Bug#290039: 3c59x problems after kernel upgrade from 2.4.27-2 to 2.4.27-6
Package: kernel-image-2.4.27-1-686
Version: 2.4.27-6
After upgrading a kernel-image-2.4.27-1-686 kernel on a Debian sarge
system from 2.4.27-2 to 2.4.27-6, the 3c59x ethernet driver was no longer
working.
Installing a 2.6.8 kernel fixes the problem.
The problem may be at least partially within the PCI subsystem.
The system is an IBM thinkpad T20 laptop with an 3C556B Mini-PC
ethernet/winmodem card.
The principle symptom of the bug is that the MAC address, product code and
revision and revision date reported by the driver are have all bits set.
00:03.0: 3Com PCI 3c556B Laptop Hurricane at 0x1400. Vers LK1.1.18-ac
PCI: Setting latency time of devive 00.03.0 to 64
ff:ff:ff:ff:ff:ff, IRQ 11
product code ffff rev ffff.15 date 15-31-127
and the eth0 device fails with messages like
eth0: command 0x5800 did not complete! Status=0xffff
(similarly for command 0x2804, 0x3002)
It would appear that whatever the driver is sending, the response always
has all bits set.
Other possible symptoms include messages such as
ACPI-1133: *** Error: Method execution failure [\.SB.LNKA._STA]
(Node c11866a0), AE_NOT_EXIST
Similar problems have been reported in the past
http://www.spinics.net/lists/kernel/old/2003-q1/threads.html#22099
These have usually been attributed to ACPI and APM or to the machine
having run Windows at some point. These problems persist if ACPI and APM
are disabled and the laptop never runs Windows...
- Jason
--
Jason Lander jason@env.leeds.ac.uk
Reply to: