Re: [PATCH 4/5] ArmPkg/CpuDxe: set DmaBufferAlignment according to CWG

[Ard Biesheuvel <ard.biesheuvel@linaro.org>]

On 1 November 2016 at 22:32, Leif Lindholm <leif.lindholm@linaro.org> wrote:
> On Mon, Oct 31, 2016 at 06:13:09PM +0000, Ard Biesheuvel wrote:
>> The DmaBufferAlignment currently defaults to 4, which is dangerously
>> small and may result in lost data on platform that perform non-coherent
>> DMA. So instead, take the CWG value from the cache info registers.
>>
>> Contributed-under: TianoCore Contribution Agreement 1.0
>> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
>> ---
>>  ArmPkg/Drivers/CpuDxe/CpuDxe.c | 4 +++-
>>  1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/ArmPkg/Drivers/CpuDxe/CpuDxe.c b/ArmPkg/Drivers/CpuDxe/CpuDxe.c
>> index d089cb2d119f..ddc64fd255a0 100644
>> --- a/ArmPkg/Drivers/CpuDxe/CpuDxe.c
>> +++ b/ArmPkg/Drivers/CpuDxe/CpuDxe.c
>> @@ -225,7 +225,7 @@ EFI_CPU_ARCH_PROTOCOL mCpu = {
>>    CpuGetTimerValue,
>>    CpuSetMemoryAttributes,
>>    0,          // NumberOfTimers
>> -  4,          // DmaBufferAlignment
>> +  2048,       // DmaBufferAlignment
>>  };
>>
>>  EFI_STATUS
>> @@ -239,6 +239,8 @@ CpuDxeInitialize (
>>
>>    InitializeExceptions (&mCpu);
>>
>> +  mCpu.DmaBufferAlignment = ArmCacheWritebackGranule ();
>> +
>
> Could we hide the internal structure of mCpu here by moving this to a
> helper function and calling
>   InitializeDma (&mCpu);
> (or something)?
>

We could, but why? The actual struct is defined 10 lines up