linux-kernel - Re: [PATCH 1/1] minmax.h: Slightly relax the type checking done by min() and max().

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAHk-=wgZCBedi_xrysY2EAsN8tQjb3K4-qYtF-FaEE+GFuuE4Q@mail.gmail.com>
Date:   Sun, 27 Nov 2022 12:36:33 -0800
From:   Linus Torvalds <torvalds@...ux-foundation.org>
To:     David Laight <David.Laight@...lab.com>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        Joe Perches <joe@...ches.com>
Subject: Re: [PATCH 1/1] minmax.h: Slightly relax the type checking done by
 min() and max().

On Fri, Nov 25, 2022 at 7:00 AM David Laight <David.Laight@...lab.com> wrote:
>
> - Skip the type test if either argument is a positive 'int' constant.
>   Instead cast the constant to 'int', the compiler may promote it
>   back to 'unsigned int' when doing the test.

No. This looks very wrong to me.

Maybe I'm mis-reading something, but it looks like this makes a
"sizeof()" essentially be compatible with an "int" variable.

That is horrendously wrong. It should warn.

If you are doing a "min(i,sizeof(X))", and "i" is a signed integer,
then something is wrong. What does that code expect? It shouldn't
silently say "this is ok", because it most definitely isn't.

So maybe I'm  mis-reading this all and it doesn't actually do what I
think it does, but this seems to relax things *much* too much.

There's a reason we require types to be compatible, and you just
removed some of the important signedness checks.

                  Linus