Messages in this thread | | | Date | Fri, 14 Mar 2003 13:26:02 +0300 | From | Oleg Drokin <> | Subject | Re: [2.4] Multiple memleaks in IBM Hot Plug Controller Driver |
| |
Hello!
On Fri, Mar 14, 2003 at 11:54:40AM +0200, Denis Vlasenko wrote:
> > + if (!str) > > + return NULL; > > memset (str, 0, 3); > This was fun, right? "Lets add second memset just in case > first failed" ;) ;)
Ah, I was half asleep when doing the change, so I missed other obvious stuff it seems, like memset of three bytes to two byte variabe :) memset on wrong thing (second memset should be applied to str1 of course.)
But as Greg KH said already, whis stuff should be replaced by version from 2.5 instead ;)
> > bit = (int)(var / 10); > > switch (bit) { > > @@ -608,13 +608,20 @@ > > return str; > > default: > > //2 digits number > > + str1 = (char *) kmalloc (2, GFP_KERNEL); > > + if (!str1) { > > + break; > > + } > > + memset (str, 0, 3); > > + memset (str, 0, 3); > > *str1 = (char)(bit + 48); > > strncpy (str, str1, 1); > > memset (str1, 0, 3); > Wow! *str1 is 2 bytes long, not 3!
yup.
> Anyway, here is the diff against some old 2.5 (sorry don't have latest > tree here at the moment). Also here are old and new functions > for easy visual diff. Completely untested:
New 2.5 code does not have this function at all. I will look into porting 2.5 changes back to 2.4 tonight. This seems to be proper solution in this case.
Bye, Oleg - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |