Date: Fri, 8 Dec 2000 17:24:11 -0800 From: Alfred Perlstein <bright@wintelcom.net> To: John Baldwin <jhb@FreeBSD.org> Cc: cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org Subject: Re: cvs commit: src/sys/sys mutex.h Message-ID: <20001208172411.W16205@fw.wintelcom.net> In-Reply-To: <XFMail.001208164811.jhb@FreeBSD.org>; from jhb@FreeBSD.org on Fri, Dec 08, 2000 at 04:48:11PM -0800 References: <20001208164153.V16205@fw.wintelcom.net> <XFMail.001208164811.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
* John Baldwin <jhb@FreeBSD.org> [001208 16:48] wrote: > > On 09-Dec-00 Alfred Perlstein wrote: > > * John Baldwin <jhb@FreeBSD.org> [001208 16:14] wrote: > >> jhb 2000/12/08 16:14:44 PST > >> > >> Modified files: > >> sys/sys mutex.h > >> Log: > >> Remove a comment that referrred to the obsolete mtxf struct. > > > > So with or without debug we waste space? or don't use any extra space? > > We don't use space. With debug, we (ab)use the description pointer to point to > a debug structure, and store the description pointer in that struct, with a > suitable #define of mtx_description. :) Due to this, you have to use a special > MUTEX_DECLARE() macro to declare mutexes initialized before malloc() is working > that statically allocates the appropriate mtx_debug and makes the mtx point to. > (Along with a special MTX_COLD flag for mtx_init() on these mutexes so that > mtx_init knows not to malloc() a debug struct but to use the one already > pointed to.) It's not the prettiest solution, but mtxf was uglier. :) I like this a lot better than mtxf. -- -Alfred Perlstein - [bright@wintelcom.net|alfred@freebsd.org] "I have the heart of a child; I keep it in a jar on my desk." To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20001208172411.W16205>