linux-kernel - Re: [PATCH] FE_GET_PROPERTY should be _IOW, because the associated structure is transferred from userspace to kernelspace. Keep the old ioctl around for compatibility so that existing code is not broken.

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: <4DE8E7D9.9000608@redhat.com>
Date:	Fri, 03 Jun 2011 10:55:37 -0300
From:	Mauro Carvalho Chehab <mchehab@...hat.com>
To:	Andreas Oberritter <obi@...uxtv.org>
CC:	Hans Petter Selasky <hselasky@....net>,
	"linux-media@...r.kernel.org" <linux-media@...r.kernel.org>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] FE_GET_PROPERTY should be _IOW, because the associated
 structure is transferred from userspace to kernelspace. Keep the old ioctl
 around for compatibility so that existing code is not broken.

Em 03-06-2011 09:44, Andreas Oberritter escreveu:
> On 06/01/2011 11:15 PM, Mauro Carvalho Chehab wrote:
>> The dvb_usercopy will do the right thing, if we use _IOR or _IORW.
> 
> It only works, because _IOC_READ triggers a copy_from_user, as a
> workaround for wrongly marked ioctls like this, according to a code
> comment. It does not really do the right thing, because in this special
> case the later call to copy_to_user isn't required. But it doesn't do
> any real harm either.

Yes, that's what I meant to say ;) The workaround for it is there already,
so maybe there are other ioctl's using the wrong _IOC_ directions.

As I said before, some ioctl's don't use _IOC_ directions, like for example
the tty ioctls like TIO* ones. This happens on several very old drivers.
So, ioctl core don't make any assumption about them. it is up to each driver
(or subsystem core) to handle it.

>> I prefer to not apply this patch, as it won't fix anything. Adding an _OLD means
>> that we'll need later to remove it, causing a regression. Ok, we may do like we did
>> with V4L _OLD ioctl's that were marked as _OLD at 2.6.5 and were removed on a late
>> 2.6.3x.
> 
> Either way is fine for me.

I'm not against fixing it, but, in this case, we'll need to validate all DVB
ioctl's and remove the IOC_READ hack for all non-_OLD controls, and writing
a notice at features-to-be-removed announcing that the _OLD controls will be
removed.

Cheers,
Mauro.
> 
> Regards,
> Andreas
> --
> To unsubscribe from this list: send the line "unsubscribe linux-media" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/