[Bug 586172] New: sysfs: cannot create duplicate filename '/bus/ieee1394/drivers/sbp2...
http://bugzilla.novell.com/show_bug.cgi?id=586172 http://bugzilla.novell.com/show_bug.cgi?id=586172#c0 Summary: sysfs: cannot create duplicate filename '/bus/ieee1394/drivers/sbp2... Classification: openSUSE Product: openSUSE 11.2 Version: Final Platform: x86-64 OS/Version: openSUSE 11.2 Status: NEW Severity: Normal Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: jpschewe@mtu.net QAContact: qa@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.8) Gecko/20100204 SUSE/3.5.8-0.1.1 Firefox/3.5.8 I have a Western Digital MyBook attached via Firewire to my desktop. With the latest 11.2 kernel I'm receiving this error on every boot. The drive works fine, but the warning is concerning. <6>[ 13.564098] input: HDA Digital PCBeep as /devices/pci0000:00/0000:00:10.1/input/input5 <7>[ 13.761068] ieee1394: Node added: ID:BUS[0-00:1023] GUID[0090a976213be423] <7>[ 13.761150] ieee1394: Host added: ID:BUS[0-01:1023] GUID[0010dc0000cf92b6] <6>[ 13.930522] scsi6 : SBP-2 IEEE-1394 <6>[ 18.409456] ieee1394: sbp2: Logged into SBP-2 device <6>[ 18.439171] ieee1394: sbp2: Node 0-00:1023: Max speed [S400] - Max payload [1024] <5>[ 18.569852] scsi 6:0:0:0: Direct-Access WD My Book 1025 PQ: 0 ANSI: 4 <5>[ 18.586911] sd 6:0:0:0: Attached scsi generic sg2 type 0 <6>[ 18.605107] scsi7 : SBP-2 IEEE-1394 <5>[ 18.643641] sd 6:0:0:0: [sdb] 976773168 512-byte logical blocks: (500 GB/465 GiB) <5>[ 18.691159] sd 6:0:0:0: [sdb] Write Protect is off <7>[ 18.707856] sd 6:0:0:0: [sdb] Mode Sense: 10 00 00 00 <5>[ 18.723828] sd 6:0:0:0: [sdb] Cache data unavailable <3>[ 18.740441] sd 6:0:0:0: [sdb] Assuming drive cache: write through <5>[ 18.841637] sd 6:0:0:0: [sdb] Cache data unavailable <3>[ 18.858351] sd 6:0:0:0: [sdb] Assuming drive cache: write through <6>[ 18.875218] sdb: sdb1 <5>[ 19.509210] sd 6:0:0:0: [sdb] Cache data unavailable <3>[ 19.526010] sd 6:0:0:0: [sdb] Assuming drive cache: write through <5>[ 19.542626] sd 6:0:0:0: [sdb] Attached SCSI disk <6>[ 22.908499] ieee1394: sbp2: Logged into SBP-2 device <6>[ 22.938881] ieee1394: sbp2: Node 0-00:1023: Max speed [S400] - Max payload [1024] <5>[ 23.074794] scsi 7:0:1:0: Enclosure WD My Book Device PQ: 0 ANSI: 4 <5>[ 23.091848] scsi 7:0:1:0: Attached scsi generic sg3 type 13 <5>[ 23.108446] ses 7:0:1:0: Attached Enclosure device <4>[ 23.126186] ------------[ cut here ]------------ <4>[ 23.142608] WARNING: at /usr/src/packages/BUILD/kernel-default-2.6.31.12/linux-2.6.31/fs/sysfs/dir.c:487 sysfs_add_one+0xde/0x150() <4>[ 23.159767] Hardware name: MS-7207PV <4>[ 23.176784] sysfs: cannot create duplicate filename '/bus/ieee1394/drivers/sbp2/0090a976213be423-1' <4>[ 23.194258] Modules linked in: ses enclosure sbp2 snd_hda_codec_realtek ppdev pcspkr ohci1394 floppy ieee1394 parport_pc parport sr_mod sg cdrom snd_hda_intel snd_hda_codec snd_hwdep snd_pcm button snd_timer amd64_edac_mod forcedeth edac_core k8temp snd snd_page_alloc nvidia(P) i2c_nforce2 preloadtrace linear raid456 raid6_pq async_xor async_memcpy async_tx xor raid0 dm_snapshot dm_mirror dm_region_hash dm_log reiserfs aic7xxx scsi_transport_spi edd dm_mod raid1 fan ide_pci_generic amd74xx ide_core ata_generic sata_nv pata_amd thermal processor thermal_sys <4>[ 23.268356] Pid: 624, comm: knodemgrd_0 Tainted: P 2.6.31.12-0.1-default #1 <4>[ 23.287049] Call Trace: <4>[ 23.305454] [<ffffffff81011749>] try_stack_unwind+0x189/0x1b0 <4>[ 23.324009] [<ffffffff8101013d>] dump_trace+0x9d/0x330 <4>[ 23.342316] [<ffffffff81011254>] show_trace_log_lvl+0x64/0x90 <4>[ 23.360635] [<ffffffff810112a3>] show_trace+0x23/0x40 <4>[ 23.378712] [<ffffffff81555268>] dump_stack+0x81/0x9e <4>[ 23.378719] [<ffffffff8106db80>] warn_slowpath_common+0x80/0xd0 <4>[ 23.378724] [<ffffffff8106dc5b>] warn_slowpath_fmt+0x4b/0x70 <4>[ 23.378730] [<ffffffff811d6f3e>] sysfs_add_one+0xde/0x150 <4>[ 23.378735] [<ffffffff811d756b>] sysfs_do_create_link+0x14b/0x1a0 <4>[ 23.378740] [<ffffffff811d7621>] sysfs_create_link+0x21/0x40 <4>[ 23.378746] [<ffffffff8135fac8>] driver_sysfs_add+0x38/0xa0 <4>[ 23.378751] [<ffffffff81360050>] device_bind_driver+0x20/0x50 <4>[ 23.378756] [<ffffffff813600cb>] device_attach+0x4b/0xc0 <4>[ 23.378761] [<ffffffff8135e1c1>] bus_rescan_devices_helper+0x61/0x90 <4>[ 23.378766] [<ffffffff8135ed68>] bus_for_each_dev+0x78/0xc0 <4>[ 23.378771] [<ffffffff8135edd7>] bus_rescan_devices+0x27/0x50 <4>[ 23.378787] [<ffffffffa0c575a2>] nodemgr_host_thread+0x272/0x330 [ieee1394] <4>[ 23.378808] [<ffffffff8108ea66>] kthread+0xb6/0xc0 <4>[ 23.378814] [<ffffffff8100d70a>] child_rip+0xa/0x20 <4>[ 23.378819] ---[ end trace 02b373f7d663109a ]--- Reproducible: Always Steps to Reproduce: 1. 2. 3. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c1
Jeff Mahoney
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c2
Jon Schewe
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c3
Jon Schewe
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c4
Jon Schewe
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c5
Stefan Richter
It looks like your device is getting detected as both a direct access disk and an enclosure device.
Indeed, several FireWire disks with a button (used as backup program launcher on other OSs) do represent themselves as two logical units, one of type Disk or RBC Disk, the other of type Enclosure Services. In this particular case, ieee1394 should add the following devices (e.g. to be found symlinked into /sys/bus/ieee1394/devices/) when the disk is plugged in: 0090a976213be423 for the node 0090a976213be423-0 for the disk (or the button) 0090a976213be423-1 for the button (or the disk) There is also a different internal organization of the node possible so that both SCSI logical units are represented by the IEEE 1394 unit 0090a976213be423-0, but this does not seem to be the case with this WD disk. I don't have a WD drive myself to reproduce the issue. The dual-unit devices which I have do not trigger this problem. Therefore and because the bug does apparently not prevent you from accessing the disk and because the ieee1394 kernel driver stack is in the process of being replaced by the alternative firewire kernel driver stack, it is not very likely that somebody is going to fix this. Does the SuSE kernel package already provide the new firewire drivers in parallel with the old ieee1394 drivers? If yes, you can try the new drivers by "modprobe -r ohci1394 && modprobe firewire-ohci && modprobe firewire-sbp2", then plug the disk in. The new drivers aren't actually that new anymore; they have been stable for FireWire storage devices for quite some time now. https://ieee1394.wiki.kernel.org/index.php/Juju_Migration -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c6
--- Comment #6 from Jon Schewe
find . -name firewire-ohci.ko -print
-- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c7
--- Comment #7 from Stefan Richter
This disk does not have a button on it. It is just a disk.
Ah, right, the MyBook. I guess the Enclosure logical unit is for the status LEDs that are used as a capacity indicator under other OSs. If you are willing to spend the effort to narrow down what the ieee1394 driver is doing wrong, you could get me some device data in one of the following two ways: /1./ Install a kernel that has new driver stack enabled, and attach the file /sys/bus/firewire/devices/fw1/config_rom here. It is a binary file. The 1 in this path assumes that the host was discovered before the disk. (host: fw0, disk: fw1, units on the disk: fw1.0 and fw1.1) /2./ Install the tool firecontrol (http://firecontrol.sourceforge.net/, formerly known as 1394commander, if neither firecontrol nor 1394commander is not packaged for SuSE, install the devel version of libraw1394 and compile firecontrol), then run { for i in {4,5,6,7}{0,1,2,3,4,5,6,7,8,9,a,b,c,d,e,f}{0,4,8,c}; do echo "r . 0 0xfffff0000$i 4"; done } | firecontrol > log and attach the log here. The 0 after r . in the command string assumes that the disk was discovered as node ID:BUS[0-00:1023] whereas the host is ID:BUS[0-01:1023]. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c8
--- Comment #8 from Jon Schewe
This disk does not have a button on it. It is just a disk.
Ah, right, the MyBook. I guess the Enclosure logical unit is for the status LEDs that are used as a capacity indicator under other OSs.
There is only 1 LED on the front that flashes with activity. No capacity indication.
If you are willing to spend the effort to narrow down what the ieee1394 driver is doing wrong, you could get me some device data in one of the following two ways:
/1./ Install a kernel that has new driver stack enabled, and attach the file /sys/bus/firewire/devices/fw1/config_rom here. It is a binary file.
The 1 in this path assumes that the host was discovered before the disk. (host: fw0, disk: fw1, units on the disk: fw1.0 and fw1.1)
Not likely to get this unless there already is one packaged for opensuse.
/2./ Install the tool firecontrol (http://firecontrol.sourceforge.net/, formerly known as 1394commander, if neither firecontrol nor 1394commander is not packaged for SuSE, install the devel version of libraw1394 and compile firecontrol), then run { for i in {4,5,6,7}{0,1,2,3,4,5,6,7,8,9,a,b,c,d,e,f}{0,4,8,c}; do echo "r . 0 0xfffff0000$i 4"; done } | firecontrol > log and attach the log here.
The 0 after r . in the command string assumes that the disk was discovered as node ID:BUS[0-00:1023] whereas the host is ID:BUS[0-01:1023]. Tried this, but got lots of "insufficient arguments for operation!" And the log is full of: Command: 0xfffff000064c 4 illegal command! Command: r . 0
Will attach -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c9
--- Comment #9 from Jon Schewe
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c10
Jon Schewe
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c
Stefan Richter
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c11
--- Comment #11 from Stefan Richter
This disk does not have a button on it. It is just a disk.
Ah, right, the MyBook. I guess the Enclosure logical unit is for the status LEDs that are used as a capacity indicator under other OSs.
There is only 1 LED on the front that flashes with activity. No capacity indication.
Strange then that there is another logical unit at all. Still, it shouldn't matter to ieee1394, it should create 0090a976213be423-0 and 0090a976213be423-1 once but not attempt to repeat the latter. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c12
Jeff Mahoney
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c13
--- Comment #13 from Stefan Richter
I'm going to enable CONFIG_FIREWIRE for the Factory kernel. Should the old stack be enabled at all?
By now, the new drivers support all of the hardware that the old drivers support, except for the moderately rare (and no longer produced) ALi PCI-FireWire controllers and the extremely rare NForce2 (which don't fully workwith the old stack either) and AFAIK the rare and ancient Apple UniNorth v1 controllers. Still, since Fedora has been the only distribution to ship the new drivers until recently, it may make sense to install the old drivers too as a fallback for unforeseen issues. Hard to say whether such a parallel installation or a single-stack installation will result in least support issues. The drivers as well as the low-level libraries libraw1394 and libdc1394 are fit to work seamlessly either way.
What enforces that the firewire modules are higher priority?
The way the kernel's makefiles are ordered, the driver core will first attempt to bind firewire-ohci before ohci1394 if both have been loaded into the kernel or statically linked into the kernel. A modular build together with shipped modprobe blacklist entries as listed at https://ieee1394.wiki.kernel.org/index.php/Juju_Migration#How_to_suppress_au... gives explicit control though. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c14
Jeff Mahoney
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c15
Stefan Richter
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c16
--- Comment #16 from Jeff Mahoney
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c17
--- Comment #17 from Michal Marek
Michal, can you add the blacklist entries? Do we have a way of doing that on a per-kernel basis so we don't break 11.2 installations using 11.3M kernels in the interim?
Do we really need this? Since we have modules.order support in module-init-tools, module priority should follow the makefile order. -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c18
Jon Schewe
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c19
--- Comment #19 from Stefan Richter
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c20
Jon Schewe
<4>[ 21.594754] Modules linked in: snd_hda_codec_realtek sbp2 snd_hda_intel firewire_ohci firewire_core snd_hda_codec ppdev amd64_edac_mod crc_itu_t sr_mod ^^^^^^^^^^^^^ This is the new module, right? parport_pc snd_hwdep cdrom parport edac_core snd_pcm i2c_nforce2 sg forcedeth ohci1394 ^^^^^^^^ And this is the obsolete one? It looks like the the obsolete module gets loaded first (the module list is single linked, so the last module is the head of the
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c21
--- Comment #21 from Michal Marek
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c22
--- Comment #22 from Jon Schewe
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c23
--- Comment #23 from Stefan Richter
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c24
Michal Marek
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c25
--- Comment #25 from Jon Schewe
In case of a dual stack installation, autoloading of the modules can be controlled thus: https://ieee1394.wiki.kernel.org/index.php/Juju_Migration#How_to_suppress_au...
I just rebooted with these blacklist settings and the error didn't appear on this reboot. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c26
--- Comment #26 from Stefan Richter
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c
Jeff Mahoney
https://bugzilla.novell.com/show_bug.cgi?id=586172
https://bugzilla.novell.com/show_bug.cgi?id=586172#c27
Jon Schewe
http://bugzilla.novell.com/show_bug.cgi?id=586172
http://bugzilla.novell.com/show_bug.cgi?id=586172#c28
--- Comment #28 from Bernhard Wiedemann
participants (1)
-
bugzilla_noreply@novell.com