mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: Andrey Smirnov <andrew.smirnov@gmail.com>
To: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Barebox List <barebox@lists.infradead.org>
Subject: Re: [PATCH 2/3] console: Add simplified 'serdev' framework from Linux kernel
Date: Mon, 9 Apr 2018 09:40:46 -0700	[thread overview]
Message-ID: <CAHQ1cqHdhHiknrnzfU_9YLc_HiyNVDSPZyGNNsHcrAhgUoK_ig@mail.gmail.com> (raw)
In-Reply-To: <20180403065436.p2vec7ntdja55hts@pengutronix.de>

On Mon, Apr 2, 2018 at 11:54 PM, Sascha Hauer <s.hauer@pengutronix.de> wrote:
> Hi Andrey,
>
> Some comments inside.
>
>
> On Mon, Mar 26, 2018 at 06:09:14AM -0700, Andrey Smirnov wrote:
>> Port 'serdev' UART-slave deivce framework found in recent Linux
>> kernels (post 4.13) in order to be able to port 'serdev' slave drivers
>> from Linux.
>>
>> Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
>> @@ -323,6 +324,17 @@ int console_register(struct console_device *newcdev)
>>               dev->parent = newcdev->dev;
>>       platform_device_register(dev);
>>
>> +     newcdev->open_count = 0;
>> +
>> +     /*
>> +      * If our console deive is a serdev, we skip the creation of
>
> s/deive/device/

Will fix in v2.

>
>> +      * corresponding entry in /dev as well as registration in
>> +      * console_list and just go straigh to populating child
>
> s/straigh/straight/

Ditto.

>
>> +      * devices.
>> +      */
>> +     if (serdev_node)
>> +             return of_platform_populate(serdev_node, NULL, dev);
>
> How is this going to be used? A serdev driver binds to the serdev_node
> and then it probably needs to get a pointer to the console device,
> right? How does the driver accomplish this?
>

Serdev slave driver doesn't hold explicit pointer to console device,
instead accessing it via point to serdev_device. The latter could
obtained by calling to_serdev_device(dev->parent), where dev is
device_d given to slave driver's probe function.


>> +/**
>> + * struct serdev_device - Basic representation of an serdev device
>> + *
>> + * @dev:             Corresponding device
>> + * @fifo:            Circular buffer used for console draining
>> + * @buf:             Buffer used to pass Rx data to consumers
>> + * @poller           Async poller used to poll this serdev
>> + * @polling_interval:        Async poller periodicity
>> + * @polling_window:  Duration of a single busy loop poll
>> + * @receive_buf:     Function called with data received from device;
>> + *                   returns number of bytes accepted;
>> + */
>> +struct serdev_device {
>> +     struct device_d *dev;
>> +     struct kfifo *fifo;
>> +     unsigned char *buf;
>> +     struct poller_async poller;
>> +     uint64_t polling_interval;
>> +     uint64_t polling_window;
>> +
>> +     int (*receive_buf)(struct serdev_device *, const unsigned char *,
>> +                        size_t);
>> +};
>> +
>> +int serdev_device_open(struct serdev_device *);
>> +unsigned int serdev_device_set_baudrate(struct serdev_device *, unsigned int);
>> +int serdev_device_write(struct serdev_device *, const unsigned char *,
>> +                     size_t, unsigned long);
>
> So a serdev driver uses serdev_device_write() to send characters out. To
> receive characters it has to implement serdev_device->receive_buf,
> right?

Right. I tried to implement exactly the same API that Linux's serdev
API provides.

> What kind of devices did you implement this for?

I ported serdev in support of porting the driver for RAVE SP which is
a small microcontroller device found many ZII board including RDU2. It
implement a whole bunch of various functionality including watchdog,
parameter EEPROM, sensor access, backlight control, button input event
generation, etc.

> For devices which send data without request (GPS?) this seems the way to go. For
> others a synchronous receive function might be good, no?
>

I didn't implement anything like that mostly because Linux serdev API
doesn't and any ported driver wouldn't have any need for those
functions. But in general, I am not sure how useful synchronous
receive function would be. In my experience, devices like that usually
implement some binary transport protocol with packetization/escape
sequences on top of UART, which usually requires a state machine
operating with byte granularity as the data comes in to parse
responses correctly and synchronous APIs are not extremely useful for
that kind of a use-case.

FWIW, since serdev API is integrated into poller infrastructure it is
pretty trivial to write blocking code with it. Here's how I use it in
my driver to implement request-response type of a function:

rave_sp_write(sp, data, data_size);
/*
* is_timeout will implicitly poll serdev via poller
* infrastructure
*/
while (!is_timeout(start, SECOND) && !reply.received)
   ;


Thanks,
Andrey Smirnov

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

  reply	other threads:[~2018-04-09 16:41 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-26 13:09 [PATCH 0/3] Linux's serdev framwork port Andrey Smirnov
2018-03-26 13:09 ` [PATCH 1/3] console: Introduce console_drain() Andrey Smirnov
2018-03-26 13:09 ` [PATCH 2/3] console: Add simplified 'serdev' framework from Linux kernel Andrey Smirnov
2018-04-03  6:54   ` Sascha Hauer
2018-04-09 16:40     ` Andrey Smirnov [this message]
2018-04-11  8:39       ` Sascha Hauer
2018-04-12 18:22         ` Andrey Smirnov
2018-03-26 13:09 ` [PATCH 3/3] serial: Check result of console_unregister() Andrey Smirnov
2018-04-03  7:04   ` Sascha Hauer
2018-04-09 16:00     ` Andrey Smirnov
2018-04-11  8:34       ` Sascha Hauer
2018-04-12 18:20         ` Andrey Smirnov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAHQ1cqHdhHiknrnzfU_9YLc_HiyNVDSPZyGNNsHcrAhgUoK_ig@mail.gmail.com \
    --to=andrew.smirnov@gmail.com \
    --cc=barebox@lists.infradead.org \
    --cc=s.hauer@pengutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox