Yes, it was in the KeStallExecutionProcessor(). I was having verry
searous performance problems with the default delay. There would be
very noticble lag on screen updates, keyboard, and mouse input.
with the shorter delay, it seems like the system speed is back to normal
without ndiswrapper loaded.
Giridhar Pemmasani wrote:
>I have been using BCM94306 without any performance problems (since 0.2
>or so). Moreover, KeStallExecutionProcessor (I hope that is what you
>tweaked) is supposed to busy-wait and NDIS description for this
>function says that the delay must be less than 50 us. Replacing it or
>reducing the delay is not correct (although it may work for you).
karl at hiramoto dot org