Bugzilla – Bug 133973
udev does not create device node for extended partition
Last modified: 2005-11-17 14:11:37 UTC
On purcell udev does not create device node for extended partition. (parted) p Disk geometry for /dev/hdb: 0kB - 9115MB Disk label type: msdos Number Start End Size Type File system Flags 1 1kB 1000MB 1000MB primary type=83 2 1000MB 2000MB 1000MB primary type=83 3 2000MB 3000MB 1000MB primary type=83 4 3000MB 8000MB 5000MB extended type=05 5 3000MB 4000MB 1000MB logical type=83 tcsh: ll /dev/hdb* brw-r----- 1 root disk 3, 64 2005-11-15 12:07 /dev/hdb brw-r----- 1 root disk 3, 65 2005-11-16 11:02 /dev/hdb1 brw-r----- 1 root disk 3, 66 2005-11-16 11:02 /dev/hdb2 brw-r----- 1 root disk 3, 67 2005-11-16 11:02 /dev/hdb3 brw-r----- 1 root disk 3, 69 2005-11-16 11:02 /dev/hdb5 Hannes said this is worth a bug report.
Kay? Do we need them? Should we have them at all?
Sure, udev creates them if and cause they are exported by the kernel. What devices are in /sys/block/hdb?
purcell:/sys/block/hdb # ll total 0 drwxr-xr-x 7 root root 0 2005-11-16 15:05 . drwxr-xr-x 36 root root 0 2005-11-15 12:07 .. -r--r--r-- 1 root root 4096 2005-11-16 15:05 dev lrwxrwxrwx 1 root root 0 2005-11-16 15:05 device -> ../../devices/pci0000:00/0000:00:07.1/ide0/0.1 drwxr-xr-x 2 root root 0 2005-11-16 11:02 hdb1 drwxr-xr-x 2 root root 0 2005-11-16 11:02 hdb2 drwxr-xr-x 2 root root 0 2005-11-16 11:02 hdb3 drwxr-xr-x 2 root root 0 2005-11-16 11:02 hdb5 drwxr-xr-x 3 root root 0 2005-11-15 12:07 queue -r--r--r-- 1 root root 4096 2005-11-16 15:05 range -r--r--r-- 1 root root 4096 2005-11-16 15:05 removable -r--r--r-- 1 root root 4096 2005-11-16 15:05 size -r--r--r-- 1 root root 4096 2005-11-16 15:05 stat
There is no kernel device to talk to, therefore no device node. Seems that the IDE layer does not create devices for extended partitions, don't know if that's the default on all IDE volumes, SCSI exports the extendend partition as a block device.
I'm closing this, you may need to work around it, I don't think, that somebody will fix the unfixable IDE subsystem in the kernel. :)
No problem with me. At least we showed that the code in libstorage is not "dubiously" after all.
Nothing to do with the kernel. It's a bug in parted. purcell:~ # cat /proc/partitions major minor #blocks name 3 0 60051600 hda 3 1 4819468 hda1 3 2 4819500 hda2 3 3 1052257 hda3 3 4 49359712 hda4 3 64 8901720 hdb 3 65 976562 hdb1 3 66 976562 hdb2 3 67 976562 hdb3 3 69 976562 hdb5 253 0 4194304 dm-0 253 1 15728640 dm-1 253 2 28835840 dm-2 253 3 598016 dm-3 purcell:~ # fdisk /dev/hdb The number of cylinders for this disk is set to 17662. There is nothing wrong with that, but this is larger than 1024, and could in certain setups cause problems with: 1) software that runs at boot time (e.g., old versions of LILO) 2) booting and partitioning software from other OSs (e.g., DOS FDISK, OS/2 FDISK) Command (m for help): w The partition table has been altered! Calling ioctl() to re-read partition table. Syncing disks. purcell:~ # cat /proc/partitions major minor #blocks name 3 0 60051600 hda 3 1 4819468 hda1 3 2 4819500 hda2 3 3 1052257 hda3 3 4 49359712 hda4 3 64 8901720 hdb 3 65 976562 hdb1 3 66 976562 hdb2 3 67 976562 hdb3 3 68 1 hdb4 3 69 976562 hdb5 253 0 4194304 dm-0 253 1 15728640 dm-1 253 2 28835840 dm-2 253 3 598016 dm-3
So, it shows up in sysfs after fdisk, not only in /proc?