From: Hugo Villeneuve <hvilleneuve@dimonoff.com>

With this driver, it is very hard to debug the registers using
the /sys/kernel/debug/regmap interface.

The main reason is that bits 0 and 1 of the register address
correspond to the channels bits, so the register address itself starts
at bit 2, so we must 'mentally' shift each register address by 2 bits
to get its offset.

Also, only channels 0 and 1 are supported, so combinations of bits
0 and 1 being 10b and 11b are invalid, and the display of these
registers is useless.

For example:

cat /sys/kernel/debug/regmap/spi0.0/registers
04: 10 -> Port 0, register offset 1
05: 10 -> Port 1, register offset 1
06: 00 -> Port 2, register offset 1 -> invalid
07: 00 -> port 3, register offset 1 -> invalid
...

Add a debug module parameter to call a custom dump function for each
port registers after the probe phase to help debug.

Signed-off-by: Hugo Villeneuve <hvilleneuve@dimonoff.com>
---
 drivers/tty/serial/sc16is7xx.c | 30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)

diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c
index 03d00b144304..693b6cc371f8 100644
--- a/drivers/tty/serial/sc16is7xx.c
+++ b/drivers/tty/serial/sc16is7xx.c
@@ -347,6 +347,10 @@ struct sc16is7xx_port {
 	struct sc16is7xx_one		p[];
 };
 
+static bool debug;
+module_param(debug, bool, 0644);
+MODULE_PARM_DESC(debug, "enable/disable debug messages");
+
 static unsigned long sc16is7xx_lines;
 
 static struct uart_driver sc16is7xx_uart = {
@@ -387,6 +391,28 @@ static void sc16is7xx_port_write(struct uart_port *port, u8 reg, u8 val)
 	regmap_write(s->regmap, (reg << SC16IS7XX_REG_SHIFT) | line, val);
 }
 
+static int sc16is7xx_port_dump(struct uart_port *port)
+{
+	int i;
+	unsigned char *buf;
+	char name[64];
+	const int regs_count_per_port = 16;
+
+	buf = devm_kzalloc(port->dev, regs_count_per_port, GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	for (i = 0; i < regs_count_per_port; i++)
+		buf[i] = sc16is7xx_port_read(port, i);
+
+	snprintf(name, sizeof(name), "sc16is7xx %s%i: dump ",
+		 sc16is7xx_uart.dev_name, port->line);
+	print_hex_dump(KERN_ERR, name, DUMP_PREFIX_OFFSET, 16, 1,
+		       &((u8 *)buf)[0], regs_count_per_port, 1);
+
+	return 0;
+}
+
 static void sc16is7xx_fifo_read(struct uart_port *port, unsigned int rxlen)
 {
 	struct sc16is7xx_port *s = dev_get_drvdata(port->dev);
@@ -1614,6 +1640,10 @@ static int sc16is7xx_probe(struct device *dev,
 	}
 #endif
 
+	if (debug)
+		for (i = 0; i < devtype->nr_uart; ++i)
+			sc16is7xx_port_dump(&s->p[i].port);
+
 	/*
 	 * Setup interrupt. We first try to acquire the IRQ line as level IRQ.
 	 * If that succeeds, we can allow sharing the interrupt as well.
-- 
2.30.2

On Thu, May 25, 2023 at 12:03:25AM -0400, Hugo Villeneuve wrote:
> From: Hugo Villeneuve <hvilleneuve@dimonoff.com>
> 
> With this driver, it is very hard to debug the registers using
> the /sys/kernel/debug/regmap interface.
> 
> The main reason is that bits 0 and 1 of the register address
> correspond to the channels bits, so the register address itself starts
> at bit 2, so we must 'mentally' shift each register address by 2 bits
> to get its offset.
> 
> Also, only channels 0 and 1 are supported, so combinations of bits
> 0 and 1 being 10b and 11b are invalid, and the display of these
> registers is useless.
> 
> For example:
> 
> cat /sys/kernel/debug/regmap/spi0.0/registers
> 04: 10 -> Port 0, register offset 1
> 05: 10 -> Port 1, register offset 1
> 06: 00 -> Port 2, register offset 1 -> invalid
> 07: 00 -> port 3, register offset 1 -> invalid
> ...
> 
> Add a debug module parameter to call a custom dump function for each
> port registers after the probe phase to help debug.
> 
> Signed-off-by: Hugo Villeneuve <hvilleneuve@dimonoff.com>
> ---
>  drivers/tty/serial/sc16is7xx.c | 30 ++++++++++++++++++++++++++++++
>  1 file changed, 30 insertions(+)
> 
> diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c
> index 03d00b144304..693b6cc371f8 100644
> --- a/drivers/tty/serial/sc16is7xx.c
> +++ b/drivers/tty/serial/sc16is7xx.c
> @@ -347,6 +347,10 @@ struct sc16is7xx_port {
>  	struct sc16is7xx_one		p[];
>  };
>  
> +static bool debug;
> +module_param(debug, bool, 0644);
> +MODULE_PARM_DESC(debug, "enable/disable debug messages");

Sorry, but no, use the normal dynamic debugging logic that the whole
rest of the kernel uses.  Do not add random per-driver module parameters
like this, that would be a regression from the existing infrastructure
that we have in place already.

thanks,

greg k-h

Thu, May 25, 2023 at 12:03:25AM -0400, Hugo Villeneuve kirjoitti:
> From: Hugo Villeneuve <hvilleneuve@dimonoff.com>
> 
> With this driver, it is very hard to debug the registers using
> the /sys/kernel/debug/regmap interface.
> 
> The main reason is that bits 0 and 1 of the register address
> correspond to the channels bits, so the register address itself starts
> at bit 2, so we must 'mentally' shift each register address by 2 bits
> to get its offset.
> 
> Also, only channels 0 and 1 are supported, so combinations of bits
> 0 and 1 being 10b and 11b are invalid, and the display of these
> registers is useless.
> 
> For example:
> 
> cat /sys/kernel/debug/regmap/spi0.0/registers
> 04: 10 -> Port 0, register offset 1
> 05: 10 -> Port 1, register offset 1
> 06: 00 -> Port 2, register offset 1 -> invalid
> 07: 00 -> port 3, register offset 1 -> invalid
> ...
> 
> Add a debug module parameter to call a custom dump function for each
> port registers after the probe phase to help debug.

Not sure about this. Can we rather create an abstract mapping on regmap?
(Something like gpio-pca953x.c has)

-- 
With Best Regards,
Andy Shevchenko

On Thu, 25 May 2023 14:26:43 +0300
andy.shevchenko@gmail.com wrote:

> Thu, May 25, 2023 at 12:03:25AM -0400, Hugo Villeneuve kirjoitti:
> > From: Hugo Villeneuve <hvilleneuve@dimonoff.com>
> > 
> > With this driver, it is very hard to debug the registers using
> > the /sys/kernel/debug/regmap interface.
> > 
> > The main reason is that bits 0 and 1 of the register address
> > correspond to the channels bits, so the register address itself starts
> > at bit 2, so we must 'mentally' shift each register address by 2 bits
> > to get its offset.
> > 
> > Also, only channels 0 and 1 are supported, so combinations of bits
> > 0 and 1 being 10b and 11b are invalid, and the display of these
> > registers is useless.
> > 
> > For example:
> > 
> > cat /sys/kernel/debug/regmap/spi0.0/registers
> > 04: 10 -> Port 0, register offset 1
> > 05: 10 -> Port 1, register offset 1
> > 06: 00 -> Port 2, register offset 1 -> invalid
> > 07: 00 -> port 3, register offset 1 -> invalid
> > ...
> > 
> > Add a debug module parameter to call a custom dump function for each
> > port registers after the probe phase to help debug.
> 
> Not sure about this. Can we rather create an abstract mapping on regmap?
> (Something like gpio-pca953x.c has)

Hi,
maybe we can, but more like they do in the driver max310x.c (single, dual and quad UART versions).

I will look into it, but it will probably be a patch that affects a lot of the code, and that I would like to submit separately after this serie, and so I will probably simply drop this current patch (11/11) since it will not be needed anymore.

Hugo.

Thu, May 25, 2023 at 03:49:46PM -0400, Hugo Villeneuve kirjoitti:
> On Thu, 25 May 2023 14:26:43 +0300
> andy.shevchenko@gmail.com wrote:
> > Thu, May 25, 2023 at 12:03:25AM -0400, Hugo Villeneuve kirjoitti:

...

> > Not sure about this. Can we rather create an abstract mapping on regmap?
> > (Something like gpio-pca953x.c has)
> 
> maybe we can, but more like they do in the driver max310x.c (single, dual and
> quad UART versions).
> 
> I will look into it, but it will probably be a patch that affects a lot of
> the code, and that I would like to submit separately after this serie, and so
> I will probably simply drop this current patch (11/11) since it will not be
> needed anymore.

Whatever, I commented on this solely because Greg KH is usually against new
module parameters. If you sell your point to him, fine to me.

-- 
With Best Regards,
Andy Shevchenko