serial: imx: Switch to nbcon console

By waiting at most 1 second for USR2_TXDC to be set, we avoid a potentital deadlock.

In case of the timeout, there is not much we can do, so we simply ignore the transmitter state and optimistically try to continue.

Signed-off-by Esben Haabendal <[email protected]>

---
 drivers/tty/serial/imx.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

drivers/tty/serial/imx.c +2 -1

index 54b760d845c0..f7e4f38f08f3 100644
@@ -26,6 +26,7 @@
 #include <linux/slab.h>
 #include <linux/of.h>
 #include <linux/io.h>
+#include <linux/iopoll.h>
 #include <linux/dma-mapping.h>
 
 #include <asm/irq.h>

@@ -1995,7 +1996,7 @@ imx_uart_console_write(struct console *co, const char *s, unsigned int count)
 	struct imx_port *sport = imx_uart_ports[co->index];
 	struct imx_port_ucrs old_ucr;
 	unsigned long flags;
-	unsigned int ucr1;
+	unsigned int ucr1, usr2;
 	int locked = 1;
 
 	if (sport->port.sysrq)

drivers/tty/serial/imx.c +0 -0

 	imx_uart_ucrs_restore(sport, &old_ucr);
 
 	if (locked)

Signature · 1 line

2.44.0

Quoted 7 lines · Esben Haabendal

By waiting at most 1 second for USR2_TXDC to be set, we avoid a potentital deadlock.

In case of the timeout, there is not much we can do, so we simply ignore the transmitter state and optimistically try to continue.

Signed-off-by Esben Haabendal <[email protected]>

For the next round you should consult "scripts/get_maintainer.pl drivers/tty/serial/imx.c" and add these people/lists on Cc.

Quoted 26 lines

---
 drivers/tty/serial/imx.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

drivers/tty/serial/imx.c +2 -1

index 54b760d845c0..f7e4f38f08f3 100644
@@ -26,6 +26,7 @@
 #include <linux/slab.h>
 #include <linux/of.h>
 #include <linux/io.h>
+#include <linux/iopoll.h>
 #include <linux/dma-mapping.h>
 
 #include <asm/irq.h>

@@ -1995,7 +1996,7 @@ imx_uart_console_write(struct console *co, const char *s, unsigned int count)
 	struct imx_port *sport = imx_uart_ports[co->index];
 	struct imx_port_ucrs old_ucr;
 	unsigned long flags;
-	unsigned int ucr1;
+	unsigned int ucr1, usr2;
 	int locked = 1;
 
 	if (sport->port.sysrq)

drivers/tty/serial/imx.c imx_uart_console_write(struct console *co, const char *s, unsigned int count) · line 2031

Quoted 7 lines

drivers/tty/serial/imx.c +2 -2

@@ -2026,8 +2027,8 @@ imx_uart_console_write(struct console *co, const char *s, unsigned int count)
 	 *	Finally, wait for transmitter to become empty
 	 *	and restore UCR1/2/3
 	 */
-	while (!(imx_uart_readl(sport, USR2) & USR2_TXDC));
-
+	read_poll_timeout_atomic(imx_uart_readl, usr2, usr2 & USR2_TXDC,
+				 0, 1000000, false, sport, USR2);

You can make use of USEC_PER_SEC here.

After fixing this, feel free to add my:

Acked-by Marc Kleine-Budde <[email protected]>

drivers/tty/serial/imx.c +32 -0

index f7e4f38f08f3..2a49880c2651 100644
+static bool imx_uart_console_write_atomic(struct console *co,
+					  struct nbcon_write_context *wctxt)
+{
+	struct imx_port *sport = imx_uart_ports[co->index];
+	struct uart_port *port = &sport->port;
+	struct imx_port_ucrs old_ucr;
+	unsigned int ucr1, usr2;
+
+	if (!nbcon_enter_unsafe(wctxt))
+		return false;
+
+	/*
+	 *	First, save UCR1/2/3 and then disable interrupts
+	 */
+	imx_uart_ucrs_save(sport, &old_ucr);
+	ucr1 = old_ucr.ucr1;
+
+	if (imx_uart_is_imx1(sport))
+		ucr1 |= IMX1_UCR1_UARTCLKEN;
+	ucr1 |= UCR1_UARTEN;
+	ucr1 &= ~(UCR1_TRDYEN | UCR1_RRDYEN | UCR1_RTSDEN);
+
+	imx_uart_writel(sport, ucr1, UCR1);
+
+	imx_uart_writel(sport, old_ucr.ucr2 | UCR2_TXEN, UCR2);
+
+	if (sport->console_newline_needed)
+		uart_console_write(port, "\n", 1, imx_uart_console_putchar);
+	uart_console_write(port, wctxt->outbuf, wctxt->len,
+			   imx_uart_console_putchar);
+
+	/*

In your imx_uart_console_write_atomic() I see lots of register usage:

ucr1, ucr2, ucr3, usr2, uts

It is critical that _all_ usage of these registers throughout the driver is protected, preferably protected by the port lock. Please go through and verify that. If imx_uart_console_write_thread() is using even more registers, you will need to check those as well.

For the 8250 I went through all uses and found several problems [0]. The imx driver may have similar issues.

I have now gone through all usage of above mentioned registers in the driver. I found a single missing lock/unlock, but other than that the driver seems to be in good shape in this regards.

In the _probe() function, the following registers are accessed
unprotected: ucr1, ucr2, ucr3, uts. But I assume this should be safe, as the uart port is not registred yet, and interrupt handlers have not been registered.

I will send out a v3 short, hoping we can get this in a shape where we can get it merged shortly after the 8250 driver nbcon patches.

Quoted 3 lines · Esben Haabendal

I have now gone through all usage of above mentioned registers in the driver. I found a single missing lock/unlock, but other than that the driver seems to be in good shape in this regards.

Great. Please send a separate patch for the missing lock/unlock to mainline. It should be a mainline problem as well.

Quoted 4 lines

In the _probe() function, the following registers are accessed
unprotected: ucr1, ucr2, ucr3, uts. But I assume this should be safe, as the uart port is not registred yet, and interrupt handlers have not been registered.

Sure.

Quoted 2 lines

I will send out a v3 short, hoping we can get this in a shape where we can get it merged shortly after the 8250 driver nbcon patches.

Great!

Note that I am currently pushing the 8250 nbcon to mainline. It is in the review process and there have already been significant changes. It is all 8250-specific changes, so it probably is not related to your work. Just letting you know. I will CC you on the next series submission.

This is a first attempt at porting another serial console driver to nbcon.

The first patch is a preparatory patch, thought to be suitable for mainline inclusion now, while the second patch is for rt only for now.

The second patch does have quite a copy-paste feeling to it.

The imx_uart_console_write_atomic() and imx_uart_console_write_thread() includes copies of code from the old imx_uart_console_write() function. This is similar to current status of nbcon support in 8250 driver.

The midle part of imx_uart_console_write_thread() is identical to the middle part of serial8250_console_write_thread() in drivers/tty/serial/8250/8250_port.c, except for the arguments to uart_console_write() function.

There is clearly potential for writing some common code to avoid both cases of copy-paste code.

I am unsure about what the plan is for keeping the legacy console code in the drivers. Are we going to drop that before submitting the code to mainline?