Date: Mon, 13 Jun 2011 16:09:52 -0600 From: "Kenneth D. Merry" <ken@FreeBSD.org> To: Kostik Belousov <kostikbel@gmail.com> Cc: svn-src-head@FreeBSD.org, svn-src-all@FreeBSD.org, "Justin T. Gibbs" <gibbs@FreeBSD.org>, src-committers@FreeBSD.org Subject: Re: svn commit: r223061 - head/sys/kern Message-ID: <20110613220952.GA47928@nargothrond.kdm.org> In-Reply-To: <20110613213150.GC48734@deviant.kiev.zoral.com.ua> References: <201106132121.p5DLL2Fp099153@svn.freebsd.org> <20110613213150.GC48734@deviant.kiev.zoral.com.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Jun 14, 2011 at 00:31:50 +0300, Kostik Belousov wrote: > On Mon, Jun 13, 2011 at 09:21:02PM +0000, Justin T. Gibbs wrote: > > Author: gibbs > > Date: Mon Jun 13 21:21:02 2011 > > New Revision: 223061 > > URL: http://svn.freebsd.org/changeset/base/223061 > > > > Log: > > Fix a couple of race conditions in devstat(9) initialization. > > > > In devstat_new_entry(), there is no need to initialize the queue > > and the mutex in this function. There are ways to do static > > initialization on both, so use STAILQ_HEAD_INITIALIZER and > > MTX_SYSINIT to initialize the queue and the mutex. > > > > In devstat_alloc(), use an atomic test and set routine to guard > > making our entry in /dev. Using just a plain static variable > > creates a race condition on multiprocessor machines. If you > > attempt to create a second entry in devfs, the kernel will panic. > Devfs returns an error if MAKEDEV_CHECKNAME flag is supplied and > attempt is made to create the existing node. The static guard is > still useful, since make_dev() call is costly, but you can remove > the atomic, since the race should be of limited scope. Done, thanks for the suggestion! Ken -- Kenneth Merry ken@FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110613220952.GA47928>