Date: Thu, 08 Apr 1999 21:09:07 +0900 From: "Daniel C. Sobral" <dcs@newsguy.com> To: NAKAGAWA Yoshihisa <y-nakaga@nwsl.mesh.ad.jp> Cc: Nate Williams <nate@mt.sri.com>, Ted Faber <faber@ISI.EDU>, Nick Sayer <nsayer@quack.kfu.com>, freebsd-mobile@FreeBSD.ORG Subject: Re: Any success with CirrusLogic 6729/6730??? Message-ID: <370C9C63.B5506516@newsguy.com> References: <199904080843.RAA02125@chandra.eatell.msr.prug.or.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
NAKAGAWA Yoshihisa wrote: > > - PCIC is an ISA device. PCIC handles bus, yet the bus is not the > bus PCIC belongs to. Please note that a PCIC is an ISA device and > a bridge between ISA bus and PC Card bus. Also, if you see codes > on NetBSD, bus dependent code and handling bridge is different. I just wish this PCIC is ISA/PCIC is not ISA would get settled... Can anyone on the not ISA camp *PROVE* it is not an ISA device? If it is not an ISA device, it ought to be possible to show a case A where an ISA device behaves in such an such way and PCIC does not. > - sysctl is the last way to go. It can't work unless the module is > loaded after boot. There are also other problems loading the > driver with current KLD. That is incorrect. It can be set through loader at load time. This is not hypothetical, it is the present state for PCIC. > - User configuration method dependent on KLD or boot loader is > wrong solution. It should be independent from them. Why? By the same token, I suggest that configuration method should be independent of anything you suggest. You did not justify your position, so I don't need to justify mine. So I suggest you justify your positions whenever you out-rule a possible solution. > - It is clearly wrong to specify the management irq by sysctl. Err... a lot of people don't see this way, so it might be wrong, but it sure is not clear. Justify. > - It is a bad way to handle resources by hard coding. A way for > the user to change the resources is must. UserConfig is the > available solution, now. Ideally, userconfig should go away. It relies on the kernel being running already, which is a bad thing. > - The method used by PAO is the right way to go, although native > probe/attach codes for PCI PCIC and CardBus controllers is yet to > be written. Justify. -- Daniel C. Sobral (8-DCS) dcs@newsguy.com dcs@freebsd.org "nothing better than the ability to perform cunning linguistics" To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-mobile" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?370C9C63.B5506516>