On Wed, Sep 14, 2016 at 07:56:55PM +0200, Christian Borntraeger wrote: > On 09/14/2016 07:51 PM, Joe Perches wrote: > > checkpatch can be a useful tool for patches. > > > > It can be a much more controversial tool when used on files with the > > -f option for style and whitespace changes for code that is relatively > > stable, obsolete, or for maintained by specific individuals. > > > > o By default, allow checkpatch to be used with the -f|--file option > > for files in drivers/staging/ > > o Add an undocumented --force command line option to be used together > > with the -f|--file option to scan any file > > > > Signed-off-by: Joe Perches <j...@perches.com> > > cc: Greg KH <g...@kroah.com> > > cc: Jonathan Corbet <cor...@lwn.net> > > cc: Josh Triplett <j...@joshtriplett.org> > > cc: Christian Borntraeger <borntrae...@de.ibm.com> > > cc: Theodore Ts'o <ty...@mit.edu> > > This will certainly help to reduce the noise. On the other hand I remember > Linus > saying something along the line that he does not like the -f parameter (and he > prefers to set this automatically). So while I like the approach I am not > happy > enough to ack right now - still looking for a better alternative :-/
This seems entirely compatible with autodetection. If checkpatch detects that it runs on a file rather than a patch, it can assume -f. It can then apply this same logic to reject that if 1) in a kernel tree and 2) running on a non-staging file and 3) not passed --force.