On Thu, Jun 04, 2020 at 12:08:49AM +0200, Linus Walleij wrote: > On Mon, Jun 1, 2020 at 8:31 PM Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote: > > On Mon, Jun 01, 2020 at 01:31:23PM +0200, Christophe JAILLET wrote: > > > Le 01/06/2020 à 10:58, Robert Jarzmik a écrit : > > > > Christophe JAILLET <christophe.jaillet@xxxxxxxxxx> writes: > > > > > > > > > Commit 6d33ee7a0534 ("pinctrl: pxa: Use devm_pinctrl_register() for pinctrl registration") > > > > > has turned a 'pinctrl_register()' into 'devm_pinctrl_register()' in > > > > > 'pxa2xx_pinctrl_init()'. > > > > > However, the corresponding 'pinctrl_unregister()' call in > > > > > 'pxa2xx_pinctrl_exit()' has not been removed. > > > > > > > > > > This is not an issue, because 'pxa2xx_pinctrl_exit()' is unused. > > > > > Remove it now to avoid some wondering in the future and save a few LoC. > > > > > > > > > > Signed-off-by: Christophe JAILLET <christophe.jaillet@xxxxxxxxxx> > > > > Acked-by: Robert Jarzmik <robert.jarzmik@xxxxxxx> > > > > > > > > Would be even a better patch with a : > > > > Fixes: 6d33ee7a0534 ("pinctrl: pxa: Use devm_pinctrl_register() for pinctrl registration") > > > > > > I was wondering it was was needed in this case. > > > The patch does not really fix anything, as the function is unused. Or it > > > fixes things on a theoretical point of view. > > > > There is no concensus... We should call a vote on this at Kernel > > Summit. :P > > Fixes means it fixes something that was wrong in that commit. > That's all. Whether syntactic or semantic or regression or > serious or not does not matter. It is also not compulsory to > add it is just helpful. Fixes tag should be compulsory for actual bug fixes. We had a the Bad Binder exploit last year because commit f5cb779ba163 ("ANDROID: binder: remove waitqueue when thread exits.") had no Fixes tag and wasn't backported to Android kernels. regards, dan carpenter