Bug 772174

Summary:	kernel 3.2 hibernation and ext4 fs write problem
Product:	[Fedora] Fedora	Reporter:	Jacek Pawlyta <cunio>
Component:	kernel	Assignee:	Eric Sandeen <esandeen>
Status:	CLOSED DUPLICATE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	high	Docs Contact:
Priority:	unspecified
Version:	rawhide	CC:	gansalmon, itamar, jonathan, kernel-maint, madhu.chinakonda
Target Milestone:	---
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2012-02-17 12:36:27 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Jacek Pawlyta 2012-01-06 09:54:37 UTC

Description of problem:
after suspend to disk (hibernation) ext4 fs is inaccessible to write

Version-Release number of selected component (if applicable):
kernel 3.2.0-2.fc17.x86_64

How reproducible:
Always

Steps to Reproduce:
1. hibernate
2. resume
3. try do write something to disk
  
Actual results:
for some programs fs is inaccessible to write 

Expected results:
everything is as before hibernation

Additional info:

37787.820619] ------------[ cut here ]------------
[37787.820630] WARNING: at fs/inode.c:884 unlock_new_inode+0x63/0x70()
[37787.820634] Hardware name: 20023               
[37787.820635] Modules linked in: wl(P) michael_mic arc4 lp parport hidp fuse lockd rfcomm bnep tpm_bios ip6t_REJECT nf_conntrack_ipv6 nf_defrag_ipv6 ip6table_filter ip6_tables coretemp nf_conntrack_netbios_ns nf_conntrack_broadcast nf_conntrack_ipv4 nf_defrag_ipv4 xt_state nf_conntrack lib80211_crypt_tkip uvcvideo videodev media v4l2_compat_ioctl32 btusb bluetooth i2c_i801 iTCO_wdt ideapad_laptop iTCO_vendor_support lib80211 snd_hda_codec_conexant snd_hda_intel snd_hda_codec snd_hwdep snd_seq snd_seq_device snd_pcm snd_timer snd soundcore snd_page_alloc microcode sparse_keymap rfkill tg3 uinput binfmt_misc sunrpc i915 drm_kms_helper drm i2c_algo_bit i2c_core video [last unloaded: wl]
[37787.820690] Pid: 5112, comm: packagekitd Tainted: P        W  O 3.2.0-2.fc17.x86_64 #1
[37787.820692] Call Trace:
[37787.820698]  [<ffffffff8106de2f>] warn_slowpath_common+0x7f/0xc0
[37787.820702]  [<ffffffff8106de8a>] warn_slowpath_null+0x1a/0x20
[37787.820705]  [<ffffffff81191463>] unlock_new_inode+0x63/0x70
[37787.820709]  [<ffffffff811f568b>] ext4_new_inode+0xabb/0x1110
[37787.820714]  [<ffffffff81202382>] ext4_create+0x92/0x160
[37787.820718]  [<ffffffff81184024>] vfs_create+0xb4/0x110
[37787.820721]  [<ffffffff8118768c>] do_last+0x61c/0x900
[37787.820724]  [<ffffffff81187a7e>] path_openat+0xce/0x3e0
[37787.820727]  [<ffffffff81187eb1>] do_filp_open+0x41/0xa0
[37787.820731]  [<ffffffff81193c9d>] ? alloc_fd+0x4d/0x140
[37787.820735]  [<ffffffff8117797c>] do_sys_open+0xec/0x1c0
[37787.820738]  [<ffffffff81177a70>] sys_open+0x20/0x30
[37787.820743]  [<ffffffff815ed2c2>] system_call_fastpath+0x16/0x1b
[37787.820745] ---[ end trace d07728137ab99ae5 ]---

Comment 1 Josh Boyer 2012-01-06 12:45:49 UTC

Eric, this is clearly the unlock_new_inode bug but it's on the final 3.2 release build in rawhide.  I'll leave it to you to dupe the bug if you want in case 3.2 makes this somehow more compelling/easier to debug.

(I'm intentionally ignoring the P taint here)

Comment 2 Jacek Pawlyta 2012-01-06 14:06:02 UTC

Josh, the P is here because I'm not able to use brcmsmac driver due to the bug:
#771477

Comment 3 Jacek Pawlyta 2012-01-13 16:21:15 UTC

After some research I have found that it has something to do with sandbox,

I switched off the sandbox as suggested here:
https://bugzilla.redhat.com/show_bug.cgi?id=742708 

rebooted and it seems for now that the hibernation works without corruption of the ext4 fs.

Comment 4 Eric Sandeen 2012-01-13 16:39:25 UTC

Funky, but good to know; ok, let me try to test that particular setup.

Comment 5 Jacek Pawlyta 2012-01-13 21:08:49 UTC

kernel 3.2.1-2 + sandbox off + resume from hibernation =
 
[  842.202641] EXT4-fs error (device dm-1) in ext4_new_inode:941: IO failure
[  842.378472] EXT4-fs error (device dm-1) in ext4_new_inode:941: IO failure
[  851.928284] EXT4-fs error (device dm-1) in ext4_new_inode:941: IO failure

Comment 6 Jacek Pawlyta 2012-01-13 21:09:57 UTC

kernel 3.2.0-2 + sandbox off = OK

Comment 7 Eric Sandeen 2012-01-13 21:10:56 UTC

The IO failure message is due to new & improved error handling - at least now it doesn't WARN_ON and backtrace :)

Comment 8 Eric Sandeen 2012-01-13 22:58:28 UTC

I've been unable thus far to repro even w/ sandbox turned on.  Argh!

Comment 9 Jacek Pawlyta 2012-02-17 12:30:23 UTC

kernel 3.2.6-4.fc16.x86_64 + sandbox off + hibernation =

EXT4-fs error (device dm-1) in ext4_new_inode:941: IO failure

The same result with sandbox on :(

The same results with previous kernels 3.2

My system: http://www.smolts.org/client/show/pub_503fa95d-8ef3-46f5-b1ae-014b47bd254d

Comment 10 Jacek Pawlyta 2012-02-17 12:36:27 UTC


*** This bug has been marked as a duplicate of bug 744275 ***