Messages in this thread | | | Date | Tue, 8 Nov 2022 15:55:06 +0200 | From | Andy Shevchenko <> | Subject | Re: [PATCH v1 1/1] mac_pton: Don't access memory over expected length |
| |
On Mon, Oct 10, 2022 at 10:45:49PM +0200, Andrew Lunn wrote: > On Mon, Oct 10, 2022 at 11:29:37PM +0300, Andy Shevchenko wrote: > > On Mon, Oct 10, 2022 at 10:19:49PM +0200, Andrew Lunn wrote: > > > On Wed, Oct 05, 2022 at 07:43:01PM +0300, Andy Shevchenko wrote: > > > > The strlen() may go too far when estimating the length of > > > > the given string. In some cases it may go over the boundary > > > > and crash the system which is the case according to the commit > > > > 13a55372b64e ("ARM: orion5x: Revert commit 4904dbda41c8."). > > > > > > > > Rectify this by switching to strnlen() for the expected > > > > maximum length of the string. > > > > > > This seems like something which should have a fixes: tag, and be > > > against net, not net-next. > > > > I can (re-)send it that way. Just need a consensus by net maintainers. > > I would probably do: > > if (strnlen(s, maxlen) != maxlen) > return false; > > I doubt anybody is removing leading zeros in MAC addresses.
I'm not sure what this change gives us. < maxlen is more readable to understand that we refuse anything that less than maxlen, with your change it's easy to misinterpret it (by missing 'n' in the non-standard call) as "it must equal to maxlen" which is obviously not true.
That said, I leave it as is.
-- With Best Regards, Andy Shevchenko
| |