>> … >>> Fix this by reordering the kfree() post the dereference. >> >> Would a wording approach (like the following) be a bit nicer? >> >> Move a kfree() call behind an assignment statement in the affected if branch. > > No, the suggested wording would make it less precise ("post the > dereference" -> "behind an assignment") and also tries to tell pointless > things about the location in the codei that is visible in the patch itself. Would you eventually like another wording variant a bit more? Thus move a kfree() call behind a dereference of an invalid pointer. Regards, Markus