Messages in this thread | | | Date | Thu, 10 May 2001 19:01:16 -0700 | From | Jonathan Lundell <> | Subject | Re: Not a typewriter |
| |
At 8:07 PM -0400 2001-05-10, Alexander Viro wrote: >On Thu, 10 May 2001, Jonathan Lundell wrote: > >> ENOTTY is used by several non-serial devices (or file systems) to >> object to an unrecognized ioctl command. There's also ENOIOCTLCMD >> (apparently supposed to be a non-user errno, but i don't see where it >> gets changed to something else) and EINVAL. I'm not sure what the >> rationale is for choosing among them; perhaps someone would elucidate? > >ENOIOCTLCMD is something I've never met in the kernel. Normal reaction >to unrecognized ioctl() is ENOTTY, for a lot of reasons, starting with >the fact that ioctls are last-ditch API to be used when you just can't >think of better one and historically TTY had the earliest (and largest) >infestation. IOW, "not a tty" used to mean "WTF are you using ioctls here?" >OTOH, EINVAL is a catch-all thing for "something is wrong with arguments".
That's pretty much what I would have said a couple of hours ago before grepping the kernel. Try it, though. ENOTTY is rarely used. ENOIOCTLCMD is all over the damned place, though its comment in errno.h warns that a user shouldn't see it. And if you browse a bunch of random ioctl handlers, you'll see EINVAL used for a bad command much more often than ENOTTY.
FWIW, the comment in errno.h under Solaris 2.6 is "Inappropriate ioctl for device". I believe that's the POSIX interpretation. -- /Jonathan Lundell. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |