linux-kernel - RE: [PATCH 0/1] 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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Sat, 26 Nov 2022 10:21:28 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     'Andrew Morton' <akpm@...ux-foundation.org>
CC:     'Andy Shevchenko' <andriy.shevchenko@...ux.intel.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        "'Joe Perches'" <joe@...ches.com>,
        Linus Torvalds <torvalds@...ux-foundation.org>
Subject: RE: [PATCH 0/1] Slightly relax the type checking done by min() and
 max().

From: Andrew Morton
> Sent: 26 November 2022 00:02
> 
> On Fri, 25 Nov 2022 19:47:14 +0000 David Laight <David.Laight@...LAB.COM> wrote:
> 
> > > Blindly relaxing the rules is not an option in my opinion.
> >
> > The point is I'm not really relaxing the rules.
> > If anything I'm actually tightening them by allowing min() to
> > be used in more cases - so letting the buggy min_t() casts be
> > removed at some point in the future.
> 
> That sounds very virtuous.

Indeed.

> It would be helpful to see a few "convert min_t to min" patches
> to see this proposal in action.

I did try building a kernel with '#define min_t min' and then
picking up the pieces.
Doable but there are still too many false positives.
That is when I found some of the 'broken' uses of min_t().

I'd rather get this patch accepted and then build on top of it.
Then it should be possible to work out which min_t() actually
need to stay - I suspect very few.

It might even be worth converting min_t(type,x,y) to
min_t_type(x,y) and then converting the min_t_u[0-9]*(x,y) to
#define min_unsigned(x,y) min((x) + 0u + 0ull, (y) + 0u + 0ull)
I think that efficiently converts both values to an unsigned
type without masking the value.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)