On Fri, Jan 22, 2010 at 9:43 AM, SeungHyun Choi sh428.choi@samsung.com wrote:

Dear. Jassi

Thank you for all the information of AC97 reforming.

I'm working for Android smart phone now as T/L with Samsung Handset division (S3C6410 --> Galaxy-spica, InstinctQ)

Actually, we are using I2S interface and external codec chip  for audio because of various functions of audio like connecting BT or modem.. etc.

Honestly  out of my mind about AC97 codec. But your works are look fine. I agree too.

thanks for agreeing.

I've reformed i2c & pcm driver in linux 2.6.29 too since android doesn't have audio library and diffrent from other OS like WM or Linux

I keep updated(after backporting latest code from Mark's GIT) Samsung linux-2.6.29/31 that you might want to try out.

It was so difficult to handle buffer management between Platform and Kernel. (noise, distortion. ...)

Next time, I'd like to share with you if there is a chance.

Regards.

------- Original Message ------- Sender : jassi brarjassisinghbrar@gmail.com Date : 2010-01-21 15:58 (GMT+09:00) Title : Re: [alsa-devel] Reforming s3c2443 ac97 driver

On Fri, Jan 8, 2010 at 6:48 PM, Mark Brown broonie@opensource.wolfsonmicro.com wrote:

...

On 8 Jan 2010, at 05:28, JASWINDERSINGH BRAR jassi.brar@samsung.com  wrote:

...

Hi,      Right now the S3C2443's AC97_ver2.0 controller driver isn't  flexible enough  to handle newer SoCs.    In order to prepare ground for supporting 64xx and S5Pxxxx the  extant driver  needs to be reformed, and that is what i am planning to do next.

I plan to do the following:-  1) User platform_driver structure to detect and manage AC97 controller  platform devices.  2) Remove hardcoded parameters(DMA, IO address, IRQ etc) and pass via  platform resources.  3) Initialize platform specific stuff by callback pointers to  platform code(cfg_gpio)

Objections, suggestions welcome.

This sounds like a good approach - please go ahead.

There has been a slight change though.

Both parts(platform_device interface and ALSA AC97 interface) of s3c2443-ac97.c needed so much reforming that I found rewriting the driver again a better option. The new driver is called 's3c-ac97.c' ... hoping the AC97 controller doesn't change in SoCs released after you read this mail :)

The s3c2443-ac97.c is used by SMDK2443 and LN2440SBC machines. My new driver is at least as good as old one for SMDK2443(both detect controller fine but none produce any sound.. might be some h/w issue on the only board I have got). I have no access to LN2440SBC but there is no reason for it to complain. Besides, the new AC97 controller driver(plus a machine driver for SMDKs with WM9713 attached to AC97 port) has been tested on all SoCs that I could get my hands on(6410, C100, C110 and V210).

Trying to avoid one outright rejected series of patches, I wanted to confirm if my approach is acceptable.

---

최 승 현 책임

삼성전자주식회사

Embedded S/W Center

System LSI사업부

반도체총괄

031-209-8330

sh428.choi@samsung.com

---