Andy Green wrote: > I'm not sure faking module removal is ideal way but whatever you can get > working will be good.
I think what I'm doing now is basically how the Linux SDIO stack wants this to work: MMC suspend equals MMC card removal and re-insertion. I don't fake the full module removal on suspend, I'm just telling the ar6k stack that the device is gone. We'll have to revisit this concept if we want to do things with wakeup on wireless. > entirely in the stack with bringup after resume. But the Atheros stack > did suspend and resume OK so there must be a way to signal to the AR6001 > driver end that it is suspended, maybe it's worth comparing the actions > into the AR6001 from the Atheros stack. >From what I see, the Atheros SDIO stack seems to be pretty light on what it does on suspend. But I'm actually trying to avoid digging very deep there. Anyway, I saw that s3cmci doesn't like suspend/resume even without loading anything WLAN-related, and oopses on insert/remove/insert. So I think I can stay on familiar ground for this chase :-) - Werner
