.0 with a kernel upgrade
> > (i.e.,
> > > >    2.4.20).
> > > > 
> > > > 
> > > > 
> > > >    Any ideas about what I may be missing or doing wrong? Thanks much,
> > > > 
> > > > 
> > > > 
> > > >    Gurol
> > > 
> > > -- 
> > > Brian F. G. Bidulock    � The reasonable man adapts himself to the �
> > > [EMAIL PROTECTED]    � world; the unreasonable one persists in  �
> > > http://www.openss7.org/ � trying  to adapt the  world  to himself. �
> > >                         � Therefore  all  progress  depends on the �
> > >                         � unreasonable man. -- George Bernard Shaw �
> > 
> > -- 
> > Brian F. G. Bidulock    � The reasonable man adapts himself to the �
> > [EMAIL PROTECTED]    � world; the unreasonable one persists in  �
> > http://www.openss7.org/ � trying  to adapt the  world  to himself. �
> >                         � Therefore  all  progress  depends on the �
> >                         � unreasonable man. -- George Bernard Shaw �
> 
> -- 
> Brian F. G. Bidulock    � The reasonable man adapts himself to the �
> [EMAIL PROTECTED]    � world; the unreasonable one persists in  �
> http://www.openss7.org/ � trying  to adapt the  world  to himself. �
>                         � Therefore  all  progress  depends on the �
>                         � unreasonable man. -- George Bernard Shaw �

-- 
Brian F. G. Bidulock    � The reasonable man adapts himself to the �
[EMAIL PROTECTED]    � world; the unreasonable one persists in  �
http://www.openss7.org/ � trying  to adapt the  world  to himself. �
                        � Therefore  all  progress  depends on the �
                        � unreasonable man. -- George Bernard Shaw �

_______________________________________________
Linux-streams mailing list
[EMAIL PROTECTED]
http://gsyc.escet.urjc.es/mailman/listinfo/linux-streams

Reply via email to