[alsa-devel] [PATCH] ALSA: Support Media Vision Jazz16 chips

Rene Herman rene.herman at gmail.com
Wed Mar 28 23:34:26 CEST 2007

On 03/28/2007 10:03 PM, Rask Ingemann Lambertsen wrote:

>> MPU-401 IRQ:		*2, 3, 5, 7 (<-- you're also listing 10 and 15?)
>    I don't recall having ever managed to generate an MPU interrupt. I have
> no MIDI hardware.

I do, so I'll do some testing here. Need to set that up though...

>> CD ROM Port:		320, *340
>> That "CD ROM Port" seems to be a version glitch; mine only has an
>> NCR SCSI chip with a 2x8 pin connector but the install asked  me if
>> I had a Sony, Panasonic or Mitsumi CD-ROM (those old proprietary
>> CD-ROM interfaces).
> Mine has an IDE port, jumper A for 0x1f0/0x3f6, jumper B for
> 0x1f0/0x376, jumper "CDROM IRQ" for 10, 11, 14 or 15. J9 & J10: 1-2
> AMP OFF, 2-3: AMP ON.
> Note that my board is rather anonymous. For all I know, it could be
> an OEM one.

I misidentified the chip. It's NCR, but not SCSI. It's a proprietary 
CD-ROM controller for a Philips CM260 CD-ROM drive. I remembered I 
actually had a standalone Philips CD-ROM controller lying around and 
while that has a motorola chip, it's the same type: a "DIGBIE2 LMSI". 
Don't have a drive to go with that...

On the back of my Jazz16 is a sticker saying "Philips".

> # echo 'set irq 10 dma 5 dma 1' > /sys/bus/pnp/devices/03\:01/resources

[ ... ]

> Playing WAVE '/tmp/jazz16ting/TESTDMA.WAV' : Unsigned 8 bit, Rate 11025 Hz, Mono

[ ... ]

>   buffer_size  : 5495
>   period_size  : 1374

[ ... ]

> pointer (0x1578) for DMA #5 is greater than transfer size (0x1577) 
> (repeats four times - note 0x1577 = 5495 = buffer size above)
> Playback sounds normal except for a click for each "pointer ..."
> message. Any ideas? /proc/interrupts increments by 11.

Well, not _really_ but I've just helped debug a problem (on a rather 
more modern card) with buffer alignment and I do remember some issues 
with ISA-DMA buffer alignment from the old DOS days; normally the OS 
should be taking care of it all, but maybe something on the hardware 
side is also confused about 8 versus 16-bit DMA. Long shot, but try 
playing with aplay --buffer-size=4096 and/or --period-size=1024 and 
maybe varying that a bit....

> Playing WAVE '/tmp/male.wav' : Signed 16 bit Little Endian, Rate 8000 Hz, Mono

[ ... ]

>    No sound, no interrupts and this message from /var/log/kernel:
> kernel: playback write error (DMA or IRQ trouble?)

Oh good. "doesn't work" is a nice and definitive diagnosis :)

>> Would you happen to have a description of the DIP switches/jumpers? On 
>> mine, there's a red block of 8 DIP switches marked
>> MAB0 | MAB1 | DISMPU | DAB2 | DAB1 | DAB0 | DISJS | <empty>
>> which with the exception of the two DIS switches isn't completely 
>> self-evident;
>    Guesswork: MAB0/1 to select SCSI BIOS memory address, DAB0-2 to select
> NCR chip SCSI device ID - does it happen to be set to 7? DAB0-2 could also
> tbe the NCR chip base address.

I actually expect MAB0-1 is MPU Adress Base (00, 01, 10, 11 = 300, 310, 
320, 330) and DAB0-2 is DSP Address Base (210, ..., 260). The other two 
jumper blocks are close to the NRC chip, not at all to the "SPK" out (my 
card also has a "soundblaster-like" wheel on the back for hardware 
volume control, and I believe that's the only way to influence the 
onboard, as opposed to onchip, amp). All switches just have two 
positions "ON" and "OFF" and all are currently OFF.

I'll let you know about the MPU IRQ.


More information about the Alsa-devel mailing list