[Linux-kernel-mentees] checkpatch: add fixes for BAD_SIGN_OFF

[Lukas Bulwahn]

On Tue, Nov 3, 2020 at 1:11 PM Aditya <yashsri421 at gmail.com> wrote:
>
> On 3/11/20 4:11 pm, Lukas Bulwahn wrote:
>
> >> 1) WARNING:BAD_SIGN_OFF: Co-developed-by: should not be used to
> >> attribute nominal patch author '"Zhang, Jun" <jun.zhang at intel.com>'
> >>
> >
> > I think the fix here might be to drop the Co-developed-by: line,
> > right?
> > Can you confirm this fix on your evaluation?
> >
>
> Yes, we will drop the Co-developed-by line for the author. As he has
> already signed off for the patch, co-developed doesn't make sense.
>
> >> 2) WARNING:BAD_SIGN_OFF: email address '"Matthew Wilcox (Oracle)"
> >> <willy at infradead.org>' might be better as 'Matthew Wilcox (Oracle)
> >> <willy at infradead.org>'
> >>
> >
> > I think Dwaipayan and Joe are discussing the proper handling of
> > quotes. I suggest to understand the resolution that Dwaipayan and
> > Joe
> > are working out before thinking about a fix for this option. They
> > might decide that only a few specific cases should be warned about
> > or not at all in case of quote mismatches.
> >
>
> Alright, I'll omit this one for now.
>
> >> 3) WARNING:BAD_SIGN_OFF: Duplicate signature
> >> etc.
> >>
> >
> > I do not see the fix option for this case 3; the duplicate signatures
> > are mostly due to the integration Signed-off-by trail.
> >
>
> Sir, mostly the duplicate signatures are occurring when the user signs
> off 2 or more times for the same role. For eg,
>
> For the commit 6dd47d9754ff ("mac80211: fix
> ieee80211_txq_setup_flows() failure path"), we get
>
> WARNING:BAD_SIGN_OFF: Duplicate signature
> #16:
> Signed-off-by: Johannes Berg <johannes.berg at intel.com>
>
> as the user has used 'Signed-off-by' line 2 times. The fix for this
> can be to remove the repeated line, similar to co-developed-by one
> we're doing for the (1) case.
>

Bad example here as in many others probably as well. Johannes signed
off in two different roles, once as author and once as maintainer
(integrating it into the kernel tree).

It is NOT recommended to simply delete the second sign-off.

Lukas