linux-kernel - Re: [KJ][PATCH] Correct misc_register return code handling in several drivers

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] [thread-next>] [day] [month] [year] [list]

Message-Id: <1161729762.10524.660.camel@localhost.localdomain>
Date:	Wed, 25 Oct 2006 08:42:42 +1000
From:	Benjamin Herrenschmidt <benh@...nel.crashing.org>
To:	Neil Horman <nhorman@...driver.com>
Cc:	kernel-janitors@...ts.osdl.org, kjhall@...ibm.com, akpm@...l.org,
	maxk@...lcomm.com, linux-kernel@...r.kernel.org
Subject: Re: [KJ][PATCH] Correct misc_register return code handling in
	several drivers

On Tue, 2006-10-24 at 08:53 -0400, Neil Horman wrote:
> On Tue, Oct 24, 2006 at 01:34:34PM +1000, Benjamin Herrenschmidt wrote:
> > On Mon, 2006-10-23 at 13:19 -0400, Neil Horman wrote:
> > > Hey All-
> > > 	Janitor patch to clean up return code handling and exit from failed
> > > calls to misc_register accross several modules.
> > 
> > The patch doesn't match the description... What are those INIT_LIST_HEAD
> > things ? Is this something I've missed or is this a new requirement for
> > all misc devices ? Can't it be statically initialized instead ?
> > 
> 
> The INIT_LIST_HEAD is there to prevent a potential oops on module removal.
> misc_register, if it fails, leaves miscdevice.list unchanged.  That means its
> next and prev pointers contain NULL or garbage, when both pointers should contain
> &miscdevice.list. If we don't do that, then there is a chance we will oops on
> module removal when we do a list_del in misc_deregister on the moudule_exit
> routine.  I could have done this statically, but I thought it looked cleaner to
> do it with the macro in the code.

Hrm... I see, but I still for some reason don't like it that much.. I'd
rather have misc_register() do the initialisation unconditionally before
it can fail, don't you think ?

We would theorically have a similar problem with any driver that does


xxxx_register(&static_struct)

and

xxxx_unregister(&static_struct)

(pci, usb, etc...)

As long as there are list heads involved. I think the proper solution
here is to have either the unregister be smart and test for NULL/NULL or
the register initialize those fields before it has a chance to fail.

Ben.


-
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/