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

Message-ID: <0b38ced96519407e95962aef2771bbc6@AcuMS.aculab.com>
Date:   Fri, 25 Nov 2022 16:14:58 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     'Andy Shevchenko' <andriy.shevchenko@...ux.intel.com>
CC:     LKML <linux-kernel@...r.kernel.org>,
        Andrew Morton <akpm@...ux-foundation.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: 'Andy Shevchenko'
> Sent: 25 November 2022 15:58
> 
> On Fri, Nov 25, 2022 at 03:27:07PM +0000, David Laight wrote:
> > From: Andy Shevchenko
> > > Sent: 25 November 2022 15:21
> > > On Fri, Nov 25, 2022 at 03:00:40PM +0000, David Laight wrote:
> > > > The min() and max() defines include a type check to avoid the unexpected
> > > >   behaviour when a negative value is compared against and unsigned value.
> > > > However a lot of code hits this check and uses min_t() to avoid the error.
> > > > Many of these are just plain wrong.
> > > >
> > > > Those casting to u8 or u16 are particularly suspect, eg:
> > > > drivers/usb/misc/usb251xb.c:528:
> > > > 		hub->max_current_sp = min_t(u8, property_u32 / 2000, 50);
> > >
> > > I don't buy this. What's exactly wrong with this code?
> >
> > Consider what happens if propery_u32 is 512000.
> > The returned value is 0 not 50.
> 
> I considered that and there are two things to consider on your side:
> 1) it's coming from device property;
> 2) device property is validated using YAML schema.
> 
> On top of that, the wrong property is on the user. We have a lot of stuff that
> user may put wrongly, but it's user's choice.
> 
> Any better example, please?

How about:

data_size = min_t(u16, buf_size, len);

https://elixir.bootlin.com/linux/v6.1-rc6/source/kernel/printk/printk_ringbuffer.c#L1738


Now, maybe, you could claim that buf_size > 64k never happens.
But the correct cast here is u32 to match buf_size.
len (being u16) will be promoted to int before the compare.

Just search the kernel for "min_t(u8," or "min_t(u16," while some might
be ok, I really wouldn't want to verify each case.

If you look hard enough there are also some:
	u32_var = min_t(u32, u32_val, u64_val);
where the intent is to limit values that might be invalid for u32.

I did try compiling a kernel with min_t() defined to be min() (no casts)
but there were too many false positives without allowing all
unsigned v unsigned compares.

	David

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