On Tue, Apr 13, 2010 at 10:33:04PM -0400, Phillip Susi wrote: > On Tue, 13 Apr 2010 18:29:56 +0200 > Petr Uzel <[email protected]> wrote: > > > Just a blind guess: did you run the test with parter with 'reintroduce > > BLKPG' patch applied? AFAIK the bug doesn't show with BLKRRPART. > > The two issues are not related. This is a confirmed bug in the kernel > where it always sets the size of the extended partition to 2 sectors, > even when the second sector is not there since it is the first sector of > the logical partition.
ACK > The BLKPG patch just causes the error to show > up immediately instead of after a reboot, since without the patch > BLKRRPART fails and the partition table is not read until reboot. I don't think so. If the offset between extended and first logical partition is only one sector, BLKPG definitely fails. But BLKRRART seems to handle such situation just fine: #> modprobe scsi_debug dev_size_mb=1 #> parted -s /dev/sdd mklabel msdos #> parted -s /dev/sdd mkpart extended 64s 128s #> strace -o strace.out parted -s /dev/sdd mkpart logical 65s 128s #> parted -s /dev/sdd unit s print Model: Linux scsi_debug (scsi) Disk /dev/sdd: 2048s Sector size (logical/physical): 512B/512B Partition Table: msdos Number Start End Size Type File system Flags 1 64s 128s 65s extended lba 5 65s 128s 64s logical #> grep BLKRRPART strace.out ioctl(3, BLKRRPART, 0x6c) = 0 So unless I missed something, I believe the partition table will be successfully read after reboot (I didn't try, though). Nonetheless, if parted is going to switch back from BLKRRPART to BLKPG (I hope it will), your patch is definitely needed. Petr -- Petr Uzel, openSUSE Boosters Team IRC: ptr_uzl @ freenode
pgphoUtzWRVUg.pgp
Description: PGP signature
_______________________________________________ parted-devel mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/parted-devel
