Hi,
On Fri, Oct 14, 2011 at 2:14 PM, Martin Guy <
[hidden email]> wrote:
> I think I've spotted a bug in the calculation of the SPI clock
> frequency on AVR32:
>
> diff --git a/src/platform/avr32/platform.c b/src/platform/avr32/platform.c
> index 75a2d65..899a19a 100644
> --- a/src/platform/avr32/platform.c
> +++ b/src/platform/avr32/platform.c
> @@ -213,10 +213,10 @@ int platform_init()
> spiopt.modfdis = TRUE;
> spiopt.pcs_decode = FALSE;
> spiopt.delay = 0;
> - spi_initMaster(&AVR32_SPI0, &spiopt, REQ_CPU_FREQ);
> + spi_initMaster(&AVR32_SPI0, &spiopt, REQ_PBA_FREQ);
>
> #if NUM_SPI > 4
> - spi_initMaster(&AVR32_SPI1, &spiopt, REQ_CPU_FREQ);
> + spi_initMaster(&AVR32_SPI1, &spiopt, REQ_PBA_FREQ);
> #endif
>
> #endif
>
> where
> int spi_initMaster(volatile avr32_spi_t *spi, const
> spi_master_options_t *opt, U32 pba_hz)
> {
> ...
>
> which I expect would really set a frequency of one quarter of that
> which was asked for (and reported) since CPU=60000000 and PBA=15000000
>
> but I don't have any SPI hardware to test this, so before I make this
> fairly obvious change, which affects all AVR32 boards, has anyone
> tried SPI on any of these boards? It seems a fairly obvious error but
> I wouldn't want to break something that was working... if it was
> working...
I didn't try SPI on AVR32, but your fix is obviously right because of the AVR32 system architecture (as described in the datasheet) and because of this line from spi.c:
int spi_initMaster(volatile avr32_spi_t *spi, const spi_master_options_t *opt, U32
pba_hz)
I'd go ahead with it.
Best,
Bogdan
>
> M
>
_______________________________________________
eLua-dev mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/elua-dev
Locked
