[U-Boot] [PATCH 2/2] arm: socfpga: enable data/inst prefetch and shared override in the L2

[Dinh Nguyen]

On 10/15/2015 09:32 AM, Marek Vasut wrote:
> On Wednesday, October 14, 2015 at 06:32:42 PM, Pavel Machek wrote:
>> On Mon 2015-10-12 09:59:57, dinguyen at opensource.altera.com wrote:
>>> From: Dinh Nguyen <dinguyen at opensource.altera.com>
>>>
>>> Update the L2 AUX CTRL settings for the SoCFPGA.
>>>
>>> Enabling D and I prefetch bits helps improve SDRAM performance on the
>>> platform.
>>>
>>> Also, we need to enable bit 22 of the L2. By not having bit 22 set in the
>>> PL310 Auxiliary Control register (shared attribute override enable) has
>>> the side effect of transforming Normal Shared Non-cacheable reads into
>>> Cacheable no-allocate reads.
>>>
>>> Coherent DMA buffers in Linux always have a Cacheable alias via the
>>> kernel linear mapping and the processor can speculatively load cache
>>> lines into the PL310 controller. With bit 22 cleared, Non-cacheable
>>> reads would unexpectedly hit such cache lines leading to buffer
>>> corruption.
>>>
>>> Signed-off-by: Dinh Nguyen <dinguyen at opensource.altera.com>
>>> ---
>>>
>>>  arch/arm/include/asm/pl310.h |  2 ++
>>>  arch/arm/mach-socfpga/misc.c | 12 ++++++++++++
>>>  2 files changed, 14 insertions(+)
>>>
>>> diff --git a/arch/arm/include/asm/pl310.h b/arch/arm/include/asm/pl310.h
>>> index 18b90b7..7a11405 100644
>>> --- a/arch/arm/include/asm/pl310.h
>>> +++ b/arch/arm/include/asm/pl310.h
>>> @@ -17,6 +17,8 @@
>>>
>>>  #define L2X0_CTRL_EN				1
>>>  
>>>  #define PL310_SHARED_ATT_OVERRIDE_ENABLE		(1 << 22)
>>>
>>> +#define PL310_AUX_CTRL_DATA_PREFETCH_MASK	(1 << 28)
>>> +#define PL310_AUX_CTRL_INST_PREFETCH_MASK	(1 << 29)
>>
>> These would be
>>
>> arch/arm/include/asm/hardware/cache-l2x0.h:#define
>> L310_PREFETCH_CTRL_DATA_PREFETCH     BIT(28)
>> arch/arm/include/asm/hardware/cache-l2x0.h:#define
>> L310_PREFETCH_CTRL_INSTR_PREFETCH    BIT(29)
>>
>> ...in kernel. So maybe staying with L310_ prefix makes sense?
>> Otherwise it looks ok.
> 
> Why is it L... in one and PL... in the other one ? What does the "PL"
> prefix stand for anyway ?
> 

As Pavel pointed out, it should be L310_x as this is how the Linux
kernel is defining it. It's my mistake in the previous patch to change
the define to PL310.

I'll respin this patch with L310_PREFETCH_CTRL_DATA_PREFETCH and
L310_PREFETCH_CTRL_INSTR_PREFETCH.


Dinh