From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756817Ab1JRATV (ORCPT ); Mon, 17 Oct 2011 20:19:21 -0400 Received: from mail-qw0-f46.google.com ([209.85.216.46]:38683 "EHLO mail-qw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751017Ab1JRATU (ORCPT ); Mon, 17 Oct 2011 20:19:20 -0400 Date: Tue, 18 Oct 2011 02:19:09 +0200 From: Frederik Deweerdt To: "Tomas M." Cc: Linux Kernel Mailing List , laurent.pinchart@ideasonboard.com, hverkuil@xs4all.nl, mchehab@redhat.com Subject: [patch] Avoid NULL deref in v4l2_device_release (was Re: kernel OOPS when releasing usb webcam (random)) Message-ID: <20111018001909.GA29706@gambetta> References: <4E9CB0C2.3030902@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4E9CB0C2.3030902@gmail.com> User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org [Adding relevant people to CCs] Hi Tomas, On Mon, Oct 17, 2011 at 07:48:34PM -0300, Tomas M. wrote: > im getting the following null pointer dereference from time to time > when releasing a usb camera. > [...] > BUG: unable to handle kernel NULL pointer dereference at 0000006c > IP: [] v4l2_device_release+0xa2/0xf0 [videodev] > *pde = 00000000 > Oops: 0000 [#1] PREEMPT SMP > Modules linked in: fuse arc4 rt73usb rt2x00usb rt2x00lib mac80211 > cfg80211 rfkill gspca_zc3xx gspca_main videodev joydev > snd_hda_codec_si3054 sg 8139too snd_hda_codec_realtek firewire_ohci > firewire_core mmc_core snd_hda_intel snd_hda_codec snd_hwdep snd_pcm > snd_timer snd soundcore mii crc_itu_t snd_page_alloc iTCO_wdt > iTCO_vendor_support i2c_i801 evdev psmouse thermal battery serio_raw > ac cpufreq_ondemand acpi_cpufreq freq_table processor mperf usbhid > hid ext3 jbd mbcache sd_mod sr_mod cdrom pata_acpi uhci_hcd ata_piix > ehci_hcd libata scsi_mod usbcore [last unloaded: sdhci] > > Pid: 171, comm: khubd Not tainted 3.1.0-rc9 #66 Everex Systems, Inc. > Everex StepNote Series/Everex StepNote Series > EIP: 0060:[] EFLAGS: 00010292 CPU: 0 > EIP is at v4l2_device_release+0xa2/0xf0 [videodev] > EAX: 00000000 EBX: f5636004 ECX: 00000000 EDX: 00000000 > ESI: f5636000 EDI: 00000000 EBP: f563600c ESP: f5627e38 > DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 > Process khubd (pid: 171, ti=f5626000 task=f554dc00 task.ti=f5626000) > Stack: > ef000480 c1433780 f5474b00 c12343f8 f54e7e1c 00000000 c114737a f563600c > f5636028 c114605d f5636028 c1146020 f91512d4 00000000 c114737a f54e7e1c > f54e7e00 f81623f4 f56d4000 f54e7e1c f91512d4 f56d4064 00000001 c12373b7 > Call Trace: > [] ? device_release+0x18/0x80 > [] ? kref_put+0x2a/0x60 > [] ? kobject_release+0x3d/0xa0 > [] ? kobject_del+0x30/0x30 > [] ? kref_put+0x2a/0x60 > [] ? usb_unbind_interface+0x34/0x130 [usbcore] > [] ? __device_release_driver+0x57/0xb0 > [] ? device_release_driver+0x1d/0x30 > [] ? bus_remove_device+0x72/0x90 > [] ? device_del+0xdf/0x150 > [] ? usb_disable_device+0x81/0x180 [usbcore] > [] ? usb_disconnect+0x8b/0x110 [usbcore] > [] ? hub_thread+0x97c/0x1180 [usbcore] > [] ? pick_next_task_fair+0x8b/0xe0 > [] ? abort_exclusive_wait+0x90/0x90 > [] ? usb_remote_wakeup+0x40/0x40 [usbcore] > [] ? kthread+0x69/0x70 > [] ? kthread_worker_fn+0x150/0x150 > [] ? kernel_thread_helper+0x6/0xd > Code: 83 94 01 00 00 c7 83 60 01 00 00 00 00 00 00 0f b7 93 9c 01 00 > 00 c1 e0 05 f0 0f b3 90 c0 e7 0c f9 b8 20 e1 0c f9 e8 4e cf 24 c8 > <8b> 57 6c 89 f0 85 d2 74 25 ff 93 c8 01 00 00 85 ff 74 21 89 f8 > EIP: [] v4l2_device_release+0xa2/0xf0 [videodev] SS:ESP > 0068:f5627e38 > CR2: 000000000000006c > ---[ end trace 39522f0f1757c8f8 ]--- The trace hints at a v4l2 being NULL in a newly introduced v4l2_dev->release check. Attached patch below. Regards, Frederik [media] v4l: Avoid NULL pointer dereference 8280b662df96f4172c4972b14a4aec0daf272b8f introduced a potential NULL deref in the case v4l2_dev is NULL. Reported-by: Tomas M. Signed-off-by: Frederik Deweerdt diff --git a/drivers/media/video/v4l2-dev.c b/drivers/media/video/v4l2-dev.c index d721565..5c0fa64 100644 --- a/drivers/media/video/v4l2-dev.c +++ b/drivers/media/video/v4l2-dev.c @@ -181,7 +181,7 @@ static void v4l2_device_release(struct device *cd) * TODO: In the long run all drivers that use v4l2_device should use the * v4l2_device release callback. This check will then be unnecessary. */ - if (v4l2_dev->release == NULL) + if (v4l2_dev != NULL && v4l2_dev->release == NULL) v4l2_dev = NULL; /* Release video_device and perform other