Skip to contentSkip to conversation

spi: sh: Switch to using core message queue

PATCH In review Mark Brown 1 participant 2 messages Jun 10, 2022 – Jun 27, 2022 SPI

lore
Share & tools
lore

Mark Brown

author maintainer
#0 lore raw

PATCH 1/1 spi: sh: Switch to using core message queue

We deprecated open coding of the transfer queue back in 2017 so it's high time we finished up converting drivers to use the standard message queue code. The SH driver is fairly straightforward so convert to use transfer_one_message(), it looks like the driver would be a good fit for transfer_one() with a little bit of updating but this smaller change seems safer.

I'm not actually clear how the driver worked robustly previously, it clears SSA and CR1 when queueing a transfer which looks like it would interfere with any running transfer. This clearing has been moved to the start of the message transfer function.

I'm also unclear how exactly the chip select is managed with this driver.

---
 drivers/spi/spi-sh.c | 94 +++++++++++++-------------------------------
 1 file changed, 28 insertions(+), 66 deletions(-)
drivers/spi/spi-sh.c +28 -66
index 45f304935332..3e72fad99adf 100644
@@ -73,11 +73,8 @@ struct spi_sh_data {
 	void __iomem *addr;
 	int irq;
 	struct spi_master *master;
-	struct list_head queue;
-	struct work_struct ws;
 	unsigned long cr1;
 	wait_queue_head_t wait;
-	spinlock_t lock;
 	int width;
 };
 
@@ -271,47 +268,39 @@ static int spi_sh_receive(struct spi_sh_data *ss, struct spi_message *mesg,
 	return 0;
 }
 
-static void spi_sh_work(struct work_struct *work)
+static int spi_sh_transfer_one_message(struct spi_controller *ctlr,
+					struct spi_message *mesg)
 {
-	struct spi_sh_data *ss = container_of(work, struct spi_sh_data, ws);
-	struct spi_message *mesg;
+	struct spi_sh_data *ss = spi_controller_get_devdata(ctlr);
 	struct spi_transfer *t;
-	unsigned long flags;
 	int ret;
 
 	pr_debug("%s: enter\n", __func__);
 
-	spin_lock_irqsave(&ss->lock, flags);
-	while (!list_empty(&ss->queue)) {
-		mesg = list_entry(ss->queue.next, struct spi_message, queue);
-		list_del_init(&mesg->queue);
-
-		spin_unlock_irqrestore(&ss->lock, flags);
-		list_for_each_entry(t, &mesg->transfers, transfer_list) {
-			pr_debug("tx_buf = %p, rx_buf = %p\n",
-					t->tx_buf, t->rx_buf);
-			pr_debug("len = %d, delay.value = %d\n",
-					t->len, t->delay.value);
-
-			if (t->tx_buf) {
-				ret = spi_sh_send(ss, mesg, t);
-				if (ret < 0)
-					goto error;
-			}
-			if (t->rx_buf) {
-				ret = spi_sh_receive(ss, mesg, t);
-				if (ret < 0)
-					goto error;
-			}
-			mesg->actual_length += t->len;
-		}
-		spin_lock_irqsave(&ss->lock, flags);
+	spi_sh_clear_bit(ss, SPI_SH_SSA, SPI_SH_CR1);
 
-		mesg->status = 0;
-		if (mesg->complete)
-			mesg->complete(mesg->context);
+	list_for_each_entry(t, &mesg->transfers, transfer_list) {
+		pr_debug("tx_buf = %p, rx_buf = %p\n",
+			 t->tx_buf, t->rx_buf);
+		pr_debug("len = %d, delay.value = %d\n",
+			 t->len, t->delay.value);
+
+		if (t->tx_buf) {
+			ret = spi_sh_send(ss, mesg, t);
+			if (ret < 0)
+				goto error;
+		}
+		if (t->rx_buf) {
+			ret = spi_sh_receive(ss, mesg, t);
+			if (ret < 0)
+				goto error;
+		}
+		mesg->actual_length += t->len;
 	}
 
+	mesg->status = 0;
+	spi_finalize_current_message(ctlr);
+
 	clear_fifo(ss);
 	spi_sh_set_bit(ss, SPI_SH_SSD, SPI_SH_CR1);
 	udelay(100);
@@ -321,12 +310,11 @@ static void spi_sh_work(struct work_struct *work)
 
 	clear_fifo(ss);
 
-	spin_unlock_irqrestore(&ss->lock, flags);
-
-	return;
+	return 0;
 
  error:
 	mesg->status = ret;
+	spi_finalize_current_message(ctlr);
 	if (mesg->complete)
 		mesg->complete(mesg->context);
 
@@ -334,6 +322,7 @@ static void spi_sh_work(struct work_struct *work)
 			 SPI_SH_CR1);
 	clear_fifo(ss);
 
+	return ret;
 }
 
 static int spi_sh_setup(struct spi_device *spi)
@@ -355,29 +344,6 @@ static int spi_sh_setup(struct spi_device *spi)
 	return 0;
 }
 
-static int spi_sh_transfer(struct spi_device *spi, struct spi_message *mesg)
-{
-	struct spi_sh_data *ss = spi_master_get_devdata(spi->master);
-	unsigned long flags;
-
-	pr_debug("%s: enter\n", __func__);
-	pr_debug("\tmode = %02x\n", spi->mode);
-
-	spin_lock_irqsave(&ss->lock, flags);
-
-	mesg->actual_length = 0;
-	mesg->status = -EINPROGRESS;
-
-	spi_sh_clear_bit(ss, SPI_SH_SSA, SPI_SH_CR1);
-
-	list_add_tail(&mesg->queue, &ss->queue);
-	schedule_work(&ss->ws);
-
-	spin_unlock_irqrestore(&ss->lock, flags);
-
-	return 0;
-}
-
 static void spi_sh_cleanup(struct spi_device *spi)
 {
 	struct spi_sh_data *ss = spi_master_get_devdata(spi->master);
@@ -416,7 +382,6 @@ static int spi_sh_remove(struct platform_device *pdev)
 	struct spi_sh_data *ss = platform_get_drvdata(pdev);
 
 	spi_unregister_master(ss->master);
-	flush_work(&ss->ws);
 	free_irq(ss->irq, ss);
 
 	return 0;
@@ -467,9 +432,6 @@ static int spi_sh_probe(struct platform_device *pdev)
 		dev_err(&pdev->dev, "ioremap error.\n");
 		return -ENOMEM;
 	}
-	INIT_LIST_HEAD(&ss->queue);
-	spin_lock_init(&ss->lock);
-	INIT_WORK(&ss->ws, spi_sh_work);
 	init_waitqueue_head(&ss->wait);
 
 	ret = request_irq(irq, spi_sh_irq, 0, "spi_sh", ss);
@@ -481,7 +443,7 @@ static int spi_sh_probe(struct platform_device *pdev)
 	master->num_chipselect = 2;
 	master->bus_num = pdev->id;
 	master->setup = spi_sh_setup;
-	master->transfer = spi_sh_transfer;
+	master->transfer_one_message = spi_sh_transfer_one_message;
 	master->cleanup = spi_sh_cleanup;
 
 	ret = spi_register_master(master);
Signature · 1 line
2.30.2
author maintainer
#1 lore raw
Quoted 8 lines · Mark Brown

We deprecated open coding of the transfer queue back in 2017 so it's high time we finished up converting drivers to use the standard message queue code. The SH driver is fairly straightforward so convert to use transfer_one_message(), it looks like the driver would be a good fit for transfer_one() with a little bit of updating but this smaller change seems safer.

[...]

Applied to

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git for-next

Thanks!

[1/1] spi: sh: Switch to using core message queue

      commit: e2185072a4a4786eb46be046cf20494c08dcc78f

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark