By acking a received message we tell the DSP that we have processed the
message (reply or notification) and we are open to receive a new one.
The original implementation did this in a common code after the received
message got handled as reply or notification.
With right timing this opens up a small window when we have processed the
reply and let the other thread proceed to send a new message to the DSP,
which is allowed as the DSP is free to receive message.
But when the message is received and processed by the DSP and it wants to
send a reply it will still see that the previous message has not been
acked, so it fails to send a reply. Later the first reply got acked by the
kernel, but it is too late and the in-flight message got a timeout due to
firmware not responding (which it tried, but could not).
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Link: https://lore.kernel.org/r/20221018124008.6846-5-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Before attempting to send a message to the DSP we need to check if the
downstream BUSY flag has been cleared by the firmware to avoid lost IPC
messages by the firmware.
This is required by a firmware which only acks the received message after
it has sent a reply to the host.
With a bad luck, the host would send a message before the firmware gets to
the clearing the flag and thus losing a message.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Link: https://lore.kernel.org/r/20221018124008.6846-4-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Read both registers as the first step in the interrupt handler to make
sure that we are handling the event which triggered the interrupt.
The delayed reading of the target request register might reflect incorrect
information about the reason why the interrupt was risen.
Note also that the IPC3 interrupt handler is implemented in this way also.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Link: https://lore.kernel.org/r/20221018124008.6846-3-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
In preparation of the IPCv4 IPC support, this patch adds
support for SkyLake and KabyLake boot and descriptors
used when probing the PCI driver.
The work was initially contributed in 2018 by Liam Girdwood and Zhu
Yingjiang, and abandoned due to firmware signature issues. With the
upcoming support of IPC v4, and hence the Intel closed-source
firmware, it's time to re-add this capability.
The SKL ops will be added in the next patch.
Tested with the IPC4 and closed-source firmware on Dell XPS 9350
and KBL NUC with HDaudio codecs. The SSP and DMIC interfaces are not
supported at this time.
Co-developed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20220920131700.133103-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch replaces all dev_vdbg calls with tracepoints to reduce
overhead and enable use of trace collection and analysis tools.
Reviewed-by: Péter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Noah Klayman <noah.klayman@intel.com>
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20220919122108.43764-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
It is not yet clear, but it is possible to create a firmware so broken
that it will send a reply message before a FW_READY message (it is not
yet clear if FW_READY will arrive later).
Since the reply_data is allocated only after the FW_READY message, this
will lead to a NULL pointer dereference if not filtered out.
The issue was reported with IPC4 firmware but the same condition is present
for IPC3.
Reported-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20220712122357.31282-3-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add implementation of low level, platform dependent IPC4 message handling
and set the DSP ops for IPC4 for APL, CNL and TGL platforms.
Co-developed-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20220511171648.1622993-2-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Switch from the IPC dependent ipc_pcm_params() ops to the IPC neutral
set_stream_data_offset().
Remove the no longer used hda_ipc_pcm_params() function as well.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20220310042720.976809-9-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add implementation for the generic set_stream_data_offset() callback to be
used by HDA platforms.
Convert the hda_ipc_pcm_params() to a wrapper for the new function.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Signed-off-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20220310042720.976809-8-ranjani.sridharan@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The existing code maximizes confusion by using 'stream' and 'hstream'
variables of different types, e.g:
struct hdac_stream *stream;
struct hdac_ext_stream *stream;
struct hdac_stream *hstream;
struct hdac_ext_stream *hstream;
This confusion is partly inherited from legacy code but SOF
contributors added their own creative spin, e.g.
struct hdac_ext_stream *link_dev;
struct hdac_ext_stream *dsp_stream;
struct hdac_ext_stream hda_stream;
and my personal favorite:
stream = &hda_stream->hda_stream;
This patch suggests a consistent naming across all Intel code related
to HDAudio stream management. The convention is - by hierarchical
order:
struct sof_intel_hda_stream *hda_stream;
struct hdac_ext_stream *hext_stream;
struct hdac_stream *hstream;
No functionality change - just renaming of variables/members.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20220209063104.9971-1-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Some platforms use retries during firmware boot to overcome DSP startup
issues.
In these cases we might receive a DSP panic message which should not be
treated as fatal if it happens during boot.
Pass this information to snd_sof_dsp_panic() and omit the panic print if
it is not fatal or the user does not want to see all dumps.
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20211223113628.18582-6-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Make use of the generic snd_sof_ipc_process_reply() from the core instead
the local implementation.
snd_sof_ipc_process_reply() handles the reply retrieving and the ipc reply
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Reviewed-by: Rander Wang <rander.wang@intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Signed-off-by: Daniel Baluta <daniel.baluta@nxp.com>
Link: https://lore.kernel.org/r/20211116152137.52129-4-daniel.baluta@oss.nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
If an invalid stream is passed to snd_sof_ipc_msg_data() it won't
fill the provided object with data. The caller has to be able to
recognise such cases to avoid handling invalid data. Make the
function return an error when failing.
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Reviewed-by: Peter Ujfalusi <peter.ujfalusi@intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@linux.intel.com>
Link: https://lore.kernel.org/r/20210928103516.8066-1-peter.ujfalusi@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Checking that two values don't have common bits makes no sense,
strict equality is meant.
Fixes: f3b433e469 ("ASoC: SOF: Implement Probe IPC API")
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20210802151749.15417-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Remove the ambiguity with GPL-2.0 and use an explicit GPL-2.0-only
tag.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Daniel Baluta <daniel.baluta@nxp.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Link: https://lore.kernel.org/r/20200501145850.15178-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The HDA_DSP_IPC_PURGE_FW IPC from ROM is already handled in
cl_dsp_init(), and it will never be received in the IRQ thread,
so the wait condition on this IPC will never be satisfied. The
wait before loading firmware is redundant and can be removed safely.
Signed-off-by: Amery Song <chao.song@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Keyon Jie <yang.jie@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Pierre Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Link: https://lore.kernel.org/r/20200312200622.24477-6-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The HDA_DSP_IPC_PURGE_FW IPC from ROM is already handled in
cl_dsp_init(), and as IPC IRQ is disabled at this stage, this
IPC will be never received in the IRQ thread. The function
hda_dsp_ipc_is_sof for filtering the ROM IPC can be removed
safely.
Signed-off-by: Amery Song <chao.song@intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Reviewed-by: Keyon <yang.jie@intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Link: https://lore.kernel.org/r/20200312200622.24477-5-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add all required types and methods to support each and every request
that driver could sent to firmware. Probe is one of SOF firmware
features which allows for data extraction and injection directly from
or to DMA stream.
Exposes eight IPCs:
- addition and removal of injection DMAs
- addition and removal of probe points
- info retrieval of injection DMAs and probe points
- probe initialization and cleanup
Signed-off-by: Cezary Rojewski <cezary.rojewski@intel.com>
Acked-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20200218143924.10565-5-cezary.rojewski@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The existing code uses two handlers for a shared edge-based MSI interrupts.
In corner cases, interrupts are lost, leading to IPC timeouts. Those
timeouts do not appear in legacy mode.
This patch merges the two handlers and threads into a single one, and
simplifies the mask/unmask operations by using a single top-level mask
(Global Interrupt Enable). The handler only checks for interrupt
sources using the Global Interrupt Status (GIS) field, and all the
actual work happens in the thread. This also enables us to remove the
use of spin locks. Stream events are prioritized over IPC ones.
This patch was tested with HDaudio and SoundWire platforms, and all
known IPC timeout issues are solved in MSI mode. The
SoundWire-specific patches will be provided in follow-up patches,
where the SoundWire interrupts are handled in the same thread as IPC
and stream interrupts.
Signed-off-by: Bard Liao <yung-chuan.liao@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20191204212859.13239-1-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Memory windows could be powered off before receiving PM_GATE IPC reply
from FW, we can't read the mailbox to get reply.
Signed-off-by: Keyon Jie <yang.jie@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20191025224122.7718-15-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
We can use generic sof_fw_ready function and reduce code duplication.
Careful here that we need to provide the implementation for
get_mailbox_offset and get_window_offset.
Signed-off-by: Daniel Baluta <daniel.baluta@nxp.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190807150203.26359-7-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Downgrade "nothing to do in IRQ thread" message from error to a debug
message in the IPC interrupt handler thread.
The spurious wake-up can happen if a HDA stream interrupt is
raised while the IPC interrupt thread is running. IPC functionality
is not impacted by this condition, so debug is a more appropriate
trace level.
Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Link: https://lore.kernel.org/r/20190722141402.7194-21-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The HIPCCTL register controls the IPC interrupts. It can be set or
cleared to mask or enable these interrupts, but it makes no sense to
read and test its fields in an interrupt (which can only executed if
its fields are set).
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Align with Skylake driver and enable the IRQ at end of handler,
instead of at beginning.
Also add an error log if we have nothing to do in this handler.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Align with hardware recommended sequences, and read all IPC registers
before doing any other actions. Playing with BUSY and DONE bits may
invalidate values.
The values read may not actually be necessary but at least this
provides a snapshot of the IPC registers with no consistency issues.
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
We use shim registers only to notify the other
side that a message was sent. The actual information
for the message is transmitted via mailbox.
cmd information inside shim register is not used by
the DSP, so we remove it to avoid confusion.
Signed-off-by: Daniel Baluta <daniel.baluta@gmail.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Currently on all supported platforms the IPC IRQ thread first signals
the sender when an IPC response is received from the DSP, then unmasks
the IPC interrupt. Those actions are performed without holding any
locks, so the thread can be interrupted between them. IPC timeouts
have been observed in such scenarios: if the sender is woken up and it
proceeds with sending the next message without unmasking the IPC
interrupt, it can miss the next response. This patch takes a spin-lock
to prevent the IRQ thread from being preempted at that point. It also
makes sure, that the next IPC transmission by the host cannot take
place before the IRQ thread has finished updating all the required IPC
registers.
Fixes: 53e0c72d98 ("ASoC: SOF: Add support for IPC IO between DSP and Host")
Reviewed-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
When using a shared IRQ between IPC interrupt and stream IOC interrupt,
the interrupt handlers need to check the interrupt source before
scheduling their respective IRQ threads. In the case of IPC handler, it
should check if it is an IPC interrupt before waking up the IPC IRQ
thread.
The IPC IRQ thread, once scheduled, does not need to check the IRQ
source again. So, remove the superfluous check in the thread. Remove the
irq_status field from snd_sof_dev struct also as it is no longer needed.
Signed-off-by: Keyon Jie <yang.jie@linux.intel.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Peter Ujfalusi