linux-kernel - Re: [PATCH] iio: dac: adi-axi-dac: drop io

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: <20250211192819.0c0be5fb@jic23-huawei>
Date: Tue, 11 Feb 2025 19:28:19 +0000
From: Jonathan Cameron <jic23@...nel.org>
To: Nuno Sá <noname.nuno@...il.com>
Cc: Angelo Dureghello <adureghello@...libre.com>, Lars-Peter Clausen	
 <lars@...afoo.de>, Michael Hennerich <Michael.Hennerich@...log.com>, Nuno
 Sa	 <nuno.sa@...log.com>, Jonathan Cameron <Jonathan.Cameron@...wei.com>,
 linux-iio@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] iio: dac: adi-axi-dac: drop io_mode check

On Tue, 11 Feb 2025 09:56:31 +0000
Nuno Sá <noname.nuno@...il.com> wrote:

> On Mon, 2025-02-10 at 19:13 +0000, Jonathan Cameron wrote:
> > On Mon, 10 Feb 2025 10:05:47 +0000
> > Nuno Sá <noname.nuno@...il.com> wrote:
> >   
> > > On Sat, 2025-02-08 at 15:45 +0000, Jonathan Cameron wrote:  
> > > > On Thu, 06 Feb 2025 09:36:14 +0100
> > > > Angelo Dureghello <adureghello@...libre.com> wrote:
> > > >     
> > > > > From: Angelo Dureghello <adureghello@...libre.com>
> > > > > 
> > > > > Drop mode check, producing the following robot test warning:
> > > > > 
> > > > > smatch warnings:
> > > > > drivers/iio/dac/adi-axi-dac.c:731 axi_dac_bus_set_io_mode()
> > > > >   warn: always true condition '(mode >= 0) => (0-u32max >= 0)'
> > > > > 
> > > > > The range check results not useful since these are the only
> > > > > plausible modes for enum ad3552r_io_mode.
> > > > > 
> > > > > Fixes: 493122c53af1 ("iio: dac: adi-axi-dac: add bus mode setup")
> > > > > Signed-off-by: Angelo Dureghello <adureghello@...libre.com>    
> > > > Ah. I missed this.  Anyhow made the same change directly so all is well
> > > > than ends well!
> > > >     
> > > 
> > > Hi Angelo, Jonathan,
> > > 
> > > I wanted to reply to this one when I saw it but I haven't done right away
> > > and
> > > then totally forgot. Sorry about that!
> > > 
> > > I don't really agree with the "fix" in this patch. AFAIU, smatch is
> > > complaining
> > > since the enum is apparently defaulting to an unsigned type which means
> > > doing
> > > the >= 0 check is useless. But we should keep the upper bound...  
> > 
> > Why? It's an enum so unless we are messing around with deliberate casts the
> > compiler should always be able to spot this. The check may be needed on a
> > future  
> 
> I do not think the compiler will catch this:
> 
> diff --git a/drivers/iio/dac/ad3552r-hs.c b/drivers/iio/dac/ad3552r-hs.c
> index c1dae58c1975..5234dd5e227d 100644
> --- a/drivers/iio/dac/ad3552r-hs.c
> +++ b/drivers/iio/dac/ad3552r-hs.c
> @@ -293,7 +293,7 @@ static int ad3552r_hs_buffer_postenable(struct iio_dev
> *indio_dev)
>          * Back bus to simple SPI, this must be executed together with above
>          * target mode unwind, and can be done only after it.
>          */
> -       st->data->bus_set_io_mode(st->back, AD3552R_IO_MODE_SPI);
> +       st->data->bus_set_io_mode(st->back, -1);
> 
> A W=1 build (clang) did not complained at all... Maybe tools like smatch will.
> 
> > date if we add more types to that enum.
> > 
> > So I agree the check wasn't terrible and perhaps acted as hardening but it
> > isn't strictly speaking doing anything today.
> >   
> 
> It's not a very super important check, I agree... and being an enum will be
> easier to spot a raw value being passed during a review but since we already had
> the check, I don't see why we should remove it completely and not keep the upper
> bound.

ok.  I'd take a patch putting the upper bound back.   Enums checking is an interesting
hole to fall down.

Jonathan

> 
> - Nuno Sá