Re: [edk2-devel] [PATCH v1 1/1] MdePkg: Correcting EFI_ACPI_DMA_TRANSFER_TYPE_16_BIT definition

["Paul" <paul.grimes@amd.com>]

[AMD Public Use]

Thanks for the feedback, Lazlo.  I agree with your point on the optimal format for these #defines.  I think it would be best to submit the current patch as is given that the same feedback could (should?) be applied to various other #defines in the file,
eg: 
  EFI_ACPI_DMA_BUS_MASTER_MASK  0x04, which only applies to Bit 2 and
  EFI_ACPI_IRQ_POLARITY_MASK            0x08, ... Bit 3 and
  EFI_ACPI_IRQ_MODE                               0x01, ... bit 0
 
IMO if these defines were to be updated for clarity, it should probably be done for the whole file in a separate commit.    

Thanks,
Paul

-----Original Message-----
From: Laszlo Ersek <lersek@redhat.com> 
Sent: Friday, August 28, 2020 10:06 AM
To: devel@edk2.groups.io; Grimes, Paul <Paul.Grimes@amd.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>; Liming Gao <gaoliming@byosoft.com.cn>; Zhiguang Liu <zhiguang.liu@intel.com>
Subject: Re: [edk2-devel] [PATCH v1 1/1] MdePkg: Correcting EFI_ACPI_DMA_TRANSFER_TYPE_16_BIT definition

[CAUTION: External Email]

On 08/27/20 22:40, Paul wrote:
> In Acpi10.h, EFI_ACPI_DMA_TRANSFER_TYPE_16_BIT is defined as 0x10, but 
> should be 0x02 per the ACPI Specification.
>
> REF:https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2F
> bugzilla.tianocore.org%2Fshow_bug.cgi%3Fid%3D2937&amp;data=02%7C01%7Cp
> aul.grimes%40amd.com%7C82b28bb6544a4612fc1108d84b749dc6%7C3dd8961fe488
> 4e608e11a82d994e183d%7C0%7C0%7C637342311528396385&amp;sdata=7vHYIHHaJU
> 4yrXzAWtv5xTf%2BQfclAUBusz278%2F6I%2BRY%3D&amp;reserved=0
>
> Cc: Michael D Kinney <michael.d.kinney@intel.com>
> Cc: Liming Gao <gaoliming@byosoft.com.cn>
> Cc: Zhiguang Liu <zhiguang.liu@intel.com>
> Signed-off-by: Paul G <paul.grimes@amd.com>
> ---
>  MdePkg/Include/IndustryStandard/Acpi10.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/MdePkg/Include/IndustryStandard/Acpi10.h 
> b/MdePkg/Include/IndustryStandard/Acpi10.h
> index fa06eefbb6e6..adeb5ae8c219 100644
> --- a/MdePkg/Include/IndustryStandard/Acpi10.h
> +++ b/MdePkg/Include/IndustryStandard/Acpi10.h
> @@ -358,7 +358,7 @@ typedef struct {
>  #define EFI_ACPI_DMA_TRANSFER_TYPE_MASK                 0x03
>
>  #define   EFI_ACPI_DMA_TRANSFER_TYPE_8_BIT              0x00
>
>  #define   EFI_ACPI_DMA_TRANSFER_TYPE_8_BIT_AND_16_BIT   0x01
>
> -#define   EFI_ACPI_DMA_TRANSFER_TYPE_16_BIT             0x10
>
> +#define   EFI_ACPI_DMA_TRANSFER_TYPE_16_BIT             0x02
>
>
>
>  //
>
>  // IO Information
>

Good catch. The ACPI spec text was likely cut n' pasted into the edk2 source, and then prefixed with "0x". The spec says,

"""
Bits [1:0] DMA transfer type preference, _SIZ
  00 8-bit only
  01 8- and 16-bit
  10 16-bit only
  11 Reserved
"""

but that's in binary, not in hexadecimal.

In fact, the leading zero on *all four* macros (including
EFI_ACPI_DMA_TRANSFER_TYPE_MASK) is misleading. In hex, the leading zero in the current macros stands for bits [7:4], which are completely irrelevant for the _SIZ bit-field in the DMA Descriptor. So optimally we'd have

#define EFI_ACPI_DMA_TRANSFER_TYPE_MASK                 0x3
#define   EFI_ACPI_DMA_TRANSFER_TYPE_8_BIT              0x0
#define   EFI_ACPI_DMA_TRANSFER_TYPE_8_BIT_AND_16_BIT   0x1
#define   EFI_ACPI_DMA_TRANSFER_TYPE_16_BIT             0x2

But I agree the current patch is OK too:

Reviewed-by: Laszlo Ersek <lersek@redhat.com>

I also agree it's a bugfix and should be merged now.

Thanks
Laszlo