Skip to contentSkip to conversation

[2.6.26] kobject_add_internal failed for 2:0 with -EEXIST / unable to handle kernel NULL pointer dereference in sysfs_create_link

DISCUSSION Folkert van Heusden 4 participants 17 messages Oct 28, 2008 – Nov 3, 2008

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e
[   42.509823]  [<ffffffff8031b46e>] kobject_add+0x74/0x7c
[   42.509969]  [<ffffffff802e2470>] sysfs_addrm_finish+0x19/0x1ea
[   42.510141]  [<ffffffff802e21b4>] sysfs_find_dirent+0x1b/0x2f
[   42.510331]  [<ffffffff802e2741>] create_dir+0x5a/0x87
[   42.510466]  [<ffffffff8031ae88>] kobject_get+0x12/0x17
[   42.510614]  [<ffffffff80382771>] get_device+0x17/0x20
[   42.510754]  [<ffffffff80382d81>] device_add+0x9b/0x53f
[   42.510915]  [<ffffffff8031acf2>] kobject_init+0x41/0x69
[   42.511374]  [<ffffffff803832d1>] device_create_vargs+0x9a/0xc6
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4
[   42.511657]  [<ffffffff8030ac34>] elv_register_queue+0x67/0x6f
[   42.511818]  [<ffffffff8030e54e>] blk_register_queue+0x77/0x9b
[   42.511818]  [<ffffffff80311ff4>] add_disk+0x64/0x87
[   42.511818]  [<ffffffffa0071f04>] :floppy:floppy_module_init+0xdf3/0xea8
[   42.511818]  [<ffffffff8022c184>] try_to_wake_up+0x118/0x129
[   42.511840]  [<ffffffff80254e9b>] sys_init_module+0x190e/0x1aa4
[   42.511992]  [<ffffffff8030cc77>] blk_init_queue+0x0/0x8
[   42.512148]  [<ffffffff8020be9a>] system_call_after_swapgs+0x8a/0x8f
[   42.512290]
[   42.512410] BUG: unable to handle kernel NULL pointer dereference at 0000000000000128
[   42.512703] IP: [<ffffffff802e2f17>] sysfs_create_link+0x44/0x105
[   42.512889] PGD 1bdc12067 PUD 1ba9d5067 PMD 0
[   42.513198] Oops: 0000 [1] SMP
[   42.513422] CPU 2
[   42.513576] Modules linked in: floppy(+) output ide_cd_mod serio_raw dm_snapshot ata_generic snd_pcm snd_timer i2c_piix4 dm_mirror ehci_hcd battery usbhid ff_memless pcspkr(-) fan thermal_sys libata loop i2c_core joydev dm_log hid cdrom snd_page_alloc evdev netconsole configfs ipv6 snd soundcore ext3 jbd mbcache dm_mod dock enclosure sd_mod serverworks aacraid scsi_mod tg3 ide_pci_generic ide_core [last unloaded: psmouse]
[   42.519374] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.519374] RIP: 0010:[<ffffffff802e2f17>]  [<ffffffff802e2f17>] sysfs_create_link+0x44/0x105
[   42.519374] RSP: 0018:ffff8101ba991d48  EFLAGS: 00010246
[   42.519374] RAX: 0000000000009292 RBX: 00000000000000f0 RCX: ffffffff804fe088
[   42.519374] RDX: ffffffff804b341d RSI: 00000000000000f0 RDI: ffffffff80653d80
[   42.519374] RBP: ffffffff804b341d R08: 0000000000000000 R09: ffff8101bb44d000
[   42.519374] R10: 0000000000000001 R11: 0000000000000046 R12: 00000000fffffff2
[   42.519374] R13: ffff8101bac6c5f0 R14: 0000000000000008 R15: 0000000000000000
[   42.519374] FS:  00007f97800126e0(0000) GS:ffff8101bf0ad0c0(0000) knlGS:0000000000000000
[   42.519374] CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
[   42.519374] CR2: 0000000000000128 CR3: 00000001a9c89000 CR4: 00000000000006e0
[   42.519374] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   42.519374] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
[   42.519374] Process modprobe (pid: 5301, threadinfo ffff8101ba990000, task ffff8101bf20e3c0)
[   42.519374] Stack:  0000000000000000 ffffffff8030e54e ffff8101bc0d9c00 0000000000000000
[   42.519374]  ffff8101bc0d9d58 0000000000000000 0000000000000000 ffffffffa01fee30
[   42.522582]  ffffffffa01fee20 ffffffffa0071f04 ffffffff8022c184 0000000000000282
[   42.522582] Call Trace:
[   42.522582]  [<ffffffff8030e54e>] ? blk_register_queue+0x77/0x9b
[   42.522582]  [<ffffffffa0071f04>] ? :floppy:floppy_module_init+0xdf3/0xea8
[   42.522582]  [<ffffffff8022c184>] ? try_to_wake_up+0x118/0x129
[   42.522582]  [<ffffffff80254e9b>] ? sys_init_module+0x190e/0x1aa4
[   42.522582]  [<ffffffff8030cc77>] ? blk_init_queue+0x0/0x8
[   42.522582]  [<ffffffff8020be9a>] ? system_call_after_swapgs+0x8a/0x8f
[   42.522582]
[   42.522582]
[   42.522582] Code: 48 85 ff 49 c7 c5 a0 ba 50 80 74 13 4c 8b 6f 38 41 bc f2 ff ff ff 4d 85 ed 0f 84 bc 00 00 00 48 c7 c7 80 3d 65 80 e8 fb 68 14 00 <48> 8b 5b 38 48 85 db 74 19 83 3b 00 75 11 be 81 00 00 00 48 c7
[   42.526571] RIP  [<ffffffff802e2f17>] sysfs_create_link+0x44/0x105
[   42.526571]  RSP <ffff8101ba991d48>
[   42.526571] CR2: 0000000000000128
[   42.531559] ---[ end trace 4eb65a6452398ce5 ]---

Folkert van Heusden

Signature · 6 lines
Multitail est un outil permettant la visualisation de fichiers de
journalisation et/ou le suivi de l'exécution de commandes. Filtrage,
mise en couleur de mot-clé, fusions, visualisation de différences
(diff-view), etc.  http://www.vanheusden.com/multitail/
----------------------------------------------------------------------
Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com
#1 lore raw
Quoted 5 lines · Folkert van Heusden

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

Quoted 16 lines
[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e
[   42.509823]  [<ffffffff8031b46e>] kobject_add+0x74/0x7c
[   42.509969]  [<ffffffff802e2470>] sysfs_addrm_finish+0x19/0x1ea
[   42.510141]  [<ffffffff802e21b4>] sysfs_find_dirent+0x1b/0x2f
[   42.510331]  [<ffffffff802e2741>] create_dir+0x5a/0x87
[   42.510466]  [<ffffffff8031ae88>] kobject_get+0x12/0x17
[   42.510614]  [<ffffffff80382771>] get_device+0x17/0x20
[   42.510754]  [<ffffffff80382d81>] device_add+0x9b/0x53f
[   42.510915]  [<ffffffff8031acf2>] kobject_init+0x41/0x69
[   42.511374]  [<ffffffff803832d1>] device_create_vargs+0x9a/0xc6
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry.

What does: ls -l "/sys/class/bdi/" print?

How many floppies (or emulated floppies) does this system have?

Kay

#2 lore raw
Quoted 22 lines
Quoted 5 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

Quoted 16 lines
[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e
[   42.509823]  [<ffffffff8031b46e>] kobject_add+0x74/0x7c
[   42.509969]  [<ffffffff802e2470>] sysfs_addrm_finish+0x19/0x1ea
[   42.510141]  [<ffffffff802e21b4>] sysfs_find_dirent+0x1b/0x2f
[   42.510331]  [<ffffffff802e2741>] create_dir+0x5a/0x87
[   42.510466]  [<ffffffff8031ae88>] kobject_get+0x12/0x17
[   42.510614]  [<ffffffff80382771>] get_device+0x17/0x20
[   42.510754]  [<ffffffff80382d81>] device_add+0x9b/0x53f
[   42.510915]  [<ffffffff8031acf2>] kobject_init+0x41/0x69
[   42.511374]  [<ffffffff803832d1>] device_create_vargs+0x9a/0xc6
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4
Quoted 3 lines

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry.
What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
total 0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:1
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:10
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:11
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:12
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:13
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:14
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:15
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:2
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:3
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:4
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:5
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:6
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:7
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:8
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:9
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:1
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:2
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:3
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:4
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:5
drwxr-xr-x 3 root root 0 2008-10-28 18:32 3:0
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:0
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:1
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:2
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:3
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:4
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:5
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:6
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:7
drwxr-xr-x 3 root root 0 2008-10-28 18:32 8:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 default
Quoted 1 line

How many floppies (or emulated floppies) does this system have?

1 physical.

Folkert van Heusden

Signature · 4 lines
Ever wonder what is out there? Any alien races? Then please support
the seti@home project: setiathome.ssl.berkeley.edu
----------------------------------------------------------------------
Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com
#3 lore raw
Quoted 65 lines · Folkert van Heusden
Quoted 22 lines
Quoted 5 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

Quoted 16 lines
[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e
[   42.509823]  [<ffffffff8031b46e>] kobject_add+0x74/0x7c
[   42.509969]  [<ffffffff802e2470>] sysfs_addrm_finish+0x19/0x1ea
[   42.510141]  [<ffffffff802e21b4>] sysfs_find_dirent+0x1b/0x2f
[   42.510331]  [<ffffffff802e2741>] create_dir+0x5a/0x87
[   42.510466]  [<ffffffff8031ae88>] kobject_get+0x12/0x17
[   42.510614]  [<ffffffff80382771>] get_device+0x17/0x20
[   42.510754]  [<ffffffff80382d81>] device_add+0x9b/0x53f
[   42.510915]  [<ffffffff8031acf2>] kobject_init+0x41/0x69
[   42.511374]  [<ffffffff803832d1>] device_create_vargs+0x9a/0xc6
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4
Quoted 3 lines

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry.
What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
total 0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:1
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:10
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:11
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:12
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:13
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:14
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:15
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:2
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:3
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:4
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:5
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:6
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:7
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:8
drwxr-xr-x 3 root root 0 2008-10-28 18:32 1:9
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:1
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:2
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:3
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:4
drwxr-xr-x 3 root root 0 2008-10-28 18:32 254:5
drwxr-xr-x 3 root root 0 2008-10-28 18:32 3:0
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:0
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:1
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:2
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:3
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:4
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:5
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:6
drwxr-xr-x 3 root root 0 2008-10-29 11:39 7:7
drwxr-xr-x 3 root root 0 2008-10-28 18:32 8:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 default
Quoted 5 lines
Quoted 1 line

How many floppies (or emulated floppies) does this system have?

1 physical.

Hi guys,

I found similar issue on Blackfin board and I believe it is a common bug for bdi
http://lkml.org/lkml/2008/10/16/126

But there is no response about this bug, although I am working on it.

-Bryan

#5 lore raw
In reply to Folkert van Heusden
Quoted 26 lines · Folkert van Heusden
Quoted 22 lines
Quoted 5 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

Quoted 16 lines
[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e
[   42.509823]  [<ffffffff8031b46e>] kobject_add+0x74/0x7c
[   42.509969]  [<ffffffff802e2470>] sysfs_addrm_finish+0x19/0x1ea
[   42.510141]  [<ffffffff802e21b4>] sysfs_find_dirent+0x1b/0x2f
[   42.510331]  [<ffffffff802e2741>] create_dir+0x5a/0x87
[   42.510466]  [<ffffffff8031ae88>] kobject_get+0x12/0x17
[   42.510614]  [<ffffffff80382771>] get_device+0x17/0x20
[   42.510754]  [<ffffffff80382d81>] device_add+0x9b/0x53f
[   42.510915]  [<ffffffff8031acf2>] kobject_init+0x41/0x69
[   42.511374]  [<ffffffff803832d1>] device_create_vargs+0x9a/0xc6
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4
Quoted 3 lines

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry.
What does: ls -l "/sys/class/bdi/" print?

Quoted 6 lines

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
total 0

drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

Thanks,
Kay

#4 lore raw
In reply to Bryan Wu
Quoted 27 lines · Bryan Wu
Quoted 25 lines · Folkert van Heusden
Quoted 25 lines
Quoted 5 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

Quoted 16 lines
[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e
[   42.509823]  [<ffffffff8031b46e>] kobject_add+0x74/0x7c
[   42.509969]  [<ffffffff802e2470>] sysfs_addrm_finish+0x19/0x1ea
[   42.510141]  [<ffffffff802e21b4>] sysfs_find_dirent+0x1b/0x2f
[   42.510331]  [<ffffffff802e2741>] create_dir+0x5a/0x87
[   42.510466]  [<ffffffff8031ae88>] kobject_get+0x12/0x17
[   42.510614]  [<ffffffff80382771>] get_device+0x17/0x20
[   42.510754]  [<ffffffff80382d81>] device_add+0x9b/0x53f
[   42.510915]  [<ffffffff8031acf2>] kobject_init+0x41/0x69
[   42.511374]  [<ffffffff803832d1>] device_create_vargs+0x9a/0xc6
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry.
What does: ls -l "/sys/class/bdi/" print?

...

Quoted 2 lines
Quoted 2 lines

drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1

...

Quoted 4 lines
Quoted 4 lines

drwxr-xr-x 3 root root 0 2008-10-28 18:32 default

Quoted 1 line

How many floppies (or emulated floppies) does this system have?

Quoted 8 lines
Quoted 1 line

1 physical.

Hi guys,

I found similar issue on Blackfin board and I believe it is a common bug for bdi
http://lkml.org/lkml/2008/10/16/126

But there is no response about this bug, although I am working on it.

Peter, any idea? BDI seems to try to create duplicate devices.

Thanks,
Kay

#6 lore raw
In reply to Kay Sievers
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

Quoted 7 lines
Quoted 4 lines

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory

Folkert van Heusden

Signature · 6 lines
Multitail es una herramienta flexible que permite visualizar los "log
file" y seguir la ejecución de comandos. Permite filtrar, añadir
colores, combinar archivos, la visualización de diferencias (diff-
view), etc.  http://www.vanheusden.com/multitail/
----------------------------------------------------------------------
Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com
#7 lore raw
Quoted 22 lines · Folkert van Heusden
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
Quoted 6 lines
Quoted 2 lines

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory

Ah, sorry. Seems the bdi stuff never got to pass the usual parent device with the device registration, to let the bdi device show up at the right place in the device tree.

Let's see what current devices on your box have the major 2:
find /sys -name dev | xargs grep '^2:'

Thanks,
Kay

#8 lore raw
Quoted 29 lines
Quoted 22 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
Quoted 6 lines
Quoted 2 lines

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory
Quoted 6 lines

Ah, sorry. Seems the bdi stuff never got to pass the usual parent device with the device registration, to let the bdi device show up at the right place in the device tree.

Let's see what current devices on your box have the major 2:
find /sys -name dev | xargs grep '^2:'

/sys/block/fd0/dev:2:0
/sys/block/fd1/dev:2:1

As my script does modprobe/rmmod in parallel (4 processes) maybe it is a conflict of one process doing an modprobe of floppy while the other does an rmmod? Or both a modprobe?

Folkert van Heusden

Signature · 5 lines
Multi tail barnamaj mowahib li mora9abat attasjilat wa nataij awamir
al 7asoub. damj, talwin, mora9abat attarchi7 wa ila akhirih.
http://www.vanheusden.com/multitail/
----------------------------------------------------------------------
Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com
#9 lore raw
Quoted 36 lines · Folkert van Heusden
Quoted 29 lines
Quoted 22 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
Quoted 6 lines
Quoted 2 lines

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory
Quoted 6 lines

Ah, sorry. Seems the bdi stuff never got to pass the usual parent device with the device registration, to let the bdi device show up at the right place in the device tree.

Let's see what current devices on your box have the major 2:
find /sys -name dev | xargs grep '^2:'

Quoted 6 lines

/sys/block/fd0/dev:2:0
/sys/block/fd1/dev:2:1

As my script does modprobe/rmmod in parallel (4 processes) maybe it is a conflict of one process doing an modprobe of floppy while the other does an rmmod? Or both a modprobe?

Might be, yes. If you just bootup, and don't run your modprobe/rmmod script, does the box have 2 floppy devices in /sys too?

Kay

#10 lore raw
Quoted 39 lines
Quoted 39 lines
Quoted 29 lines
Quoted 22 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
Quoted 6 lines
Quoted 2 lines

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory
Quoted 6 lines

Ah, sorry. Seems the bdi stuff never got to pass the usual parent device with the device registration, to let the bdi device show up at the right place in the device tree.

Let's see what current devices on your box have the major 2:
find /sys -name dev | xargs grep '^2:'

/sys/block/fd0/dev:2:0
/sys/block/fd1/dev:2:1

Quoted 6 lines
Quoted 3 lines

As my script does modprobe/rmmod in parallel (4 processes) maybe it is a conflict of one process doing an modprobe of floppy while the other does an rmmod? Or both a modprobe?

Might be, yes. If you just bootup, and don't run your modprobe/rmmod script, does the box have 2 floppy devices in /sys too?

Yes it does. One physical drive.

Folkert van Heusden

Signature · 6 lines
Multitail est un outil permettant la visualisation de fichiers de
journalisation et/ou le suivi de l'exécution de commandes. Filtrage,
mise en couleur de mot-clé, fusions, visualisation de différences
(diff-view), etc.  http://www.vanheusden.com/multitail/
----------------------------------------------------------------------
Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com
#11 lore raw
Quoted 39 lines · Folkert van Heusden
Quoted 39 lines
Quoted 39 lines
Quoted 29 lines
Quoted 22 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
Quoted 6 lines
Quoted 2 lines

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory
Quoted 6 lines

Ah, sorry. Seems the bdi stuff never got to pass the usual parent device with the device registration, to let the bdi device show up at the right place in the device tree.

Let's see what current devices on your box have the major 2:
find /sys -name dev | xargs grep '^2:'

/sys/block/fd0/dev:2:0
/sys/block/fd1/dev:2:1

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

As my script does modprobe/rmmod in parallel (4 processes) maybe it is a conflict of one process doing an modprobe of floppy while the other does an rmmod? Or both a modprobe?

Might be, yes. If you just bootup, and don't run your modprobe/rmmod script, does the box have 2 floppy devices in /sys too?

Yes it does. One physical drive.

Seems that always happens with multiple floppies. I can reproduce it here with qemu. It seems not related to modprobing. Also mtd devices suffer from the same problem, as bug reports show.

It might be a bug in bdi. Looks like floppies share a single queue, the bdi structure lives in the queue. Now we register for every device a bdi device, but the queue is shared and the former recorded dev_t in the bdi structure is overwritten. At unregistering the bdi device, all earlier devices using the same queue are not removed.

Peter, please check, if something like this can happen?

Thanks,
Kay

#12 lore raw
Quoted 54 lines · Kay Sievers
Quoted 39 lines · Folkert van Heusden
Quoted 39 lines
Quoted 39 lines
Quoted 29 lines
Quoted 22 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
Quoted 6 lines
Quoted 2 lines

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory
Quoted 6 lines

Ah, sorry. Seems the bdi stuff never got to pass the usual parent device with the device registration, to let the bdi device show up at the right place in the device tree.

Let's see what current devices on your box have the major 2:
find /sys -name dev | xargs grep '^2:'

/sys/block/fd0/dev:2:0
/sys/block/fd1/dev:2:1

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

As my script does modprobe/rmmod in parallel (4 processes) maybe it is a conflict of one process doing an modprobe of floppy while the other does an rmmod? Or both a modprobe?

Might be, yes. If you just bootup, and don't run your modprobe/rmmod script, does the box have 2 floppy devices in /sys too?

Yes it does. One physical drive.

Seems that always happens with multiple floppies. I can reproduce it here with qemu. It seems not related to modprobing. Also mtd devices suffer from the same problem, as bug reports show.

Quoted 7 lines

It might be a bug in bdi. Looks like floppies share a single queue, the bdi structure lives in the queue. Now we register for every device a bdi device, but the queue is shared and the former recorded dev_t in the bdi structure is overwritten. At unregistering the bdi device, all earlier devices using the same queue are not removed.

Peter, please check, if something like this can happen?

Ok, I get annoyed by these sysfs bugs. :)

Peter, it looks like bdi does not work for devices which share a single queue. If I add:
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -184,6 +184,8 @@ int bdi_register(struct backing_dev_info *bdi,
struct device *parent,

                  goto exit;
          }

+ printk("XXXXXXX old bdidev is %p\n", bdi->dev);

  +       printk("XXXXXXX new bdidev is %p\n", dev);
          bdi->dev = dev;
          bdi_debug_register(bdi, dev_name(dev));

I get:

  $ modprobe floppy
  Floppy drive(s): fd0 is 1.44M, fd1 is 1.44M
  FDC 0 is a S82078B
  XXXXXXX old bdidev is 0000000000000000
  XXXXXXX new bdidev is ffff88001f20cd10
  XXXXXXX old bdidev is ffff88001f20cd10
  XXXXXXX new bdidev is ffff88001f20de30

which very much looks like bdi will not remove any earlier registered device, only the last one, right?

Thanks,
Kay

#13 lore raw
Quoted 88 lines · Kay Sievers
Quoted 54 lines · Kay Sievers
Quoted 39 lines · Folkert van Heusden
Quoted 39 lines
Quoted 39 lines
Quoted 29 lines
Quoted 22 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines

While running my http://vanheusden.com/pyk/ script (which randomly inserts and removes modules) I triggered the folllowing oops in a 2.6.26 kernel on an IBM xSeries 260. This oops (in fact no oops at all) did not get triggered in a 2.6.18 kernel on that system.

[   42.507375] FDC 0 is a National Semiconductor PC87306
[   42.509057] kobject_add_internal failed for 2:0 with -EEXIST, don't try to register things with the same name in the same directory.
[   42.509291] Pid: 5301, comm: modprobe Not tainted 2.6.26-1-amd64 #1
[   42.509431]
[   42.509433] Call Trace:
[   42.509685]  [<ffffffff8031b031>] kobject_add_internal+0x13f/0x17e

...

Quoted 10 lines
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
[   42.511519]  [<ffffffff8027d23b>] bdi_register+0x57/0xb4

Looks like bdi sees two devices with the same devnum, or didn't cleanup an old entry. What does: ls -l "/sys/class/bdi/" print?

The following:

folkert@debiantesthw:~$ ls -l /sys/class/bdi/
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:0
drwxr-xr-x 3 root root 0 2008-10-28 18:32 2:1
Quoted 6 lines
Quoted 2 lines

Oh, you are running the old sysfs layout without symlinks. Care to tell where the "device" link in these directories points to?

None exist:

folkert@debiantesthw:~$ ls -la /sys/class/bdi/*/device
ls: cannot access /sys/class/bdi/*/device: No such file or directory
Quoted 6 lines

Ah, sorry. Seems the bdi stuff never got to pass the usual parent device with the device registration, to let the bdi device show up at the right place in the device tree.

Let's see what current devices on your box have the major 2:
find /sys -name dev | xargs grep '^2:'

/sys/block/fd0/dev:2:0
/sys/block/fd1/dev:2:1

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

As my script does modprobe/rmmod in parallel (4 processes) maybe it is a conflict of one process doing an modprobe of floppy while the other does an rmmod? Or both a modprobe?

Might be, yes. If you just bootup, and don't run your modprobe/rmmod script, does the box have 2 floppy devices in /sys too?

Yes it does. One physical drive.

Seems that always happens with multiple floppies. I can reproduce it here with qemu. It seems not related to modprobing. Also mtd devices suffer from the same problem, as bug reports show.

Quoted 7 lines

It might be a bug in bdi. Looks like floppies share a single queue, the bdi structure lives in the queue. Now we register for every device a bdi device, but the queue is shared and the former recorded dev_t in the bdi structure is overwritten. At unregistering the bdi device, all earlier devices using the same queue are not removed.

Peter, please check, if something like this can happen?

Ok, I get annoyed by these sysfs bugs. :)

Peter, it looks like bdi does not work for devices which share a single queue. If I add:
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -184,6 +184,8 @@ int bdi_register(struct backing_dev_info *bdi,
struct device *parent,

                  goto exit;
          }

+ printk("XXXXXXX old bdidev is %p\n", bdi->dev);

  +       printk("XXXXXXX new bdidev is %p\n", dev);
          bdi->dev = dev;
          bdi_debug_register(bdi, dev_name(dev));

I get:

  $ modprobe floppy
  Floppy drive(s): fd0 is 1.44M, fd1 is 1.44M
  FDC 0 is a S82078B
  XXXXXXX old bdidev is 0000000000000000
  XXXXXXX new bdidev is ffff88001f20cd10
  XXXXXXX old bdidev is ffff88001f20cd10
  XXXXXXX new bdidev is ffff88001f20de30
Quoted 2 lines

which very much looks like bdi will not remove any earlier registered device, only the last one, right?

This fixes it for me.

Thanks,
Kay

From: Kay Sievers <[email protected]>
Subject: bdi: register sysfs bdi device only once per queue

Devices which share the same queue, like floppies and mtd devices, get registered multiple times in the bdi interface, but bdi accounts only the last registered device of the devices sharing one queue.

On remove, all earlier registered devices leak, stay around in sysfs, and cause "duplicate filename" errors if the devices are recreated.

This prevents the creation of multiple bdi interfaces per queue, and the bdi device will carry the dev_t name of the block device which is the first one registered, of the pool of devices using the same queue.

---

mm/backing-dev.c +3 -0
index f2e574d..e6676e5 100644
@@ -176,6 +176,9 @@ int bdi_register(struct backing_dev_info *bdi, struct device *parent,
 	int ret = 0;
 	struct device *dev;
 
+	if (bdi->dev)
+		goto exit;
+
 	va_start(args, fmt);
 	dev = device_create_vargs(bdi_class, parent, MKDEV(0, 0), bdi, fmt, args);
 	va_end(args);
#14 lore raw
In reply to Kay Sievers
Quoted 23 lines · Kay Sievers

Peter, it looks like bdi does not work for devices which share a single queue. If I add:
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -184,6 +184,8 @@ int bdi_register(struct backing_dev_info *bdi,
struct device *parent,

                  goto exit;
          }

+ printk("XXXXXXX old bdidev is %p\n", bdi->dev);

  +       printk("XXXXXXX new bdidev is %p\n", dev);
          bdi->dev = dev;
          bdi_debug_register(bdi, dev_name(dev));

I get:

  $ modprobe floppy
  Floppy drive(s): fd0 is 1.44M, fd1 is 1.44M
  FDC 0 is a S82078B
  XXXXXXX old bdidev is 0000000000000000
  XXXXXXX new bdidev is ffff88001f20cd10
  XXXXXXX old bdidev is ffff88001f20cd10
  XXXXXXX new bdidev is ffff88001f20de30
Quoted 2 lines

which very much looks like bdi will not remove any earlier registered device, only the last one, right?

Sharing a bdi is odd to begin with, let me poke at this a little.

#15 lore raw
Quoted 25 lines · Peter Zijlstra
Quoted 23 lines · Kay Sievers

Peter, it looks like bdi does not work for devices which share a single queue. If I add:
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -184,6 +184,8 @@ int bdi_register(struct backing_dev_info *bdi,
struct device *parent,

                  goto exit;
          }

+ printk("XXXXXXX old bdidev is %p\n", bdi->dev);

  +       printk("XXXXXXX new bdidev is %p\n", dev);
          bdi->dev = dev;
          bdi_debug_register(bdi, dev_name(dev));

I get:

  $ modprobe floppy
  Floppy drive(s): fd0 is 1.44M, fd1 is 1.44M
  FDC 0 is a S82078B
  XXXXXXX old bdidev is 0000000000000000
  XXXXXXX new bdidev is ffff88001f20cd10
  XXXXXXX old bdidev is ffff88001f20cd10
  XXXXXXX new bdidev is ffff88001f20de30
Quoted 4 lines
Quoted 2 lines

which very much looks like bdi will not remove any earlier registered device, only the last one, right?

Sharing a bdi is odd to begin with, let me poke at this a little.

Yeah, it's odd, but I'm not sure if you want to touch floppy.c. :) Any objection to this patch, until you possible fix it differently?
http://marc.info/?l=linux-kernel&m=122541569310798&w=4

Kay

#16 lore raw
Quoted 26 lines · Kay Sievers
Quoted 25 lines · Peter Zijlstra
Quoted 23 lines · Kay Sievers

Peter, it looks like bdi does not work for devices which share a single queue. If I add:
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -184,6 +184,8 @@ int bdi_register(struct backing_dev_info *bdi,
struct device *parent,

                  goto exit;
          }

+ printk("XXXXXXX old bdidev is %p\n", bdi->dev);

  +       printk("XXXXXXX new bdidev is %p\n", dev);
          bdi->dev = dev;
          bdi_debug_register(bdi, dev_name(dev));

I get:

  $ modprobe floppy
  Floppy drive(s): fd0 is 1.44M, fd1 is 1.44M
  FDC 0 is a S82078B
  XXXXXXX old bdidev is 0000000000000000
  XXXXXXX new bdidev is ffff88001f20cd10
  XXXXXXX old bdidev is ffff88001f20cd10
  XXXXXXX new bdidev is ffff88001f20de30
Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

which very much looks like bdi will not remove any earlier registered device, only the last one, right?

Sharing a bdi is odd to begin with, let me poke at this a little.

Yeah, it's odd, but I'm not sure if you want to touch floppy.c. :) Any objection to this patch, until you possible fix it differently?
http://marc.info/?l=linux-kernel&m=122541569310798&w=4

Non at all, touching floppy.c is dangerous, folks might assume you know something about it ;-)