Received-SPF: pass (google.com: domain of linux-kernel+bounces-72858-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Message-ID: <f0ec7c23aabc531f609f57b5e7ffb0c00047dfc4.camel@gmail.com>
Subject: Re: [PATCH v2 2/5] spi: move splitting transfers to
 spi_optimize_message()
From: Nuno =?ISO-8859-1?Q?S=E1?= <noname.nuno@gmail.com>
To: David Lechner <dlechner@baylibre.com>, Mark Brown <broonie@kernel.org>
Cc: Martin Sperl <kernel@martin.sperl.org>, David Jander
 <david@protonic.nl>,  Jonathan Cameron <jic23@kernel.org>, Michael
 Hennerich <michael.hennerich@analog.com>, Nuno =?ISO-8859-1?Q?S=E1?=
 <nuno.sa@analog.com>, Alain Volmat <alain.volmat@foss.st.com>, Maxime
 Coquelin <mcoquelin.stm32@gmail.com>, Alexandre Torgue
 <alexandre.torgue@foss.st.com>,  linux-spi@vger.kernel.org,
 linux-kernel@vger.kernel.org,  linux-stm32@st-md-mailman.stormreply.com,
 linux-arm-kernel@lists.infradead.org,  linux-iio@vger.kernel.org, Jonathan
 Cameron <Jonathan.Cameron@huawei.com>,  Julien Stephan
 <jstephan@baylibre.com>
Date: Tue, 20 Feb 2024 11:52:59 +0100
In-Reply-To: <20240219-mainline-spi-precook-message-v2-2-4a762c6701b9@baylibre.com>
References: 
	<20240219-mainline-spi-precook-message-v2-0-4a762c6701b9@baylibre.com>
	 <20240219-mainline-spi-precook-message-v2-2-4a762c6701b9@baylibre.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.50.3 (3.50.3-1.fc39) 
Precedence: bulk
MIME-Version: 1.0

On Mon, 2024-02-19 at 16:33 -0600, David Lechner wrote:
> Splitting transfers is an expensive operation so we can potentially
> optimize it by doing it only once per optimization of the message
> instead of repeating each time the message is transferred.
>=20
> The transfer splitting functions are currently the only user of
> spi_res_alloc() so spi_res_release() can be safely moved at this time
> from spi_finalize_current_message() to spi_unoptimize_message().
>=20
> The doc comments of the public functions for splitting transfers are
> also updated so that callers will know when it is safe to call them
> to ensure proper resource management.
>=20
> Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
> Signed-off-by: David Lechner <dlechner@baylibre.com>
> ---

Acked-by: Nuno Sa <nuno.sa@analog.com>

>=20
> v2 changes:
> - Changed line break for multiline if condition
> - Removed kernel doc inclusion (/** -> /*) from static members
> - Picked up Jonathan's Reviewed-by
>=20
> =C2=A0drivers/spi/spi.c | 110 +++++++++++++++++++++++++++++++++----------=
-----------
> =C2=A01 file changed, 68 insertions(+), 42 deletions(-)
>=20
> diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
> index f68d92b57543..ba4d3fde2054 100644
> --- a/drivers/spi/spi.c
> +++ b/drivers/spi/spi.c
> @@ -1747,38 +1747,6 @@ static int __spi_pump_transfer_message(struct spi_=
controller
> *ctlr,
> =C2=A0
> =C2=A0	trace_spi_message_start(msg);
> =C2=A0
> -	/*
> -	 * If an SPI controller does not support toggling the CS line on each
> -	 * transfer (indicated by the SPI_CS_WORD flag) or we are using a GPIO
> -	 * for the CS line, we can emulate the CS-per-word hardware function by
> -	 * splitting transfers into one-word transfers and ensuring that
> -	 * cs_change is set for each transfer.
> -	 */
> -	if ((msg->spi->mode & SPI_CS_WORD) && (!(ctlr->mode_bits & SPI_CS_WORD)=
 ||
> -					=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 spi_is_csgpiod(msg->spi))) {
> -		ret =3D spi_split_transfers_maxwords(ctlr, msg, 1);
> -		if (ret) {
> -			msg->status =3D ret;
> -			spi_finalize_current_message(ctlr);
> -			return ret;
> -		}
> -
> -		list_for_each_entry(xfer, &msg->transfers, transfer_list) {
> -			/* Don't change cs_change on the last entry in the list */
> -			if (list_is_last(&xfer->transfer_list, &msg->transfers))
> -				break;
> -			xfer->cs_change =3D 1;
> -		}
> -	} else {
> -		ret =3D spi_split_transfers_maxsize(ctlr, msg,
> -						=C2=A0 spi_max_transfer_size(msg-
> >spi));
> -		if (ret) {
> -			msg->status =3D ret;
> -			spi_finalize_current_message(ctlr);
> -			return ret;
> -		}
> -	}
> -
> =C2=A0	if (ctlr->prepare_message) {
> =C2=A0		ret =3D ctlr->prepare_message(ctlr, msg);
> =C2=A0		if (ret) {
> @@ -2124,6 +2092,8 @@ static void __spi_unoptimize_message(struct spi_mes=
sage *msg)
> =C2=A0	if (ctlr->unoptimize_message)
> =C2=A0		ctlr->unoptimize_message(msg);
> =C2=A0
> +	spi_res_release(ctlr, msg);
> +
> =C2=A0	msg->optimized =3D false;
> =C2=A0	msg->opt_state =3D NULL;
> =C2=A0}
> @@ -2169,15 +2139,6 @@ void spi_finalize_current_message(struct spi_contr=
oller
> *ctlr)
> =C2=A0
> =C2=A0	spi_unmap_msg(ctlr, mesg);
> =C2=A0
> -	/*
> -	 * In the prepare_messages callback the SPI bus has the opportunity
> -	 * to split a transfer to smaller chunks.
> -	 *
> -	 * Release the split transfers here since spi_map_msg() is done on
> -	 * the split transfers.
> -	 */
> -	spi_res_release(ctlr, mesg);
> -
> =C2=A0	if (mesg->prepared && ctlr->unprepare_message) {
> =C2=A0		ret =3D ctlr->unprepare_message(ctlr, mesg);
> =C2=A0		if (ret) {
> @@ -3819,6 +3780,10 @@ static int __spi_split_transfer_maxsize(struct
> spi_controller *ctlr,
> =C2=A0 * @msg:=C2=A0=C2=A0 the @spi_message to transform
> =C2=A0 * @maxsize:=C2=A0 the maximum when to apply this
> =C2=A0 *
> + * This function allocates resources that are automatically freed during=
 the
> + * spi message unoptimize phase so this function should only be called f=
rom
> + * optimize_message callbacks.
> + *
> =C2=A0 * Return: status of transformation
> =C2=A0 */
> =C2=A0int spi_split_transfers_maxsize(struct spi_controller *ctlr,
> @@ -3857,6 +3822,10 @@ EXPORT_SYMBOL_GPL(spi_split_transfers_maxsize);
> =C2=A0 * @msg:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 the @spi_message to transfor=
m
> =C2=A0 * @maxwords: the number of words to limit each transfer to
> =C2=A0 *
> + * This function allocates resources that are automatically freed during=
 the
> + * spi message unoptimize phase so this function should only be called f=
rom
> + * optimize_message callbacks.
> + *
> =C2=A0 * Return: status of transformation
> =C2=A0 */
> =C2=A0int spi_split_transfers_maxwords(struct spi_controller *ctlr,
> @@ -4231,6 +4200,57 @@ static int __spi_validate(struct spi_device *spi, =
struct
> spi_message *message)
> =C2=A0	return 0;
> =C2=A0}
> =C2=A0
> +/*
> + * spi_split_transfers - generic handling of transfer splitting
> + * @msg: the message to split
> + *
> + * Under certain conditions, a SPI controller may not support arbitrary
> + * transfer sizes or other features required by a peripheral. This funct=
ion
> + * will split the transfers in the message into smaller transfers that a=
re
> + * supported by the controller.
> + *
> + * Controllers with special requirements not covered here can also split
> + * transfers in the optimize_message() callback.
> + *
> + * Context: can sleep
> + * Return: zero on success, else a negative error code
> + */
> +static int spi_split_transfers(struct spi_message *msg)
> +{
> +	struct spi_controller *ctlr =3D msg->spi->controller;
> +	struct spi_transfer *xfer;
> +	int ret;
> +
> +	/*
> +	 * If an SPI controller does not support toggling the CS line on each
> +	 * transfer (indicated by the SPI_CS_WORD flag) or we are using a GPIO
> +	 * for the CS line, we can emulate the CS-per-word hardware function by
> +	 * splitting transfers into one-word transfers and ensuring that
> +	 * cs_change is set for each transfer.
> +	 */
> +	if ((msg->spi->mode & SPI_CS_WORD) &&
> +	=C2=A0=C2=A0=C2=A0 (!(ctlr->mode_bits & SPI_CS_WORD) || spi_is_csgpiod(=
msg->spi))) {
> +		ret =3D spi_split_transfers_maxwords(ctlr, msg, 1);
> +		if (ret)
> +			return ret;
> +
> +		list_for_each_entry(xfer, &msg->transfers, transfer_list) {
> +			/* Don't change cs_change on the last entry in the list */
> +			if (list_is_last(&xfer->transfer_list, &msg->transfers))
> +				break;
> +
> +			xfer->cs_change =3D 1;
> +		}
> +	} else {
> +		ret =3D spi_split_transfers_maxsize(ctlr, msg,
> +						=C2=A0 spi_max_transfer_size(msg-
> >spi));
> +		if (ret)
> +			return ret;
> +	}
> +
> +	return 0;
> +}
> +
> =C2=A0/*
> =C2=A0 * __spi_optimize_message - shared implementation for spi_optimize_=
message()
> =C2=A0 *=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=
=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0 and spi_maybe_optimize_message()
> @@ -4254,10 +4274,16 @@ static int __spi_optimize_message(struct spi_devi=
ce *spi,
> =C2=A0	if (ret)
> =C2=A0		return ret;
> =C2=A0
> +	ret =3D spi_split_transfers(msg);
> +	if (ret)
> +		return ret;
> +
> =C2=A0	if (ctlr->optimize_message) {
> =C2=A0		ret =3D ctlr->optimize_message(msg);
> -		if (ret)
> +		if (ret) {
> +			spi_res_release(ctlr, msg);
> =C2=A0			return ret;
> +		}
> =C2=A0	}
> =C2=A0
> =C2=A0	msg->optimized =3D true;
>=20