ARM: Cortex-M/R: improvements for v2.7.0 release

[ioannisg]

Is your enhancement proposal related to a problem? Please describe. Enhancements for the Cortex-M architecture support

List of open issues

• [x] Optimize SPLIM programming, potentially removing SPLIM clearing to zero
• [x] An AARC32 user guide, similar to ARC and x86 (#36388)
• [ ] Validate MSP in HardFault and NMI handlers
• [x] Use labels for constants in assembly (#36333)
• [ ] Investigate if barrier instructions are needed for semaphore operations
• [ ] Investigate if aligned access can be enforced, or at least set by default
• [ ] Investigate using unprivileged load-store instructions in user mode assembly code
• [x] Re-work the VTOR offset register for the different variants of Cortex-M (#36333)
• [x] Implement FPU support in QEMU and enable it in CI (#35381)
• [ ] Clean up SoC-specific code from arch/arm/aarch32 and include/arch/arm/aarch32
• [x] Switch arch_irq_lock to use the BASEPRI max instead of BASEPRI setting (#36333)
• [ ] Configure non-banked faults in accordance with TF-M, if building with TF-M
• [ ] Store FP callee-saved context in stack instead of allocating a placeholder in kernel memory
• [x] Move save/restore of FPU context in arch, as common internal API (#35846)
• [ ] Zero latency IRQs for Baseline Cortex-M
  • [x] [Cortex-R] MPU support (#28231 )
• [ ] [Enh. Kernel] Rename userlocal_datato something that does not contain "user" in the name, as it is used under CONFIG_USERSPACE but beyond user threads
• [ ] [Enh. Cortex-R] Support floating-point (#TBD)
• [ ] [Enh. ARM] Cross-ARCH API for setting IRQ priority (RFC, #25910)
• [ ] [Enh. ARM Cortex-M] Add HW stack protection during boot, too.
• [ ] [Bug ARM Cortex-M] Support for HW stack protection features together with CONFIG_ARM_SECURE_FIRMWARE=y and CONFIG_FP_SHARING (currently broken)
• [x] Dump fault information during HardFault escalation in Cortex-M (#36333)
• [x] Standardize API for ARM's z_platform_init (#36333)
• [x] Initialize unused core exceptions to prevent undefined behavior (#36333)

[theotherjimmy]

Do we have a reproducing case for the FP_SHARING bug? I would like to take a look into that soon.

[carlescufi]

Assigned to @ithinuel. Feel free to close this one.

[ithinuel]

@theotherjimmy I know it's been a while but do you remember if you had time to look into that at the time and if there was a reproducible issue ?