#1 "Device or resource busy" on 2.6.8 + 2.6.10

[Anonymous]

Hi!

This sf.net account doesn't seem to be very active, but
before filing a bug against the kernel I'll give it a
try ;-)

Here's the problem:
Having recompiled the kernel my network card (onboard
nForce controller, using the forcedeth kernel driver)
won't start any more:

---
stefan@stefan:~$ ifconfig eth0 up
SIOCSIFFLAGS: Device or resource busy
---

I used the debian packaged kernel-source-2.6.8-11 but
get the same problem with 2.6.8-12 and 2.6.10-3 from
debian unstable. With 2.6.6 it works fine.

With debug messages on, dmesg gives:

modprobe forcedeth:
---
forcedeth.c: Reverse Engineered nForce ethernet driver.
Version 0.30.
ACPI: PCI interrupt 0000:00:04.0[A] -> GSI 10 (level,
low) -> IRQ 10
PCI: Setting latency timer of device 0000:00:04.0 to 64
0000:00:04.0: resource 0 start e0082000 len 1024 flags
0x00000200.
0000:00:04.0: MAC Address 00:10:dc:27:7d:94
0000:00:04.0: link timer on.
eth%d: mii_rw read from reg 2 at PHY 1: 0x15.
eth%d: mii_rw read from reg 3 at PHY 1: 0xf441.
0000:00:04.0: open: Found PHY 0540:003d at address 1.
eth%d: mii_rw read from reg 4 at PHY 1: 0x1e1.
eth%d: mii_rw wrote 0xde1 to reg 4 at PHY 1
eth%d: mii_rw read from reg 1 at PHY 1: 0x782d.
eth%d: mii_rw read from reg 0 at PHY 1: 0x3000.
eth%d: mii_rw wrote 0xb000 to reg 0 at PHY 1
eth%d: mii_rw read from reg 0 at PHY 1: 0x3000.
eth%d: mii_rw of reg 0 at PHY 1 failed.
eth%d: mii_rw read from reg 0 at PHY 1: 0x3000.
eth0: forcedeth.c: subsystem: 01462:373c bound to
0000:00:04.0
---

ifconfig eth0 up:
---
nv_open: begin
eth0: nv_alloc_rx: Packet 0 marked as Available
eth0: nv_alloc_rx: Packet 1 marked as Available
eth0: nv_alloc_rx: Packet 2 marked as Available
...
eth0: nv_alloc_rx: Packet 127 marked as Available
eth0: nv_txrx_reset
---

I tried to use an older driver version (0.25, works for
2.6.6) with 2.6.10: doesn't work. So it isn't sure that
it really is a forcedeth problem.

Networking in general is ok, eth1 (a Realtek) works fine.

Thx

Stefan