From: Stefan Gohr <sgohr@la...> - 2004-12-03 06:49:36
I had the same problem on a suse 9.1 with the bind rpm package,
the suse bind version "sees" the cpus from the host and tries to use =
them, though it runs in the uml machine. You have to build a new bind =
version from source using the --disable-threads configure option. I have =
done this- now binds 9 works fine on my uml.
Von: user-mode-linux-user-admin@... =
[mailto:user-mode-linux-user-admin@...] Im Auftrag von =
Gesendet: Donnerstag, 2. Dezember 2004 18:21
Betreff: [uml-user] bind9 utils seg fault
I'm running a 2.4.27 host + host-skas3-2.4.25.patch and have tried guest =
kernels 220.127.116.11 (patched), vanilla 2.6.9, and 2.6.10-rc2-mm4. On each of =
these guests, everything works fine except for tools that come with=20
bind9 (host, dig, nslookup, named, rndc, etc) from Debian sid. They all=20
seg fault immediately... strace and ltrace indicate the problem=20
happening at a slightly different point each time. Strace shows it's=20
most often after a munmap(), and ltrace seems to point to libisc as the=20
culprit. The same tools work fine on a real host. There's plenty of=20
memory available, only about 20M used out of 128M available to each =
The most unusual thing about the setup is that the host is an SMP=20
system, but that doesn't seem to be affecting anything else in the UML.
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.=20
User-mode-linux-user mailing list
On Friday 03 December 2004 07:47, Stefan Gohr wrote:
> I had the same problem on a suse 9.1 with the bind rpm package,
> the suse bind version "sees" the cpus from the host and tries to use them,
> though it runs in the uml machine.
> You have to build a new bind version
> from source using the --disable-threads configure option. I have done this-
> now binds 9 works fine on my uml.
> Stefan Gohr
Sorry, could you explain what is happening more clearly - I want to fix UML
bugs... but I'd also like to know what's happening (and your description does
not seems clear to me). Is --disable-threads about disabling some kind of
NPTL support? Does that option requires a 2.6 host kernel (which would
probably mean something similar to the previous question)? Any pointer to
give me? Thanks
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729