Date: Mon, 16 Jul 2012 01:28:16 -0700 From: Colin Percival <cperciva@freebsd.org> To: Konstantin Belousov <kib@FreeBSD.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r237847 - head/usr.bin/killall Message-ID: <5003D0A0.3050209@freebsd.org> In-Reply-To: <201206301636.q5UGaNCB029813@svn.freebsd.org> References: <201206301636.q5UGaNCB029813@svn.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 06/30/12 09:36, Konstantin Belousov wrote: > New Revision: 237847 > URL: http://svn.freebsd.org/changeset/base/237847 > > Log: > Once in a month, when the moon is full, killall mistakenly considers > living process as a zombie and refuses to kill it. The cause is that > the code masks ki_stat with SZOMB to compare with SZOMB, but ki_stat > is not a mask. > > Possibly reported by: cperciva For the archives: Judging by the frequency with which I was seeing killall failing to send a signal and the fact I haven't seen it happen since I applied this patch, I'm 99.9999% confident that this was indeed the bug I was running into. Hopefully everybody else is sending signals to specific pids rather than using killall during log rotation like I am. :-) -- Colin Percival Security Officer Emeritus, FreeBSD | The power to serve Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?5003D0A0.3050209>