RE: [PATCH net-next 3/4] fec: add FIXME to move 'mac_managed_pm' to probe

From: Wei Fang <wei.fang@nxp.com>
To: Wolfram Sang <wsa+renesas@sang-engineering.com>,
	"linux-renesas-soc@vger.kernel.org" 
	<linux-renesas-soc@vger.kernel.org>,
	"kernel@pengutronix.de" <kernel@pengutronix.de>,
	Shenwei Wang <shenwei.wang@nxp.com>,
	Clark Wang <xiaoning.wang@nxp.com>,
	dl-linux-imx <linux-imx@nxp.com>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>
Subject: RE: [PATCH net-next 3/4] fec: add FIXME to move 'mac_managed_pm' to probe
Date: Thu, 16 Mar 2023 09:10:23 +0000	[thread overview]
Message-ID: <AM5PR04MB313943C24A61C48D761B452688BC9@AM5PR04MB3139.eurprd04.prod.outlook.com> (raw)
In-Reply-To: <ZBLNCYgeTtNBSaMi@ninjato>

> -----Original Message-----
> From: Wolfram Sang <wsa+renesas@sang-engineering.com>
> Sent: 2023年3月16日 16:02
> To: Wei Fang <wei.fang@nxp.com>; linux-renesas-soc@vger.kernel.org;
> kernel@pengutronix.de; Shenwei Wang <shenwei.wang@nxp.com>; Clark
> Wang <xiaoning.wang@nxp.com>; dl-linux-imx <linux-imx@nxp.com>; David S.
> Miller <davem@davemloft.net>; Eric Dumazet <edumazet@google.com>;
> Jakub Kicinski <kuba@kernel.org>; Paolo Abeni <pabeni@redhat.com>;
> linux-kernel@vger.kernel.org; netdev@vger.kernel.org
> Subject: Re: [PATCH net-next 3/4] fec: add FIXME to move 'mac_managed_pm'
> to probe
> 
> 
> > Yes, I will resend the series as RFC with more explanations.
> 
> Because I was able to fix SMSC myself, I'll just describe the procedure
> here:
> 
> 1) apply this debug patch:
> 
> diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c index
> 1b2e253fce75..7b79c5979486 100644
> --- a/drivers/net/phy/phy_device.c
> +++ b/drivers/net/phy/phy_device.c
> @@ -310,6 +310,8 @@ static __maybe_unused int
> mdio_bus_phy_suspend(struct device *dev)
>  	if (phydev->mac_managed_pm)
>  		return 0;
> 
> +printk(KERN_INFO "****** MDIO suspend\n");
> +
>  	/* Wakeup interrupts may occur during the system sleep transition when
>  	 * the PHY is inaccessible. Set flag to postpone handling until the PHY
>  	 * has resumed. Wait for concurrent interrupt handler to complete.
> 
> 2) boot the device without bringing the interface (and thus the PHY) up.
>    Bringing it down after it was up is not the same! It is important
>    that it was never up before.
> 
> 3) do a suspend-to-ram/resume cycle
> 
> 4) your log should show the above debug message. If not, I was wrong
> 
> 5) If yes, apply a similar fix to the one I did for the Renesas drivers
>    in this series
> 
> 6) suspend/resume should not show the debug message anymore
> 
> 7) test for regressions and send out :)
> 
> I hope this was understandable. If not, feel free to ask.
> 
> Happy hacking,
> 

Thank you Wolfram. But I'm not sure whether it's really an issue. The flag
" mac_managed_pm" indicates that the MAC driver will take care of
suspending/resuming the PHY, that is to say the MAC driver calls
phy_stop()/phy_start() in its PM callbacks. If a ethernet interface never
brings up, the MAC PM callbacks do nothing and just return 0 directly. So I
think it's fine for the MDIO PM to do suspend/resume the PHY unless the MDIO
bus can't be accessed.