preparing for 4.2.1 and 4.1.4

preparing for 4.2.1 and 4.1.4

[Jan Beulich]

All,

with the aim of getting both stable releases out the door in the
second half of November, please check the respective branches
for backports of bug fixes in -unstable that you think would
validly belong into those trees (an initial backport of some tools
side patches is known to be still missing).

My intention is to tag -rc1 on both trees towards the end of this
week (of course provided the trees get pushes from their staging
counterparts of everything that has got put there by then).

Jan

Re: preparing for 4.2.1 and 4.1.4

[Olaf Hering]

On Mon, Oct 29, Jan Beulich wrote:

> with the aim of getting both stable releases out the door in the
> second half of November, please check the respective branches
> for backports of bug fixes in -unstable that you think would
> validly belong into those trees (an initial backport of some tools
> side patches is known to be still missing).

What about these tools patches for 4.2.1?
The first one is also important for 4.1.4.

changeset:   26079:b3b03536789a
hotplug/Linux: close lockfd after lock attempt


changeset:   25941:795c493fe561
pygrub: always append --args
changeset:   26018:ecc7627ca6d7
pygrub: correct typo in --args assignment


changeset:   26006:8b6870d686d6
hotplug/Linux: Remove tracing (bash -x) from network-nat script

changeset:   26007:fe756682cc7f
xenballoond.init: remove 4 from default runlevel

changeset:   26008:eecb528583d7
xend/pvscsi: fix passing of SCSI control LUNs
changeset:   26009:2dbfa4d2e107
xend/pvscsi: fix usage of persistant device names for SCSI devices
changeset:   26010:cff10030c6ea
xend/pvscsi: update sysfs parser for Linux 3.0

changeset:   26011:b6fb4e63b946
stubdom: fix parallel build by expanding CROSS_MAKE

changeset:   26077:33348baecf37
stubdom: fix compile errors in grub

changeset:   26081:02064298ebcb
stubdom: fix rpmlint warning spurious-executable-perm

changeset:   26082:8cf26ace9ca0
blktap2/libvhd: fix rpmlint warning spurious-executable-perm

changeset:   26083:3fbeb019d522
blktap: fix rpmlint warning spurious-executable-perm

changeset:   26084:fe9a0eb9aaaa
hotplug: install hotplugpath.sh as data file

changeset:   26085:e32f4301f384
stubdom: install stubdompath.sh as data file

changeset:   26086:ba6b1db89ec8
hotplug/Linux: correct sysconfig tag in xendomains

changeset:   26087:6239ace16749
hotplug/Linux: install sysconfig files as data files



There are also two other HVM related patches, nice to have:

changeset:   26108:79185dcdf558
hvmloader: Reserve FE700000-FE800000 in physical memory map for guest use.

changeset:   26093:4ae08ca5500f
hvm: handle PoD and grant pages in HVMOP_get_mem_type


Olaf

Re: preparing for 4.2.1 and 4.1.4

[Jan Beulich]

>>> On 30.10.12 at 14:29, Olaf Hering <olaf@aepfle.de> wrote:
> On Mon, Oct 29, Jan Beulich wrote:
> 
>> with the aim of getting both stable releases out the door in the
>> second half of November, please check the respective branches
>> for backports of bug fixes in -unstable that you think would
>> validly belong into those trees (an initial backport of some tools
>> side patches is known to be still missing).
> 
> What about these tools patches for 4.2.1?

Tools patches get taken care of by IanJ (i.e. you should
probably have Cc-ed him) , but first of all - did you check what
is in 4.2-staging already (seems to me you didn't)?

Stubdom ones I'm not sure about - Ian, Ian?

> The first one is also important for 4.1.4.
> 
> changeset:   26079:b3b03536789a
> hotplug/Linux: close lockfd after lock attempt
> 
> 
> changeset:   25941:795c493fe561
> pygrub: always append --args
> changeset:   26018:ecc7627ca6d7
> pygrub: correct typo in --args assignment
> 
> 
> changeset:   26006:8b6870d686d6
> hotplug/Linux: Remove tracing (bash -x) from network-nat script
> 
> changeset:   26007:fe756682cc7f
> xenballoond.init: remove 4 from default runlevel
> 
> changeset:   26008:eecb528583d7
> xend/pvscsi: fix passing of SCSI control LUNs
> changeset:   26009:2dbfa4d2e107
> xend/pvscsi: fix usage of persistant device names for SCSI devices
> changeset:   26010:cff10030c6ea
> xend/pvscsi: update sysfs parser for Linux 3.0
> 
> changeset:   26011:b6fb4e63b946
> stubdom: fix parallel build by expanding CROSS_MAKE
> 
> changeset:   26077:33348baecf37
> stubdom: fix compile errors in grub
> 
> changeset:   26081:02064298ebcb
> stubdom: fix rpmlint warning spurious-executable-perm
> 
> changeset:   26082:8cf26ace9ca0
> blktap2/libvhd: fix rpmlint warning spurious-executable-perm
> 
> changeset:   26083:3fbeb019d522
> blktap: fix rpmlint warning spurious-executable-perm
> 
> changeset:   26084:fe9a0eb9aaaa
> hotplug: install hotplugpath.sh as data file
> 
> changeset:   26085:e32f4301f384
> stubdom: install stubdompath.sh as data file
> 
> changeset:   26086:ba6b1db89ec8
> hotplug/Linux: correct sysconfig tag in xendomains
> 
> changeset:   26087:6239ace16749
> hotplug/Linux: install sysconfig files as data files
> 
> 
> 
> There are also two other HVM related patches, nice to have:
> 
> changeset:   26108:79185dcdf558
> hvmloader: Reserve FE700000-FE800000 in physical memory map for guest use.

Isn't this just cosmetic?

> changeset:   26093:4ae08ca5500f
> hvm: handle PoD and grant pages in HVMOP_get_mem_type

This one even made it out of staging already (on both 4.2 and
4.1).

Jan

Re: preparing for 4.2.1 and 4.1.4

[Ian Jackson]

Jan Beulich writes ("Re: [Xen-devel] preparing for 4.2.1 and 4.1.4"):
> >>> On 30.10.12 at 14:29, Olaf Hering <olaf@aepfle.de> wrote:
> > On Mon, Oct 29, Jan Beulich wrote:
> > 
> >> with the aim of getting both stable releases out the door in the
> >> second half of November, please check the respective branches
> >> for backports of bug fixes in -unstable that you think would
> >> validly belong into those trees (an initial backport of some tools
> >> side patches is known to be still missing).
> > 
> > What about these tools patches for 4.2.1?
> 
> Tools patches get taken care of by IanJ (i.e. you should
> probably have Cc-ed him) , but first of all - did you check what
> is in 4.2-staging already (seems to me you didn't)?

I think many of these are.  I haven't been through Olaf's whole list.

> Stubdom ones I'm not sure about - Ian, Ian?

Yes, that's tools and within my remit.

I did a huge pile of backports late last week which seem to have
broken some things.  The bisector is working on it and I'm hoping it
will report fairly shortly ...

Ian.

Re: preparing for 4.2.1 and 4.1.4

[Ian Jackson]

Ian Jackson writes ("Re: [Xen-devel] preparing for 4.2.1 and 4.1.4"):
> I did a huge pile of backports late last week which seem to have
> broken some things.  The bisector is working on it and I'm hoping it
> will report fairly shortly ...

I'm sorry to report that the bisector has been having some internal
troubles.  It's taking less-trodden paths, mainly because it hasn't
tackled a stable branch like 4.2 for a long time (if at all).

I've been fixing the problems with the bisector.  If we don't have a
result from it by tomorrow morning I'll investigate the bug by
traditional methods...

Ian.

Re: preparing for 4.2.1 and 4.1.4

[Jan Beulich]

>>> On 30.10.12 at 15:18, Ian Jackson <Ian.Jackson@eu.citrix.com> wrote:
> I did a huge pile of backports late last week which seem to have
> broken some things.  The bisector is working on it and I'm hoping it
> will report fairly shortly ...

So what about the 4.1 tree? Should we cut RC1 there without any
tools backports?

Also, for both releases - what about their qemu tree(s)? There was
a regression reported for 4.1.3 compared to 4.1.2, fixing of which
supposedly involved a revert that apparently had been done in
4.2 already.

Jan

Re: preparing for 4.2.1 and 4.1.4

[Olaf Hering]

On Mon, Nov 05, Jan Beulich wrote:

> >>> On 30.10.12 at 15:18, Ian Jackson <Ian.Jackson@eu.citrix.com> wrote:
> > I did a huge pile of backports late last week which seem to have
> > broken some things.  The bisector is working on it and I'm hoping it
> > will report fairly shortly ...
> 
> So what about the 4.1 tree? Should we cut RC1 there without any
> tools backports?

At least the locking change should be backported to 4.1 as a bugfix:

changeset:   26079:b3b03536789a
hotplug/Linux: close lockfd after lock attempt


Olaf

Re: preparing for 4.2.1 and 4.1.4

[Olaf Hering]

On Tue, Oct 30, Ian Jackson wrote:

> Jan Beulich writes ("Re: [Xen-devel] preparing for 4.2.1 and 4.1.4"):
> > >>> On 30.10.12 at 14:29, Olaf Hering <olaf@aepfle.de> wrote:
> > Tools patches get taken care of by IanJ (i.e. you should
> > probably have Cc-ed him) , but first of all - did you check what
> > is in 4.2-staging already (seems to me you didn't)?
> 
> I think many of these are.  I haven't been through Olaf's whole list.

The locking.sh changes are not in 4.2 yet:

changeset:   26079:b3b03536789a
hotplug/Linux: close lockfd after lock attempt

This changeset fixes perl 5.8, the failure does not happen for me with
perl 5.10+:

changeset:   26146:92c52eb4b270
hotplug/Linux: Fix locking in tools/hotplug/Linux/locking.sh


Olaf