From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263291AbTJKMBO (ORCPT ); Sat, 11 Oct 2003 08:01:14 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263292AbTJKMBN (ORCPT ); Sat, 11 Oct 2003 08:01:13 -0400 Received: from imladris.demon.co.uk ([193.237.130.41]:64236 "EHLO imladris.demon.co.uk") by vger.kernel.org with ESMTP id S263291AbTJKMBK (ORCPT ); Sat, 11 Oct 2003 08:01:10 -0400 From: David Woodhouse To: Ingo Molnar Cc: linux-kernel@vger.kernel.org Message-Id: <1065873664.30987.431.camel@imladris.demon.co.uk> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.5 (1.4.5-2.dwmw2.3) Date: Sat, 11 Oct 2003 13:01:05 +0100 X-SA-Exim-Mail-From: dwmw2@infradead.org Subject: Signals and kernel threads. Content-Type: text/plain Content-Transfer-Encoding: 7bit X-SA-Exim-Version: 3.0+cvs (built Mon Aug 18 15:53:30 BST 2003) X-SA-Exim-Scanned: Yes Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Once upon a time, kernel threads could make use of signals for receiving control from userspace. Now, the NPTL code looks at the recipient thread's userspace handlers for signals, even in the case of a kernel thread, and if there's not a handler installed will either discard the signal or convert it to SIGKILL. Could we disable this behaviour in the case where the recipient is a kernel thread please? Or at least make allow_signal() set sa_handler to some non-zero dummy value so that the signal is delivered as intended. -- dwmw2