From: Hugo Villeneuve hvilleneuve@dimonoff.com
When using a high speed clock with a low baud rate, the 4x prescaler is automatically selected if required. In that case, sc16is7xx_set_baud() properly configures the chip registers, but returns an incorrect baud rate by not taking into account the prescaler value. This incorrect baud rate is then fed to uart_update_timeout().
For example, with an input clock of 80MHz, and a selected baud rate of 50, sc16is7xx_set_baud() will return 200 instead of 50.
Fix this by first changing the prescaler variable to hold the selected prescaler value instead of the MCR bitfield. Then properly take into account the selected prescaler value in the return value computation.
Also add better documentation about the divisor value computation.
Fixes: dfeae619d781 ("serial: sc16is7xx") Cc: stable@vger.kernel.org Signed-off-by: Hugo Villeneuve hvilleneuve@dimonoff.com --- drivers/tty/serial/sc16is7xx.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-)
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c index 03cf30e20b75..dcd6c5615401 100644 --- a/drivers/tty/serial/sc16is7xx.c +++ b/drivers/tty/serial/sc16is7xx.c @@ -555,16 +555,28 @@ static bool sc16is7xx_regmap_noinc(struct device *dev, unsigned int reg) return reg == SC16IS7XX_RHR_REG; }
+/* + * Configure programmable baud rate generator (divisor) according to the + * desired baud rate. + * + * From the datasheet, the divisor is computed according to: + * + * XTAL1 input frequency + * ----------------------- + * prescaler + * divisor = --------------------------- + * baud-rate x sampling-rate + */ static int sc16is7xx_set_baud(struct uart_port *port, int baud) { struct sc16is7xx_one *one = to_sc16is7xx_one(port, port); u8 lcr; - u8 prescaler = 0; + int prescaler = 1; unsigned long clk = port->uartclk, div = clk / 16 / baud;
if (div >= BIT(16)) { - prescaler = SC16IS7XX_MCR_CLKSEL_BIT; - div /= 4; + prescaler = 4; + div /= prescaler; }
/* Enable enhanced features */ @@ -574,9 +586,10 @@ static int sc16is7xx_set_baud(struct uart_port *port, int baud) SC16IS7XX_EFR_ENABLE_BIT); sc16is7xx_efr_unlock(port);
+ /* If bit MCR_CLKSEL is set, the divide by 4 prescaler is activated. */ sc16is7xx_port_update(port, SC16IS7XX_MCR_REG, SC16IS7XX_MCR_CLKSEL_BIT, - prescaler); + prescaler == 1 ? 0 : SC16IS7XX_MCR_CLKSEL_BIT);
/* Backup LCR and access special register set (DLL/DLH) */ lcr = sc16is7xx_port_read(port, SC16IS7XX_LCR_REG); @@ -592,7 +605,7 @@ static int sc16is7xx_set_baud(struct uart_port *port, int baud) /* Restore LCR and access to general register set */ sc16is7xx_port_write(port, SC16IS7XX_LCR_REG, lcr);
- return DIV_ROUND_CLOSEST(clk / 16, div); + return DIV_ROUND_CLOSEST((clk / prescaler) / 16, div); }
static void sc16is7xx_handle_rx(struct uart_port *port, unsigned int rxlen,
base-commit: 660a708098569a66a47d0abdad998e29e1259de6
On Fri, Apr 26, 2024 at 4:59 PM Hugo Villeneuve hugo@hugovil.com wrote:
From: Hugo Villeneuve hvilleneuve@dimonoff.com
When using a high speed clock with a low baud rate, the 4x prescaler is automatically selected if required. In that case, sc16is7xx_set_baud() properly configures the chip registers, but returns an incorrect baud rate by not taking into account the prescaler value. This incorrect baud rate is then fed to uart_update_timeout().
For example, with an input clock of 80MHz, and a selected baud rate of 50, sc16is7xx_set_baud() will return 200 instead of 50.
Fix this by first changing the prescaler variable to hold the selected prescaler value instead of the MCR bitfield. Then properly take into account the selected prescaler value in the return value computation.
Also add better documentation about the divisor value computation.
Fixes: dfeae619d781 ("serial: sc16is7xx") Cc: stable@vger.kernel.org Signed-off-by: Hugo Villeneuve hvilleneuve@dimonoff.com
drivers/tty/serial/sc16is7xx.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-)
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c index 03cf30e20b75..dcd6c5615401 100644 --- a/drivers/tty/serial/sc16is7xx.c +++ b/drivers/tty/serial/sc16is7xx.c @@ -555,16 +555,28 @@ static bool sc16is7xx_regmap_noinc(struct device *dev, unsigned int reg) return reg == SC16IS7XX_RHR_REG; }
+/*
- Configure programmable baud rate generator (divisor) according to the
- desired baud rate.
- From the datasheet, the divisor is computed according to:
XTAL1 input frequency
-----------------------
prescaler
- divisor = ---------------------------
baud-rate x sampling-rate- */
static int sc16is7xx_set_baud(struct uart_port *port, int baud) { struct sc16is7xx_one *one = to_sc16is7xx_one(port, port); u8 lcr;
u8 prescaler = 0;
int prescaler = 1; unsigned long clk = port->uartclk, div = clk / 16 / baud; if (div >= BIT(16)) {
prescaler = SC16IS7XX_MCR_CLKSEL_BIT;div /= 4;
prescaler = 4;div /= prescaler; } /* Enable enhanced features */@@ -574,9 +586,10 @@ static int sc16is7xx_set_baud(struct uart_port *port, int baud) SC16IS7XX_EFR_ENABLE_BIT); sc16is7xx_efr_unlock(port);
/* If bit MCR_CLKSEL is set, the divide by 4 prescaler is activated. */ sc16is7xx_port_update(port, SC16IS7XX_MCR_REG, SC16IS7XX_MCR_CLKSEL_BIT,
prescaler);
prescaler == 1 ? 0 : SC16IS7XX_MCR_CLKSEL_BIT); /* Backup LCR and access special register set (DLL/DLH) */ lcr = sc16is7xx_port_read(port, SC16IS7XX_LCR_REG);@@ -592,7 +605,7 @@ static int sc16is7xx_set_baud(struct uart_port *port, int baud) /* Restore LCR and access to general register set */ sc16is7xx_port_write(port, SC16IS7XX_LCR_REG, lcr);
return DIV_ROUND_CLOSEST(clk / 16, div);
return DIV_ROUND_CLOSEST((clk / prescaler) / 16, div);}
static void sc16is7xx_handle_rx(struct uart_port *port, unsigned int rxlen,
base-commit: 660a708098569a66a47d0abdad998e29e1259de6
2.39.2
For me, looks normal. Does not cause problems on my vendored kernel with my XR20M1172 patches. Do I need to integrate those inside my patch? Or how should I do?
On 26. 04. 24, 15:59, Hugo Villeneuve wrote:
From: Hugo Villeneuve hvilleneuve@dimonoff.com
When using a high speed clock with a low baud rate, the 4x prescaler is automatically selected if required. In that case, sc16is7xx_set_baud() properly configures the chip registers, but returns an incorrect baud rate by not taking into account the prescaler value. This incorrect baud rate is then fed to uart_update_timeout().
For example, with an input clock of 80MHz, and a selected baud rate of 50, sc16is7xx_set_baud() will return 200 instead of 50.
Fix this by first changing the prescaler variable to hold the selected prescaler value instead of the MCR bitfield. Then properly take into account the selected prescaler value in the return value computation.
Also add better documentation about the divisor value computation.
Fixes: dfeae619d781 ("serial: sc16is7xx") Cc: stable@vger.kernel.org Signed-off-by: Hugo Villeneuve hvilleneuve@dimonoff.com
drivers/tty/serial/sc16is7xx.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-)
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c index 03cf30e20b75..dcd6c5615401 100644 --- a/drivers/tty/serial/sc16is7xx.c +++ b/drivers/tty/serial/sc16is7xx.c @@ -555,16 +555,28 @@ static bool sc16is7xx_regmap_noinc(struct device *dev, unsigned int reg) return reg == SC16IS7XX_RHR_REG; } +/*
- Configure programmable baud rate generator (divisor) according to the
- desired baud rate.
- From the datasheet, the divisor is computed according to:
XTAL1 input frequency
-----------------------
prescaler
- divisor = ---------------------------
baud-rate x sampling-rate- */ static int sc16is7xx_set_baud(struct uart_port *port, int baud) { struct sc16is7xx_one *one = to_sc16is7xx_one(port, port); u8 lcr;
- u8 prescaler = 0;
- int prescaler = 1;
Ugh, why do you move to signed arithmetics?
regards,
From: Jiri Slaby
Sent: 29 April 2024 07:39
...
- u8 prescaler = 0;
- int prescaler = 1;
Ugh, why do you move to signed arithmetics?
Any arithmetic would always have been signed. u8 is promoted to 'signed int' before being used for pretty much anything.
'unsigned int prescaler' might have changed arithmetic to be unsigned.
OTOH you probably don't want a u8 - that might require the compiler mask an arithmetic result to 8 bits.
David
- Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
On 29. 04. 24, 11:14, David Laight wrote:
From: Jiri Slaby
Sent: 29 April 2024 07:39
...
- u8 prescaler = 0;
- int prescaler = 1;
Ugh, why do you move to signed arithmetics?
Any arithmetic would always have been signed. u8 is promoted to 'signed int' before being used for pretty much anything.
Sorry, what?
C99 §6.3.8.1 states:
If both operands have the same type, then no further conversion is needed.
Otherwise, if both operands have signed integer types or both have unsigned integer types, the operand with the type of lesser integer conversion rank is converted to the type of the operand with greater rank.
=====
I.e. u8 is converted according to that to ulong in this case. So unsigned arithmetic happens.
'unsigned int prescaler' might have changed arithmetic to be unsigned.
The same as u8.
OTOH you probably don't want a u8 - that might require the compiler mask an arithmetic result to 8 bits.
Pardon? Not at all.
Am I missing something?
thanks,
On Mon, 29 Apr 2024 08:39:22 +0200 Jiri Slaby jirislaby@kernel.org wrote:
On 26. 04. 24, 15:59, Hugo Villeneuve wrote:
From: Hugo Villeneuve hvilleneuve@dimonoff.com
When using a high speed clock with a low baud rate, the 4x prescaler is automatically selected if required. In that case, sc16is7xx_set_baud() properly configures the chip registers, but returns an incorrect baud rate by not taking into account the prescaler value. This incorrect baud rate is then fed to uart_update_timeout().
For example, with an input clock of 80MHz, and a selected baud rate of 50, sc16is7xx_set_baud() will return 200 instead of 50.
Fix this by first changing the prescaler variable to hold the selected prescaler value instead of the MCR bitfield. Then properly take into account the selected prescaler value in the return value computation.
Also add better documentation about the divisor value computation.
Fixes: dfeae619d781 ("serial: sc16is7xx") Cc: stable@vger.kernel.org Signed-off-by: Hugo Villeneuve hvilleneuve@dimonoff.com
drivers/tty/serial/sc16is7xx.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-)
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c index 03cf30e20b75..dcd6c5615401 100644 --- a/drivers/tty/serial/sc16is7xx.c +++ b/drivers/tty/serial/sc16is7xx.c @@ -555,16 +555,28 @@ static bool sc16is7xx_regmap_noinc(struct device *dev, unsigned int reg) return reg == SC16IS7XX_RHR_REG; } +/*
- Configure programmable baud rate generator (divisor) according to the
- desired baud rate.
- From the datasheet, the divisor is computed according to:
XTAL1 input frequency
-----------------------
prescaler
- divisor = ---------------------------
baud-rate x sampling-rate- */ static int sc16is7xx_set_baud(struct uart_port *port, int baud) { struct sc16is7xx_one *one = to_sc16is7xx_one(port, port); u8 lcr;
- u8 prescaler = 0;
- int prescaler = 1;
Ugh, why do you move to signed arithmetics?
Hi Jiri, before this patch, the variable prescaler was used to store an 8 bit bitfield. Now the variable meaning is changed to be used as the prescaler value, which can be 1 or 4 in this case. Leaving it as u8 would still be ok, or making it "unsigned int" maybe?
Hugo.
On 29. 04. 24, 15:47, Hugo Villeneuve wrote:
On Mon, 29 Apr 2024 08:39:22 +0200 Jiri Slaby jirislaby@kernel.org wrote:
On 26. 04. 24, 15:59, Hugo Villeneuve wrote:
From: Hugo Villeneuve hvilleneuve@dimonoff.com
When using a high speed clock with a low baud rate, the 4x prescaler is automatically selected if required. In that case, sc16is7xx_set_baud() properly configures the chip registers, but returns an incorrect baud rate by not taking into account the prescaler value. This incorrect baud rate is then fed to uart_update_timeout().
For example, with an input clock of 80MHz, and a selected baud rate of 50, sc16is7xx_set_baud() will return 200 instead of 50.
Fix this by first changing the prescaler variable to hold the selected prescaler value instead of the MCR bitfield. Then properly take into account the selected prescaler value in the return value computation.
Also add better documentation about the divisor value computation.
Fixes: dfeae619d781 ("serial: sc16is7xx") Cc: stable@vger.kernel.org Signed-off-by: Hugo Villeneuve hvilleneuve@dimonoff.com
drivers/tty/serial/sc16is7xx.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-)
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c index 03cf30e20b75..dcd6c5615401 100644 --- a/drivers/tty/serial/sc16is7xx.c +++ b/drivers/tty/serial/sc16is7xx.c @@ -555,16 +555,28 @@ static bool sc16is7xx_regmap_noinc(struct device *dev, unsigned int reg) return reg == SC16IS7XX_RHR_REG; } +/*
- Configure programmable baud rate generator (divisor) according to the
- desired baud rate.
- From the datasheet, the divisor is computed according to:
XTAL1 input frequency
-----------------------
prescaler
- divisor = ---------------------------
baud-rate x sampling-rate- */ static int sc16is7xx_set_baud(struct uart_port *port, int baud) { struct sc16is7xx_one *one = to_sc16is7xx_one(port, port); u8 lcr;
- u8 prescaler = 0;
- int prescaler = 1;
Ugh, why do you move to signed arithmetics?
Hi Jiri, before this patch, the variable prescaler was used to store an 8 bit bitfield. Now the variable meaning is changed to be used as the prescaler value, which can be 1 or 4 in this case. Leaving it as u8 would still be ok, or making it "unsigned int" maybe?
Both :). What you prefer -- uint matches more IMO, given it's now a value and not a register...
thanks,
On Tue, 30 Apr 2024 07:22:54 +0200 Jiri Slaby jirislaby@kernel.org wrote:
On 29. 04. 24, 15:47, Hugo Villeneuve wrote:
On Mon, 29 Apr 2024 08:39:22 +0200 Jiri Slaby jirislaby@kernel.org wrote:
On 26. 04. 24, 15:59, Hugo Villeneuve wrote:
From: Hugo Villeneuve hvilleneuve@dimonoff.com
When using a high speed clock with a low baud rate, the 4x prescaler is automatically selected if required. In that case, sc16is7xx_set_baud() properly configures the chip registers, but returns an incorrect baud rate by not taking into account the prescaler value. This incorrect baud rate is then fed to uart_update_timeout().
For example, with an input clock of 80MHz, and a selected baud rate of 50, sc16is7xx_set_baud() will return 200 instead of 50.
Fix this by first changing the prescaler variable to hold the selected prescaler value instead of the MCR bitfield. Then properly take into account the selected prescaler value in the return value computation.
Also add better documentation about the divisor value computation.
Fixes: dfeae619d781 ("serial: sc16is7xx") Cc: stable@vger.kernel.org Signed-off-by: Hugo Villeneuve hvilleneuve@dimonoff.com
drivers/tty/serial/sc16is7xx.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-)
diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c index 03cf30e20b75..dcd6c5615401 100644 --- a/drivers/tty/serial/sc16is7xx.c +++ b/drivers/tty/serial/sc16is7xx.c @@ -555,16 +555,28 @@ static bool sc16is7xx_regmap_noinc(struct device *dev, unsigned int reg) return reg == SC16IS7XX_RHR_REG; } +/*
- Configure programmable baud rate generator (divisor) according to the
- desired baud rate.
- From the datasheet, the divisor is computed according to:
XTAL1 input frequency
-----------------------
prescaler
- divisor = ---------------------------
baud-rate x sampling-rate- */ static int sc16is7xx_set_baud(struct uart_port *port, int baud) { struct sc16is7xx_one *one = to_sc16is7xx_one(port, port); u8 lcr;
- u8 prescaler = 0;
- int prescaler = 1;
Ugh, why do you move to signed arithmetics?
Hi Jiri, before this patch, the variable prescaler was used to store an 8 bit bitfield. Now the variable meaning is changed to be used as the prescaler value, which can be 1 or 4 in this case. Leaving it as u8 would still be ok, or making it "unsigned int" maybe?
Both :). What you prefer -- uint matches more IMO, given it's now a value and not a register...
Hi Jiri, I will go with uint.
Thank you, Hugo.
thanks,
js suse labs
linux-stable-mirror@lists.linaro.org
-
David Laight -
Hugo Villeneuve -
Jiri Slaby -
Konstantin P.