From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <53107343.9010707@xenomai.org> Date: Fri, 28 Feb 2014 12:30:11 +0100 From: Gilles Chanteperdrix MIME-Version: 1.0 References: <530F96D3.4040407@xenomai.org> <53104C58.3040903@xenomai.org> <53106E73.9030702@xenomai.org> <53106EF2.2020109@xenomai.org> In-Reply-To: <53106EF2.2020109@xenomai.org> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Subject: Re: [Xenomai] [Xenomai-git] Philippe Gerum : copperplate/registry: reap sysregd selectively List-Id: Discussions about the Xenomai project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Philippe Gerum Cc: xenomai@xenomai.org On 02/28/2014 12:11 PM, Philippe Gerum wrote: > On 02/28/2014 12:09 PM, Gilles Chanteperdrix wrote: >> On 02/28/2014 09:44 AM, Philippe Gerum wrote: >>> On 02/27/2014 08:49 PM, Gilles Chanteperdrix wrote: >>>> On 02/27/2014 04:54 PM, git repository hosting wrote: >>>>> memset(&sa, 0, sizeof(sa)); >>>>> sa.sa_handler = SIG_IGN; >>>>> sigaction(SIGCHLD, &sa, NULL); >>>> >>>> Why not masking the signal here instead of ignoring it? So that the >>>> notifications which would happen during the short interval are not lost? >>>> >>>> >>> >>> We only want the child to be reaped instead of going zombie, we don't >>> actually care about the exit status. >> >> What if the application wants the exit status of another child process >> which would be created at the same time? >> > > This can't be, since the application has not started yet at this point. > Even if it loads xenomai libraries with dlopen? -- Gilles.