From owner-freebsd-hackers Mon Jul 17 15:12:32 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id PAA27194 for hackers-outgoing; Mon, 17 Jul 1995 15:12:32 -0700 Received: from rover.village.org (rover.village.org [198.137.146.49]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id PAA27174 for ; Mon, 17 Jul 1995 15:12:02 -0700 Received: from localhost (localhost [127.0.0.1]) by rover.village.org (8.6.11/8.6.6) with SMTP id QAA08206 for ; Mon, 17 Jul 1995 16:11:52 -0600 Message-Id: <199507172211.QAA08206@rover.village.org> To: hackers@freebsd.org Subject: ENOTTY???? Date: Mon, 17 Jul 1995 16:11:51 -0600 From: Warner Losh Sender: hackers-owner@freebsd.org Precedence: bulk Under what cases will a read return ENOTTY in FreeBSD 2.0R? I can't seem to find it in the kernel sources... That's not a ioctl (which I can find), but a read. Any ideas? What I'm seeing that is that, under heavy load (56700bps) TIA is exiting because a read returned -1 and errno is set to ENOTTY. I've placed a breakpoint at close, and I know I'm not closing the TTY fd. The fd is in raw mode. I didn't see a ENOTTY as a possible error code in the read(2) man page, and the entry for ENOTTY in the errno.h file just says the ioctl is only for ttys and this fd isn't a tty (but it is!). No signals appear to be generated, and I'm fairly sure that carrier isn't being dropped (it is a direct line, null modem). Normally, I don't like to ask these sorts of questions, but this smells like a kernel bug. Comments? Warner