Re: Kernel 2.6.16-1.2107_FC5: eth0 dead

[Jack Howarth]

Dave,
   Doesn't the fact that the msi changes exist in the current FC4
kernel as well yet there are no networking problems in FC4 argue
against it being msi related? Wouldn't it more likely be caused
by some patch or change unique to the FC5 kernel?
               Jack

Kernel 2.6.16-1.2107_FC5: eth0 dead

[Dave Jones]

On Thu, May 04, 2006 at 12:14:38PM -0400, Jack Howarth wrote:
>  Dave,
>     Doesn't the fact that the msi changes exist in the current FC4
>  kernel as well yet there are no networking problems in FC4 argue
>  against it being msi related? Wouldn't it more likely be caused
>  by some patch or change unique to the FC5 kernel? 
There was a Xen update in the FC5 which wasn't changelogged for some reason.
It's believed this is responsible for the broken networking (even on non-xen
kernels).
(Which explains why the FC4 kernel didnt get the same bugs).

I'm working on getting a 2109 built with the xen changes reverted.

		Dave

-- 
http://www.codemonkey.org.uk

Kernel 2.6.16-1.2107_FC5: eth0 dead

[Wong Kwok-hon]

On 5/5/06, Dave Jones <davej(a)redhat.com&gt; wrote:
>  On Thu, May 04, 2006 at 12:14:38PM -0400, Jack Howarth wrote:
>   > Dave,
>   >    Doesn't the fact that the msi changes exist in the current FC4
>   > kernel as well yet there are no networking problems in FC4 argue
>   > against it being msi related? Wouldn't it more likely be caused
>   > by some patch or change unique to the FC5 kernel?
>
>  There was a Xen update in the FC5 which wasn't changelogged for some reason.
>  It's believed this is responsible for the broken networking (even on non-xen
> kernels).
>  (Which explains why the FC4 kernel didnt get the same bugs).
>
>  I'm working on getting a 2109 built with the xen changes reverted.
>  
Would new version release to fix the error ?
Because lots of people complaining it... :-)

Kernel 2.6.16-1.2107_FC5: eth0 dead

[Dave Jones]

On Fri, May 05, 2006 at 01:34:28AM +0800, Wong Kwok-hon wrote:
>  On 5/5/06, Dave Jones <davej(a)redhat.com&gt; wrote:
>  >On Thu, May 04, 2006 at 12:14:38PM -0400, Jack Howarth wrote:
>  > > Dave,
>  > >    Doesn't the fact that the msi changes exist in the current FC4
>  > > kernel as well yet there are no networking problems in FC4 argue
>  > > against it being msi related? Wouldn't it more likely be caused
>  > > by some patch or change unique to the FC5 kernel?
>  >
>  >There was a Xen update in the FC5 which wasn't changelogged for some 
>  >reason.
>  >It's believed this is responsible for the broken networking (even on 
>  >non-xen kernels).
>  >(Which explains why the FC4 kernel didnt get the same bugs).
>  >
>  >I'm working on getting a 2109 built with the xen changes reverted.
>  
>  Would new version release to fix the error ?
>  Because lots of people complaining it... :-) 
Yes. It's building right now.
I've also asked for 2107 to be pulled from updates in the meantime.

		Dave
-- 
http://www.codemonkey.org.uk

Kernel 2.6.16-1.2107_FC5: eth0 dead

[Jonathan Berry]

On 5/4/06, Dave Jones <davej(a)redhat.com&gt; wrote:
>  On Fri, May 05, 2006 at 01:34:28AM +0800, Wong Kwok-hon wrote:
>   > On 5/5/06, Dave Jones <davej(a)redhat.com&gt; wrote:
>   > >There was a Xen update in the FC5 which wasn't changelogged for some
>   > >reason.
>   > >It's believed this is responsible for the broken networking (even on
>   > >non-xen kernels).
>   > >(Which explains why the FC4 kernel didnt get the same bugs).
>   > >
>   > >I'm working on getting a 2109 built with the xen changes reverted.
>   >
>   > Would new version release to fix the error ?
>   > Because lots of people complaining it... :-)
>
>  Yes. It's building right now.
>  I've also asked for 2107 to be pulled from updates in the meantime. 
Hi Dave,
Just installed, and am writing from, 2111 from your repo.  It is
working great!  No problems with RHGB or Ethernet.  Good detective
work and excellent response.  Because we don't say it enough, thanks
for what you do for Fedora!
Now, who should we slap around for commiting the Xen patch without
mentioning it in the changelog? ;)
Jonathan