linux-kernel - Re: [PATCH 2/2] Input: atmel_mxt_ts - fix double free of input device

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 for Android: free password hash cracker in your pocket

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

Date:	Tue, 9 Sep 2014 16:49:22 -0700
From:	Dmitry Torokhov <dmitry.torokhov@...il.com>
To:	nick.dyer@...ev.co.uk
Cc:	Stephen Warren <swarren@...dotorg.org>,
	Yufeng Shen <miletus@...gle.com>,
	Daniel Kurtz <djkurtz@...omium.org>,
	Henrik Rydberg <rydberg@...omail.se>,
	Joonyoung Shim <jy0922.shim@...sung.com>,
	Alan Bowens <Alan.Bowens@...el.com>,
	linux-input@...r.kernel.org, linux-kernel@...r.kernel.org,
	Peter Meerwald <pmeerw@...erw.net>,
	Benson Leung <bleung@...omium.org>,
	Olof Johansson <olofj@...omium.org>,
	Sekhar Nori <nsekhar@...com>
Subject: Re: [PATCH 2/2] Input: atmel_mxt_ts - fix double free of input device

On Tue, Sep 09, 2014 at 03:50:49PM +0100, nick.dyer@...ev.co.uk wrote:
> From: Stephen Warren <swarren@...dotorg.org>
> 
> [reworked after comments by Dmitry Torokhov. Move free of input device into
> separate function. Only call in paths that require it. Move mxt_initialize
> after sysfs init, because otherwise an error in the sysfs init may interfere
> with the async return from the firmware loader. Add guards for sysfs
> functions. ]

Ugh... there is still problem with asycn firmware loading: you need to
make sure it is done before you try to unbind the dveice. I also do not
see what stops several firmware update requests to happen
simultaneously. Once you add proper handling for that you can use the
same lock in sysfs read methods.

Another option is wait a bit and see what's the outcome of async probing
discussion on LKML is and maybe we can stop using
request_firmware_nowait() in probe path but rather have device core fire
off probe asynchronously.

I'd rather have fix for input device freeing be separate from
sysfs/firmware/config loading changes.

Thanks.

-- 
Dmitry
--
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/