All of lore.kernel.org
 help / color / mirror / Atom feed
From: Takashi Sakamoto <o-takashi@sakamocchi.jp>
To: clemens@ladisch.de, tiwai@suse.de
Cc: alsa-devel@alsa-project.org, ffado-devel@lists.sf.net
Subject: [PATCH 06/25] ALSA: firewire-lib: add throttle for MIDI data rate
Date: Sat, 22 Aug 2015 18:19:22 +0900	[thread overview]
Message-ID: <1440235181-18181-7-git-send-email-o-takashi@sakamocchi.jp> (raw)
In-Reply-To: <1440235181-18181-1-git-send-email-o-takashi@sakamocchi.jp>

Typically, the target devices have internal buffer to adjust output of
received MIDI messages for MIDI serial bus, while the capacity of the
buffer is limited. IEEE 1394 transactions can transfer more MIDI messages
than MIDI serial bus can. This can cause buffer over flow in device side.

This commit adds throttle to limit MIDI data rate by counting intervals
by jiffies between two MIDI messages. Usual MIDI messages consists of two
or three bytes. This requires 1.302 to 1.953 mili-seconds interval between
these messages. Using jiffies for this purpose is not perfect idea.
Further work is needed for better implementation.

Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
---
 sound/firewire/lib.c | 26 +++++++++++++++++++++++++-
 sound/firewire/lib.h |  1 +
 2 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/sound/firewire/lib.c b/sound/firewire/lib.c
index 83d20b6..5d04632 100644
--- a/sound/firewire/lib.c
+++ b/sound/firewire/lib.c
@@ -76,6 +76,9 @@ static void async_midi_port_callback(struct fw_card *card, int rcode,
 
 	if (rcode == RCODE_COMPLETE && substream != NULL)
 		snd_rawmidi_transmit_ack(substream, port->consume_bytes);
+	else if (!rcode_is_permanent_error(rcode))
+		/* To start next transaction immediately for recovery. */
+		port->next_tick = 0;
 
 	port->idling = true;
 
@@ -99,6 +102,12 @@ static void midi_port_tasklet(unsigned long data)
 	if (substream == NULL || snd_rawmidi_transmit_empty(substream))
 		return;
 
+	/* Do it in next chance. */
+	if (time_before64(get_jiffies_64(), port->next_tick)) {
+		tasklet_schedule(&port->tasklet);
+		return;
+	}
+
 	/*
 	 * Fill the buffer. The callee must use snd_rawmidi_transmit_peek().
 	 * Later, snd_rawmidi_transmit_ack() is called.
@@ -107,8 +116,10 @@ static void midi_port_tasklet(unsigned long data)
 	port->consume_bytes = port->packetize(substream, port->buf);
 	if (port->consume_bytes <= 0) {
 		/* Do it in next chance, immediately. */
-		if (port->consume_bytes == 0)
+		if (port->consume_bytes == 0) {
+			port->next_tick = 0;
 			tasklet_schedule(&port->tasklet);
+		}
 		return;
 	}
 
@@ -118,6 +129,18 @@ static void midi_port_tasklet(unsigned long data)
 	else
 		type = TCODE_WRITE_BLOCK_REQUEST;
 
+	/*
+	 * Set interval to next transaction.
+	 *
+	 * TODO: Using jiffies to count 31,250 bits per second may be a bad
+	 * idea. The jiffies is assumed to count up constantly according to HZ,
+	 * while recent Linux goes to dynamic ticks or tickless system. Further
+	 * work is required and the interval should be set with the other
+	 * services, i.e. by using monotonic time with higher resolution.
+	 */
+	port->next_tick = jiffies_64 +
+		msecs_to_jiffies(port->consume_bytes * 8 * 1000 / 31250);
+
 	/* Start this transaction. */
 	port->idling = false;
 
@@ -163,6 +186,7 @@ int snd_fw_async_midi_port_init(struct snd_fw_async_midi_port *port,
 	port->addr = addr;
 	port->packetize = packetize;
 	port->idling = true;
+	port->next_tick = 0;
 
 	tasklet_init(&port->tasklet, midi_port_tasklet, (unsigned long)port);
 
diff --git a/sound/firewire/lib.h b/sound/firewire/lib.h
index 9d76f5c..bfc353b 100644
--- a/sound/firewire/lib.h
+++ b/sound/firewire/lib.h
@@ -26,6 +26,7 @@ struct snd_fw_async_midi_port {
 	struct fw_device *parent;
 	struct tasklet_struct tasklet;
 	bool idling;
+	__u64 next_tick;
 
 	__u64 addr;
 	struct fw_transaction transaction;
-- 
2.1.4

  parent reply	other threads:[~2015-08-22  9:19 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-08-22  9:19 [PATCH 00/25 v2] ALSA: support AMDTP variants Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 01/25] ALSA: firewire-lib: rename 'amdtp' to 'amdtp-stream' for functional separation Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 02/25] ALSA: firewire-lib: functional separation for packet transmission layer and data processing layer Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 03/25] ALSA: firewire-lib: add helper functions for asynchronous MIDI port Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 04/25] ALSA: firewire-lib: add a restriction for a transaction at once Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 05/25] ALSA: firewire-lib: schedule tasklet again when MIDI substream has rest of MIDI messages Takashi Sakamoto
2015-08-22  9:19 ` Takashi Sakamoto [this message]
2015-08-22  9:19 ` [PATCH 07/25] ALSA: firewire-lib: avoid endless loop to transfer MIDI messages at fatal error Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 08/25] ALSA: firewire-digi00x: add skeleton for Digi 002/003 family Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 09/25] ALSA: firewire-digi00x: add data processing layer Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 10/25] ALSA: firewire-digi00x: add stream functionality Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 11/25] ALSA: firewire-digi00x: add proc node to show clock status Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 12/25] ALSA: firewire-digi00x: add PCM functionality Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 13/25] ALSA: firewire-digi00x: add MIDI functionality Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 14/25] ALSA: firewire-digi00x: add hwdep interface Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 15/25] ALSA: firewire-digi00x: add support for asynchronous messaging Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 16/25] ALSA: firewire-digi00x: add support for MIDI ports for machine control Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 17/25] ALSA: firewire-tascam: add skeleton for TASCAM FireWire series Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 18/25] ALSA: firewire-tascam: add a structure for model-dependent parameters Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 19/25] ALSA: firewire-tascam: add proc node to show firmware information Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 20/25] ALSA: firewire-tascam: add data processing layer Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 21/25] ALSA: firewire-tascam: add streaming functionality Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 22/25] ALSA: firewire-tascam: add PCM functionality Takashi Sakamoto
2015-08-24 21:12   ` Takashi Iwai
2015-08-25 11:56     ` Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 23/25] ALSA: firewire-tascam: add transaction functionality Takashi Sakamoto
2015-08-22  9:19 ` [PATCH 24/25] ALSA: firewire-tascam: add MIDI functionality Takashi Sakamoto
2015-08-24 20:59   ` Takashi Iwai
2015-08-22  9:19 ` [PATCH 25/25] ALSA: firewire-tascam: add hwdep interface Takashi Sakamoto
2015-08-23  8:04 ` [PATCH 00/25 v2] ALSA: support AMDTP variants Takashi Iwai
2015-08-23 14:58   ` Takashi Sakamoto
2015-08-24  6:26     ` Takashi Iwai
2015-08-25  5:24       ` Takashi Iwai
2015-08-25 11:47         ` Takashi Sakamoto
2015-08-25 12:03           ` Takashi Iwai
2015-08-29  6:00             ` Takashi Sakamoto
  -- strict thread matches above, loose matches on Subject: below --
2015-08-13  0:19 [PATCH 00/25] " Takashi Sakamoto
2015-08-13  0:20 ` [PATCH 06/25] ALSA: firewire-lib: add throttle for MIDI data rate Takashi Sakamoto
2015-08-13 21:34   ` Takashi Sakamoto

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=1440235181-18181-7-git-send-email-o-takashi@sakamocchi.jp \
    --to=o-takashi@sakamocchi.jp \
    --cc=alsa-devel@alsa-project.org \
    --cc=clemens@ladisch.de \
    --cc=ffado-devel@lists.sf.net \
    --cc=tiwai@suse.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.