linux-input.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Aniroop Mathur <aniroop.mathur@gmail.com>
To: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: "linux-input@vger.kernel.org" <linux-input@vger.kernel.org>,
	a.mathur@samsung.com, Thomas Gleixner <tglx@linutronix.de>,
	John Stultz <john.stultz@linaro.org>
Subject: Re: [PATCH] Input: Add support for CLOCK_BOOTTIME
Date: Thu, 18 Dec 2014 03:03:26 +0530	[thread overview]
Message-ID: <CADYu30_n-mc8X8h+v4PX7Mq9TSyKY1Hh=zEeKBcHMDrikB9xmA@mail.gmail.com> (raw)
In-Reply-To: <20141217182011.GA32399@dtor-ws>

Hello Mr. Torokhov,

On Wed, Dec 17, 2014 at 11:50 PM, Dmitry Torokhov
<dmitry.torokhov@gmail.com> wrote:
> Hi Aniroop,
>
> On Tue, Dec 16, 2014 at 10:26:34PM +0530, Aniroop Mathur wrote:
>> This patch adds support for CLOCK_BOOTTIME for input event timestamp.
>> CLOCK_BOOTTIME includes suspend time, so it would allow aplications
>> to get correct time difference between two events even when system
>> resumes from suspend state.
>>
>> Signed-off-by: Aniroop Mathur <a.mathur@samsung.com>
>> ---
>>  drivers/input/evdev.c | 33 +++++++++++++++++++++------------
>>  1 file changed, 21 insertions(+), 12 deletions(-)
>>
>> diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c
>> index de05545..7825794 100644
>> --- a/drivers/input/evdev.c
>> +++ b/drivers/input/evdev.c
>> @@ -28,6 +28,13 @@
>>  #include <linux/cdev.h>
>>  #include "input-compat.h"
>>
>> +enum clock_type {
>> +     REAL = 0,
>> +     MONO,
>> +     BOOT,
>> +     CLK_MAX
>> +};
>> +
>>  struct evdev {
>>       int open;
>>       struct input_handle handle;
>> @@ -108,8 +115,9 @@ static void evdev_queue_syn_dropped(struct evdev_client *client)
>>       struct input_event ev;
>>       ktime_t time;
>>
>> -     time = (client->clkid == CLOCK_MONOTONIC) ?
>> -             ktime_get() : ktime_get_real();
>> +     time = (client->clkid == CLOCK_REALTIME) ?
>> +             ktime_get_real() : (client->clkid == CLOCK_MONOTONIC) ?
>> +             ktime_get() : ktime_get_boottime();
>>

Is this okay ?
To improve readabilty,
how about adding a new funtion to get time depending upon clkid like below ?
time = get_clk_time(client->clkid);

static ktime_t get_clk_time(int id)
{
        switch (id) {

        case CLOCK_REALTIME:     return ktime_get_real();
        case CLOCK_MONOTONIC: return ktime_get();
        case CLOCK_BOOTTIME:    return ktime_get_bootime();
        }
}

>>       ev.time = ktime_to_timeval(time);
>>       ev.type = EV_SYN;
>> @@ -159,7 +167,7 @@ static void __pass_event(struct evdev_client *client,
>>
>>  static void evdev_pass_values(struct evdev_client *client,
>>                       const struct input_value *vals, unsigned int count,
>> -                     ktime_t mono, ktime_t real)
>> +                     ktime_t *ev_time)
>>  {
>>       struct evdev *evdev = client->evdev;
>>       const struct input_value *v;
>> @@ -169,8 +177,9 @@ static void evdev_pass_values(struct evdev_client *client,
>>       if (client->revoked)
>>               return;
>>
>> -     event.time = ktime_to_timeval(client->clkid == CLOCK_MONOTONIC ?
>> -                                   mono : real);
>> +     event.time = ktime_to_timeval(client->clkid == CLOCK_REALTIME ?
>> +                     ev_time[REAL] : client->clkid == CLOCK_MONOTONIC ?
>> +                     ev_time[MONO] : ev_time[BOOT]);
>
> This becomes unwieldy. Should we have client->clk_type instead of
> client->clkid and convert CLOCK_* into your EVDEV_CLK_{REAL|MONO|BOOT}
> when setting it in ioctl and then do
>
>         event.time = ktime_to_timeval(ev_time[client->clk_offset]);
>

Sure, using clk_type would be better and improve code readability.

I am afraid how converting CLOCK_* into your
EVDEV_CLK_{REAL|MONO|BOOT} will help
because we cannot have ev_time array size more than CLK_MAX (3).

Did you mean to have following changes:

1. Add clk_type variable
struct evdev_client {
         int clkid:
      + int clk_type;
}

2. Add new funtion for setting clk_type to improve readability
static int set_clk_type(struct evdev_client *client, int id)
{
        switch (id) {

        case CLOCK_REALTIME:     client->clk_type = EV_REAL;  break;
        case CLOCK_MONOTONIC: client->clk_type = EV_MONO; break;
        case CLOCK_BOOTTIME:    client->clk_type = EV_BOOT;  break;
        default: return -1
        }

        return 0;
}

3. Addition in ioctl call
    case EVIOCSCLOCKID:
    ...
    if (!set_clk_type())
           return -EINVAL;

     client->clkid = i;
     return 0;

4. Change in evdev_pass_values function
event.time = ktime_to_timeval(ev_time[client->clk_type]);

>>
>>       /* Interrupts are disabled, just acquire the lock. */
>>       spin_lock(&client->buffer_lock);
>> @@ -198,21 +207,21 @@ static void evdev_events(struct input_handle *handle,
>>  {
>>       struct evdev *evdev = handle->private;
>>       struct evdev_client *client;
>> -     ktime_t time_mono, time_real;
>> +     ktime_t ev_time[CLK_MAX];
>>
>> -     time_mono = ktime_get();
>> -     time_real = ktime_mono_to_real(time_mono);
>> +     ev_time[MONO] = ktime_get();
>> +     ev_time[REAL] = ktime_mono_to_real(ev_time[MONO]);
>> +     ev_time[BOOT] = ktime_get_boottime();
>
> I do not think we want to use ktime_get_boottime() here; I'd rather you
> use ktime_mono_to_anY(ev_time[MONO], TK_OFFS_BOOT) so that all 3 times
> are consistent.
>

Sure. I will send the updated patch with this change.

>>
>>       rcu_read_lock();
>>
>>       client = rcu_dereference(evdev->grab);
>>
>>       if (client)
>> -             evdev_pass_values(client, vals, count, time_mono, time_real);
>> +             evdev_pass_values(client, vals, count, ev_time);
>>       else
>>               list_for_each_entry_rcu(client, &evdev->client_list, node)
>> -                     evdev_pass_values(client, vals, count,
>> -                                       time_mono, time_real);
>> +                     evdev_pass_values(client, vals, count, ev_time);
>>
>>       rcu_read_unlock();
>>  }
>> @@ -874,7 +883,7 @@ static long evdev_do_ioctl(struct file *file, unsigned int cmd,
>>       case EVIOCSCLOCKID:
>>               if (copy_from_user(&i, p, sizeof(unsigned int)))
>>                       return -EFAULT;
>> -             if (i != CLOCK_MONOTONIC && i != CLOCK_REALTIME)
>> +             if (i != CLOCK_MONOTONIC && i != CLOCK_REALTIME && i != CLOCK_BOOTTIME)
>>                       return -EINVAL;
>>               client->clkid = i;
>>               return 0;
>> --
>> 1.9.1
>>
>
> Thanks.
>
> --
> Dmitry

Thanks,
Aniroop Mathur

  reply	other threads:[~2014-12-17 21:33 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-12-16 16:56 [PATCH] Input: Add support for CLOCK_BOOTTIME Aniroop Mathur
2014-12-17 18:20 ` Dmitry Torokhov
2014-12-17 21:33   ` Aniroop Mathur [this message]
2014-12-17 22:03     ` Dmitry Torokhov
2014-12-17 23:11       ` Aniroop Mathur
  -- strict thread matches above, loose matches on Subject: below --
2014-12-16  8:23 Aniroop, Mathur

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='CADYu30_n-mc8X8h+v4PX7Mq9TSyKY1Hh=zEeKBcHMDrikB9xmA@mail.gmail.com' \
    --to=aniroop.mathur@gmail.com \
    --cc=a.mathur@samsung.com \
    --cc=dmitry.torokhov@gmail.com \
    --cc=john.stultz@linaro.org \
    --cc=linux-input@vger.kernel.org \
    --cc=tglx@linutronix.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;
as well as URLs for NNTP newsgroup(s).