On Fri, 2005-01-14 at 12:09 -0500, Dan Williams wrote: > I think the reason you may be having problems now is that I took out > lots of code that took cards down and then up at various points to clear > out settings. This code had two problems: > > 1) Caused lots of hotplug events because cards with loaded-firmware > (prism54, atmel, etc) have to re-load the firmware whenever you bring > the device "down". > 2) Caused kernel panics on some cards (prism54 and Cisco Aironet) in > some ipv6 code, or at least greatly aggravated them. > > The main reason for this was #2. Unfortunately, that means that this > problem occurs much more often on prism54 cards than it used to... I > would suggest talking to the prism54 people about this. Sorry to still not be convinced: When I /don't/ do anything (not taking down and the up again) of my card I don't have any problem. Now you say, old NM did this at various (regular? I don't know exactly what you meant) points in time to prevent lockups, something new NM does not do any longer. You say that doing nothing anymore means that the problems the card has anyway will now shine through. But why is the card running without a problem without NM running at all? I never had a single problem with it the whole year I had it running without knowing about NM at all... Sorry to bother you again with this, jtl
Attachment:
signature.asc
Description: This is a digitally signed message part