libc(picolibc): fix locks.c K_MUTEX and _LOCK_T incompatibilities (clean)#53
Closed
bhoot1234567890 wants to merge 10000 commits intozmkfirmware:mainfrom
Closed
libc(picolibc): fix locks.c K_MUTEX and _LOCK_T incompatibilities (clean)#53bhoot1234567890 wants to merge 10000 commits intozmkfirmware:mainfrom
bhoot1234567890 wants to merge 10000 commits intozmkfirmware:mainfrom
Conversation
Use the host defined roles, rather than the HCI defined ones. Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Since the entire cid checking function is conditionally included based on KConfig, the usage site also needs to use #if instead of IS_ENABLED, otherwise you get the following warning during compilation when the option is disabled: `error: implicit declaration of function ‘bt_mesh_vnd_mod_msg_cid_check’` Signed-off-by: Ludvig Jordet <ludvig.jordet@nordicsemi.no>
This deprecates CONFIG_BT_MESH_MODEL_VND_MSG_CID_FORCE. This option was added to allow an optimization that a) only can apply in edge case networks that use multiple vendor models with different CIDs, b) doesn't really have much potential gains even in this case unless there is a large number of opcodes in these vendor models, and c) makes it impractical to implement support for vendor models handling SIG opcodes. Signed-off-by: Ludvig Jordet <ludvig.jordet@nordicsemi.no>
This fixes a limitation where the stack assumed that vendor models could not specify handlers for SIG-defined (1- or 2-byte) opcodes. This assumption does not exist in the specification. In fact, the specification expects this very use-case to be possible, in MeshPRT 1.1, section 3.8.3: > To exchange messages with a Bluetooth SIG adopted model, a Vendor Model > shall use the Access message defined for the Bluetooth SIG model. Signed-off-by: Ludvig Jordet <ludvig.jordet@nordicsemi.no>
This adds a test case to the access BSIM tests to check that vendor models can specify a handler for and receive messages using a SIG-defined 1-byte opcode. Signed-off-by: Ludvig Jordet <ludvig.jordet@nordicsemi.no>
Since the STM32MP2 I3C controller uses a combined interrupt line for I3C events and errors, add the support for handling this type of interrupt in the I3C driver. Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
Add clock bindings for I3C peripherals for the STM32MP2 series. Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
Add reset bindings for I3C peripherals for the STM32MP2 series. Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
Add I3C clocks rate reading to the STM32MP2 clock driver. Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
Add the node I3C4 in non-secure context to dtsi. Others I3C nodes cannot be added as they rely on EXTI1 interrupt lines that is not supported yet for the STM32MP2. Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
The I3C is now supported on the STM32MP257F-DK and STM32MP257F-EV1 boards. Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
The I2C and SPI features have been added to the supported features for the stm32mp257f_ev1 but were missing from the stm32mp257f_dk board file. Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
This commit adds the i2c8 node to the device tree for the STM32MP257F-DK board. This node corresponds to the i2c exposed by the GPIO expansion connector [1] and not used by the cortex A35. [1]: https://www.st.com/resource/en/user_manual/um3385-discovery-kit-with-stm32mp257f-mpu-stmicroelectronics.pdf Table 24. GPIO connector pinout Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
This commit adds the i3c4 node to the device tree for the STM32MP257F-EV1 and the STM32MP257F-DK boards. This node corresponds to the i3c exposed by the GPIO expansion connector [1], [2] and not used by the cortex A35. According to the STM32MP257F-EV1 and STM32MP257F-DK schematics, the I3C4 are connected to the same pins as the I2C8. Therefore, both nodes cannot be enabled together. [1]: https://wiki.st.com/stm32mpu/wiki/STM32MP257x-EV1_-_hardware_description [1]: https://www.st.com/resource/en/user_manual/um3385-discovery-kit-with-stm32mp257f-mpu-stmicroelectronics.pdf Table 24. GPIO connector pinout Co-authored-by: Charles Wenger <charles.wenger@savoirfairelinux.com> Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
This driver uses math functions that require the math.h header. If not included a lot of implicit declaration warnings are generated during build. Signed-off-by: Tanguy Raufflet <tanguy.raufflet@savoirfairelinux.com>
Previously, eventfd file descriptors were not being counted against the required size for the global file descriptor table, which would result in the function `eventfd()` (and `zvfs_eventfd()`) failing due to insufficient resources. Signed-off-by: Chris Friedt <chris@fr4.co>
This patch adds support for the QSPI memory controller (for the SiWG917 SoC) in order to enable PSRAM testing. It also adds board overlays for this family that include PSRAM. Signed-off-by: Martin Hoff <martin.hoff@silabs.com>
Enable the adc node to prevent undefined node reference error in adc_api test. Signed-off-by: Fabian Blatz <fabianblatz@gmail.com>
Product photo from https://learn.adafruit.com/assets/121788, with the license CC BY-SA 3.0 Tested with the commands mentioned in index.rst Signed-off-by: Jonas Berg <jonas.s.t.berg@gmail.com>
Add corresponding overlay file Signed-off-by: Jonas Berg <jonas.s.t.berg@gmail.com>
set_iir_config() dereferences the device pointer before checking it against NULL, making the defensive check ineffective. Remove the redundant check. Signed-off-by: Gaetan Perrot <gaetan.perrot@spacecubics.com>
This patch enables the output sample for the board siwx917_rb4342a by adding an overlay. Signed-off-by: Martin Hoff <martin.hoff@silabs.com>
Raspberry pi based boards are now using binary info by default (96b9b0f), and that limits pinctrl groups to 4 or 6 nodes. Split this node in multiple groups. Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
This commit adds the main DTS configurations required to enable MSPI/OSPI/QSPI support on STM32. Signed-off-by: Sara Touqan <zephyr@exalt.ps> Signed-off-by: Sarah Younis <zephyr@exalt.ps> Signed-off-by: Mohammad Odeh <zephyr@exalt.ps>
This commit introduces support for the mspi and ospi drivers on STM32, enabling functionality APIs for MSPI/OSPI/QSPI host controllers.. Signed-off-by: Sara Touqan <zephyr@exalt.ps> Signed-off-by: Sarah Younis <zephyr@exalt.ps> Signed-off-by: Mohammad Odeh <zephyr@exalt.ps>
this commit enables building and running the sample on stm32h573i_dk board,stm32h735g_disco board, arduino_giga_r1 board, and b_u585i_iot02a board by providing the required overlay and configuration updates. Signed-off-by: Sara Touqan <zephyr@exalt.ps> Signed-off-by: Sarah Younis <zephyr@exalt.ps> Signed-off-by: Mohammad Odeh <zephyr@exalt.ps>
- Moves nxp,imx-gpt.yaml from the timer to the counter folder, as this is the binding for the counter driver drivers/counter/counter_mcux_gpt.c - Fixes: zephyrproject-rtos#100947 Signed-off-by: Andrej Butok <andrey.butok@nxp.com>
UART0 is not currently used by Zephyr running on RPU. Remove it. Signed-off-by: Yasushi SHOJI <yashi@spacecubics.com>
Use the convenient DT_FREQ_M() macro from freq.h to improve readability. Signed-off-by: Yasushi SHOJI <yashi@spacecubics.com>
1. Add power state nodes for mcxa156, including sleep, deepsleep and powerdown. 2. Add power management related peripheral nodes for mcxa156, including spc, cmc, vbat and wuu. 3. Add 'zephyr,cortex-m-idle-timer' node in frdm_mcxa156. Signed-off-by: Zhaoxiang Jin <Zhaoxiang.Jin_1@nxp.com>
The device is not able to place the entropy into stack or heap allocated buffers. This uses a bounce buffer to be able to use any buffer with virtio entropy. Signed-off-by: Juliusz Sosinowicz <juliusz@wolfssl.com>
Add Renode simulation support for `mpfs_icicle`. Signed-off-by: Marek Slowinski <mslowinski@antmicro.com>
Tune quantum parameter for selected kernel tests targeting mpfs_icicle platform. Those tests require higher fidelity of the virtual time flow which is achievable on multi-core platforms in Renode by reducing the quantum. Increase time resolution to 10us (default is 100us). It is a maximum drift of the virtual time between simulated processors in SMP configuration. Signed-off-by: Marek Slowinski <mslowinski@antmicro.com>
Tested with the commands in index.rst Product photo from https://github.com/CytronTechnologies/MAKER-NANO-RP2040/ blob/main/MAKER-NANO-RP2040-Image.PNG Signed-off-by: Jonas Berg <jonas.s.t.berg@gmail.com>
Run clang-format on drivers/i2c/i2c_cdns.c. Signed-off-by: Simon Maurer <mail@maurer.systems>
Switch the Cadence I2C device driver over to the use of the DEVICE_MMIO_... macros instead of just using the physical base address from the device tree. Signed-off-by: Simon Maurer <mail@maurer.systems>
Clear the hold flag during the final data interrupt; otherwise the Transfer Complete interrupt is never asserted. Signed-off-by: Simon Maurer <mail@maurer.systems>
…unter Add support for configuring and enabling the internal timestamp counter of the Bosch M_CAN IP core. Frontend drivers can overwrite this configuration for using a SoC-specific, external timestamp counter. Signed-off-by: Henrik Brix Andersen <henrik@brixandersen.dk>
Enable support for configuring and enabling the external timestamp counter of the NXP LPC MCAN. Signed-off-by: Henrik Brix Andersen <henrik@brixandersen.dk>
Add initial support for the Allwinner H3 SoC, commonly found in development boards like the Orange Pi series. This commit introduces the intial SoC support files: - Basic Kconfig configuration and SoC definition - MMU region setup for memory management - SoC-specific headers and device tree source include Signed-off-by: Muhammad Waleed Badar <walid.badar@gmail.com>
Add initial board support for the Xunlong Orange Pi Zero, single-board computer based on the Allwinner H2+/H3 SoC. Signed-off-by: Muhammad Waleed Badar <walid.badar@gmail.com>
The MCXN236 can use ostimer as the kernel timer. We enable it in the dts, and which specific timer to use as the system timer can be controlled through Kconfig option based on application requirements. Signed-off-by: Zhaoxiang Jin <Zhaoxiang.Jin_1@nxp.com>
MCUX_LPTMR_TIMER is default to n, no need to set it again. Signed-off-by: Zhaoxiang Jin <Zhaoxiang.Jin_1@nxp.com>
Add MCUX_OS_TIMER configuration with CPU_FREQ-based default selection. Update CORTEX_M_SYSTICK to be disabled when either LPTMR or OSTIMER is selected as the system timer, ensuring only one timer is active. Signed-off-by: Zhaoxiang Jin <Zhaoxiang.Jin_1@nxp.com>
Cmake complains about this. Signed-off-by: Wilfried Chauveau <wilfried.chauveau@arm.com>
Fix macro parameter typo (##inst → ##index) in CAN_RENESAS_RA_INIT to ensure unique IRQ configure function names per instance. Signed-off-by: Khai Cao <khai.cao.xk@renesas.com>
BIT_MASK(32) cannot do (1 << 32) - 1 without causing integer overflow. Hence, for these macros BIT64_MASK(32) is required. Additionally, remove the unnecessary (and presently unused) macros for the register space. All of these can be reliably derived from the present register masks. Signed-off-by: Amneesh Singh <amneesh@ti.com>
Disable all interrupts when the VIM interrupt controller is initialized so that it doesn't encounter any stray interrupts that were not enabled on Zephyr. Signed-off-by: Amneesh Singh <amneesh@ti.com>
List USB device controller as a supported feature in order to increase test coverage. Signed-off-by: Henrik Brix Andersen <henrik@brixandersen.dk>
Reneamed photograph of Renesas US159 DA14531EVZ board to match convention documented in the Board Porting Guide (resolving issue with old image not appearing in list of supported boards on Zephyr website). Signed-off-by: Ian Morris <ian.d.morris@outlook.com>
In z_vrfy_k_poll, there is a memory access check K_SYSCALL_MEMORY_WRITE which is wrapped in a spinlock, the same spinlock used in z_handle_obj_poll_events which is called from k_sem_give() for example. The K_SYSCALL_MEMORY_WRITE() macro conditionally calls LOG_ERR() which may call the UART console, which may call an API like k_sem_give(). This will cause a deadlock since the locked spinlock will be relocked, and a recursive lock if SPINLOCK_VALIDATE and ASSERTS are enabled as the validation will fail, causing a LOG_ERR, causing a k_sem_give() causing a relock... until stack overflows. To solve the issue, only protect the copy of events to events_copy with the spinlock, the content of events is not actually checked, and bound is not shared, so there is no need to do this validation in a critical section. The contents of events is shared so that must be copied in atomically. Signed-off-by: Bjarki Arge Andreasen <bjarki.andreasen@nordicsemi.no>
The current implementation of k_sleep(), when multi-threading is disabled, busy waits using k_busy_wait() until the sleep timeout has expired. This patch aims to improve power efficiency of k_sleep() for single-threaded applications by starting a timer (k_timer) and idling the CPU until the timer interrupt wakes it up, thus avoiding busy-looping. Signed-off-by: Emanuele Di Santo <emdi@nordicsemi.no>
Instead of performing a linear search to determine if a given thread is running on another CPU, or if it is marked as being preempted by a metaIRQ on any CPU do this in O(1) time. On SMP systems, Zephyr already tracks the CPU on which a thread executes (or lasted executed). This information is leveraged to do the search in O(1) time. Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
Add support for clause 45 MDIO transactions to the GPIO MDIO driver. Signed-off-by: Tim Pambor <tim.pambor@codewrights.de>
This problem didn't surface earlier, as different sizes weren't permitted. Signed-off-by: Benedek Kupper <kupper.benedek@gmail.com>
bhoot1234567890
force-pushed
the
clean/fix-picolibc-locks
branch
from
40c0617 to
0fdd5de
Compare
Author
|
Closing this duplicate PR — the clean single-commit fix is at zephyrproject-rtos#101988. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
20 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary:
This patch fixes several issues in zephyr/lib/libc/picolibc/locks.c discovered when building ZMK with picolibc enabled. The changes make the file compatible with newlib/newlib-style expectations and Zephyr userspace handling.
Problems found:
K_MUTEX_DEFINE(__lock___libc_recursive_mutex)conflicts with newlib'sstruct __lockand Zephyr iterable section macros, causing "conflicting types" compilation errors._LOCK_Tpreviously defined asvoid *mismatched newlib expectations (newlib expects_LOCK_Tto be a pointer tostruct __lock). This caused conflicting function signatures for__retarget_lock_*.K_MUTEX_DEFINE, which cannot be granted properly as an object pointer in userspace contexts.struct k_mutexpointers (or returned raw k_mutex pointers), which didn’t match newlib's_LOCK_Tusage.What this patch does:
struct __lockwrapper matching newlib expectations and typedefs_LOCK_Tasstruct __lock *so signatures match newlib.K_MUTEX_DEFINEstatic declaration with astruct __lock __lock___libc_recursive_mutexwrapper and initializes the embedded mutex in aSYS_INIThandler.CONFIG_USERSPACEis enabled, the static wrapper allocates a kernel mutex object, initializes it, and grants access withk_object_access_all_grant.struct __lockwrappers and to allocate/initialize the underlying k_mutex appropriately for both kernel and userspace builds.__retarget_lock_acquire/_release/_try_acquireto operate on the wrapper correctly in both userspace and non-userspace builds.Testing performed:
zephyr/zmk.elfis generated.CONFIG_NEWLIB_LIBCswitches the build to newlib (no longer uses this file), and observed expected size differences.Notes:
K_MUTEX_DEFINEapproach, an alternate patch could expose the object properly to userspace; this patch prefers an explicit wrapper to match newlib ABI expectations.Files changed:
Please review for style and userspace semantics; happy to adjust to match upstream conventions.