ALSA: dice: perform sequence replay for media clock recovery

This commit takes ALSA dice driver to perform sequence replay for media
clock recovery.

Unlike the other types of device, DICE-based devices interpret the value
of syt field of CIP header in rx packets as presentation time for audio
playback, thus it's required for driver to compute value for outgoing
packet adequate to the device. It's done by media clock recovery by
handling tx packets.

The device starts packet transmission immediately at operation to
GLOBAL_ENABLE thus on-the-fly mode is not required.

DICE ASICs supports several pairs of isochronous packet streams.
Actually, maximum two pairs of streams are supported by devices.
We have three cases regarding to the number of streams:

1. a pair of streams
2. two tx packet streams and one rx packet streams
3. one tx packet streams and two rx packet streams
4. two pair of streams

The decision of playback timing is slightly different in the four cases.

In the case 1, sequence replay in the pair results in suitable playback
timing.

In the case 2, sequence replay from the first tx packet stream to rx
packet stream results in suitable playback timing.

In the case 3, sequence replay from tx packet stream to all of rx packet
stream results in suitable playback timing. Furthermore, the cycle to
start receiving packets should be the same between all rx packet streams.

In the case 4, sequence replay in each pair results in suitable playback
timing. Furthermore, the cycle to start receiving packets should be the
same between all rx packet streams.

The sequence replay is tested with below models:

* For case 1:
  * TC Electronic Konnekt 24d (DiceII)
  * TC Electronic Konnekt 8 (DiceII)
  * TC Electronic Konnekt Live (DiceII)
  * TC Electronic Impact Twin (DiceII)
  * TC Electronic Digital Konnekt X32 (DiceII)
  * TC Electronic Desktop Konnekt 6 (TCD2220)
  * Solid State Logic Duende Classic (DiceII)
  * Solid State Logic Duende Mini (DiceII)
  * PreSonus FireStudio Project (TCD2210)
  * PreSonus FireStudio Mobile (TCD2210)
  * Lexicon I-ONIX FW810s (TCD2220)
  * Avid Mbox 3 Pro (TCD2220)

* For case 2 (but case 1 depends on sampling transfer frequency):
  * Alesis iO 26 (DiceII)
  * Alesis iO 14 (DiceII)
  * Alesis MultiMix 12 FireWire (DiceII)
  * Focusrite Saffire Pro 26 (TCD2220)

* For case 3 (but case 1 depends on sampling transfer frequency):
  * M-Audio Profire 610 (TCD2220)
  * Loud Technology Mackie Onyx Blackbird (TCD2210)

* For case 4:
  * TC Electronic Studio Konnekt 48 (DiceII + TCD2220)
  * PreSonus FireStudio (DiceII)
  * M-Audio Profire 2626 (TCD2220)
  * Focusrite Liquid Saffire 56 (TCD2220)
  * Focusrite Saffire Pro 40 (TCD2220)
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Link: https://lore.kernel.org/r/20210601081753.9191-3-o-takashi@sakamocchi.jpSigned-off-by: Takashi Iwai <tiwai@suse.de>

sound/firewire/dice/dice-stream.c

@@ -444,7 +444,11 @@ int snd_dice_stream_start_duplex(struct snd_dice *dice)
 			goto error;
 		}
 
-		err = amdtp_domain_start(&dice->domain, 0, false, false);
+		// MEMO: The device immediately starts packet transmission when enabled. Some
+		// devices are strictly to generate any discontinuity in the sequence of tx packet
+		// when they receives invalid sequence of presentation time in CIP header. The
+		// sequence replay for media clock recovery can suppress the behaviour.
+		err = amdtp_domain_start(&dice->domain, 0, true, false);
 		if (err < 0)
 			goto error;