Re: [CHECKER] 32 Memory Leaks on Error Paths

[Valdis . Kletnieks]

On Tue, 16 Sep 2003 10:52:06 EDT, Timothy Miller said:
> 
> >> 276:	/* OK, it's not open. Create cache info for it */
> >>START -->
> >> 277:	mtdblk = kmalloc(sizeof(struct mtdblk_dev), GFP_KERNEL);
> >> 278:	if (!mtdblks)
> >>END -->
> >> 279:		return -ENOMEM;
> 
> > 
> > Invalid.  This is quite an obvious false positive, at least if your
> > algorithm checks for possible value ranges.
> 
> Wait... one is "mtdblk", and the other is "mtdblks".  One has an extra 
> 's' on it.  Unless there is some kind of aliasing going on, they would 
> appear to be different variables.  Naturally, I didn't check the 
> original code, so I could be full of it.  :)

On the other hand, even if the report was invalid (which another posting says isn't
the case), this code is just *begging* for somebody to "fix the typo", due to how
much the kernel uses
	foo = function(..)
	if (!foo)
Yeah, they're different variables - but we alloc mktblk and then return -ENOMEM
if the OTHER variable is null?? ;)

Attachment: pgp00001.pgp
Description: PGP signature