I pulled the latest xeno-1.2.bk this morning and am getting
a glitch in xc.domain_create
The domain ID it returns is a large bogus 64 bit number
This prints ''created 4613842385142171136 (4007a934400a4600)''
id = xc.domain_create( mem_kb=mem_size*1024, name=domain_name )
output (''created %d (%x) '' % (id, id))
A sane domain id appears from ''xc_dom_control.py list'', but
the domain
scripts fail because they use the bogus id returned from xc.domain_create
If I know the next ID that will get created (eg. 13), I can successfully
start up a new ID with something like this in xc_dom_create.py:
id = xc.domain_create( mem_kb=mem_size*1024, name=domain_name )
id = 13
The sources actually look OK. xc_domain_create says
u64 *pdomid
*pdomid = (u64)op.u.createdomain.domain;
and pyxc_domain_create says
u64 dom;
xc_domain_create(xc->xc_handle, mem_kb, name, &dom)
return PyLong_FromUnsignedLongLong(dom);
so I would expect xc.domain_create to return the right number.
This is all with xen, xenolinux and the utils built from the current 1.2
tree with gcc version 3.3.2 (Debian) and Python 2.3.2
-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel
> I pulled the latest xeno-1.2.bk this morning and am getting > a glitch in xc.domain_create > > The domain ID it returns is a large bogus 64 bit number > This prints ''created 4613842385142171136 (4007a934400a4600)'' > id = xc.domain_create( mem_kb=mem_size*1024, name=domain_name ) > output (''created %d (%x) '' % (id, id)) > > A sane domain id appears from ''xc_dom_control.py list'', but the domain > scripts fail because they use the bogus id returned from xc.domain_createOops. Yesterday we decided to backport a changeset from 1.3 to 1.2 to make domain id''s 64 bit. The intention was to ensure that the user-space tools remained identical in 1.3 and 1.2. Looks like we bungled it... You can probably do a temporary fix by reverting the last checkin in the BK repository (bk cset -x). I''ll take a look tomorrow. Ian ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel
> > > I pulled the latest xeno-1.2.bk this morning and am getting > > a glitch in xc.domain_create > > > > The domain ID it returns is a large bogus 64 bit number > > This prints ''created 4613842385142171136 (4007a934400a4600)'' > > id = xc.domain_create( mem_kb=mem_size*1024, name=domain_name ) > > output (''created %d (%x) '' % (id, id)) > > > > A sane domain id appears from ''xc_dom_control.py list'', but the domain > > scripts fail because they use the bogus id returned from xc.domain_create > > Oops. Yesterday we decided to backport a changeset from 1.3 to 1.2 > to make domain id''s 64 bit. The intention was to ensure that the > user-space tools remained identical in 1.3 and 1.2. Looks like we > bungled it... > > You can probably do a temporary fix by reverting the last > checkin in the BK repository (bk cset -x). I''ll take a look > tomorrow.Most likely the correct fix is to upgrade your installed toolset. The Xen interface hasn''t changed, but the interface between the C and Python user-space libraries _has_ changed. -- Keir ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel
> > > > > I pulled the latest xeno-1.2.bk this morning and am getting > > > a glitch in xc.domain_create > > > > > > The domain ID it returns is a large bogus 64 bit number > > > This prints ''created 4613842385142171136 (4007a934400a4600)'' > > > id = xc.domain_create( mem_kb=mem_size*1024, name=domain_name ) > > > output (''created %d (%x) '' % (id, id)) > > > > > > A sane domain id appears from ''xc_dom_control.py list'', but the domain > > > scripts fail because they use the bogus id returned from xc.domain_create > > > > Oops. Yesterday we decided to backport a changeset from 1.3 to 1.2 > > to make domain id''s 64 bit. The intention was to ensure that the > > user-space tools remained identical in 1.3 and 1.2. Looks like we > > bungled it... > > > > You can probably do a temporary fix by reverting the last > > checkin in the BK repository (bk cset -x). I''ll take a look > > tomorrow. > > Most likely the correct fix is to upgrade your installed toolset. The > Xen interface hasn''t changed, but the interface between the C and > Python user-space libraries _has_ changed.Ah... Actually there was a wafeur-thin change to the Xen interface. I''ve checked it into the 1.2 repository now. If you pull the fix and rebuild Xen then all should be fine. :-) -- Keir ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel
> Ah... Actually there was a wafeur-thin change to the Xen > interface. I''ve checked it into the 1.2 repository now. If you pull > the fix and rebuild Xen then all should be fine. :-)I''ve verified that the fault that was introduced yesterday into the tools in the 1.2 tree is now fixed. I''ve updated the tar ball on the downloads page too. Apologies to all who were effected by the bug. We decided that it would be sensible to backport a couple of minor internal interface changes from the 1.3 tree so that we could keep all of the user-space python tools identical in both trees. Although its very late in the release cycle, we thought the change wouldn''t be visible to anyone in user space, and the benefits of maintaining future tool compatibility would be well worthwhile. Unfortunately, the cset didn''t quite get this right... Ian ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click _______________________________________________ Xen-devel mailing list Xen-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/xen-devel