linux/sound/firewire/oxfw/oxfw-stream.c

832 lines
18 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* oxfw_stream.c - a part of driver for OXFW970/971 based devices
*
* Copyright (c) 2014 Takashi Sakamoto
*/
#include "oxfw.h"
#include <linux/delay.h>
#define AVC_GENERIC_FRAME_MAXIMUM_BYTES 512
#define CALLBACK_TIMEOUT 200
/*
* According to datasheet of Oxford Semiconductor:
* OXFW970: 32.0/44.1/48.0/96.0 Khz, 8 audio channels I/O
* OXFW971: 32.0/44.1/48.0/88.2/96.0/192.0 kHz, 16 audio channels I/O, MIDI I/O
*/
static const unsigned int oxfw_rate_table[] = {
[0] = 32000,
[1] = 44100,
[2] = 48000,
[3] = 88200,
[4] = 96000,
[5] = 192000,
};
/*
* See Table 5.7 Sampling frequency for Multi-bit Audio
* in AV/C Stream Format Information Specification 1.1 (Apr 2005, 1394TA)
*/
static const unsigned int avc_stream_rate_table[] = {
[0] = 0x02,
[1] = 0x03,
[2] = 0x04,
[3] = 0x0a,
[4] = 0x05,
[5] = 0x07,
};
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
static int set_rate(struct snd_oxfw *oxfw, unsigned int rate)
{
int err;
err = avc_general_set_sig_fmt(oxfw->unit, rate,
AVC_GENERAL_PLUG_DIR_IN, 0);
if (err < 0)
goto end;
if (oxfw->has_output)
err = avc_general_set_sig_fmt(oxfw->unit, rate,
AVC_GENERAL_PLUG_DIR_OUT, 0);
end:
return err;
}
static int set_stream_format(struct snd_oxfw *oxfw, struct amdtp_stream *s,
unsigned int rate, unsigned int pcm_channels)
{
u8 **formats;
struct snd_oxfw_stream_formation formation;
enum avc_general_plug_dir dir;
unsigned int len;
int i, err;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
if (s == &oxfw->tx_stream) {
formats = oxfw->tx_stream_formats;
dir = AVC_GENERAL_PLUG_DIR_OUT;
} else {
formats = oxfw->rx_stream_formats;
dir = AVC_GENERAL_PLUG_DIR_IN;
}
/* Seek stream format for requirements. */
for (i = 0; i < SND_OXFW_STREAM_FORMAT_ENTRIES; i++) {
err = snd_oxfw_stream_parse_format(formats[i], &formation);
if (err < 0)
return err;
if ((formation.rate == rate) && (formation.pcm == pcm_channels))
break;
}
if (i == SND_OXFW_STREAM_FORMAT_ENTRIES)
return -EINVAL;
/* If assumed, just change rate. */
if (oxfw->assumed)
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
return set_rate(oxfw, rate);
/* Calculate format length. */
len = 5 + formats[i][4] * 2;
err = avc_stream_set_format(oxfw->unit, dir, 0, formats[i], len);
if (err < 0)
return err;
/* Some requests just after changing format causes freezing. */
msleep(100);
return 0;
}
static int start_stream(struct snd_oxfw *oxfw, struct amdtp_stream *stream)
{
struct cmp_connection *conn;
int err;
if (stream == &oxfw->rx_stream)
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
conn = &oxfw->in_conn;
else
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
conn = &oxfw->out_conn;
err = cmp_connection_establish(conn);
if (err < 0)
return err;
err = amdtp_domain_add_stream(&oxfw->domain, stream,
conn->resources.channel, conn->speed);
if (err < 0) {
cmp_connection_break(conn);
return err;
}
return 0;
}
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
static int check_connection_used_by_others(struct snd_oxfw *oxfw,
struct amdtp_stream *stream)
{
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
struct cmp_connection *conn;
bool used;
int err;
if (stream == &oxfw->tx_stream)
conn = &oxfw->out_conn;
else
conn = &oxfw->in_conn;
err = cmp_connection_check_used(conn, &used);
if ((err >= 0) && used && !amdtp_stream_running(stream)) {
dev_err(&oxfw->unit->device,
"Connection established by others: %cPCR[%d]\n",
(conn->direction == CMP_OUTPUT) ? 'o' : 'i',
conn->pcr_index);
err = -EBUSY;
}
return err;
}
static int init_stream(struct snd_oxfw *oxfw, struct amdtp_stream *stream)
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
{
struct cmp_connection *conn;
enum cmp_direction c_dir;
enum amdtp_stream_direction s_dir;
int err;
if (stream == &oxfw->tx_stream) {
conn = &oxfw->out_conn;
c_dir = CMP_OUTPUT;
s_dir = AMDTP_IN_STREAM;
} else {
conn = &oxfw->in_conn;
c_dir = CMP_INPUT;
s_dir = AMDTP_OUT_STREAM;
}
err = cmp_connection_init(conn, oxfw->unit, c_dir, 0);
if (err < 0)
return err;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
err = amdtp_am824_init(stream, oxfw->unit, s_dir, CIP_NONBLOCKING);
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
if (err < 0) {
cmp_connection_destroy(conn);
return err;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
}
/*
* OXFW starts to transmit packets with non-zero dbc.
* OXFW postpone transferring packets till handling any asynchronous
* packets. As a result, next isochronous packet includes more data
* blocks than IEC 61883-6 defines.
*/
ALSA: oxfw: add Mackie Onyx Satellite quirk to inform wrong value in 'dbs' field in tx CIP header Mackie Onyx Satellite has two usage; standalone and with base station. These two modes has different stream formats. In standalone mode, rx data block includes 2 Multi Bit Linear Audio (MBLA) data channels and tx data block includes 2. With base station, they're 6 and 2. Although, with base station, the actual tx packet include wrong value in 'dbs' field in its CIP header. This quirk causes packet streaming layer to detect packet discontinuity and to stop PCM substream. This is a response of 'single' subfunction 'extended stream format information' command in AV/C Stream Format Information Specification 1.1. It means that a data block in tx packets includes 2 MBLA data channels. $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffffffff response: 000: 0c ff bf c0 01 00 00 00 ff 00 90 40 03 02 01 02 response: 010: 06 Current OXFW driver parses the response and detects stream formats correctly. $ cat /proc/asound/card1/firewire/formation ... Output Stream from device: Rate PCM MIDI * 48000 2 0 44100 2 0 88200 2 0 96000 2 0 On the other hand, in actual tx CIP, the 'dbs' field has 6. But the number of quadlets in CIP payload is not multiple of 6. Seeing the subtraction of two successive payload quadlets, it should be multiple of 2. payload CIP CIP quadlets header0 header1 24 00060052 9002ffff 24 0006005e 9002ffff 26 0006006a 9002ffff 24 00060077 9002ffff 24 00060083 9002ffff 26 0006008f 9002ffff 24 0006009c 9002ffff 24 000600a8 9002ffff 26 000600b4 9002ffff 24 000600c1 9002ffff This commit adds support for this quirk to OXFW driver, by using CIP_WRONG_DBS flag. When this flag is set, packet streaming layer uses the value of its 'data_block_quadlets' member instead of the value in CIP header. This value is already set by OXFW driver and no discontinuity is detected. Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-09-20 20:18:55 +08:00
if (stream == &oxfw->tx_stream) {
oxfw->tx_stream.flags |= CIP_JUMBO_PAYLOAD;
ALSA: oxfw: add Mackie Onyx Satellite quirk to inform wrong value in 'dbs' field in tx CIP header Mackie Onyx Satellite has two usage; standalone and with base station. These two modes has different stream formats. In standalone mode, rx data block includes 2 Multi Bit Linear Audio (MBLA) data channels and tx data block includes 2. With base station, they're 6 and 2. Although, with base station, the actual tx packet include wrong value in 'dbs' field in its CIP header. This quirk causes packet streaming layer to detect packet discontinuity and to stop PCM substream. This is a response of 'single' subfunction 'extended stream format information' command in AV/C Stream Format Information Specification 1.1. It means that a data block in tx packets includes 2 MBLA data channels. $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffffffff response: 000: 0c ff bf c0 01 00 00 00 ff 00 90 40 03 02 01 02 response: 010: 06 Current OXFW driver parses the response and detects stream formats correctly. $ cat /proc/asound/card1/firewire/formation ... Output Stream from device: Rate PCM MIDI * 48000 2 0 44100 2 0 88200 2 0 96000 2 0 On the other hand, in actual tx CIP, the 'dbs' field has 6. But the number of quadlets in CIP payload is not multiple of 6. Seeing the subtraction of two successive payload quadlets, it should be multiple of 2. payload CIP CIP quadlets header0 header1 24 00060052 9002ffff 24 0006005e 9002ffff 26 0006006a 9002ffff 24 00060077 9002ffff 24 00060083 9002ffff 26 0006008f 9002ffff 24 0006009c 9002ffff 24 000600a8 9002ffff 26 000600b4 9002ffff 24 000600c1 9002ffff This commit adds support for this quirk to OXFW driver, by using CIP_WRONG_DBS flag. When this flag is set, packet streaming layer uses the value of its 'data_block_quadlets' member instead of the value in CIP header. This value is already set by OXFW driver and no discontinuity is detected. Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2015-09-20 20:18:55 +08:00
if (oxfw->wrong_dbs)
oxfw->tx_stream.flags |= CIP_WRONG_DBS;
}
return 0;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
}
static int keep_resources(struct snd_oxfw *oxfw, struct amdtp_stream *stream)
{
enum avc_general_plug_dir dir;
u8 **formats;
struct snd_oxfw_stream_formation formation;
struct cmp_connection *conn;
int i;
int err;
if (stream == &oxfw->rx_stream) {
dir = AVC_GENERAL_PLUG_DIR_IN;
formats = oxfw->rx_stream_formats;
conn = &oxfw->in_conn;
} else {
dir = AVC_GENERAL_PLUG_DIR_OUT;
formats = oxfw->tx_stream_formats;
conn = &oxfw->out_conn;
}
err = snd_oxfw_stream_get_current_formation(oxfw, dir, &formation);
if (err < 0)
return err;
for (i = 0; i < SND_OXFW_STREAM_FORMAT_ENTRIES; i++) {
struct snd_oxfw_stream_formation fmt;
if (formats[i] == NULL)
break;
err = snd_oxfw_stream_parse_format(formats[i], &fmt);
if (err < 0)
return err;
if (fmt.rate == formation.rate && fmt.pcm == formation.pcm &&
fmt.midi == formation.midi)
break;
}
if (i == SND_OXFW_STREAM_FORMAT_ENTRIES)
return -EINVAL;
// The stream should have one pcm channels at least.
if (formation.pcm == 0)
return -EINVAL;
err = amdtp_am824_set_parameters(stream, formation.rate, formation.pcm,
formation.midi * 8, false);
if (err < 0)
return err;
return cmp_connection_reserve(conn, amdtp_stream_get_max_payload(stream));
}
int snd_oxfw_stream_reserve_duplex(struct snd_oxfw *oxfw,
struct amdtp_stream *stream,
unsigned int rate, unsigned int pcm_channels,
unsigned int frames_per_period,
unsigned int frames_per_buffer)
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
{
struct snd_oxfw_stream_formation formation;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
enum avc_general_plug_dir dir;
int err;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
// Considering JACK/FFADO streaming:
// TODO: This can be removed hwdep functionality becomes popular.
err = check_connection_used_by_others(oxfw, &oxfw->rx_stream);
if (err < 0)
return err;
if (oxfw->has_output) {
err = check_connection_used_by_others(oxfw, &oxfw->tx_stream);
if (err < 0)
return err;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
}
if (stream == &oxfw->tx_stream)
dir = AVC_GENERAL_PLUG_DIR_OUT;
else
dir = AVC_GENERAL_PLUG_DIR_IN;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
err = snd_oxfw_stream_get_current_formation(oxfw, dir, &formation);
if (err < 0)
return err;
if (rate == 0) {
rate = formation.rate;
pcm_channels = formation.pcm;
}
if (formation.rate != rate || formation.pcm != pcm_channels) {
amdtp_domain_stop(&oxfw->domain);
cmp_connection_break(&oxfw->in_conn);
cmp_connection_release(&oxfw->in_conn);
if (oxfw->has_output) {
cmp_connection_break(&oxfw->out_conn);
cmp_connection_release(&oxfw->out_conn);
}
}
if (oxfw->substreams_count == 0 ||
formation.rate != rate || formation.pcm != pcm_channels) {
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
err = set_stream_format(oxfw, stream, rate, pcm_channels);
if (err < 0) {
dev_err(&oxfw->unit->device,
"fail to set stream format: %d\n", err);
return err;
}
err = keep_resources(oxfw, &oxfw->rx_stream);
if (err < 0)
return err;
if (oxfw->has_output) {
err = keep_resources(oxfw, &oxfw->tx_stream);
if (err < 0) {
cmp_connection_release(&oxfw->in_conn);
return err;
}
}
err = amdtp_domain_set_events_per_period(&oxfw->domain,
frames_per_period, frames_per_buffer);
if (err < 0) {
cmp_connection_release(&oxfw->in_conn);
if (oxfw->has_output)
cmp_connection_release(&oxfw->out_conn);
return err;
}
}
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
return 0;
}
int snd_oxfw_stream_start_duplex(struct snd_oxfw *oxfw)
{
int err;
if (oxfw->substreams_count == 0)
return -EIO;
if (amdtp_streaming_error(&oxfw->rx_stream) ||
amdtp_streaming_error(&oxfw->tx_stream)) {
amdtp_domain_stop(&oxfw->domain);
cmp_connection_break(&oxfw->in_conn);
if (oxfw->has_output)
cmp_connection_break(&oxfw->out_conn);
}
if (!amdtp_stream_running(&oxfw->rx_stream)) {
err = start_stream(oxfw, &oxfw->rx_stream);
if (err < 0) {
dev_err(&oxfw->unit->device,
"fail to prepare rx stream: %d\n", err);
goto error;
}
if (oxfw->has_output &&
!amdtp_stream_running(&oxfw->tx_stream)) {
err = start_stream(oxfw, &oxfw->tx_stream);
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
if (err < 0) {
dev_err(&oxfw->unit->device,
"fail to prepare tx stream: %d\n", err);
goto error;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
}
}
err = amdtp_domain_start(&oxfw->domain, 0);
if (err < 0)
goto error;
// Wait first packet.
if (!amdtp_stream_wait_callback(&oxfw->rx_stream,
CALLBACK_TIMEOUT)) {
err = -ETIMEDOUT;
goto error;
}
if (oxfw->has_output) {
if (!amdtp_stream_wait_callback(&oxfw->tx_stream,
CALLBACK_TIMEOUT)) {
err = -ETIMEDOUT;
goto error;
}
}
}
return 0;
error:
amdtp_domain_stop(&oxfw->domain);
cmp_connection_break(&oxfw->in_conn);
if (oxfw->has_output)
cmp_connection_break(&oxfw->out_conn);
return err;
}
void snd_oxfw_stream_stop_duplex(struct snd_oxfw *oxfw)
{
if (oxfw->substreams_count == 0) {
amdtp_domain_stop(&oxfw->domain);
cmp_connection_break(&oxfw->in_conn);
cmp_connection_release(&oxfw->in_conn);
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
if (oxfw->has_output) {
cmp_connection_break(&oxfw->out_conn);
cmp_connection_release(&oxfw->out_conn);
}
}
}
static void destroy_stream(struct snd_oxfw *oxfw, struct amdtp_stream *stream)
{
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
struct cmp_connection *conn;
if (stream == &oxfw->tx_stream)
conn = &oxfw->out_conn;
else
conn = &oxfw->in_conn;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
amdtp_stream_destroy(stream);
cmp_connection_destroy(conn);
}
int snd_oxfw_stream_init_duplex(struct snd_oxfw *oxfw)
{
int err;
err = init_stream(oxfw, &oxfw->rx_stream);
if (err < 0)
return err;
if (oxfw->has_output) {
err = init_stream(oxfw, &oxfw->tx_stream);
if (err < 0) {
destroy_stream(oxfw, &oxfw->rx_stream);
return err;
}
}
err = amdtp_domain_init(&oxfw->domain);
if (err < 0) {
destroy_stream(oxfw, &oxfw->rx_stream);
if (oxfw->has_output)
destroy_stream(oxfw, &oxfw->tx_stream);
}
return err;
}
// This function should be called before starting the stream or after stopping
// the streams.
void snd_oxfw_stream_destroy_duplex(struct snd_oxfw *oxfw)
{
amdtp_domain_destroy(&oxfw->domain);
destroy_stream(oxfw, &oxfw->rx_stream);
if (oxfw->has_output)
destroy_stream(oxfw, &oxfw->tx_stream);
}
void snd_oxfw_stream_update_duplex(struct snd_oxfw *oxfw)
{
amdtp_domain_stop(&oxfw->domain);
cmp_connection_break(&oxfw->in_conn);
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
amdtp_stream_pcm_abort(&oxfw->rx_stream);
if (oxfw->has_output) {
cmp_connection_break(&oxfw->out_conn);
amdtp_stream_pcm_abort(&oxfw->tx_stream);
}
}
int snd_oxfw_stream_get_current_formation(struct snd_oxfw *oxfw,
enum avc_general_plug_dir dir,
struct snd_oxfw_stream_formation *formation)
{
u8 *format;
unsigned int len;
int err;
len = AVC_GENERIC_FRAME_MAXIMUM_BYTES;
format = kmalloc(len, GFP_KERNEL);
if (format == NULL)
return -ENOMEM;
err = avc_stream_get_format_single(oxfw->unit, dir, 0, format, &len);
if (err < 0)
goto end;
if (len < 3) {
err = -EIO;
goto end;
}
err = snd_oxfw_stream_parse_format(format, formation);
end:
kfree(format);
return err;
}
/*
* See Table 6.16 - AM824 Stream Format
* Figure 6.19 - format_information field for AM824 Compound
* in AV/C Stream Format Information Specification 1.1 (Apr 2005, 1394TA)
* Also 'Clause 12 AM824 sequence adaption layers' in IEC 61883-6:2005
*/
int snd_oxfw_stream_parse_format(u8 *format,
struct snd_oxfw_stream_formation *formation)
{
unsigned int i, e, channels, type;
memset(formation, 0, sizeof(struct snd_oxfw_stream_formation));
/*
* this module can support a hierarchy combination that:
* Root: Audio and Music (0x90)
* Level 1: AM824 Compound (0x40)
*/
if ((format[0] != 0x90) || (format[1] != 0x40))
return -ENXIO;
/* check the sampling rate */
for (i = 0; i < ARRAY_SIZE(avc_stream_rate_table); i++) {
if (format[2] == avc_stream_rate_table[i])
break;
}
if (i == ARRAY_SIZE(avc_stream_rate_table))
return -ENXIO;
formation->rate = oxfw_rate_table[i];
for (e = 0; e < format[4]; e++) {
channels = format[5 + e * 2];
type = format[6 + e * 2];
switch (type) {
/* IEC 60958 Conformant, currently handled as MBLA */
case 0x00:
/* Multi Bit Linear Audio (Raw) */
case 0x06:
formation->pcm += channels;
break;
/* MIDI Conformant */
case 0x0d:
formation->midi = channels;
break;
/* IEC 61937-3 to 7 */
case 0x01:
case 0x02:
case 0x03:
case 0x04:
case 0x05:
/* Multi Bit Linear Audio */
case 0x07: /* DVD-Audio */
case 0x0c: /* High Precision */
/* One Bit Audio */
case 0x08: /* (Plain) Raw */
case 0x09: /* (Plain) SACD */
case 0x0a: /* (Encoded) Raw */
case 0x0b: /* (Encoded) SACD */
/* SMPTE Time-Code conformant */
case 0x0e:
/* Sample Count */
case 0x0f:
/* Anciliary Data */
case 0x10:
/* Synchronization Stream (Stereo Raw audio) */
case 0x40:
/* Don't care */
case 0xff:
default:
return -ENXIO; /* not supported */
}
}
if (formation->pcm > AM824_MAX_CHANNELS_FOR_PCM ||
formation->midi > AM824_MAX_CHANNELS_FOR_MIDI)
return -ENXIO;
return 0;
}
static int
assume_stream_formats(struct snd_oxfw *oxfw, enum avc_general_plug_dir dir,
unsigned int pid, u8 *buf, unsigned int *len,
u8 **formats)
{
struct snd_oxfw_stream_formation formation;
unsigned int i, eid;
int err;
/* get format at current sampling rate */
err = avc_stream_get_format_single(oxfw->unit, dir, pid, buf, len);
if (err < 0) {
dev_err(&oxfw->unit->device,
"fail to get current stream format for isoc %s plug %d:%d\n",
(dir == AVC_GENERAL_PLUG_DIR_IN) ? "in" : "out",
pid, err);
goto end;
}
/* parse and set stream format */
eid = 0;
err = snd_oxfw_stream_parse_format(buf, &formation);
if (err < 0)
goto end;
formats[eid] = devm_kmemdup(&oxfw->card->card_dev, buf, *len,
GFP_KERNEL);
if (!formats[eid]) {
err = -ENOMEM;
goto end;
}
/* apply the format for each available sampling rate */
for (i = 0; i < ARRAY_SIZE(oxfw_rate_table); i++) {
if (formation.rate == oxfw_rate_table[i])
continue;
err = avc_general_inquiry_sig_fmt(oxfw->unit,
oxfw_rate_table[i],
dir, pid);
if (err < 0)
continue;
eid++;
formats[eid] = devm_kmemdup(&oxfw->card->card_dev, buf, *len,
GFP_KERNEL);
if (formats[eid] == NULL) {
err = -ENOMEM;
goto end;
}
formats[eid][2] = avc_stream_rate_table[i];
}
err = 0;
oxfw->assumed = true;
end:
return err;
}
static int fill_stream_formats(struct snd_oxfw *oxfw,
enum avc_general_plug_dir dir,
unsigned short pid)
{
u8 *buf, **formats;
unsigned int len, eid = 0;
struct snd_oxfw_stream_formation dummy;
int err;
buf = kmalloc(AVC_GENERIC_FRAME_MAXIMUM_BYTES, GFP_KERNEL);
if (buf == NULL)
return -ENOMEM;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
if (dir == AVC_GENERAL_PLUG_DIR_OUT)
formats = oxfw->tx_stream_formats;
else
formats = oxfw->rx_stream_formats;
/* get first entry */
len = AVC_GENERIC_FRAME_MAXIMUM_BYTES;
err = avc_stream_get_format_list(oxfw->unit, dir, 0, buf, &len, 0);
if (err == -ENXIO) {
/* LIST subfunction is not implemented */
len = AVC_GENERIC_FRAME_MAXIMUM_BYTES;
err = assume_stream_formats(oxfw, dir, pid, buf, &len,
formats);
goto end;
} else if (err < 0) {
dev_err(&oxfw->unit->device,
"fail to get stream format %d for isoc %s plug %d:%d\n",
eid, (dir == AVC_GENERAL_PLUG_DIR_IN) ? "in" : "out",
pid, err);
goto end;
}
/* LIST subfunction is implemented */
while (eid < SND_OXFW_STREAM_FORMAT_ENTRIES) {
/* The format is too short. */
if (len < 3) {
err = -EIO;
break;
}
/* parse and set stream format */
err = snd_oxfw_stream_parse_format(buf, &dummy);
if (err < 0)
break;
formats[eid] = devm_kmemdup(&oxfw->card->card_dev, buf, len,
GFP_KERNEL);
if (!formats[eid]) {
err = -ENOMEM;
break;
}
/* get next entry */
len = AVC_GENERIC_FRAME_MAXIMUM_BYTES;
err = avc_stream_get_format_list(oxfw->unit, dir, 0,
buf, &len, ++eid);
/* No entries remained. */
if (err == -EINVAL) {
err = 0;
break;
} else if (err < 0) {
dev_err(&oxfw->unit->device,
"fail to get stream format %d for isoc %s plug %d:%d\n",
eid, (dir == AVC_GENERAL_PLUG_DIR_IN) ? "in" :
"out",
pid, err);
break;
}
}
end:
kfree(buf);
return err;
}
int snd_oxfw_stream_discover(struct snd_oxfw *oxfw)
{
u8 plugs[AVC_PLUG_INFO_BUF_BYTES];
struct snd_oxfw_stream_formation formation;
u8 *format;
unsigned int i;
int err;
/* the number of plugs for isoc in/out, ext in/out */
err = avc_general_get_plug_info(oxfw->unit, 0x1f, 0x07, 0x00, plugs);
if (err < 0) {
dev_err(&oxfw->unit->device,
"fail to get info for isoc/external in/out plugs: %d\n",
err);
goto end;
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
} else if ((plugs[0] == 0) && (plugs[1] == 0)) {
err = -ENXIO;
goto end;
}
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
/* use oPCR[0] if exists */
if (plugs[1] > 0) {
err = fill_stream_formats(oxfw, AVC_GENERAL_PLUG_DIR_OUT, 0);
ALSA: oxfw: fix for Stanton SCS.1d Stanton SCS.1d uses Oxford Semiconductor FW 971 ASIC (FW971) for communication. Although the unit is bound to ALSA oxfw driver, the instance of sound card can not be added due to its quirk of plug information. This bug was added when snd-scs1x is merged into snd-oxfw at commit 9e2004f9cedf ("ALSA: oxfw: obsolete scs1x module"). This commit fixes the driver for the quirk. In cases that the unit returns NOT IMPLEMENTED for some AV/C commands, the sound card is added without any PCM/MIDI interfaces for packet streaming. For SCS.1d, model dependent operation adds MIDI interface and applications can use it to operate according to HSS1394 protocol from reverse-engineering work by Sean M. Pappalardo. Plug Control Register (PCR) has information that the unit has a pair of plugs for isochronous communication: (oMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000900 result: 80ff0001 (iMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000980 result: 80ff0001 AV/C PLUG INFO also returns information that the unit has a pair of plugs for isochronous communication. (AV/C PLUG INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff0200ffffffff response: 000: 0c ff 02 00 01 01 02 02 However, AV/C PLUG SIGNAL INFO command is rejected for both plugs. (AV/C OUTPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1800ffffffff response: 000: 0a ff 18 00 ff ff ff ff (AV/C INPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1900ffffffff response: 000: 0a ff 19 00 ff ff ff ff Furthermore, AV/C EXTENDED STREAM FORMAT INFO is not implemented. (AV/C EXTENDED STREAM FORMAT INFO list subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc000000000ffff00ff response: 000: 08 ff bf c0 00 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO list subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffff00ff response: 000: 08 ff bf c0 01 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc100000000ffffffff response: 000: 08 ff bf c1 00 00 00 00 ff ff ff ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc101000000ffffffff response: 000: 08 ff bf c1 01 00 00 00 ff ff ff ff Reference: https://mailman.alsa-project.org/pipermail/alsa-devel/2012-May/052264.html Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Link: https://lore.kernel.org/r/20200113073418.24622-4-o-takashi@sakamocchi.jp Signed-off-by: Takashi Iwai <tiwai@suse.de>
2020-01-13 15:34:18 +08:00
if (err < 0) {
if (err != -ENXIO)
return err;
ALSA: oxfw: fix for Stanton SCS.1d Stanton SCS.1d uses Oxford Semiconductor FW 971 ASIC (FW971) for communication. Although the unit is bound to ALSA oxfw driver, the instance of sound card can not be added due to its quirk of plug information. This bug was added when snd-scs1x is merged into snd-oxfw at commit 9e2004f9cedf ("ALSA: oxfw: obsolete scs1x module"). This commit fixes the driver for the quirk. In cases that the unit returns NOT IMPLEMENTED for some AV/C commands, the sound card is added without any PCM/MIDI interfaces for packet streaming. For SCS.1d, model dependent operation adds MIDI interface and applications can use it to operate according to HSS1394 protocol from reverse-engineering work by Sean M. Pappalardo. Plug Control Register (PCR) has information that the unit has a pair of plugs for isochronous communication: (oMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000900 result: 80ff0001 (iMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000980 result: 80ff0001 AV/C PLUG INFO also returns information that the unit has a pair of plugs for isochronous communication. (AV/C PLUG INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff0200ffffffff response: 000: 0c ff 02 00 01 01 02 02 However, AV/C PLUG SIGNAL INFO command is rejected for both plugs. (AV/C OUTPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1800ffffffff response: 000: 0a ff 18 00 ff ff ff ff (AV/C INPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1900ffffffff response: 000: 0a ff 19 00 ff ff ff ff Furthermore, AV/C EXTENDED STREAM FORMAT INFO is not implemented. (AV/C EXTENDED STREAM FORMAT INFO list subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc000000000ffff00ff response: 000: 08 ff bf c0 00 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO list subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffff00ff response: 000: 08 ff bf c0 01 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc100000000ffffffff response: 000: 08 ff bf c1 00 00 00 00 ff ff ff ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc101000000ffffffff response: 000: 08 ff bf c1 01 00 00 00 ff ff ff ff Reference: https://mailman.alsa-project.org/pipermail/alsa-devel/2012-May/052264.html Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Link: https://lore.kernel.org/r/20200113073418.24622-4-o-takashi@sakamocchi.jp Signed-off-by: Takashi Iwai <tiwai@suse.de>
2020-01-13 15:34:18 +08:00
// The oPCR is not available for isoc communication.
err = 0;
} else {
for (i = 0; i < SND_OXFW_STREAM_FORMAT_ENTRIES; i++) {
format = oxfw->tx_stream_formats[i];
if (format == NULL)
continue;
err = snd_oxfw_stream_parse_format(format,
&formation);
if (err < 0)
continue;
/* Add one MIDI port. */
if (formation.midi > 0)
oxfw->midi_input_ports = 1;
}
ALSA: oxfw: fix for Stanton SCS.1d Stanton SCS.1d uses Oxford Semiconductor FW 971 ASIC (FW971) for communication. Although the unit is bound to ALSA oxfw driver, the instance of sound card can not be added due to its quirk of plug information. This bug was added when snd-scs1x is merged into snd-oxfw at commit 9e2004f9cedf ("ALSA: oxfw: obsolete scs1x module"). This commit fixes the driver for the quirk. In cases that the unit returns NOT IMPLEMENTED for some AV/C commands, the sound card is added without any PCM/MIDI interfaces for packet streaming. For SCS.1d, model dependent operation adds MIDI interface and applications can use it to operate according to HSS1394 protocol from reverse-engineering work by Sean M. Pappalardo. Plug Control Register (PCR) has information that the unit has a pair of plugs for isochronous communication: (oMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000900 result: 80ff0001 (iMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000980 result: 80ff0001 AV/C PLUG INFO also returns information that the unit has a pair of plugs for isochronous communication. (AV/C PLUG INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff0200ffffffff response: 000: 0c ff 02 00 01 01 02 02 However, AV/C PLUG SIGNAL INFO command is rejected for both plugs. (AV/C OUTPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1800ffffffff response: 000: 0a ff 18 00 ff ff ff ff (AV/C INPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1900ffffffff response: 000: 0a ff 19 00 ff ff ff ff Furthermore, AV/C EXTENDED STREAM FORMAT INFO is not implemented. (AV/C EXTENDED STREAM FORMAT INFO list subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc000000000ffff00ff response: 000: 08 ff bf c0 00 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO list subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffff00ff response: 000: 08 ff bf c0 01 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc100000000ffffffff response: 000: 08 ff bf c1 00 00 00 00 ff ff ff ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc101000000ffffffff response: 000: 08 ff bf c1 01 00 00 00 ff ff ff ff Reference: https://mailman.alsa-project.org/pipermail/alsa-devel/2012-May/052264.html Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Link: https://lore.kernel.org/r/20200113073418.24622-4-o-takashi@sakamocchi.jp Signed-off-by: Takashi Iwai <tiwai@suse.de>
2020-01-13 15:34:18 +08:00
oxfw->has_output = true;
}
ALSA: oxfw: Add support AMDTP in-stream Previous commit adds support for some devices which can capture PCM samples. These devices transmit AMDTP stream in non-blocking mode. This commit adds functionality to handle AMDTP incoming stream. OXFW seems to have two quirks: - Transmits packets with non-zero dbc in its beginning - Transmits packets with wrong values in syt field For the first quirk, this commit adds CIP_SKIP_INIT_DBC_CHECK flag for incoming stream to skip first check of dbc. For the second quirk, this commit doesn't add duplex stream which Fireworks/BeBoB drivers use. So OXFW driver generates syt value for outgoing stream. Here are examples of a sequence of packets transmitted by Behringer F-Control Audio 202. There are differences between sequences of syt value when OXFW driver transfers outgoing stream or not. When driver gives no outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 14 00020092 900103D1 39 12 00020098 900102FF 40 12 0002009D 9001027F 41 14 000200A2 90010396 42 14 000200A8 900102E8 43 12 000200AE 90010219 44 14 000200B3 90010331 45 12 000200B9 9001025F 46 14 000200BE 90010376 47 12 000200C4 900102A1 00 12 000200C9 9001023E 01 14 000200CE 90010358 02 12 000200D4 90010289 03 16 000200D9 900103A3 04 12 000200E0 900102DD 05 14 000200E5 900103F1 06 12 000200EB 90010335 07 12 000200F0 90010263 08 14 000200F5 9001037C 09 12 000200FB 900102AE When driver gives outgoing stream: Index Payload CIP_Header_0 CIP_Header_1 38 12 000200BD 900104A8 39 14 000200C2 900104A8 40 12 000200C8 900104AC 41 14 000200CD 900104A9 42 12 000200D3 900104B1 43 14 000200D8 900104A8 44 12 000200DE 900104AA 45 14 000200E3 900104A9 46 14 000200E9 900104AE 47 12 000200EF 900104A8 00 14 000200F4 900104AD 01 12 000200FA 900104A7 02 14 000200FF 900104A9 03 12 00020005 900104A9 04 14 0002000A 900104B1 05 12 00020010 900104AA 06 14 00020015 900104AD 07 12 0002001B 900104A7 08 14 00020020 900104AC 09 12 00020026 900104A7 Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Acked-by: Clemens Ladisch <clemens@ladisch.de> Signed-off-by: Takashi Iwai <tiwai@suse.de>
2014-12-08 23:10:46 +08:00
}
/* use iPCR[0] if exists */
if (plugs[0] > 0) {
err = fill_stream_formats(oxfw, AVC_GENERAL_PLUG_DIR_IN, 0);
ALSA: oxfw: fix for Stanton SCS.1d Stanton SCS.1d uses Oxford Semiconductor FW 971 ASIC (FW971) for communication. Although the unit is bound to ALSA oxfw driver, the instance of sound card can not be added due to its quirk of plug information. This bug was added when snd-scs1x is merged into snd-oxfw at commit 9e2004f9cedf ("ALSA: oxfw: obsolete scs1x module"). This commit fixes the driver for the quirk. In cases that the unit returns NOT IMPLEMENTED for some AV/C commands, the sound card is added without any PCM/MIDI interfaces for packet streaming. For SCS.1d, model dependent operation adds MIDI interface and applications can use it to operate according to HSS1394 protocol from reverse-engineering work by Sean M. Pappalardo. Plug Control Register (PCR) has information that the unit has a pair of plugs for isochronous communication: (oMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000900 result: 80ff0001 (iMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000980 result: 80ff0001 AV/C PLUG INFO also returns information that the unit has a pair of plugs for isochronous communication. (AV/C PLUG INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff0200ffffffff response: 000: 0c ff 02 00 01 01 02 02 However, AV/C PLUG SIGNAL INFO command is rejected for both plugs. (AV/C OUTPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1800ffffffff response: 000: 0a ff 18 00 ff ff ff ff (AV/C INPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1900ffffffff response: 000: 0a ff 19 00 ff ff ff ff Furthermore, AV/C EXTENDED STREAM FORMAT INFO is not implemented. (AV/C EXTENDED STREAM FORMAT INFO list subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc000000000ffff00ff response: 000: 08 ff bf c0 00 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO list subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffff00ff response: 000: 08 ff bf c0 01 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc100000000ffffffff response: 000: 08 ff bf c1 00 00 00 00 ff ff ff ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc101000000ffffffff response: 000: 08 ff bf c1 01 00 00 00 ff ff ff ff Reference: https://mailman.alsa-project.org/pipermail/alsa-devel/2012-May/052264.html Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Link: https://lore.kernel.org/r/20200113073418.24622-4-o-takashi@sakamocchi.jp Signed-off-by: Takashi Iwai <tiwai@suse.de>
2020-01-13 15:34:18 +08:00
if (err < 0) {
if (err != -ENXIO)
return err;
ALSA: oxfw: fix for Stanton SCS.1d Stanton SCS.1d uses Oxford Semiconductor FW 971 ASIC (FW971) for communication. Although the unit is bound to ALSA oxfw driver, the instance of sound card can not be added due to its quirk of plug information. This bug was added when snd-scs1x is merged into snd-oxfw at commit 9e2004f9cedf ("ALSA: oxfw: obsolete scs1x module"). This commit fixes the driver for the quirk. In cases that the unit returns NOT IMPLEMENTED for some AV/C commands, the sound card is added without any PCM/MIDI interfaces for packet streaming. For SCS.1d, model dependent operation adds MIDI interface and applications can use it to operate according to HSS1394 protocol from reverse-engineering work by Sean M. Pappalardo. Plug Control Register (PCR) has information that the unit has a pair of plugs for isochronous communication: (oMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000900 result: 80ff0001 (iMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000980 result: 80ff0001 AV/C PLUG INFO also returns information that the unit has a pair of plugs for isochronous communication. (AV/C PLUG INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff0200ffffffff response: 000: 0c ff 02 00 01 01 02 02 However, AV/C PLUG SIGNAL INFO command is rejected for both plugs. (AV/C OUTPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1800ffffffff response: 000: 0a ff 18 00 ff ff ff ff (AV/C INPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1900ffffffff response: 000: 0a ff 19 00 ff ff ff ff Furthermore, AV/C EXTENDED STREAM FORMAT INFO is not implemented. (AV/C EXTENDED STREAM FORMAT INFO list subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc000000000ffff00ff response: 000: 08 ff bf c0 00 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO list subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffff00ff response: 000: 08 ff bf c0 01 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc100000000ffffffff response: 000: 08 ff bf c1 00 00 00 00 ff ff ff ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc101000000ffffffff response: 000: 08 ff bf c1 01 00 00 00 ff ff ff ff Reference: https://mailman.alsa-project.org/pipermail/alsa-devel/2012-May/052264.html Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Link: https://lore.kernel.org/r/20200113073418.24622-4-o-takashi@sakamocchi.jp Signed-off-by: Takashi Iwai <tiwai@suse.de>
2020-01-13 15:34:18 +08:00
// The iPCR is not available for isoc communication.
err = 0;
} else {
for (i = 0; i < SND_OXFW_STREAM_FORMAT_ENTRIES; i++) {
format = oxfw->rx_stream_formats[i];
if (format == NULL)
continue;
err = snd_oxfw_stream_parse_format(format,
&formation);
if (err < 0)
continue;
/* Add one MIDI port. */
if (formation.midi > 0)
oxfw->midi_output_ports = 1;
}
ALSA: oxfw: fix for Stanton SCS.1d Stanton SCS.1d uses Oxford Semiconductor FW 971 ASIC (FW971) for communication. Although the unit is bound to ALSA oxfw driver, the instance of sound card can not be added due to its quirk of plug information. This bug was added when snd-scs1x is merged into snd-oxfw at commit 9e2004f9cedf ("ALSA: oxfw: obsolete scs1x module"). This commit fixes the driver for the quirk. In cases that the unit returns NOT IMPLEMENTED for some AV/C commands, the sound card is added without any PCM/MIDI interfaces for packet streaming. For SCS.1d, model dependent operation adds MIDI interface and applications can use it to operate according to HSS1394 protocol from reverse-engineering work by Sean M. Pappalardo. Plug Control Register (PCR) has information that the unit has a pair of plugs for isochronous communication: (oMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000900 result: 80ff0001 (iMPR) $ ./firewire-request /dev/fw1 read 0xfffff0000980 result: 80ff0001 AV/C PLUG INFO also returns information that the unit has a pair of plugs for isochronous communication. (AV/C PLUG INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff0200ffffffff response: 000: 0c ff 02 00 01 01 02 02 However, AV/C PLUG SIGNAL INFO command is rejected for both plugs. (AV/C OUTPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1800ffffffff response: 000: 0a ff 18 00 ff ff ff ff (AV/C INPUT PLUG SIGNAL INFO command) $ ./firewire-request /dev/fw1 fcp 0x01ff1900ffffffff response: 000: 0a ff 19 00 ff ff ff ff Furthermore, AV/C EXTENDED STREAM FORMAT INFO is not implemented. (AV/C EXTENDED STREAM FORMAT INFO list subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc000000000ffff00ff response: 000: 08 ff bf c0 00 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO list subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc001000000ffff00ff response: 000: 08 ff bf c0 01 00 00 00 ff ff 00 ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for input plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc100000000ffffffff response: 000: 08 ff bf c1 00 00 00 00 ff ff ff ff (AV/C EXTENDED STREAM FORMAT INFO single subfunction for output plug) $ ./firewire-request /dev/fw1 fcp 0x01ffbfc101000000ffffffff response: 000: 08 ff bf c1 01 00 00 00 ff ff ff ff Reference: https://mailman.alsa-project.org/pipermail/alsa-devel/2012-May/052264.html Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp> Link: https://lore.kernel.org/r/20200113073418.24622-4-o-takashi@sakamocchi.jp Signed-off-by: Takashi Iwai <tiwai@suse.de>
2020-01-13 15:34:18 +08:00
oxfw->has_input = true;
}
}
end:
return err;
}
void snd_oxfw_stream_lock_changed(struct snd_oxfw *oxfw)
{
oxfw->dev_lock_changed = true;
wake_up(&oxfw->hwdep_wait);
}
int snd_oxfw_stream_lock_try(struct snd_oxfw *oxfw)
{
int err;
spin_lock_irq(&oxfw->lock);
/* user land lock this */
if (oxfw->dev_lock_count < 0) {
err = -EBUSY;
goto end;
}
/* this is the first time */
if (oxfw->dev_lock_count++ == 0)
snd_oxfw_stream_lock_changed(oxfw);
err = 0;
end:
spin_unlock_irq(&oxfw->lock);
return err;
}
void snd_oxfw_stream_lock_release(struct snd_oxfw *oxfw)
{
spin_lock_irq(&oxfw->lock);
if (WARN_ON(oxfw->dev_lock_count <= 0))
goto end;
if (--oxfw->dev_lock_count == 0)
snd_oxfw_stream_lock_changed(oxfw);
end:
spin_unlock_irq(&oxfw->lock);
}