Re: Suspend/hibernate wireless regression with 2.6.33 kernel

[Chris Vine <chris@cvine.freeserve.co.uk>]

On Wed, 17 Mar 2010 01:51:13 +0000
Chris Vine <chris@cvine.freeserve.co.uk> wrote:
> Hi,
> 
> I have a Lenovo S12 (Ideapad) netbook with a BCM4312 802.11b/g
> [14e4:4315] wireless device.  Although this works using the PIO option
> in kernel 2.6.33/2.6.33.1 (not with DMA) using the in-kernel b43
> driver, and also works with the broadcom proprietary wl driver, it
> breaks after a suspend or hibernate. Attempts to bring up the wlan0
> interface with 'ifconfig wlan0 up' after suspension or hibernation
> results in the following message (although nothing is revealed by
> dmesg):
> 
>   SIOCSIFFLAGS: Unknown error 132
> 
> More worryingly, when suspending or hibernating, acpi sometimes
> but not always appears to write to the CMOS.  On two occasions I
> have observed that although a reboot into 2.6.33 would work (until
> there is another suspension or hibernation), wireless becomes
> permanently switched off when rebooting into 2.6.32 or earlier - when
> reading from the BIOS, kernel 2.6.32 and below appears to think that
> the radio has been disabled even though the BIOS set-up screen
> disagrees. When this occurs, I have to restore all defaults in the
> BIOS to get wireless to come on again in kernel 2.6.32 and less.

[snip]

This regression between 2.6.32 and 2.6.33 is still present in
2.6.34-rc2, with an added twist.  In 2.6.34-rc2, on resuming from
suspension or hibernation, my rfkill button is completely disabled -
that is, I cannot switch wireless on, but I also cannot switch bluetooth
off.

Chris