Date: Tue, 20 Jun 2006 16:20:37 -0600 From: Scott Long <scottl@samsco.org> To: Maxim Konovalov <maxim@macomnet.ru> Cc: cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/share/man/man9 lock.9 Message-ID: <449874B5.50909@samsco.org> In-Reply-To: <20060621020909.N56083@mp2.macomnet.net> References: <200606202141.k5KLfETG075895@repoman.freebsd.org> <20060621014634.U55744@mp2.macomnet.net> <44987135.1070007@samsco.org> <20060621020909.N56083@mp2.macomnet.net>
next in thread | previous in thread | raw e-mail | index | archive | help
Maxim Konovalov wrote: > On Tue, 20 Jun 2006, 16:05-0600, Scott Long wrote: > > >>Maxim Konovalov wrote: >> >>>On Tue, 20 Jun 2006, 21:41-0000, Maxim Konovalov wrote: >>> >>> >>> >>>>maxim 2006-06-20 21:41:14 UTC >>>> >>>> FreeBSD src repository >>>> >>>> Modified files: >>>> share/man/man9 lock.9 >>>> Log: >>>> o Remove LK_REENABLE and LK_NOPAUSE lockinit(9) flags, add LK_NOSHARE. >>> >>> >>>Btw, as I see there are not many consumers of lock.9 infrastructure >>>comparing to mutex.9, sx.9 etc in our tree. Is it something derecated? >>> >> >>I guess that VFS doesn't count as being important? > > > As an average user I just want to know what pros and contras for using > lock.9 vs all other locking primitivies. It is not clear from our man > pages and arch book. Sorry if I'm asking something completely stupid. > lockmgr is somewhat analogous to an SX lock. However, due to some of the special considerations needed by VFS in terms of the vnode life cycle and blocking for I/O, they aren't completely functionally equivalent to an SX lock. It would be nice to fix VFS to not need these special considerations, but that is a very complicated and risky task. Scott
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?449874B5.50909>