NetBSD Problem Report #114
Date: Tue, 8 Feb 1994 09:35:08 -0800
From: jimj@na1.cs.uwm.edu
Reply-To: jimj@na1.cs.uwm.edu
To: gnats-admin
Cc: gnats-admin, netbsd-bugs
Subject: kern/114: fifofs
>Number: 114
>Category: kern
>Synopsis: fifofs bugs/features
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: gnats-admin
>State: closed
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Tue Feb 08 09:35:02 +0000 1994
>Closed-Date: Wed Mar 16 18:29:12 +0000 1994
>Last-Modified: Thu Jan 02 18:12:54 +0000 2014
>Originator: James Jegers
>Release:
>Organization:
" "
>Environment:
System: NetBSD na1.cs.uwm.edu 0.9a JIMJ#0 i386
>Description:
The fifofs (named pipes in the filesystem) have some strange behavious
which I'm not sure if they are bugs or features.
1. A pipe can only be open as read or as write, not as read/write.
I see this comparison is the first line in the fifo_open code. Other
machines I've worked on, AIX, HPUX, Solaris, Sun4.1 support
opening of the pipe for both reading/writing.
2. If one opens the pipe for reading, and there is not a writter at
the other end: Doing a select for read will return true, but
when the read is actually done it will block as there really
isn't any data there. This might be a because pipes are implmented
as sockets, maybe there should be a check in there for select
to see if the other end of the socket exists instead of just
calling the socket_select.
>How-To-Repeat:
>Fix:
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->closed
State-Changed-By: jtc
State-Changed-When: Wed Mar 16 18:29:12 1994
State-Changed-Why:
The history of this bug was lost when the PR database was corrupted.
This was classified as not a bug.
1. FIFO's are unidirectional, so opening them R/W makes no sense
2. ???
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.