rt2500pci BUG schedule while atomic
I've been unable to boot rt2x00.git kernels lately because I get BUG messages that go something like
BUG scheduling while atomic NetworkManager 0x10001010000
affected kernels 2.6.25-wl-31379-g64e4500 and 2.6.26-rc1-wl-39106-g311fd1d
I'll try booting with NetworkManager disabled to see how far I can get or whether I'm able to pin down where exactly the bug occurs.
Could you post the stacktrace that should be printed to the log?
Sorry, so far I have been unable to get any log whatsoever
As soon as I [code2vmtqlcq]ifconfig wlan0 up[/code2vmtqlcq], my machine gives two messages
[quote2vmtqlcq]BUG scheduling while atomic ifconfig/1710/0x10010000
BUG scheduling while atomic ifconfig/1710/0x00010000
and then freezes up hard. Not even the MagicSysRq keys work anymore
Well I'll see if I can reproduce this, but seeing how at least rt61pci was unaffected on my machine, this might become tricky.
I seem to be getting this too with latest git.
On top of the call trace, I get something along the lines of
... rt2500pcirt2500pci_config_intf +0x24/0x140
I'll try to get some netconsole stuff going later in order to get the full warn text.
And here's a bit of output of a netconsole session. Haven't had time to look at it myself yet, though...
Fixed in latest rt2x00.git (patch "rt2x00 Use atomic interface iteration in irq context")
Fixed indeed, Thanks