Date: Mon, 30 Apr 2001 08:28:58 +0200 From: Mark Murray <mark@grondar.za> To: Bruce Evans <bde@zeta.org.au> Cc: smp@FreeBSD.ORG Subject: Re: sys/mutex.h sys/lock.h (and other) cleanups. Commit Candidate #2 Message-ID: <200104300627.f3U6Rdp31851@gratis.grondar.za> In-Reply-To: <Pine.BSF.4.21.0104301046560.22502-100000@besplex.bde.org> ; from Bruce Evans <bde@zeta.org.au> "Mon, 30 Apr 2001 11:23:08 %2B1000." References: <Pine.BSF.4.21.0104301046560.22502-100000@besplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> On Sun, 29 Apr 2001, Mark Murray wrote: > > > Please review > > > > http://people.freebsd.org/~markm/patches/sys.SYS_MUTEX.diff.3 > > > > Nothing has changed much since the last Commit Candidate; I have > > just merged in some big commits. > > > > Reminder - the primary purpose of these is to untangle the requirement > > that "major" system includes (<sys/*.h> files) have on sys/lock.h > > and sys/mutex.h. Also included is a further deprecation of > > sys/lockmgr.h (by including it in sys/lock.h). Lesser purposes of > > Actually, the deprecation consists of including <sys/lockmgr.h> instead > of <sys/lock.h> in some headers that only need the former. <sys/lock.h> > is not such a header, but it still includes <sys/lockmgr.h> although this > is more bogus than before. The verbose comment before this include has > been copied to to many places (more than 0). That is based on something you asked me to do? > > this include sorting of the sys/*.h includes in affected files, > > and fixing of the problems that these sorts (may) create. > > Please use separate commit(s) to sort the files. Committing these first > is best I guess. OK - I can do this, except that a) this lock/mutex dethreading is going to affect it (in some cases significantly), and b) its going to delay (perhaps by a month) the whole commit (due to other (social) pressures). > The differences between SYS_MUTEX.diff.2 and SYS_MUTEX.diff.3 are > about half for moving around bogus includes of <net/radix.h> and > <sys/socket.h>. They were included recently to work around breakage > of <sys/mount.h> and only about 90% of them have been backed out so > far. At the time they were included, they had to be included before > <sys/mount.h> and sorting them would been wrong. Now sorting them > has no effect (except to clean things up) because including them has > no effect (except to trigger bogometers). OK - I'll clean them out too. > Have you tested this for alphas? Earlier I did. Before the commit, it will be. (Build, not run). > Committing this is taking too long. At least we avoided <net/radix.h>- > like spam (I hope). Agreed. :-( See above. M -- Mark Murray Warning: this .sig is umop ap!sdn To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200104300627.f3U6Rdp31851>