This commit adds a functionality to manage packet streaming for MOTU
FireWire series.
The streaming is not controlled by CMP, thus against IEC 61883-1. Write
transaction to certain addresses start/stop packet streaming.
Transactions to 0x'ffff'f000'0b00 results in isochronous channel number for
both directions and starting/stopping transmission of packets. The
isochronous channel number is represented in 6 bit field, thus units can
identify the channels up to 64, as IEEE 1394 bus specification described.
Transactions to 0x'ffff'f000'0b10 results in packet format for both
directions and transmission speed. When each of data block includes fixed
part of data chunks only, corresponding flags stand.
When bus reset occurs, the units continue to transmit packets with
non-contiguous data block counter. This causes discontinuity detection in
packet streaming engine and ALSA PCM applications receives EPIPE from any
I/O operation. In this case, typical applications manage to recover
corresponding PCM substream. This behaviour is kicked much earlier than
callback of bus reset handler by Linux FireWire subsystem, therefore
status of packet streaming is not changed in the handler.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
All models of MOTU FireWire series can be controlled by write transaction
to addresses in a range from 0x'ffff'f0000'0b00 to 0x'ffff'f000'0cff.
The models support asynchronous notification. This notification has 32 bit
field data, and is transferred when status of clock changes. Meaning of
the value is not enough clear yet.
Drivers can register its address to receive the notification. Write
transaction to 0x'ffff'f000'0b04 registers higher 16 bits of the address.
Write transaction to 0x'ffff'f0000'0b08 registers the rest of bits. The
address includes node ID, thus it should be registered every time of bus
reset.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
MOTU FireWire series uses blocking transmission for AMDTP packet streaming.
They transmit/receive 8,000 packets per second, to handle the same number
of data blocks as current sampling transmission frequency. Thus,
IEC 61883-1/6 packet streaming engine of ALSA firewire stack is available
for them.
However, the sequence of packet and data blocks includes some quirks.
Below sample is a sequence of CIP headers of packets received by 828mk2,
at 44.1kHz of sampling transmission frequency.
quads CIP1 CIP2
488 0x020F04E8 0x8222FFFF
8 0x020F04F8 0x8222FFFF
488 0x020F0400 0x8222FFFF
488 0x020F0408 0x8222FFFF
8 0x020F04E8 0x8222FFFF
488 0x020F04F0 0x8222FFFF
488 0x020F04F8 0x8222FFFF
The SID (source node ID), DBS (data block size), SPH (source packet header),
FMT (format ID), FDF (format dependent field) and SYT (time stamp) fields
are in IEC 61883-1. Especially, FMT is 0x02, FDF is 0x22 and SYT is 0xffff
to define MOTU specific protocol. In an aspect of dbc field, the value
represents accumulated number of data blocks included the packet. This
is against IEC 61883-1, because according to the specification this value
should be the number of data blocks already transferred.
In ALSA IEC 61883-1/6 engine, this quirk is already supported by
CIP_DBC_IS_END_EVENT flag, because Echo Audio Fireworks has.
Each data block includes SPH as its first quadlet field, to represent its
presentation time stamp. Actual value of SPH is compliant to IEC 61883-1;
lower 25 bits of 32 bits width consists of 13 bits cycle count and 12 bits
cycle offset.
The rest of each data block consists of 24 bit chunks. All of PCM samples,
MIDI messages, status and control messages are transferred by the chunks.
This is similar to '24-bit * 4 Audio Pack' in IEC 61883-6. The position of
each kind of data depends on generations of each model. The number of
whole chunks in a data block is a multiple of 4, to consists of
quadlet-aligned packets.
This commit adds data block processing layer specific for the MOTU
protocol. The remarkable point is the way to generate SPH header. Time
stamps for each data blocks are generated by below calculation:
* Using pre-computed table for the number of ticks per event
* 44,1kHz: (557 + 123/441)
* 48.0kHz: (512 + 0/441)
* 88.2kHz: (278 + 282/441)
* 96.0kHz: (256 + 0/441)
* 176.4kHz: (139 + 141/441)
* 192.0kHz: (128 + 0/441)
* Accumulate the ticks and set the value to SPH for every events.
* This way makes sense only for blocking transmission because this mode
transfers fixed number or none of events.
This calculation assumes that each data block has a PCM frame which is
sampled according to event timing clock. Current packet streaming layer
has the same assumption.
Although this sequence works fine for MOTU FireWire series at sampling
transmission frequency based on 48.0kHz, it is not enough at the frequency
based on 44.1kHz. The units generate choppy noise every few seconds.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Commit c8bdf49b9935("ALSA: fireworks/firewire-lib: Add a quirk for the
meaning of dbc") adds CIP_DBC_IS_END_EVENT flag just for tx packets.
However, MOTU FireWire series has this quirk for rx packets.
This commit allows both directions with the flag.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In IEC 61883-1, CIP headers can have a SPH field. When a packet has 1 in
SPH field of its CIP header, the packet has a source packet headers. A
source packet header consists of 32 bit field (= 1 quadlet) and it
transfers time stamp, which is the same value as the lower 25 bits of the
IEEE 1394 CYCLE_TIMER register and the rest is zero.
This commit just supports source packet header field because IEC 61883-1
includes ambiguity the position of this header and its count. Each
protocol layer is allowed to have actual implementation according its
requirements.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Currently, packet streaming layer passes generated SYT value to data block
processing layer. However, this is not enough in a case that the data block
processing layer generates time stamps by its own ways.
For out-packet stream, the packet streaming layer guarantees 8,000 times
calls of data block processing layers per sec. Therefore, when cycle count
of the first packet is recorded, data block processing layers can calculate
own time stamps with the recorded value.
For the reason, this commit allows packet streaming layer to record the
first cycle count. Each data block processing layer can read the count by
accessing a member of structure for packet streaming layer.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
In an aspect of used protocols to communicate, models of MOTU FireWire
units are categorized to three generations.
This commit adds an abstraction layer of the protocols for features
related to packet streaming functionality. This layer includes 5
operations.
When configuring packet streaming functionality with sampling rate and
sampling transmission frequency, .get_clock_rate and .set_clock_rate are
called with proper arguments. MOTU FireWire series supports up to 192.0kHz.
When checking current source of sampling clock (not clock for packetization
layer), .get_clock_source is used. Enumeration is added to represent the
sources supported by this series. This operation can be used to expose
available sampling rate to user space applications when the unit is
configured to use any input signal as source of clock instead of crystal
clock.
In the protocols, the path between packet processing layer and digital
signal processing layer can be controlled. This looks a functionality to
'mute' the unit. For this feature, .switch_fetching_mode is added. This
can be used to suppress noises every time packet streaming starts/stops.
In a point of the size of data blocks at a certain sampling transmission
frequency, the most units accept several modes. This is due to usage of
optical interfaces. The size differs depending on which modes are
configured to the interfaces; None, S/PDIF and ADAT. Additionally, format
of packet is different depending on protocols. To cache current size of
data blocks and its format, .cache_packet_formats is added. This is used
by PCM functionality, packet streaming functionality and data block
processing layer.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
MOTU FireWire series doesn't tell drivers their capabilities, thus
the drivers should have and apply model-dependent parameters to detected
models.
This commit adds a structure to represent such parameters. Capabilities
are represented by enumeration except for the number of analog line
in/out. Identification name also be in the structure because the units has
no registers for this purpose.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Just after appearing on IEEE 1394 bus, this unit generates several bus
resets. This is due to loading firmware from on-board flash memory and
initialize hardware. It's better to postpone sound card registration.
This commit applies this idea.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
This commit adds an new driver for MOTU FireWire series. In this commit,
this driver just creates/removes card instance according to bus event.
More functionalities will be added in following commits.
Signed-off-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
When trying to add a gpio-hog, we enter a weird loop where the gpio-ranges
is needed when gpiochip_add_data() is called but in the current implementation
the ranges are added from the driver afterwards.
A simple solution is to rely on the DR gpio-ranges attribute and remove the
call to gpiochip_add_pin_range().
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Add a Git tree on @kernel.org for maintaining the Samsung pinctrl
drivers. The tree will be maintained in a shared model between current
Samsung pinctrl maintainers. Pull requests will be going to Linus
Walleij.
Also add the patchwork for linux-samsung-soc mailing list which will be
used for handling the patches.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Acked-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Fix some inverted bit numbers in some pinctrl groups and add missing pins
and groups to be in pair with the GXBB pinctrl pins definition.
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
With real-time preemption, regmap functions cannot be used in the
implementation of irq_chip since they use spinlocks which may sleep.
Move the setting of the mux for IRQs to an irq_bus_sync_unlock handler
where we are allowed to sleep.
Signed-off-by: John Keeping <john@metanate.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
We need to avoid calling regmap functions from irq handlers, so the next
commit is going to move the call to rockchip_set_mux() into an
irq_bus_sync_unlock handler. But we can't return an error from there so
we still need to check the settings from rockchip_irq_set_type() and we
will use this new rockchip_verify_mux() function from there.
Signed-off-by: John Keeping <john@metanate.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This lock is used from rockchip_irq_set_type() which is part of the
irq_chip implementation and thus must use raw_spinlock_t as documented
in Documentation/gpio/driver.txt.
Signed-off-by: John Keeping <john@metanate.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
regmap_update_bits does its own locking and everything else accessed
here is a local variable so there is no need to lock around it.
Signed-off-by: John Keeping <john@metanate.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Setup each port to use the first bulk-out endpoint in calc_num_ports so
that core allocates the corresponding port resources for us.
Signed-off-by: Johan Hovold <johan@kernel.org>
Setup each port to use the first bulk-out endpoint in calc_num_ports so
that core allocates the corresponding port resources for us.
Signed-off-by: Johan Hovold <johan@kernel.org>
Add an explicit sanity check to make sure we have the expected
endpoints. This will provide a descriptive error message in case an
expected endpoint is missing when probing.
Note that the driver already gracefully fails to probe (albeit with a
less descriptive error message) if a bulk-in endpoint is missing, and an
attempt to write to a port whose device lack a bulk-out endpoint would
fail with -ENODEV.
Signed-off-by: Johan Hovold <johan@kernel.org>
Print a message and register two ports for interfaces for which we do
not know how many ports there are instead of binding, allocating
resources, but not register any ports.
This provides a hint that anyone adding a dynamic device id must also
provide a reference id (driver info) from which the port count can be
retrieved, for example:
echo <vid> <pid> 0 0x110A 0x1410 > new_id
Signed-off-by: Johan Hovold <johan@kernel.org>
Use the new endpoint-remap functionality to configure the ports for
treo devices instead of poking around in the port structures after the
ports have been setup.
Signed-off-by: Johan Hovold <johan@kernel.org>
Use the new endpoint-remap functionality to configure the ports for
clie_5 devices.
Note that the same bulk-out endpoint is being used for both ports.
Signed-off-by: Johan Hovold <johan@kernel.org>
Drop the redundant calc_num_ports callback from the clie_5 type, for
which the callback always returns zero and hence falls back to the type
num_ports value (2).
Signed-off-by: Johan Hovold <johan@kernel.org>
These devices always require at least one bulk-out endpoint so let core
verify that.
This avoids attempting to send bulk data to the default pipe when
downloading firmware in boot mode.
Note that further endpoints are still needed when not in boot mode.
Signed-off-by: Johan Hovold <johan@kernel.org>
The pci-idio-16 gpio driver currently implements an irq_chip for handling
interrupts; due to how irq_chip handling is done, it's necessary for the
irq_chip methods to be invoked from hardirq context, even on a a
real-time kernel. Because the spinlock_t type becomes a "sleeping"
spinlock w/ RT kernels, it is not suitable to be used with irq_chips.
A quick audit of the operations under the lock reveal that they do only
minimal, bounded work, and are therefore safe to do under a raw spinlock.
Signed-off-by: Julia Cartwright <julia@ni.com>
Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Now that the endpoint-port mapping has been properly set up during
probe, we can switch to using the more efficient generic write
implementation.
Note that this currently means that chars_in_buffer now overcounts
slightly as we always write a full endpoint-sized packet.
Also add a copyright entry.
Signed-off-by: Johan Hovold <johan@kernel.org>
The 104-idio-16 gpio driver currently implements an irq_chip for handling
interrupts; due to how irq_chip handling is done, it's necessary for the
irq_chip methods to be invoked from hardirq context, even on a a
real-time kernel. Because the spinlock_t type becomes a "sleeping"
spinlock w/ RT kernels, it is not suitable to be used with irq_chips.
A quick audit of the operations under the lock reveal that they do only
minimal, bounded work, and are therefore safe to do under a raw spinlock.
Signed-off-by: Julia Cartwright <julia@ni.com>
Acked-by: William Breathitt Gray <vilhelm.gray@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
These devices use the second bulk-out endpoint for writing. Instead of
using the resources of the second port structure setup by core, use the
new endpoint-remap functionality to simply ignore the first bulk-out
endpoint. This specifically avoids allocating resources for the unused
endpoint.
Note that the disconnect callback was always redundant as all URBs would
have been killed by USB core on disconnect.
Signed-off-by: Johan Hovold <johan@kernel.org>
Clean up the endpoint sanity check by letting core verify the single
interrupt endpoint, and verifying the bulk endpoints in calc_num_ports
after having determined the number of ports.
Note that the static type num_ports field was neither correct or used
(since calc_num_ports never returns zero).
Signed-off-by: Johan Hovold <johan@kernel.org>
This driver have treated the interrupt endpoint as optional despite it
always being present (according to the datasheet). Let's consider it
mandatory instead.
Signed-off-by: Johan Hovold <johan@kernel.org>
Clean up the mcs7715 port setup by using the new endpoint-remap
functionality provided by core. Instead of poking around in internal
port-structure fields, simply swap the endpoint descriptors of the two
ports in calc_num_ports before the port structures are even allocated.
Note that we still need to override the default interrupt completion
handler.
Signed-off-by: Johan Hovold <johan@kernel.org>
Drop the redundant read-urb check from open. The presence of a bulk-in
endpoint is now verified during probe and core has allocated the
corresponding resources.
Signed-off-by: Johan Hovold <johan@kernel.org>
Verify that the required interrupt endpoint is present at probe rather
than at open to avoid allocating resources for an unusable device.
Note that the endpoint is only required when in download mode.
Signed-off-by: Johan Hovold <johan@kernel.org>
These devices always require at least one bulk-out endpoint so let core
verify that.
This avoids attempting to send bulk data to the default pipe when
downloading firmware in boot mode.
Note that further endpoints are still needed when not in boot mode.
Signed-off-by: Johan Hovold <johan@kernel.org>
Use the calc_num_ports rather than attach callback to verify that the
required endpoints are present when in download mode.
This avoids allocating port resources for interfaces that won't be bound.
Signed-off-by: Johan Hovold <johan@kernel.org>
Use the calc_num_ports callback to ignore unused endpoints.
The driver binds to any interface with at least one bulk-in and one
bulk-out endpoint, but some devices can have three or more endpoints of
which only either the first or second pair of endpoints is needed.
This avoids allocating resources for unused endpoints, and specifically
a port is no longer registered for the unused first endpoint pair when
there are more than three endpoints.
Signed-off-by: Johan Hovold <johan@kernel.org>
Use the calc_num_ports rather than attach callback to determine which
interface to bind to in order to avoid allocating port-resources for
interfaces that won't be bound.
Signed-off-by: Johan Hovold <johan@kernel.org>
We can now abort probe early after an error in calc_num_ports by
returning an errno instead of attempting to continue probing but not
register any ports.
Signed-off-by: Johan Hovold <johan@kernel.org>
Simplify the endpoint sanity check by letting core verify that the
required endpoints are present and moving the max-packet check to
calc_num_ports.
Signed-off-by: Johan Hovold <johan@kernel.org>
Use the calc_num_ports rather than probe callback to determine which
interface to bind to.
This allows us to remove some duplicate code.
Signed-off-by: Johan Hovold <johan@kernel.org>
Implement the "horrible endpoint hack" for some legacy devices as a
quirk and clean up the code somewhat.
Note that the bulk-endpoint check can be removed as core will already
have verified this.
Signed-off-by: Johan Hovold <johan@kernel.org>
Some pl2303 devices require the use of the interrupt endpoint of an
unrelated interface. This has so far been dealt with in usb-serial core,
but can now be moved to a driver calc_num_ports callback.
Note that we relax the endpoint requirements checked by core and instead
verify that we have an interrupt-in endpoint in calc_num_ports for all
devices so that the hack can first be applied.
Signed-off-by: Johan Hovold <johan@kernel.org>
Relax the generic driver bulk-endpoint requirement. The driver handles
devices without bulk-out endpoints just fine these days.
Signed-off-by: Johan Hovold <johan@kernel.org>
Add a calc_num_ports callback to the generic driver and verify that the
device has the required endpoints there instead of in core.
Note that the generic driver num_ports field was never used.
Signed-off-by: Johan Hovold <johan@kernel.org>
Add a probe callback to the generic driver and print the
only-for-testing message there.
This is a first step in getting rid of the CONFIG_USB_SERIAL_GENERIC
ifdef from usb-serial core.
Signed-off-by: Johan Hovold <johan@kernel.org>
Allow subdrivers to modify the port-endpoint mapping by passing the
endpoint descriptors to calc_num_ports.
The callback can now also be used to verify that the required endpoints
exists and abort probing otherwise.
This will allow us to get rid of a few hacks in subdrivers that are
already modifying the port-endpoint mapping (or aborting probe due to
missing endpoints), but only after the port structures have been setup.
Signed-off-by: Johan Hovold <johan@kernel.org>
On Power8 & Power9 the early CPU inititialisation in __init_HFSCR()
turns on HFSCR[TM] (Hypervisor Facility Status and Control Register
[Transactional Memory]), but that doesn't take into account that TM
might be disabled by CPU features, or disabled by the kernel being built
with CONFIG_PPC_TRANSACTIONAL_MEM=n.
So later in boot, when we have setup the CPU features, clear HSCR[TM] if
the TM CPU feature has been disabled. We use CPU_FTR_TM_COMP to account
for the CONFIG_PPC_TRANSACTIONAL_MEM=n case.
Without this a KVM guest might try use TM, even if told not to, and
cause an oops in the host kernel. Typically the oops is seen in
__kvmppc_vcore_entry() and may or may not be fatal to the host, but is
always bad news.
In practice all shipping CPU revisions do support TM, and all host
kernels we are aware of build with TM support enabled, so no one should
actually be able to hit this in the wild.
Fixes: 2a3563b023 ("powerpc: Setup in HFSCR for POWER8")
Cc: stable@vger.kernel.org # v3.10+
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Tested-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
[mpe: Rewrite change log with input from Sam, add Fixes/stable]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>