linux-kernel - Re: [PATCH v1 1/1] pktcdvd: Use clamp

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]

Message-ID: <ZJG/tfh7kKauoMZJ@smile.fi.intel.com>
Date:   Tue, 20 Jun 2023 18:03:17 +0300
From:   'Andy Shevchenko' <andriy.shevchenko@...ux.intel.com>
To:     David Laight <David.Laight@...lab.com>
Cc:     Jens Axboe <axboe@...nel.dk>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "linux-block@...r.kernel.org" <linux-block@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1 1/1] pktcdvd: Use clamp_val() instead of min()+max()

On Tue, Jun 20, 2023 at 01:35:11PM +0000, David Laight wrote:
> From: 'Andy Shevchenko'
> > Sent: 20 June 2023 14:25
> > On Tue, Jun 20, 2023 at 12:06:49PM +0000, David Laight wrote:

...

> > > > +		*hi = clamp_val(*hi, 500, 1000000);
> > >
> > > (standard rant about minmax.h)
> > >
> > > clamp_val() is pretty much broken by design.
> > > It MIGHT be ok here but it casts both limits to the
> > > type of the value being compared.
> > > In general that is just plain wrong.
> > >
> > > Like min_t() it is generally ok because the kernel only uses
> > > unsigned values between 0 and MAXINT.
> > >
> > > If min/max were ok, then using clamp() should also be ok.
> > 
> > Submit a patch to fix it, if you think you can make it better.
> > Obviously your comment can be addressed separately if we even
> > need that.
> 
> Did you try using clamp() ?
> 
> To see why clamp_val() is broken consider?
> 	unsigned char val = 200;
> 	...
> 	xxx = clamp_val(val, 10, 300);
>  
> This sets xxx to 44 - not exactly expected.

Right, clamp_val() has to be improved.
However this is not the case in this driver.

I have just sent a v2 with clamp().

-- 
With Best Regards,
Andy Shevchenko