Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 7 Apr 2023 16:14:36 +0200
From:   Andrew Lunn <andrew@lunn.ch>
To:     "Radu Pirea (OSS)" <radu-nicolae.pirea@oss.nxp.com>
Cc:     hkallweit1@gmail.com, linux@armlinux.org.uk, davem@davemloft.net,
        edumazet@google.com, kuba@kernel.org, pabeni@redhat.com,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        stable@vger.kernel.org
Subject: Re: [PATCH net] net: phy: nxp-c45-tja11xx: disable port and global
 interrupts
Message-ID: <47d0347d-02a1-48a3-8553-d6ab2be731e8@lunn.ch>
References: <20230406095546.74351-1-radu-nicolae.pirea@oss.nxp.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230406095546.74351-1-radu-nicolae.pirea@oss.nxp.com>
Precedence: bulk

On Thu, Apr 06, 2023 at 12:55:46PM +0300, Radu Pirea (OSS) wrote:
> Disabling only the link event irq is not enough to disable the
> interrupts. PTP will still be able to generate interrupts.
> 
> The interrupts are organised in a tree on the C45 TJA11XX PHYs. To
> completely disable the interrupts, they are disable from the top of the
> interrupt tree.
> 
> Fixes: 514def5dd339 ("phy: nxp-c45-tja11xx: add timestamping support")
> CC: stable@vger.kernel.org # 5.15+
> Signed-off-by: Radu Pirea (OSS) <radu-nicolae.pirea@oss.nxp.com>
> ---
>  drivers/net/phy/nxp-c45-tja11xx.c | 22 ++++++++++++++++------
>  1 file changed, 16 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/net/phy/nxp-c45-tja11xx.c b/drivers/net/phy/nxp-c45-tja11xx.c
> index 029875a59ff8..ce718a5865a4 100644
> --- a/drivers/net/phy/nxp-c45-tja11xx.c
> +++ b/drivers/net/phy/nxp-c45-tja11xx.c
> @@ -31,6 +31,10 @@
>  #define DEVICE_CONTROL_CONFIG_GLOBAL_EN	BIT(14)
>  #define DEVICE_CONTROL_CONFIG_ALL_EN	BIT(13)
>  
> +#define VEND1_PORT_IRQ_ENABLES		0x0072
> +#define PORT1_IRQ			BIT(1)
> +#define GLOBAL_IRQ			BIT(0)

I find the PORT1 confusing there, it suggests there is a port0? What
is port0? There is no other reference to numbered ports in the driver.

> -static bool nxp_c45_poll_txts(struct phy_device *phydev)
> +static bool nxp_c45_poll(struct phy_device *phydev)
>  {
>  	return phydev->irq <= 0;
>  }

Maybe as a new patch, but phy_interrupt_is_valid() can be used here.

Maybe also extend the commit message to include a comment that
functions names are changed to reflect that all interrupts are now
disabled, not just _txts interrupts. Otherwise this rename might be
considered an unrelated change.

> @@ -448,7 +452,7 @@ static void nxp_c45_process_txts(struct nxp_c45_phy *priv,
>  static long nxp_c45_do_aux_work(struct ptp_clock_info *ptp)
>  {
>  	struct nxp_c45_phy *priv = container_of(ptp, struct nxp_c45_phy, caps);
> -	bool poll_txts = nxp_c45_poll_txts(priv->phydev);
> +	bool poll_txts = nxp_c45_poll(priv->phydev);
>  	struct skb_shared_hwtstamps *shhwtstamps_rx;
>  	struct ptp_clock_event event;
>  	struct nxp_c45_hwts hwts;
> @@ -699,7 +703,7 @@ static void nxp_c45_txtstamp(struct mii_timestamper *mii_ts,
>  		NXP_C45_SKB_CB(skb)->header = ptp_parse_header(skb, type);
>  		skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
>  		skb_queue_tail(&priv->tx_queue, skb);
> -		if (nxp_c45_poll_txts(priv->phydev))
> +		if (nxp_c45_poll(priv->phydev))
>  			ptp_schedule_worker(priv->ptp_clock, 0);
>  		break;
>  	case HWTSTAMP_TX_OFF:
> @@ -772,7 +776,7 @@ static int nxp_c45_hwtstamp(struct mii_timestamper *mii_ts,
>  				 PORT_PTP_CONTROL_BYPASS);
>  	}
>  
> -	if (nxp_c45_poll_txts(priv->phydev))
> +	if (nxp_c45_poll(priv->phydev))
>  		goto nxp_c45_no_ptp_irq;
>  
>  	if (priv->hwts_tx)
> @@ -892,10 +896,12 @@ static int nxp_c45_config_intr(struct phy_device *phydev)
>  {
>  	if (phydev->interrupts == PHY_INTERRUPT_ENABLED)
>  		return phy_set_bits_mmd(phydev, MDIO_MMD_VEND1,
> -					VEND1_PHY_IRQ_EN, PHY_IRQ_LINK_EVENT);
> +					VEND1_PORT_IRQ_ENABLES,
> +					PORT1_IRQ | GLOBAL_IRQ);
>  	else
>  		return phy_clear_bits_mmd(phydev, MDIO_MMD_VEND1,
> -					  VEND1_PHY_IRQ_EN, PHY_IRQ_LINK_EVENT);
> +					  VEND1_PORT_IRQ_ENABLES,
> +					  PORT1_IRQ | GLOBAL_IRQ);
>  }
>  
>  static irqreturn_t nxp_c45_handle_interrupt(struct phy_device *phydev)
> @@ -1290,6 +1296,10 @@ static int nxp_c45_config_init(struct phy_device *phydev)
>  	phy_set_bits_mmd(phydev, MDIO_MMD_VEND1, VEND1_PORT_FUNC_ENABLES,
>  			 PTP_ENABLE);
>  
> +	if (!nxp_c45_poll(phydev))
> +		phy_set_bits_mmd(phydev, MDIO_MMD_VEND1,
> +				 VEND1_PHY_IRQ_EN, PHY_IRQ_LINK_EVENT);
> +

It seems odd to be touching interrupt configuration outside of
nxp_c45_config_intr(). Is there a reason this cannot be part of
phydev->interrupts == PHY_INTERRUPT_ENABLED ?

	Andrew