linux-kernel - Re: [PATCH] staging: r8723au: This patch tries to fix some byte order issues that is found by sparse check.

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: <wrfjmvpxp3jw.fsf@redhat.com>
Date:	Thu, 17 Mar 2016 10:44:19 -0400
From:	Jes Sorensen <Jes.Sorensen@...hat.com>
To:	Julian Calaby <julian.calaby@...il.com>
Cc:	Jandy Gou <qingsong.gou@...telecom.com>,
	Larry Finger <Larry.Finger@...inger.net>,
	Greg KH <gregkh@...uxfoundation.org>,
	linux-wireless <linux-wireless@...r.kernel.org>,
	"devel\@driverdev.osuosl.org" <devel@...verdev.osuosl.org>,
	"linux-kernel\@vger.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] staging: r8723au: This patch tries to fix some byte order issues that is found by sparse check.

Julian Calaby <julian.calaby@...il.com> writes:
> Hi Jandy,
>
> On Thu, Mar 17, 2016 at 7:03 PM, Jandy Gou <qingsong.gou@...telecom.com> wrote:
>> make C=1 M=drivers/staging/rtl8723au/
>>
>> drivers/staging/rtl8723au/hal/rtl8723a_cmd.c:96:38: warning: cast to
>> restricted __le16
>> drivers/staging/rtl8723au/hal/rtl8723a_cmd.c:100:27: warning: cast to
>> restricted __le32
>>
>> Signed-off-by: Jandy Gou <qingsong.gou@...telecom.com>
>
> I'm not sure your change is correct. Perhaps you should add temporary
> variables for h2c_cmd_ex and h2c_cmd while they're cpu-endian?
>
> Jes,
>
> I'm pretty sure this isn't the first time this has come up. Do you
> have any ideas for a solution? Or should we ignore this as this driver
> will eventually be going away?

Temp variables as you suggest would be a clean way to accomplish this.

Technically this might work, but esthetically this is so gross I wish I
had never seen it. There is a reason why we have the byteorder specific
types, and le{16,32}_to_cpus() violates that.

I am surprised we still have these macros around, tbh I didn't even know
they existed. A quick search for le16_to_cpus() shows that it's really
very old drivers and some more recent Intel Ethernet drivers that use
them - maybe this would be a good time to get rid of them completely.

Cheers,
Jes