Re: [Qemu-riscv] [Qemu-devel] [PATCH v6 00/26] Invert Endian bit in SPARCv9 MMU TTE

[Philippe Mathieu-Daudé]

On 8/7/19 10:23 AM, address@hidden wrote:
[...]
> v3:
> - Like v1, the entire TCGMemOp enum is now MemOp.
> - MemOp target dependant attributes are conditional upon NEED_CPU_H
> 
> v4:
> - Added Paolo Bonzini as include/exec/memop.h maintainer
> 
> v5:
> - Improved commit messages to clarify how interface to access
>   MemoryRegion will be converted from "unsigned size" to "MemOp op".
> - Moved cpu_transaction_failed() MemOp conversion from patch #11 to #9
>   to make review easier.
> 
> v6:
> - Improved commit messages.
> - Include as patch #1 an earlier posted TARGET_ALIGNED_ONLY configure patch.
> - Typeless macro SIZE_MEMOP is now inline.
> - size_memop now includes CONFIG_DEBUG_TCG code.
> - size_memop now also encodes endianness via MO_TE.
> - Reverted size_memop operand "unsigned long" back to "unsigned".
> - Second pass of size_memop to replace no-op place holder with 
> MO_{8|16|32|64}.
> - Delay memory_region_dispatch_{read,write} operand conversion until no-op
>   size_memop is implemented so we have proper typing at all points in between.
> - Fixed bug where not all memory_region_dispatch_{read,write} callers where
>   encoding endianness into the MemOp operand, see patch #20.
> - Fixed bug where not all memory_region_dispatch_{read,write} callers were
>   collapsing their byte swap into adjust_endianness, see patch #20 and #22.
> - Split byte swap collapsing patch (v5 #11) into #21 and #22.
> - Corrected non-common *-common-obj to *-obj.
> - Replaced enum device_endian with MemOp to simplify endianness checks. A
>   straight forward sed but touched *alot* of files. See patch #16 and #17.
> - Deleted enum device_endian.
> - Deleted DEVICE_HOST_ENDIAN definition.
> - Generalized the description of introduced MemTxAttrs attribute byte_swap.

I'm confused I think I already reviewed various patches of your previous
series but don't see my Reviewed-by tags.