linux-kernel - Re: [PATCH v5] vfat: Deduplicate hex2bin()

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: <1501592205.29303.320.camel@linux.intel.com>
Date:   Tue, 01 Aug 2017 15:56:45 +0300
From:   Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To:     OGAWA Hirofumi <hirofumi@...l.parknet.co.jp>,
        Andy Shevchenko <andy.shevchenko@...il.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Joe Perches <joe@...ches.com>
Subject: Re: [PATCH v5] vfat: Deduplicate hex2bin()

On Tue, 2017-08-01 at 05:54 +0900, OGAWA Hirofumi wrote:
> Andy Shevchenko <andy.shevchenko@...il.com> writes:
> 
> > > +
> > > +                               *(wchar_t *)op = uc[0] << 8 |
> > > uc[1];
> > > +
> > > +                               op += 2;
> > 
> > This had been in the original patch 6 years ago and had been refused
> > because of endianess issues.
> 
> Sorry, I forgot what I said completely. Maybe I changed my mind? 
> 
> 			if (uni_xlate == 1) {
> 				*op++ = ':';
> 				op = hex_byte_pack(op, ec >> 8);
> 				op = hex_byte_pack(op, ec);
> 				len -= 5;
> 
> Here is output. So "uc[0] << 8 | uc[1]" is right code, isn't it?

Yes, the right part of the expression is correct.

If *(wchar_t *)op is what we are expecting, that it's okay.

> > >                                 charlen = nls->char2uni(ip, len -
> > > i,
> > > -                                                                 
> > >       (wchar_t *)op);
> > > +                                                       (wchar_t
> > > *)op);
> > 
> > It perfectly fits one line.
> 
> It over 80 column.

For one character? :-)

>>> > +                          fill = hex2bin(hc, ip + 1, 2);
>>> > +                          if (fill)
>>> > +                                  return fill;
 
>>> This should not use random errno (in this case, it is -1 (EPERM)).

>> You perhaps missed the side note I put after --- line.
>> It reflects this change.

> Sure, I missed to read it. But same here, hex2bin() doesn't care FS's
> errno, it is what "random errno" I meant (I.e. hex2bin() might change
it
> to bool or any other errno again).

I see your point. 

I'm fine with shadowing in cases when it's strictly needed, otherwise
how can it be changed to bool without compile time issues? Any other
error code changes for a such widely used helper might be a disaster not
only for this driver, so it's quite unlikely.

At the end it's up to you to decide. So, I'm fine with the patch Andrew
took.

-- 
Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Intel Finland Oy