Date: Fri, 29 May 1998 23:19:38 -0600 (MDT) From: "Justin T. Gibbs" <gibbs@narnia.plutotech.com> To: Doug Rabson <dfr@nlsystems.com> Cc: current@FreeBSD.ORG Subject: Re: FreeBSD/alpha status report (2) Message-ID: <199805300519.XAA02620@narnia.plutotech.com>
next in thread | raw e-mail | index | archive | help
> Along the way, I have successfully probed for simulated devices on a > simulated PCI bus and attached simulated SCSI disks :-). I am *not* > currently using NetBSD's bus_space stuff to handle accesses to device i/o > ports and memory. Given that 99% of the machines that the port will work > on don't need the complexity of bus_space, I have taken the Linux route > and each chipset will supply versions of inb etc which perform the > relavent contortions. I suppose I don't understand the rational here. The i386 port doesn't have to go through any contortions in it's bus space implemenation and implementing bus space for FreeBSD x86 (look in i386/include/bus.h) was *trivial*. So why not use bus space? The CAM drivers already use it, you say that 99% of Alphas can use a "simple" implementation, and it buys us the ability to more easily port code from NetBSD? Just because you seem to believe that NetBSD's implementation of the bus space and bus DMA interfaces for Alpha are overly complex, doesn't mean that the interfaces themselves are a bad idea. -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805300519.XAA02620>