Re: [PATCH 1/2] net: ravb: Always process TX descriptor ring

[Niklas Söderlund]

On 2024-03-26 09:54:04 +0000, Paul Barker wrote:
> On 26/03/2024 09:38, Niklas Söderlund wrote:
> > Hi Paul,
> > 
> > Thanks for your work.
> > 
> > On 2024-03-26 08:37:39 +0000, Paul Barker wrote:
> >> The TX queue should be serviced each time the poll function is called,
> >> even if the full RX work budget has been consumed. This prevents
> >> starvation of the TX queue when RX bandwidth usage is high.
> > 
> > Is this not a design decision? That the driver should prioritize Rx over 
> > Tx if there is contention. I have no opinion on if this design is good 
> > or bad, I let Sergey judge that.
> > 
> >>
> >> Fixes: a0d2f20650e8 ("Renesas Ethernet AVB PTP clock driver")
> > 
> > However, I do not think it is a bug and should not have a fixes tag.  
> > Also this fixes tag is incorrect, this points to the commit where ravb.c 
> > was renamed ravb_main.c. ravb_poll() is not touched by this commit.
> 
> Sergey identified these as bugfixes in the following emails:
>   https://lore.kernel.org/netdev/a364963f-4e4f-dba5-cb59-b2125c14e8fc@xxxxxx/
>   https://lore.kernel.org/netdev/c58ab319-222b-5ab0-0924-7774a473e276@xxxxxx/

I see, I missed that. I do not agree, this is not a bugfix, it changes a 
design decision and the behavior of the driver.

@Sergey: What do you think? If you feel strongly about this being a bug 
I will yield.

> 
> I got the wrong fixes tag though, it should be:
>   Fixes: c156633f1353 ("Renesas Ethernet AVB driver proper")
> 
> Thanks,
> 
> -- 
> Paul Barker






-- 
Kind Regards,
Niklas Söderlund