Received-SPF: pass (google.com: domain of linux-kernel+bounces-42944-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Mon, 29 Jan 2024 16:56:32 +0200
From: Vladimir Oltean <olteanv@gmail.com>
To: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Cc: =?utf-8?B?QXLEsW7DpyDDnE5BTA==?= <arinc.unal@arinc9.com>,
	Andrew Lunn <andrew@lunn.ch>,
	Florian Fainelli <f.fainelli@gmail.com>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	mithat.guner@xeront.com, erkin.bozoglu@xeront.com,
	Alvin =?utf-8?Q?=C5=A0ipraga?= <ALSI@bang-olufsen.dk>,
	netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH net-next] net: dsa: remove OF-based MDIO bus registration
 from DSA core
Message-ID: <20240129145632.d2mu6vbwcqhjpyty@skbuf>
References: <20240122053348.6589-1-arinc.unal@arinc9.com>
 <20240123154431.gwhufnatxjppnm64@skbuf>
 <d32d17ed-87b5-4032-b310-f387cea72837@arinc9.com>
 <CAJq09z6pidHvtv=3F_yKHDdY89kzYSF+xh89pzg1raAiQPMyMg@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <CAJq09z6pidHvtv=3F_yKHDdY89kzYSF+xh89pzg1raAiQPMyMg@mail.gmail.com>

On Sat, Jan 27, 2024 at 11:23:33PM -0300, Luiz Angelo Daros de Luca wrote:
> > > IIUC, Luiz made the original patch for the realtek switches. Shouldn't
> > > we wait until realtek registers ds->user_mii_bus on its own, before
> > > reverting? Otherwise, you're basically saying that Luiz made the DSA
> > > core patch without needing it.
> >
> > My findings point to that. Luiz made the patch to optionally register the
> > MDIO bus of the MDIO controlled Realtek switches OF-based. So it's not
> > necessary to wait.
> 
> Back in the time when I wrote that code, with the phy_read/write in
> dsa_switch_ops, the OF node was only required to associate IRQ to each
> port. Until my patch to register its own mdiobus driver lands (I hope
> that happens before the next version), the port status will fall back
> to polling. I don't think it is a critical feature but I'll let the
> maintainers decide. ACK for me.
> 
> Regards,
> 
> Luiz

It isn't really great that this loses IRQ support for Realtek internal PHYs,
especially since Arınç's commit message did not estimate this would happen.

I don't see why this patch could not wait until you resubmit the realtek
consolidation set and it gets accepted.