serial: add NPCM UART driver
Tomer Maimon
maintainerv2 cover serial: add NPCM UART driver
Changes since version 1:
- rewritten the driver, the new driver based on 8250 UART driver with modifications to support NPCM UART functionality
This patch set adds UART support for the Nuvoton NPCM Baseboard Management Controller (BMC).
The NPCM UART based on 8250 UART driver
The NPCM7xx BMC contains four UART blocks and accessory logic.
To enable the driver please add CONFIG_SERIAL_8250_NPCM configuration (except to 8250 serial configuration).
Tomer Maimon (2):
dt-binding: serial: document NPCM UART DT bindings
serial: npcm: add NPCM UART driver
.../bindings/serial/nuvoton,npcm-uart.txt | 34 +++ drivers/tty/serial/8250/8250_npcm.c | 280 +++++++++++++++++++++ drivers/tty/serial/8250/Kconfig | 10 + drivers/tty/serial/8250/Makefile | 1 + 4 files changed, 325 insertions(+) create mode 100644 Documentation/devicetree/bindings/serial/nuvoton,npcm-uart.txt create mode 100644 drivers/tty/serial/8250/8250_npcm.c
Signature · 1 line
2.14.1
Tomer Maimon
maintainerPATCH v2 1/2 dt-binding: serial: document NPCM UART DT bindings
Added device tree binding documentation for Nuvoton NPCM UART.
--- .../bindings/serial/nuvoton,npcm-uart.txt | 34 ++++++++++++++++++++++ 1 file changed, 34 insertions(+) create mode 100644 Documentation/devicetree/bindings/serial/nuvoton,npcm-uart.txt
Documentation/devicetree/bindings/serial/nuvoton,npcm-uart.txt new file +34 -0
@@ -0,0 +1,34 @@ +Nuvoton NPCM Universal Asynchronous Receiver/Transmitter (UART) + +Required properties: +- compatible : "nuvoton,npcm750-uart" for Poleg NPCM750. +- reg : Offset and length of the register set for the device. +- interrupts : Contain the UART interrupt with flags for falling edge. + +Required clocking property, have to be one of: +- clocks : phandle of UART reference clock. +- clock-frequency : The frequency in Hz of the clock that drives the NPCM + UART (usually 24000000). + +Optional properties: +- reg-shift : quantity to shift the register offsets by (default 2). +- fifo-size : the fifo size of the UART (default 1). + +Note: Each uart controller should have an alias correctly numbered +in "aliases" node. + +Example: + +aliases { + serial0 = &serial0; +}; + +serial0: serial0@f0001000 { + compatible = "nuvoton,npcm750-uart"; + reg = <0xf0001000 0x1000>; + clocks = <&clk NPCM7XX_CLK_UART_CORE>; + interrupts = <GIC_SPI 2 IRQ_TYPE_LEVEL_HIGH>; + reg-shift = <2>; + fifo-size = <14>; + status = "disabled"; +};
Signature · 1 line
2.14.1
Tomer Maimon
maintainerPATCH v2 2/2 serial: npcm: add NPCM UART driver
In reply to Tomer MaimonAdd Nuvoton BMC NPCM UART driver.
The NPCM7xx BMC contains four UART blocks and accessory logic.
NPCM UART based on 8250 driver.
--- drivers/tty/serial/8250/8250_npcm.c | 280 ++++++++++++++++++++++++++++++++++++ drivers/tty/serial/8250/Kconfig | 10 ++ drivers/tty/serial/8250/Makefile | 1 + 3 files changed, 291 insertions(+) create mode 100644 drivers/tty/serial/8250/8250_npcm.c
drivers/tty/serial/8250/8250_npcm.c new file +280 -0
@@ -0,0 +1,280 @@ +/* +* Copyright (c) 2018 Nuvoton Technology corporation. +* +* Released under the GPLv2 only. +* SPDX-License-Identifier: GPL-2.0 +*/ + +#include <linux/clk.h> +#include <linux/io.h> +#include <linux/module.h> +#include <linux/of_irq.h> +#include <linux/of_platform.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/serial_8250.h> +#include <linux/serial_reg.h> + +#include "8250.h" + +#define MSB (parm16) ((u8)((u16)(parm16) >> 8)) +#define LSB (parm16) ((u8)(parm16)) + +#define UART_TOR 7 +#define NPCM_SERIAL_RX_TIMEOUT 0x20 +#define NPCM_SERIAL_RX_TIMEOUT_EN 0x80 + +/* Common baudrate definitions */ +typedef enum { +NPCM_UART_BAUDRATE_110 =110 , +NPCM_UART_BAUDRATE_300 =300 , +NPCM_UART_BAUDRATE_600 =600 , +NPCM_UART_BAUDRATE_1200 =1200 , +NPCM_UART_BAUDRATE_2400 =2400 , +NPCM_UART_BAUDRATE_4800 =4800 , +NPCM_UART_BAUDRATE_9600 =9600 , +NPCM_UART_BAUDRATE_14400 =14400 , +NPCM_UART_BAUDRATE_19200 =19200 , +NPCM_UART_BAUDRATE_38400 =38400 , +NPCM_UART_BAUDRATE_57600 =57600 , +NPCM_UART_BAUDRATE_115200 =115200 , +NPCM_UART_BAUDRATE_230400 =230400 , +NPCM_UART_BAUDRATE_380400 =380400 , +NPCM_UART_BAUDRATE_460800 =460800 , +}NPCM_UART_BAUDRATE_T ; + +struct npcm_data { +int line; +struct clk *uart_clk; +}; + +static int npcm_uart_set_baud_rate (struct uart_port *port, +NPCM_UART_BAUDRATE_T baudrate) +{ +struct uart_8250_port *up =up_to_u8250p (port); +int divisor =0 ; +int ret =0 ; +unsigned long flags; + + divisor = ((int )port->uartclk / ((int )baudrate *16 )) -2 ; + +/* since divisor is rounded down check +* if it is better when rounded up +*/ +if (((int )port->uartclk / (16 * (divisor +2 )) - (int )baudrate) > + ((int )baudrate - + (int )port->uartclk / (16 * ((divisor +1 ) +2 )))) + divisor++; + +if (divisor <0 ) { + divisor =0 ; + ret =EINVAL ; + } + +spin_lock_irqsave (&port->lock, flags); + +/* Set baud rate to baudRate bps */ +serial_port_out (port,UART_LCR , up->lcr |UART_LCR_DLAB ); +serial_port_out (port,UART_DLL ,LSB (divisor)); +serial_port_out (port,UART_DLM ,MSB (divisor)); +serial_port_out (port,UART_LCR , up->lcr); + +spin_unlock_irqrestore (&port->lock, flags); + +return ret; +} + +static void npcm_set_termios (struct uart_port *port,struct ktermios *termios, +struct ktermios *old) +{ +unsigned int baud =0 ; + +serial8250_do_set_termios (port, termios, old); + +switch (termios->c_cflag &CBAUD ) { +case B110 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_110 ); + baud =110 ; +break ; +case B300 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_300 ); + baud =300 ; +break ; +case B600 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_600 ); + baud =600 ; +break ; +case B1200 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_1200 ); + baud =1200 ; +break ; +case B2400 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_2400 ); + baud =2400 ; +break ; +case B4800 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_4800 ); + baud =4800 ; +break ; +case B9600 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_9600 ); + baud =9600 ; +break ; +case B19200 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_19200 ); + baud =19200 ; +break ; +case B38400 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_38400 ); + baud =38400 ; +break ; +case B57600 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_57600 ); + baud =57600 ; +break ; +default : +case B115200 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_115200 ); + baud =115200 ; +break ; +case B230400 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_230400 ); + baud =230400 ; +break ; +case B460800 : +npcm_uart_set_baud_rate (port,NPCM_UART_BAUDRATE_460800 ); + baud =460800 ; +break ; + } + +pr_info ("NPCM Serial: Baudrate %d\n" , baud); + +if (tty_termios_baud_rate (termios)) +tty_termios_encode_baud_rate (termios, baud, baud); +} + +static int npcm_serial_startup (struct uart_port *port) +{ +int rc; + + rc =serial8250_do_startup (port); +if (rc) +return rc; + +serial_port_out (port,UART_TOR ,NPCM_SERIAL_RX_TIMEOUT | +NPCM_SERIAL_RX_TIMEOUT_EN ); + +return 0 ; +} + +static int npcm_probe_of (struct platform_device *pdev,struct uart_port *p, +struct npcm_data *data) +{ +u32 clk_rate, prop; +struct device_node *np; +u32 ret; + + np = pdev->dev.of_node; + + data->uart_clk =devm_clk_get (&pdev->dev,NULL ); +if (IS_ERR (data->uart_clk)) { + ret =of_property_read_u32 (np,"clock-frequency" , &clk_rate); +if (ret) +return ret; + }else { +clk_prepare_enable (data->uart_clk); + clk_rate =clk_get_rate (data->uart_clk); + } + + p->uartclk = clk_rate; + +/* Check for registers offset within the devices address range */ +if (of_property_read_u32 (np,"reg-shift" , &prop) ==0 ) + p->regshift = prop; +else + p->regshift =2 ; + +/* Check for fifo size */ +if (of_property_read_u32 (np,"fifo-size" , &prop) ==0 ) + p->fifosize = prop; +else + p->fifosize =1 ; + +return 0 ; +} + +static int npcm_probe (struct platform_device *pdev) +{ +struct uart_8250_port uart = {}; +struct device_node *np = pdev->dev.of_node; +struct resource *regs =platform_get_resource (pdev,IORESOURCE_MEM ,0 ); +struct npcm_data *data; +int err; + +if (!regs) { +dev_err (&pdev->dev,"no registers defined\n" ); +return -EINVAL ; + } + + uart.port.membase =devm_ioremap (&pdev->dev, regs->start, +resource_size (regs)); +if (!uart.port.membase) +return -ENOMEM ; + + data =devm_kzalloc (&pdev->dev,sizeof (*data),GFP_KERNEL ); +if (!data) +return -ENOMEM ; + + err =npcm_probe_of (pdev, &uart.port, data); +if (err) +return err; + + uart.port.mapbase = regs->start; + uart.port.mapsize =resource_size (regs); + uart.port.irq =irq_of_parse_and_map (np,0 ); + uart.port.type =PORT_16550 ; + uart.port.flags =UPF_BOOT_AUTOCONF |UPF_FIXED_TYPE ; + uart.port.dev = &pdev->dev; + uart.port.iotype =UPIO_MEM ; + uart.port.private_data = data; + uart.port.set_termios = npcm_set_termios; + uart.port.startup = npcm_serial_startup; + +platform_set_drvdata (pdev, data); + + data->line =serial8250_register_8250_port (&uart); + +if (data->line <0 ) +return data->line; + +return 0 ; +} + +static int npcm_remove (struct platform_device *pdev) +{ +struct npcm_data *data =platform_get_drvdata (pdev); + +serial8250_unregister_port (data->line); + +return 0 ; +} + +static const struct of_device_id npcm_of_match[] = { + { .compatible ="nuvoton,npcm750-uart" }, + {/* Sentinel */ } +}; +MODULE_DEVICE_TABLE (of, npcm_of_match); + +static struct platform_driver npcm_platform_driver = { + .driver = { + .name ="npcm750-uart" , + .of_match_table = npcm_of_match, + }, + .probe = npcm_probe, + .remove = npcm_remove, +}; +module_platform_driver (npcm_platform_driver); + +MODULE_AUTHOR ("Tomer Maimon" ); +MODULE_LICENSE ("GPL" ); +MODULE_DESCRIPTION ("Driver for Nuvoton device based on 16550 serial driver" );
drivers/tty/serial/8250/Kconfig +10 -0
@@ -468,6 +468,16 @@ config SERIAL_8250_PXA applicable to both devicetree and legacy boards, and early console is part of its support. +config SERIAL_8250_NPCM + tristate "NPCM tty serial support" + depends on SERIAL_8250 + help + Say Y here if you have a NPCM tty serial support. + If unsure, say N. + + This driver can also be built as a module. The module will be called + 8250_nuvoton. If you want to do that, say M here. + config SERIAL_OF_PLATFORM tristate "Devicetree based probing for 8250 ports" depends on SERIAL_8250 && OF
drivers/tty/serial/8250/Makefile +1 -0
@@ -34,6 +34,7 @@ obj-$(CONFIG_SERIAL_8250_LPSS) += 8250_lpss.o obj-$(CONFIG_SERIAL_8250_MID) += 8250_mid.o obj-$(CONFIG_SERIAL_8250_MOXA) += 8250_moxa.o obj-$(CONFIG_SERIAL_8250_PXA) += 8250_pxa.o +obj-$(CONFIG_SERIAL_8250_NPCM) += 8250_npcm.o obj-$(CONFIG_SERIAL_OF_PLATFORM) += 8250_of.o CFLAGS_8250_ingenic.o += -I$(srctree)/scripts/dtc/libfdt
Signature · 1 line
2.14.1
Re: [PATCH linux dev-4.13 v2 2/2] serial: npcm: add NPCM UART driver
Quoted 5 lines · Tomer Maimon
Add Nuvoton BMC NPCM UART driver.
The NPCM7xx BMC contains four UART blocks and accessory logic.
NPCM UART based on 8250 driver.
Nice! Much less code to review and maintain than before :)
I did some more digging, to see if we could add even less code. I'll sent a patch to the list, but I need to clarify the divisor calculation first.
drivers/tty/serial/8250/8250_npcm.c line 59
Quoted 9 lines
drivers/tty/serial/8250/8250_npcm.c new file +9
-0+static int npcm_uart_set_baud_rate (struct uart_port *port, +NPCM_UART_BAUDRATE_T baudrate) +{ +struct uart_8250_port *up =up_to_u8250p (port); +int divisor =0 ; +int ret =0 ; +unsigned long flags; + + divisor = ((int )port->uartclk / ((int )baudrate *16 )) -2 ;
The 8250 core uses this calculation:
divisor = port->uartclk / (baudrate * 16).
Checking the datasheet, this is the calculation that is documented for Poleg:
divisor = port->uartclk / (baudrate * 16 + 2).
If I change the calculation to this, the UART doesn't work (due to the precision of the integer division, there's no change from the 8250 core code). If I subtract 2 from the number, as you do in your patch, it works!
Can you explain what is happening here?
drivers/tty/serial/8250/8250_npcm.c line 67
Quoted 8 lines
drivers/tty/serial/8250/8250_npcm.c new file +8
-0+ +/* since divisor is rounded down check +* if it is better when rounded up +*/ +if (((int )port->uartclk / (16 * (divisor +2 )) - (int )baudrate) > + ((int )baudrate - + (int )port->uartclk / (16 * ((divisor +1 ) +2 )))) + divisor++;
This bit confused me when I tried to understand what it was checking. I couldn't find a mention of it in the datasheet either.
Can you explain to me what it is trying to do? Do we need it?
Cheers,
Joel
Tomer Maimon
maintainerRe: [PATCH linux dev-4.13 v2 2/2] serial: npcm: add NPCM UART driver
Quoted 5 lines · Joel Stanley
Quoted 4 lines · Tomer Maimon
Add Nuvoton BMC NPCM UART driver.
The NPCM7xx BMC contains four UART blocks and accessory logic.
Quoted 8 lines
Quoted 1 line
NPCM UART based on 8250 driver.
Nice! Much less code to review and maintain than before :)
I did some more digging, to see if we could add even less code. I'll sent a patch to the list, but I need to clarify the divisor calculation first.
Thanks!
Quoted 34 lines
drivers/tty/serial/8250/8250_npcm.c line 59
Quoted 9 lines
drivers/tty/serial/8250/8250_npcm.c new file +9
-0+static int npcm_uart_set_baud_rate (struct uart_port *port, +NPCM_UART_BAUDRATE_T baudrate) +{ +struct uart_8250_port *up =up_to_u8250p (port); +int divisor =0 ; +int ret =0 ; +unsigned long flags; + + divisor = ((int )port->uartclk / ((int )baudrate *16 )) -2 ;The 8250 core uses this calculation:
divisor =
port->uartclk/ (baudrate * 16).Checking the datasheet, this is the calculation that is documented for Poleg:
divisor =
port->uartclk/ (baudrate * 16 + 2).If I change the calculation to this, the UART doesn't work (due to the precision of the integer division, there's no change from the 8250 core code). If I subtract 2 from the number, as you do in your patch, it works!
Can you explain what is happening here?
drivers/tty/serial/8250/8250_npcm.c line 67
Quoted 8 lines
drivers/tty/serial/8250/8250_npcm.c new file +8
-0+ +/* since divisor is rounded down check +* if it is better when rounded up +*/ +if (((int )port->uartclk / (16 * (divisor +2 )) - (int )baudrate) > + ((int )baudrate - + (int )port->uartclk / (16 * ((divisor +1 ) +2 )))) + divisor++;
Quoted 4 lines
This bit confused me when I tried to understand what it was checking. I couldn't find a mention of it in the datasheet either.
Can you explain to me what it is trying to do? Do we need it?
hope i will explain it right...
we are calculating the nearest round up of the divisor according to the distance from the real divisor (with point) to the value we get from the round up and round down. (and it different from using DIV_ROUND_CLOSEST)
for example:
If the baud rate is 230,400
the divisor is 4.510 (it means round up)
if we doing the calculation above:
(24Mhz/(16*(4+2))) - 230,400 = 19,600
230,400 - (24Mhz/(16*(4+3))) = 16,114
also in here we see round - up, but the distance from the second (16,114) it better and the 4.510 not represent it correctly because 2 result we get are not that the same.
so maybe in clock rate of 24Mhz it will be fine but if we will modify the clock rate the calculation could be wrong with a use of DIV_ROUND_CLOSEST.
so we think it is better to use the calculation above and not round up.
Is it problematic to use it in the modify driver you sent?
Quoted 4 lines
Cheers,
Joel
Cheers,
Tomer
Re: [PATCH linux dev-4.13 v2 2/2] serial: npcm: add NPCM UART driver
Quoted 8 lines · Tomer Maimon
Quoted 7 lines · Joel Stanley
Quoted 5 lines · Tomer Maimon
Add Nuvoton BMC NPCM UART driver.
The NPCM7xx BMC contains four UART blocks and accessory logic.
NPCM UART based on 8250 driver.
Quoted 31 lines
Quoted 6 lines
Nice! Much less code to review and maintain than before :)
I did some more digging, to see if we could add even less code. I'll sent a patch to the list, but I need to clarify the divisor calculation first.
Thanks!
Quoted 24 lines
drivers/tty/serial/8250/8250_npcm.c line 59
Quoted 9 lines
drivers/tty/serial/8250/8250_npcm.c new file +9
-0+static int npcm_uart_set_baud_rate (struct uart_port *port, +NPCM_UART_BAUDRATE_T baudrate) +{ +struct uart_8250_port *up =up_to_u8250p (port); +int divisor =0 ; +int ret =0 ; +unsigned long flags; + + divisor = ((int )port->uartclk / ((int )baudrate *16 )) -2 ;The 8250 core uses this calculation:
divisor =
port->uartclk/ (baudrate * 16).Checking the datasheet, this is the calculation that is documented for Poleg:
divisor =
port->uartclk/ (baudrate * 16 + 2).If I change the calculation to this, the UART doesn't work (due to the precision of the integer division, there's no change from the 8250 core code). If I subtract 2 from the number, as you do in your patch, it works!
Can you explain what is happening here?
Hi Tomer, you missed this question. Can you explain to me what is happening here?
Quoted 35 lines
Quoted 10 lines
drivers/tty/serial/8250/8250_npcm.c line 67
Quoted 9 lines
drivers/tty/serial/8250/8250_npcm.c new file +8
-0+ +/* since divisor is rounded down check +* if it is better when rounded up +*/ +if (((int )port->uartclk / (16 * (divisor +2 )) - (int )baudrate) > + ((int )baudrate - + (int )port->uartclk / (16 * ((divisor +1 ) +2 )))) + divisor++;Quoted 4 lines
This bit confused me when I tried to understand what it was checking. I couldn't find a mention of it in the datasheet either.
Can you explain to me what it is trying to do? Do we need it?
hope i will explain it right...
we are calculating the nearest round up of the divisor according to the distance from the real divisor (with point) to the value we get from the round up and round down. (and it different from using
DIV_ROUND_CLOSEST)
for example:If the baud rate is 230,400
the divisor is 4.510 (it means round up)
if we doing the calculation above:
(24Mhz/(16*(4+2))) - 230,400 = 19,600
230,400 - (24Mhz/(16*(4+3))) = 16,114also in here we see round - up, but the distance from the second (16,114) it better and the 4.510 not represent it correctly because 2 result we get are not that the same.
Quoted 7 lines
so maybe in clock rate of 24Mhz it will be fine but if we will modify the clock rate the calculation could be wrong with a use of
DIV_ROUND_CLOSEST.so we think it is better to use the calculation above and not round up.
Is it problematic to use it in the modify driver you sent?
Thanks for the explanation. As I mention above, I need to understand the first calculation in order to understand this part.
It would be easy enough to add this calculation if it was strictly necessary. However, to know if it is required, I first need to understand what is happening.
Cheers,
Joel
A message that is not in the archive (sent off-list or not yet received): CAP6Zq1gMb5VjLpFBCWoiYkRjSKdpqa9mZMiU5qUadBtaUvPicw@mail.gmail.com
Tomer Maimon
maintainerRe: [PATCH linux dev-4.13 v2 2/2] serial: npcm: add NPCM UART driver
Forget to add CC. Jeremy Kerr <[email protected]>,
OpenBMC Maillist <[email protected]>
Quoted 117 lines · Tomer Maimon
Quoted 39 lines · Joel Stanley
Quoted 38 lines · Tomer Maimon
Quoted 5 lines · Joel Stanley
Quoted 4 lines · Tomer Maimon
Add Nuvoton BMC NPCM UART driver.
The NPCM7xx BMC contains four UART blocks and accessory logic.
Quoted 8 lines
Quoted 1 line
NPCM UART based on 8250 driver.
Nice! Much less code to review and maintain than before :)
I did some more digging, to see if we could add even less code. I'll sent a patch to the list, but I need to clarify the divisor calculation first.
Thanks!
Quoted 23 lines
drivers/tty/serial/8250/8250_npcm.c line 59
Quoted 9 lines
drivers/tty/serial/8250/8250_npcm.c new file +9
-0+static int npcm_uart_set_baud_rate (struct uart_port *port, +NPCM_UART_BAUDRATE_T baudrate) +{ +struct uart_8250_port *up =up_to_u8250p (port); +int divisor =0 ; +int ret =0 ; +unsigned long flags; + + divisor = ((int )port->uartclk / ((int )baudrate *16 )) -2 ;The 8250 core uses this calculation:
divisor =
port->uartclk/ (baudrate * 16).Checking the datasheet, this is the calculation that is documented for Poleg:
divisor =
port->uartclk/ (baudrate * 16 + 2).If I change the calculation to this, the UART doesn't work (due to the precision of the integer division, there's no change from the 8250 core code). If I subtract 2 from the number, as you do in your patch, it works!
Quoted 5 lines
Quoted 1 line
Quoted 1 line
Can you explain what is happening here?
Hi Tomer, you missed this question. Can you explain to me what is happening here?
Opps sorry!
I think you have little mistake...
From the data sheet calculation:
baudrate =
port->uartclk/ (16* (divisor + 2))So if we isolate the divisor
divisor = (
port->uartclk/ (16* baudrate)) - 2this is why if you subtract 2 from the divisor you get the right divisor.
Quoted 46 lines
Quoted 35 lines
Quoted 10 lines
drivers/tty/serial/8250/8250_npcm.c line 67
Quoted 9 lines
drivers/tty/serial/8250/8250_npcm.c new file +8
-0+ +/* since divisor is rounded down check +* if it is better when rounded up +*/ +if (((int )port->uartclk / (16 * (divisor +2 )) - (int )baudrate) > + ((int )baudrate - + (int )port->uartclk / (16 * ((divisor +1 ) +2 )))) + divisor++;Quoted 4 lines
This bit confused me when I tried to understand what it was checking. I couldn't find a mention of it in the datasheet either.
Can you explain to me what it is trying to do? Do we need it?
hope i will explain it right...
we are calculating the nearest round up of the divisor according to the distance from the real divisor (with point) to the value we get from the round up and round down. (and it different from using
DIV_ROUND_CLOSEST)
for example:If the baud rate is 230,400
the divisor is 4.510 (it means round up)
if we doing the calculation above:
(24Mhz/(16*(4+2))) - 230,400 = 19,600
230,400 - (24Mhz/(16*(4+3))) = 16,114also in here we see round - up, but the distance from the second (16,114) it better and the 4.510 not represent it correctly because 2 result we get are not that the same.
Quoted 7 lines
so maybe in clock rate of 24Mhz it will be fine but if we will modify the clock rate the calculation could be wrong with a use of
DIV_ROUND_CLOSEST.so we think it is better to use the calculation above and not round up.
Is it problematic to use it in the modify driver you sent?
Thanks for the explanation. As I mention above, I need to understand the first calculation in order to understand this part.
Quoted 7 lines
It would be easy enough to add this calculation if it was strictly necessary. However, to know if it is required, I first need to understand what is happening.
Cheers,
Joel
Cheers,
Tomer