Re: [uml-user] UML locked up with high cpu usage

[BlaisorBlade <bla...@ya...>]

Alle 20:21, marted=EC 6 luglio 2004, David Houlden ha scritto:
> Further to my original post on this problem. I've gone back to 2.4.24-1um
> for my uml and applied the following from BlaisorBlade's split patches
>
> mini-changes.patch
> use_setjmp_wrapper.patch
> use-gettimeofday.patch
> remapper-fix.patch
> SIG-fileHandle-changes.patch
>
> I am now getting the uml locking up using large amounts of cpu in
> gettimeofday(). Backtrace below. Can anyone help me with this problem? Is
> this backtrace useful to anyone? BlaisorBlade?

I'm in a hurry for now, but try unapplying use-gettimeofday.patch and testi=
ng=20
if there is still any problem. Alternatively, SIG-fileHandle-changes.patch=
=20
could be the buggy one - maybe even remapper-fix.patch, but I would leave=20
this out for now. Also, Frank Sorenson <fr...@tu...> identified=20
critical problems even in mini-changes.patch, even if I don't believe it's=
=20
the problem here.

However, I assume that for if you use 2.4.24-1 without any further patches,=
=20
your UML works fine (and in this case, you can probably run that one; I sti=
ll=20
ask you to try the other patch combinations, if you can, to help identifyin=
g=20
the cause of the bug) - correct me if I am wrong, in that case maybe the fi=
x=20
will be some other one.
=2D-=20
Paolo Giarrusso, aka Blaisorblade
Linux registered user n. 292729