This is part of a series of moving memory management related
stuff out of the Z_ namespace and into its own namespace.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The board defconfig forces CONFIG_USERSPACE to be disabled
due to small number of available MPU regions which are not
enough for userspace testing.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Also any demand paging and page frame related bits are
renamed.
This is part of a series to move memory management related
stuff out of the Z_ namespace into its own namespace.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This is part of a series to move memory management related
stuff from Z_ namespace into its own namespace.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Renames:
Z_KERNEL_VIRT_START to K_MEM_KERNEL_VIRT_START
Z_KERNEL_VIRT_SIZE to K_MEM_KERNEL_VIRT_SIZE
Z_KERNEL_VIRT_END to K_MEM_KERNEL_VIRT_END
This is part of a series to move memory management related
stuff from Z_ namespace into its own namespace.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Renames:
Z_PHYS_RAM_START to K_MEM_PHYS_RAM_START
Z_PHYS_RAM_SIZE to K_MEM_PHYS_RAM_SIZE
Z_PHYS_RAM_END to K_MEM_PHYS_RAM_END
This is part of a series to move memory management related
stuff from Z_ namespace into its own namespace.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This is part of a series to move memory management functions
away from the z_ namespace and into its own namespace. Also
make documentation available via doxygen.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This renames z_phys_map() and z_phys_unmap() to
k_mem_map_phys_bare() and k_mem_unmap_phys_bare()
respectively. This is part of the series to move memory
management functions away from the z_ namespace.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Add tests for
DT_REG_HAS_NAME,
DT_REG_ADDR_BY_NAME_OR,
DT_REG_SIZE_BY_NAME_OR,
DT_INST_REG_HAS_NAME,
DT_INST_REG_SIZE_BY_NAME_OR and
DT_INST_REG_ADDR_BY_NAME_OR.
Signed-off-by: Fin Maaß <f.maass@vogl-electronic.com>
In arm64 linker script, create a memory region and section for each
device tree node with compatible string "zephyr,memory-region".
Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
The 3-byte value suits the assigned number much better,
and also allows for less memory copies when getting and
setting the values.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Remove the "stream" part of the value and functions to
better fit with the name in the assigned numbers document.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
This PR adds support for new Synopsys nSIM RMX100 platform.
New platform based on RISC-V ISA instead of classic ARC.
Signed-off-by: Nikolay Agishev <agishev@synopsys.com>
Change source tree organization for Synopsys nSIM platform.
Classical ARC architectures arc_v2 arc_v3 moved to arc_classic
SoC and boards family.
nSIM SoCs were separated regarding series: EM, HS, SEM, VPX2.
This PR sould be seeing as the preparation for
addition new nSIM platform based on the RISC-V architecture.
Signed-off-by: Nikolay Agishev <agishev@synopsys.com>
Rename all Silicon Labs radio boards to always use the official kit name of
the kit containing only the radio board as board target name. Disentangle
boards using Series 1 and Series 2 SoCs that have almost nothing in common
in terms of devicetree content, but still occupied the same directory.
As part of the initial HWMv2 port, all radio boards were merged into a
single target efr32_radio, using the SoC as a board qualifier to target
specific boards. This was unfortunate, as the SoC name is not sufficient
to identify a specific board -- the same SoC may be used on multiple
different boards. This commit reinstates every radio board as an individual
board target.
Signed-off-by: Aksel Skauge Mellbye <aksel.mellbye@silabs.com>
Rename all Silicon Labs dev kits to always use the official kit name
as board target name. Previous kit names used various naming schemes,
including putting part of the SoC name in the board name. With HWMv2,
SoC names (if needed) should go in the SoC board qualifier.
Use HWMv2 revisions for the two variants of SLTB010A.
Split the xG27 Dev Kit from the EFR32BG22 Thunderboard. This attempt
at deduplication is confusing due to the very different kit names
and use of different ICs (BG22 vs BG27), and was not continued with
the xG24 Dev Kit, meaning that it wasn't consistent.
Signed-off-by: Aksel Skauge Mellbye <aksel.mellbye@silabs.com>
We don't need the TX thread anymore.
Generalizing the pull-based architecture (ie. `tx_processor`) to HCI
commands makes it possible to run the whole TX path from the the system
workqueue, or any workqueue really.
There is an edge-case, where we call `bt_hci_cmd_send_sync()` from the
syswq, stalling the system. The proposed mitigation is to attempt to drain
the command queue from within `bt_hci_cmd_send_sync()`.
My spidey sense tingles however, and it would be better to just remove the
capability of calling this fn from the syswq. But doing this requires
refactoring a bunch of synchronous procedures in the stack (e.g. stack
init, connection establishment, address setting etc), dragging in more
work. I will do it, but in a later patch.
Signed-off-by: Jonathan Rico <jonathan.rico@nordicsemi.no>
The current TX pattern in the host is to try to push a buffer through all
the layers up until it is ingested by the controller.
Since sending can fail at any layer, we need error-handling and separate
retry logic on pretty much all layers. That logic obscures the "happy path"
for people trying ot understand the code.
This commit inverts the control, in a way that doesn't require changing the
host or HCI driver API (yet):
Layers don't send buffers synchronously, they instead put their buffer in a
private queue of their own and raise a TX flag on the lower layer. Think of
it as a `READY` interrupt line that has to be serviced by the lower layer.
Sending is now non-blocking, rate depends on the size of buffer pools.
There is a single TX processing function. This can be thought as the
Interrupt Service Routine that will handle the `READY` interrupt from the
layers above.
That `tx_processor()` will then attempt to allocate enough resources in
order to send the buffer through to the controller. This allocation logic
does not block.
After acquiring all the resources, the TX processor will attempt to pull
data from the upper layer. The upper layer has to figure out which buffer
to pass to the controller. This is a good spot to put scheduling or QoS
logic in the upper layer.
Notes:
- user-facing API for tuning QoS will be implemented in a future patch
- this scheme could (and probably will) be extended to upper layers (e.g.
ATT, L2CAP CoC segmentation).
- this patch removes the `pending_no_cb()` memory optimization for
clarity/correctness. It might get re-implemented after a stabilization
period. Hopefully with more documentation.
Signed-off-by: Jonathan Rico <jonathan.rico@nordicsemi.no>
Co-authored-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
Instead of allocating segments/fragments and copying data into them, we
allocate segments as "views" (or slices) into the original buffer.
The view also gives access to the headroom of the original buffer, allowing
lower layers to push their headers.
We choose not to allow multiple views into the same buffer as the headroom
of a view would overlap with the data of the previous view.
We mark a buffer as locked (or "in-view") by temporarily setting its
headroom to zero. This effectively stops create_view because the requested
headroom is not available.
Each layer that does some kind of fragmentation and wants to use views for
that needs to maintain a buffer pool (bufsize 0, count = max views) and a
metadata array (size = max views) for the view mechanism to work.
Maximum number of views: number of parallel buffers from the upper layer,
e.g. number of L2CAP channels for L2CAP segmentation or number of ACL
connections for HCI fragmentation.
Reason for the change:
1. prevent deadlocks or (ATT/SMP) requests timing out
2. save time (zero-copy)
3. save memory (gets rid of frag pools)
L2CAP CoC: would either allocate from the `alloc_seg` application callback,
or worse _steal_ from the same pool, or allocate from the global ACL pool.
Conn/HCI: would either allocate from `frag_pool` or the global ACL pool.
Signed-off-by: Jonathan Rico <jonathan.rico@nordicsemi.no>
Co-authored-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
Add the following new macros:
- DT_FOREACH_NODELABEL
- DT_FOREACH_NODELABEL_VARGS
- DT_INST_FOREACH_NODELABEL
- DT_INST_FOREACH_NODELABEL_VARGS
These are for-each helpers for iterating over the node labels of a
devicetree node. Since node labels are unique in the entire
devicetree, their token representations can be useful as unique IDs in
code as well.
As a first user of these, add:
- DT_NODELABEL_STRING_ARRAY
- DT_INST_NODELABEL_STRING_ARRAY
The motivating use case for these macros is to allow looking up a
struct device by devicetree node label in Zephyr shell utilities.
The work on the shells themselves is deferred to other patches.
To make working with the string array helpers easier, add:
- DT_NUM_NODELABELS
- DT_INST_NUM_NODELABELS
Signed-off-by: Martí Bolívar <mbolivar@amperecomputing.com>
Restructures the latency_measure benchmark configurations. Most of them
do not need to be executed by default by CI. However, we still want to
make it easy to run the benchmark with various configurations so that
they can be easily compared against the default.
The updated documentation shows how this can be done.
Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
For sake of memory footprint it might be required to shrink
down the memory footprint as much as possible. Unfortunately Mbed TLS
PSA interface brings in some extra code for key management which makes
it larger than the TinyCrypt counterpart when it comes to p256-m
interfacing. For this reason it might be useful to directly access
the p256-m driver directly.
This commit adds this support and it also updates the corresponding
test in order to make use of this condition.
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
Add a simple test to compare the footprint of MbedTLS vs
TinyCrypt for the secp256r1 curve. On the MbedTLS side
the test uses the P256M driver developed by the MbedTLS
team. This is an optimized implementation which targets
the secp256r1 curve.
The example itself is pretty easy:
- generate 2 EC keys
- perform key agreement
Signed-off-by: Valerio Setti <vsetti@baylibre.com>
Since commit 0c23cf94a4 ("include/kernel: check type of object passed
to K_POLL_EVENT_INITIALIZER") it is no longer possible to use arbitrary
type literals with the initializer macro. Misconstructing a k_poll_event
type for test purpose must be done explicitly.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
Update references to devicetree chosen properties and how devicetree nodes
should be defined in order to enable the new HCI drivers.
Signed-off-by: Johan Hedberg <johan.hedberg@silabs.com>
Convert all Bluetooth tests that define dummy HCI driver instances to use
the new HCI driver API. This requires both a custom DTS binding as well
as an app-specific overlay file.
Signed-off-by: Johan Hedberg <johan.hedberg@gmail.com>
Add support for using an UART instead of the native controller. This is
accomplished with a custom DTS overlay for the tests that require it.
Signed-off-by: Johan Hedberg <johan.hedberg@gmail.com>
Convert the H4 driver to the new HCI driver API. This includes updating
also any boards that use the driver, i.e. adding the appropriate
devicetree node and chosen property to them.
Signed-off-by: Johan Hedberg <johan.hedberg@gmail.com>
- Tests that the API of the uart-emul also works with the async API
- Tests that `uart_emul_set_release_buffer_on_timeout()` configures the
behavior as expected
- The tests are now splited into three different configurations: only
polling API enabled, additionally interrupt driven API enabled and
additionally async API enabled
- Enabled `native_sim` board
Signed-off-by: Ruben Völl <ruben.voell@grandcentrix.net>
Remove nSIM platform filter for tests/lib/heap. The filter is no
longer needed as nSIM performence was improved and test timeout
was increased because of another platforms.
Signed-off-by: Nikolay Agishev <agishev@synopsys.com>
Refactored the unicast audio client sample to use a
separate thread for TXing, rather than the system workqueue.
This allows us to do blocking waiting operations like
buf = net_buf_alloc(&tx_pool, K_FOREVER);
and also splits the responsibility of TXing to a new
file as well.
LC3 handling have also been moved to a new file, so
that it does not pollute the source code for non-LC3
supported builds. This also fixes various issues and
improves the LC3 handling in the sample.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
ull_adv_sync_pdu_set_clear does pretty much everything, causing
it to be very complex and awkward to use. In addition, it only handles
a single PDU, meaning callers have to handle the chain.
It has been replaced with simpler, more complete functions for handling
the relevant operations
Fixed issues include:
- Fragmentation of adv data over HCI is now decoupled from PDU
fragmentation, fixing HCI/DDI/BI-13-C, LL/DDI/ADV/BV-26-C and
LL/DDI/ADV/BV-55-C
- Adding BigInfo now preserves the PDU chain
- Enabling periodic advertising with ADI on would sometimes fail
due to insufficient space in a single PDU to add ADI
Signed-off-by: Troels Nilsson <trnn@demant.com>
For slow processors it is possible that setting 1 ms alarm and
then canceling it could lead to alarm expiring before it is canceled
and that was not expected in the test. Increasing timeout to 1.5ms.
Signed-off-by: Krzysztof Chruściński <krzysztof.chruscinski@nordicsemi.no>
Distinguish between test case dependencies where a board has a test
feature "usbd" and where no dependency is required other than the
devicetree overlay file, as here to build the UDC skeleton driver.
Signed-off-by: Johann Fischer <johann.fischer@nordicsemi.no>
