Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 14 Sep 2022 17:03:09 +0300
From:   Andy Shevchenko <andriy.shevchenko@intel.com>
To:     "Farber, Eliav" <farbere@amazon.com>
Cc:     jdelvare@suse.com, linux@roeck-us.net, robh+dt@kernel.org,
        p.zabel@pengutronix.de, rtanwar@maxlinear.com,
        linux-hwmon@vger.kernel.org, devicetree@vger.kernel.org,
        linux-kernel@vger.kernel.org, hhhawa@amazon.com, jonnyc@amazon.com
Subject: Re: [PATCH v5 20/21] hwmon: (mr75203) add debugfs to read and write
 temperature coefficients
Message-ID: <YyHfHSIz9F+1SGnX@smile.fi.intel.com>
References: <20220908152449.35457-1-farbere@amazon.com>
 <20220908152449.35457-21-farbere@amazon.com>
 <YxowTBIODMLjf1Ek@smile.fi.intel.com>
 <581a4a0b-8e0e-b7a2-f873-77ed74b54e96@amazon.com>
 <3b121ab4-dd64-68b3-ee89-8571b5d3651e@amazon.com>
 <YyC3hsNhbQGIlReU@smile.fi.intel.com>
 <bdb73546-f309-60dd-3c40-d749654228fe@amazon.com>
 <YyGfvzlCu9qgtgA0@smile.fi.intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <YyGfvzlCu9qgtgA0@smile.fi.intel.com>
Organization: Intel Finland Oy - BIC 0357606-4 - Westendinkatu 7, 02160 Espoo
Precedence: bulk

On Wed, Sep 14, 2022 at 12:32:47PM +0300, Andy Shevchenko wrote:
> On Wed, Sep 14, 2022 at 07:26:36AM +0300, Farber, Eliav wrote:
> > On 9/13/2022 8:01 PM, Andy Shevchenko wrote:
> > > On Tue, Sep 13, 2022 at 05:40:16PM +0300, Farber, Eliav wrote:
> > > > On 9/13/2022 4:06 PM, Farber, Eliav wrote:

...

> > > > It seems like debugfs_attr_write() calls simple_attr_write() and it uses
> > > > kstrtoull(), which is why it fails when setting a negative value.
> > > > This is the same also in v6.0-rc5.
> > > > 
> > > > debugfs_attr_read() on the other hand does show the correct value also
> > > > when j is negative.
> > > 
> > > Which puzzles me since there is a few drivers that use %lld.
> > > Yeah, changing it to
> > > 
> > >        ret = sscanf(attr->set_buf, attr->fmt, &val);
> > >        if (ret != 1)
> > >                ret = -EINVAL;
> > > 
> > > probably can fix that. Dunno if debugfs maintainer is okay with this.
> > > 
> > > P.S. This needs revisiting all format strings to see if there are no
> > > additional
> > > characters, otherwise that needs to be addressed first, if feasible.
> > 
> > I was thinking of making such a correction:
> > 
> > -       ret = kstrtoull(attr->set_buf, 0, &val);
> > +       if (attr->set_buf[0] == '-')
> > +               ret = kstrtoll(attr->set_buf, 0, &val);
> > +       else
> > +               ret = kstrtoull(attr->set_buf, 0, &val);
> > 
> > and when I tested the change it worked, but then I noticed this commit:
> > 
> > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/fs/libfs.c?h=v6.0-rc5&id=488dac0c9237647e9b8f788b6a342595bfa40bda
> > 
> > According to this, it previously used simple_strtoll() which supports
> > negative values, but was changed to use kstrtoull() to deliberately
> > return '-EINVAL' if it gets a negative value.
> > 
> > So I’m not sure debugfs maintainers will be okay with a fix that
> > basically reverts the commit I mentioned.
> > Hence, what do you suggest to do with my commit?
> > Is it ok to leave it as it is today?
> 
> Meanwhile asking is not a problem, at least we will know for sure.
> And yes, leave it as is, but point to the thread where you asking
> the clarification.

For the record:

$ git grep -n -A1 -w DEFINE_DEBUGFS_ATTRIBUTE | grep ');' | sed 's,.*\(".*%.*"\).*,\1,' | sort | uniq -c
  1 "%08llx\n"
  5 "0x%016llx\n"
  5 "0x%02llx\n"
  5 "0x%04llx\n"
 13 "0x%08llx\n"
  1 "0x%4.4llx\n"
  3 "0x%.4llx\n"
  4 "0x%llx\n"
  1 "%1lld\n"
 40 "%lld\n"
  2 "%lli\n"
129 "%llu\n"
  1 "%#llx\n"
  2 "%llx\n"

means that sscanf() should work and fix the issue. You may even propose a patch
as a starter for a discussion.

-- 
With Best Regards,
Andy Shevchenko