Big network problem on recent kernel update from xen/stable2.6.32.x

Big network problem on recent kernel update from xen/stable2.6.32.x

[Fantu]

I see only today how with last build of kernel do from last commit
(27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to connect
with rdp to domU, unable to use network from domU and viceversa
Also with build with 2.6.32.14 (from commit
e8734951b7a8d838050277696541a7e57183bdce) have problem
but my last build with 2.6.32.13 not have network problem, I can not
remember committing but i have build on 26 may if just remember
All these build have same config: 
http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
-- 
View this message in context: http://old.nabble.com/Big-network-problem-on-recent-kernel-update-from-xen-stable2.6.32.x-tp28805128p28805128.html
Sent from the Xen - Dev mailing list archive at Nabble.com.

RE: Big network problem on recent kernel update fromxen/stable2.6.32.x

[djmagee]

I discovered the same issue, with PV drivers only.  On Xen 3.4.3 and
4.0.0, I experienced the same problems.  Using one fully HVM linux domu,
and fully PV linux domu, and one windows domu with gplpv drivers, I
tested a 2.6.32.13 build from two weeks ago, and a 2.6.32.15 build from
Saturday with the same config.  PV network does not work properly with
2.6.32.15, but qemu emulated device in the HVM domain without PV drivers
works fine.  All three work just fine with 2.6.32.13.  Symptom is that
packets never seem to get back to the domu.  For example, I have VMs on
192.168.2.0/24 subnet, and dom0 routes traffic to local physical net on
192.168.1.0/24.  If I ping 192.168.1.1 from a domu with PV drivers, I
can see the echo request and reply in tcpdump on the bridge device on
dom0, but domu does not seem to see the reply.  In some other cases,
domu sees the reply, but reports 999-1001ms response times, which I know
are inaccurate by watching tcpdump output.  Traffic other than ICMP is
completely unreliable; the packets that do get through have incredibly
high latency.  I'm 100% sure I don't have iptables or other network
config problems.


-----Original Message-----
From: xen-devel-bounces@lists.xensource.com
[mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Fantu
Sent: Monday, June 07, 2010 9:20 AM
To: xen-devel@lists.xensource.com
Subject: [Xen-devel] Big network problem on recent kernel update
fromxen/stable2.6.32.x


I see only today how with last build of kernel do from last commit
(27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
connect
with rdp to domU, unable to use network from domU and viceversa
Also with build with 2.6.32.14 (from commit
e8734951b7a8d838050277696541a7e57183bdce) have problem
but my last build with 2.6.32.13 not have network problem, I can not
remember committing but i have build on 26 may if just remember
All these build have same config: 
http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
-- 
View this message in context:
http://old.nabble.com/Big-network-problem-on-recent-kernel-update-from-x
en-stable2.6.32.x-tp28805128p28805128.html
Sent from the Xen - Dev mailing list archive at Nabble.com.


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: Big network problem on recent kernel update fromxen/stable2.6.32.x

[Pasi Kärkkäinen]

On Mon, Jun 07, 2010 at 09:53:09AM -0400, djmagee@mageenet.net wrote:
> I discovered the same issue, with PV drivers only.  On Xen 3.4.3 and
> 4.0.0, I experienced the same problems.  Using one fully HVM linux domu,
> and fully PV linux domu, and one windows domu with gplpv drivers, I
> tested a 2.6.32.13 build from two weeks ago, and a 2.6.32.15 build from
> Saturday with the same config.  PV network does not work properly with
> 2.6.32.15, but qemu emulated device in the HVM domain without PV drivers
> works fine.  All three work just fine with 2.6.32.13.  Symptom is that
> packets never seem to get back to the domu.  For example, I have VMs on
> 192.168.2.0/24 subnet, and dom0 routes traffic to local physical net on
> 192.168.1.0/24.  If I ping 192.168.1.1 from a domu with PV drivers, I
> can see the echo request and reply in tcpdump on the bridge device on
> dom0, but domu does not seem to see the reply.  In some other cases,
> domu sees the reply, but reports 999-1001ms response times, which I know
> are inaccurate by watching tcpdump output.  Traffic other than ICMP is
> completely unreliable; the packets that do get through have incredibly
> high latency.  I'm 100% sure I don't have iptables or other network
> config problems.
> 

What bridge are you using? Is it virbr0, or some bridge with a physical NIC attached to it? 

-- Pasi

> 
> -----Original Message-----
> From: xen-devel-bounces@lists.xensource.com
> [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Fantu
> Sent: Monday, June 07, 2010 9:20 AM
> To: xen-devel@lists.xensource.com
> Subject: [Xen-devel] Big network problem on recent kernel update
> fromxen/stable2.6.32.x
> 
> 
> I see only today how with last build of kernel do from last commit
> (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
> with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
> connect
> with rdp to domU, unable to use network from domU and viceversa
> Also with build with 2.6.32.14 (from commit
> e8734951b7a8d838050277696541a7e57183bdce) have problem
> but my last build with 2.6.32.13 not have network problem, I can not
> remember committing but i have build on 26 may if just remember
> All these build have same config: 
> http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
> -- 
> View this message in context:
> http://old.nabble.com/Big-network-problem-on-recent-kernel-update-from-x
> en-stable2.6.32.x-tp28805128p28805128.html
> Sent from the Xen - Dev mailing list archive at Nabble.com.
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

Re: Big network problem on recent kernel update fromxen/stable2.6.32.x

[Michal Novotny]

On 06/07/2010 05:04 PM, Pasi Kärkkäinen wrote:
> On Mon, Jun 07, 2010 at 09:53:09AM -0400, djmagee@mageenet.net wrote:
>    
>> I discovered the same issue, with PV drivers only.  On Xen 3.4.3 and
>> 4.0.0, I experienced the same problems.  Using one fully HVM linux domu,
>> and fully PV linux domu, and one windows domu with gplpv drivers, I
>> tested a 2.6.32.13 build from two weeks ago, and a 2.6.32.15 build from
>> Saturday with the same config.  PV network does not work properly with
>> 2.6.32.15, but qemu emulated device in the HVM domain without PV drivers
>> works fine.  All three work just fine with 2.6.32.13.  Symptom is that
>> packets never seem to get back to the domu.  For example, I have VMs on
>> 192.168.2.0/24 subnet, and dom0 routes traffic to local physical net on
>> 192.168.1.0/24.  If I ping 192.168.1.1 from a domu with PV drivers, I
>> can see the echo request and reply in tcpdump on the bridge device on
>> dom0, but domu does not seem to see the reply.  In some other cases,
>> domu sees the reply, but reports 999-1001ms response times, which I know
>> are inaccurate by watching tcpdump output.  Traffic other than ICMP is
>> completely unreliable; the packets that do get through have incredibly
>> high latency.  I'm 100% sure I don't have iptables or other network
>> config problems.
>>
>>      
> What bridge are you using? Is it virbr0, or some bridge with a physical NIC attached to it?
>
> -- Pasi
>
>    

The bridge is one automatically created using the xen daemon but the 
configuration broke now when I rebooted so I need to make it working 
first since now even `xm list` is not working and the xend.log is 
completely empty :(

Michal
>> -----Original Message-----
>> From: xen-devel-bounces@lists.xensource.com
>> [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Fantu
>> Sent: Monday, June 07, 2010 9:20 AM
>> To: xen-devel@lists.xensource.com
>> Subject: [Xen-devel] Big network problem on recent kernel update
>> fromxen/stable2.6.32.x
>>
>>
>> I see only today how with last build of kernel do from last commit
>> (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
>> with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
>> connect
>> with rdp to domU, unable to use network from domU and viceversa
>> Also with build with 2.6.32.14 (from commit
>> e8734951b7a8d838050277696541a7e57183bdce) have problem
>> but my last build with 2.6.32.13 not have network problem, I can not
>> remember committing but i have build on 26 may if just remember
>> All these build have same config:
>> http://old.nabble.com/file/p28805128/config-2.6.32.15 config
>> -- 
>> View this message in context:
>> http://old.nabble.com/Big-network-problem-on-recent-kernel-update-from-x
>> en-stable2.6.32.x-tp28805128p28805128.html
>> Sent from the Xen - Dev mailing list archive at Nabble.com.
>>
>>
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.xensource.com
>> http://lists.xensource.com/xen-devel
>>
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.xensource.com
>> http://lists.xensource.com/xen-devel
>>      
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
>    


-- 
Michal Novotny<minovotn@redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat

RE: Big network problem on recent kernel updatefromxen/stable2.6.32.x

[djmagee]

I don't use the xend scripts to create a bridge, I use fedora/red hat config scripts to create the bridge at boot.  I use a bridge called xenbr0 that has no physical interface, but it does have an IP assigned in dom0, which handles routing for now.

-----Original Message-----
From: xen-devel-bounces@lists.xensource.com [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Pasi Kärkkäinen
Sent: Monday, June 07, 2010 11:04 AM
To: djmagee@mageenet.net
Cc: Jeremy Fitzhardinge; xen-devel@lists.xensource.com; Fantu
Subject: Re: [Xen-devel] Big network problem on recent kernel updatefromxen/stable2.6.32.x

On Mon, Jun 07, 2010 at 09:53:09AM -0400, djmagee@mageenet.net wrote:
> I discovered the same issue, with PV drivers only.  On Xen 3.4.3 and
> 4.0.0, I experienced the same problems.  Using one fully HVM linux domu,
> and fully PV linux domu, and one windows domu with gplpv drivers, I
> tested a 2.6.32.13 build from two weeks ago, and a 2.6.32.15 build from
> Saturday with the same config.  PV network does not work properly with
> 2.6.32.15, but qemu emulated device in the HVM domain without PV drivers
> works fine.  All three work just fine with 2.6.32.13.  Symptom is that
> packets never seem to get back to the domu.  For example, I have VMs on
> 192.168.2.0/24 subnet, and dom0 routes traffic to local physical net on
> 192.168.1.0/24.  If I ping 192.168.1.1 from a domu with PV drivers, I
> can see the echo request and reply in tcpdump on the bridge device on
> dom0, but domu does not seem to see the reply.  In some other cases,
> domu sees the reply, but reports 999-1001ms response times, which I know
> are inaccurate by watching tcpdump output.  Traffic other than ICMP is
> completely unreliable; the packets that do get through have incredibly
> high latency.  I'm 100% sure I don't have iptables or other network
> config problems.
> 

What bridge are you using? Is it virbr0, or some bridge with a physical NIC attached to it? 

-- Pasi

> 
> -----Original Message-----
> From: xen-devel-bounces@lists.xensource.com
> [mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Fantu
> Sent: Monday, June 07, 2010 9:20 AM
> To: xen-devel@lists.xensource.com
> Subject: [Xen-devel] Big network problem on recent kernel update
> fromxen/stable2.6.32.x
> 
> 
> I see only today how with last build of kernel do from last commit
> (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
> with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
> connect
> with rdp to domU, unable to use network from domU and viceversa
> Also with build with 2.6.32.14 (from commit
> e8734951b7a8d838050277696541a7e57183bdce) have problem
> but my last build with 2.6.32.13 not have network problem, I can not
> remember committing but i have build on 26 may if just remember
> All these build have same config: 
> http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
> -- 
> View this message in context:
> http://old.nabble.com/Big-network-problem-on-recent-kernel-update-from-x
> en-stable2.6.32.x-tp28805128p28805128.html
> Sent from the Xen - Dev mailing list archive at Nabble.com.
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: Big network problem on recent kernel update from xen/stable2.6.32.x

[Fantu]

i rebuild from commit 930edab1aba24f1807fd9ab8127006115bbd3422 and network
work, i think how the network problem was introduced by this commit:
commit	528ea798f02ccb03023e72cbda665ed5c3eec6b1
Merge branch 'xen/dom0/backend/netback-tasklet' into xen/next

* xen/dom0/backend/netback-tasklet:
  xen/netback: Use Kernel thread to replace the tasklet.
  xen/netback: Multiple tasklets support.
  xen/netback: Introduce a new struct type page_ext.
  xen/netback: Move global/static variables into struct xen_netbk.
 
Maybe some problem with new commit not see in xen/next but see in xen/stable
because in xen/stable there are merging of 2.6.32.x upstream and in xen/next
no, there are some differences between xen/next and xen/stable who cause
some unexpected behaviours


Fantu wrote:
> 
> I see only today how with last build of kernel do from last commit
> (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
> with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
> connect with rdp to domU, unable to use network from domU and viceversa
> Also with build with 2.6.32.14 (from commit
> e8734951b7a8d838050277696541a7e57183bdce) have problem
> but my last build with 2.6.32.13 not have network problem, I can not
> remember committing but i have build on 26 may if just remember
> All these build have same config: 
> http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
> 

-- 
View this message in context: http://old.nabble.com/Big-network-problem-on-recent-kernel-update-from-xen-stable2.6.32.x-tp28805128p28831366.html
Sent from the Xen - Dev mailing list archive at Nabble.com.

Re: Big network problem on recent kernel update from xen/stable2.6.32.x

[Pasi Kärkkäinen]

On Wed, Jun 09, 2010 at 07:45:01AM -0700, Fantu wrote:
> 
> i rebuild from commit 930edab1aba24f1807fd9ab8127006115bbd3422 and network
> work, i think how the network problem was introduced by this commit:
> commit	528ea798f02ccb03023e72cbda665ed5c3eec6b1
> Merge branch 'xen/dom0/backend/netback-tasklet' into xen/next
> 
> * xen/dom0/backend/netback-tasklet:
>   xen/netback: Use Kernel thread to replace the tasklet.
>   xen/netback: Multiple tasklets support.
>   xen/netback: Introduce a new struct type page_ext.
>   xen/netback: Move global/static variables into struct xen_netbk.
>  

At least this merge touches dom0 netback.. so if there was a bug in it, it could definitely cause the problems..

-- Pasi

> Maybe some problem with new commit not see in xen/next but see in xen/stable
> because in xen/stable there are merging of 2.6.32.x upstream and in xen/next
> no, there are some differences between xen/next and xen/stable who cause
> some unexpected behaviours
> 
> 
> Fantu wrote:
> > 
> > I see only today how with last build of kernel do from last commit
> > (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
> > with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
> > connect with rdp to domU, unable to use network from domU and viceversa
> > Also with build with 2.6.32.14 (from commit
> > e8734951b7a8d838050277696541a7e57183bdce) have problem
> > but my last build with 2.6.32.13 not have network problem, I can not
> > remember committing but i have build on 26 may if just remember
> > All these build have same config: 
> > http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
> > 
> 
> -- 
> View this message in context: http://old.nabble.com/Big-network-problem-on-recent-kernel-update-from-xen-stable2.6.32.x-tp28805128p28831366.html
> Sent from the Xen - Dev mailing list archive at Nabble.com.
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel

Re: Big network problem on recent kernel update from xen/stable2.6.32.x

[Jeremy Fitzhardinge]

On 06/09/2010 07:45 AM, Fantu wrote:
> i rebuild from commit 930edab1aba24f1807fd9ab8127006115bbd3422 and network
> work, i think how the network problem was introduced by this commit:
> commit	528ea798f02ccb03023e72cbda665ed5c3eec6b1
> Merge branch 'xen/dom0/backend/netback-tasklet' into xen/next
>
> * xen/dom0/backend/netback-tasklet:
>   xen/netback: Use Kernel thread to replace the tasklet.
>   xen/netback: Multiple tasklets support.
>   xen/netback: Introduce a new struct type page_ext.
>   xen/netback: Move global/static variables into struct xen_netbk.
>  
> Maybe some problem with new commit not see in xen/next but see in xen/stable
> because in xen/stable there are merging of 2.6.32.x upstream and in xen/next
> no, there are some differences between xen/next and xen/stable who cause
> some unexpected behaviours
>   

OK, that makes sense.  I hadn't had any problems with it, but it could
be v. dependent on how your system is configured.

Does "git revert -m 1 528ea798f02ccb03023e72cbda665ed5c3eec6b1" and
rebuilding fix the problem? If so, I'll revert it in stable until we can
sort it out.

Dongxiao, can you investigate this?

Thanks,
    J

>
> Fantu wrote:
>   
>> I see only today how with last build of kernel do from last commit
>> (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp domU
>> with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
>> connect with rdp to domU, unable to use network from domU and viceversa
>> Also with build with 2.6.32.14 (from commit
>> e8734951b7a8d838050277696541a7e57183bdce) have problem
>> but my last build with 2.6.32.13 not have network problem, I can not
>> remember committing but i have build on 26 may if just remember
>> All these build have same config: 
>> http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
>>
>>     
>   

RE: Big network problem on recent kernel update fromxen/stable2.6.32.x

[djmagee]

I can confirm that reverting 528ea798f02... fixes the problem for me.

Doug Magee

-----Original Message-----
From: xen-devel-bounces@lists.xensource.com
[mailto:xen-devel-bounces@lists.xensource.com] On Behalf Of Jeremy
Fitzhardinge
Sent: Wednesday, June 09, 2010 2:31 PM
To: Fantu
Cc: Xu, Dongxiao; xen-devel@lists.xensource.com
Subject: Re: [Xen-devel] Big network problem on recent kernel update
fromxen/stable2.6.32.x

On 06/09/2010 07:45 AM, Fantu wrote:
> i rebuild from commit 930edab1aba24f1807fd9ab8127006115bbd3422 and
network
> work, i think how the network problem was introduced by this commit:
> commit	528ea798f02ccb03023e72cbda665ed5c3eec6b1
> Merge branch 'xen/dom0/backend/netback-tasklet' into xen/next
>
> * xen/dom0/backend/netback-tasklet:
>   xen/netback: Use Kernel thread to replace the tasklet.
>   xen/netback: Multiple tasklets support.
>   xen/netback: Introduce a new struct type page_ext.
>   xen/netback: Move global/static variables into struct xen_netbk.
>  
> Maybe some problem with new commit not see in xen/next but see in
xen/stable
> because in xen/stable there are merging of 2.6.32.x upstream and in
xen/next
> no, there are some differences between xen/next and xen/stable who
cause
> some unexpected behaviours
>   

OK, that makes sense.  I hadn't had any problems with it, but it could
be v. dependent on how your system is configured.

Does "git revert -m 1 528ea798f02ccb03023e72cbda665ed5c3eec6b1" and
rebuilding fix the problem? If so, I'll revert it in stable until we can
sort it out.

Dongxiao, can you investigate this?

Thanks,
    J

>
> Fantu wrote:
>   
>> I see only today how with last build of kernel do from last commit
>> (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp
domU
>> with gplpv 0.11.0.213 unusable, ping lost or over 1 second unable to
>> connect with rdp to domU, unable to use network from domU and
viceversa
>> Also with build with 2.6.32.14 (from commit
>> e8734951b7a8d838050277696541a7e57183bdce) have problem
>> but my last build with 2.6.32.13 not have network problem, I can not
>> remember committing but i have build on 26 may if just remember
>> All these build have same config: 
>> http://old.nabble.com/file/p28805128/config-2.6.32.15 config 
>>
>>     
>   


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

RE: Big network problem on recent kernel update from xen/stable2.6.32.x

[Xu, Dongxiao]

Jeremy Fitzhardinge wrote:
> On 06/09/2010 07:45 AM, Fantu wrote:
>> i rebuild from commit 930edab1aba24f1807fd9ab8127006115bbd3422 and
>> network work, i think how the network problem was introduced by this
>> commit: commit	528ea798f02ccb03023e72cbda665ed5c3eec6b1
>> Merge branch 'xen/dom0/backend/netback-tasklet' into xen/next
>> 
>> * xen/dom0/backend/netback-tasklet:
>>   xen/netback: Use Kernel thread to replace the tasklet.
>>   xen/netback: Multiple tasklets support.
>>   xen/netback: Introduce a new struct type page_ext.
>>   xen/netback: Move global/static variables into struct xen_netbk.
>> 
>> Maybe some problem with new commit not see in xen/next but see in
>> xen/stable because in xen/stable there are merging of 2.6.32.x
>> upstream and in xen/next no, there are some differences between
>> xen/next and xen/stable who cause some unexpected behaviours 
>> 
> 
> OK, that makes sense.  I hadn't had any problems with it, but it could
> be v. dependent on how your system is configured.
> 
> Does "git revert -m 1 528ea798f02ccb03023e72cbda665ed5c3eec6b1" and
> rebuilding fix the problem? If so, I'll revert it in stable until we
> can sort it out.
> 
> Dongxiao, can you investigate this?

Sorry for the trouble cause by the patchset.
I will address the problem for windows ASAP.

Thanks,
Dongxiao

> 
> Thanks,
>     J
> 
>> 
>> Fantu wrote:
>> 
>>> I see only today how with last build of kernel do from last commit
>>> (27ed1b0e0dae5f1d5da5c76451bc84cb529128bd) do network of windows xp
>>> domU with gplpv 0.11.0.213 unusable, ping lost or over 1 second
>>> unable to connect with rdp to domU, unable to use network from domU
>>> and viceversa Also with build with 2.6.32.14 (from commit
>>> e8734951b7a8d838050277696541a7e57183bdce) have problem
>>> but my last build with 2.6.32.13 not have network problem, I can not
>>> remember committing but i have build on 26 may if just remember
>>> All these build have same config:
>>> http://old.nabble.com/file/p28805128/config-2.6.32.15 config

Re: Big network problem on recent kernel update from xen/stable2.6.32.x

[Pasi Kärkkäinen]

On Thu, Jun 10, 2010 at 09:26:07AM +0800, Xu, Dongxiao wrote:
> Jeremy Fitzhardinge wrote:
> > On 06/09/2010 07:45 AM, Fantu wrote:
> >> i rebuild from commit 930edab1aba24f1807fd9ab8127006115bbd3422 and
> >> network work, i think how the network problem was introduced by this
> >> commit: commit	528ea798f02ccb03023e72cbda665ed5c3eec6b1
> >> Merge branch 'xen/dom0/backend/netback-tasklet' into xen/next
> >> 
> >> * xen/dom0/backend/netback-tasklet:
> >>   xen/netback: Use Kernel thread to replace the tasklet.
> >>   xen/netback: Multiple tasklets support.
> >>   xen/netback: Introduce a new struct type page_ext.
> >>   xen/netback: Move global/static variables into struct xen_netbk.
> >> 
> >> Maybe some problem with new commit not see in xen/next but see in
> >> xen/stable because in xen/stable there are merging of 2.6.32.x
> >> upstream and in xen/next no, there are some differences between
> >> xen/next and xen/stable who cause some unexpected behaviours 
> >> 
> > 
> > OK, that makes sense.  I hadn't had any problems with it, but it could
> > be v. dependent on how your system is configured.
> > 
> > Does "git revert -m 1 528ea798f02ccb03023e72cbda665ed5c3eec6b1" and
> > rebuilding fix the problem? If so, I'll revert it in stable until we
> > can sort it out.
> > 
> > Dongxiao, can you investigate this?
> 
> Sorry for the trouble cause by the patchset.
> I will address the problem for windows ASAP.
> 

It's not really Windows.. all the Linux _PV_ guests stop to work!
(Afaik some HVM guests do work with the patch).

-- Pasi