[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <1b2a5353-d7c5-dd9c-3478-27825ec1f52a@osg.samsung.com>
Date: Mon, 14 Nov 2016 16:41:25 -0700
From: Shuah Khan <shuahkh@....samsung.com>
To: Max Kellermann <max.kellermann@...il.com>,
linux-media@...r.kernel.org, mchehab@....samsung.com
Cc: linux-kernel@...r.kernel.org, shuah Khan <shuahkh@....samsung.com>
Subject: Re: [PATCH 01/12] [media] rc-main: clear rc_map.name in
ir_free_table()
On 08/09/2016 03:32 PM, Max Kellermann wrote:
> rc_unregister_device() will first call ir_free_table(), and later
> device_del(); however, the latter causes a call to rc_dev_uevent(),
> which prints rc_map.name, which at this point has already bee freed.
>
> This fixes a use-after-free bug found with KASAN.
>
> Signed-off-by: Max Kellermann <max.kellermann@...il.com>
> ---
> drivers/media/rc/rc-main.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/media/rc/rc-main.c b/drivers/media/rc/rc-main.c
> index 8e7f292..1e5a520 100644
> --- a/drivers/media/rc/rc-main.c
> +++ b/drivers/media/rc/rc-main.c
> @@ -159,6 +159,7 @@ static void ir_free_table(struct rc_map *rc_map)
> {
> rc_map->size = 0;
> kfree(rc_map->name);
> + rc_map->name = NULL;
> kfree(rc_map->scan);
> rc_map->scan = NULL;
> }
>
Hi Mauro,
Could you please get this fix into 4.9. I am seeing the following when I do
rmmod on au0828
[ 179.010878] ==================================================================
[ 179.010895] BUG: KASAN: use-after-free in string+0x170/0x1f0 at addr ffff8801bd513000
[ 179.010900] Read of size 1 by task rmmod/1831
[ 179.010908] CPU: 1 PID: 1831 Comm: rmmod Tainted: G W 4.9.0-rc5 #5
[ 179.010910] Hardware name: Hewlett-Packard HP ProBook 6475b/180F, BIOS 68TTU Ver. F.04 08/03/2012
[ 179.010914] ffff8801aea2f680 ffffffff81b37ad3 ffff8801fa403b80 ffff8801bd513000
[ 179.010922] ffff8801aea2f6a8 ffffffff8156c301 ffff8801aea2f738 ffff8801bd513000
[ 179.010930] ffff8801fa403b80 ffff8801aea2f728 ffffffff8156c59a ffff8801aea2f770
[ 179.010937] Call Trace:
[ 179.010944] [<ffffffff81b37ad3>] dump_stack+0x67/0x94
[ 179.010950] [<ffffffff8156c301>] kasan_object_err+0x21/0x70
[ 179.010954] [<ffffffff8156c59a>] kasan_report_error+0x1fa/0x4d0
[ 179.010968] [<ffffffffa116f05f>] ? au0828_exit+0x10/0x21 [au0828]
[ 179.010973] [<ffffffff8156c8b3>] __asan_report_load1_noabort+0x43/0x50
[ 179.010978] [<ffffffff81b58b20>] ? string+0x170/0x1f0
[ 179.010982] [<ffffffff81b58b20>] string+0x170/0x1f0
[ 179.010987] [<ffffffff81b621c4>] vsnprintf+0x374/0x1c50
[ 179.010992] [<ffffffff81b61e50>] ? pointer+0xa80/0xa80
[ 179.010996] [<ffffffff8156b676>] ? save_stack+0x46/0xd0
[ 179.011001] [<ffffffff81566faa>] ? __kmalloc+0x14a/0x2a0
[ 179.011006] [<ffffffff81b3d70a>] ? kobject_get_path+0x9a/0x200
[ 179.011010] [<ffffffff81b408c2>] ? kobject_uevent_env+0x282/0xca0
[ 179.011014] [<ffffffff81b412eb>] ? kobject_uevent+0xb/0x10
[ 179.011020] [<ffffffff81f10104>] ? device_del+0x434/0x6d0
[ 179.011029] [<ffffffffa0fea717>] ? rc_unregister_device+0x177/0x240 [rc_core]
[ 179.011037] [<ffffffffa116eeb0>] ? au0828_rc_unregister+0x60/0xb0 [au0828]
The problem is fixed with this patch on Linux 4.9-rc4
thanks,
-- Shuah
Powered by blists - more mailing lists