The > should be >= or otherwise we potentially read one element beyond
the end of the ff->tx_midi_substreams[] array.
Fixes: 73f5537fb2 ("ALSA: fireface: support tx MIDI functionality of Fireface UCX")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch changes the parent pointer assignment of snd_info_entry
object to be always non-NULL. More specifically,check the parent
argument in snd_info_create_module_entry() & co, and assign
snd_proc_root if NULL is passed there.
This assures that the proc object is always freed when the root is
freed, so avoid possible memory leaks. For example, some error paths
(e.g. snd_info_register() error at snd_minor_info_init()) may leave
snd_info_entry object although the proc file itself is freed.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The proc files are recursively freed by calling with the root
snd_info_entry object, so we don't have to keep each object for
releasing one by one. Move the release of the PCM stream proc root at
the beginning, so that we can remove the redundant code and resource.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In Fireface series, registration of higher 4 bytes of destination
address for asynchronous transaction of MIDI messages is done by
a write transaction to model-specific register.
On the other hand, registration of lower 4 bytes of the address is
selectable from 4 options. A register for this registration includes
the other purpose options such as input attenuation. Thus this
driver expects userspace applications to configure the register.
Actual behaviour for the asynchronous transaction is different
depending on protocols. In former protocol, destination offset
of each transaction is the same as the registered address even if
it is block request. In latter models, destination offset of each
transaction is the offset of previous transaction plus 4 byte
and the transaction is quadlet request.
This commit cleanups comments about the above mechanism.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
After the previous code refactoring, the PCM stream locking code
became nothing but the PCM group lock with self_group object. Use the
existing helper function for simplifying the code.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Remove the hackish down_write_nonfifo() that was introduced as a
workaround of rwsem deadlock.
It used to be a problem for non-atomic PCM streams that take the rwsem
for the locking and hit the high lock contention. Since the current
PCM locking refactoring, we'll no longer hit it as the hot code-paths
don't take global locks.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
We have currently two global locks, a rwlock and a rwsem, that are
used for managing linking the PCM streams. Due to these global locks,
once when a linked stream is used, the lock granularity suffers a
lot.
This patch attempts to eliminate the former global lock for atomic
ops. The latter rwsem needs remaining because of the loosy way of the
loop calls in snd_pcm_action_nonatomic(), as well as for avoiding the
deadlock at linking. However, these are used far rarely, actually
only by two actions (prepare and reset), where both are no timing
critical ones. So this can be still seen as a good improvement.
The basic strategy to eliminate the rwlock is to assure group->lock at
adding or removing a stream to / from the group. Since we already
takes the group lock whenever taking the all substream locks under the
group, this shouldn't be a big problem. The reference to group
pointer in snd_pcm_substream object is protected by the stream lock
itself.
However, there are still pitfalls: a race window at re-locking and the
lifecycle of group object. The former is a small race window for
dereferencing the substream group object opened while snd_pcm_action()
performs re-locking to avoid ABBA deadlocks. This includes the unlink
of group during that window, too. And the latter is the kfree
performed after all streams are removed from the group while it's
still dereferenced.
For addressing these corner cases, two new tricks are introduced:
- After re-locking, the group assigned to the stream is checked again;
if the group is changed, we retry the whole procedure.
- Introduce a refcount to snd_pcm_group object, so that it's freed
only when it's empty and really no one refers to it.
(Some readers might wonder why not RCU for the latter. RCU in this
case would cost more than refcounting, unfortunately. We take the
group lock sooner or later, hence the performance improvement by RCU
would be negligible. Meanwhile, because we need to deal with
schedulable context depending on the pcm->nonatomic flag, it'll become
dynamic RCU/SRCU switch, and the grace period may become too long.)
Along with these changes, there are a significant amount of code
refactoring. The complex group re-lock & ref code is factored out to
snd_pcm_stream_group_ref() function, for example.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In latter model of Fireface series, asynchronous transaction includes
a prefix byte to indicate the way to decode included MIDI bytes.
Upper 4 bits of the prefix byte indicates port number, and the rest 4
bits indicate the way to decode rest of bytes for MIDI messages.
Basically the rest bits indicates the number of bytes for MIDI message.
However, if the last byte of each MIDi message is included, the rest
bits are 0xf. For example:
message: f0 00 00 66 14 20 00 00 f7
offset: content (big endian, port 0)
'0030: 0x02f00000
'0030: 0x03006614
'0030: 0x03200000
'0030: 0x0ff70000
This commit supports encoding scheme for the above and allows
applications to transfer MIDI messages via ALSA rawmidi interface.
An unused member (running_status) is reused to keep state of
transmission of system exclusive messages.
For your information, this is a dump of config rom.
$ sudo ./hinawa-config-rom-printer /dev/fw1
{ 'bus-info': { 'bmc': False,
'chip_ID': 13225063715,
'cmc': False,
'cyc_clk_acc': 0,
'imc': False,
'isc': True,
'max_rec': 512,
'name': '1394',
'node_vendor_ID': 2613},
'root-directory': [ [ 'NODE_CAPABILITIES',
{ 'addressing': {'64': True, 'fix': True, 'prv': False},
'misc': {'int': False, 'ms': False, 'spt': True},
'state': { 'atn': False,
'ded': False,
'drq': True,
'elo': False,
'init': False,
'lst': True,
'off': False},
'testing': {'bas': False, 'ext': False}}],
['VENDOR', 2613],
['DESCRIPTOR', 'RME!'],
['EUI_64', 2873037108442403],
[ 'UNIT',
[ ['SPECIFIER_ID', 2613],
['VERSION', 4],
['MODEL', 1054720],
['DESCRIPTOR', 'Fireface UCX']]]]}
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Between former and latter models, content of asynchronous transaction
for MIDI messages from driver to device is different.
This commit is a preparation to support latter models. A protocol-specific
operation is added to encode MIDI messages to the transaction.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Between former and latter models, destination address to receive
asynchronous transactions for MIDI messages is different.
This commit adds model-dependent parameter for the addresses.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fireface UCX transfers asynchronous transactions for MIDI messages.
One transaction includes quadlet data therefore it can transfer 3
message bytes as maximum. Base address of the destination is
configured by two settings; a register for higher 8 byte of the
address, and a bitflag to option register indicates lower 8byte.
The register for higher address is 0x'ffff'0000'0034. Unfortunately,
firmware v24 includes a bug to ignore registered value for the
destination address and transfers to 0x0001xxxxxxxx always. This
driver doesn't work well if the bug exists, therefore users should
install the latest firmware (v27).
The bitflag is a part of value to be written to option register
(0x'ffff'0000'0014).
lower addr: bitflag (little endian)
'0000'0000: 0x00002000
'0000'0080: 0x00004000
'0000'0100: 0x00008000
'0000'0180: 0x00010000
This register includes more options but they are not relevant to
packet streaming or MIDI functionality. This driver don't touch it.
Furthermore, the transaction is sent to address offset incremented
by 4 byte to the offset in previous time. When it reaches base address
plus 0x7c, next offset is the base address.
Content of the transaction includes a prefix byte. Upper 4 bits of
the byte indicates port number, and the rest 4 bits indicate the way
to decode rest of bytes for MIDI message.
Except for system exclusive messages, the rest bits are the same as
status bits of the message without channel bits. For system exclusive
messages, the rest bits are encoded according to included message bytes.
For example:
message: f0 7e 7f 09 01 f7
offset: content (little endian, port 0)
'0000: 0x04f07e7f
'0004: 0x070901f7
message: f0 00 00 66 14 20 00 00 00 f7
offset: content (little endian, port 1)
'0014: 0x14f00000
'0018: 0x14661420
'001c: 0x14000000
'0020: 0x15f70000
message: f0 00 00 66 14 20 00 00 f7
offset: content (little endian, port 0)
'0078: 0x04f00000
'007c: 0x04661420
'0000: 0x070000f7
This commit supports decoding scheme for the above and allows
applications to receive MIDI messages via ALSA rawmidi interface.
The lower 8 bytes of destination address is fixed to 0x'0000'0000,
thus this driver expects userspace applications to configure option
register with bitflag 0x00002000 in advance.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In Fireface series, drivers can register destination address for
asynchronous transaction which transfers MIDI messages from device.
In former models, all of the transactions arrive at the registered
address without any offset. In latter models, each of the transaction
arrives at the registered address with sequential offset within 0x00
to 0x7f. This seems to be for discontinuity detection.
This commit adds model-dependent member for the address range.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In a series of Fireface, devices transfer asynchronous transaction with
MIDI messages. In the transaction, content is different depending on
models. ALSA fireface driver has protocol-dependent handler to pick up
MIDI messages from the content.
In latter models of the series, the transaction is transferred to range
of address sequentially. This seems to check continuity of transferred
messages.
This commit changes prototype of the handler to receive offset of
address for received transactions.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
AZX_DCAPS_PM_RUNTIME flag is added to indicate support for runtime PM.
azx_has_pm_runtime() is used to check if above is enabled and thus
forbid runtime PM calls if needed.
Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
Reviewed-by: Mohan Kumar D <mkumard@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch moves clock enable/disable from system resume/suspend to
runtime resume/suspend respectively. Along with this hda controller
chip init or stop is also moved. System resume/suspend can invoke
runtime callbacks and do necessary setup.
chip->running can be used to check for probe completion and device
access during runtime_resume or runtime_suspend can be avoided if
probe is not yet finished. This helps to avoid kernel panic during
boot where runtime PM callbacks can happen from system PM.
Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
Reviewed-by: Mohan Kumar D <mkumard@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Explicit clock enable is not required during probe, as this would be
managed by runtime PM calls. Clock can be enabled/disabled in runtime
resume/suspend. This way it is easier to balance clock enable/disable
counts.
Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
Reviewed-by: Mohan Kumar D <mkumard@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Moved devm_clk_get() API calls to a separate function and the same
can be called early in the probe. This is done before runtime PM
for the device is enabled. The runtime resume/suspend callbacks can
later enable/disable clocks respectively(the support would be added
in subsequent patches). Clock handles should be available by the
time runtime suspend/resume calls can happen.
Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
Reviewed-by: Mohan Kumar D <mkumard@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This patch enables runtime power management(runtime PM) support for
hda. pm_runtime_enable() and pm_runtime_disable() are added during
device probe and remove respectively. The runtime PM callbacks will
be forbidden if hda controller does not have support for runtime PM.
pm_runtime_get_sync() and pm_runtime_put() are added for hda register
access. The callbacks for above will be added in subsequent patches.
Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
Reviewed-by: Mohan Kumar D <mkumard@nvidia.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The snd_pcm_group_for_each_entry() loop found in snd_pcm_unlink() is
only for taking the first list entry. Use list_first_entry() to make
clearer.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Make a common helper to re-assign the PCM link using list_move() instead
of open code with manual list_del() and list_add_tail(). This assures
the consistency and we can get rid of snd_pcm_group.count field -- its
purpose is only to check whether the list is singular, and we can know
it by list_is_singular() call now.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
There are multiple open codes that initialize the same object.
Create a common helper function instead.
Also, use kzalloc() to be safer at creating a group object, and move
the initialization out of the critical section.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The snd_card_unref() call in snd_pcm_link() looks suspicious through a
quick glance, but it's a correct usage; this is needed just because
the file descriptor check in is_pcm_file() calls the helper
snd_lookup_minor_data() that keeps the card refcount.
Despite of the correctness, the code still looks confusing.
Basically, keeping the card ref for the whole code isn't needed
as fdget() blocks the release of the opened file. Hence it's more
understandable if snd_card_unref() is moved into is_pcm_file(), then
the caller doesn't have to take care after the call.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Fireface UFX was shipped by RME GmbH in 2012. This model supports later
protocol for management of isochronous communication and synchronization
of sampling transmission frequency.
This commit adds support for the model. At present, it's not clear how
to encode MIDI messages and decide destination address for asynchronous
transaction, thus this commit adds support for isochronous communication
for PCM frames only.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
A procedure to retrieve clock configuration is used by two callers.
Each of caller has duplicated code to parse bits.
This commit adds refactoring to remove the duplicated code.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds refactoring for dump of sync status by adding
tables for check bits.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds a member for a callback function to get clock status
to former protocol.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds a member for a callback function to switch frame
fetching mode to former protocol.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds a member for a callback function to dump status and
move existing code to former protocol.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In a series of Fireface, latter protocol has no way for drivers to
retrieve current clock configuration. On the other hand, this driver
has proc node for it.
This commit removes a proc node to dump both clock configuration
and synchronization status in one proc node.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit moves codes for Fireface 400 to a file of former protocol.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In a series of Fireface, later model supports different protocol
from former models.
This commit is a preparation to support both of protocols.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The unused variable was forgotten to be removed and now we get a
compiler warning:
sound/pci/hda/hda_codec.c: In function 'hda_codec_runtime_suspend':
sound/pci/hda/hda_codec.c:2926:18: warning: unused variable 'pcm'
Fixes: 17bc4815de ("ALSA: pci: Remove superfluous snd_pcm_suspend*() calls")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Pull the PCM suspend improvement / cleanup.
This moves the most of snd_pcm_suspend*() calls into PCM's own device
PM ops. There should be no change from the functionality POV.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
snd_pcm_suspend() is no longer called from outside, so let's make it
local static. Also drop a superfluous NULL check there.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The call of snd_pcm_suspend_all() & co became superfluous since we
call it in the PCM PM ops. Let's remove them.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
ATIIXP driver supports the full PCM resume and saves/restores the
running PCM pointer. This used to be done in the suspend and resume
callbacks together with snd_pcm_suspend() call. But since we moved
the snd_pcm_supsend*() call in PCM device PM ops, this should be moved
to a more appropriate place, i.e. the trigger callback.
Along with the movement of the PCM suspend/resume code, remove the
superfluous snd_pcm_suspend_all() call, too.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Until now we rely on each driver calling snd_pcm_suspend*() explicitly
at its own PM handling. However, this can be done far more easily by
setting the PM ops to each actual snd_pcm device object.
This patch adds the device_type object for PCM stream and assigns to
each PCM stream object. The type contains only the PM ops for system
suspend; we don't need to deal with the resume in general.
The suspend hook simply calls snd_pcm_suspend_all() for the given PCM
streams. This implies that the PM order is correctly put, i.e. PCM is
suspended before the main (or codec) driver, which should be true in
general. If a special ordering is needed, you'd need to adjust the
device PM order manually later.
This patch introduces a new flag, snd_pcm.no_device_suspend, too.
With this flag set, the PCM device object won't invoke
snd_pcm_suspend_all() by itself. This is needed for ASoC who wants to
manage the PM call orders in its serialized way, and the flag is set
in soc_new_pcm() as default.
For the non-ASoC world, we can get rid of the manual snd_pcm_suspend
calls. This will be done in the later patches.
Reviewed-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Program codec stripe through AC_VERB_SET_STRIPE_CONTROL to use multiple
sdo lines if supported. Audio needs to be striped across number of sdo
lines for simultaneous playbacks of higher resolutions to work.
This needs to be implemented only for an Audio Output Converter and only
if the stripe bit(AC_WCAP_STRIPE) of Audio Widget Capabilities parameter
is 1.
Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Reviewed-by: Mohan Kumar D <mkumard@nvidia.com>
Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Platforms having multiple SORs and hdmi/dp sinks require higher
bandwidth to support simultaneous playbacks of higher resolution.
If hda controller supports multiple SDO lines, STRIPE can be used
to indicate how many of the SDO lines the stream should be striped
across.
During stream start stripe control bits are programmed to use given
number of sdo lines and the same is cleared during stream stop.
Signed-off-by: Sameer Pujar <spujar@nvidia.com>
Reviewed-by: Mohan Kumar D <mkumard@nvidia.com>
Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
