Date: Sat, 25 May 2002 00:25:59 +1000 (EST) From: Bruce Evans <bde@zeta.org.au> To: Ruslan Ermilov <ru@FreeBSD.org> Cc: John Baldwin <jhb@FreeBSD.org>, <cvs-committers@FreeBSD.org>, <cvs-all@FreeBSD.org> Subject: Re: cvs commit: src/sys/i386/include cpufunc.h src/sys/kern kern_mutex.c Message-ID: <20020525001424.H3875-100000@gamplex.bde.org> In-Reply-To: <20020523073437.GC35465@sunbay.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 23 May 2002, Ruslan Ermilov wrote: > On Thu, May 23, 2002 at 05:21:49PM +1000, Bruce Evans wrote: > > <machine/cpufunc.h> is really a kernel-only header, but it sometimes gets > > abused in userland. I think this shouldn't be supported. It contains > > mostly very machine-dependent interfaces, so it should be used in at most > > the implementation of libvgl. libvgl uses i386 interfaces unconditionally, > > so it only works on i386's, but this limitation is not documented. > > > Would you like me to fix this by adding the share/man/man3/i386, and > installing this manpage to there? No, I think libvgl can/should be as machine-independent as syscons. libvgl seems to be mainly isa/vga dependent, and I don't want man3/isa :-). On alphas, inb()'s for isa/vga support are actually bus_space_read_1()'s on alphas, so they work in the kernel but not in libvgl. This leads back to the issue of how very machine-dependent interfaces like inb() should be exposed to applications. Everything in the kernel except the lowest levels should use the bus space interfaces instead of inb(), but bus space may be a bit much for applications. Bruce 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?20020525001424.H3875-100000>