Bug#1117959: ipv6_route flags RTF_ADDRCONF and RTF_PREFIX_RT are not cleared when static on-link routes are added during IPv6 address configuration
Hi Salvatore,
On Fri, 2025-11-21 at 12:07 +0100, Salvatore Bonaccorso wrote:
> HI Garri,
>
> On Fri, Nov 21, 2025 at 11:07:39AM +0100, Garri Djavadyan wrote:
> > On Sat, 2025-11-15 at 14:02 +0100, Garri Djavadyan wrote:
> > > On Mon, 2025-11-10 at 17:54 +0100, Fernando Fernandez Mancera
> > > wrote:
> > > >
> > > >
> > > > On 10/25/25 11:21 PM, Garri Djavadyan wrote:
> > > > > On Sat, 2025-10-25 at 16:53 +0200, Salvatore Bonaccorso
> > > > > wrote:
> > > > > > Hi Garri,
> > > > > >
> > > > > > On Sat, Oct 18, 2025 at 01:39:02AM -0700, Stephen Hemminger
> > > > > > wrote:
> > > > > > > On Thu, 16 Oct 2025 00:12:40 +0200
> > > > > > > Garri Djavadyan <g.djavadyan@gmail.com> wrote:
> > > > > > >
> > > > > > > > Hi Everyone,
> > > > > > > >
> > > > > > > > A year ago I noticed a problem with handling ipv6_route
> > > > > > > > flags
> > > > > > > > that in
> > > > > > > > some scenarios can lead to reachability issues. It was
> > > > > > > > reported
> > > > > > > > here:
> > > > > > > >
> > > > > > > > https://bugzilla.kernel.org/show_bug.cgi?id=219205
> > > > > > > >
> > > > > > > >
> > > > > > > > Also it was recently reported in the Debian tracker
> > > > > > > > after
> > > > > > > > checking if
> > > > > > > > the latest Debian stable is still affected:
> > > > > > > >
> > > > > > > > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1117959
> > > > > > > >
> > > > > > > >
> > > > > > > > Unfortunately, the Debian team cannot act on the report
> > > > > > > > because
> > > > > > > > no one
> > > > > > > > from the upstream kernel team has confirmed if the
> > > > > > > > report
> > > > > > > > in
> > > > > > > > the
> > > > > > > > upstream tracker is valid or not. Therefore, I am
> > > > > > > > checking
> > > > > > > > if
> > > > > > > > anyone
> > > > > > > > can help confirm if the observed behavior is indeed a
> > > > > > > > bug.
> > > > > > > >
> > > > > > > > Many thanks in advance!
> > > > > > > >
> > > > > > > > Regards,
> > > > > > > > Garri
> > > > > > > >
> > > > > > >
> > > > > > > Linux networking does not actively use kernel bugzilla.
> > > > > > > I forward the reports to the mailing list, that is all.
> > > > > > > After than sometimes developers go back and update
> > > > > > > bugzilla
> > > > > > > but it is not required or expected.
> > > > > >
> > > > > > Garri, best action would likely be to really post your full
> > > > > > report on
> > > > > > netdev directly.
> > > > > >
> > > > > > Regards,
> > > > > > Salvatore
> > > > >
> > > > >
> > > > > Thank you for your suggestions Stephen and Salvatore.
> > > > >
> > > > > Below is the full report that was originally posted to the
> > > > > kernel
> > > > > bugzilla a year ago. It is still reproducible with fresher
> > > > > kernels.
> > > > >
> > > > > -----BEGIN REPORT-----
> > > > > I noticed that the ipv6_route flags RTF_ADDRCONF and
> > > > > RTF_PREFIX_RT
> > > > > are
> > > > > not cleared when static on-link routes are added during IPv6
> > > > > address
> > > > > configuration, and it leads to situations when the kernel
> > > > > updates
> > > > > the
> > > > > static on-link routes with expiration time.
> > > > >
> > > >
> > > > This is indeed a bug, I have a patch already and I am doing
> > > > some
> > > > testing
> > > > before sending it to net.git. I hope it can be sent tomorrow.
> > > >
> > > > Thanks,
> > > > Fernando.
> > >
> > >
> > > For the record, Fernando submitted the patch for review to net-
> > > next:
> > >
> > > https://lore.kernel.org/netdev/20251115095939.6967-1-fmancera@suse.de/
> > >
> > >
> > > Regards,
> > > Garri
> >
> > The patch has landed on linux-next:
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=f72514b3c5698e4b900b25345e09f9ed33123de6
>
> That is great :)
>
> > Salvatore, could you please clarify a few questions to which I
> > could
> > not find clear answers in the Debian Linux kernel handbook?
> >
> > - Should the patch first make it to the mainline or stable upstream
> > trees before it is considered for acceptance in the Debian trees?
> > - Will it be acceptable for both stable and oldstable Debian
> > releases
> > to include the fix considering that it can be seen as a security
> > issue
> > in some corner cases?
>
> There is really no need for a Debian specific approach here in my
> opinion. The patch has the correct fixes tags upstream so once it is
> in mainline it should make it the way down to the stable series.
>
> As you know Debian follows the upstream stable series. In case a fix
> is now known to be okay (needs to be at least in mainline, ideally
> already queued for stable), we might pick the commit as well for a
> next upload in case it is urgent enough. But usually the best thing
> is
> monitor the situation and help that the patch goes accepted in the
> desired stable series and it will land in Debian correspondly as well
> with the next rebased upload (both done at point release times but as
> well in DSAs via a security upload).
>
> The maybe bit shorter answer is, that the fix should be ideally at
> least in mainline and well vetted that to go to the stable series,
> and
> might then be picked as well in advance bit earlier than the actual
> release in the upstream series.
>
> Does this answers your question?
Yes, it answers the question in the best possible way! :)
Thank you so much for the details, Salvatore. Everything is clear now.
Regards,
Garri
Reply to: