From mboxrd@z Thu Jan  1 00:00:00 1970
From: Dirk Behme <dirk.behme@googlemail.com>
Date: Tue, 10 Apr 2012 18:20:06 +0200
Subject: [U-Boot] [PATCH] FAT: Properly align buffers to allow cache
	operations
In-Reply-To: <4F844CAE.1010502@boundarydevices.com>
References: <1333953903-26335-1-git-send-email-marex@denx.de>
	<201204100036.07124.vapier@gentoo.org>
	<201204100700.48644.marex@denx.de>
	<4F844CAE.1010502@boundarydevices.com>
Message-ID: <4F845DB6.7050002@googlemail.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

On 10.04.2012 17:07, Eric Nelson wrote:
> On 04/09/2012 10:00 PM, Marek Vasut wrote:
>> Dear Mike Frysinger,
>>
>>> On Monday 09 April 2012 02:45:03 Marek Vasut wrote:
>>>> @@ -765,9 +767,6 @@
>>>>
>>>> -__attribute__ ((__aligned__ (__alignof__ (dir_entry))))
>>>> -__u8 do_fat_read_block[MAX_CLUSTSIZE];
>>>> -
>>>>
>>>> @@ -788,6 +787,7 @@ do_fat_read (const char *filename, void *buffer,
>>>>
>>>> int rootdir_size = 0;
>>>> int j;
>>>>
>>>> + uint8_t do_fat_read_block[MAX_CLUSTSIZE]
>>>> __attribute__((aligned(32)));
>>>
>>> what's going on here exactly ? the old code had the advantage of
>>> being in
>>> the bss and the linker taking care of the alignment. this new code
>>> has an
>>> incorrectly hard-coded "32", and puts a 64KiB array onto the *stack*.
>>
>> This will be probably fixed in Eric's patch
>>
> Yep. I left it in bss space.

Sorry guys, but I'm confused now.

We have two patches, Eric's [1] and Marek's [2]. Which one should we 
take? With the discussion here and [3] I'm somehow under the 
impression that both patches [1] [2] are not complete?

Do I miss anything? Or do we need a new version ("best of [1] & [2]")?

Best regards

Dirk

[1] http://lists.denx.de/pipermail/u-boot/2012-March/119311.html

[2] http://lists.denx.de/pipermail/u-boot/2012-April/122100.html

[3] http://lists.denx.de/pipermail/u-boot/2012-April/122126.html