Date: Fri, 17 May 2002 10:02:08 -0400 (EDT) From: John Baldwin <jhb@FreeBSD.org> To: Peter Wemm <peter@wemm.org> Cc: Perforce Change Reviews <perforce@freebsd.org>, Julian Elischer <julian@elischer.org>, Jonathan Mini <mini@freebsd.org> Subject: Re: PERFORCE change 11120 for review Message-ID: <XFMail.20020517100208.jhb@FreeBSD.org> In-Reply-To: <20020517060856.CF499380A@overcee.wemm.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 17-May-2002 Peter Wemm wrote: > Jonathan Mini wrote: >> Peter Wemm [peter@wemm.org] wrote : >> >> > Jonathan Mini wrote: >> > >> > > The only problem he foresaw was that the init/fini functions could be >> > > cal > led >> > > by the pager daemon, but I don't see any problem there either. >> > >> > Somewhere along the way we were planning to put in code that checked for >> > things that *might* call tsleep() and trap mutexes being held. I dont >> > know >> > if the UMA stuff calls tsleep (directly or indirectly) or not, but it was >> > my understanding that it is a Bad Idea(TM) to call anything that can >> > tsleep >> > with a mutex held. >> > >> >> I think maybe you misunderstand. The problem isn't that the pager calls the >> uma init/fini functions, but rather that *my* init/fini functions may block >> inside the VM. > > Ah. Thanks for the clarification. If that block within VM is a tsleep block > instead of a mutex block then we probably do have problems.. Assuming uma > calls your init/fini functions with some of its locks held. Yes, I think that is the problem. I think it has to do with setting up/tearing down the thread stacks. If uma could do this w/o holding the zone locks that would probably be sufficient. -- John Baldwin <jhb@FreeBSD.org> <>< http://www.FreeBSD.org/~jhb/ "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe p4-projects" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?XFMail.20020517100208.jhb>