At Mon, 3 Dec 2007 00:33:25 +0100, Ctirad Fertr wrote:
Dne Tuesday 27 of November 2007 17:43:05 jste napsal(a):
The changes look almost good to me. But, if I understand correctly, the input source put isn't changed properly for EMU1616.
No, it's correct, because emu1010_input_dst[] structure is generic for all EMU PCI cards.
OK, but is emu1010_src_regs used for EMU1616 input, too? Then the value range would be incorrect for input_source, because the same info callback is used for both input and output that gives items = 49 for emu1616.
In my patch, I modified the input put callback to use emu1616_src_regs. Check if it's a wrong implementation.
The patch below is the revised version with bunch of coding style fixes (at the next time, try scripts/checkpatch.pl in linux kernel tree before submission).
I will do that next time. Thank you.
Let me know if it works for you. I'll apply to ALSA tree after your confirmation.
I works very well :)
Unfortunatelly, now I found that driver (or host card) reload causes kernel ooops on 2.6.23 (2.6.22 is unaffected). This is NOT related to my patch. It crashes even with 1.0.15 version. This is bad :( It crashes immeditely after a firmware for the host card is loaded. Can someone confirm that?
Try HG version whether the same problem occurs. Many bugs related with emu10k1 have been done after 1.0.15.
If this still happens, please paste the Oops log to see more details.
thanks,
Takashi