[U-Boot-Users] [PATCH] Fix dm9000 receive status and len little endian issue

[Liew Tsi Chung]

Ben,

	The data (packet data) and the status/len are different in
LE/BE, cannot use just the macro dm9000_x to cover both. In my case,
packet data is already in BE, but the status/len is in LE; therefore,
status/len require byteswap. This has to do with cpu endian mode.

Regards,
TsiChung

-----Original Message-----
From: Ben Warren [mailto:biggerbadderben at gmail.com] 
Sent: Wednesday, June 25, 2008 4:01 PM
To: Liew Tsi Chung
Cc: U-Boot-Users; Rigby John
Subject: Re: [PATCH] Fix dm9000 receive status and len little endian
issue

Tsi-Chung Liew wrote:
> From: TsiChung Liew <Tsi-Chung.Liew at freescale.com>
>
> The received status and len was in little endian format and caused the

> ethernet unable to proceed further. Add __le16_to_cpu() in 
> dm9000_rx_status_8/16/32bit().
>
> Signed-off-by: TsiChung Liew <Tsi-Chung.Liew at freescale.com>
> ---
>  drivers/net/dm9000x.c |   16 ++++++++++------
>  1 files changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/net/dm9000x.c b/drivers/net/dm9000x.c index 
> 844fb76..e7365fe 100644
> --- a/drivers/net/dm9000x.c
> +++ b/drivers/net/dm9000x.c
> @@ -214,24 +214,28 @@ static void dm9000_rx_status_32bit(u16
*RxStatus, u16 *RxLen)
>  	DM9000_outb(DM9000_MRCMD, DM9000_IO);
>  
>  	tmpdata = DM9000_inl(DM9000_DATA);
> -	*RxStatus = tmpdata;
> -	*RxLen = tmpdata >> 16;
> +	*RxStatus = __le16_to_cpu(tmpdata);
> +	*RxLen = __le16_to_cpu(tmpdata >> 16);
>  }
>  
>  static void dm9000_rx_status_16bit(u16 *RxStatus, u16 *RxLen)  {
>  	DM9000_outb(DM9000_MRCMD, DM9000_IO);
>  
> -	*RxStatus = DM9000_inw(DM9000_DATA);
> -	*RxLen = DM9000_inw(DM9000_DATA);
> +	*RxStatus = __le16_to_cpu(DM9000_inw(DM9000_DATA));
> +	*RxLen = __le16_to_cpu(DM9000_inw(DM9000_DATA));
>  }
>  
>  static void dm9000_rx_status_8bit(u16 *RxStatus, u16 *RxLen)  {
>  	DM9000_outb(DM9000_MRCMD, DM9000_IO);
>  
> -	*RxStatus = DM9000_inb(DM9000_DATA) + (DM9000_inb(DM9000_DATA)
<< 8);
> -	*RxLen = DM9000_inb(DM9000_DATA) + (DM9000_inb(DM9000_DATA) <<
8);
> +	*RxStatus =
> +	    __le16_to_cpu(DM9000_inb(DM9000_DATA) +
> +			  (DM9000_inb(DM9000_DATA) << 8));
> +	*RxLen =
> +	    __le16_to_cpu(DM9000_inb(DM9000_DATA) +
> +			  (DM9000_inb(DM9000_DATA) << 8));
>  }
>  
>  /*
>   
Sorry for opening Pandora's box, but wouldn't it be better to redefine
the DM9000_x macros instead of where they're called?

regards,
Ben