From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sukadev Bhattiprolu Subject: Re: [RFC][PATCH 4/5] Protect cinit from fatal signals Date: Mon, 1 Dec 2008 12:21:12 -0800 Message-ID: <20081201202112.GC12493@us.ibm.com> References: <20081126034242.GA23120@us.ibm.com> <20081126034634.GD23238@us.ibm.com> <20081127010753.GB13545@wavehammer.waldi.eu.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20081127010753.GB13545@wavehammer.waldi.eu.org> Sender: linux-kernel-owner@vger.kernel.org To: Bastian Blank , oleg@redhat.com, ebiederm@xmission.com, roland@redhat.com, containers@lists.osdl.org, linux-kernel@vger.kernel.org, xemul@openvz.org List-Id: containers.vger.kernel.org Bastian Blank [bastian@waldi.eu.org] wrote: | On Tue, Nov 25, 2008 at 07:46:34PM -0800, Sukadev Bhattiprolu wrote: | > To protect container-init from fatal signals, set SIGNAL_UNKILLABLE but | > clear it if it receives SIGKILL from parent namespace - so it is still | > killable from ancestor namespace. | | This sounds like a workaround. yes... | | > Note that container-init is still somewhat special compared to 'normal | > processes' - unhandled fatal signals like SIGUSR1 to a container-init | > are dropped even if they are from ancestor namespace. SIGKILL from an | > ancestor namespace is the only reliable way to kill a container-init. | | It sounds not right to make this special case for a "normal" process. | | However, no idea how to do this better. ... like I mentioned in the other message, we have tried different approaches and they were either intrusive or required more drastic changes in semantics. Container-inits are special in some ways and this change requires SIGKILL to terminate them.