[PATCH v2 04/47] opcodes: add intelgt as a configuration

[Jan Beulich]

On 03.02.2025 18:13, Aktemur, Tankut Baris wrote:
> Hello Jan,
> 
> On Friday, January 3, 2025 5:46 AM, Simon Marchi wrote:
>> On 2024-12-20 04:47, Aktemur, Tankut Baris wrote:
>>> On Wednesday, December 18, 2024 8:22 AM, Jan Beulich wrote:
>>>> On 17.12.2024 19:47, Aktemur, Tankut Baris wrote:
>>>>> On Monday, December 16, 2024 8:44 AM, Jan Beulich wrote:
>>>>>> On 13.12.2024 16:59, Tankut Baris Aktemur wrote:
>>>>>>> Intel GT does not use the opcodes lib for disassembling instructions.
>>>>>>
>>>>>> How else is disassembly going to work?
>>>>>
>>>>> For disassembly, libiga (IGA library from Intel Graphics Compiler [1]) is used.
>>>>> The patch that adds disassembly support is
>>>>>
>>>>>   https://sourceware.org/pipermail/gdb-patches/2024-December/214036.html
>>>>
>>>> That's (a) interesting and (b) covering only gdb; what about objdump?
>>>>
>>>> Jan
>>>
>>> Our case is similar to AMD GPU (ROCm) debugging, for which there is
>>> already support in upstream GDB.  For disassembling, the amdgpu target
>>> also uses a separate disassembler that is part of their debug API
>>> library:
>>>
>>>   https://inbox.sourceware.org/binutils/20220315194303.3716792-3-simon.marchi@polymtl.ca/
>>>
>> https://rocm.docs.amd.com/projects/ROCdbgapi/en/latest/doxygen/html/group__architecture__gro
>> up.html#ga1958899dde4a98c13980e1cb7d657acd
>>>
>>> I believe the amdgpu target has similar configuration restrictions
>>> with our target.  Adding Simon to Cc to confirm.
>>>
>>> Thanks,
>>> -Baris
>>
>> As far as I know, what Baris says is true.
>>
>> Simon
> 
> Any further comments on this?

As with the other patch - since you copy pre-existing behavior, I'm okay(ish).

Jan