>-----Original Message-----
>From: Mike Frysinger [mailto:[email protected]] 
>Sent: Thursday, November 26, 2009 6:02 PM
>To: Cai, Cliff
>Cc: [email protected]; 
>[email protected]
>Subject: Re: [Linux-kernel-commits] [7883] 
>trunk/drivers/spi/bfin_sport_spi.c: bug[#5722] use the new 
>gpio cs scheme
>
>On Thu, Nov 26, 2009 at 01:02, Cai, Cliff wrote:
>>From: Mike Frysinger [mailto:[email protected]]
>>>On Wed, Nov 25, 2009 at 22:06,  <[email protected]> wrote:
>>>> Revision 7883 Author cliff Date 2009-11-25 22:06:52 -0500
>>>(Wed, 25 Nov
>>>> 2009)
>>>>
>>>> Log Message
>>>>
>>>> bug[#5722]use the new gpio cs scheme
>>>>
>>>> trunk/drivers/spi/bfin_sport_spi.c
>>>>
>>>>  static void bfin_sport_spi_cs_active(struct master_data 
>*drv_data, 
>>>> struct slave_data *chip)  {
>>>> -    gpio_direction_output(chip->cs_gpio, 0);
>>>> +    if (chip->chip_select_num > MAX_CTRL_CS)
>>>> +            gpio_direction_output(chip->cs_gpio, 0);
>>>>  }
>>>
>>>these MAX_CTRL_CS checks make no sense at all.  the CS with the 
>>>SPORT/SPI bus can only ever be GPIOs.
>>
>> Not really,the sport can provide cs by itself.
>
>except that MAX_CTRL_CS is specific to the Blackfin SPI 
>controller -- the SPORT does not support 7 different 
>"internal" cs's.  the driver also doesnt explain how exactly 
>"the sport is in charge of it".
>before your commit, a GPIO CS was always used.  after your 
>commit, the GPIO CS is sometimes used, but otherwise, no 
>"internal" cs changes are apparent.
>-mike

If we don't use GPIO as CS, then SPORT can provide cs-change-per-word kind of 
CS.
And we can't control it.I will add the comments in the driver.

Cliff

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

Reply via email to