Re: [Xen-devel] [OSSTEST PATCH 4/9] mg-schema-test-database: Borrow shares properly

[Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>]

Ian Campbell writes ("Re: [OSSTEST PATCH 4/9] mg-schema-test-database: Borrow 
shares properly"):
> On Thu, 2015-12-17 at 17:06 +0000, Ian Jackson wrote:
> >     # As we copy, we note everything we're not borrowing as
> > -   # belonging to the parent db.
> > +   # belonging to the parent db.  We borrow shares of a shared
> > +   # resource.  If we borrow only some rather than all of the
> > +   # shares, neither DB will be able to unshare it.
> 
> This is what actually happens, rather than this comment explaining a thing
> we are avoiding, right?

Yes.

> I ask because although the text (and the use of "properly" in the subject)
> seems pretty clear it's the former I was surprised to find this code was
> borrowing partial shares and therefore setting up such a situation. Given
> the caveat below would it not be better to just never allow this?

I have added this to the commit message:

 Previously, the test database would be generated in a broken state:
 resources share-host/foo/{1,2,...} exist but the resource host/foo/0
 is allocated to magic/xdbref rather than to magic/shared.  This causes
 various resource allocation machinery to crash.  (Even if the host is
 entirely un-borrowed.)

I guess it might be possible for mg-schema-test-database to check that
either none or all of the shares of a host are being borrowed, but I
don't think it's worth it.  The osstest git hash in the sharetype will
in practice mostly prevent undesirable sharing of the same resource
between test and real db instances.

> I suppose that given the list of tasks to borrow came from the user
> this might be considered a "keep both pieces" situation.

Yes.  This partial sharing can only affect you if you have allocated
individual shares to the task you say you want to borrow from.

Ian.

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