> > +mutex_init(\(&mm->fld\|&(mm->fld)\)) > > > An extra SmPL disjunction is probably unnecessary because of an isomorphism. > https://gitlab.inria.fr/coccinelle/coccinelle/-/blob/3f7496ff9c2c5d4fadae1e585aa458e1a0037972/standard.iso#L382 > https://github.com/coccinelle/coccinelle/blob/19ee1697bf152d37a78a20cefe148775bf4b0e0d/standard.iso#L382 This is correct. It would be better to only have the version with the parentheses, because Coccinelle will consider the case without parentheses anyway. > > > +mutex_init(&(mm->fld)) > > > > … > > > +@r3@ > > +identifier s, fld; > > +position p != {r2.p}; > > > Why do you think that such a SmPL constraint would be required? > > > > +@@ > > + > > +struct s { > > + ... > > + struct mutex fld@p; > > + ... > > +}; > > > Why would the source code search repetition matter here? He is searching for a structure that is different from the ones matched previously. > > > … > > > +@r5 depends on r4@ > > +identifier r3.s; > > +struct s *mm; > > +position p; > > +@@ > > +* mm@p = \(kmalloc\|kzalloc\|devm_kmalloc\|devm_kzalloc\)(...) > > > I would expect that the usage of the asterisk in the first column should belong > to the operation mode “context”. This is correct. Either the context mode should be fully supported or the asterisk should be removed. julia
