Date: Mon, 15 Jun 1998 09:58:52 -0600 From: Nate Williams <nate@mt.sri.com> To: Jun-ichiro itojun Itoh <itojun@iijlab.net> Cc: Michael Hancock <michaelh@cet.co.jp>, Bill Fenner <fenner@parc.xerox.com>, Donald Burr <dburr@POBoxes.com>, Mike Smith <mike@smith.net.au>, bjc23@hermes.cam.ac.uk, freebsd-current@FreeBSD.ORG, Frank Mayhar <frank@exit.com> Subject: Re: Auto power-off? Message-ID: <199806151558.JAA04217@mt.sri.com> In-Reply-To: <11163.897801516@coconut.itojun.org> References: <Pine.SV4.3.95.980614115946.8470A-100000@parkplace.cet.co.jp> <11163.897801516@coconut.itojun.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> >> After doing the appropriate ioctl, the APM in 2.2.6-PAO shuts off my > >> Sharp Widenote when I halt the system. I understand that APM in PAO > >> is fairly different from that in 2.2.6; I couldn't find any explicit > >> "turn off on halt" code in 2.2.6-RELEASE. > > APM code in 2.2.6 does not conform to (some part of) APM BIOS spec, > because in the past some report said there are some BIOS that does > not conform to the spec. Actually, you've got your facts backwards. The APM code in PAO does not conform to the specification because of bugs in the BIOS, but the code in the releases have. (Although, I haven't reviewed PHK's recent APM 1.2 additions, which I believe are in violation of the spec. since we don't support 1.2, just claim to support it.;( ) > I believe APM code has to be updated to conform to the spec, and we > should separate workarounds as "workarounds" (optional, maybe > enabled by device flags), not mainstream code. This is exactly my problem with the PAO code, which makes *everything* a work-around. Unfortunately, for the average user it's not obvious which of the 5-6 different workarounds to use, and some workarounds break existing hardware. IMHO, I believe many of the workarounds are workarounds broken-ness in some of the PAO extensions that can be fixed correctly so that the code works in agreement with the specification. In the past, I spent a large amount of time 'removing workarounds' introduced in PAO and replaced them with 'fixes'. That is why I know the FreeBSD code is/was not in violation of the specification. Nate > The largest difference between PAO and wildboar is implementation of > socket/card service. While PAO implements it as pccardd (daemon), > wildboar implements it in kernel. PAO used the code that was in FreeBSD written Andrew McRae (of Cisco-Australia), so this isn't a PAO specific item. Are there any pointers to the WildBoar code? Nate 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?199806151558.JAA04217>