On Sat, 4 Apr 2015, Dan Carpenter wrote: > On Sat, Apr 04, 2015 at 06:20:53PM +0200, Julia Lawall wrote: > > Couldn't one say: > > > > x = NULL; > > y = &x->whatever; > > z = container_of(y, struct blah, whatever); > > > > and end up with z being NULL? > > That is crazy person code. It looks deliberately wrong. If we start > merging deliberate mistakes then we're already screwed. > > I have a smatch check which warns on container_of() but I should update > it to not complain if it's the first struct member. I have looked at > quite a few of these warnings and I worry that there are some places > where it relies on container_of() to be a no-op... That's also crazy > but it's the kind of crazy that people do in real life. :P OK. Should I update the patch to remove the test? I find 57 NULL tests on the result of container_of in the kernel as a whole. julia _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
