I noticed an important thing about the famous RMB. I initialized 4
disksin this way:
230MB disk (512 bytes/sector) a single partition covering the whole
disk
230MB disk (512 bytes/sector) 2 partitions covering 1/2 disk each
640MB disk (2048 bytes/sector) a single partition covering the whole
disk
640MB disk (2048 bytes/sector) 2 partitions covering 1/2 disk each
When I mount a partition from a 640MB disk, then I change disk and I
try to mount
the new one (640MB) I get some errors about ext2 corruption(bsd inode
etc. etc.. But if I do the same with 230MB disk, all goes right. No
errors, no warning, no data loss.
So we can definitely say that all problems come from the non-standard
sector size. After all, support for non-512 bytes sized sectors come
only with 2.2.1/2 as far I remember (before that I wasn't able to use
640MB disks at all), so the presence of some hidden bug is somewhat
normal...
Anyway, how can I trace the bug and fix it ? My knowledge of the
kernel is far too spare to locate the bug by myself :(
Bye.
P.S.: The drive do not send UNIT_ATTENTION, so I still need to try to
mount the device with no disk insented before inserting the new disk to
force the kernel to consider a disk_change event, but this is another
story...
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/