PCI 2.2 card hangs the system. Need driver writer's advice

From: Denis Vlasenko
Date: Tue Mar 09 2004 - 17:30:00 EST

Hi folks,

I'm going to tap on the collective experience of lkml
wrt dealing with hardware bugs and quirks.

I have a pair of SMC 2802w PCI 2.2 wireless NICs built
around Prism GT chipset (with ARM processor lurking inside).
There is a Linux driver for it at prism54.org.

I was going to deploy several 11g routers with these cards but
wanted to stress test them first. I fixed one easy bug but now
I've hit another one. Under bidirectional UDP flood + heavy
IDE write activity system freeze instantly and needs powercycling.
It turned out that card does something nasty to PCI bus
when driver does a readl() from PCI mapped space.
Even reset button does not return the card to sanity,
modprobe prism54 and you're toast.

All the details are at

I need to prevent this freeze by reprogramming card
and/or mainboard chipset.

Does anybody seen similar quirks before? Any ideas? Hints?
Does anybody have PDFs about my chipset? Here's the lspci:

00:00.0 Host bridge: VIA Technologies, Inc. VT8363/8365 [KT133/KM133] (rev 02)
00:01.0 PCI bridge: VIA Technologies, Inc. VT8363/8365 [KT133/KM133 AGP]
00:07.0 ISA bridge: VIA Technologies, Inc. VT82C686 [Apollo Super South] (rev 22)
00:07.1 IDE interface: VIA Technologies, Inc. Bus Master IDE (rev 10)
00:07.4 Host bridge: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] (rev 30)
00:07.5 Multimedia audio controller: VIA Technologies, Inc. AC97 Audio Controller (rev 20)
00:0a.0 Network controller: Harris Semiconductor: Unknown device 3890 (rev 01)
00:0f.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139 (rev 10)
01:00.0 VGA compatible controller: nVidia Corporation Vanta [NV6] (rev 15)


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/