linux-um archives
 help / color / mirror / Atom feed
From: Ralph Passgang <ralph@debianbase.de>
To: user-mode-linux-devel@lists.sourceforge.net
Subject: Re: [uml-devel] uml modules broken in 2.6.8 (has something to do with memmove)
Date: Tue, 7 Sep 2004 11:22:40 +0200	[thread overview]
Message-ID: <200409071122.40804.ralph@debianbase.de> (raw)
In-Reply-To: <200409062005.08014.blaisorblade_spam@yahoo.it>

Am Montag, 6. September 2004 20:05 schrieben Sie:
> On Monday 06 September 2004 00:05, Ralph Paßgang wrote:
> > Hi all,
> >
> > I am not a kernel hacker, but I found with google that something changed
> > with memmove in 2.6.8, but I don't know enough of the kernel stuff to
> > know what it means:
> >
> > http://kerneltrap.org/node/view/3443 -> search for memmove and you find:
> > "i386: uninline memmove"
>
> Well, smart catch. If you know C, I can explain this. 

I know C a bit, but I am not were good in it... But I always want to learn to 
get better :)

> Before memmove was an 
> inline function defined in kernel headers, while in 2.6.8. I hope you are
> using 2.6.8.1, actually, because 2.6.8 is known-buggy for nfs. Well, no,
> you say 2.6.8. However, I guess that Debian-source includes the fix.

Yes, the debian package should include the fix. it is already the 5th version 
of the 2.6.8 kernel-source package in debian and the changelog says that a 
nfs-bugfix is included (with some more fixes).

I guess the debian package maintainer didn't used the 2.6.8.1 versionnumber, 
because it could produce a lot of problems in other packages/scripts that 
expect X.Y.Z format.

> > Has nobody besides me this problem? I can't use my uml without nat, so I
> > downgraded again to 2.6.6...
>
> Solutions in such cases are (waiting for the fix):

I waited for some days to see if somebody is saying something about memmove in 
the uml-mailinglists or if I find a 2.6.8.1-2 patch in the uml homepage. But 
because I didn't found anything I started asking myself, espacially because 
some other people seems to used the 2.6.8.1 without any problem...

> - compiling what you need in the kernel rather than as a module

I noticed that myself for another module (a bit ago)... But I don't want all 
iptables modules + ppp,pppoe + a lot other modules to be in the kernel rather 
than as module. If the problem just occours for one module, that is a fix 
indeed :)

> - if you get complaints about a certain symbol, i.e. memmove, just add a
> EXPORT_SYMBOL(memmove) and the prototype (or the appropriate include) in
> arch/um/kernel/ksyms.c and recompile the kernel (you don't need to
> recompile the modules).

I will try to fix the next problem (hopefully there will be no errors at all 
in upcoming versions *g*) on my own. Exporting kernelsymbols is not what I am 
used to do every day, but that seems not to be too difficult.

> In this case, apply this patch with patch -p1 -l:
>
> diff -puN arch/um/os-Linux/user_syms.c~uml-export-Symbols
> arch/um/os-Linux/user_syms.c ---
> uml-linux-2.6.8.1/arch/um/os-Linux/user_syms.c~uml-export-Symbols  
> 2004-08-29 14:40:56.000000000 +0200 +++
> uml-linux-2.6.8.1-paolo/arch/um/os-Linux/user_syms.c        2004-08-29
> 14:40:56.000000000 +0200 @@ -14,12 +14,14 @@
>
>  extern size_t strlen(const char *);
>  extern void *memcpy(void *, const void *, size_t);
> +extern void *memmove(void *, const void *, size_t);
>  extern void *memset(void *, int, size_t);
>  extern int printf(const char *, ...);
>
>  EXPORT_SYMBOL(strlen);
>  EXPORT_SYMBOL(memcpy);
>  EXPORT_SYMBOL(memset);
> +EXPORT_SYMBOL(memmove);
>  EXPORT_SYMBOL(printf);
>
>  EXPORT_SYMBOL(strstr);
>
> If you experience other module-related problems, you may want to use the
> attached patch, however I'm not sure it will apply on your tree.

It applied without a problem and I also have compiled it for debian sid and 
sarge. Now my 2.6.8 uml runs without any problems, even iptables_nat and 
pppoe works like charm :)

Thank you for the patch... Keep the good work up, I love UML :)...

--Ralph


-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_idP47&alloc_id\x10808&opÌk
_______________________________________________
User-mode-linux-devel mailing list
User-mode-linux-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-devel

      reply	other threads:[~2004-09-07  9:23 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2004-09-05 22:05 [uml-devel] uml modules broken in 2.6.8 (has something to do with memmove) Ralph Paßgang
2004-09-06 18:05 ` BlaisorBlade
2004-09-07  9:22   ` Ralph Passgang [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200409071122.40804.ralph@debianbase.de \
    --to=ralph@debianbase.de \
    --cc=user-mode-linux-devel@lists.sourceforge.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox