Re: [0/3] 2.6.19-rc2: known regressions

[Russell King <rmk+lkml@arm.linux.org.uk>]

On Sun, Oct 15, 2006 at 02:42:10PM +0200, Adrian Bunk wrote:
> On Sun, Oct 15, 2006 at 01:24:54PM +0100, Russell King wrote:
> > On Sat, Oct 14, 2006 at 01:14:58PM +0200, Adrian Bunk wrote:
> > > As usual, we are swamped with bug reports for regressions after -rc1.
> > > 
> > > For an easier reading (and hoping linux-kernel might not eat the emails), 
> > > I've splitted the list of known regressions in three emails:
> > >   [1/3] known unfixed regressions
> > >   [2/3] knwon regressions with workarounds
> > >   [3/3] known regressions with patches
> > 
> > There's a raft of ARM regressions as well (see
> > http://armlinux.simtec.co.uk/kautobuild/2.6.19-rc2/index.html), mostly
> > related to the IRQ changes, as well as this error:
> 
> Thanks, I'll look at them before preparing the next version of my 
> regressions list.
> 
> > sysctl_net.c:(.text+0x64a8c): undefined reference to `highest_possible_node_id'
> 
> This problem already got an entry a few hours ago:
> 
> Subject    : undefined reference to highest_possible_node_id
> References : http://lkml.org/lkml/2006/9/4/233
>              http://lkml.org/lkml/2006/10/15/11
> Submitter  : Olaf Hering <olaf@aepfle.de>
> Caused-By  : Greg Banks <gnb@melbourne.sgi.com>
>              commit 0f532f3861d2c4e5aa7dcd33fb18e9975eb28457
> Status     : unknown

Looking at this commit and the mails, it was known on the 4th September
that this patch caused build errors while this change was in -mm, yet it
still found its way into mainline on 2nd October.

Is anyone going to look at fixing this problem, or should we be asking
for the commit to be reverted?

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core