From mboxrd@z Thu Jan  1 00:00:00 1970
From: Josh Wu <josh.wu@atmel.com>
Date: Fri, 16 May 2014 17:17:54 +0800
Subject: [U-Boot] [PATCH 2/2] ARMv7: at91: enable ICache and DCache.
In-Reply-To: <5375CC82.3010705@atmel.com>
References: <1400228130-18113-1-git-send-email-josh.wu@atmel.com>
	<1400228130-18113-3-git-send-email-josh.wu@atmel.com>
	<5375CC82.3010705@atmel.com>
Message-ID: <5375D7C2.9040002@atmel.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, Bo

Thanks for the review  ;-)

On 5/16/2014 4:29 PM, Bo Shen wrote:
> Hi Josh,
>
> On 05/16/2014 04:15 PM, Josh Wu wrote:
>> As SAMA5D3 is a at91 ARMv7 cpu, so we should enable LCD DCache
>> flush for SAMA5D3X-EK.
>>
>> Signed-off-by: Josh Wu <josh.wu@atmel.com>
>> ---
>>   arch/arm/cpu/armv7/at91/cpu.c       |    6 ++++++
>>   board/atmel/sama5d3xek/sama5d3xek.c |    6 +++++-
>>   2 files changed, 11 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/arm/cpu/armv7/at91/cpu.c 
>> b/arch/arm/cpu/armv7/at91/cpu.c
>> index 2fbf60d..6f2164f 100644
>> --- a/arch/arm/cpu/armv7/at91/cpu.c
>> +++ b/arch/arm/cpu/armv7/at91/cpu.c
>> @@ -61,6 +61,12 @@ int print_cpuinfo(void)
>>
>>   void enable_caches(void)
>>   {
>> +#ifndef CONFIG_SYS_ICACHE_OFF
>> +    icache_enable();
>> +#endif
>
> No need this one. Ifdef CONFIG_SYS_ICACHE_OFF, it will be disabled.

ok, good. I will remove this #if.

>
>> +#ifndef CONFIG_SYS_DCACHE_OFF
>> +    dcache_enable();
>> +#endif

This will be removed too.

>>   }
>>
>>   unsigned int get_chip_id(void)
>> diff --git a/board/atmel/sama5d3xek/sama5d3xek.c 
>> b/board/atmel/sama5d3xek/sama5d3xek.c
>> index c835c12..1fc91d9 100644
>> --- a/board/atmel/sama5d3xek/sama5d3xek.c
>> +++ b/board/atmel/sama5d3xek/sama5d3xek.c
>> @@ -200,8 +200,12 @@ int board_init(void)
>>           at91_gmac_hw_init();
>>   #endif
>>   #ifdef CONFIG_LCD
>> -    if (has_lcdc())
>> +    if (has_lcdc()) {
>>           sama5d3xek_lcd_hw_init();
>> +
>> +        /* Enable flushing as we enabled dcache */
>> +        lcd_set_flush_dcache(1);
>
> Why not implement in driver, or else, each SoC enable dcache will need 
> add this.

hmm, the thing is some at91 board, for example at91sam9x5ek board, not 
defined CONFIG_SYS_DCACHE_OFF, that means DCache is enabled. But 
actually it doesn't implement dcache_enable().
In above situation, if we implement it in driver, I can't tell whether 
dcache is really enabled or not.
For this reason, I manually call lcd_set_flush_dcache(1) in the board file.
In the future if all boards (use hlcd driver) enabled dcache I will move 
this function to hlcd driver.

Best Regards,
Josh Wu

>
>> +    }
>>   #endif
>>       return 0;
>>   }
>>
>
> Best Regards,
> Bo Shen
>