Re: [PATCHSET 22/25] add support for PC-9800 architecture (sound alsa)

[Takashi Iwai <tiwai@suse.de>]

Hi,

At Fri, 25 Oct 2002 11:47:55 +0900,
Osamu Tomita wrote:
> 
> Hi.
> 
> -----Original Message-----
> From: Takashi Iwai
> To: Osamu Tomita
> Cc: LKML
> Sent: 2002/10/24 0:46
> Subject: Re: [PATCHSET 22/25] add support for PC-9800 architecture (sound
> alsa)
> 
> > the question is, whether cs4232 module works on PC9800, or not.
> > i guess the control-port is not used on this card.  in such a case,
> > you can deactivate the control-port via module option (or even add
> > ifdef for the specific kernel config).
> Test results on some PC-9800 (including not CS4232 chip),
> snd-cs4232 driver works fine with snd_isapnp=0 snd_cport=-1 options.
> Thanks for your advice.
> 
> Another problem, I'm tring to separate mpu401 driver for PC-9800.
> To do this, I need to patch many files. So I think previus patch is
> better than separated driver. How do you think?

for which card (chip), is the modification of mpu401 necessary?
if it's only for the cs4232 (on-board?) above, then we can make a
new top-level module, such like snd-pc98-somewhat, which uses the
cs4231 lowlevel module without isapnp stuff, and the modified mpu401
routines, so that the original cs4231 and mpu401 are not changed too
much.

even if cs4231 low-level routines must be changed for pc98, we can
pass the hardware-type CS4231_HW_PC98 in snd_cs4231_create(), and as
Alan suggested, the pc98-specific codes run on this condition.

although this is true for mpu401, the large initialization-code found
in the last patch is definitely not for the "common" module like
mpu401_uart.c.  such a thing should go into the device-specific
module.
since both cs4231 and mpu401_uart are shared among veraious drivers,
changing such a common module results in breakage of other drivers.
(please note that even some pci drivers use mpu401 low-level module.)

or, the change for mpu401 is always necessary on PC9800 regardless of
which device is used?


ciao,

Takashi