Re: [PATCH net] net: ena: avoid double-counting TX DMA mappings on failure

[Jakub Kicinski <kuba@kernel.org>]

On Thu, 12 Nov 2026 01:30:29 +0000 Kiyanovski, Arthur wrote:
> > On Wed, 11 Nov 2026 23:39:58 +0000 Arthur Kiyanovski wrote:  
> > > Therefore please do the following:
> > >
> > > 1. Add a From: tag to give credit to Evgeny Ostrovsky, who wrote this fix
> > >    (without leading "-"):
> > >
> > > - From: Evgeny Ostrovsky <evostrov@amazon.com>  
> > 
> > Unclear whether that's warranted TBH.
> > Avi is making the effort to upstream the fix, they deserve the credit.
> >   
> > > This should be first line before "ena_tx_map_skb() increme..."
> > >
> > > 2. Evgeny has agreed, please also add (without leading "-"):
> > >
> > > - Signed-off-by: Evgeny Ostrovsky <evostrov@amazon.com>  
> > 
> > SoB indeed seems appropriate if the patch is based on prior work.  
> 
> The commit code (as well as the other 2 that were already merged)
> Are copies of fixes that we already published in our github driver. 

I get that, but it sat on your GH for 3+ years. At some point the code
sitting somewhere on GH becomes just a GPL(?) code drop and the history
is irrelevant.

> This is the first time I encountered such a case so I searched
> how to give credit to the person who originally implemented the fixes.
> 
> If you feel like SOB is enough I'm OK with it.

Thanks. Also first time I'm seeing this. I'd prefer to give the credit
to the person who cared enough to push the code upstream. 
Or differently put - the best solution going forward would be if you
made sure to upstream ENA fixes from your team in a timely fashion :)