On Sun, Dec 30, 2001 at 06:44:06PM +0000, Philip Blundell wrote: > I think there are a couple of gotchas if you want to use the vidc.c driver > on a 7500.
I've always wondered why you meed to modify the driver. It's fully functional as it stands. > First up, it wants to reserve a DMA channel even though it isn't actually > going to use it for anything. Yes, we could drop the "virtual sound dma channel" since we don't actually use the DMA support functions. > At some point I also had an "outb(DMA_CR_E | 0x10, IOMD_SD0CR);" in > attach_vidc(). Eww, that means you're eating system bus bandwidth while the driver is loaded, and at high sample rates this is a _significant_ proportion of the available bus bandwidth. The correct way the DMA is started is when the audio driver calls the "trigger" method. If you're having problems with the existing driver as it stands, you might like to send a detailed bug report so that any problem that may exist can be solved properly. _______________________________________________ http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm Please visit the above address for information on this list.
