Date: Sat, 25 Oct 2014 15:28:22 -0400 From: John Baldwin <jhb@freebsd.org> To: Andrey Chernov <ache@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r273601 - head/lib/libc/gen Message-ID: <94327674.WeLfOcjE9r@ralph.baldwin.cx> In-Reply-To: <544B1C34.1020204@freebsd.org> References: <201410241955.s9OJtnbm014234@svn.freebsd.org> <544B0930.8010004@freebsd.org> <544B1C34.1020204@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Saturday, October 25, 2014 07:42:44 AM Andrey Chernov wrote: > On 25.10.2014 6:21, Andrey Chernov wrote: > > On 24.10.2014 23:55, John Baldwin wrote: > >> Author: jhb > >> Date: Fri Oct 24 19:55:49 2014 > >> New Revision: 273601 > >> URL: https://svnweb.freebsd.org/changeset/base/273601 > >> > >> Log: > >> Don't reference sem(4) from the POSIX semaphore pages. POSIX > >> semaphores > >> were reimplemented using umtx in FreeBSD 9 and no longer use sem(4). > > > > sem kernel module referenced in sem(4) uses the same uipc_sem.c file as > > kernel's P1003_1B_SEMAPHORES option, so they both use the same code and > > better be cross-linked since sem(4) explains how to turn it on. > > From the first glance I can't determine, is libc implementation better > than kernel one or not. BTW, firefox port still recommends sem_load="YES" The sem_init/open/unlike/destroy/post/*wait() API used to use sem(4) (which is sem.ko and the kernel option) from FreeBSD 5 up through FreeBSD 8. These functions have used umtx_op() instead of sem(4) since FreeBSD 9.0. Note that libc/gen/sem_new.c does not call ksem_*. -- John Baldwin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?94327674.WeLfOcjE9r>