On Thu, Jul 05, 2001 at 10:30:10PM +0200, Ivan Popov wrote:
> 35, sometimes 36
> 
> Happens once, always(??), rather soon during the first /coda/....
> accesses.
> 
> -------------------------------------------
> 22:11:55 /coda now mounted.
> 
> 22:12:12 worker::main Got a bogus opcode 36
> -------------------------------------------
> 
> It doesn't seem to do any harm, but I would like to know for sure :)

Yup, no harm at all. The linux-2.4.4 kernel module just tries to use a
new close (opcodes 35 & 36) that would allow errorcodes during close to
be passed back to userspace. However, venus doesn't support this yet.

We used to call close when the in kernel open count dropped to zero, but
by that time it is too late to return an error code to userspace. The
existing close has to be split up into pretty much everything that close
currently does, but decrementing of the owrite refcount has to be done
later.

When venus returns EOPNOTSUPP the kernel falls back to using the old
CODA_CLOSE upcall.

Jan