public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: Leif Lindholm <leif.lindholm@linaro.org>
To: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: edk2-devel@lists.01.org
Subject: Re: [PATCH v2] ArmPlatformPkg/NorFlashDxe: use strictly aligned CopyMem()
Date: Fri, 9 Sep 2016 13:34:29 +0100	[thread overview]
Message-ID: <20160909123429.GP16080@bivouac.eciton.net> (raw)
In-Reply-To: <1473423502-7061-1-git-send-email-ard.biesheuvel@linaro.org>

On Fri, Sep 09, 2016 at 01:18:22PM +0100, Ard Biesheuvel wrote:
> The UEFI spec stipulates that unaligned accesses should be enabled
> on CPUs that support them, which means all of them, given that we
> no longer support pre-v7 ARM cores, and the AARCH64 bindings mandate
> support for unaligned accesses unconditionally.
> 
> This means that one should not assume that CopyMem () is safe to call
> on regions that may be mapped using device attributes, which is the
> case for the NOR flash. Since we have no control over the mappings when
> running under the OS, and given that write accesses require device
> mappings, we should not call CopyMem () in the read path either, but
> use our own implementation that is guaranteed to take alignment into
> account.
> 
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>

That's a lot less complex, thanks.
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>

> ---
> v2:
> - simplify the AlignedCopyMem () implementation, given that we do not require
>   memmove() semantics for copying between NOR flash and memory
> - document the above
> - add macro to perform alignment check
> - drop redundant/unnecessary casts and parentheses
> 
>  ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe.c | 63 +++++++++++++++++++-
>  1 file changed, 61 insertions(+), 2 deletions(-)
> 
> diff --git a/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe.c b/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe.c
> index 3abbe5cb32bc..ca61ac5e1983 100644
> --- a/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe.c
> +++ b/ArmPlatformPkg/Drivers/NorFlashDxe/NorFlashDxe.c
> @@ -744,6 +744,65 @@ NorFlashWriteBlocks (
>    return Status;
>  }
>  
> +#define BOTH_ALIGNED(a, b, align) ((((UINTN)(a) | (UINTN)(b)) & ((align) - 1)) == 0)
> +
> +/**
> +  Copy Length bytes from Source to Destination, using aligned accesses only.
> +  Note that this implementation uses memcpy() semantics rather then memmove()
> +  semantics, i.e., SourceBuffer and DestinationBuffer should not overlap.
> +
> +  @param  DestinationBuffer The target of the copy request.
> +  @param  SourceBuffer      The place to copy from.
> +  @param  Length            The number of bytes to copy.
> +
> +  @return Destination
> +
> +**/
> +STATIC
> +VOID *
> +AlignedCopyMem (
> +  OUT     VOID                      *DestinationBuffer,
> +  IN      CONST VOID                *SourceBuffer,
> +  IN      UINTN                     Length
> +  )
> +{
> +  UINT8             *Destination8;
> +  CONST UINT8       *Source8;
> +  UINT32            *Destination32;
> +  CONST UINT32      *Source32;
> +  UINT64            *Destination64;
> +  CONST UINT64      *Source64;
> +
> +  if (BOTH_ALIGNED(DestinationBuffer, SourceBuffer, 8) && Length >= 8) {
> +    Destination64 = DestinationBuffer;
> +    Source64 = SourceBuffer;
> +    while (Length >= 8) {
> +      *Destination64++ = *Source64++;
> +      Length -= 8;
> +    }
> +
> +    Destination8 = (UINT8 *)Destination64;
> +    Source8 = (CONST UINT8 *)Source64;
> +  } else if (BOTH_ALIGNED(DestinationBuffer, SourceBuffer, 4) && Length >= 4) {
> +    Destination32 = DestinationBuffer;
> +    Source32 = SourceBuffer;
> +    while (Length >= 4) {
> +      *Destination32++ = *Source32++;
> +      Length -= 4;
> +    }
> +
> +    Destination8 = (UINT8 *)Destination32;
> +    Source8 = (CONST UINT8 *)Source32;
> +  } else {
> +    Destination8 = DestinationBuffer;
> +    Source8 = SourceBuffer;
> +  }
> +  while (Length-- != 0) {
> +    *Destination8++ = *Source8++;
> +  }
> +  return DestinationBuffer;
> +}
> +
>  EFI_STATUS
>  NorFlashReadBlocks (
>    IN NOR_FLASH_INSTANCE   *Instance,
> @@ -791,7 +850,7 @@ NorFlashReadBlocks (
>    SEND_NOR_COMMAND (Instance->DeviceBaseAddress, 0, P30_CMD_READ_ARRAY);
>  
>    // Readout the data
> -  CopyMem(Buffer, (UINTN *)StartAddress, BufferSizeInBytes);
> +  AlignedCopyMem (Buffer, (VOID *)StartAddress, BufferSizeInBytes);
>  
>    return EFI_SUCCESS;
>  }
> @@ -832,7 +891,7 @@ NorFlashRead (
>    SEND_NOR_COMMAND (Instance->DeviceBaseAddress, 0, P30_CMD_READ_ARRAY);
>  
>    // Readout the data
> -  CopyMem (Buffer, (UINTN *)(StartAddress + Offset), BufferSizeInBytes);
> +  AlignedCopyMem (Buffer, (VOID *)StartAddress + Offset, BufferSizeInBytes);
>  
>    return EFI_SUCCESS;
>  }
> -- 
> 2.7.4
> 


  reply	other threads:[~2016-09-09 12:34 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-09-09 12:18 [PATCH v2] ArmPlatformPkg/NorFlashDxe: use strictly aligned CopyMem() Ard Biesheuvel
2016-09-09 12:34 ` Leif Lindholm [this message]
2016-09-09 12:59   ` Ard Biesheuvel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-list from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160909123429.GP16080@bivouac.eciton.net \
    --to=devel@edk2.groups.io \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox