From: James L. R. <ji...@ho...> - 2005-02-28 09:37:08
|
On Mon, 28 Feb 2005 03:58:38 +0100 (MET) Peter Karlsson <pet...@st...> wrote: > > Could this be a timing/order problem? I.e. if the registers has to be > "poked" in a certain order or within a certain time... Could be. According to the i2c dumps, the windows code seems to stop and start the firmware a lot. The linux driver is written to allow this but the mode change routines seem to follow a "stop firmware...change a lot of stuff...start it back again" pattern. This is what I'm used to in hardware drivers, but it could very well be an ordering or timing thing. Since the other conexant parts seem not to exhibit these symptoms, I'm skeptical, but anything's possible--and I'm certainly no expert! ----- Jim Reese, WD5IYT ji...@ho... |