Date: Sun, 22 Apr 2018 16:55:13 +0200 From: Tijl Coosemans <tijl@FreeBSD.org> To: Konstantin Belousov <kostikbel@gmail.com> Cc: jeff@FreeBSD.org, "freebsd-current@freebsd.org" <freebsd-current@freebsd.org> Subject: Re: i386 hangs during halt "vnodes remaining... 0 time out" Message-ID: <20180422165513.711579c3@kalimero.tijl.coosemans.org> In-Reply-To: <20180422132855.GU6887@kib.kiev.ua> References: <YQBPR0101MB1042BF37F603335C9CE6346BDD8B0@YQBPR0101MB1042.CANPRD01.PROD.OUTLOOK.COM> <20180421234934.10d7dfab@kalimero.tijl.coosemans.org> <20180422120521.GS6887@kib.kiev.ua> <20180422151500.1608af96@kalimero.tijl.coosemans.org> <20180422132855.GU6887@kib.kiev.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 22 Apr 2018 16:28:55 +0300 Konstantin Belousov <kostikbel@gmail.com> wrote: > On Sun, Apr 22, 2018 at 03:15:00PM +0200, Tijl Coosemans wrote: >> Thanks for the review. There's just one concern I have. With this patch >> the bufspace_daemon threads appear to shutdown after the buf_daemon and >> after the syncer because the event handlers are registered later. Are >> there any dependencies between these processes that require the bufspace >> threads to be stopped earlier? > > I think for correctness bufdaemon must stop after the syncer, since syncer > operation can cause a situation where bufdaemon help is needed to proceed. > Other than this, the stop order is irrelevant, because after syncer > finished, there should be no any further filesystem activity. A quick way to do that would be to use SHUTDOWN_PRI_LAST + 100 for the event handlers in the patch, like shutdown_conf in kern_shutdown.c already does. Is that acceptable here as well?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20180422165513.711579c3>