From mboxrd@z Thu Jan  1 00:00:00 1970
From: Steve Rae <srae@broadcom.com>
Date: Thu, 29 May 2014 10:58:22 -0700
Subject: [U-Boot] [PATCH] mmc: add wrappers for MMC block_{read, write,
 erase}
In-Reply-To: <53875F6D.5010708@wwwdotorg.org>
References: <1401315346-30231-1-git-send-email-srae@broadcom.com>
	<53875F6D.5010708@wwwdotorg.org>
Message-ID: <5387753E.9060106@broadcom.com>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

Hi, Stephen

On 14-05-29 09:25 AM, Stephen Warren wrote:
> On 05/28/2014 04:15 PM, Steve Rae wrote:
>> Each wrapper function:
>> - switches to the specified physical partition, then
>> - performs the original function, and then
>> - switches back to the original physical partition
>> where the physical partition (aka HW partition) is
>>    0=User, 1=Boot1, 2=Boot2, etc.
>
> This feels wrong; why wouldn't mmc_get_dev() return a block_dev_desc_t
> containing block_read/block_write functions that do the HW partition
> switching. That way, this is all completely hidden, and all client code
> only knows about block devices, rather than having to know about
> MMC-specific mmc_block_read/write/erase_hwpart() functions.
>
This goes back to the initial discussion on this mailing list (which was 
never resolved):
   http://lists.denx.de/pipermail/u-boot/2014-April/178171.html
This issue is that the three callback functions defined in 
'block_desc_t' do not accept the "partition number" as an argument.
It was suggested that we could overwrite those functions; but the 
"partition number" still needs to be passed in by:
(1) overloading the "int dev_num" argument, or
(2) adding another argument to the callback functions
I assumed that neither of these was acceptable, so I have proposed these 
wrappers...

Thanks, Steve