[alsa-devel] [PATCH 0/3] ALSA: bebob: misc improvements for starting/stopping packet streaming
Hi,
This patchset improves model-specific issues in a point of protocol version supported by the model. The changes are relevant to packet streaming, thus it can brings some regressions.
For testers, I prepare for a remote branch to backport the latest codes to Linux v4.17 or later. I'm happy to receive any report. https://github.com/takaswie/snd-firewire-improve/tree/topic/bebob-connection...
Takashi Sakamoto (3): ALSA: bebob: expand sleep just after breaking connections for protocol version 1 ALSA: bebob: expand delay of start for IR context just for version 3 firmware ALSA: bebob: link the order of establishing connections and Syt-match clock mode
sound/firewire/bebob/bebob_stream.c | 76 ++++++++++++++--------------- 1 file changed, 38 insertions(+), 38 deletions(-)
As long as I investigated, some devices with BeBoB protocol version 1 can be freezed during several hundreds milliseconds after breaking connections. When accessing during the freezed time, any transaction is corrupted. In the worst case, the device is going to reboot.
I can see this issue in: * Roland FA-66 * M-Audio FireWire Solo
This commit expands sleep just after breaking connections to avoid the freezed time as much as possible. I note that the freeze/reboot behaviour is similar to below models: * Focusrite Saffire Pro 10 I/O * Focusrite Saffire Pro 26 I/O
The above models certainly reboot after breaking connections.
Signed-off-by: Takashi Sakamoto o-takashi@sakamocchi.jp --- sound/firewire/bebob/bebob_stream.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/sound/firewire/bebob/bebob_stream.c b/sound/firewire/bebob/bebob_stream.c index f7f0db5aa811..1b264d0d63eb 100644 --- a/sound/firewire/bebob/bebob_stream.c +++ b/sound/firewire/bebob/bebob_stream.c @@ -415,15 +415,16 @@ static int make_both_connections(struct snd_bebob *bebob) return 0; }
-static void -break_both_connections(struct snd_bebob *bebob) +static void break_both_connections(struct snd_bebob *bebob) { cmp_connection_break(&bebob->in_conn); cmp_connection_break(&bebob->out_conn);
- /* These models seems to be in transition state for a longer time. */ - if (bebob->maudio_special_quirk != NULL) - msleep(200); + // These models seem to be in transition state for a longer time. When + // accessing in the state, any transactions is corrupted. In the worst + // case, the device is going to reboot. + if (bebob->version < 2) + msleep(600); }
static int
On Fri, Nov 01, 2019 at 10:13:21PM +0900, Takashi Sakamoto wrote:
As long as I investigated, some devices with BeBoB protocol version 1 can be freezed during several hundreds milliseconds after breaking connections. When accessing during the freezed time, any transaction is corrupted. In the worst case, the device is going to reboot.
I can see this issue in:
- Roland FA-66
- M-Audio FireWire Solo
This commit expands sleep just after breaking connections to avoid the freezed time as much as possible. I note that the freeze/reboot behaviour is similar to below models:
- Focusrite Saffire Pro 10 I/O
- Focusrite Saffire Pro 26 I/O
The above models certainly reboot after breaking connections.
Signed-off-by: Takashi Sakamoto o-takashi@sakamocchi.jp
sound/firewire/bebob/bebob_stream.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/sound/firewire/bebob/bebob_stream.c b/sound/firewire/bebob/bebob_stream.c index f7f0db5aa811..1b264d0d63eb 100644 --- a/sound/firewire/bebob/bebob_stream.c +++ b/sound/firewire/bebob/bebob_stream.c @@ -415,15 +415,16 @@ static int make_both_connections(struct snd_bebob *bebob) return 0; }
-static void -break_both_connections(struct snd_bebob *bebob) +static void break_both_connections(struct snd_bebob *bebob) { cmp_connection_break(&bebob->in_conn); cmp_connection_break(&bebob->out_conn);
- /* These models seems to be in transition state for a longer time. */
- if (bebob->maudio_special_quirk != NULL)
msleep(200);
- // These models seem to be in transition state for a longer time. When
- // accessing in the state, any transactions is corrupted. In the worst
- // case, the device is going to reboot.
- if (bebob->version < 2)
msleep(600);}
static int
2.20.1
Alsa-devel mailing list Alsa-devel@alsa-project.org https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
This patch causes a reboot when I power on my Edirol FA-101. This is consistent behavior with kernels 5.5.1 - 5.5.11.
I've changed msleep(600) back to msleep(200) and rebuilt the module and I am once again able to power on and use the interface.
I am running Arch linux with 5.5.11-arch1-1 #1 SMP PREEMPT Sun, 22 Mar 2020 16:33:15 +0000 x86_64 GNU/Linux.
I had previously posted to the list on Feb 19, 2020 with the subject:
sytem reboots when initializing Edirol FA-101 firewire audio interface
Please forgive any lapses in reporting protocol as I generally don't work at the kernel level. Hopefully this more specific report will help fix the issue.
Thanks,
Steve
On Thu, 26 Mar 2020 22:47:22 +0100, Steve Morris wrote:
On Fri, Nov 01, 2019 at 10:13:21PM +0900, Takashi Sakamoto wrote:
As long as I investigated, some devices with BeBoB protocol version 1 can be freezed during several hundreds milliseconds after breaking connections. When accessing during the freezed time, any transaction is corrupted. In the worst case, the device is going to reboot.
I can see this issue in:
- Roland FA-66
- M-Audio FireWire Solo
This commit expands sleep just after breaking connections to avoid the freezed time as much as possible. I note that the freeze/reboot behaviour is similar to below models:
- Focusrite Saffire Pro 10 I/O
- Focusrite Saffire Pro 26 I/O
The above models certainly reboot after breaking connections.
Signed-off-by: Takashi Sakamoto o-takashi@sakamocchi.jp
sound/firewire/bebob/bebob_stream.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/sound/firewire/bebob/bebob_stream.c b/sound/firewire/bebob/bebob_stream.c index f7f0db5aa811..1b264d0d63eb 100644 --- a/sound/firewire/bebob/bebob_stream.c +++ b/sound/firewire/bebob/bebob_stream.c @@ -415,15 +415,16 @@ static int make_both_connections(struct snd_bebob *bebob) return 0; }
-static void -break_both_connections(struct snd_bebob *bebob) +static void break_both_connections(struct snd_bebob *bebob) { cmp_connection_break(&bebob->in_conn); cmp_connection_break(&bebob->out_conn);
- /* These models seems to be in transition state for a longer time. */
- if (bebob->maudio_special_quirk != NULL)
msleep(200);
- // These models seem to be in transition state for a longer time. When
- // accessing in the state, any transactions is corrupted. In the worst
- // case, the device is going to reboot.
- if (bebob->version < 2)
msleep(600);}
static int
2.20.1
Alsa-devel mailing list Alsa-devel@alsa-project.org https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
This patch causes a reboot when I power on my Edirol FA-101. This is consistent behavior with kernels 5.5.1 - 5.5.11.
I've changed msleep(600) back to msleep(200) and rebuilt the module and I am once again able to power on and use the interface.
I am running Arch linux with 5.5.11-arch1-1 #1 SMP PREEMPT Sun, 22 Mar 2020 16:33:15 +0000 x86_64 GNU/Linux.
I had previously posted to the list on Feb 19, 2020 with the subject:
sytem reboots when initializing Edirol FA-101 firewire audio interface
Please forgive any lapses in reporting protocol as I generally don't work at the kernel level. Hopefully this more specific report will help fix the issue.
This is weird. If just changing the msleep value from 600 to 200 "fixes" the problem, it means that something is already fundamentally broken...
Takashi
On Fri, Mar 27, 2020 at 09:40:34AM +0100, Takashi Iwai wrote:
On Thu, 26 Mar 2020 22:47:22 +0100, Steve Morris wrote:
On Fri, Nov 01, 2019 at 10:13:21PM +0900, Takashi Sakamoto wrote:
As long as I investigated, some devices with BeBoB protocol version 1 can be freezed during several hundreds milliseconds after breaking connections. When accessing during the freezed time, any transaction is corrupted. In the worst case, the device is going to reboot.
I can see this issue in:
- Roland FA-66
- M-Audio FireWire Solo
This commit expands sleep just after breaking connections to avoid the freezed time as much as possible. I note that the freeze/reboot behaviour is similar to below models:
- Focusrite Saffire Pro 10 I/O
- Focusrite Saffire Pro 26 I/O
The above models certainly reboot after breaking connections.
Signed-off-by: Takashi Sakamoto o-takashi@sakamocchi.jp
sound/firewire/bebob/bebob_stream.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/sound/firewire/bebob/bebob_stream.c b/sound/firewire/bebob/bebob_stream.c index f7f0db5aa811..1b264d0d63eb 100644 --- a/sound/firewire/bebob/bebob_stream.c +++ b/sound/firewire/bebob/bebob_stream.c @@ -415,15 +415,16 @@ static int make_both_connections(struct snd_bebob *bebob) return 0; }
-static void -break_both_connections(struct snd_bebob *bebob) +static void break_both_connections(struct snd_bebob *bebob) { cmp_connection_break(&bebob->in_conn); cmp_connection_break(&bebob->out_conn);
- /* These models seems to be in transition state for a longer time. */
- if (bebob->maudio_special_quirk != NULL)
msleep(200);
- // These models seem to be in transition state for a longer time. When
- // accessing in the state, any transactions is corrupted. In the worst
- // case, the device is going to reboot.
- if (bebob->version < 2)
msleep(600);}
static int
2.20.1
Alsa-devel mailing list Alsa-devel@alsa-project.org https://mailman.alsa-project.org/mailman/listinfo/alsa-devel
This patch causes a reboot when I power on my Edirol FA-101. This is consistent behavior with kernels 5.5.1 - 5.5.11.
I've changed msleep(600) back to msleep(200) and rebuilt the module and I am once again able to power on and use the interface.
I am running Arch linux with 5.5.11-arch1-1 #1 SMP PREEMPT Sun, 22 Mar 2020 16:33:15 +0000 x86_64 GNU/Linux.
I had previously posted to the list on Feb 19, 2020 with the subject:
sytem reboots when initializing Edirol FA-101 firewire audio interface
Please forgive any lapses in reporting protocol as I generally don't work at the kernel level. Hopefully this more specific report will help fix the issue.
This is weird. If just changing the msleep value from 600 to 200 "fixes" the problem, it means that something is already fundamentally broken...
Takashi
It appears that 'fixed' was too strong a word. While I have managed to get it to work a couple of times, generally it is still broken.
I'm happy to help debug, but this is outside my scope.
Back to the LTS kernel for the moment.
Steve
As long as I investigated, there's some cases about the delay for device between establishing OUTPUT_PLUG and transmitting first isoc packet. For devices which support BeBoB protocol version 1 can transmit the packet within several hundred milliseconds, while for devices which support BeBoB protocol version 3 can transmit the packet within 2 seconds.
Devices with protocol version 1: * Edirol FA-66 * Yamaha GO46 * Terratec Phase x24 FW * M-Audio FireWire AudioPhile * M-Audio FireWire Solo * M-Audio FireWire 1814 * M-Audio FireWire 410 * Focusrite Saffire Pro 26 I/O
Devices with protocol version 3: * M-Audio Profire Lightbridge * Behringer FCA610 * Phonic Firefly 202
At present ALSA bebob driver postpones starting IR context during 1.5 sec for all of supported devices. The delay is too long for devices with protocol version 1, while it's not enough for devices with protocol version 3.
This commit improves the delay for these protocols.
Signed-off-by: Takashi Sakamoto o-takashi@sakamocchi.jp --- sound/firewire/bebob/bebob_stream.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/sound/firewire/bebob/bebob_stream.c b/sound/firewire/bebob/bebob_stream.c index 1b264d0d63eb..895a113635af 100644 --- a/sound/firewire/bebob/bebob_stream.c +++ b/sound/firewire/bebob/bebob_stream.c @@ -7,7 +7,7 @@
#include "./bebob.h"
-#define CALLBACK_TIMEOUT 2000 +#define CALLBACK_TIMEOUT 2500 #define FW_ISO_RESOURCE_DELAY 1000
/* @@ -639,6 +639,7 @@ int snd_bebob_stream_start_duplex(struct snd_bebob *bebob)
if (!amdtp_stream_running(&bebob->rx_stream)) { unsigned int curr_rate; + unsigned int ir_delay_cycle;
if (bebob->maudio_special_quirk) { err = bebob->spec->rate->get(bebob, &curr_rate); @@ -660,13 +661,18 @@ int snd_bebob_stream_start_duplex(struct snd_bebob *bebob)
// The device postpones start of transmission mostly for 1 sec // after receives packets firstly. For safe, IR context starts - // 1.5 sec (=12000 cycles) later. This is within 2.0 sec - // (=CALLBACK_TIMEOUT). + // 0.4 sec (=3200 cycles) later to version 1 or 2 firmware, + // 2.0 sec (=16000 cycles) for version 3 firmware. This is + // within 2.5 sec (=CALLBACK_TIMEOUT). // Furthermore, some devices transfer isoc packets with // discontinuous counter in the beginning of packet streaming. // The delay has an effect to avoid detection of this // discontinuity. - err = amdtp_domain_start(&bebob->domain, 12000); + if (bebob->version < 2) + ir_delay_cycle = 3200; + else + ir_delay_cycle = 16000; + err = amdtp_domain_start(&bebob->domain, ir_delay_cycle); if (err < 0) goto error;
Originally BeBeB ASICs and firmware supports clock mode to synchronizing to syt field of received isoc packet. This mode is known as 'SYT Match' slightly described in IEC 61883-6 (but no detail mechanisms). In this mode, drivers can control sampling clock in device. Driver for Windows and macOS uses this feature to perform synchronization for devices on the same bus.
In this mode, a plug of Music subunit for synchronization is connected to a plug of isoc unit for incoming packet streaming, then the order to establish connections is INPUT_PLUG first, OUTPUT_PLUG second.
This commit implements the above.
Actually each device works with its own clock for sampling, therefore the original design is hardly implemented to vendor's products.
Signed-off-by: Takashi Sakamoto o-takashi@sakamocchi.jp --- sound/firewire/bebob/bebob_stream.c | 51 +++++++++++++---------------- 1 file changed, 22 insertions(+), 29 deletions(-)
diff --git a/sound/firewire/bebob/bebob_stream.c b/sound/firewire/bebob/bebob_stream.c index 895a113635af..bbae04793c50 100644 --- a/sound/firewire/bebob/bebob_stream.c +++ b/sound/firewire/bebob/bebob_stream.c @@ -398,23 +398,6 @@ check_connection_used_by_others(struct snd_bebob *bebob, struct amdtp_stream *s) return err; }
-static int make_both_connections(struct snd_bebob *bebob) -{ - int err = 0; - - err = cmp_connection_establish(&bebob->out_conn); - if (err < 0) - return err; - - err = cmp_connection_establish(&bebob->in_conn); - if (err < 0) { - cmp_connection_break(&bebob->out_conn); - return err; - } - - return 0; -} - static void break_both_connections(struct snd_bebob *bebob) { cmp_connection_break(&bebob->in_conn); @@ -427,8 +410,7 @@ static void break_both_connections(struct snd_bebob *bebob) msleep(600); }
-static int -start_stream(struct snd_bebob *bebob, struct amdtp_stream *stream) +static int start_stream(struct snd_bebob *bebob, struct amdtp_stream *stream) { struct cmp_connection *conn; int err = 0; @@ -438,18 +420,19 @@ start_stream(struct snd_bebob *bebob, struct amdtp_stream *stream) else conn = &bebob->out_conn;
- /* channel mapping */ + // channel mapping. if (bebob->maudio_special_quirk == NULL) { err = map_data_channels(bebob, stream); if (err < 0) - goto end; + return err; }
- // start amdtp stream. - err = amdtp_domain_add_stream(&bebob->domain, stream, - conn->resources.channel, conn->speed); -end: - return err; + err = cmp_connection_establish(conn); + if (err < 0) + return err; + + return amdtp_domain_add_stream(&bebob->domain, stream, + conn->resources.channel, conn->speed); }
static int init_stream(struct snd_bebob *bebob, struct amdtp_stream *stream) @@ -638,6 +621,8 @@ int snd_bebob_stream_start_duplex(struct snd_bebob *bebob) }
if (!amdtp_stream_running(&bebob->rx_stream)) { + enum snd_bebob_clock_type src; + struct amdtp_stream *master, *slave; unsigned int curr_rate; unsigned int ir_delay_cycle;
@@ -647,15 +632,23 @@ int snd_bebob_stream_start_duplex(struct snd_bebob *bebob) return err; }
- err = make_both_connections(bebob); + err = snd_bebob_stream_get_clock_src(bebob, &src); if (err < 0) return err;
- err = start_stream(bebob, &bebob->rx_stream); + if (src != SND_BEBOB_CLOCK_TYPE_SYT) { + master = &bebob->tx_stream; + slave = &bebob->rx_stream; + } else { + master = &bebob->rx_stream; + slave = &bebob->tx_stream; + } + + err = start_stream(bebob, master); if (err < 0) goto error;
- err = start_stream(bebob, &bebob->tx_stream); + err = start_stream(bebob, slave); if (err < 0) goto error;
On Fri, 01 Nov 2019 14:13:20 +0100, Takashi Sakamoto wrote:
Hi,
This patchset improves model-specific issues in a point of protocol version supported by the model. The changes are relevant to packet streaming, thus it can brings some regressions.
For testers, I prepare for a remote branch to backport the latest codes to Linux v4.17 or later. I'm happy to receive any report. https://github.com/takaswie/snd-firewire-improve/tree/topic/bebob-connection...
Takashi Sakamoto (3): ALSA: bebob: expand sleep just after breaking connections for protocol version 1 ALSA: bebob: expand delay of start for IR context just for version 3 firmware ALSA: bebob: link the order of establishing connections and Syt-match clock mode
Applies all patches now.
thanks,
Takashi
participants (3)
-
Steve Morris -
Takashi Iwai -
Takashi Sakamoto