Re: e1000 crash on 3.15.0-rc4

[Ben Hutchings]

On Wed, 2014-05-07 at 13:32 +0800, Liu Bo wrote:
> Hi,
> 
> I'm using qemu to load kernel of 3.15.0-rc4, and with CONFIG_E1000=y,
> everying is ok, but with CONFIG_E1000=m, I got the following crash:
> 
> [...]
> [    6.020409] tsc: Refined TSC clocksource calibration: 2491.912 MHz
> [    6.074925] floppy: Unknown symbol mutex_lock (err 0)
> [    6.128579] virtio_blk: Unknown symbol mutex_lock (err 0)
> [    6.141264] i2c_core: Unknown symbol mutex_lock (err 0)

Really???

> [    6.167490] e1000: Intel(R) PRO/1000 Network Driver - version 7.3.21-k8-NAPI
> [    6.168240] e1000: Copyright (c) 1999-2006 Intel Corporation.
> [    6.171078] ACPI: PCI Interrupt Link [LNKC] enabled at IRQ 10
> [    6.190295] BUG: spinlock bad magic on CPU#3, systemd-udevd/208
[...]

I think you may have rebuilt the kernel with lockdep enabled, but
haven't also rebuilt and installed the modules with lockdep enabled.

You should enable symbol versions to catch such mistakes.

Ben.

-- 
Ben Hutchings
Sturgeon's Law: Ninety percent of everything is crap.

Attachment: signature.asc
Description: This is a digitally signed message part