[U-Boot] [PATCH 1/4] net: sun8i_emac: Fix DMA alignment issues with the rx / tx buffers

[Ian Campbell]

On Wed, 2016-07-27 at 18:10 +0200, Hans de Goede wrote:
> This fixes the following CACHE warnings when using sun8i_emac:
> 
> => dhcp
> BOOTP broadcast 1
> BOOTP broadcast 2
> CACHE: Misaligned operation at range [7bf594a8, 7bf59628]
> BOOTP broadcast 3
> CACHE: Misaligned operation at range [7bf59c90, 7bf59e10]
> CACHE: Misaligned operation at range [7bf5a478, 7bf5a5f8]
> DHCP client bound to address 10.42.43.80 (1009 ms)
> 
> Cc: Chen-Yu Tsai <wens at csie.org>
> Cc: Corentin LABBE <clabbe.montjoie at gmail.com>
> Cc: Amit Singh Tomar <amittomer25 at gmail.com>
> Signed-off-by: Hans de Goede <hdegoede at redhat.com>
> ---
>  drivers/net/sun8i_emac.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/sun8i_emac.c b/drivers/net/sun8i_emac.c
> index 7c088c3..877859c 100644
> --- a/drivers/net/sun8i_emac.c
> +++ b/drivers/net/sun8i_emac.c
> @@ -32,7 +32,8 @@
>  
>  #define CONFIG_TX_DESCR_NUM	32
>  #define CONFIG_RX_DESCR_NUM	32
> -#define CONFIG_ETH_BUFSIZE	2024
> +#define CONFIG_ETH_BUFSIZE	2048
> +#define CONFIG_ETH_RXSIZE	2024 /* Note most fit in ETH_BUFSIZE */

s/most/must/?

A comment (perhaps in the commit message rather than the code) as to
how/why RXSIZE and BUFSIZE interact to affect the alignment in the
desired fasion would be useful, since it is non-obvious to me at
least. 

I was about to speculate on the difference of 14 bytes relating to the
Ethernet frame header, but then I realised it's 24 not 14 and deleted
those paragraphs, which I think underscores the need for a comment ;-)

Ian.