Bug#878811: dummy interface in bridge sticks in "configuring", leading to degraded system

[Michael Biebl]

Control: tags -1 + moreinfo

On Mon, 16 Oct 2017 21:57:15 +0200 Marc Haber
 wrote:
> Package: systemd
> Version: 235-2.0~zgSID+1
> Severity: normal
> Tags: upstream patch
> Forwarded: https://github.com/systemd/systemd/issues/6961
> 
> This is upstream issue 6961, where a dummy interface configured into a
> bridge gets stuck in "configuring" state, with the usual consequences of
> the network never getting "online", ultimately leading to a degraded
> system. Having a dummy interface in a bridge is a rather common idiom to
> force the bridge "up" even if there is nothing really "connected" yet.
> 
> I can confirm that the attached patch fixes the issue in systemd 235-2,
> and that it applies with minimal fuzz also applies to the systemd
> version in Debian stretch. Afaik, Susant Sahani, the developer of the
> patch, has submitted the patch, but it is not yet linked to the issue.

Hm, I don't see this patch applied in upstream systemd [1] but the
upstream issue has been closed.
Can you please verify if the issue still exists and if so, reopen the
upstream bug report.

Thanks,
Michael

[1] In case I missed it, can you point me at a git commit?

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#878811: dummy interface in bridge sticks in "configuring", leading to degraded system

[Marc Haber]

found #878811 235-3
thanks

On Tue, Oct 17, 2017 at 07:36:43AM +0200, Marc Haber wrote:
> I apologize, I filed this from a system with the fixed version
> installed. Both 232 from stretch and 235 from sid (and 234 which was the
> current version when the patch was developed) are affected, hence
> the suggestion to bring the fix into stretch through a point release.

I can confirm that 235-3 still has the issue, that the patch applies,
and fixes the issue.

Greetings
Marc

-- 
-
Marc Haber | "I don't trust Computers. They | Mailadresse im Header
Leimen, Germany|  lose things."Winona Ryder | Fon: *49 6224 1600402
Nordisch by Nature |  How to make an American Quilt | Fax: *49 6224 1600421

Bug#878811: dummy interface in bridge sticks in "configuring", leading to degraded system

[Marc Haber]

On Tue, Oct 17, 2017 at 12:11:07AM +0200, Michael Biebl wrote:
> Control:
> Am 16.10.2017 um 21:57 schrieb Marc Haber:
> > Package: systemd
> > Version: 235-2.0~zgSID+1
> 
> Seems the bts is not happy with that version.
> For proper version tracking, which version is affected by this?
> Is this a regression in v235 or (given your comment) is not a regression
> and happens with older releases as well.

I apologize, I filed this from a system with the fixed version
installed. Both 232 from stretch and 235 from sid (and 234 which was the
current version when the patch was developed) are affected, hence
the suggestion to bring the fix into stretch through a point release.

I do not remember, and do not have systems to test, to verify whether
this is already present in jessie.

Greetings
Marc
-- 
-
Marc Haber | "I don't trust Computers. They | Mailadresse im Header
Leimen, Germany|  lose things."Winona Ryder | Fon: *49 6224 1600402
Nordisch by Nature |  How to make an American Quilt | Fax: *49 6224 1600421

Bug#878811: dummy interface in bridge sticks in "configuring", leading to degraded system

[Michael Biebl]

Control:
Am 16.10.2017 um 21:57 schrieb Marc Haber:
> Package: systemd
> Version: 235-2.0~zgSID+1

Seems the bts is not happy with that version.
For proper version tracking, which version is affected by this?
Is this a regression in v235 or (given your comment) is not a regression
and happens with older releases as well.

Michael

-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Bug#878811: dummy interface in bridge sticks in "configuring", leading to degraded system

[Marc Haber]

Package: systemd
Version: 235-2.0~zgSID+1
Severity: normal
Tags: upstream patch
Forwarded: https://github.com/systemd/systemd/issues/6961

This is upstream issue 6961, where a dummy interface configured into a
bridge gets stuck in "configuring" state, with the usual consequences of
the network never getting "online", ultimately leading to a degraded
system. Having a dummy interface in a bridge is a rather common idiom to
force the bridge "up" even if there is nothing really "connected" yet.

I can confirm that the attached patch fixes the issue in systemd 235-2,
and that it applies with minimal fuzz also applies to the systemd
version in Debian stretch. Afaik, Susant Sahani, the developer of the
patch, has submitted the patch, but it is not yet linked to the issue.

Please consider applying this in sid and in a future stretch point
release. I can send you the patch for systemd 232 as well, but it is
really stupidly created by quilt refresh.

Greetings
Marc

-- Package-specific info:

-- System Information:
Debian Release: buster/sid
  APT prefers unstable
  APT policy: (500, 'unstable')
Architecture: amd64 (x86_64)

Kernel: Linux 4.13.7-zgsrv20080 (SMP w/4 CPU cores; PREEMPT)
Locale: LANG=de_DE.utf8, LC_CTYPE=de_DE.utf8 (charmap=UTF-8), LANGUAGE=en 
(charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages systemd depends on:
ii  adduser 3.116
ii  libacl1 2.2.52-3+b1
ii  libapparmor12.11.0-11
ii  libaudit1   1:2.8.1-1
ii  libblkid1   2.30.2-0.1
ii  libc6   2.24-17
ii  libcap2 1:2.25-1.1
ii  libcryptsetup4  2:1.7.5-1
ii  libgcrypt20 1.7.9-1
ii  libgpg-error0   1.27-3
ii  libidn111.33-2
ii  libip4tc0   1.6.1-2
ii  libkmod224-1
ii  liblz4-10.0~r131-2+b1
ii  liblzma55.2.2-1.3
ii  libmount1   2.30.2-0.1
ii  libpam0g1.1.8-3.6
ii  libseccomp2 2.3.1-2.1
ii  libselinux1 2.7-2
ii  libsystemd0 235-2.0~zgSID+1
ii  mount   2.30.2-0.1
ii  procps  2:3.3.12-3
ii  util-linux  2.30.2-0.1

Versions of packages systemd recommends:
ii  dbus1.11.20-1
ii  libpam-systemd  235-2.0~zgSID+1

Versions of packages systemd suggests:
pn  policykit-1
pn  systemd-container  

Versions of packages systemd is related to:
pn  dracut   
ii  initramfs-tools  0.130
ii  udev 235-2.0~zgSID+1

-- no debconf information
--- a/src/network/networkd-link.c
+++ b/src/network/networkd-link.c
@@ -94,6 +94,10 @@ static bool link_ipv4ll_enabled(Link *li
 }
 
 static bool link_ipv6ll_enabled(Link *link) {
+_cleanup_free_ char *disable_ipv6 = NULL;
+const char *p = NULL;
+int r;
+
 assert(link);
 
 if (!socket_ipv6_is_supported())
@@ -104,6 +108,12 @@ static bool link_ipv6ll_enabled(Link *li
 
 if (!link->network)
 return false;
+p = strjoina("/proc/sys/net/ipv6/conf/", link->ifname, 
"/disable_ipv6");
+r = read_one_line_file(p, _ipv6);
+if (r >= 0) {
+if (streq(disable_ipv6, "1"))
+return false;
+}
 
 return link->network->link_local & ADDRESS_FAMILY_IPV6;
 }
@@ -1694,6 +1704,9 @@ static int link_up_handler(sd_netlink *r
brought up later */
 log_link_warning_errno(link, r, "Could not bring up interface: 
%m");
 
+printf("link_up_handler= %s\n", link->ifname);
+link_check_ready(link);
+
 return 1;
 }