On Wed, 2002-03-20 at 18:18, Frank Trenkamp wrote: [on module loading order] > However, with kernel 2.4.18 this problem just disappeared. A good soul named > Michal Ceresna (his credits here) sent me a patch as feedback to an > installation report I published via www.linux-on-laptop.net: > > That's it: [...] > ! dir |= 0x10; /* assuming pci bus master? */ [...] > As you can see, the changes are minimal. With the stock 2.4.18 kernel this > patch was already integrated; now it does not matter in which order the irda > modules or maestro3 have been loaded. It just works.. Indeed this is correct, I have just verified that the order of module loading no longer important in 2.4.18 in this case. Loading maestro3 before irtty is just fine with this kernel version. In previous trials I recall that a cold boot was required to get irda and maestro3 working - if you just rebooted normally you would lose irda regardless of module loading order. This probably fixes that as well. What confused me was that I tried to modprobe smc-ircc, and then I stopped trying to get FIR working, and I did what you should never do: I changed two things at once. I started loading irtty instead of smc-ircc, but at the same time I moved it in /etc/modules to be loaded before maestro3. So I tricked myself into believing that the last change was the fix... :P Thanks for the explanation Frank! Cheers, fabbe
Attachment:
signature.asc
Description: This is a digitally signed message part