On Fri, Jun 17, 2011 at 04:23, Bob Liu wrote:
> On Fri, Jun 17, 2011 at 4:16 PM, Mike Frysinger wrote:
>> --- a/drivers/spi/spi_bfin5xx.c
>> +++ b/drivers/spi/spi_bfin5xx.c
>> @@ -687,7 +687,7 @@ static void bfin_spi_pump_transfers(unsigned long data)
>> drv_data->len = (transfer->len) >> 1;
>> cr_width = BIT_CTL_WORDSIZE;
>> drv_data->ops = &bfin_bfin_spi_transfer_ops_u16;
>> - } else if ((bits_per_word > 0) && (bits_per_word % 8 == 0)) {
>> + } else if (bits_per_word % 8 == 0) {
>> drv_data->n_bytes = bits_per_word/8;
>> drv_data->len = transfer->len;
>> cr_width = 0;
>>
>
> I think this change still not work. Sorry for that.
> By this change drv_data->n_bytes will be set to 0 which isn't correct.
>
> In my opinion it should be like this, but i am still have no time to test it .
>
> --- drivers/spi/spi_bfin5xx.c (revision 9946)
> +++ drivers/spi/spi_bfin5xx.c (working copy)
> @@ -682,6 +682,9 @@
>
> /* Bits per word setup */
> bits_per_word = transfer->bits_per_word ? :
> message->spi->bits_per_word;
> + if (!bits_per_word)
> + bits_per_word = 8; /* set default value */
> +
> if ((bits_per_word > 0) && (bits_per_word % 16 == 0)) {
> drv_data->n_bytes = bits_per_word/8;
> drv_data->len = (transfer->len) >> 1;
makes sense. i'll go with this tweaked version:
--- a/drivers/spi/spi_bfin5xx.c
+++ b/drivers/spi/spi_bfin5xx.c
@@ -641,8 +641,9 @@
drv_data->cs_change = transfer->cs_change;
/* Bits per word setup */
- bits_per_word = transfer->bits_per_word ? : message->spi->bits_per_word;
- if ((bits_per_word > 0) && (bits_per_word % 16 == 0)) {
+ bits_per_word = transfer->bits_per_word ? :
+ message->spi->bits_per_word ? : 8;
+ if (bits_per_word % 16 == 0) {
drv_data->n_bytes = bits_per_word/8;
drv_data->len = (transfer->len) >> 1;
cr_width = BIT_CTL_WORDSIZE;
-mike
_______________________________________________
Linux-kernel-commits mailing list
[email protected]
https://blackfin.uclinux.org/mailman/listinfo/linux-kernel-commits
