Re: [Xen-devel] [PATCH 1/2] libxl: move check for existence of qemuu device model

[Ian Campbell <ian.campbell@xxxxxxxxxx>]

On Tue, 2013-03-05 at 11:22 +0000, Ian Jackson wrote:
> Ian Campbell writes ("Re: [Xen-devel] [PATCH 1/2] libxl: move check for 
> existence of qemuu device model"):
> > On Mon, 2013-03-04 at 11:42 +0000, Ian Jackson wrote:
> > > No, because the whole thing is in an if which applies only if we're
> > > trying to use ..._QEMU_XEN.
> > 
> > Oh, right. I combined the tow patches wrongly in my head.
> > 
> > Don't we also want to check for the presence of the trad binary though?
> 
> AIUI the trad binary is always built ?

By us it is, but e.g. Debian's packaging has been known to omit it, or
the user may have overridden the path but made a typo.

>   AIUI there is a possibility
> that the new one isn't ?  And that was the rationale for this check ?
> 
> CCing Stefano.
> 
> > > > On a related note we could do with stating the device model even in the
> > > > case where the user has specified device_model_version explicitly, to
> > > > give better error reporting in that case too.
> > > 
> > > You mean logging it ?
> > 
> > Actually I meant "not trying to start the guest regardless and returning
> > success", which AIUI is what happens... Logging would be nice too
> > though.
> 
> Oh I see.  TBH once we start to think about these kinds of
> installation problems it's difficult to see where we'll stop.

This is one which has been tripping people up in practice.


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