2011/4/3 Franck JULLIEN <franck.jullien@gmail.com>

Hello Marc,

2011/4/3 Marc Kleine-Budde <mkl@pengutronix.de>

On 04/03/2011 12:25 PM, franck.jullien@gmail.com wrote:
> From: Franck JULLIEN <franck.jullien@gmail.com>

Please add your S-o-b to the driver. See Documentation/SubmittingPatches
in a Linux-kernel tree.

Ok I will. 

 

See comments inline.

>
> ---
>  drivers/serial/Kconfig         |    5 ++
>  drivers/serial/Makefile        |   25 +++++-----
>  drivers/serial/serial_altera.c |   97 ++++++++++++++++++++++++++++++++++++++++
>  3 files changed, 115 insertions(+), 12 deletions(-)
>  create mode 100644 drivers/serial/serial_altera.c
>
> diff --git a/drivers/serial/Kconfig b/drivers/serial/Kconfig
> index ffd877a..9e05f4b 100644
> --- a/drivers/serial/Kconfig
> +++ b/drivers/serial/Kconfig
> @@ -43,6 +43,11 @@ config DRIVER_SERIAL_BLACKFIN
>       default y
>       bool "Blackfin serial driver"
>
> +config DRIVER_SERIAL_ALTERA
> +     depends on NIOS2
> +     default y
> +     bool "Altera serial driver"
> +
>  config DRIVER_SERIAL_NS16550
>       default n
>       bool "NS16550 serial driver"
> diff --git a/drivers/serial/Makefile b/drivers/serial/Makefile
> index 9f0e12b..8067b74 100644
> --- a/drivers/serial/Makefile
> +++ b/drivers/serial/Makefile
> @@ -4,15 +4,16 @@
>  # serial_max3100.o
>  # serial_pl010.o
>  # serial_xuartlite.o
> -obj-$(CONFIG_DRIVER_SERIAL_ARM_DCC)          += arm_dcc.o
> -obj-$(CONFIG_SERIAL_AMBA_PL011) += amba-pl011.o
> -obj-$(CONFIG_DRIVER_SERIAL_IMX)                      += serial_imx.o
> -obj-$(CONFIG_DRIVER_SERIAL_STM378X)          += stm-serial.o
> -obj-$(CONFIG_DRIVER_SERIAL_ATMEL)            += atmel.o
> -obj-$(CONFIG_DRIVER_SERIAL_NETX)             += serial_netx.o
> -obj-$(CONFIG_DRIVER_SERIAL_LINUX_COMSOLE)    += linux_console.o
> -obj-$(CONFIG_DRIVER_SERIAL_MPC5XXX)          += serial_mpc5xxx.o
> -obj-$(CONFIG_DRIVER_SERIAL_BLACKFIN)         += serial_blackfin.o
> -obj-$(CONFIG_DRIVER_SERIAL_NS16550)          += serial_ns16550.o
> -obj-$(CONFIG_DRIVER_SERIAL_PL010)            += serial_pl010.o
> -obj-$(CONFIG_DRIVER_SERIAL_S3C24X0)          += serial_s3c24x0.o
> +obj-$(CONFIG_DRIVER_SERIAL_ARM_DCC)             += arm_dcc.o
> +obj-$(CONFIG_SERIAL_AMBA_PL011)                 += amba-pl011.o
> +obj-$(CONFIG_DRIVER_SERIAL_IMX)                 += serial_imx.o
> +obj-$(CONFIG_DRIVER_SERIAL_STM378X)             += stm-serial.o
> +obj-$(CONFIG_DRIVER_SERIAL_ATMEL)               += atmel.o
> +obj-$(CONFIG_DRIVER_SERIAL_NETX)                += serial_netx.o
> +obj-$(CONFIG_DRIVER_SERIAL_LINUX_COMSOLE)       += linux_console.o
> +obj-$(CONFIG_DRIVER_SERIAL_MPC5XXX)             += serial_mpc5xxx.o
> +obj-$(CONFIG_DRIVER_SERIAL_BLACKFIN)            += serial_blackfin.o
> +obj-$(CONFIG_DRIVER_SERIAL_NS16550)             += serial_ns16550.o
> +obj-$(CONFIG_DRIVER_SERIAL_PL010)               += serial_pl010.o
> +obj-$(CONFIG_DRIVER_SERIAL_S3C24X0)             += serial_s3c24x0.o
> +obj-$(CONFIG_DRIVER_SERIAL_ALTERA)              += serial_altera.o

Why do you reformat the whole file? Please don't do that.

Because there is a mix between tabs and spaces before the "+= ....". I think it's better to have spaces here in order

to maintain alignment no matter what the tab size is. Tell me if you're agree, if not, I'll just put my new line at the bottom..... :)

 

> diff --git a/drivers/serial/serial_altera.c b/drivers/serial/serial_altera.c
> new file mode 100644
> index 0000000..03e48d1
> --- /dev/null
> +++ b/drivers/serial/serial_altera.c
> @@ -0,0 +1,97 @@
> +/*
> + * (C) Copyright 2011, Franck JULLIEN, <elec4fun@gmail.com>
> + *
> + * See file CREDITS for list of people who contributed to this
> + * project.
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License as
> + * published by the Free Software Foundation; either version 2 of
> + * the License, or (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public License
> + * along with this program; if not, write to the Free Software
> + * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
> + * MA 02111-1307 USA
> + */
> +
> +#include <common.h>
> +#include <driver.h>
> +#include <init.h>
> +#include <malloc.h>
> +#include <asm/io.h>
> +#include <asm/nios2-io.h>
> +
> +static int altera_serial_setbaudrate(struct console_device *cdev, int baudrate)
> +{
> +     volatile struct nios_uart * uart = (struct nios_uart *)cdev->dev->map_base;

                                  ^

please remove this space.

Damn, I used the checkpatch script though, should have seen it....

 

no volatiles please.

Yeah I know but last time I removed those volatile, the driver didn't work anymore ?!? I've done a test after I read your reply and it

works without so I'll remove them.

In Documentation/volatile-considered-harmful.txt, one can read:

There are still a few rare situations where volatile makes sense in the kernel:

       - The above-mentioned accessor functions might use volatile on

         architectures where direct I/O memory access does work.  Essentially,

         each accessor call becomes a little critical section on its own and

         ensures that the access happens as expected by the programmer.

      (......)

       - Pointers to data structures in coherent memory which might be modified

         by I/O devices can, sometimes, legitimately be volatile.  A ring buffer

         used by a network adapter, where that adapter changes pointers to

         indicate which descriptors have been processed, is an example of this

         type of situation.

So is it that bad to use volatile ?

 

 

> +     unsigned div;

> +
> +     div = (CPU_FREQ / baudrate) - 1;
> +     writel(div, &uart->divisor);
> +
> +     return 0;
> +}
> +
> +static void altera_serial_putc(struct console_device *cdev, char c)
> +{
> +     volatile struct nios_uart *uart = (struct nios_uart *)cdev->dev->map_base;
> +
> +     if (c == '\n')
> +             altera_serial_putc(cdev, '\r');

Why this? Only the at91rm9200 driver does this and I don't know why....

Because I copied it from U-Boot :) Removed.

 

> +
> +     while ((readl(&uart->status) & NIOS_UART_TRDY) == 0);

> +     writel(c, &uart->txdata);
> +}
> +
> +static int altera_serial_tstc(struct console_device *cdev)
> +{
> +     volatile struct nios_uart *uart = (struct nios_uart *)cdev->dev->map_base;

No volatiles

> +
> +     return readl(&uart->status) & NIOS_UART_RRDY;
> +}
> +
> +static int altera_serial_getc(struct console_device *cdev)
> +{
> +     volatile struct nios_uart *uart = (struct nios_uart *)cdev->dev->map_base;

dito

> +
> +     while (altera_serial_tstc(cdev) == 0);

please add a comment, or at least an empty line after while

Do you guys put the semi colon on the next line also ?

 

> +     return readl(&uart->rxdata) & 0x000000FF;

> +}
> +
> +static int altera_serial_probe(struct device_d *dev)
> +{
> +     struct console_device *cdev;
> +
> +     cdev = malloc(sizeof(struct console_device));

malloc can fail, use xmalloc instead

Done.

 

> +     dev->type_data = cdev;

> +     cdev->dev = dev;
> +     cdev->f_caps = CONSOLE_STDIN | CONSOLE_STDOUT | CONSOLE_STDERR;
> +     cdev->tstc = altera_serial_tstc;
> +     cdev->putc = altera_serial_putc;
> +     cdev->getc = altera_serial_getc;
> +     cdev->setbrg = altera_serial_setbaudrate;
> +
> +     console_register(cdev);
> +
> +     return 0;
> +}
> +
> +static struct driver_d altera_serial_driver = {
> +     .name  = "altera_serial",

            ^^
one space is enough

Copied from blackfin :) Patch it ?

 

> +     .probe = altera_serial_probe,

> +};
> +
> +static int altera_serial_init(void)
> +{
> +     register_driver(&altera_serial_driver);
> +     return 0;

I think return register_driver(); is preferred.

Copied from blackfin :) Patch it ?

 

> +}
> +
> +console_initcall(altera_serial_init);
> +

cheers, Marc

--
Pengutronix e.K.                  | Marc Kleine-Budde           |
Industrial Linux Solutions        | Phone: +49-231-2826-924     |
Vertretung West/Dortmund          | Fax:   +49-5121-206917-5555 |
Amtsgericht Hildesheim, HRA 2686  | http://www.pengutronix.de   |

Do you want me to patch files each time I see a wrong coding style in the code ? Because I'm so focused

on that since I'm working on the nios2 port (and since I display tabs and white spaces in my favorite editor :))

that I can see coding style mistakes in a lot of files :) (for example, leading spaces in serial_imx line 222, 228, 

245, 313,.... checkpatch gives 9 errors and 44 warnings on this file).

Regards,

Franck.