Date: Tue, 4 May 1999 07:59:55 -0400 (EDT) From: Dan Janowski <danj@3skel.com> To: Nate Williams <nate@mt.sri.com> Cc: Randy Bush <randy@psg.com>, "Sean O'Connell" <sean@stat.Duke.EDU>, FreeBSD Laptoppers <freebsd-mobile@FreeBSD.ORG> Subject: Re: PAO+ep0 slooooow Message-ID: <Pine.BSF.4.05.9905040751160.20499-100000@fnur.3skel.com> In-Reply-To: <199905031710.LAA03833@mt.sri.com>
next in thread | previous in thread | raw e-mail | index | archive | help
If this horse is not dead yet, I have a simple contribution. I just set up a 3c589c on a Dell Inspiron 3500. Ethernet traffic was BAD, would go out fine, but coming back it was obviously missing an interrupt. One thing is for SURE, FreeBSD is currently unable to detect, some or all, win-modems AND the interrupts they steal. Since most winmodems are parked on IRQ-3 and insertion will try to use the first available IRQ... well, forget about it. I knew for sure by this, when you eject the card, the system reports what IRQ it was on. When I noticed that mine was saying IRQ-3, I smelled a rat. I was able to quickly confirm by turning off the winmodem in the BIOS, and like magic it worked. The long term solution was to do this: ignirq 3 in the pccard.conf since I don't want to have to remember to turn the winmodem on or off if I use windows. Dan On Mon, 3 May 1999, Nate Williams wrote: > > >> pcic0: <Ricoh RL5C475 PCI-CardBus Brige> rev 0x00 int a irq 255 on pci0.10.0 > > >> ... > > >> pcic0 at 0x3e0-0x3e1 irq 11 on isa > > >> PC-Card ctlr(0) Ricoh RL5C475/476 [CardBus bridge mode] (5 mem & 2 I/O windows) > > >> pcic0: slot 0 controller I/O address 0x3e0 > > >> ... > > >> cmd ntpd pid 112 tried to use non-present sched_get_priority_min > > >> cmd ntpd pid 112 tried to use non-present sched_setscheduler > > >> ep0: utp/bnc[*UTP*] address 00:a0:24:ac:89:33 > > >> > > >> and ep0 is forced to int 10 by pccard.conf > > > > > > So, is it working? > > > > see original message. it works. problem is that it takes 15 hours to do an > > rdump when as zp0 pre-pao it took 45 minutes. i.e. this is a performance > > issue. > > Sounds like the interrupt is actually not being set, so you're actually > using the watchdog timer to transfer data, which creates a fake > interrupt 1/sec. > > Try hard-codeing the interrupt to something else in /etc/pccard.conf to > see if that helps. > > > > Nate > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-mobile" in the body of the message > -- Dan Janowski danj@3skel.com Triskelion Systems, Inc. Bronx, NY 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?Pine.BSF.4.05.9905040751160.20499-100000>