This commit controls the stream busy flag of the dma channnel
Set as true : stream is busy each time the channel is
started or reloaded.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
When the channel is overriden by the HAL,
the dma irq must always be handled (even if not busy).
Only when the dma channel is not overriden by the HAL
the irq is exiting when the channel is no more busy.
Signed-off-by: Francois Ramu <francois.ramu@st.com>
SoCs can have different headers, make the build system include those if
the directory exists for this SoC.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Intel ADSP CAVS is now a proper series with all CAVS SoCs running under
it. This will give us to Intel ADSP series:
- CAVS
- ACE v1.x
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The simulator seems to drop garbage addresses (somewhere in the ROM it
looks like) into this SR at arbitrary times. I don't know if this is
a hardware exception handler that we can't turn off, or a simulator
bug, or what. But our code that assumes it will be cleared to zero or
valid is breaking. Set it every time in every context switch for now
pending someone figuring out what's going wrong.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
There is no easy way to clear event bits without
the potential for a race to exist between producer(s)
and consumer(s). The result of this race is that events
can be lost through the various resetting mechanisms
available (flag to k_event_wait(), or k_event_set()).
Add k_event_set_masked() which permits bits to be set or cleared.
This allows consumers to clear just the bits that they have read
without (accidentally) discarding any new bits.
Update unit tests to verify the functionality.
Partly Fixes#46117.
Signed-off-by: Andrew Jackson <andrew.jackson@amd.com>
Although there is nothing wrong with the existing code,
it doesn't permit individual bits to be set (or cleared).
This makes further changes slightly awkward.
Use a mask to restrict the bits set in an event.
Signed-off-by: Andrew Jackson <andrew.jackson@amd.com>
Moves sof to the module area, based on that also remove maintainer, as
this was associated with SOF.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This is not a sample of an audio subsystem in zephyr, it is a sample
that completely depends on sof and its audio support.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
The uart_pipe_isr() double-checks whether rx interrupt occured by
calling uart_irq_is_pending() first. Get rid of that function call since
it is redundant (i.e. nothing more is done or checked inside that
conditional).
Signed-off-by: Jan Kuliga <jtkuliga@gmail.com>
In order to coredump dump the ram memory, it is necessary that we define
in the linker scripts two variable to indicate the start and the end of
the ram area. Adding these variables to ace linker script.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
In order to coredump dump the ram memory, it is necessary that we define
in the linker scripts two variable to indicate the start and the end of
the ram area. Adding these variables to cavs linker script.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
cavs/version.h was never used by Zephyr, it was supposed to be needed by
SoF. This no longer seems to be the case, just removing it.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
Sensor Kconfig sybmols should be enabled if CONFIG_SENSOR=y
and the devicetree node for the sensor is enabled. We can
remove explicitly enabling specific sensor drivers in .conf
files.
Signed-off-by: Kumar Gala <galak@kernel.org>
The `binding_filename` function, which resolves the binding
documentation file name from the binding yaml file name, made a very
crude assumption that the binding yaml file name will always have the
extension name of `yaml` -- this is no longer true after the commit
c453220006, which added support for both
`yaml` or `yml`.
This commit reworks the function such that its implementation is not
dependent on the extension name of the binding yaml file.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
The VREFINT and TEMPSENSOR internal channels have a stabilization time
after they are enabled. Right now this just causes the first measure to
be a bit off, however with PR #47691 which stops the internal channels
after each readout, this is something important to respect. Fortunately
the stabilization time is available as constants in the HAL, so just
wait the time specified by the constants.
Note that the VBAT internal channel does not have any stabilization
time.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Due to added call to ull_cp_tx_nrf call in common function event_done,
encryption test test_encryption_start_periph_rem_no_ltk failed with
assert caused by unexpected number of free control procedure contexts.
That is caused by internal call to rr_check_done and lr_check_done
in the ull_cp_tx_nrf that frees last unreleased context.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
Unit tests reposnsible for verification of collisions mitigation
didn't have handling for delayed completion of PHY Update control
procedure. The PHY Update is completed on event generated when
"at-instant" connection event is on air.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
The PHY Update control procedure completion is delayed to event
"instant-on-air". That has to be handled in CTE Request control
procedure unit test.
The changes required are:
- Added handling of event done for "at-instant" connection event.
This is required to call ull_cp_tx_ntf function to send an
envet "instant-on-air" to PHY Update procedure state machine.
This is the delayed completion of PHY Update procedure.
- Remote requested PHY Update completion is also delayed and completed
on "instant-on-air" event, hence the local CTE Request is not executed
at the same connection event as PHY Update completion. This caused
changes to check_phy_update_and_cte_req_complete function. The functin
does not hanle CTE Request completion anymore. This also implies
changes to run_phy_update_central and run_phy_update_peripheral.
These function do not need to get CTE request argument anymore
because it was just passed to check_phy_update_and_cte_req_complete.
The check_phy_update_cte_req_complete was renamed to check_phy_update.
The test test_cte_req_wait_for_remote_phy_update_complete was changed
to execute and verify local CTE request procedure explicitly after
completion of PHY Update.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
The test_phy_update_central_rem_collision didn't have processing
of an "at-instant" connection event. Thas causes the remote requesed
PHY update to hand in waiting for instant on air state and whole
test to fail.
To fix the problem, proper processing of "at-instant" connection evnet
has to be added. In consequence the local PHY update procedure, that
as waiting for end of the remote procedure is delayed by one connection
event.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
Unit tests does not maintain connection events hence there is no
notification to PHY upodate state machine about instant on air
event. That has to be simulated by explicit call to ull_cp_tx_nrf.
In unit tests the event_done function is responsible for handling
all steps done at end of connection event, so it is the best place
to simulate the instant on air event and send it to LLCP state machine.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
PHY update control procedure has to send a notification towards
Host when the PHY is actually changed, when the instant happens.
Control procedures are handled in a prepare phase of a connection
event, hence notifications are send towards host to early.
The fix for that is to postpone send of a notification.
New state has been introduced to remote requested PHY update
control procedure: RP_PU_STATE_WAIT_INSTANT_ON_AIR.
Also new event has been introduced: RP_PU_EVT_NTF, that is used
to inform the PHY update state machine that notification may
be send.
There are two places where the notification events may be generated
ull_conn_rx, just before first received PDU is send towards Host
or ull_conn_done, in case there were no PDU received.
Signed-off-by: Piotr Pryga <piotr.pryga@nordicsemi.no>
HSE is a valid alternate clock source for these series,
therefore STM32_SRC_HSE is added in this commit.
Both series can use HSE as an alternate clock source for USB and FDCAN.
Signed-off-by: Thomas Stranger <thomas.stranger@outlook.com>
EPD displays currently have two compatible strings, their specific
model number and the generic driver name. This isn't ideal since we
will need to match on the driver chip model number to work around
subtle chip differences in some cases. Add the specific chip model to
the list of compatibles.
Signed-off-by: Andreas Sandberg <andreas@sandberg.uk>
GoodDisplay used to hide the real vendor of the driver chip in some of
their old data sheets. This is no longer the case and the actual chip
version can be found on their official site. Update the documentation
to show which chip is actually used.
Signed-off-by: Andreas Sandberg <andreas@sandberg.uk>
Replace Kconfig (IVSHMEM_DEV_NAME) based named device_get_binding
with DEVICE_DT_GET_ONE. Since there is only one driver for ivshmem
use the qemu,ivshmem for that.
Signed-off-by: Kumar Gala <galak@kernel.org>
Migrate the BT tests for host_long_adv_recv to use FFF instead of
ztest mock which is being deprecated.
Signed-off-by: Yuval Peress <peress@google.com>
