On Fri, Jan 20, 2012 at 11:54 AM, Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote: > It still complains about the following macros where parenthesis are > not needed. > > ERROR: Macros with complex values should be enclosed in parenthesis > #156: FILE: staging/android/pmem.c:156: > +#define PMEM_IS_FREE(id, index) !(pmem[id].bitmap[index].allocated) > > Let's just make the check look for an operator with a low > precedence. > http://en.wikipedia.org/wiki/Order_of_operations#Programming_languages > > Otherwise the submitters are going to change it to: > > #define PMEM_IS_FREE(id, index) (!(pmem[id].bitmap[index].allocated)) > > That has two pairs of unneeded paranthesis and we run the risk of > reprogramming the kernel in lisp, by mistake. Yep we want to avoid that. As -> binds tighter than many of the operators in these bands we can only safely avoid the ()'s if the operator does not potentially make a pointer, so for now I have added unary minus and the two unary not operators. I think I want to see some examples of other operators before I would be keen to have any further exceptions. Anyhow how does this work for you: http://people.canonical.com/~apw/checkpatch/checkpatch-next.pl -apw _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
