* [PATCH v2] ArmPlatformPkg/NorFlashDxe: use strictly aligned CopyMem()
@ 2016-09-09 12:18 Ard Biesheuvel
2016-09-09 12:34 ` Leif Lindholm
0 siblings, 1 reply; 3+ messages in thread
From: Ard Biesheuvel @ 2016-09-09 12:18 UTC (permalink / raw)
To: edk2-devel, leif.lindholm; +Cc: Ard Biesheuvel
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>
---
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
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH v2] ArmPlatformPkg/NorFlashDxe: use strictly aligned CopyMem()
2016-09-09 12:18 [PATCH v2] ArmPlatformPkg/NorFlashDxe: use strictly aligned CopyMem() Ard Biesheuvel
@ 2016-09-09 12:34 ` Leif Lindholm
2016-09-09 12:59 ` Ard Biesheuvel
0 siblings, 1 reply; 3+ messages in thread
From: Leif Lindholm @ 2016-09-09 12:34 UTC (permalink / raw)
To: Ard Biesheuvel; +Cc: edk2-devel
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
>
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH v2] ArmPlatformPkg/NorFlashDxe: use strictly aligned CopyMem()
2016-09-09 12:34 ` Leif Lindholm
@ 2016-09-09 12:59 ` Ard Biesheuvel
0 siblings, 0 replies; 3+ messages in thread
From: Ard Biesheuvel @ 2016-09-09 12:59 UTC (permalink / raw)
To: Leif Lindholm; +Cc: edk2-devel-01
On 9 September 2016 at 13:34, Leif Lindholm <leif.lindholm@linaro.org> wrote:
> 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>
>
Pushed, thanks.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2016-09-09 12:59 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-09-09 12:18 [PATCH v2] ArmPlatformPkg/NorFlashDxe: use strictly aligned CopyMem() Ard Biesheuvel
2016-09-09 12:34 ` Leif Lindholm
2016-09-09 12:59 ` Ard Biesheuvel
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox