[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] Commit moratorium for branching Xen 4.9



On 05/25/2017 01:46 PM, Julien Grall wrote:
> Hi Boris,
>
> On 25/05/17 18:08, Boris Ostrovsky wrote:
>> On 05/25/2017 11:54 AM, Julien Grall wrote:
>>> Hi,
>>>
>>> The patches fixing the 2 outstanding regressions have been merged.
>>
>> Is this:
>>
>> https://lists.xenproject.org/archives/html/xen-devel/2017-05/msg02340.html
>>
>>
>> being deferred to the next RC (if we are having one)?
>
> I missed that issue sorry. So far, this patch lacks of
> Acked-by/Reviewed-by but as a Tested-by.
>
> I am aware the patch that introduced the regression (efa9596e9d
> "x86/mm: fix incorrect unmapping of 2MB and 1GB pages") was attempting
> to prevent breaking down an EPT structure irreversibly when removing
> 2MB and 1GB pages. AFAIU, this commit will improve performance, but a
> boot regression was introduced on AMD platforms.


The regression only affects AMD guests that have maxmem > memory.

Since we have a fix in flight my suggestion would be to release-note
this issue but not revert the offending patch.

-boris

>
> As the first regression was not introduced by Xen 4.9 and we lack
> review bandwidth, I would suggest to revert the commit efa9596e9d
> until we get the patch [1] reviewed.
>
> If we managed to review it before the release, then we can consider to
> include both of them. But, at the moment, I would not consider the
> regression fixed by the commit efa9596e9d as a blocker for the release.
>
> Cheers,
>
> [1]
> https://lists.xenproject.org/archives/html/xen-devel/2017-05/msg02340.html
>


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.