From: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Cc: linux-input@vger.kernel.org, linux-kernel@vger.kernel.org,
aniroop.mathur@gmail.com, Aniroop Mathur <a.mathur@samsung.com>,
Benjamin Tissoires <benjamin.tissoires@gmail.com>,
David Herrmann <dh.herrmann@gmail.com>,
Peter Hutterer <peter.hutterer@who-t.net>,
Henrik Rydberg <rydberg@bitmath.org>
Subject: Re: [PATCH] Input: evdev - drop partial events after emptying the buffer
Date: Sat, 2 Jan 2016 20:39:21 -0800 [thread overview]
Message-ID: <20160103043921.GB23680@dtor-ws> (raw)
In-Reply-To: <1451513207-712-1-git-send-email-a.mathur@samsung.com>
On Thu, Dec 31, 2015 at 03:36:47AM +0530, Aniroop Mathur wrote:
> This patch introduces concept to drop partial events in evdev handler
> itself after emptying the buffer which are dropped by all evdev
> clients in userspace after SYN_DROPPED occurs.
> This in turn saves space of partial events in evdev handler buffer
> and reduces evdev client reading requests.
Let's add a few people who write consumer code to see if this is
something that they consider useful.
>
> Signed-off-by: Aniroop Mathur <a.mathur@samsung.com>
> ---
> drivers/input/evdev.c | 49 +++++++++++++++++++++++++++++++++++++++++++++----
> 1 file changed, 45 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c
> index e9ae3d5..e7b612e 100644
> --- a/drivers/input/evdev.c
> +++ b/drivers/input/evdev.c
> @@ -58,6 +58,7 @@ struct evdev_client {
> struct list_head node;
> unsigned int clk_type;
> bool revoked;
> + bool drop_pevent; /* specifies whether partial events need to be dropped */
> unsigned long *evmasks[EV_CNT];
> unsigned int bufsize;
> struct input_event buffer[];
> @@ -192,6 +193,7 @@ static int evdev_set_clk_type(struct evdev_client *client, unsigned int clkid)
> {
> unsigned long flags;
> unsigned int clk_type;
> + struct input_event *ev;
>
> switch (clkid) {
>
> @@ -218,6 +220,17 @@ static int evdev_set_clk_type(struct evdev_client *client, unsigned int clkid)
> spin_lock_irqsave(&client->buffer_lock, flags);
>
> if (client->head != client->tail) {
> +
> + /*
> + * Set drop_pevent to true if last event packet is
> + * not stored completely in buffer.
> + */
> + client->head--;
> + client->head &= client->bufsize - 1;
> + ev = &client->buffer[client->head];
> + if (!(ev->type == EV_SYN && ev->code == SYN_REPORT))
> + client->drop_pevent = true;
> +
> client->packet_head = client->head = client->tail;
> __evdev_queue_syn_dropped(client);
> }
> @@ -228,31 +241,51 @@ static int evdev_set_clk_type(struct evdev_client *client, unsigned int clkid)
> return 0;
> }
>
> -static void __pass_event(struct evdev_client *client,
> +static bool __pass_event(struct evdev_client *client,
> const struct input_event *event)
> {
> + struct input_event *prev_ev;
> +
> client->buffer[client->head++] = *event;
> client->head &= client->bufsize - 1;
>
> if (unlikely(client->head == client->tail)) {
> /*
> - * This effectively "drops" all unconsumed events, leaving
> - * EV_SYN/SYN_DROPPED plus the newest event in the queue.
> + * This effectively "drops" all unconsumed events, storing
> + * EV_SYN/SYN_DROPPED and the newest event in the queue but
> + * only if it is not part of partial packet.
> + * Set drop_pevent to true if last event packet is not stored
> + * completely in buffer and set to false if SYN_REPORT occurs.
> */
> +
> client->tail = (client->head - 2) & (client->bufsize - 1);
>
> + prev_ev = &client->buffer[client->tail];
> + if (!(prev_ev->type == EV_SYN && prev_ev->code == SYN_REPORT)) {
> + client->drop_pevent = true;
> + client->head--;
> + client->head &= client->bufsize - 1;
> + }
> +
> client->buffer[client->tail].time = event->time;
> client->buffer[client->tail].type = EV_SYN;
> client->buffer[client->tail].code = SYN_DROPPED;
> client->buffer[client->tail].value = 0;
>
> client->packet_head = client->tail;
> +
> + if (event->type == EV_SYN && event->code == SYN_REPORT) {
> + client->drop_pevent = false;
> + return true;
> + }
> }
>
> if (event->type == EV_SYN && event->code == SYN_REPORT) {
> client->packet_head = client->head;
> kill_fasync(&client->fasync, SIGIO, POLL_IN);
> }
> +
> + return false;
> }
>
> static void evdev_pass_values(struct evdev_client *client,
> @@ -284,10 +317,18 @@ static void evdev_pass_values(struct evdev_client *client,
> wakeup = true;
> }
>
> + /* drop partial events until SYN_REPORT occurs */
> + if (client->drop_pevent) {
> + if (v->type == EV_SYN && v->code == SYN_REPORT)
> + client->drop_pevent = false;
> + continue;
> + }
> +
> event.type = v->type;
> event.code = v->code;
> event.value = v->value;
> - __pass_event(client, &event);
> + if (__pass_event(client, &event))
> + wakeup = false;
> }
>
> spin_unlock(&client->buffer_lock);
> --
> 2.6.2
>
--
Dmitry
next prev parent reply other threads:[~2016-01-03 4:39 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-30 22:06 [PATCH] Input: evdev - drop partial events after emptying the buffer Aniroop Mathur
2016-01-03 4:39 ` Dmitry Torokhov [this message]
2016-01-03 23:38 ` Peter Hutterer
2016-01-04 16:32 ` 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=20160103043921.GB23680@dtor-ws \
--to=dmitry.torokhov@gmail.com \
--cc=a.mathur@samsung.com \
--cc=aniroop.mathur@gmail.com \
--cc=benjamin.tissoires@gmail.com \
--cc=dh.herrmann@gmail.com \
--cc=linux-input@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=peter.hutterer@who-t.net \
--cc=rydberg@bitmath.org \
/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).