On Sun, Nov 16, 2014 at 12:48:37PM +0100, SF Markus Elfring wrote:
> > An example of a bug introduced is here:
> >
> > https://lkml.org/lkml/2014/11/3/505
>
> It seems that we try to clarify a different interpretation of "bugs", don't
> we?
>
You removed the statement from "if (foo) kfree_fsm(foo);" so now it
prints a warning.
drivers/s390/net/fsm.c
71 void
72 kfree_fsm(fsm_instance *this)
73 {
74 if (this) {
75 if (this->f) {
76 kfree(this->f->jumpmatrix);
77 kfree(this->f);
78 }
79 kfree(this);
80 } else
81 printk(KERN_WARNING
82 "fsm: kfree_fsm called with NULL argument\n");
83 }
> It is an usual software development challenge to decide on the best source
> code places
> where to put input parameter validation (and when it can be omitted), isn't
> it?
No, it's not. You should just try to write the most readable software
you can instead of removing if statements because you can.
But that's not my point. My point is that these patches are not always
welcome so we should not merge them through the trivial tree.
regards,
dan carpenter
_______________________________________________
Cocci mailing list
[email protected]
https://systeme.lip6.fr/mailman/listinfo/cocci
