Sigmund Augdal Helberg wrote: > > On Wed, 2006-01-11 at 15:07 +0100, Edgar Toernig wrote: > > > > Well it does improve that situation in so far as the card is unusable > > in my system without these changes. Without lowering the fifo trigger > > level (but with the other changes) I got it working but occasionally, > > especially when using the USB bus, I still got fifo overflows. With > > the lowered fifo trigger level that's gone. > > Is there some messages I can look for that appear when the fifo > overflows? That I perhaps could count with and without the change? You get something like this in the system log: | Jul 16 21:07:24 smutje kernel: bt878(0): irq FDSR risc_pc=21720110 If you've never seen that then you don't need the patch (but won't hurt neither). > If you are saying that without these changes your card doesn't work at > all, then that should be a strong argument for keeping card specific > stuff, because I've got quite a few cards working well without these > changes. It's not the card but the interaction between the Bt878 and the mother- board's PCI chipset. Some motherboards are fine and never show the problem, others only when using SATA DMA, others (like mine, ATI IXP) all the time. Ciao, ET.
