From owner-freebsd-current@FreeBSD.ORG Sun Nov 21 18:29:08 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id B198B16A4CE for ; Sun, 21 Nov 2004 18:29:08 +0000 (GMT) Received: from bombadil.mebtel.net (bombadil.mebtel.net [64.40.67.44]) by mx1.FreeBSD.org (Postfix) with ESMTP id 6F52143D45 for ; Sun, 21 Nov 2004 18:29:08 +0000 (GMT) (envelope-from dlt@mebtel.net) Received: from localhost (localhost [127.0.0.1]) by bombadil.mebtel.net (Postfix) with ESMTP id BFAE0208686 for ; Sun, 21 Nov 2004 13:29:07 -0500 (EST) Received: from bombadil.mebtel.net ([127.0.0.1]) by localhost (bombadil [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 19453-08 for ; Sun, 21 Nov 2004 13:29:07 -0500 (EST) Received: from lorne.arm.org (66-79-79-177.dsl.mebtel.net [66.79.79.177]) by bombadil.mebtel.net (Postfix) with ESMTP id 4C48D208684 for ; Sun, 21 Nov 2004 13:29:07 -0500 (EST) Received: from lorne.arm.org (localhost [127.0.0.1]) by lorne.arm.org (8.13.1/8.12.9) with ESMTP id iALIT67p001027 for ; Sun, 21 Nov 2004 13:29:07 -0500 (EST) (envelope-from dlt@lorne.arm.org) Received: (from dlt@localhost) by lorne.arm.org (8.13.1/8.13.1/Submit) id iALIT6Vi001024; Sun, 21 Nov 2004 13:29:06 -0500 (EST) (envelope-from dlt) Date: Sun, 21 Nov 2004 13:29:06 -0500 (EST) Message-Id: <200411211829.iALIT6Vi001024@lorne.arm.org> From: Derek Tattersall To: current@FreeBSD.org X-Virus-Scanned: by amavisd-new at mebtel.net Subject: Error in write to named pipe X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list Reply-To: dlt@mebtel.net List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 21 Nov 2004 18:29:08 -0000 I have an application that reads a file and writes some lines to a named pipe. It's just a sigmonster. On a current kernel/world from 11/11 the signature gets created and written to the pipe quite nicely. On a kernel/world from this morning (1600 GMT) the app fails to write the signature. The pipe and write portion of the ktrace is as follows: 929 signature CALL fork 929 signature RET fork 930/0x3a2 929 signature CALL exit(0) 931 signature RET open 3 931 signature CALL pipe 931 signature RET pipe 4 931 signature CALL fork 931 signature RET fork 959/0x3bf 931 signature CALL wait4(0x3bf,0,0,0) 931 signature RET wait4 959/0x3bf 931 signature CALL close(0x5) 931 signature RET close 0 931 signature CALL read(0x4,0xbfbfdbd4,0x200) 931 signature GIO fd 4 read 64 bytes "Don't abandon hope: your Tom Mix decoder ring arrives tomorrow. " 931 signature RET read 64/0x40 931 signature CALL close(0x4) 931 signature RET close 0 931 signature CALL __sysctl(0xbfbfd1f0,0x2,0xbfbfd630,0xbfbfd1ec,0,0) 931 signature RET __sysctl 0 931 signature CALL __sysctl(0xbfbfd1f0,0x2,0xbfbfd730,0xbfbfd1ec,0,0) 931 signature RET __sysctl 0 931 signature CALL __sysctl(0xbfbfd1f0,0x2,0xbfbfd830,0xbfbfd1ec,0,0) 931 signature RET __sysctl 0 931 signature CALL __sysctl(0xbfbfd1f0,0x2,0xbfbfd930,0xbfbfd1ec,0,0) 931 signature RET __sysctl 0 931 signature CALL __sysctl(0xbfbfd1f0,0x2,0xbfbfda30,0xbfbfd1ec,0,0) 931 signature RET __sysctl 0 931 signature CALL write(0x3,0xbfbfe3d4,0xad) 931 signature RET write -1 errno 45 Operation not supported 931 signature CALL close(0x3) 931 signature RET close 0 931 signature CALL nanosleep(0xbfbfdb38,0xbfbfdb30) 931 signature RET nanosleep 0 931 signature CALL access(0x804b220,0) 931 signature NAMI "/home/dlt/.signature" 931 signature RET access 0 931 signature CALL open(0x804b220,0x1,0x400) 931 signature NAMI "/home/dlt/.signature" Why would write suddenly be returning "operation not supported" when it's to a pipe? -- Derek Tattersall dlt@mebtel.net dlt666@yahoo.com