Date:   Mon, 29 Nov 2021 20:01:11 +0100
From:   Wolfram Sang <wsa@kernel.org>
To:     Chris Packham <chris.packham@alliedtelesis.co.nz>
Cc:     rjui@broadcom.com, sbranden@broadcom.com,
        bcm-kernel-feedback-list@broadcom.com, linux-i2c@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
        Richard Laing <richard.laing@alliedtelesis.co.nz>
Subject: Re: [PATCH v3] i2c: bcm-iproc: Add i2c recovery support
Message-ID: <YaUjd747t94Ge/VA@kunai>
Mail-Followup-To: Wolfram Sang <wsa@kernel.org>,
        Chris Packham <chris.packham@alliedtelesis.co.nz>,
        rjui@broadcom.com, sbranden@broadcom.com,
        bcm-kernel-feedback-list@broadcom.com, linux-i2c@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org,
        Richard Laing <richard.laing@alliedtelesis.co.nz>
References: <20210603052515.11178-1-chris.packham@alliedtelesis.co.nz>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
        protocol="application/pgp-signature"; boundary="KsO1SrJz6NmBFB0C"
Content-Disposition: inline
In-Reply-To: <20210603052515.11178-1-chris.packham@alliedtelesis.co.nz>
Precedence: bulk


--KsO1SrJz6NmBFB0C
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Hi Chris,

sorry for the long delay.

All looks good except for this:

> @@ -806,6 +954,7 @@ static int bcm_iproc_i2c_xfer_wait(struct bcm_iproc_i=
2c_dev *iproc_i2c,
>  		/* flush both TX/RX FIFOs */
>  		val =3D BIT(M_FIFO_RX_FLUSH_SHIFT) | BIT(M_FIFO_TX_FLUSH_SHIFT);
>  		iproc_i2c_wr_reg(iproc_i2c, M_FIFO_CTRL_OFFSET, val);
> +		iproc_i2c_lockup_recover(iproc_i2c);
>  		return -ETIMEDOUT;
>  	}
> =20
> @@ -814,6 +963,7 @@ static int bcm_iproc_i2c_xfer_wait(struct bcm_iproc_i=
2c_dev *iproc_i2c,
>  		/* flush both TX/RX FIFOs */
>  		val =3D BIT(M_FIFO_RX_FLUSH_SHIFT) | BIT(M_FIFO_TX_FLUSH_SHIFT);
>  		iproc_i2c_wr_reg(iproc_i2c, M_FIFO_CTRL_OFFSET, val);
> +		iproc_i2c_lockup_recover(iproc_i2c);
>  		return ret;
>  	}

I2C specs suggest recovery at the beginning of a transfer when SDA is
detected low. Not at the end of a transfer.

Happy hacking,

   Wolfram


--KsO1SrJz6NmBFB0C
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
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=qV9d
-----END PGP SIGNATURE-----

--KsO1SrJz6NmBFB0C--