Re: [Xen-devel] [PATCH 1/5] xen: add tainted state and show warning is gcov is enabled

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Fri, Sep 02, 2016 at 01:30:40PM +0100, Andrew Cooper wrote:
> On 02/09/16 13:26, Jan Beulich wrote:
> >>>> On 02.09.16 at 14:13, <andrew.cooper3@xxxxxxxxxx> wrote:
> >> On 02/09/16 13:06, Jan Beulich wrote:
> >>>>>> On 02.09.16 at 14:01, <wei.liu2@xxxxxxxxxx> wrote:
> >>>> On Fri, Sep 02, 2016 at 05:56:49AM -0600, Jan Beulich wrote:
> >>>>>>>> On 02.09.16 at 13:47, <wei.liu2@xxxxxxxxxx> wrote:
> >>>>> Since this is a config option - why bother issuing a warning and
> >>>>> tainting the hypervisor?
> >>>>>
> >>>> Because there isn't a clear indicator if gcov is enabled.
> >>>>
> >>>> I think it would be valuable to just tell from the backtrace or console
> >>>> log that gcov is enabled, then we can legitimately refuse to provide
> >>>> (security) support for such builds.
> >>> Then perhaps making it match the "debug=" would be the better
> >>> approach for a feature not controlled on the command line?
> >> I would prefer not to make it depend on debug=
> >>
> >> Coverage on a release hypervisor is equally important, and will be
> >> different from a debug hypervisor.
> > I didn't say "depend on", but "match" (which I mean just logging wise).
> >
> >> I am on the fence as to whether a taint is right to use, but I do think
> >> that a "with GCOV" is needed somewhere obvious on the banner line.
> > Right, hence the matching goal with "debug=".
> 
> Ah - I see what you mean.  Yes - that would be fine by me.
> 

Fine by me, too.

I will replace this patch with a new one.

Wei.

> ~Andrew

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