Re: [PATCH v2 1/1] opcodes: microblaze: Add new bit-field instructions

[Michael Eager <eager@eagercon.com>]

Reverted patch.

On 10/7/23 12:23, Maciej W. Rozycki wrote:
> On Fri, 6 Oct 2023, Michael Eager wrote:
> 
>>> This patch has been tested for years of AMD Xilinx Yocto
>>> releases as part of the following patch set:
>>>
>>> https://github.com/Xilinx/meta-xilinx/tree/master/meta-microblaze/recipes-devtools/binutils/binutils
>>>
>>> Signed-off-by: nagaraju <nagaraju.mekala@amd.com>
>>> Signed-off-by: Ibai Erkiaga <ibai.erkiaga-elorza@amd.com>
>>> Signed-off-by: Neal Frager <neal.frager@amd.com>
>>
>> Committed.
> 
>   Yet it has caused numerous regressions:
> 
> microblaze-elf  +FAIL: unordered .debug_info references to .debug_ranges
> microblaze-elf  +FAIL: binutils-all/pr26548
> microblaze-elf  +FAIL: readelf -Wwi pr26548e (reason: unexpected output)
> microblaze-elf  +FAIL: readelf --debug-dump=loc locview-1 (reason: unexpected output)
> microblaze-elf  +FAIL: readelf --debug-dump=loc locview-2 (reason: unexpected output)
> microblaze-elf  +FAIL: readelf -wiaoRlL dw5
> microblaze-elf  +FAIL: readelf -wi dwarf-attributes (reason: unexpected output)
> microblaze-elf  +FAIL: readelf -wKis -P debuglink (reason: unexpected output)
> microblaze-elf  +FAIL: readelf --debug-dump=links --debug-dump=no-follow-links dwo
> microblaze-elf  +FAIL: DWARF2 1
> microblaze-elf  +FAIL: DWARF2 2
> microblaze-elf  +FAIL: DWARF2 3
> microblaze-elf  +FAIL: DWARF2 5
> microblaze-elf  +FAIL: DWARF2 6
> microblaze-elf  +FAIL: DWARF2 7
> microblaze-elf  +FAIL: DWARF2 11
> microblaze-elf  +FAIL: DWARF2 12
> microblaze-elf  +FAIL: DWARF2 13
> microblaze-elf  +FAIL: DWARF2 14
> microblaze-elf  +FAIL: DWARF2 15
> microblaze-elf  +FAIL: DWARF2 16
> microblaze-elf  +FAIL: DWARF2 17
> microblaze-elf  +FAIL: DWARF2 18
> microblaze-elf  +FAIL: DWARF2 19
> microblaze-elf  +FAIL: DWARF2_20: debug ranges ignore non-code sections
> microblaze-elf  +FAIL: DWARF2 21
> microblaze-elf  +FAIL: DWARF5 .loc 0
> microblaze-elf  +FAIL: DWARF4 CU
> microblaze-elf  +FAIL: DWARF5 CU
> microblaze-elf  +FAIL: Check line table is produced with .nops
> microblaze-elf  +FAIL: line number entries for section changes inside .irp
> microblaze-elf  +FAIL: line number entries for .macro expansions
> microblaze-elf  +FAIL: line number entries for expansions of .macro coming from .include
> microblaze-elf  +FAIL: lns-duplicate
> microblaze-elf  +FAIL: lns-common-1
> microblaze-elf  +FAIL: ld-elf/pr22450
> 
> They all seem to follow a similar pattern, e.g:
> 
> exited abnormally with 1, output:readelf: Warning: Corrupt unit length (got 0x4e00004e expected at most 0x4e) in section .debug_info
> 
> FAIL: DWARF2 1
> 
> or:
> 
> exited abnormally with 0, output:readelf: Warning: Corrupt unit length (got 0x20000020 expected at most 0x20) in section .debug_info
> 
> FAIL: Check line table is produced with .nops
> 
> Configured with:
> 
> $ /path/to/configure --disable-nls --disable-gdb --disable-gdbserver --disable-gprofng --disable-libbacktrace --disable-libdecnumber --disable-readline --disable-sim --enable-obsolete --enable-plugins --build=powerpc64le-linux --target=microblaze-elf
> 
>   Just reporting in case it's useful as it has popped up in unrelated
> verification; I won't do anything else here.
> 
>   NB I've skimmed over the change and noticed it does not follow the GNU
> Coding Standards in many places, most prominently in `get_field_imm5width'
> and `get_field_rfsl', but also elsewhere (which is easy to spot looking
> through the diff).
> 
>    Maciej
> 

-- 
Michael Eager