From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263791AbTDXTN7 (ORCPT ); Thu, 24 Apr 2003 15:13:59 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263813AbTDXTN7 (ORCPT ); Thu, 24 Apr 2003 15:13:59 -0400 Received: from sol.cobite.com ([208.222.80.183]:16620 "EHLO sol.cobite.com") by vger.kernel.org with ESMTP id S263791AbTDXTN4 (ORCPT ); Thu, 24 Apr 2003 15:13:56 -0400 Date: Thu, 24 Apr 2003 15:26:00 -0400 (EDT) From: David Mansfield X-X-Sender: david@sol.cobite.com To: Andrew Morton , Subject: odd gnome-terminal behavior in 2.5.67-mm3 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi Andrew, list I've been experiencing some odd behavior running 2.5.67-mm3 on my RH9 based desktop. It's probably an application bug, but something strange is happening anyway that doesn't happen in 'stable' kernels. What happens is that gnome-terminal gets stuck in some sort of 'infinite loop' when a lot of output is going to the screen and also keypresses are going in (like paging through a large file - holding down pgup/pgdown). Xterm doesn't seem to be affected. If I kill off the shell running inside the terminal, it recovers (that is, when gnome-terminal is running as a 'factory' the window that is hung disappears but the process and all of the other windows are fine). I captured some 'strace' of this while it was happening (sorry about the wrapping, the lines are quite long): ioctl(3, FIONREAD, [0]) = 0 poll([{fd=17, events=POLLIN}, {fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=8, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=11, events=POLLIN|POLLPRI}, {fd=12, ev ents=POLLIN|POLLPRI}, {fd=13, events=POLLIN|POLLPRI}, {fd=15, events=POLLIN|POLLPRI}, {fd=14, events=POLLIN|POLLPRI}], 10, 0) = 0 write(3, "+\20\1\0", 4) = 4 read(3, "\1\1-\205\0\0\0\0&\0\300\3\0\0\0\0\1\0\0\0\0\0\0\0\370"..., 32) = 32 ioctl(3, FIONREAD, [0]) = 0 write(3, "5\20\4\0.\356)\2q\266#\2\24\0\25\0007\0\5\0/\356)\2.\356"..., 328) = 328 read(3, "\1\1A\205\0\0\0\0&\0\300\3\0\0\0\0\0\0\0\0\0\0\0\0\370"..., 32) = 32 ioctl(3, FIONREAD, [0]) = 0 poll([{fd=17, events=POLLIN}, {fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=8, events=POLLIN}, {fd=10, events=POLLIN|POLLPRI}, {fd=11, events=POLLIN|POLLPRI}, {fd=12, ev ents=POLLIN|POLLPRI}, {fd=13, events=POLLIN|POLLPRI}, {fd=15, events=POLLIN|POLLPRI}, {fd=14, events=POLLIN|POLLPRI}, {fd=22, events=POLLIN}, {fd=23, events=POLLIN}, {fd=21, events=POLLIN, revents=POLLIN}, {fd=24, events=POLLIN}, {fd=16, events=POLLIN}, {fd=25, events=POLLIN}], 16, 0) = 1 read(21, 0xbfffe8f0, 4294965249) = -1 EFAULT (Bad address) getpid() = 31085 write(2, "\n** (gnome-terminal:31085): WARN"..., 79) = 79 This exact sequence repeats ad infinitum. The gnome-terminal is apparently doing an invalid read of -1 bytes after poll is saying there is data available. I don't know what this could mean. David -- /==============================\ | David Mansfield | | lkml@dm.cobite.com | \==============================/