Received-SPF: pass (google.com: domain of linux-kernel+bounces-69943-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Date: Sat, 17 Feb 2024 16:42:40 +0100
From: Andrew Lunn <andrew@lunn.ch>
To: Ciprian Regus <ciprian.regus@analog.com>
Cc: linux-kernel@vger.kernel.org, "David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Simon Horman <horms@kernel.org>,
	Dell Jin <dell.jin.code@outlook.com>,
	Amit Kumar Mahapatra <amit.kumar-mahapatra@amd.com>,
	Yang Yingliang <yangyingliang@huawei.com>, netdev@vger.kernel.org
Subject: Re: [net-next] net: ethernet: adi: adin1110: Reduce the MDIO_TRDONE
 poll interval
Message-ID: <b3c8d287-3715-4c03-be3e-e94253f6d2ac@lunn.ch>
References: <20240216103636.1231815-1-ciprian.regus@analog.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240216103636.1231815-1-ciprian.regus@analog.com>

On Fri, Feb 16, 2024 at 12:36:32PM +0200, Ciprian Regus wrote:
> In order to do a clause 22 access to the PHY registers of the ADIN1110,
> we have to write the MDIO frame to the ADIN1110_MDIOACC register, and
> then poll the MDIO_TRDONE bit (for a 1) in the same register. The
> device will set this bit to 1 once the internal MDIO transaction is
> done. In practice, this bit takes ~50 - 60 us to be set.
> 
> The first attempt to poll the bit is right after the ADIN1110_MDIOACC
> register is written, so it will always be read as 0. The next check will
> only be done after 10 ms, which will result in the MDIO transactions
> taking a long time to complete. Reduce this polling interval to 100 us.
> 
> Signed-off-by: Ciprian Regus <ciprian.regus@analog.com>
> ---
>  drivers/net/ethernet/adi/adin1110.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/adi/adin1110.c b/drivers/net/ethernet/adi/adin1110.c
> index d7c274af6d4d..6fca19e6ae67 100644
> --- a/drivers/net/ethernet/adi/adin1110.c
> +++ b/drivers/net/ethernet/adi/adin1110.c
> @@ -467,3 +467,3 @@ static int adin1110_mdio_read(struct mii_bus *bus, int phy_id, int reg)
>  	ret = readx_poll_timeout(adin1110_read_mdio_acc, priv, val,
> -				 (val & ADIN1110_MDIO_TRDONE), 10000, 30000);
> +				 (val & ADIN1110_MDIO_TRDONE), 100, 30000);
>  	if (ret < 0)
> @@ -498,3 +498,3 @@ static int adin1110_mdio_write(struct mii_bus *bus, int phy_id,
>  	return readx_poll_timeout(adin1110_read_mdio_acc, priv, val,
> -				  (val & ADIN1110_MDIO_TRDONE), 10000, 30000);
> +				  (val & ADIN1110_MDIO_TRDONE), 100, 30000);

The kernel can have trouble sleeping for such short times. It might
make sense to swap to the _atomic version which spins rather than
sleeps. If you are just reading a few PHY registers every so often, it
might not be worth it. But if you have an Ethernet switch and need to
access a lot of registers, it could make it a bit faster.

       Andrew