Return-Path: <cyrus@holtmann.org>
MIME-Version: 1.0
In-Reply-To: <CABBYNZ+HiG58wPbXkgikR_7m9APwQLYn4E6VhzBdcbmSVPGc6Q@mail.gmail.com>
References: <1337940693-3417-1-git-send-email-mikel.astiz.oss@gmail.com>
	<1337940693-3417-4-git-send-email-mikel.astiz.oss@gmail.com>
	<CABBYNZ+HiG58wPbXkgikR_7m9APwQLYn4E6VhzBdcbmSVPGc6Q@mail.gmail.com>
Date: Fri, 25 May 2012 12:45:18 +0200
Message-ID: <CANT-zCX32WtZbHpPRdWjU114wvJAwfj0JZyGU9sA_Dp-RkhBYA@mail.gmail.com>
Subject: Re: [PATCH obexd v1 03/16] client: Add progress property to transfer
From: Mikel Astiz <mikel.astiz.oss@gmail.com>
To: Luiz Augusto von Dentz <luiz.dentz@gmail.com>
Cc: linux-bluetooth@vger.kernel.org,
	Mikel Astiz <mikel.astiz@bmw-carit.de>
Content-Type: text/plain; charset=ISO-8859-1
Sender: linux-bluetooth-owner@vger.kernel.org
List-ID: <linux-bluetooth.vger.kernel.org>

Hi Luiz,

On Fri, May 25, 2012 at 12:35 PM, Luiz Augusto von Dentz
<luiz.dentz@gmail.com> wrote:
> Hi Mikel,
>
> On Fri, May 25, 2012 at 1:11 PM, Mikel Astiz <mikel.astiz.oss@gmail.com> wrote:
>
>> +static void transfer_notify_progress(struct obc_transfer *transfer)
>> +{
>> + ? ? ? gint64 now;
>> + ? ? ? gint64 notify;
>> +
>> + ? ? ? DBG("Transfer %p progress: %lu bytes", transfer, transfer->transferred);
>> +
>> + ? ? ? if (transfer->path == NULL)
>> + ? ? ? ? ? ? ? return;
>> +
>> + ? ? ? if (transfer->transferred == transfer->transferred_dbus)
>> + ? ? ? ? ? ? ? return;
>> +
>> + ? ? ? now = g_get_monotonic_time();
>> + ? ? ? notify = transfer->transferred_dbus_time +
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? TRANSFER_PROGRESS_PERIOD * 1000;
>> +
>> + ? ? ? if ((transfer->transferred != transfer->size) && (now < notify))
>> + ? ? ? ? ? ? ? return;
>> +
>> + ? ? ? transfer->transferred_dbus = transfer->transferred;
>> + ? ? ? transfer->transferred_dbus_time = now;
>> +
>> + ? ? ? obex_dbus_signal_property_changed(transfer->conn,
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? transfer->path,
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? TRANSFER_INTERFACE, "Progress",
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? DBUS_TYPE_INT64,
>> + ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? &transfer->transferred_dbus);
>> +}
>
> I would suggest doing it in a different manner using
> g_timeout_add_seconds and storing the id, so you just need to know if
> the timer is running by checking the id, if there is no timer running
> you start it, if there is you just update the bytes transferred and
> wait the timeout to be fired where you emit the signal. If in the
> meantime the transfer is completed you just have to cleanup using
> g_source_remove, this should be simpler because you don't have to
> bother how much time has passed on every packet.

That would certainly avoid a call to g_get_monotonic_time() every time
we receive a packet, but on the other hand we would be waking up the
process one additional time per second. I'm not sure if there is any
performance improvement there.

>From the implementation complexity point of view both approaches equally simple.

Cheers,
Mikel