Hi Alan, Thanks for your answer. On Sat, 18 Apr 2009 10:48:28 -0400 (EDT), Alan Stern wrote: > On Sat, 18 Apr 2009, Jean Delvare wrote: > > While I was testing kernel 2.6.30-rc2, I lost one of my USB sticks. The > > stick is no longer usable in any machine, all I get is "sdb: unknown > > partition table". So apparently the partition table was corrupted by > > the problem I hit. > > > > Here is what the logs have to say about what happened: > > > > Apr 17 10:48:16 hyperion kernel: usb 2-2.2: new full speed USB device using uhci_hcd and address 3 > > Apr 17 10:48:16 hyperion kernel: usb 2-2.2: not running at top speed; connect to a high speed hub > > Apr 17 10:48:16 hyperion kernel: usb 2-2.2: configuration #1 chosen from 1 choice > > Apr 17 10:48:16 hyperion kernel: Initializing USB Mass Storage driver... > > Apr 17 10:48:16 hyperion kernel: scsi4 : SCSI emulation for USB Mass Storage devices > > Apr 17 10:48:16 hyperion kernel: usbcore: registered new interface driver usb-storage > > Apr 17 10:48:16 hyperion kernel: USB Mass Storage support registered. > > Apr 17 10:48:16 hyperion kernel: usb-storage: device found at 3 > > Apr 17 10:48:16 hyperion kernel: usb-storage: waiting for device to settle before scanning > > Apr 17 10:48:21 hyperion kernel: scsi 4:0:0:0: Direct-Access USB007 mini-USB2BU 0.00 PQ: 0 ANSI: 2 > > Apr 17 10:48:21 hyperion kernel: sd 4:0:0:0: Attached scsi generic sg2 type 0 > > Apr 17 10:48:21 hyperion kernel: usb-storage: device scan complete > > Apr 17 10:48:21 hyperion kernel: sd 4:0:0:0: [sdb] 2015231 512-byte hardware sectors: (1.03 GB/983 MiB) > > Apr 17 10:48:21 hyperion kernel: sd 4:0:0:0: [sdb] Write Protect is off > > Apr 17 10:48:21 hyperion kernel: sd 4:0:0:0: [sdb] Mode Sense: 00 00 00 00 > > Apr 17 10:48:21 hyperion kernel: sd 4:0:0:0: [sdb] Assuming drive cache: write through > > Apr 17 10:48:21 hyperion kernel: sd 4:0:0:0: [sdb] Assuming drive cache: write through > > Apr 17 10:48:21 hyperion kernel: sdb: sdb1 > > So at the time of this log, the stick was working correctly (valid > partition table and valid size). Yes, I managed to mount it once. > > Apr 17 10:48:21 hyperion kernel: sd 4:0:0:0: [sdb] Attached SCSI removable disk > > Apr 17 10:48:25 hyperion hald: mounted /dev/sdb1 on behalf of uid 501 > > Apr 17 10:50:41 hyperion kernel: FAT: Filesystem panic (dev sdb1) > > Apr 17 10:50:41 hyperion kernel: fat_get_cluster: invalid cluster chain (i_pos 0) > > Apr 17 10:50:41 hyperion kernel: File system has been set read-only > > Sounds like a bad FAT or directory entry. Could be the stored data was > corrupted or a read didn't return the right information. Could this be caused by bad mount options, for example incorrect charset? As I wrote the stick isn't mine, it has probably been written to under Windows not Linux. I'm also curious if FAT has a limit to the number of files it can store? 8500 files seem a lot for a 1 GB stick. > ... lots of errors ... > > Apr 17 10:51:55 hyperion kernel: usb 2-2.2: reset full speed USB device using uhci_hcd and address 3 > > Apr 17 10:51:58 hyperion kernel: sd 4:0:0:0: [sdb] Assuming drive cache: write through > > Apr 17 10:51:58 hyperion kernel: sdb: unknown partition table > > And presumably you now see something like this every time you plug in > the stick? Yes. > > This happened as I was asking Thunar (Xfce's file manager) the > > properties of the USB stick. So it was counting all the files and there > > were a lot of them (about 8500 if memory serves). As the stick isn't > > mine, I have no idea if the file count was correct or not. I remember > > that the cumulative size looked plain wrong to me, something like 400 > > GB while the stick is supposed to be 1 GB total. > > > > I would also like to mention that my system had been hit by bug #13135 > > one hour before that. I have no idea if this may be related or not: > > http://bugzilla.kernel.org/show_bug.cgi?id=13135 > > Andrew Morton seems to think that one is a false positive, so probably > > it isn't related, but I thought I'd mention it just in case. > > > > I presume that the USB stick might just have been defective, either > > physically or at the filesystem level... It isn't mine, so I can't tell > > much about this. But in any case I find it weird that I ended up losing > > the partition table due to these errors. > > The thing is, the log doesn't really show any USB errors, although no > doubt something strange was going on. No way to tell what it was, now. I agree I don't really know who to blame. Looking at the logs again, it may be a filesystem problem rather than USB problem. That being said... > It's hard to imagine how Linux could have destroyed the stick merely by > reading a bunch of files from it. Most likely this was some sort of > hardware failure. ... this is indeed hard to imagine. I can't see how a filesystem bug could have overwritten data outside the partition. > If you still have the stick, and if it still is in working condition, > you should be able to get it going again by repartitioning and > reformatting it. Yes I still have it and it is still seen as an USB device. I will try repartitioning it and reformatting it, let's see where it gets me. I'm not sure exactly how I can restore the boot sector though. Maybe I'll copy it from another USB stick. And then I could try to reproduce the bug... who knows. Thanks, -- Jean Delvare -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
