Skip to contentSkip to conversation

struct page_info

PATCH New Mukesh Rathor 9 participants 28 messages Apr 24, 2009 – May 12, 2009 1 message not in the archive

lore
Share & tools
lore
#0 lore raw

Hi,
I'm trying to figure out all the incarnations of a page. That led me to page_info{}. Now, looking at it, the shadow page is confusing.

	/* Page is in use: ((count_info & PGC_count_mask) != 0).  */
	..

        /* Page is in use as a shadow: count_info == 0. */
	..
	/* Page is on a free list: ((count_info & PGC_count_mask) == 0). */
	..

If count_info == 0, and it implies shadow page, how's that different from an unallocated page which also would've count_info=0 I presume?

Also, if "Page is on a free list:", I assume one of the PGC_ fields would be set?

thanks,
Mukesh

#1 lore raw
Quoted 11 lines
Quoted 1 line
Quoted 1 line

Mukesh Rathor <[email protected]> 24.04.09 05:15 >>>

Now, looking at it, the shadow page is confusing.

	/* Page is in use: ((count_info & PGC_count_mask) != 0).  */
	..

        /* Page is in use as a shadow: count_info == 0. */
	..
	/* Page is on a free list: ((count_info & PGC_count_mask) == 0). */
	..
Quoted 2 lines

If count_info == 0, and it implies shadow page, how's that different from an unallocated page which also would've count_info=0 I presume?

These are not if-and-only-if conditions, so you can't conclude the use of a page from just looking at count_info; only the other way around: from e.g. knowing the page is in use as a shadow, you'll know count_info must be zero.

Quoted 2 lines

Also, if "Page is on a free list:", I assume one of the PGC_ fields would be set?

Not would, but might.

Jan

#2 lore raw
Quoted 23 lines

From: [email protected] [mailto:[email protected]] On Behalf Of Jan Beulich
Sent: Thursday, April 23, 2009 11:51 PM

Quoted 11 lines
Quoted 1 line
Quoted 1 line

Mukesh Rathor <[email protected]> 24.04.09 05:15 >>>

Now, looking at it, the shadow page is confusing.

    /* Page is in use: ((count_info & PGC_count_mask) != 0).  */
    ..

        /* Page is in use as a shadow: count_info == 0. */
    ..
    /* Page is on a free list: ((count_info & PGC_count_mask) == 0). */
    ..
Quoted 2 lines

If count_info == 0, and it implies shadow page, how's that different from an unallocated page which also would've count_info=0 I presume?

These are not if-and-only-if conditions, so you can't conclude the use of a page from just looking at count_info; only the other way around: from e.g. knowing the page is in use as a shadow, you'll know count_info must be zero.

Quoted 4 lines
Quoted 2 lines

Also, if "Page is on a free list:", I assume one of the PGC_ fields would be set?

Not would, but might.

Would:

        ((page->count_info & PGC_count_mask) != 0 || page->count_info == 0)

definitively indicate if a page was in use?

Joe

#3 lore raw
Quoted 8 lines · Joseph
Quoted 4 lines
Quoted 2 lines

Also, if "Page is on a free list:", I assume one of the PGC_ fields would be set?

Not would, but might.

Would:

        ((page->count_info & PGC_count_mask) != 0 || page->count_info == 0)

definitively indicate if a page was in use?

Only PGC_broken|PGC_offlined may be set for a free page. Most free pages have count_info==0.

-- Keir

#27 lore raw
In reply to Jan Beulich
Quoted 11 lines · Jan Beulich
Quoted 11 lines
Quoted 1 line
Quoted 1 line

Mukesh Rathor <[email protected]> 24.04.09 05:15 >>>

Now, looking at it, the shadow page is confusing.

	/* Page is in use: ((count_info & PGC_count_mask) != 0).  */
	..

        /* Page is in use as a shadow: count_info == 0. */
	..
	/* Page is on a free list: ((count_info & PGC_count_mask) == 0). */
	..
Quoted 7 lines
Quoted 2 lines

If count_info == 0, and it implies shadow page, how's that different from an unallocated page which also would've count_info=0 I presume?

These are not if-and-only-if conditions, so you can't conclude the use of a page from just looking at count_info; only the other way around: from e.g. knowing the page is in use as a shadow, you'll know count_info must be zero.

Ahh, so I found a mfn on the floor no easy way of knowing where it belongs. Okey.

thanks
Mukesh

Quoted 6 lines
Quoted 2 lines

Also, if "Page is on a free list:", I assume one of the PGC_ fields would be set?

Not would, but might.

Jan

#4 lore raw

Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

In reply to Keir Fraser

Compilation logs for :-

# make install-xen
# make install-tools
are attached

Compiled  pvops enabled kernel 2.6.30-rc3-tip no errors After booting xen 3.4.gz and kernel system fails to start daemon xend:- Multiple messages on screen output:-
******************************************************************
File "/usr/sbin/xend" line 36 in <module> from xen.xend.server import SrvDaemon.
ImportError No module named xen.xend.server.
******************************************************************

Gnome desktop packages are dependent on python 2.6 ( vs 2.5 on Ubuntu Intrepid Server ) where the load procedure runs fine and xend also starts smoothly , if /etc/fstab gets properly edited

I found SrvDaemon on 9.04 in location:-

root@ServerJaunty:/# find /usr/lib -name "SrvDaemon*" -print

/usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.py /usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.pyc

On Ubuntu Server 8.10 location is :-

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.py

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.pyc

Boris

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#5 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Boris,

Try this:

apt-get install python-xml
echo oldxml >> /usr/lib/python2.6/site-packages/oldxml.pth

It works for me on Hardy (python2.5)...

Thiago

2009/4/25 Boris Derzhavets <[email protected]>

Quoted 36 lines

Compilation logs for :-

# make install-xen
# make install-tools
are attached

Compiled pvops enabled kernel 2.6.30-rc3-tip no errors After booting xen 3.4.gz and kernel system fails to start daemon xend:- Multiple messages on screen output:-
******************************************************************
File "/usr/sbin/xend" line 36 in <module> from xen.xend.server import SrvDaemon.
ImportError No module named xen.xend.server.
******************************************************************

Gnome desktop packages are dependent on python 2.6 ( vs 2.5 on Ubuntu Intrepid Server ) where the load procedure runs fine and xend also starts smoothly , if /etc/fstab gets properly edited

I found SrvDaemon on 9.04 in location:-

root@ServerJaunty:/# find /usr/lib -name "SrvDaemon*" -print

/usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.py /usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.pyc

On Ubuntu Server 8.10 location is :-

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.py /usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.pyc

Boris

Quoted 6 lines

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#6 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Thiago,

Reproduced instruction.
Uninstalled xen&tools, rebuild xen&tools, installed again.
Attempted to boot into Xen Unstable with pvops kernel.
Same failure to start xend.
Interpid (8.10)  Server does have python 2.5 and everything goes pretty straight as far as get compiled ( all packages in place).

Boris,

--- On Sat, 4/25/09, Thiago Camargo Martins Cordeiro <[email protected]> wrote:
From: Thiago Camargo Martins Cordeiro <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "Boris Derzhavets" <[email protected]>
Cc: "xen-devel" <[email protected]>, "Keir Fraser" <[email protected]>
Date: Saturday, April 25, 2009, 9:20 AM

Boris,

 Try this:

 apt-get install python-xml
 echo oldxml >> /usr/lib/python2.6/site-packages/oldxml.pth

It works for me on Hardy (python2.5)...

Thiago

2009/4/25 Boris Derzhavets <[email protected]>

Compilation logs for :-

# make install-xen
# make install-tools
are attached

Compiled  pvops enabled kernel 2.6.30-rc3-tip no errors After booting xen 3.4.gz and kernel system fails to start daemon xend:-

Multiple messages on screen output:-
******************************************************************
File "/usr/sbin/xend" line 36 in <module> from xen.xend.server import SrvDaemon.
ImportError No module named xen.xend.server.

******************************************************************

Gnome desktop packages are dependent on python 2.6 ( vs 2.5 on Ubuntu Intrepid Server ) where the load procedure runs fine and xend also starts smoothly , if /etc/fstab gets properly edited

I found SrvDaemon on 9.04 in location:-

root@ServerJaunty:/# find /usr/lib -name "SrvDaemon*"
 -print

/usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.py /usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.pyc

On Ubuntu Server 8.10 location is :-

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.py

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.pyc

Boris

_______________________________________________

Xen-devel mailing list

[email protected]

http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#7 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

In reply to Boris Derzhavets

Mounted remotely root file system of Ubuntu Server 9.04 on /mnt on Ubuntu Intrepid Server (8.10) Compiled xen on this box as follows:-

# cd /usr/src/xen-unstable.hg
# make dist
# ./install.sh /mnt

Boris
--- On Sat, 4/25/09, Boris Derzhavets <[email protected]> wrote:
From: Boris Derzhavets <[email protected]>
Subject: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "Keir Fraser" <[email protected]>
Cc: "xen-devel" <[email protected]>
Date: Saturday, April 25, 2009, 8:54 AM

Compilation logs for :-

# make install-xen
# make install-tools
are attached

Compiled  pvops enabled kernel 2.6.30-rc3-tip no errors After booting xen 3.4.gz and kernel system fails to start daemon xend:- Multiple messages on screen output:-
******************************************************************
File "/usr/sbin/xend" line 36 in <module> from xen.xend.server import SrvDaemon.
ImportError No module named xen.xend.server.
******************************************************************

Gnome desktop packages are dependent on python 2.6 ( vs 2.5 on Ubuntu Intrepid Server ) where the load procedure runs fine and xend also starts smoothly , if /etc/fstab gets properly edited

I found SrvDaemon on 9.04 in location:-

root@ServerJaunty:/# find /usr/lib -name "SrvDaemon*"
 -print

/usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.py /usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.pyc

On Ubuntu Server 8.10 location is :-

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.py

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.pyc

Boris

      _______________________________________________

Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#8 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Is there any way to point xen 3.4-rc3-pre source (xen-unstbale.hg) to
pick up for build phase python 2.5 , installed on  the system, instead of 2.6
used by gnome environment as default.
I believe, that SrvDaemon.py is a core of xend daemon itself
   It looks like it's not possible to perform make install-xen, make install-tools properly with python 2.6, regardless build itself seems to be completed with no errors. Through my experience it was possible to build Xen from source only on systems with python 2.4 (5) installed as default.

Looking at page :-
http://www.mail-archive.com/[email protected]/msg261640.html

I found statement :-

The reason for this is that Xen 3.0 is bound to Python 2.4 whereas Xen 3.2 is bound to Python 2.5.

Please, advise.

Boris.

--- On Sun, 4/26/09, Boris Derzhavets <[email protected]> wrote:
From: Boris Derzhavets <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "Keir Fraser" <[email protected]>
Cc: "xen-devel" <[email protected]>
Date: Sunday, April 26, 2009, 8:15 AM

Mounted remotely root file system of Ubuntu Server 9.04 on /mnt on Ubuntu Intrepid Server (8.10) Compiled xen on this box as follows:-

# cd /usr/src/xen-unstable.hg
# make dist
# ./install.sh /mnt

Boris
--- On Sat, 4/25/09, Boris Derzhavets <[email protected]> wrote:
From: Boris Derzhavets <[email protected]>
Subject: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "Keir Fraser" <[email protected]>
Cc: "xen-devel" <[email protected]>
Date: Saturday, April 25, 2009, 8:54 AM

Compilation logs for :-

# make install-xen
# make install-tools
are attached

Compiled  pvops enabled kernel 2.6.30-rc3-tip no errors After booting xen 3.4.gz and kernel system fails to start daemon xend:- Multiple messages on screen output:-
******************************************************************
File "/usr/sbin/xend" line 36 in <module> from xen.xend.server import SrvDaemon.
ImportError No module named xen.xend.server.
******************************************************************

Gnome desktop packages are dependent on python 2.6 ( vs 2.5 on Ubuntu Intrepid Server ) where the load procedure runs fine and xend also starts smoothly , if /etc/fstab gets properly edited

I found SrvDaemon on 9.04 in
location:-

root@ServerJaunty:/# find /usr/lib -name "SrvDaemon*"
 -print

/usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.py /usr/lib/python2.6/site-packages/xen/xend/server/SrvDaemon.pyc

On Ubuntu Server 8.10 location is :-

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.py

/usr/lib/python2.5/site-packages/xen/xend/server/SrvDaemon.pyc

Boris

      _______________________________________________

Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

      _______________________________________________

Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#9 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Boris Derzhavets writes ("Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 3 lines

Is there any way to point xen 3.4-rc3-pre source (xen-unstbale.hg) to pick up for build phase python 2.5 , installed on  the system, instead of 2.6 used by gnome environment as default.

You need to change the default python version used by setup.py' which calls the standard Python setup machinery. I'm no python expert but I think you might be able to get the Xen build system to do the right thing by making python' on $PATH point to the python version you want it to use.

It seems like a bug to me that it uses python' rather than having a variable PYTHON' but we probably won't fix that in 3.4 now.

Quoted 4 lines

I found statement :-

The reason for this is that Xen 3.0 is bound to Python 2.4 whereas Xen 3.2 is bound to Python 2.5.

That comment is about the Debian Xen packages and isn't relevant if you're building it yourself. When you build Xen your build version gets the python version you used baked into it.

Ian.

#10 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Could you please point to an exact location of setup.py to be edited. I found several files under xen-unstable.hg.

export PYTHONPATH=/usr/lib/python2.5/site-packages doesn't help.

Boris.

--- On Wed, 4/29/09, Ian Jackson <[email protected]> wrote:
From: Ian Jackson <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: [email protected]
Cc: "xen-devel" <[email protected]>
Date: Wednesday, April 29, 2009, 11:59 AM

Boris Derzhavets writes ("Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 2 lines

Is there any way to point xen 3.4-rc3-pre source (xen-unstbale.hg) to pick up for build phase python 2.5 , installed on  the system, instead of

2.6

Quoted 1 line

used by gnome environment as default.

You need to change the default python version used by setup.py' which calls the standard Python setup machinery. I'm no python expert but I think you might be able to get the Xen build system to do the right thing by making python' on $PATH point to the python version you want it to use.

It seems like a bug to me that it uses python' rather than having a variable PYTHON' but we probably won't fix that in 3.4 now.

Quoted 4 lines

I found statement :-

The reason for this is that Xen 3.0 is bound to Python 2.4 whereas Xen 3.2 is bound to Python 2.5.

That comment is about the Debian Xen packages and isn't relevant if you're building it yourself. When you build Xen your build version gets the python version you used baked into it.

Ian.

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#11 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Boris Derzhavets writes ("Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 2 lines

Could you please point to an exact location of setup.py to be edited. I found several files under xen-unstable.hg.

No, I mean, either change /usr[/local]/bin/python (which is probably a symlink) to point to the desired version of python, or set your main PATH variable to have a directory at the front containing `python' pointing to the version you want.

Quoted 2 lines

export PYTHONPATH=/usr/lib/python2.5/site-packages doesn't help.

No, I don't think it would.

Ian.

#26 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

In reply to Boris Derzhavets

Disregard this question.
Boris.

--- On Wed, 4/29/09, Boris Derzhavets <[email protected]> wrote:
From: Boris Derzhavets <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "Ian Jackson" <[email protected]>
Cc: "xen-devel" <[email protected]>
Date: Wednesday, April 29, 2009, 12:15 PM

Could you please point to an exact location of setup.py to be edited. I found several files under xen-unstable.hg.

export PYTHONPATH=/usr/lib/python2.5/site-packages doesn't help.

Boris.

--- On Wed, 4/29/09, Ian Jackson <[email protected]> wrote:
From: Ian Jackson <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: [email protected]
Cc: "xen-devel" <[email protected]>
Date: Wednesday, April 29, 2009, 11:59 AM

Boris Derzhavets writes ("Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 1 line

Is there any way to point xen

3.4-rc3-pre source (xen-unstbale.hg) to

Quoted 1 line

pick up for build phase python 2.5 , installed on  the system, instead of

2.6

Quoted 1 line

used by gnome environment as default.

You need to change the default python version used by setup.py' which calls the standard Python setup machinery. I'm no python expert but I think you might be able to get the Xen build system to do the right thing by making python' on $PATH point to the python version you want it to use.

It seems like a bug to me that it uses python' rather than having a variable PYTHON' but we probably won't fix that in 3.4 now.

Quoted 4 lines

I found statement :-

The reason for this is that Xen 3.0 is bound to Python 2.4 whereas Xen 3.2 is bound to Python 2.5.

That comment is about the Debian Xen packages and isn't relevant if you're building it yourself. When you build Xen your build version gets the python version
you used baked into it.

Ian.

      _______________________________________________

Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#12 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

In reply to Ian Jackson

--- On Wed, 4/29/09, Ian Jackson <[email protected]> wrote:
From: Ian Jackson <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "[email protected]" <[email protected]>
Cc: "xen-devel" <[email protected]>
Date: Wednesday, April 29, 2009, 12:19 PM

Boris Derzhavets writes ("Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 2 lines

Could you please point to an exact location of setup.py to be edited. I found several files under xen-unstable.hg.

No, I mean, either change /usr[/local]/bin/python (which is probably a symlink)
*******************
I did it. No luck
*******************
to point to the desired version of python, or set your main PATH variable to have a directory at the front containing `python' pointing to the version you want.
********************
Will try.
Thanks.
********************

Quoted 2 lines

export PYTHONPATH=/usr/lib/python2.5/site-packages doesn't help.

No, I don't think it would.

Ian.

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#13 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Hi,

It looks like the problem is that if you have python2.6 installed then xen-unstable.hg installs into /usr/lib/python2.6/site-packages/xen, but site-packages is no longer in the default path in python2.6.

A hack which fixes this is to add the following two lines to /etc/python2.6/sitecustomize.py:

import sys sys.path.append('/usr/lib/python2.6/site-packages')

but I'm not sure what the proper fix would be...

Regards,

Alex

PS This is also relevant https://bugs.launchpad.net/ubuntu/+source/python2.6/+bug/362570

#14 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Alex Zeffertt writes ("Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 3 lines

It looks like the problem is that if you have python2.6 installed then xen-unstable.hg installs into /usr/lib/python2.6/site-packages/xen, but site-packages is no longer in the default path in python2.6.

This is a bit odd.

We don't do anything special with our uses of the Python distutils module. The string "site-packages" does not even appear in xen-unstable the source[1] !

Quoted 2 lines

PS This is also relevant
https://bugs.launchpad.net/ubuntu/+source/python2.6/+bug/362570

This looks like exactly our bug. So I think this is a bug in the Ubuntu python packaging. But the Ubuntu Python maintainer seems not to have understood the problem and has called the bug a wishlist bug.

Ian.

[1] Except for once in tools/xenstat/libxenstat/Makefile which I assume is irrelevant.

#15 lore raw

Re: [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Quoted 9 lines · Ian Jackson

Alex Zeffertt writes ("Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 1 line

It looks like the problem is that if you have python2.6 installed then xen-unstable.hg installs into /usr/lib/python2.6/site-packages/xen, but site-packages is no longer in the default path in python2.6.

This is a bit odd.

We don't do anything special with our uses of the Python distutils module. The string "site-packages" does not even appear in xen-unstable the source[1] !

Quoted 7 lines
Quoted 1 line

PS This is also relevant https://bugs.launchpad.net/ubuntu/+source/python2.6/+bug/362570

This looks like exactly our bug. So I think this is a bug in the Ubuntu python packaging. But the Ubuntu Python maintainer seems not to have understood the problem and has called the bug a wishlist bug.

Ian.

This patch appears to fix the problem. Just don't use --prefix. It doesn't seem to make any difference for python2.5, but it installs to the correct place under python 2.6.

-- Alex

Install xen python modules to correct path under python2.6.

Do not use "--prefix" with setup.py scripts. In python2.6 this causes modules to be installed into /usr/lib/python2.6/site-packages, which is not on the default python2.6 sys.path.

Without the --prefix option modules get installed to /usr/local/lib/python2.6/dist-packages, which is on the default python2.6 sys.path. If the installed version of python is python2.5, modules get installed to /usr/lib/python2.5/site-packages, which is on the default
python2.5 sys.path.

Signed-off-by <[email protected]>

tools/pygrub/Makefile +1 -1
diff -r 474e93610de3 tools/pygrub/Makefile
@@ -11,7 +12,7 @@
 .PHONY: install
 install: all
 	CC="$(CC)" CFLAGS="$(CFLAGS)" python setup.py install \
-		--prefix="$(PREFIX)" --root="$(DESTDIR)" --force
+		--root="$(DESTDIR)" --force
 	$(INSTALL_DIR) $(DESTDIR)/var/run/xend/boot
 
 .PHONY: clean
tools/python/Makefile +1 -1
diff -r 474e93610de3 tools/python/Makefile
@@ -56,7 +56,7 @@
 .PHONY: install
 install: install-messages install-dtd
 	CC="$(CC)" CFLAGS="$(CFLAGS)" python setup.py install \
-		--prefix="$(PREFIX)" --root="$(DESTDIR)" --force
+		--root="$(DESTDIR)" --force
 
 install-dtd: all
 	$(INSTALL_DIR) $(DESTDIR)$(DOCDIR)
tools/security/Makefile +1 -1
diff -r 474e93610de3 tools/security/Makefile
@@ -61,7 +61,7 @@
 	$(INSTALL_DIR) $(DESTDIR)$(ACM_SECGEN_CGIDIR)
 	$(INSTALL_PROG) $(ACM_INST_CGI) $(DESTDIR)$(ACM_SECGEN_CGIDIR)
 	python python/setup.py install \
-		--prefix="$(PREFIX)" --root="$(DESTDIR)" --force
+		--root="$(DESTDIR)" --force
 else
 .PHONY: all
 all:
#16 lore raw

Re: [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Quoted 3 lines · Alex Zeffertt

This patch appears to fix the problem. Just don't use --prefix. It doesn't seem to make any difference for python2.5, but it installs to the correct place under python 2.6.

Wouldn't it be correct to specify --prefix (even if that means installing to blah/site-packages) if the user has overridden prefix to make a personal installation avoiding conflicting with distro packages? Otherwise PREFIX allows a user to install Xen bits somewhere out of the way, except the Python bits which will still blat the default install path.

Perhaps config/*.mk should define DEFAULT_PREFIX and PREFIX, and we omit --prefix= to Python installer if the two are equal?

-- Keir

#17 lore raw

Re: [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Keir Fraser writes ("Re: [Xen-devel] [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 10 lines · Alex Zeffertt
Quoted 3 lines

This patch appears to fix the problem. Just don't use --prefix. It doesn't seem to make any difference for python2.5, but it installs to the correct place under python 2.6.

Wouldn't it be correct to specify --prefix (even if that means installing to blah/site-packages) if the user has overridden prefix to make a personal installation avoiding conflicting with distro packages? Otherwise PREFIX allows a user to install Xen bits somewhere out of the way, except the Python bits which will still blat the default install path.

Quite so.

Quoted 2 lines

Perhaps config/*.mk should define DEFAULT_PREFIX and PREFIX, and we omit --prefix= to Python installer if the two are equal?

That would be one way to do it but sadly it's not sufficient. On pre-jaunty Ubuntu systems (and many others), omitting --prefix puts everything in /usr which is also wrong.

I think the right answer is to provide a way to suppress the --prefix explicitly. That will provide a workaround to people with the broken python setup, without making things go wrong for anyone else.

Ian.

#18 lore raw

Re: [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Quoted 10 lines · Ian Jackson
Quoted 2 lines

Perhaps config/*.mk should define DEFAULT_PREFIX and PREFIX, and we omit --prefix= to Python installer if the two are equal?

That would be one way to do it but sadly it's not sufficient. On pre-jaunty Ubuntu systems (and many others), omitting --prefix puts everything in /usr which is also wrong.

I think the right answer is to provide a way to suppress the --prefix explicitly. That will provide a workaround to people with the broken python setup, without making things go wrong for anyone else.

Is this not a generic Python26 problem then, but an Ubuntu (not even a Debian upstream) customisation?

-- Keir

#19 lore raw

Re: [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Keir Fraser writes ("Re: [Xen-devel] [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 2 lines

Is this not a generic Python26 problem then, but an Ubuntu (not even a Debian upstream) customisation?

I think it's Ubuntu-specific. I think we'd have heard if recent Debian unstable has it, but I could be wrong. Perhaps the Debian users are using the official Debian packages.

Keir Fraser writes ("Re: [Xen-devel] [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 1 line

If this is specific to Ubuntu, will it be worth patching for 3.4?

It'll be a simple fix which I'll test to make sure it doesn't have any effect on the build when you don't turn it on :-).

Ian.

#20 lore raw

PATCH 1/1 Permit user to suppress passing --prefix to setup.py

iwj writes ("Re: [Xen-devel] [PATCH] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)"):

Quoted 2 lines

It'll be a simple fix which I'll test to make sure it doesn't have any effect on the build when you don't turn it on :-).

We change all invocations of setup.py as follows:

* use $(PYTHON) instead of python' so that the user can specify an alternative python version if they need to. If not set it defaults to python' in Config.mk.

* pass --prefix=$(PREFIX) via a new make variable $(PYTHON_PREFIX_ARG). This allows a user to suppress the --prefix=... argument entirely by setting PYTHON_PREFIX_ARG=''. This will work around the bug described here

     https://bugs.launchpad.net/ubuntu/+bug/362570     

where passing --prefix=/usr/local (which ought to have no effect as /usr/local is the default prefix) changes which subdirectory distutils chooses, and results in the files being installed in site-packages which is not on the default search path.

Users not affected by this python packaging bug should not set PYTHON_PREFIX_ARG and their builds will not be affected. (Provided PREFIX did not contain spaces. People who put spaces in PREFIX are being quite optimistic.)

Config.mk +6 -0
diff -r a1a0bd345acc Config.mk
@@ -34,6 +34,12 @@ EXTRA_INCLUDES += $(EXTRA_PREFIX)/includ
 EXTRA_INCLUDES += $(EXTRA_PREFIX)/include
 EXTRA_LIB += $(EXTRA_PREFIX)/$(LIBLEAFDIR)
 endif
+
+PYTHON    ?= python
+PYTHON_PREFIX_ARG ?= --prefix="$(PREFIX)"
+# Sorry, if PREFIX contains spaces, you lose.  This variable is here
+# to permit the user to set PYTHON_PREFIX_ARG to '' to workaround this
+# bug:  https://bugs.launchpad.net/ubuntu/+bug/362570
 
 # cc-option: Check if compiler supports first option, else fall back to second.
 # Usage: cflags-y += $(call cc-option,$(CC),-march=winchip-c6,-march=i586)
tools/pygrub/Makefile +3 -3
diff -r a1a0bd345acc tools/pygrub/Makefile
@@ -6,12 +6,12 @@ all: build
 all: build
 .PHONY: build
 build:
-	CC="$(CC)" CFLAGS="$(CFLAGS)" python setup.py build
+	CC="$(CC)" CFLAGS="$(CFLAGS)" $(PYTHON) setup.py build
 
 .PHONY: install
 install: all
-	CC="$(CC)" CFLAGS="$(CFLAGS)" python setup.py install \
-		--prefix="$(PREFIX)" --root="$(DESTDIR)" --force
+	CC="$(CC)" CFLAGS="$(CFLAGS)" $(PYTHON) setup.py install \
+		$(PYTHON_PREFIX_ARG) --root="$(DESTDIR)" --force
 	$(INSTALL_DIR) $(DESTDIR)/var/run/xend/boot
 
 .PHONY: clean
tools/python/Makefile +3 -3
diff -r a1a0bd345acc tools/python/Makefile
@@ -16,7 +16,7 @@ NLSDIR = $(SHAREDIR)/locale
 
 .PHONY: build buildpy
 buildpy: 
-	CC="$(CC)" CFLAGS="$(CFLAGS)" python setup.py build
+	CC="$(CC)" CFLAGS="$(CFLAGS)" $(PYTHON) setup.py build
 
 build: buildpy refresh-pot refresh-po $(CATALOGS)
 
@@ -55,8 +55,8 @@ refresh-po: $(POTFILE)
 
 .PHONY: install
 install: install-messages install-dtd
-	CC="$(CC)" CFLAGS="$(CFLAGS)" python setup.py install \
-		--prefix="$(PREFIX)" --root="$(DESTDIR)" --force
+	CC="$(CC)" CFLAGS="$(CFLAGS)" $(PYTHON) setup.py install \
+		$(PYTHON_PREFIX_ARG) --root="$(DESTDIR)" --force
 
 install-dtd: all
 	$(INSTALL_DIR) $(DESTDIR)$(DOCDIR)
tools/security/Makefile +3 -3
diff -r a1a0bd345acc tools/security/Makefile
@@ -60,8 +60,8 @@ install: all $(ACM_CONFIG_FILE)
 	$(INSTALL_DATA) $(ACM_INST_HTML) $(DESTDIR)$(ACM_SECGEN_HTMLDIR)
 	$(INSTALL_DIR) $(DESTDIR)$(ACM_SECGEN_CGIDIR)
 	$(INSTALL_PROG) $(ACM_INST_CGI) $(DESTDIR)$(ACM_SECGEN_CGIDIR)
-	python python/setup.py install \
-		--prefix="$(PREFIX)" --root="$(DESTDIR)" --force
+	$(PYTHON) python/setup.py install $(PYTHON_PREFIX_ARG) \
+		--root="$(DESTDIR)" --force
 else
 .PHONY: all
 all:
@@ -72,7 +72,7 @@ endif
 
 .PHONY: build
 build: $(ACM_INST_TOOLS) $(ACM_NOINST_TOOLS)
-	python python/setup.py build
+	$(PYTHON) python/setup.py build
 	chmod 700 $(ACM_SCRIPTS)
 
 xensec_tool: $(OBJS_TOOL)
#21 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

In reply to Alex Zeffertt

Follow your suggestion i was able to build Xen Host on Ubuntu 9.04 Server
with the most recent Xen-3.4-rc4-pre and 2.6.30-rc3-tip pvops kernel via
Jeremy's git repo.
Due to newly built Xen Host dual boot with similar one on top of Ubuntu 8.10 Server i just copied F10 PV profile to new Xen Host and attempted to run F10 PV :-

# xm create -c f10.pyrun

Name                                        ID   Mem VCPUs      State   Time(s) Domain-0                                     0  6034     4     r-----    138.5 F10PV                                        5  2048     1     --p---      0.0

root@ServerJaunty:/home/boris/vm# xm unpause F10PV
/usr/lib/python2.6/site-packages/xen/util/acmpolicy.py:20: DeprecationWarning: the sha module is deprecated; use the hashlib module instead
  import sha
# xm unpause F10PV

PCI: Fatal: No config space access function found
IO APIC resources could be not be allocated.
Loading /lib/kbd/keymaps/i386/qwerty/us.map

root@ServerJaunty:/home/boris/vm# Error: Device 0 (vif) could not be connected. Hotplug scripts not working.

Then i've tried to comment out vif in profile:

PCI: Fatal: No config space access function found
IO APIC resources could be not be allocated.
Loading /lib/kbd/keymaps/i386/qwerty/us.map

BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
IP: [<ffffffffa00010e5>] backend_changed+0x2d5/0x3af [xen_blkfront]
PGD 7d1a0067 PUD 7d1a1067 PMD 7d1b0067 PTE 0
Oops: 0000 [1] SMP 
CPU 1 
Modules linked in: xen_blkfront
Pid: 19, comm: xenwatch Not tainted 2.6.27.19-170.2.35.fc10.x86_64 #1 
RIP: e030:[<ffffffffa00010e5>]  [<ffffffffa00010e5>] backend_changed+0x2d5/0x3af [xen_blkfront]
RSP: e02b:ffff88007f40fdf0  EFLAGS: 00010246
RAX: 0000000000000000 RBX: ffff88007d09b2c0 RCX: 0000000000000000
RDX: ffff88007f40fe28 RSI: 0000000000000005 RDI: ffff88007ff5f400
RBP: ffff88007f40fe80 R08: ffff88007f40fdd0 R09: ffff88007f40fcc0
R10: 0000000000000001 R11: ffff880000000018 R12: ffff88007ff5f400
R13: ffff88007d0be000 R14: 000000000000000a R15: ffffffff817e6eb4
FS:  00007f80e570a6f0(0000) GS:ffff88007fc04880(0000) knlGS:0000000000000000
CS:  e033 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 000000007d1a4000 CR4: 0000000000002660
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process xenwatch (pid: 19, threadinfo ffff88007f40e000, task ffff88007ffd5c40)
Stack:  ffff88007f40fe00 ffff88007f40fe40 ffffffff816e2500 ffffffff816e2500
 ffffffff81458716 ffffffff8146545f ffff88007f40fe7c 0000000000000000
 000000000000000a ffff88007d09b2c0 ffff88007ff5f400 ffffffffa0002ac0
Call Trace:
 [<ffffffff811ca80a>] otherend_changed+0x81/0x8a
 [<ffffffff811c9438>] xenwatch_thread+0x101/0x134
 [<ffffffff8105553d>] ? autoremove_wake_function+0x0/0x38
 [<ffffffff811c9337>] ? xenwatch_thread+0x0/0x134
 [<ffffffff810551f7>] kthread+0x49/0x76
 [<ffffffff81011719>] child_rip+0xa/0x11
 [<ffffffff81010a37>] ? restore_args+0x0/0x30
 [<ffffffff8101170f>] ? child_rip+0x0/0x11


Code: 58 08 e9 97 00 00 00 49 8b 7d 00 48 c7 c2 b9 15 00 a0 be ed ff ff ff 48 8b 4f 10 31 c0 e8 fe 75 1c e1 e9 cf 00 00 00 49 8b 45 08 <8b> 38 c1 e7 14 0b 78 04 e8 5a 5c 0e e1 48 85 c0 48 89 c3 75 16 
RIP  [<ffffffffa00010e5>] backend_changed+0x2d5/0x3af [xen_blkfront]
 RSP <ffff88007f40fdf0>
CR2: 0000000000000000
---[ end trace 06219b74fe5c0c01 ]---
root@ServerJaunty:/home/boris/vm# Error: Device 51712 (vbd) could not be connected. Hotplug scripts not work

The image been used (on /dev/sdc) works fine with the same F10 PV @ Xen
Host ( absolutly identical new one) , but been built on top of Ubuntu 8.10 Server ( all servers 64-bit).

Boris.

--- On Wed, 5/6/09, Alex Zeffertt <[email protected]> wrote:
From: Alex Zeffertt <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "[email protected]" <[email protected]>
Cc: "Ian Jackson" <[email protected]>, "xen-devel" <[email protected]>
Date: Wednesday, May 6, 2009, 12:37 PM

Hi,

It looks like the problem is that if you have python2.6 installed then xen-unstable.hg installs into /usr/lib/python2.6/site-packages/xen, but site-packages is no longer in the default path in python2.6.

A hack which fixes this is to add the following two lines to /etc/python2.6/sitecustomize.py:

import sys sys.path.append('/usr/lib/python2.6/site-packages')

but I'm not sure what the proper fix would be...

Regards,

Alex

PS This is also relevant
https://bugs.launchpad.net/ubuntu/+source/python2.6/+bug/362570

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#22 lore raw

Re: Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)

     Notice, that direct port Xen 3.4 ( with 2.6.30-rc3-tip)  to F11 (preview) has been tested with no problems.  Just no xen-lib and xen-lib dependent packages: libvirt,libvirt-python,..  is a must . Problem been described seems to me as python 2.6 implementation issue on Ubuntu 9.04 Server.

Boris.

--- On Fri, 5/8/09, Boris Derzhavets <[email protected]> wrote:
From: Boris Derzhavets <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "Alex Zeffertt" <[email protected]>
Cc: "xen-devel" <[email protected]>, "Ian Jackson" <[email protected]>
Date: Friday, May 8, 2009, 11:45 AM

Follow your suggestion i was able to build Xen Host on Ubuntu 9.04 Server
with the most recent Xen-3.4-rc4-pre and 2.6.30-rc3-tip pvops kernel via
Jeremy's git repo.
Due to newly built Xen Host dual boot with similar one on top of Ubuntu 8.10 Server i just copied F10 PV profile to new Xen Host and attempted to run F10 PV :-

# xm create -c f10.pyrun

Name                                        ID   Mem VCPUs      State
Time(s)
Domain-0                                     0  6034     4     r-----    138.5 F10PV                                        5  2048     1     --p---      0.0

root@ServerJaunty:/home/boris/vm# xm unpause F10PV
/usr/lib/python2.6/site-packages/xen/util/acmpolicy.py:20: DeprecationWarning: the sha module is deprecated; use the hashlib module instead
  import sha
# xm unpause F10PV

PCI: Fatal: No config space access
function found
IO APIC resources could be not be allocated.
Loading /lib/kbd/keymaps/i386/qwerty/us.map

root@ServerJaunty:/home/boris/vm# Error: Device 0 (vif) could not be connected. Hotplug scripts not working.

Then i've tried to comment out vif in profile:

PCI: Fatal: No config space access function found
IO APIC resources could be not be allocated.
Loading /lib/kbd/keymaps/i386/qwerty/us.map

BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
IP: [<ffffffffa00010e5>] backend_changed+0x2d5/0x3af [xen_blkfront]
PGD 7d1a0067 PUD 7d1a1067 PMD 7d1b0067 PTE 0
Oops: 0000 [1] SMP 
CPU 1 
Modules linked in: xen_blkfront
Pid: 19, comm: xenwatch Not tainted 2.6.27.19-170.2.35.fc10.x86_64 #1 
RIP: e030:[<ffffffffa00010e5>]  [<ffffffffa00010e5>] backend_changed+0x2d5/0x3af [xen_blkfront]
RSP: e02b:ffff88007f40fdf0  EFLAGS: 00010246

RAX:
0000000000000000 RBX: ffff88007d09b2c0 RCX: 0000000000000000
RDX: ffff88007f40fe28 RSI: 0000000000000005 RDI: ffff88007ff5f400
RBP: ffff88007f40fe80 R08: ffff88007f40fdd0 R09: ffff88007f40fcc0
R10: 0000000000000001 R11: ffff880000000018 R12: ffff88007ff5f400
R13: ffff88007d0be000 R14: 000000000000000a R15: ffffffff817e6eb4 FS:  00007f80e570a6f0(0000) GS:ffff88007fc04880(0000) knlGS:0000000000000000 CS:  e033 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 000000007d1a4000 CR4: 0000000000002660
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process xenwatch (pid: 19, threadinfo ffff88007f40e000, task ffff88007ffd5c40)

Stack:  ffff88007f40fe00 ffff88007f40fe40 ffffffff816e2500 ffffffff816e2500
 ffffffff81458716 ffffffff8146545f ffff88007f40fe7c 0000000000000000
 000000000000000a
 ffff88007d09b2c0 ffff88007ff5f400 ffffffffa0002ac0
Call Trace:
 [<ffffffff811ca80a>] otherend_changed+0x81/0x8a
 [<ffffffff811c9438>] xenwatch_thread+0x101/0x134
 [<ffffffff8105553d>] ? autoremove_wake_function+0x0/0x38
 [<ffffffff811c9337>] ? xenwatch_thread+0x0/0x134
 [<ffffffff810551f7>] kthread+0x49/0x76
 [<ffffffff81011719>] child_rip+0xa/0x11
 [<ffffffff81010a37>] ? restore_args+0x0/0x30
 [<ffffffff8101170f>] ? child_rip+0x0/0x11


Code: 58 08 e9 97 00 00 00 49 8b 7d 00 48 c7 c2 b9 15 00 a0 be ed ff ff ff 48 8b 4f 10 31 c0 e8 fe 75 1c e1 e9 cf 00 00 00 49 8b 45 08 <8b> 38 c1 e7 14 0b 78 04 e8 5a 5c 0e e1 48 85 c0 48 89 c3 75 16 
RIP  [<ffffffffa00010e5>] backend_changed+0x2d5/0x3af [xen_blkfront]
 RSP <ffff88007f40fdf0>
CR2: 0000000000000000
---[ end trace 06219b74fe5c0c01
 ]---
root@ServerJaunty:/home/boris/vm# Error: Device 51712 (vbd) could not be connected. Hotplug scripts not work

The image been used (on /dev/sdc) works fine with the same F10 PV @ Xen
Host ( absolutly identical new one) , but been built on top of Ubuntu 8.10 Server ( all servers 64-bit).

Boris.

--- On Wed, 5/6/09, Alex Zeffertt <[email protected]> wrote:
From: Alex Zeffertt <[email protected]>
Subject: Re: [Xen-devel] Attempt to build xen-unstable on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: "[email protected]" <[email protected]>
Cc: "Ian Jackson" <[email protected]>, "xen-devel" <[email protected]>
Date: Wednesday, May 6, 2009, 12:37 PM

Hi,

It looks like the problem is that if you have python2.6
installed then
xen-unstable.hg installs into /usr/lib/python2.6/site-packages/xen, but site-packages is no longer in the default path in python2.6.

A hack which fixes this is to add the following two lines to /etc/python2.6/sitecustomize.py:

import sys sys.path.append('/usr/lib/python2.6/site-packages')

but I'm not sure what the proper fix would be...

Regards,

Alex

PS This is also relevant
https://bugs.launchpad.net/ubuntu/+source/python2.6/+bug/362570

      _______________________________________________

Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#23 lore raw

2.6.30-rc3-tip kernel's backend network driver response when loading Xen 3.4 Dom0 on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Last lines of dmesg report Xen 3.4 (2.6.30-rc3-tip) instance been built on Ubuntu 9.04 Server ( IP6v by default)

[   29.214154] skge eth0: enabling interface
[   29.218498] ADDRCONF(NETDEV_UP): eth0: link is not ready
[   29.630012] eth1: no IPv6 routers present
[   38.012544] sky2 eth1: disabling interface
[   38.092791] sky2 peth1: enabling interface
[   38.093450] ADDRCONF(NETDEV_UP): peth1: link is not ready
[   39.776997] sky2 peth1: Link is up at 100 Mbps, full duplex, flow control both [   39.777643] ADDRCONF(NETDEV_CHANGE): peth1: link becomes ready
[   40.145017] device peth1 entered promiscuous mode
[   40.182778] eth1: topology change detected, propagating
[   40.182783] eth1: port 1(peth1) entering forwarding state
[   49.820430] xenbus_probe wake_waiting
[   49.820491] xenbus_probe wake_waiting
[   49.829859] XENBUS: Unable to read cpu state
[   49.831443] XENBUS: Unable to read cpu state
[   49.831633] XENBUS: Unable to read cpu state
[   49.831851] XENBUS: Unable to read cpu state
[   49.832132] xenbus_probe_devices backend
[   49.832283] xenbus_probe_devices failed xenbus_directory
[   49.832331] backend_probe_and_watch devices probed ok
[   49.832494] backend_probe_and_watch watch add ok ok
[   49.832541] backend_probe_and_watch all done
[   49.832581] xenbus_probe_devices device
[   49.832731] xenbus_probe_devices failed xenbus_directory
[   49.832777] frontend_probe_and_watch devices probed ok
[   49.832948] frontend_probe_and_watch watch add ok ok
[   49.832995] frontend_probe_and_watch all done
[   49.910018] peth1: no IPv6 routers present
[   51.062517] eth1: no IPv6 routers present
[  173.299352] vif1.0 (vif): not using net_device_ops yet

Last lines of dmesg report Xen 3.4 (2.6.30-rc3-tip) instance been built on Ubuntu 8.10

[   54.856147] sky2 peth0: enabling interface
[   54.856810] ADDRCONF(NETDEV_UP): peth0: link is not ready
[   55.777286] skge eth1: enabling interface
[   55.781784] ADDRCONF(NETDEV_UP): eth1: link is not ready
[   56.521131] sky2 peth0: Link is up at 100 Mbps, full duplex, flow control both [   56.521765] ADDRCONF(NETDEV_CHANGE): peth0: link becomes ready
[   56.905093] device peth0 entered promiscuous mode
[   56.917357] eth0: topology change detected, propagating
[   56.917362] eth0: port 1(peth0) entering forwarding state
[   64.531928] xenbus_probe wake_waiting
[   64.531984] xenbus_probe wake_waiting
[   64.542317] XENBUS: Unable to read cpu state
[   64.542486] XENBUS: Unable to read cpu state
[   64.542661] XENBUS: Unable to read cpu state
[   64.542828] XENBUS: Unable to read cpu state
[   64.543137] xenbus_probe_devices backend
[   64.543313] xenbus_probe_devices failed xenbus_directory
[   64.543356] backend_probe_and_watch devices probed ok
[   64.543514] backend_probe_and_watch watch add ok ok
[   64.543557] backend_probe_and_watch all done
[   64.543599] xenbus_probe_devices device
[   64.543776] xenbus_probe_devices failed xenbus_directory
[   64.543818] frontend_probe_and_watch devices probed ok
[   64.543972] frontend_probe_and_watch watch add ok ok
[   64.544021] frontend_probe_and_watch all done
[   67.040013] peth0: no IPv6 routers present
[   67.890015] eth0: no IPv6 routers present
[  118.012469] vif1.0 (vif): not using net_device_ops yet
[  118.127252] device vif1.0 entered promiscuous mode
[  118.130694] eth0: topology change detected, propagating
[  118.130701] eth0: port 2(vif1.0) entering forwarding state
[  122.657298] blkback: ring-ref 8, event-channel 9, protocol 1 (x86_64-abi) [  128.700025] vif1.0: no IPv6 routers present
[  183.761252] eth0: port 2(vif1.0) entering disabled state
[  183.800702] eth0: port 2(vif1.0) entering disabled state

Line:-

[  122.657298] blkback: ring-ref 8, event-channel 9, protocol 1 (x86_64-abi)

is absent in first report. I believe it's a Dom0's kernel message from network backend driver. Instances dual boot on the same box.

Boris.

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

#24 lore raw

Re: 2.6.30-rc3-tip kernel's backend network driver response when loading Xen 3.4 Dom0 on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Quoted 72 lines · Boris Derzhavets

Last lines of dmesg report Xen 3.4 (2.6.30-rc3-tip) instance been built on Ubuntu 9.04 Server ( IP6v by default)

[   29.214154] skge eth0: enabling interface
[   29.218498] ADDRCONF(NETDEV_UP): eth0: link is not ready
[   29.630012] eth1: no IPv6 routers present
[   38.012544] sky2 eth1: disabling interface
[   38.092791] sky2 peth1: enabling interface
[   38.093450] ADDRCONF(NETDEV_UP): peth1: link is not ready
[   39.776997] sky2 peth1: Link is up at 100 Mbps, full duplex, flow control both
[   39.777643] ADDRCONF(NETDEV_CHANGE): peth1: link becomes ready
[   40.145017] device peth1 entered promiscuous mode
[   40.182778] eth1: topology change detected, propagating
[   40.182783] eth1: port 1(peth1) entering forwarding state
[   49.820430] xenbus_probe wake_waiting
[   49.820491] xenbus_probe wake_waiting
[   49.829859] XENBUS: Unable to read cpu state
[   49.831443] XENBUS: Unable to read cpu state
[   49.831633] XENBUS: Unable to read cpu state
[   49.831851] XENBUS: Unable to read cpu state
[   49.832132] xenbus_probe_devices backend
[   49.832283] xenbus_probe_devices failed xenbus_directory
[   49.832331] backend_probe_and_watch devices probed ok
[   49.832494] backend_probe_and_watch watch add ok ok
[   49.832541] backend_probe_and_watch all done
[   49.832581] xenbus_probe_devices device
[   49.832731] xenbus_probe_devices failed xenbus_directory
[   49.832777] frontend_probe_and_watch devices probed ok
[   49.832948] frontend_probe_and_watch watch add ok ok
[   49.832995] frontend_probe_and_watch all done
[   49.910018] peth1: no IPv6 routers present
[   51.062517] eth1: no IPv6 routers present
[  173.299352] vif1.0 (vif): not using net_device_ops yet

Last lines of dmesg report Xen 3.4 (2.6.30-rc3-tip) instance been built on Ubuntu 8.10

[   54.856147] sky2 peth0: enabling interface
[   54.856810] ADDRCONF(NETDEV_UP): peth0: link is not ready
[   55.777286] skge eth1: enabling interface
[   55.781784] ADDRCONF(NETDEV_UP): eth1: link is not ready
[   56.521131] sky2 peth0: Link is up at 100 Mbps, full duplex, flow control both
[   56.521765] ADDRCONF(NETDEV_CHANGE): peth0: link becomes ready
[   56.905093] device peth0 entered promiscuous mode
[   56.917357] eth0: topology change detected, propagating
[   56.917362] eth0: port 1(peth0) entering forwarding state
[   64.531928] xenbus_probe wake_waiting
[   64.531984] xenbus_probe wake_waiting
[   64.542317] XENBUS: Unable to read cpu state
[   64.542486] XENBUS: Unable to read cpu state
[   64.542661] XENBUS: Unable to read cpu state
[   64.542828] XENBUS: Unable to read cpu state
[   64.543137] xenbus_probe_devices backend
[   64.543313] xenbus_probe_devices failed xenbus_directory
[   64.543356] backend_probe_and_watch devices probed ok
[   64.543514] backend_probe_and_watch watch add ok ok
[   64.543557] backend_probe_and_watch all done
[   64.543599] xenbus_probe_devices device
[   64.543776] xenbus_probe_devices failed xenbus_directory
[   64.543818] frontend_probe_and_watch devices probed ok
[   64.543972] frontend_probe_and_watch watch add ok ok
[   64.544021] frontend_probe_and_watch all done
[   67.040013] peth0: no IPv6 routers present
[   67.890015] eth0: no IPv6 routers present
[  118.012469] vif1.0 (vif): not using net_device_ops yet
[  118.127252] device vif1.0 entered promiscuous mode
[  118.130694] eth0: topology change detected, propagating
[  118.130701] eth0: port 2(vif1.0) entering forwarding state
[  122.657298] blkback: ring-ref 8, event-channel 9, protocol 1 (x86_64-abi)
[  128.700025] vif1.0: no IPv6 routers present
[  183.761252] eth0: port 2(vif1.0) entering disabled state
[  183.800702] eth0: port 2(vif1.0) entering disabled state
Quoted 7 lines

Line:-

[  122.657298] blkback: ring-ref 8, event-channel 9, protocol 1 (x86_64-abi)

is absent in first report. I believe it's a Dom0's kernel message from network backend driver. Instances dual boot on the same box.

I've seen similar things when building xenbus as modules; the backend modules don't get auto-loaded. Either compiling them non-modular or manually loading the modules works.

If that doesn't work, I'm not sure what's going on.

J

#25 lore raw

Re: Re: 2.6.30-rc3-tip kernel's backend network driver response when loading Xen 3.4 Dom0 on Ubuntu 9.04 Server (Ubuntu Desktop installed)

Jeremy,

Backend support always gets compiled built-in either on Ubuntu 9.04 or Ubuntu 8.10. But on 8.10 it works (gcc 4.3) on 9.04 it doesn't (gcc 4.4). In the last case hotplug scripts refuse to work. DomU paused at boot up and failed with messages:-
****************************************************************
XENBUS: Timeout connecting to device: /device/vbd/51712
vif(0)  could not be connected . Hotplug scripts not working.
*****************************************************************
Please, advise how to try to load backend support modules manually.

Boris

--- On Mon, 5/11/09, Jeremy Fitzhardinge <[email protected]> wrote:
From: Jeremy Fitzhardinge <[email protected]>
Subject: [Xen-devel] Re: 2.6.30-rc3-tip kernel's backend network driver response when loading Xen 3.4 Dom0 on Ubuntu 9.04 Server (Ubuntu Desktop installed)
To: [email protected]
Cc: "xen-devel" <[email protected]>
Date: Monday, May 11, 2009, 6:51 PM

Quoted 1 line · Boris Derzhavets

Last lines of dmesg report Xen 3.4 (2.6.30-rc3-tip) instance been built on

Ubuntu 9.04 Server ( IP6v by default)

Quoted 8 lines
[   29.214154] skge eth0: enabling interface
[   29.218498] ADDRCONF(NETDEV_UP): eth0: link is not ready
[   29.630012] eth1: no IPv6 routers present
[   38.012544] sky2 eth1: disabling interface
[   38.092791] sky2 peth1: enabling interface
[   38.093450] ADDRCONF(NETDEV_UP): peth1: link is not ready
[   39.776997] sky2 peth1: Link is up at 100 Mbps, full duplex, flow

control both

Quoted 24 lines
[   39.777643] ADDRCONF(NETDEV_CHANGE): peth1: link becomes ready
[   40.145017] device peth1 entered promiscuous mode
[   40.182778] eth1: topology change detected, propagating
[   40.182783] eth1: port 1(peth1) entering forwarding state
[   49.820430] xenbus_probe wake_waiting
[   49.820491] xenbus_probe wake_waiting
[   49.829859] XENBUS: Unable to read cpu state
[   49.831443] XENBUS: Unable to read cpu state
[   49.831633] XENBUS: Unable to read cpu state
[   49.831851] XENBUS: Unable to read cpu state
[   49.832132] xenbus_probe_devices backend
[   49.832283] xenbus_probe_devices failed xenbus_directory
[   49.832331] backend_probe_and_watch devices probed ok
[   49.832494] backend_probe_and_watch watch add ok ok
[   49.832541] backend_probe_and_watch all done
[   49.832581] xenbus_probe_devices device
[   49.832731] xenbus_probe_devices failed xenbus_directory
[   49.832777] frontend_probe_and_watch devices probed ok
[   49.832948] frontend_probe_and_watch watch add ok ok
[   49.832995] frontend_probe_and_watch all done
[   49.910018] peth1: no IPv6 routers present
[   51.062517] eth1: no IPv6 routers present
[  173.299352] vif1.0 (vif): not using net_device_ops yet
Quoted 1 line

Last lines of dmesg report Xen 3.4 (2.6.30-rc3-tip) instance been built on

Ubuntu 8.10

Quoted 7 lines

[   54.856147] sky2 peth0: enabling interface
[   54.856810] ADDRCONF(NETDEV_UP): peth0: link is not ready
[   55.777286] skge eth1: enabling interface
[   55.781784] ADDRCONF(NETDEV_UP): eth1: link is not ready
[   56.521131] sky2 peth0: Link is up at 100 Mbps, full duplex, flow

control both

Quoted 19 lines
[   56.521765] ADDRCONF(NETDEV_CHANGE): peth0: link becomes ready
[   56.905093] device peth0 entered promiscuous mode
[   56.917357] eth0: topology change detected, propagating
[   56.917362] eth0: port 1(peth0) entering forwarding state
[   64.531928] xenbus_probe wake_waiting
[   64.531984] xenbus_probe wake_waiting
[   64.542317] XENBUS: Unable to read cpu state
[   64.542486] XENBUS: Unable to read cpu state
[   64.542661] XENBUS: Unable to read cpu state
[   64.542828] XENBUS: Unable to read cpu state
[   64.543137] xenbus_probe_devices backend
[   64.543313] xenbus_probe_devices failed xenbus_directory
[   64.543356] backend_probe_and_watch devices probed ok
[   64.543514] backend_probe_and_watch watch add ok ok
[   64.543557] backend_probe_and_watch all done
[   64.543599] xenbus_probe_devices device
[   64.543776] xenbus_probe_devices failed xenbus_directory
[   64.543818] frontend_probe_and_watch devices probed ok
[   64.543972] frontend_probe_and_watch watch add ok ok
Quoted 8 lines
[   64.544021] frontend_probe_and_watch all done
[   67.040013] peth0: no IPv6 routers present
[   67.890015] eth0: no IPv6 routers present
[  118.012469] vif1.0 (vif): not using net_device_ops yet
[  118.127252] device vif1.0 entered promiscuous mode
[  118.130694] eth0: topology change detected, propagating
[  118.130701] eth0: port 2(vif1.0) entering forwarding state
[  122.657298] blkback: ring-ref 8, event-channel 9, protocol 1

(x86_64-abi)

Quoted 7 lines
[  128.700025] vif1.0: no IPv6 routers present
[  183.761252] eth0: port 2(vif1.0) entering disabled state
[  183.800702] eth0: port 2(vif1.0) entering disabled state

Line:-

[  122.657298] blkback: ring-ref 8, event-channel 9, protocol 1

(x86_64-abi)

Quoted 2 lines

is absent in first report. I believe it's a Dom0's kernel message

from network

Quoted 2 lines

backend driver. Instances dual boot on the same box.

I've seen similar things when building xenbus as modules; the backend modules don't get auto-loaded. Either compiling them non-modular or manually loading the modules works.

If that doesn't work, I'm not sure what's going on.

J

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel