Skip to contentSkip to conversation

Mainline commit "serial: Test for no tx data on tx restart" not correct for i.MX6?

DISCUSSION Dirk Behme 3 participants 6 messages Jul 23, 2014 – Aug 6, 2014

Subsystems (2)
Arm i.MX and Layerscape TTY layer and serial drivers
lore
Share & tools
lore

Dirk Behme

author
#0 lore raw

Hi,

this is a question regarding the i.MX6 part (drivers/tty/serial/imx.c) of the commit

serial: Test for no tx data on tx restart

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c557d392fbf5badd693ea1946a4317c87a26a716

Talking with some i.MX6 experts about this, I've got the comment

-- cut --
The imx serial driver part of this commit isn't correct as imx.c sends x_char in irq handler, not in imx_start_tx(), -- cut --

What do you think?

Best regards

Dirk

P.S.: I don't have the initial post in my inbox, so sorry for the new thread.

#1 lore raw

Hi Dirk,

Quoted 8 lines · Dirk Behme

Hi,

this is a question regarding the i.MX6 part (drivers/tty/serial/imx.c) of the commit

serial: Test for no tx data on tx restart

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c557d392fbf5badd693ea1946a4317c87a26a716

Quoted 8 lines

Talking with some i.MX6 experts about this, I've got the comment

-- cut --
The imx serial driver part of this commit isn't correct as imx.c sends x_char in irq handler, not in imx_start_tx(), -- cut --

What do you think?

Thanks for the comment.

Yeah, I missed that the imx driver handled x_char _at all_, because how the imx driver is handling x_char looks broken.

For example, if data is already in the tx ring buffer, imx_start_tx() will send that data before the x_char, and if all the tx ring buffer data is sent successfully, the tx interrupt is switched back off, so the x_char won't be sent.

Also, the imx driver doesn't send x_char in dma mode?

Regards,
Peter Hurley

#2 lore raw

Hi Peter

Quoted 21 lines · Peter Hurley

Hi Dirk,

Quoted 8 lines · Dirk Behme

Hi,

this is a question regarding the i.MX6 part (drivers/tty/serial/imx.c) of the commit

serial: Test for no tx data on tx restart

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c557d392fbf5badd693ea1946a4317c87a26a716

Quoted 8 lines

Talking with some i.MX6 experts about this, I've got the comment

-- cut --
The imx serial driver part of this commit isn't correct as imx.c sends x_char in irq handler, not in imx_start_tx(), -- cut --

What do you think?

Thanks for the comment.

Quoted 7 lines

Yeah, I missed that the imx driver handled x_char _at all_, because how the imx driver is handling x_char looks broken.

For example, if data is already in the tx ring buffer, imx_start_tx() will send that data before the x_char, and if all the tx ring buffer data is sent successfully, the tx interrupt is switched back off, so the x_char won't be sent.

imx_start_tx() doesn't do any actual data transfer, it only switches tx interrupt on, and tx interrupt does the actual data transfer, so if x_char is pending, it will always be sent before the data in ring buffer.

Quoted 1 line

Also, the imx driver doesn't send x_char in dma mode?

Yes, I think so.

Thanks,
Jiada

Quoted 3 lines

Regards,
Peter Hurley

#3 lore raw

Hi Jiada,

Quoted 27 lines · jiwang

Hi Peter

Quoted 24 lines · Peter Hurley

Hi Dirk,

Quoted 8 lines · Dirk Behme

Hi,

this is a question regarding the i.MX6 part (drivers/tty/serial/imx.c) of the commit

serial: Test for no tx data on tx restart

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c557d392fbf5badd693ea1946a4317c87a26a716

Quoted 8 lines

Talking with some i.MX6 experts about this, I've got the comment

-- cut --
The imx serial driver part of this commit isn't correct as imx.c sends x_char in irq handler, not in imx_start_tx(), -- cut --

What do you think?

Thanks for the comment.

Yeah, I missed that the imx driver handled x_char _at all_, because how the imx driver is handling x_char looks broken.

Quoted 6 lines
Quoted 4 lines

For example, if data is already in the tx ring buffer, imx_start_tx() will send that data before the x_char, and if all the tx ring buffer data is sent successfully, the tx interrupt is switched back off, so the x_char won't be sent.

imx_start_tx() doesn't do any actual data transfer

???

   565	static void imx_start_tx(struct uart_port *port)
   566	{
   ...
   610	
   611		if (readl(sport->port.membase + uts_reg(sport)) & UTS_TXEMPTY)
   612			imx_transmit_buffer(sport);
   613	}


   463	static inline void imx_transmit_buffer(struct imx_port *sport)
   464	{
   465		struct circ_buf *xmit = &sport->port.state->xmit;
   466	
   467		while (!uart_circ_empty(xmit) &&
   468				!(readl(sport->port.membase + uts_reg(sport))
   469					& UTS_TXFULL)) {
   470			/* send xmit->buf[xmit->tail]
   471			 * out the port here */
   472			writel(xmit->buf[xmit->tail], sport->port.membase + URTX0);
   473			xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
   474			sport->port.icount.tx++;
   475		}
   476	
   477		if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
   478			uart_write_wakeup(&sport->port);
   479	
   480		if (uart_circ_empty(xmit))
   481			imx_stop_tx(&sport->port);
   482	}

Regards,
Peter Hurley

#4 lore raw

Hi Peter

Quoted 58 lines · Peter Hurley

Hi Jiada,

Quoted 24 lines · jiwang

Hi Peter

Quoted 21 lines · Peter Hurley

Hi Dirk,

Quoted 8 lines · Dirk Behme

Hi,

this is a question regarding the i.MX6 part (drivers/tty/serial/imx.c) of the commit

serial: Test for no tx data on tx restart

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c557d392fbf5badd693ea1946a4317c87a26a716

Quoted 8 lines

Talking with some i.MX6 experts about this, I've got the comment

-- cut --
The imx serial driver part of this commit isn't correct as imx.c sends x_char in irq handler, not in imx_start_tx(), -- cut --

What do you think?

Thanks for the comment.

Quoted 8 lines
Quoted 7 lines

Yeah, I missed that the imx driver handled x_char _at all_, because how the imx driver is handling x_char looks broken.

For example, if data is already in the tx ring buffer, imx_start_tx() will send that data before the x_char, and if all the tx ring buffer data is sent successfully, the tx interrupt is switched back off, so the x_char won't be sent.

imx_start_tx() doesn't do any actual data transfer

???

    565	static void imx_start_tx(struct uart_port *port)
    566	{
    ...
    610	
    611		if (readl(sport->port.membase + uts_reg(sport)) & UTS_TXEMPTY)
    612			imx_transmit_buffer(sport);
    613	}


    463	static inline void imx_transmit_buffer(struct imx_port *sport)
    464	{
    465		struct circ_buf *xmit = &sport->port.state->xmit;
    466	
    467		while (!uart_circ_empty(xmit) &&
    468				!(readl(sport->port.membase + uts_reg(sport))
    469					& UTS_TXFULL)) {
    470			/* send xmit->buf[xmit->tail]
    471			 * out the port here */
    472			writel(xmit->buf[xmit->tail], sport->port.membase + URTX0);
    473			xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
    474			sport->port.icount.tx++;
Quoted 8 lines
    475		}
    476	
    477		if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
    478			uart_write_wakeup(&sport->port);
    479	
    480		if (uart_circ_empty(xmit))
    481			imx_stop_tx(&sport->port);
    482	}

sorry, our imx driver has been modified differently,
yes, you are right, mainline imx driver's handling of x_char looks broken

@Shijie,
do you have any comment?

Thanks,
Jiada

Quoted 2 lines

Regards,
Peter Hurley

#5 lore raw
Quoted 23 lines · jiwang
Quoted 22 lines · Peter Hurley
Quoted 21 lines · jiwang
Quoted 20 lines · Peter Hurley
Quoted 8 lines · Dirk Behme

Hi,

this is a question regarding the i.MX6 part (drivers/tty/serial/imx.c) of the commit

serial: Test for no tx data on tx restart

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=c557d392fbf5badd693ea1946a4317c87a26a716

Quoted 8 lines

Talking with some i.MX6 experts about this, I've got the comment

-- cut --
The imx serial driver part of this commit isn't correct as imx.c sends x_char in irq handler, not in imx_start_tx(), -- cut --

What do you think?

Yeah, I missed that the imx driver handled x_char _at all_, because how the imx driver is handling x_char looks broken.

[...]

Quoted 2 lines

sorry, our imx driver has been modified differently,
yes, you are right, mainline imx driver's handling of x_char looks broken

Jiada or Dirk or whomever,

Would someone please test the patch below on actual imx hardware?

You can download a standalone x_char unit test from
http://blog.hurleysoftware.com/uploads/tty_xchar.c

Instructions for building, setting up and testing are included in the file.

--- >% ---
Subject: [PATCH] serial: imx: Fix x_char handling and tx flow control

The serial core expects the UART driver to transmit x_char (START/STOP chars) even if tx is stopped and before data already in the tx ring buffer if possible. Also, sending x_char must not cause additional data in the tx ring buffer to transmit if tx is stopped.

Cause x_char to be transmitted before any other data is sent. Auto-stop tx if the tx ring buffer is empty or tx should be stopped. Only perform one write wakeup if tx ring buffer space is below threshold.

x_char handling in DMA mode is still broken; add FIXME.

Signed-off-by Peter Hurley <[email protected]>
---
 drivers/tty/serial/imx.c | 39 ++++++++++++++++-----------------------
 1 file changed, 16 insertions(+), 23 deletions(-)
drivers/tty/serial/imx.c +16 -23
index be1c842..00a37e9 100644
@@ -460,9 +460,19 @@ static inline void imx_transmit_buffer(struct imx_port *sport)
 {
 	struct circ_buf *xmit = &sport->port.state->xmit;
 
+	if (sport->port.x_char) {
+		/* Send next char */
+		writel(sport->port.x_char, sport->port.membase + URTX0);
+		return;
+	}
+
+	if (uart_circ_empty(xmit) || uart_tx_stopped(&sport->port)) {
+		imx_stop_tx(&sport->port);
+		return;
+	}
+
 	while (!uart_circ_empty(xmit) &&
-			!(readl(sport->port.membase + uts_reg(sport))
-				& UTS_TXFULL)) {
+	       !(readl(sport->port.membase + uts_reg(sport)) & UTS_TXFULL)) {
 		/* send xmit->buf[xmit->tail]
 		 * out the port here */
 		writel(xmit->buf[xmit->tail], sport->port.membase + URTX0);

@@ -563,9 +573,6 @@ static void imx_start_tx(struct uart_port *port)
 	struct imx_port *sport = (struct imx_port *)port;
 	unsigned long temp;
 
-	if (uart_circ_empty(&port->state->xmit))
-		return;
-
 	if (USE_IRDA(sport)) {
 		/* half duplex in IrDA mode; have to disable receive mode */
 		temp = readl(sport->port.membase + UCR4);

@@ -600,7 +607,10 @@ static void imx_start_tx(struct uart_port *port)
 	}
 
 	if (sport->dma_is_enabled) {
-		imx_dma_tx(sport);
+		/* FIXME: port->x_char must be transmitted if != 0 */
+		if (!uart_circ_empty(&port->state->xmit) &&
+		    !uart_tx_stopped(port))
+			imx_dma_tx(sport);
 		return;
 	}
 

@@ -628,27 +638,10 @@ static irqreturn_t imx_rtsint(int irq, void *dev_id)
 static irqreturn_t imx_txint(int irq, void *dev_id)
 {
 	struct imx_port *sport = dev_id;
-	struct circ_buf *xmit = &sport->port.state->xmit;
 	unsigned long flags;
 
 	spin_lock_irqsave(&sport->port.lock, flags);
-	if (sport->port.x_char) {
-		/* Send next char */
-		writel(sport->port.x_char, sport->port.membase + URTX0);
-		goto out;
-	}
-
-	if (uart_circ_empty(xmit) || uart_tx_stopped(&sport->port)) {
-		imx_stop_tx(&sport->port);
-		goto out;
-	}
-
 	imx_transmit_buffer(sport);
-
-	if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
-		uart_write_wakeup(&sport->port);
-
-out:
 	spin_unlock_irqrestore(&sport->port.lock, flags);
 	return IRQ_HANDLED;
 }
Signature · 1 line
2.0.4