From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753420Ab3KSXxH (ORCPT ); Tue, 19 Nov 2013 18:53:07 -0500 Received: from mailout3.w2.samsung.com ([211.189.100.13]:26063 "EHLO usmailout3.samsung.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752228Ab3KSXxD (ORCPT ); Tue, 19 Nov 2013 18:53:03 -0500 X-AuditID: cbfec37b-b7fa96d00000716e-4a-528bf9deef36 Message-id: <528BF9DB.6050308@samsung.com> Date: Tue, 19 Nov 2013 16:52:59 -0700 From: Shuah Khan Reply-to: shuah.kh@samsung.com User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.1.0 MIME-version: 1.0 To: "Rafael J. Wysocki" , Greg KH Cc: len.brown@intel.com, pavel@ucw.cz, anton@enomsg.org, dwmw2@infradead.org, linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org, shuahkhan@gmail.com, Shuah Khan Subject: Re: [PATCH 2/2] power: Change device_wakeup_enable() to WARN_ON on null dev_name(dev) References: <622a097f9780dc721354d0e2b6cb512280da1ebc.1384875691.git.shuah.kh@samsung.com> <20131119231401.GC24054@kroah.com> <4885530.zkZVX2zzgu@vostro.rjw.lan> In-reply-to: <4885530.zkZVX2zzgu@vostro.rjw.lan> Content-type: text/plain; charset=UTF-8; format=flowed Content-transfer-encoding: 7bit X-Originating-IP: [105.144.21.63] X-Brightmail-Tracker: H4sIAAAAAAAAA+NgFmpjkeLIzCtJLcpLzFFi42I5/e+wn+69n91BBt0XuC0ObtW0mLhyMrNF 8+L1bBazpuxlsri8aw6bxefeI4wWd08dZbM4c/oSq8XXnw4OnB4T+j8xeuycdZfdY/MKLY/F e14yeeyfu4bdY8vVdhaPFau/s3t83iQXwBHFZZOSmpNZllqkb5fAlXF/4nT2gpNmFVNmTmBs YHyp28XIySEhYCLx5u5/NghbTOLCvfVANheHkMAyRoltDZPYIZxeJomFW9cwQTibGCXunLzC DtLCK6AlsfrVaUYQm0VAVWLazmlgcTYBdYnPr3eA2UICchJNS1Yzg9iiAhESR1c/Y4XoFZT4 MfkeC4gtIhAscet9P9gCZoEDjBJfjy0ESwgLJEi8aLvNBjHoIqPEweceIDangIHEj0+dYHFm ATOJRy3rmCFseYnNa94yQ9QrS/y5fApoKAfQb4oS21dbTWAUmYVk9Swk3bOQdC9gZF7FKFZa nFxQnJSeWmGkV5yYW1yal66XnJ+7iRESd9U7GO9+tTnEKMDBqMTDK7GgO0iINbGsuDL3EKME B7OSCO+aF0Ah3pTEyqrUovz4otKc1OJDjEwcnFINjHZvSmX37r30Urf169YdLeJKrpezbhfe rZz8OVsu9Pb+5RU7Lt28yFZ+1MD82+vvoTevP56g8snlwaFd7RxnNI2l3e0+vX89p29q2v6L f9MbN6ZfX2xTzHMiWTPDu+Dv4y2HgtzDMpYlv3ogsUxzy7Sc9CfF7euy6t+X7nDSdlnzL0nL iumS50wlluKMREMt5qLiRABhnGTPmQIAAA== Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 11/19/2013 04:56 PM, Rafael J. Wysocki wrote: > On Tuesday, November 19, 2013 03:14:01 PM Greg KH wrote: >> On Tue, Nov 19, 2013 at 09:05:46AM -0700, Shuah Khan wrote: >>> device_wakeup_enable() uses dev_name(dev) as the wakeup source name. >>> When it gets called with a device with its name not yet set, ws structure >>> with ws->name = NULL gets created. >>> >>> When kernel is booted with wakeup_source_activate enabled, it will panic >>> when the trace point code tries to dereferences ws->name. >>> >>> Change device_wakeup_enable() to WARN_ON on dev_name(dev) null condition >>> to detect driver bugs that result in calling device_wakeup_enable() before >>> device is fully initialized with its name in device_wakeup_enable(). >>> >>> This change without the power_supply_register() fix will result in early >>> boot panics when AC Adapter and Battery device drivers try to register >>> wakeup source. >>> >>> The following panic resulted from power_supply_register() registering >>> wakeup source with a null device name. >>> >>> [ 819.769934] device: 'BAT1': device_add >>> [ 819.770078] PM: Adding info for No Bus:BAT1 >>> [ 819.770235] BUG: unable to handle kernel NULL pointer dereference at (null) >>> [ 819.770435] IP: [] skip_spaces+0x30/0x30 >>> [ 819.770572] PGD 3efd90067 PUD 3eff61067 PMD 0 >>> [ 819.770716] Oops: 0000 [#1] SMP >>> [ 819.770829] Modules linked in: arc4 iwldvm mac80211 x86_pkg_temp_thermal coretemp kvm_intel joydev i915 kvm uvcvideo ghash_clmulni_intel videobuf2_vmalloc aesni_intel videobuf2_memops videobuf2_core aes_x86_64 ablk_helper cryptd videodev iwlwifi lrw rfcomm gf128mul glue_helper bnep btusb media bluetooth parport_pc hid_generic ppdev snd_hda_codec_hdmi drm_kms_helper snd_hda_codec_realtek cfg80211 drm tpm_infineon samsung_laptop snd_hda_intel usbhid snd_hda_codec hid snd_hwdep snd_pcm microcode snd_page_alloc snd_timer psmouse i2c_algo_bit lpc_ich tpm_tis video wmi mac_hid serio_raw ext2 lp parport r8169 mii >>> [ 819.771802] CPU: 0 PID: 2167 Comm: bash Not tainted 3.12.0+ #25 >>> [ 819.771876] Hardware name: SAMSUNG ELECTRONICS CO., LTD. 900X3C/900X3D/900X4C/900X4D/SAMSUNG_NP1234567890, BIOS P03AAC 07/12/2012 >>> [ 819.772022] task: ffff88002e6ddcc0 ti: ffff8804015ca000 task.ti: ffff8804015ca000 >>> [ 819.772119] RIP: 0010:[] [] skip_spaces+0x30/0x30 >>> [ 819.772242] RSP: 0018:ffff8804015cbc70 EFLAGS: 00010046 >>> [ 819.772310] RAX: 0000000000000003 RBX: ffff88040cfd6d40 RCX: 0000000000000018 >>> [ 819.772397] RDX: 0000000000020001 RSI: 0000000000000000 RDI: 0000000000000000 >>> [ 819.772484] RBP: ffff8804015cbcc0 R08: 0000000000000000 R09: ffff8803f0768d40 >>> [ 819.772570] R10: ffffea001033b800 R11: 0000000000000000 R12: ffffffff81c519c0 >>> [ 819.772656] R13: 0000000000020001 R14: 0000000000000000 R15: 0000000000020001 >>> [ 819.772744] FS: 00007ff98309b740(0000) GS:ffff88041f200000(0000) knlGS:0000000000000000 >>> [ 819.772845] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 >>> [ 819.772917] CR2: 0000000000000000 CR3: 00000003f59dc000 CR4: 00000000001407f0 >>> [ 819.773001] Stack: >>> [ 819.773030] ffffffff81114003 ffff8804015cbcb0 0000000000000000 0000000000000046 >>> [ 819.773146] ffff880409757a18 ffff8803f065a160 0000000000000000 0000000000020001 >>> [ 819.773273] 0000000000000000 0000000000000000 ffff8804015cbce8 ffffffff8143e388 >>> [ 819.773387] Call Trace: >>> [ 819.773434] [] ? ftrace_raw_event_wakeup_source+0x43/0xe0 >>> [ 819.773520] [] wakeup_source_report_event+0xb8/0xd0 >>> [ 819.773595] [] __pm_stay_awake+0x2d/0x50 >>> [ 819.773724] [] power_supply_changed+0x3c/0x90 >>> [ 819.773795] [] power_supply_register+0x18c/0x250 >>> [ 819.773869] [] sysfs_add_battery+0x61/0x7b >>> [ 819.773935] [] battery_notify+0x37/0x3f >>> [ 819.774001] [] notifier_call_chain+0x4c/0x70 >>> [ 819.774071] [] __blocking_notifier_call_chain+0x4d/0x70 >>> [ 819.774149] [] blocking_notifier_call_chain+0x16/0x20 >>> [ 819.774227] [] pm_notifier_call_chain+0x1a/0x40 >>> [ 819.774316] [] hibernate+0x66/0x1c0 >>> [ 819.774407] [] state_store+0x71/0xa0 >>> [ 819.774507] [] kobj_attr_store+0xf/0x20 >>> [ 819.774613] [] sysfs_write_file+0x128/0x1c0 >>> [ 819.774735] [] vfs_write+0xbd/0x1e0 >>> [ 819.774841] [] SyS_write+0x49/0xa0 >>> [ 819.774939] [] system_call_fastpath+0x16/0x1b >>> [ 819.775055] Code: 89 f8 48 89 e5 f6 82 c0 a6 84 81 20 74 15 0f 1f 44 00 00 48 83 c0 01 0f b6 10 f6 82 c0 a6 84 81 20 75 f0 5d c3 66 0f 1f 44 00 00 <80> 3f 00 55 48 89 e5 74 15 48 89 f8 0f 1f 40 00 48 83 c0 01 80 >>> [ 819.775760] RIP [] skip_spaces+0x30/0x30 >>> [ 819.775881] RSP >>> [ 819.775949] CR2: 0000000000000000 >>> [ 819.794175] ---[ end trace c4ef25127039952e ]--- >>> >>> Signed-off-by: Shuah Khan >>> --- >>> drivers/base/power/wakeup.c | 3 +++ >>> 1 file changed, 3 insertions(+) >>> >>> diff --git a/drivers/base/power/wakeup.c b/drivers/base/power/wakeup.c >>> index 2d56f41..a605f0d 100644 >>> --- a/drivers/base/power/wakeup.c >>> +++ b/drivers/base/power/wakeup.c >>> @@ -223,6 +223,9 @@ int device_wakeup_enable(struct device *dev) >>> if (!dev || !dev->power.can_wakeup) >>> return -EINVAL; >>> >>> + if (WARN_ON(!dev_name(dev))) >>> + return -EINVAL; >>> + >>> ws = wakeup_source_register(dev_name(dev)); >> >> Shouldn't wakeup_source_register() just handle a NULL for a name better? > > In fact, it does. > > The bug is in wakeup_source_activate() and that's because it passes ws->name > to trace_wakeup_source_activate() which then doesn't bother to check it against > NULL. > > So the solution here is to check ws->name before attempting to pass > it to trace_wakeup_source_activate() in wakeup_source_activate(). It is > actually valid to have a wakeup source with a NULL name and it shouldn't > blow up like this. > > Thanks, > Rafael > Rafael beat me to it. Yes adding the following to DECLARE_EVENT_CLASS(wakeup_source, should take care of the problem: const char *tmp_i = name ? name : "ws no name"; __assign_str(name, tmp_i); should make trace_wakeup_source_activate() not blow up. I did consider this as one of the fixes to the oops. Would you rather see this fix instead of this change to device_wakeup_enable()? -- Shuah -- Shuah Khan Senior Linux Kernel Developer - Open Source Group Samsung Research America(Silicon Valley) shuah.kh@samsung.com | (970) 672-0658