Date: Thu, 08 Feb 2001 21:12:04 -0700 From: Warner Losh <imp@harmony.village.org> To: Jonathan Chen <jon@FreeBSD.ORG> Cc: John Baldwin <jhb@FreeBSD.ORG>, freebsd-mobile@FreeBSD.ORG, Jose Gabriel J Marcelino <gabriel@maquina.com> Subject: Re: Aironet under NEWCARD Message-ID: <200102090412.f194C4955701@harmony.village.org> In-Reply-To: Your message of "Thu, 08 Feb 2001 15:52:10 EST." <20010208155210.A93326@spock.org> References: <20010208155210.A93326@spock.org> <200102050642.f156g6971659@harmony.village.org> <XFMail.010205143119.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <20010208155210.A93326@spock.org> Jonathan Chen writes: : Under the current implementation of newcard, 16-bit pcmcia cards will get : the bridge's interrupt if it is hooked up to a pccbb (as opposed to : pcic) bridge. This is because when I hacked up this mess, I a) was unsure : of how to do pci-isa irq routing on the pccbb, and b) saw that there : appeared to be a way for 16-bit pcmcia cards to share PCI interrupts : through checking the card's ccr status register. This should probably : still work under isa bridges as the bridge would give the pccard bus a : different interrupt, and the bus would share that amongst the functions on : the same card. There's issues with sharing on ISA cards. I've been through them and you can kinda sorta do it if you use pulse level interrupts. But then things get very unreliable. And sometimes you get hangs anyway. Forget about the idea of sharing on the ISA bus. Maybe the bridge could multiplex the interrupt with the management irq and with its sister card in the other slot, but that's about it. And there are issues with interrupt ordering with OLDCARD, which is very fussy about attempts to do sharing. Maybe with NEWCARD we can do better, but that's still dicy. Warner 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?200102090412.f194C4955701>