CentOS virt - Jan 2018 - Xen 4.6.6-9 (with XPTI meltdown mitigation) packages making their way to centos-virt-xen-testing

On Tue, Jan 23, 2018 at 06:20:39PM -0600, Kevin Stange
wrote:
> On 01/23/2018 05:57 PM, Karl Johnson wrote:
> > 
> > 
> > On Tue, Jan 23, 2018 at 4:50 PM, Nathan March <nathan at gt.net
> > <mailto:nathan at gt.net>> wrote:
> > 
> >     Hi,
> > 
> >     > Hmm.. isn't this the ldisc bug that was discussed a few
months ago on this
> >     list,
> >     > and a patch was applied to virt-sig kernel aswell?
> >     >
> >     > Call trace looks similar..
> > 
> >     Good memory! I'd forgotten about that despite being the one
who ran
> >     into it.
> > 
> >     Looks like that patch was just removed in 4.9.75-30 which I just
> >     upgraded
> >     this system to: http://cbs.centos.org/koji/buildinfo?buildID=21122
> >     <http://cbs.centos.org/koji/buildinfo?buildID=21122>
> >     Previously I was on 4.9.63-29 which does not have this problem,
and does
> >     have the ldisc patch. So I guess the question is for Johnny, why
was it
> >     removed?
> > 
> >     In the meantime, I'll revert the kernel and follow up if I see
any
> >     further
> >     problems.
> > 
> > 
> > IIRC the patch has been removed from the spec file because it has been
> > merged upstream in 4.9.71.
> 
> The IRC discussion I found in my log indicates that it was removed
> because it didn't apply cleanly due to changes when updating to 4.9.75,
> yet I don't think anyone independently validated that the changes made
> are equivalent to the patch that was removed.  I was never able to
> reproduce this issue, so I didn't investigate it myself.
>
Sounds like the patch is still needed :)

Anyone up to re-porting it to 4.9.75+ ?


-- Pasi
 
> -- 
> Kevin Stange
> Chief Technology Officer
> Steadfast | Managed Infrastructure, Datacenter and Cloud Services
> 800 S Wells, Suite 190 | Chicago, IL 60607
> 312.602.2689 X203 | Fax: 312.602.2688
> kevin at steadfast.net | www.steadfast.net

On 01/24/2018 01:01 AM, Pasi K?rkk?inen wrote:
> On Tue, Jan 23, 2018 at 06:20:39PM -0600, Kevin Stange wrote:
>> On 01/23/2018 05:57 PM, Karl Johnson wrote:
>>>
>>>
>>> On Tue, Jan 23, 2018 at 4:50 PM, Nathan March <nathan at gt.net
>>> <mailto:nathan at gt.net>> wrote:
>>>
>>>     Hi,
>>>
>>>     > Hmm.. isn't this the ldisc bug that was discussed a
few months ago on this
>>>     list,
>>>     > and a patch was applied to virt-sig kernel aswell?
>>>     >
>>>     > Call trace looks similar..
>>>
>>>     Good memory! I'd forgotten about that despite being the one
who ran
>>>     into it.
>>>
>>>     Looks like that patch was just removed in 4.9.75-30 which I
just
>>>     upgraded
>>>     this system to:
http://cbs.centos.org/koji/buildinfo?buildID=21122
>>>     <http://cbs.centos.org/koji/buildinfo?buildID=21122>
>>>     Previously I was on 4.9.63-29 which does not have this problem,
and does
>>>     have the ldisc patch. So I guess the question is for Johnny,
why was it
>>>     removed?
>>>
>>>     In the meantime, I'll revert the kernel and follow up if I
see any
>>>     further
>>>     problems.
>>>
>>>
>>> IIRC the patch has been removed from the spec file because it has
been
>>> merged upstream in 4.9.71.
>>
>> The IRC discussion I found in my log indicates that it was removed
>> because it didn't apply cleanly due to changes when updating to
4.9.75,
>> yet I don't think anyone independently validated that the changes
made
>> are equivalent to the patch that was removed.  I was never able to
>> reproduce this issue, so I didn't investigate it myself.
>>
> 
> Sounds like the patch is still needed :)
> 
> Anyone up to re-porting it to 4.9.75+ ?
It looked, at first glance, like 4.9.71 fixed it .. I guess not in all cases



-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: OpenPGP digital signature
URL:
<http://lists.centos.org/pipermail/centos-virt/attachments/20180124/f3ccc7dd/attachment-0002.sig>

> -----Original Message-----
> From: CentOS-virt [mailto:centos-virt-bounces at centos.org] On Behalf Of
> Johnny Hughes
> Sent: Wednesday, January 24, 2018 6:39 AM
> To: centos-virt at centos.org
> Subject: Re: [CentOS-virt] Xen 4.6.6-9 (with XPTI meltdown mitigation)
> packages making their way to centos-virt-xen-testing
> 
> On 01/24/2018 01:01 AM, Pasi K?rkk?inen wrote:
> > On Tue, Jan 23, 2018 at 06:20:39PM -0600, Kevin Stange wrote:
> >> On 01/23/2018 05:57 PM, Karl Johnson wrote:
> >>>
> >>>
> >>> On Tue, Jan 23, 2018 at 4:50 PM, Nathan March <nathan at
gt.net
> >>> <mailto:nathan at gt.net>> wrote:
> >>>
> >>>     Hi,
> >>>
> >>>     > Hmm.. isn't this the ldisc bug that was discussed
a few months ago
> on this
> >>>     list,
> >>>     > and a patch was applied to virt-sig kernel aswell?
> >>>     >
> >>>     > Call trace looks similar..
> >>>
> >>>     Good memory! I'd forgotten about that despite being
the one who ran
> >>>     into it.
> >>>
> >>>     Looks like that patch was just removed in 4.9.75-30 which
I just
> >>>     upgraded
> >>>     this system to:
http://cbs.centos.org/koji/buildinfo?buildID=21122
> >>>     <http://cbs.centos.org/koji/buildinfo?buildID=21122>
> >>>     Previously I was on 4.9.63-29 which does not have this
problem, and
> does
> >>>     have the ldisc patch. So I guess the question is for
Johnny, why was it
> >>>     removed?
> >>>
> >>>     In the meantime, I'll revert the kernel and follow up
if I see any
> >>>     further
> >>>     problems.
> >>>
> >>>
> >>> IIRC the patch has been removed from the spec file because it
has been
> >>> merged upstream in 4.9.71.
> >>
> >> The IRC discussion I found in my log indicates that it was removed
> >> because it didn't apply cleanly due to changes when updating
to 4.9.75,
> >> yet I don't think anyone independently validated that the
changes made
> >> are equivalent to the patch that was removed.  I was never able to
> >> reproduce this issue, so I didn't investigate it myself.
> >>
> >
> > Sounds like the patch is still needed :)
> >
> > Anyone up to re-porting it to 4.9.75+ ?
> 
> It looked, at first glance, like 4.9.71 fixed it .. I guess not in all
cases
I'm happy to do testing here if anyone's able to help with a patch, does
look like reverting to 4.9.63-29 solved it for me in the interm.