Re: Serious fat + ramdisk/loopback bug

Gordon Chaffee (chaffee@odie.cs.berkeley.edu)
Sat, 15 Jun 1996 15:23:45 -0700

In <hpa.31c09311.Linux.is.free@freya.yggdrasil.com> H. Peter Anvin writes:
> There seems to be a bad interaction between the FAT filesystem driver
> and the ramdisk/loopback drivers. This presumably indicates a buffer
> cache problem, possible something having to do with the new buffer
> cache code?

Take a look at your kernel log messages. You may see some messages like

Jun 15 15:14:52 zeego kernel: Directory sread (sector 32) failed
Jun 15 15:14:52 zeego kernel: ll_rw_block: device 07:00: only 1024-char blocks
implemented (512)

Then try mounting your loopback device with an explicit setting of the
block size:

mount -t msdos -o loop,blocksize=1024 -r floppy /mnt

or for the ram disk

mount -t msdos -o blocksize=1024 -r /dev/ram /mnt

This allowed it to work for me.

Gordon Chaffee
chaffee@bugs-bunny.cs.berkeley.edu