Re: [PATCH] block i2c tuner reads for Avermedia Twinstar in the af9035 driver

[Antti Palosaari <crope@iki.fi>]

On 03/11/2013 01:51 PM, Jose Alberto Reguero wrote:
> On Lunes, 11 de febrero de 2013 14:48:18 Jose Alberto Reguero escribió:
>> On Domingo, 10 de febrero de 2013 22:11:53 Antti Palosaari escribió:
>>> On 02/10/2013 09:43 PM, Jose Alberto Reguero wrote:
>>>> This patch block the i2c tuner reads for Avermedia Twinstar. If it's
>>>> needed other pids can be added.
>>>>
>>>> Signed-off-by: Jose Alberto Reguero <jareguero@telefonica.net>
>>>>
>>>> diff -upr linux/drivers/media/usb/dvb-usb-v2/af9035.c
>>>> linux.new/drivers/media/usb/dvb-usb-v2/af9035.c ---
>>>> linux/drivers/media/usb/dvb-usb-v2/af9035.c	2013-01-07
>>>> 05:45:57.000000000 +0100 +++
>>>> linux.new/drivers/media/usb/dvb-usb-v2/af9035.c	2013-02-08
>>>> 22:55:08.304089054 +0100 @@ -232,7 +232,11 @@ static int
>>>> af9035_i2c_master_xfer(struct
>>>>
>>>>    			buf[3] = 0x00; /* reg addr MSB */
>>>>    			buf[4] = 0x00; /* reg addr LSB */
>>>>    			memcpy(&buf[5], msg[0].buf, msg[0].len);
>>>>
>>>> -			ret = af9035_ctrl_msg(d, &req);
>>>> +			if (state->block_read) {
>>>> +				msg[1].buf[0] = 0x3f;
>>>> +				ret = 0;
>>>> +			} else
>>>> +				ret = af9035_ctrl_msg(d, &req);
>>>>
>>>>    		}
>>>>    	
>>>>    	} else if (num == 1 && !(msg[0].flags & I2C_M_RD)) {
>>>>    	
>>>>    		if (msg[0].len > 40) {
>>>>
>>>> @@ -638,6 +642,17 @@ static int af9035_read_config(struct dvb
>>>>
>>>>    	for (i = 0; i < ARRAY_SIZE(state->af9033_config); i++)
>>>>    	
>>>>    		state->af9033_config[i].clock = clock_lut[tmp];
>>>>
>>>> +	state->block_read = false;
>>>> +
>>>> +	if (le16_to_cpu(d->udev->descriptor.idVendor) == USB_VID_AVERMEDIA &&
>>>> +		le16_to_cpu(d->udev->descriptor.idProduct) ==
>>>> +			USB_PID_AVERMEDIA_TWINSTAR) {
>>>> +		dev_dbg(&d->udev->dev,
>>>> +				"%s: AverMedia Twinstar: block i2c read from tuner\n",
>>>> +				__func__);
>>>> +		state->block_read = true;
>>>> +	}
>>>> +
>>>>
>>>>    	return 0;
>>>>
>>>>    err:
>>>> diff -upr linux/drivers/media/usb/dvb-usb-v2/af9035.h
>>>> linux.new/drivers/media/usb/dvb-usb-v2/af9035.h ---
>>>> linux/drivers/media/usb/dvb-usb-v2/af9035.h	2013-01-07
>>>> 05:45:57.000000000 +0100 +++
>>>> linux.new/drivers/media/usb/dvb-usb-v2/af9035.h	2013-02-08
>>>> 22:52:42.293842710 +0100 @@ -54,6 +54,7 @@ struct usb_req {
>>>>
>>>>    struct state {
>>>>
>>>>    	u8 seq; /* packet sequence number */
>>>>    	bool dual_mode;
>>>>
>>>> +	bool block_read;
>>>>
>>>>    	struct af9033_config af9033_config[2];
>>>>
>>>>    };
>>>
>>> Could you test if faking tuner ID during attach() is enough?
>>>
>>> Also, I would like to know what is returned error code from firmware
>>> when it fails. Enable debugs to see it. It should print something like
>>> that: af9035_ctrl_msg: command=03 failed fw error=2
>>>
>>>
>>> diff --git a/drivers/media/usb/dvb-usb-v2/af9035.c
>>> b/drivers/media/usb/dvb-usb-v2/af9035.c
>>> index a1e953a..5a4f28d 100644
>>> --- a/drivers/media/usb/dvb-usb-v2/af9035.c
>>> +++ b/drivers/media/usb/dvb-usb-v2/af9035.c
>>> @@ -1082,9 +1082,22 @@ static int af9035_tuner_attach(struct
>>> dvb_usb_adapter *adap)
>>>
>>>                           tuner_addr = 0x60 | 0x80; /* I2C bus hack */
>>>
>>>                   }
>>>
>>> +               // fake used tuner for demod firmware / i2c adapter
>>> +               if (adap->id == 0)
>>> +                       ret = af9035_wr_reg(d, 0x00f641,
>>> AF9033_TUNER_FC0011);
>>> +               else
>>> +                       ret = af9035_wr_reg(d, 0x10f641,
>>> AF9033_TUNER_FC0011);
>>> +
>>>
>>>                   /* attach tuner */
>>>                   fe = dvb_attach(mxl5007t_attach, adap->fe[0],
>>>                   &d->i2c_adap,
>>>
>>>                                   tuner_addr,
>>>
>>> &af9035_mxl5007t_config[adap->id]);
>>> +
>>> +               // return correct tuner
>>> +               if (adap->id == 0)
>>> +                       ret = af9035_wr_reg(d, 0x00f641,
>>> AF9033_TUNER_MXL5007T);
>>> +               else
>>> +                       ret = af9035_wr_reg(d, 0x10f641,
>>> AF9033_TUNER_MXL5007T);
>>> +
>>>
>>>                   break;
>>>
>>>           case AF9033_TUNER_TDA18218:
>>>                   /* attach tuner */
>>>
>>> regards
>>> Antti
>>
>> I will try with fake tuner, but I can't test unil next weekend.
>> If I remember, the read operation is performed, and return good value,
>> but after that, all the i2c transfers fail. Seee:
>>
>> http://www.mail-archive.com/linux-media@vger.kernel.org/msg56346.html
>>
>> Jose Alberto
>>
>>
>
> I tried with fake tuner without success:
>
> [ 1346.707405] DVB: registering new adapter (AVerMedia Twinstar (A825))
> [ 1346.959043] i2c i2c-1: af9033: firmware version: LINK=11.5.9.0
> OFDM=5.17.9.1
> [ 1346.962920] usb 1-2: DVB: registering adapter 0 frontend 0 (Afatech AF9033
> (DVB-T))...
> [ 1347.439354] mxl5007t 1-0060: creating new instance
> [ 1347.440644] mxl5007t_get_chip_id: unknown rev (3f)
> [ 1347.440652] mxl5007t_get_chip_id: MxL5007T detected @ 1-0060
> [ 1347.443023] mxl5007t_write_reg: 472: failed!
> [ 1347.443031] mxl5007t_attach: error -121 on line 903
> [ 1347.443790] usb 1-2: dvb_usb_v2: 'AVerMedia Twinstar (A825)' error while
> loading driver (-19)
> [ 1347.446624] usb 1-2: dvb_usb_v2: 'AVerMedia Twinstar (A825)' successfully
> deinitialized and disconnected

I don't see how the hell it could even go to the mxl5007t_write_reg() 
during attach. Any idea?

I have some thoughts that mxl5007t do not use repeated condition. Driver 
still does that. Could you test to perform register read without a 
repeated I2C condition?

regards
Antti

-- 
http://palosaari.fi/