Skip site navigation (1)Skip section navigation (2)
Date:      10 Feb 1998 13:01 EST
From:      "Andrew Atrens" <atrens@nortel.ca>
To:        hackers@FreeBSD.ORG
Subject:   pci-ide-dma side effects
Message-ID:  <199802101859.KAA19606@hub.freebsd.org>

next in thread | raw e-mail | index | archive | help
Hi All,

I'm running "1/30/98 - current", on an AMD-K6/200 w. ASUS TP97 mobo.

Last night curiousity got the best of me, and I was tempted to try out the
wd driver's DMA support.

To do this I changed my kernel config from:

controller wdc0 at isa? port "IO_WD1" bio irq 14 flags 0x80ff80ff vector wdintr

to:

controller wdc0 at isa? port "IO_WD1" bio irq 14 flags 0xa0ffa0ff vector wdintr


To make a long story short, I launched the new kernel, DMA *was* detected:

Probing for devices on PCI bus 0:
chip0: <Intel 82439TX System Controller (MTXC)> rev 0x01 on pci0.0.0
chip1: <Intel 82371AB PCI to ISA bridge> rev 0x01 on pci0.1.0
ide_pci0: <Intel PIIX4 Bus-master IDE controller> rev 0x01 on pci0.1.1
chip2: <Intel 82371AB USB host controller> rev 0x01 int d irq 255 on pci0.1.2
chip3: <Intel 82371AB power management> rev 0x01 on pci0.1.3
vga0: <ATI Mach64-GX graphics accelerator> rev 0x03 on pci0.9.0

wdc0 at 0x1f0-0x1f7 irq 14 flags 0xa0ffa0ff on isa
wdc0: unit 0 (wd0): <ST52520A>, DMA, 32-bit, multi-block-16
wd0: 2446MB (5009760 sectors), 4970 cyls, 16 heads, 63 S/T, 512 B/S 


launched X, then ran `Bonnie'. 

What happened next is hard to describe. :)

My X screen was filled with (for want of a better description) `marching ants'.
It was as though the screen was `swimming'. I suspended `Bonnie' and the
swimming stopped. :)  Lots of garbage on the screen, so I did a screen refresh
and everything went back to normal. I resumed `Bonnie' and swimming started 
again on cue. Then my monitor appeared to lose the graphics card sync signal
and began behaving very strangely, then it went into power-save mode and
shut down. I re-suspended `Bonnie', did some CTRL-ALT-F? things and back it
came! I killed and re-started Bonnie again, quickly switched to a cons screen,
started `top' and waited for Bonnie to complete. I then unmounted and fsck'd
the filesystem - it was clean as a whistle! And so ended the experiment.
Needless to say, I've backed away from ide DMA for now. :)

Comments?  Have I done something glaringly stupid here to expect this result? :)

(The purpose of this post, actually, is to provide some feedback to the
 IDE DMA folks - to hopefully help them find a bugster, if one exists. I am
 willing to be a guinea pig if someone requires more data about this
 anomoly. If hackers is the wrong forum for such a post please let me know so
 that I may correctly address future sightings. )

No flames please.

Andrew.
(opinions mine, not Nortel's)

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199802101859.KAA19606>