On Tue, Jul 27, 2010 at 5:00 AM, Mike Frysinger <[email protected]> wrote:
> On Mon, Jul 26, 2010 at 07:05,  <[email protected]> wrote:
>> Log Message
>>
>> fix some small problem to make adau1761 can work now.
>
> when you make commits, please observe some basic rules:
>  - prefix the commit message with a short identifier to narrow down
> what you're working on (so like "sigma firmware:" or "adau1761:")
>  - include information on what you're actually fixing, how you're
> fixing it, and anything else relevant.  i cant take this change and
> send it upstream with "fix some small problem".  the maintainers would
> reject that in a heart beat.
>  - split your commits up into logical pieces.  so this should have
> been one commit to the sigma firmware driver and a different commit to
> the adau1761 driver.
>

I am sorry for that.
It won't happen next time :-)

>> @@ -30,7 +30,8 @@
>>                       if (ssfw->fw->size < ssfw->pos + len)
>>                               return -EINVAL;
>>                       ret = i2c_master_send(client, (void *)&sa->addr, len);
>> -                     len -= 2;
>
> the length field is currently defined as including the two byte i2c
> address.  so why is this line being removed ?
>

Because I think the sigma_action_len() is just the payload len.
It needn't to add the 2bytes i2c addr len which can be included
by sizeof(struct sigma_action).

Of course it's no affect leaving it the former way, just change the
sigma firmware util to adapt it.

>> +/*
>> + * read ADAU1761 hw register and update cache
>> + */
>> +static int adau1761_read_reg_byte(struct snd_soc_codec *codec,
>> +     unsigned int reg)
>> ...
>> -     if (codec->hw_read(codec, ADAU_DSP_ENA) != 0x01)
>> +     if (adau1761_read_reg_byte(codec, ADAU_DSP_ENA) != 0x01)
>
> if this is a generic i2c hook func for this device driver, then
> shouldnt you still be using the hw function pointers ?

The generic i2c hook hw_read func will transfer register
and data through one i2c_transfer() func call.

But the data can't be read out in this way on our platform,
so I broke it out to two i2c_transfer() call.

Thanks
-- 
Regards,
--Bob
_______________________________________________
Linux-kernel-commits mailing list
[email protected]
https://blackfin.uclinux.org/mailman/listinfo/linux-kernel-commits

Reply via email to