[PATCH i-g-t] tools/gputop: Fix engine columns with amdgpu

[Lucas De Marchi]

On Mon, Apr 15, 2024 at 12:29:06PM +0100, Tvrtko Ursulin wrote:
>
>On 03/04/2024 21:27, Lucas De Marchi wrote:
>>On Wed, Apr 03, 2024 at 04:00:57PM +0100, Tvrtko Ursulin wrote:
>>>From: Tvrtko Ursulin <tvrtko.ursulin at igalia.com>
>>>
>>>Amdgpu kernel driver skips output of fdinfo keys for unused engines. That
>>>is completely legal but corrupts the column formatting in the current
>>>code.
>>>
>>>Fix it by simply treating change in detected engines used by a client
>>>as trigger to re-emit a new header. This ensures columns are always
>>
>>That would be kind of ugly if each client uses a different engine.
>>
>>Why not outputing all the engine classes regardless of what the
>>clients are using?
>
>It is a bit ugly but gputop doesn't know what are all the engine 
>classes - it simply auto-detects them from the ones listed in fdinfo. 
>To know "all" would need more code to be added. See ->
>
>>
>>Lucas De Marchi
>>
>>>correctly aligned, albeit with a cost of potentially duplicating the
>>>header for the same DRM minor.
>>>
>>>This is considered good enough for a reference implementation. The
>>>alternative would be to add some real per DRM minor state tracking which
>>>sounds like an overkill, at least until gputop gains a nicer (any) UI.
>
>-> this paragraph.
>
>Can you live with it for now? Until someone decides to write a nicer 
>UI for it?

I think it's ugly to repeat the same header over and over when
there are different engines used, but I'm also unaffected by this
change, so fine by me.

Lucas De Marchi