Potential ethernet bug

122 views
Skip to first unread message

pag...@gmail.com

unread,
Jan 28, 2016, 8:25:35 PM1/28/16
to qubes-devel
I have a supermicro motherboard with 2 onboard intel I210 nics.

Using 4.1.13-8.pvops.qubes.x86_64 and 4.1.13-7.pvops.qubes.x86_64 the ethernet in sys-net will never get an ip address.
It will constantly be resetting like the bug report here: https://bugs.centos.org/view.php?id=7034

However, if I use the 4.1.13-6.pvops.qubes.x86_64 kernel, it connects without errors and no NIC resets.

Marek Marczykowski-Górecki

unread,
Jan 28, 2016, 8:29:49 PM1/28/16
to pag...@gmail.com, qubes-devel
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
Are you sure it is the only difference? There is no change in any
network driver between those versions. The only difference is a security
fix[1].

https://github.com/QubesOS/qubes-secpack/blob/master/QSBs/qsb-023-2015.txt

- --
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJWqsCEAAoJENuP0xzK19csceYH/0w4MeRQ/W6ZmskrMTbFcLLb
MzjiNj8CTMhFVNZsuMHP9CZqFYDK9Lcpa0nbwS2lIk5/unSN5mBBvSde5m4DwgW4
6C+hb0n4mR4S1Lu3uWHykAfI1mBcKsainPpc+mYiZoZ93zLV5R/4mNKdXG43puLK
ZCK0gjf0HwZSzx98HqBKqXmrRHiEupbYvr7TfiOHTdXozDRofstefKmKY3Df5xnO
k9RytH+Fuat2RaxiK6Mwz1Cc2KinZG+WndxuNzWWhTQb6gMJeLE5k8K4D1d8Pqmd
jWnwM+SAKDBsXJwpSgZMbaBJxgzqivVRho2HiZxSocyPbqOFDmGpNeMqpMvm32Q=
=rP43
-----END PGP SIGNATURE-----

pag...@gmail.com

unread,
Jan 28, 2016, 8:35:22 PM1/28/16
to qubes-devel, pag...@gmail.com

Yes, that is the only difference. It just depends on which kernel I use in the Qubes grub list. Using the same sys-net vm.

Marek Marczykowski-Górecki

unread,
Jan 31, 2016, 5:52:52 PM1/31/16
to pag...@gmail.com, qubes-devel
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On Thu, Jan 28, 2016 at 05:33:41PM -0800, pag...@gmail.com wrote:
>
>
> On Thursday, January 28, 2016 at 8:29:49 PM UTC-5, Marek
> Marczykowski-Górecki wrote:
> > On Thu, Jan 28, 2016 at 05:20:25PM -0800, pag...@gmail.com <javascript:>
> > wrote:
> > > I have a supermicro motherboard with 2 onboard intel I210 nics.
> > >
> > > Using 4.1.13-8.pvops.qubes.x86_64 and 4.1.13-7.pvops.qubes.x86_64 the
> > > ethernet in sys-net will never get an ip address.
> > > It will constantly be resetting like the bug report here:
> > > https://bugs.centos.org/view.php?id=7034
> > >
> > > However, if I use the 4.1.13-6.pvops.qubes.x86_64 kernel, it connects
> > > without errors and no NIC resets.
> >
> > Are you sure it is the only difference? There is no change in any
> > network driver between those versions. The only difference is a security
> > fix[1].
> >
> > https://github.com/QubesOS/qubes-secpack/blob/master/QSBs/qsb-023-2015.txt
> >
>
> Yes, that is the only difference. It just depends on which kernel I use in
> the Qubes grub list. Using the same sys-net vm.

"grub list" - so it's about kernel in dom0? Just to make it clear, can
you send output of `uname -r` from working version and not working, from
*both* dom0 and sys-net? And preferably also kernel messages (both dom0
and sys-net, both versions) - so it can be compared.

- --
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJWrpA7AAoJENuP0xzK19cs8DwH/0Y/u3VvqWNxQKNzuCHq6OKb
SVsfykNaTJrl2HeKJ9/Rye3BaqCoF8j7WH1SMHBCL9sKEEoYBecujlE1XpYlC60O
3OlZ9gzt6FdhZUhfYa5Zbs+vW8unxQKgpbc1VcQZq03yuXTi2cs/RvmZ6KiYyhKJ
coFC+CNWKeS4yzV9wlIU1/KOw5cmv2eRZ3DbDvrcdMsVUhD0nQXfx/ekW5BQW9qG
32Hw8B1uxrU1XHKAk4/2f2jCcbf629+NSG9QqEX0VxjtD1PozfNSDNMpewkEhCq3
zmdObP5bhyazEWvfvls1IV9FRh5//oWvwcdwaSJ/dg5rUO5ynS4eFkxThh8oIpI=
=sGcj
-----END PGP SIGNATURE-----

pag...@gmail.com

unread,
Jan 31, 2016, 8:53:21 PM1/31/16
to qubes-devel, pag...@gmail.com

 

Here is the information you requested.


Working

dom0:

4.1.13-6.pvops.qubes.x86_64


sys-net:

4.1.13-8.pvops.qubes.x86_64



Non Working

dom0:

4.1.13-8.pvops.qubes.x86_64


sys-net:

4.1.13-8.pvops.qubes.x86_64


nonworking-dom0.txt
nonworking-sys-net.txt
working-dom0.txt
working-sys-net.txt

Marek Marczykowski-Górecki

unread,
Jan 31, 2016, 9:19:50 PM1/31/16
to pag...@gmail.com, qubes-devel, Eric Shelton
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On Sun, Jan 31, 2016 at 05:25:22PM -0800, pag...@gmail.com wrote:
>
>
> On Sunday, January 31, 2016 at 5:52:52 PM UTC-5, Marek Marczykowski-Górecki
> wrote:
> >
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA256
> >
> > On Thu, Jan 28, 2016 at 05:33:41PM -0800, pag...@gmail.com <javascript:>
> Here is the information you requested.
>
> Working
>
> dom0:
>
> 4.1.13-6.pvops.qubes.x86_64
>
> sys-net:
>
> 4.1.13-8.pvops.qubes.x86_64
>
>
> Non Working
>
> dom0:
>
> 4.1.13-8.pvops.qubes.x86_64
>
> sys-net:
>
> 4.1.13-8.pvops.qubes.x86_64


Ok, this is really interesting. I see those errors in sys-net:
IPv6: ADDRCONF(NETDEV_UP): enp0s0: link is not ready
xen:events: Failed to obtain physical IRQ 109
xen:events: Failed to obtain physical IRQ 110
xen:events: Failed to obtain physical IRQ 111
xen:events: Failed to obtain physical IRQ 112
xen:events: Failed to obtain physical IRQ 113
IPv6: ADDRCONF(NETDEV_UP): enp0s0: link is not ready
IPv6: ADDRCONF(NETDEV_UP): enp0s1: link is not ready
xen:events: Failed to obtain physical IRQ 115
xen:events: Failed to obtain physical IRQ 116
xen:events: Failed to obtain physical IRQ 117
xen:events: Failed to obtain physical IRQ 118
xen:events: Failed to obtain physical IRQ 119
IPv6: ADDRCONF(NETDEV_UP): enp0s1: link is not ready
IPv6: ADDRCONF(NETDEV_UP): vif2.0: link is not ready

Which is very similar to errors got with USB 3.0 controller in this
thread:
https://groups.google.com/d/msgid/qubes-users/20160129225226.GG1702%40mail-itl

And also here (network adapter):
http://lists.xen.org/archives/html/xen-devel/2016-01/msg03553.html
And according to the next message there, downgrade of dom0 kernel also
fixes the problem.

I have no idea what dom0 kernel version have to do here, because those
IRQ fails are returned directly from Xen, but apparently there is some
link between them. Since kernel changes between those versions are not
that big, it's feasible to just review them and try to understand what
is going on.

@Eric maybe you've already figured this out? Anyway letting you know
about this one, since you track Xen related issues (Big Thanks!).

- --
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJWrsC9AAoJENuP0xzK19cs9EcH/RRQ1H2eLta9+WyR1yZJWHPR
6Y6iYOP5LDBuoeDX+TDAPioXTO+Xb8loHyzB8DBcLMDcOuUWXDFOSSNolOxJyWx7
4FldK0U/tz9I6wrvMbUPcQD+KktdlIvKzkyEs6TYVrttyzY5J+sQ1+Vfwh597eHo
nzYWTP2kTT3ue2JBNIw5y/ETnkTB5jVlF8XT8SKSNLM5IWSWHvMIFcrrKL456xiU
S8twJzY8MF+Z5Y2oG/OMrVdAR0Zx0Zr9Wd97lRw8n69G9q//ZorkcLAEnXX1DP+l
OM5GofIL/i6c0JqOaZ5Albd0V9BnhtIoEvgJPOblKQ4xgKtWv0laleh+lhCG8g8=
=P+C3
-----END PGP SIGNATURE-----

Eric Shelton

unread,
Jan 31, 2016, 9:54:09 PM1/31/16
to qubes-devel, pag...@gmail.com, knock...@gmail.com
No.  But it sounds like there may be a problem in the XSA-155 related patches (the official Xen ones or the additional ones Qubes did), since that seems to be the only difference between the working and non-working kernels discussed above.

Eric

Marek Marczykowski-Górecki

unread,
Feb 9, 2016, 6:43:54 PM2/9/16
to Eric Shelton, qubes-devel, pag...@gmail.com
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On Sun, Jan 31, 2016 at 06:54:09PM -0800, Eric Shelton wrote:
> On Sunday, January 31, 2016 at 9:19:50 PM UTC-5, Marek Marczykowski-Górecki
> wrote:
> >
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA256
> >
> > On Sun, Jan 31, 2016 at 05:25:22PM -0800, pag...@gmail.com <javascript:>
> > <http://www.google.com/url?q=http%3A%2F%2Flists.xen.org%2Farchives%2Fhtml%2Fxen-devel%2F2016-01%2Fmsg03553.html&sa=D&sntz=1&usg=AFQjCNFqW62V4ZOdaeBDZocS1BJzgRC82w>
> > And according to the next message there, downgrade of dom0 kernel also
> > fixes the problem.
> >
> > I have no idea what dom0 kernel version have to do here, because those
> > IRQ fails are returned directly from Xen, but apparently there is some
> > link between them. Since kernel changes between those versions are not
> > that big, it's feasible to just review them and try to understand what
> > is going on.
> >
> > @Eric maybe you've already figured this out? Anyway letting you know
> > about this one, since you track Xen related issues (Big Thanks!).
> >
>
> No. But it sounds like there may be a problem in the XSA-155 related
> patches (the official Xen ones or the additional ones Qubes did), since
> that seems to be the only difference between the working and non-working
> kernels discussed above.

Just an update also in this thread: there (hopefully) fixed
kernel-4.1.13-8.3 package in qubes-dom0-update repo for R3.1.
Some details/links:
https://github.com/QubesOS/qubes-issues/issues/1734

- --
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJWunmxAAoJENuP0xzK19csOAMH/jE59o00DYgdhNCP3HZVg3EF
ny8UbG+NKt7FgSB+YT4mVRnVVsFMGK+dOz6Zw8PIgrawsgTCtVlW0eA7XSHQ41+P
+/OU/ufZwxXcr2lHLBkwz0zj3lyh+fFUXVVlTRId1SPma92kBnG4/pGTZ8kDv1Ou
udTtf5mPvfNhSq4mDFhja1xhR9baKHa823Kakje5pJzRccTtscVDn722TxIYP34Y
RysgALW3RpBdRUEla8Yyw9AU1zNf26LRE+RD4++XAO1IjcRavESkGY8t61IxOd6I
oedldR+KjRVoWM1rcHYVjKn7vJg2X8tvR4U7BIgoDrSxSK/YEDTVUEqgdn/q8II=
=QW/s
-----END PGP SIGNATURE-----

pag...@gmail.com

unread,
Feb 24, 2016, 12:25:27 PM2/24/16
to qubes-devel, knock...@gmail.com, pag...@gmail.com


Just wanted to let you know that updating dom0, I am now successfully running 4.1.13-9 kernel and the ethernet is working properly.

Outback Dingo

unread,
Feb 24, 2016, 6:20:56 PM2/24/16
to pag...@gmail.com, qubes-devel, knock...@gmail.com


Just wanted to let you know that updating dom0, I am now successfully running 4.1.13-9 kernel and the ethernet is working properly. ]



Boy if it were only that easy to fix the pci passthru issue im still having, to the best i can tell the issue is the vm kernel crashed on r8169 nic, with shared pci bus.
again, it works fine on fedora 23 / kernel Linux localhost.localdomain 4.3.5-300.fc23.x86_64 #1 SMP Mon Feb 1 03:18:41 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
with XEN xen_version            : 4.5.2

Im close to pretty convinced it is the kernel in the vm itself... always segfaulting, preventing networking from coming alive. 

--
You received this message because you are subscribed to the Google Groups "qubes-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qubes-devel...@googlegroups.com.
To post to this group, send email to qubes...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-devel/64d02b96-d386-4a10-a8da-80a64ac89d84%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Reply all
Reply to author
Forward
0 new messages