panic: mount: lost mount

[David Schultz]

On Wed, May 21, 2003, Bruce Evans wrote:
> On Wed, 21 May 2003, Jun Kuriyama wrote:
> 
> > With today's current, I've got a panic with bad floppy.
> >
> > fd0: hard error cmd=read fsbn 57 of 56-63 (ST0 44<abnrml,top_head> ST1 20<bad_crc> ST2 20<bad_crc> cyl 1 hd 1 sec 4)
> > panic: mount: lost mount
> > cpuid = 1; lapic.id = 01000000
> > Debugger("panic")
> > Stopped at      Debugger+0x55:  xchgl   %ebx,in_Debugger.0
> > db> trace
> > Debugger(c03eb40e,1000000,c03f085e,ebc05b80,1) at Debugger+0x55
> > panic(c03f085e,ebc05ba4,c03f0812,46a,c896b720) at panic+0x11f
> > vfs_mount(c896b720,c7eafd40,c8a32b80,0,bfbfece0) at vfs_mount+0xa80
> > mount(c896b720,ebc05d10,c0407bf6,3fb,4) at mount+0xb8
> > syscall(2f,2f,2f,80a210f,bfbff6a4) at syscall+0x26e
> > Xint0x80_syscall() at Xint0x80_syscall+0x1d
> 
> Unfortunately, this is fairly normal file system behaviour when a critical
> block is unreadable or damaged.  Here vfs detects a problem that it knows
> it cannot handle, and panics.

I've run into this as well while testing other properties of how
removable media is handled.  Is there an easy way to get slightly
more graceful behavior, such as forcing a downgrade to r/o and
zapping the vnodes for any unrecoverable files a la 'umount -f'?