On Tuesday, 6 April 1999 at 15:44:51 +0100, David Malone wrote: > I'm having trouble with a kernel built from this mornings make world. > It seems to be related to the new SiS 5591 ide chipset support. It > gets as far as the automatic reboot in progress and then says: > > wd0: interrupt timeout (status 50<rdy,seekdone> error 0) > wd0: wdtimeout() DMA status 0 > wd0: interrupt timeout (status 50<rdy,seekdone> error 1<no_dam>) > wd0: wdtimeout() DMA status 0 > wd0: interrupt timeout (status 50<rdy,seekdone> error 1<no_dam>) > wd0: wdtimeout() DMA status 0 > wd0: interrupt timeout (status 50<rdy,seekdone> error 1<no_dam>) > wd0: wdtimeout() DMA status 0 > . > . > . > > The first thing is that I guess the <no_dam> should be <no_dma>,
My understanding is that it stands for "no damage". > the second is that it seems to sit there doing that for longer than > I was willing to wait (it eventually prints a message saying that it > presumes it is a laptop and shouldn't print any more of these > messages). Right, it's timing out on something. > Backing out the last change to ide_pci.c seems to fix the problem. > I've lots of flags turned on in the kernel config file (0xa0ffa0ff), > the old kernel complains a little but works fine, the new kernel > spots that it is a SiS 5591, but grinds to a halt. Interesting. What motherboard do you have? Does it also have a 5595 on board? > If it helps any, the older working kernel prints one timeout message > when booting after the bad kernel has been booted, and my root > filesystem is actually on wd1. Is my drive/controller not up to the > flags I've set, or is this a problem with the new chipset support? One way or another it's a problem with the new chipset support if it makes a previously working system no longer work. > In config file: > --------------- > controller wdc0 at isa? port "IO_WD1" bio irq 14 flags 0xa0ffa0ff > > The old kernel says: > -------------------- > ide_pci0: <PCI IDE controller (busmaster capable)> rev 0xc1 int a irq 14 on > pci0.1.1 > ide_pci: generic_dmainit 01f0:0: warning, IDE controller timing not set > > The new kernel reports this: > ---------------------------- > ide_pci0: <SiS 5591 Bus-Master IDE controler> rev 0xc1 int a irq 14 on > pci0.1.1 Hmm. This is an older version than mine. I have: ide_pci0: <SiS 5591 Bus-master IDE Controller> rev 0xd0 int a irq 14 on pci0.0.1 chip1: <SiS 85c503> rev 0x01 on pci0.1.0 chip2: <PCI to PCI bridge (vendor=1039 device=0001)> rev 0x00 on pci0.2.0 Do you have the other chip reports as well? > New kernel with -v: > ------------------- > SiS 5591 dmainit: primary drive 0 setting ultra DMA mode 2 > wd0: wdsetmode() setting transfer mode to 42 > SiS 5591 status: CRTC 12 PCICLK, CATC12 PCICLK, applies to all IDE drives > SiS 5591 status: burst cycles enabeled, fast post write control enabeled > SiS 5591 status: primary drive 0 DRTC 1 PCICLK, DATC 3 PCICLK > SiS 5591 status: primary drive 0 Ultra DMA enabeled, 1PCICLK data out > SiS 5591 status: primary drive 0 postwrite enabeled, prefetch enabeled > prefetch count is 512 > SiS 5591 status: primary drive 0 has been configured for DMA > wdc0: unit 0 (wd0) <ST32122A>, DMA, 32-bit, multi-block-16 > wd0: 2014MB (4124736 sectors), 4092 cyls, 16 heads, 63 S/T, 512 B/S > wd0: ATA INQUIRE valid = 0007, dmaword = 0007, apio = 0003, vdma = 0407 OK. Can I assume that you have only one drive on the machine? Have you tried a smaller DMA transfer size (say, to start with, 0xa001a001)? Greg -- See complete headers for address, home page and phone numbers finger g...@lemis.com for PGP public key To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-current" in the body of the message
