DNM: Ble stack and init #20205
DNM: Ble stack and init #20205
Conversation
frkv
requested review from
a team,
trantanen,
tokangas,
carlescufi,
Frodevan,
adamkondraciuk,
lemrey,
wentong-li,
bama-nordic,
gmarull,
D-Triveni,
muraliThokala and
krish2718
as code owners
BLE logging size requirements have increased a lot. We need to reduce logging to fit the application. Signed-off-by: Maximilian Deubel <[email protected]>
Switched from speed to size optimization for nRF54H20 DK target in debug configurations to reduce memory footprint. Jira: NCSDK-31374 Signed-off-by: Mateusz Kapala <[email protected]>
Switched from speed to size optimizations for nRF54L15 DK (nRF54L05 emulated SoC) and enabled LTO in debug and release configurations to reduce memory footprint. Jira: NCSDK-31374 Signed-off-by: Mateusz Kapala <[email protected]>
Remove reference of z_impl_sys_rand_get(). Signed-off-by: Eivind Jølsgard <[email protected]>
Updates required by changes in zephyr upstream for the BT HCI driver. Signed-off-by: Tommi Rantanen <[email protected]>
Update pm policy event usage after API was changed from using relative time in cycles to absolute time in ticks. Signed-off-by: Knut Eldhuset <[email protected]>
This test is not really useful in NCS context (checking if the nRF Wi-Fi driver can be built successfully is done by other tests and samples), and it requires the hal_atmel module to be pulled in (because the WINC1500 driver requires two header files that are located in that module). And the module weights ~87 MB, so it's not reasonable to pull it in just for this one test. Signed-off-by: Andrzej Głąbek <[email protected]>
Fixes an issue with stream writer by providing the size of the partition Signed-off-by: Jamie McCrae <[email protected]>
Change posix names to zsock in nrf_cloud lib. Signed-off-by: Maximilian Deubel <[email protected]>
Some sdk-zephyr tests and samples reports issues due to RAM overflow issues. Refers: NCSDK-31667 NCSDK-31671 Signed-off-by: Katarzyna Giądła <[email protected]>
Applications `applications.nrf_desktop.zrelease` and `applications.nrf_desktop.zdebug` report RAM overflow. Therefore, applications on listed platforms are moved to quaratine. Refers: NCSDK-31675 Signed-off-by: Katarzyna Giądła <[email protected]>
PSA requires large stacks, especially on 54. Signed-off-by: Sean Madigan <[email protected]> Signed-off-by: Frank Audun Kvamtrø <[email protected]>
Test configurations with secure_storage should be move to quarantine due to issues on native_sim platform. Signed-off-by: Katarzyna Giądła <[email protected]>
Clang is not available in our toolchain, therefore test kernel.common.toolchain will fail. Signed-off-by: Katarzyna Giądła <[email protected]>
`nanopb_BUILD_RUNTIME` is not an option to be passed to nanopb, but rather a cmake one, so set it to off in the proper way. The previous invalid form was not working as intended and only by chance not causing any problems. With a correction that was applied in the nanopb module (removal of leading space before the first -I directive passed to nanopb), this behavior changed and because of that the generation could not be performed correctly. Signed-off-by: Andrzej Głąbek <[email protected]>
Use zsock_addrinfo to avoid dependency on the POSIX subsystem. Remove CONFIG_POSIX_API from provisioning tests. Signed-off-by: Juha Ylinen <[email protected]>
... because this causes conflicting types compilation errors in library files it uses. Signed-off-by: Andrzej Głąbek <[email protected]>
Test scenarios `shell.device_filter` and `init.check_init_priorities` should be fixed in Zephyr first. Signed-off-by: Katarzyna Giądła <[email protected]>
-This added SYS_INIT will ensure that psa_crypto_init is called in kernel *after* Hw_CC3XX is executed to ensure that mutexes are ready to be used. Signed-off-by: Frank Audun Kvamtrø <[email protected]>
-Changing default y for nrf_security to use DT_HAS_ZEPHYR_PSA_CRYPTO_RNG_ENABLED instead of depending on ENTROPY_PSA_CRYPTO_RNG. Doing this avoids a dependency loop on ENTROPY_GENERATOR Signed-off-by: Frank Audun Kvamtrø <[email protected]>
-This commit is only used to enact testing. The dependency for PR nrfconnect#2464 to the previous commit is inherent, but we would like to have a commit that is pickable in the upmerge Signed-off-by: Frank Audun Kvamtrø <[email protected]>
This is required by PSA in NCS, so make sure it is selected NRF_SECURITY_ENABLER. Signed-off-by: Sean Madigan <[email protected]>
-Enabling ENTROPY_GENERATOR for BT_HOST_CRYPTO_PRNG and BT_ECC as a workaround of having ENTROPY_GENERATOR in sample/tests Also the heap size required for bsim is larger than on actual target, reason unknown. Signed-off-by: Sean Madigan <[email protected]> Signed-off-by: Frank Audun Kvamtrø <[email protected]>
The NRF_RNG_ENTROPY_DRIVER provides entropy though the PSA APIs. This is used to seed/reseed the Oberon PSA random drivers (CTR_DRBG/HMAC_DRBG). Before this was only used to allow enabling this PSA driver with the devices which use the DT label 'rng'. We extend that now to allow this driver to be used with the fake entropy node which uses the DT label 'prng'. This is needed to unblock development until real RNG is enabled for all the needed applications/samples. In practice this means that when 'prng' is enabled a device can generate random numbers with the psa_generate_random call. Signed-off-by: Georgios Vasilakis <[email protected]>
-Adding CONFIG_MBEDTLS_FORCE_LEGACY_MD for cc3xx/oberon legacy psa_tls to avoid that MBEDTLS_MD_SOME_PSA is getting set Signed-off-by: Frank Audun Kvamtrø <[email protected]>
-Fixes issues with nRF54H20 and nRF5340dk in upmerge. This can be seen as a workaround -Adjusting board-files for central_uart and peripheral_uart for nRF5340 and nRFH20 cpuapp Signed-off-by: Frank Audun Kvamtrø <[email protected]>
-HKDF reports errors when MBEDTLS_MD_SOME_PSA is set. Setting this configuration to avoid trying to resolve with PSA APIs. This configurations and these tests to be removed with more standardization on PSA crypto Signed-off-by: Frank Audun Kvamtrø <[email protected]>
Increased stack sizes to prevent the stack overflows. Jira: NCSDK-31374 Signed-off-by: Mateusz Kapala <[email protected]>
frkv
force-pushed
the
ble_stack_and_init
branch
from
7087cfb to
8fa454b
Compare
Memory footprint analysis revealed the following potential issuesapplications.nrf_desktop.zdebug.usb_next[nrf52840gmouse/nrf52840]: RAM size increased by 5900[B] in comparison to the main[a3b7e57] branch. - link (cc: @MarekPieta) Note: This message is automatically posted and updated by the CI (latest/sdk-nrf/PR-20205/16) |
Just for test
Just for testing. Changes will be taken in #19720 if this works
test_low_level: PR-1765
test_rs: PR-1483
test_fem: PR-1483
test_ble_mesh: ble_mesh_upmerge_test
test_crypto: PR-750
test_suit_dfu: PR-450
test_tfm: PR-194
test_secdom_samples_public: NCSDK-NONE_upmerge_jan2025
test_audio: PR-294
test_wifi: sdk-nrf-pr-19720