lkml.org 
[lkml]   [1996]   [Oct]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: Solution to problem with loading modules in 2.1.4/2.1.5
    Well, to get libc, go to:

    ftp://sunsite.unc.edu/pub/linux/packages/GCC/private/tofu

    Get the file libc-5.4.10.bin.tar.gz...

    i know i said 5.4.9, but 5.4.10 just came out today or yesterday, so grab
    that ;>

    Also, get the file release.libc-5.4.10, which is also in the directory. it
    contains detailed instructions on how to install the new libc.

    These are the instructions on installing it:

    Assuming the libc binary tarfile is in /, and as root, execute:

    cd /
    rm -f /lib/libc.so /lib/libm.so
    rm -f /usr/lib/libc.so /usr/lib/libm.so
    rm -f /usr/include/iolibio.h /usr/include/iostdio.h
    rm -f /usr/include/wcstr.h /usr/include/mbstr.h
    rm -f /usr/include/ld_so_config.h /usr/include/localeinfo.h
    rm -rf /usr/include/netinet /usr/include/net /usr/include/pthread
    gzip -dc libc-5.4.10.bin.tar.gz | tar xvf -

    Then run 'ldconfig', and you are all done. :>

    -Rob

    On Sun, 20 Oct 1996, babydr wrote:

    >
    > Hello All, Is the below mentioned libc-* the one to
    > use under 2.1.x kernel tree & is there a good
    > page or Faq for installing these 'wonderful' changes
    > to the core of a kernel/filesystem. The last time I
    > played around with installing a -new- libc (pre-linux2.0)
    > to be able to compile the new & greatly improved linux-2.0.x
    > the 'scheme' I used, seemed to break 1/3 of all the 'tools'
    > (ie: net, file, a couple more ), I'm just a little gun shy
    > after that last episode.
    >
    > The only real cure (for me) was to purchase the slackware96.
    > This time I beleive that Slackware96(1/2) will be a little
    > ways off to wait for...
    >
    > PS. I've learned more from these mailing lists & the 5-7 times
    > I've had to reload because I've broken something than .... ;-)
    >
    > Any helpful suggestions, Please JimL
    >
    > On Sat, 19 Oct 1996, Rob Glover wrote:
    >
    > > Date: Sat, 19 Oct 1996 21:38:20 -0700 (PDT)
    > > From: Rob Glover <potato@dsnet.com>
    > > To: Al Longyear <longyear@netcom.com>
    > > Cc: linux-kernel@vger.rutgers.edu
    > > Subject: Re: Solution to problem with loading modules in 2.1.4/2.1.5
    > >
    > > upgrade to libc-5.4.9. simple as that :>
    > >
    > > On Sat, 19 Oct 1996, Al Longyear wrote:
    > >
    > > > For the past couple of days I have been trying to make the 2.1.4
    > > > kernel work on my laptop system.
    > > >
    > > > The laptop needs to use modules because the ethernet and modems are
    > > > both PCMCIA cards and the PCMCIA support needs modules.
    > > >
    > > > All attempts to load modules would fail with a message saying that the
    > > > module load had some really strange error number such as 1032201031.
    > > > (Clearly, this is not an error.)
    > > >
    > > > This was with the 2.0.0 version of the modules utilities. I checked
    > > > and this is the most current one as of last night.
    > > >
    > > > It would fail to load _any_ module, including the ones built by the
    > > > kernel's software such as the minix file system or the iso9660 file
    > > > system. The problem was not with David Hinds' PCMCIA code, but was
    > > > generic in nature.
    > > >
    > > > It turns out to be a problem with syscall interface which changed at
    > > > 2.1.4.
    > > >
    > > > I debugged the insmod program and looked at the source for the module
    > > > loader in the kernel. Yes, the return from the syscall(...) subroutine
    > > > in the libc library was returning the error condition.
    > > >
    > > > The version of the libc on that system is 5.4.7.
    > > >
    > > > The reason that I brought libc into this message is that the insmod
    > > > procedure uses syscall, not _syscall2 to interface to the operating
    > > > system. The problem is that the syscall function was built with the
    > > > old macros and only worked reliably on the the earlier 2.1 series
    > > > kernels.
    > > >
    > > > A solution, "just to make it work", was to change the insmod to use
    > > > the macros _syscall2 and _syscall5 directly to build the interface
    > > > procedures rather than to go through the syscall procedure.
    > > >
    > > > Lo and behold, modules were loaded and everything worked.
    > > >
    > > > That is not a recommended solution to anyone. The approach should be
    > > > to correct the syscall interface in the libc file so that it follows
    > > > the lines of the _syscall macros now in the kernel.
    > > >
    > > > I am only saying that this is what I found and this is the problem and
    > > > here is how I solved it for me for the time being. If you have similar
    > > > problems with loading modules then perhaps you have the same problem.
    > > >
    > > > In the mean time, I guess that I will just have to limp along with my
    > > > patched version of insmod that seems to work now and can get back to
    > > > my work on ppp.
    > > >
    > > > --
    > > > Al Longyear longyear@netcom.com Finger for PGP key
    > > > Design is about changing your mind until you get it correct.
    > > >
    > > >
    > >
    > >
    > >
    >
    >


    \
     
     \ /
      Last update: 2005-03-22 13:38    [W:0.024 / U:4.596 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site