linux-toradex.git/sound/firewire, branch v4.7

ALSA: firewire-lib: change a member of event structure to suppress sparse wanings to bool type

2016-05-18T14:32:09+00:00

Commit a9c4284bf5a9 ("ALSA: firewire-lib: add context information to
tracepoints") adds new members to tracepoint events of this module, to
represent context information. One of the members is bool type and
this causes sparse warnings.

16:1: warning: expression using sizeof bool
60:1: warning: expression using sizeof bool
16:1: warning: odd constant _Bool cast (ffffffffffffffff becomes 1)
60:1: warning: odd constant _Bool cast (ffffffffffffffff becomes 1)

This commit suppresses the warnings, by changing type of the member
to 'unsigned int'. Additionally, this commit applies '!!' idiom to
get 0/1 from 'in_interrupt()'.

Fixes: a9c4284bf5a9 ("ALSA: firewire-lib: add context information to tracepoints")
Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: firewire-lib: drop skip argument from helper functions to queue a packet

2016-05-11T18:36:28+00:00

On most of audio and music units on IEEE 1394 bus which ALSA firewire stack
supports (or plans to support), CIP with two quadlets header is used.
Thus, there's no cases to queue packets with blank payload. If such packets
are going to be queued, it means that they're for skips of the cycle.

This commit simplifies helper functions to queue a packet.

Signed-off-by: Takashi Sakamoto 
Acked-by: Clemens Ladisch 
Signed-off-by: Takashi Iwai

ALSA: firewire-lib: add context information to tracepoints

2016-05-11T18:36:00+00:00

In current implementation, packet processing is done in both of software
IRQ contexts of IR/IT contexts and process contexts.

This is usual interrupt handling of IR/IT context for 1394 OHCI.
(in hardware IRQ context)
irq_handler() (drivers/firewire/ohci.c)
->tasklet_schedule()
(in software IRQ context)
handle_it_packet() or handle_ir_packet_per_buffer() (drivers/firewire/ohci.c)
->flush_iso_completions()
  ->struct fw_iso_context.callback.sc()
  = out_stream_callback() or in_stream_callback()

However, we have another chance for packet processing. It's done in PCM
frame handling via ALSA PCM interfaces.
(in process context)
ioctl(i.e. SNDRV_PCM_IOCTL_HWSYNC)
->snd_pcm_hwsync() (sound/core/pcm_native.c)
  ->snd_pcm_update_hw_ptr() (sound/core/pcm_lib.c)
    ->snd_pcm_update_hw_ptr0()
      ->struct snd_pcm_ops.pointer()
      = amdtp_stream_pcm_pointer()
        ->fw_iso_context_flush_completions() (drivers/firewire/core-iso.c)
          ->struct fw_card_driver.flush_iso_completions()
          = ohci_flush_iso_completions() (drivers/firewire/ohci.c)
            ->flush_iso_completions()
              ->struct fw_iso_context.callback.sc()
              = out_stream_callback() or in_stream_callback()

This design is for a better granularity of PCM pointer. When ioctl(2) is
executed with some commands for ALSA PCM interface, queued packets are
handled at first. Then, the latest number of handled PCM frames is
reported. The number can represent PCM frames transferred in most near
isochronous cycle.

Current tracepoints include no information to distinguish running contexts.
When tracing the interval of software IRQ context, this is not good.

This commit adds more information for current context. Additionally, the
index of packet processed in one context is added in a case that packet
processing is executed in continuous context of the same kind,

As a result, the output includes 11 fields with additional two fields
to commit 0c95c1d6197f ("ALSA: firewire-lib: add tracepoints to dump a part
of isochronous packet data"):
17131.9186: out_packet: 07 7494 ffc0 ffc1 00 000700c0 9001a496 058 45 1 13
17131.9186: out_packet: 07 7495 ffc0 ffc1 00 000700c8 9001ba00 058 46 1 14
17131.9186: out_packet: 07 7496 ffc0 ffc1 00 000700d0 9001ffff 002 47 1 15
17131.9189: out_packet: 07 7497 ffc0 ffc1 00 000700d0 9001d36a 058 00 0 00
17131.9189: out_packet: 07 7498 ffc0 ffc1 00 000700d8 9001e8d4 058 01 0 01
17131.9189: out_packet: 07 7499 ffc0 ffc1 00 000700e0 9001023e 058 02 0 00
17131.9206: in_packet:  07 7447 ffc1 ffc0 01 3f070072 9001783d 058 32 1 00
17131.9206: in_packet:  07 7448 ffc1 ffc0 01 3f070072 90ffffff 002 33 1 01
17131.9206: in_packet:  07 7449 ffc1 ffc0 01 3f07007a 900191a8 058 34 1 02
(Here, some common fields are omitted so that a line is within 80
characters.)

The legend is:
 - The second of cycle scheduled for the packet
 - The count of cycle scheduled for the packet
 - The ID of node as source (hex)
 - The ID of node as destination (hex)
 - The value of isochronous channel
 - The first quadlet of CIP header (hex)
 - The second quadlet of CIP header (hex)
 - The number of included quadlets
 - The index of packet in a buffer maintained by this module
 - 0 in process context, 1 in IRQ context
 - The index of packet processed in the context

Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: firewire-lib: permit to flush queued packets only in process context for better PCM period granularity

2016-05-11T18:34:53+00:00

These three commits were merged to improve PCM pointer granularity.
commit 76fb87894828 ("ALSA: firewire-lib: taskletize the snd_pcm_period_elapsed() call")
commit e9148dddc3c7 ("ALSA: firewire-lib: flush completed packets when reading PCM position")
commit 92b862c7d685 ("ALSA: firewire-lib: optimize packet flushing")

The point of them is to handle queued packets not only in software IRQ
context of IR/IT contexts, but also in process context. As a result of
handling packets, period tasklet is scheduled when acrossing PCM period
boundary. This is to prevent recursive call of
'struct snd_pcm_ops.pointer()' in the same context.

When the pointer callback is executed in the process context, it's
better to avoid the second callback in the software IRQ context. The
software IRQ context runs immediately after scheduled in the process
context because few packets are queued yet.

For the aim, 'pointer_flush' is used, however it causes a race condition
between the process context and software IRQ context of IR/IT contexts.

Practically, this race is not so critical because it influences process
context to skip flushing queued packet and to get worse granularity of
PCM pointer. The race condition is quite rare but it should be improved
for stable service.

The similar effect can be achieved by using 'in_interrupt()' macro. This
commit obsoletes 'pointer_flush' with it.

Acked-by: Clemens Ladisch 
Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: firewire-lib: enable the same feature as CIP_SKIP_INIT_DBC_CHECK flag

2016-05-10T15:04:01+00:00

In former commit, drivers in ALSA firewire stack always starts IT context
before IR context. If IR context starts after packets are transmitted by
peer unit, packet discontinuity may be detected because the context starts
in the middle of packet streaming. This situation is rare because IT
context usually starts immediately. However, it's better to solve this
issue. This is suppressed with CIP_SKIP_INIT_DBC_CHECK flag.

This commit enables the same feature as CIP_SKIP_INIT_DBC_CHECK.

Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: firewire-lib: code cleanup for outgoing packet handling

2016-05-10T15:04:01+00:00

In previous commit, this module has no need to reuse parameters of
incoming packets for outgoing packets anymore. This commit arranges some
needless codes for outgoing packet processing.

Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: firewire-lib: code cleanup for incoming packet handling

2016-05-10T15:04:00+00:00

In previous commit, this module has no need to reuse parameters of
incoming packets for outgoing packets anymore. This commit arranges some
needless codes for incoming packet processing.

Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: firewire-lib: handle IT/IR contexts in each software interrupt context

2016-05-10T15:03:59+00:00

In clause 6.3 of IEC 61883-6:2000, there's an explanation about processing
of presentation timestamp. In the clause, we can see "If a function block
receives a CIP, processes it and subsequently re-transmits it, then the
SYT of the outgoing CIP shall be the sum of the incoming SYT and the
processing delay." ALSA firewire stack has an implementation to partly
satisfy this specification. Developers assumed the stack to perform as an
Audio function block[1].

Following to the assumption, current implementation of ALSA firewire stack
use one software interrupt context to handle both of in/out packets. In
most case, this is processed in 1394 OHCI IR context independently of the
opposite context. Thus, this implementation uses longer CPU time in the
software interrupt context. This is not better for whole system.

Against the assumption, I confirmed that each ASIC for IEC 61883-1/6
doesn't necessarily expect it to the stack. Thus, current implementation
of ALSA firewire stack includes over-engineering.

This commit purges the implementation. As a result, packets of one
direction are handled in one software interrupt context and spends
minimum CPU time.

[1] [alsa-devel] [PATCH 0/8] [RFC] new driver for Echo Audio's Fireworks based devices
http://mailman.alsa-project.org/pipermail/alsa-devel/2013-June/062660.html

Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: firewire-tascam: drop reuse of incoming packet parameter for outgoing packet parameter

2016-05-10T15:03:59+00:00

In packet streaming protocol applied to TASCAM FireWire series, the value
of SYT field in CIP header is always zero, therefore it has no meaning.
There's no need to synchronize packets in both direction for the series.

In current implementation of ALSA firewire stack, driver for the series
uses incoming packet parameter for outgoing packet parameter to calculate
the number of data blocks. This can be simplified because the task of
corresponding driver is to transfer data blocks enough to sampling transfer
frequency.

This commit purges support of full duplex synchronization to prevent
over-engineering implementation.

Signed-off-by: Takashi Sakamoto 
Signed-off-by: Takashi Iwai

ALSA: fireworks: drop reuse of incoming packet parameter for ougoing packet parameter

2016-05-10T15:03:58+00:00

On Fireworks board module of Echo Audio, TSB43Cx43A (IceLynx Micro, iCEM)
is used to process payload of isochronous packets. There's an public
document of this chip[1]. This document is for firmware programmers to
transfer/receive AMDTP with IEC60958 data format, however in clause 2.5,
2.6 and 2.7, we can see system design to utilize the sequence of value in
SYT field of CIP header. In clause 2.3, we can see the specification of
Audio Master Clock (MCLK) from iCEM.

Well, this clock is actually not used for sampling clock. This can be
confirmed when corresponding driver transfer random value as the sequence
of SYT field. Even if in this case, the unit generates proper sound.

Additionally, in unique command set for this board module, the format
of CIP is changed; for IEC 61883-6 mode which we use, and for Windows
Operating System. In the latter mode, the whole 32 bit field in second CIP
header from Windows driver is used to represent counter of packets (NO-DATA
code is still used for packets without data blocks). If the master clock
was physically used by DSP on the board module, the Windows driver must
have transferred correct sequence of SYT field.

Furthermore, as long as seeing capacities of AudioFire2, AudioFire4,
AudioFire8, AudioFirePre8 and AudioFire12, these models don't support
SYT-Match clock source.

Summary, we have no need to relate incoming/outgoing packets. This commit
drops reusing SYT sequence of incoming packets for outgoing packets.

[1] Using TSB43Cx43A: S/PDIF over 1394 (2003, Texus Instruments
Incorporated)
http://www.ti.com/analog/docs/litabsmultiplefilelist.tsp?literatureNumber=slla148&docCategoryId=1&familyId=361

Signed-off-by: Takashi Sakamoto
Signed-off-by: Takashi Iwai