On Wednesday 04 June 2003 13:25, Simon Burge wrote:

According to:

http://www.opengroup.org/onlinepubs/007904975/basedefs/errno.h.html#tag_1=
3_10

=2E..both ENOTTY and EOPNOTSUP are required by POSIX.

> We have around 16 ioctl handlers that return EOPNOTSUPP for an error
> instead of ENOTTY.  SUSv3/Posix doesn't list EOPNOTSUPP as an error,
> either does our man page for ioctl(2).  This will fix one of Matt
> Thomas' PCTS Posix test failures.
>
> A bit of mkid playing shows potential problems in:
>
> =09arch/amiga/dev/kbd.c
> =09arch/atari/dev/kbd.c
> =09arch/dreamcast/dev/maple/mms.c
> =09arch/x68k/dev/kbd.c
> =09arch/x68k/dev/pow.c
> =09coda/coda_vnops.c
> =09dev/mm.c
> =09dev/radio.c
> =09miscfs/fdesc/fdesc_vnops.c
> =09net/route.c
> =09netatalk/at_control.c
> =09netinet/if_arp.c
> =09netinet/in.c
> =09netinet6/in6.c
> =09netiso/iso.c
> =09netns/ns.c
>
> No one have any problems with changing any of these handlers to return
> ENOTTY?
>
> Simon.

// wbr