From owner-freebsd-mobile Mon Aug 7 2:12:55 2000 Delivered-To: freebsd-mobile@freebsd.org Received: from telinco.net (internal.mail.telinco.net [212.1.128.4]) by hub.freebsd.org (Postfix) with ESMTP id D188137BD50 for ; Mon, 7 Aug 2000 02:12:51 -0700 (PDT) (envelope-from b.candler@pobox.com) Received: from [212.74.96.2] (helo=bloodhound.uk.worldonline.com) by telinco.net with esmtp (Exim 3.02 #1) id 13Lixm-0000fF-00; Mon, 07 Aug 2000 10:12:50 +0100 Received: from brian by bloodhound.uk.worldonline.com with local (Exim 3.15 #1) id 13Lixh-0003f9-00; Mon, 07 Aug 2000 10:12:45 +0100 Date: Mon, 7 Aug 2000 10:12:45 +0100 From: Brian Candler To: Warner Losh Cc: freebsd-mobile@FreeBSD.ORG, shigeru@iij.ad.jp Subject: Re: Ricoh RL5C475 PCI-PCMCIA adaptor and interrupts Message-ID: <20000807101245.C13919@linnet.org> References: <20000806204051.A805@linnet.org> <200008062052.OAA26562@harmony.village.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 1.0.1i In-Reply-To: <200008062052.OAA26562@harmony.village.org>; from imp@village.org on Sun, Aug 06, 2000 at 02:52:12PM -0600 Sender: owner-freebsd-mobile@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Sun, Aug 06, 2000 at 02:52:12PM -0600, Warner Losh wrote: > : (2) Worse, what if there are other boards based on the Ricoh RL5C475 which > : need the bit to be set? It would be hard or impossible to determine this > : from the PCI ID, so you'd have to make it a user-settable flag :-( > > Yes. That's going to make things tough. In the future, we'll not > kick these cards into legacy '365 emulation mode, so a temp hack like > this is OK for now. However, I don't think this hack works :-(. > > I think you'll have to check to see if an interrupt is allocated to > the card and do this trick if one isn't. My laptop has a 475 in it > and it works 100% of the time w/o this bit set, so something better > must be done. I'm fairly certain that enabling this code on my laptop > would make it fail 100% of the time (since I recall having to add it > to make it work there). We may have a clash of terminology; this bit is actually set to 1 in the existing (unmodified) code, but I had to clear it for my board. So, in your laptop, does the 475 appear as a PCI device? But then they have wired the ISA interrupt pins directly to the ISA interrupt controller? Yeuch. In that case, we would need a clean way to distinguish between: (1) a 475 which is actually sitting on a PCI card (2) a 475 which is directly connected and as you say, short of probing interrupts, that's difficult to do. BTW, a note in the Linux pcmcia-cs-3.1.19 code says that interrupt probing for Ricoh chips seems to be unreliable, and it is disabled (except for checking for a stuck-on interrupt) > What does dmesg say for you when you boot? I'm interested only in the > pcic-pci* line(s). I'll dig this out when I'm next in front of the machine. It said something along the lines of Ricoh 5C475 at slot 8.0 on irq 11. Cheers, Brian. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-mobile" in the body of the message