* reduced reserved memory size for system stack; added temporary markup to monitor usage
* fbt: relink elf file on linker script change; removed debug memory fill
* Make PVS Happy
* Make doxygen happy
Co-authored-by: Aleksandr Kutuzov <alleteam@gmail.com>
* Now 4a listener invokes upper level callback on Halt and FieldOff
* Added new method for drawing mirrored XBM bitmaps
* iso14443_4a poller logic enhanced
* Function renamed accroding to review suggestions
* Rename #2
* Api adjustements
* Correct API bump
Co-authored-by: あく <alleteam@gmail.com>
- BT Service: cleanup code
- Dialog: correct release order in file browser
- Rpc: rollback to pre #3881 state
- Kernel: fix inverted behavior in furi_kernel_is_running
- Log: properly take mutex when kernel is not running
- Thread: rework tread control block scrubbing procedure, ensure that we don't do stupid things in idle task, add new priority for init task
- Timer: add control queue flush method, force flush on stop
- Furi: system init task now performs thread scrubbing
- BleGlue: add some extra checks
- FreeRTOSConfig: fix bunch of issues that were preventing configuration from being properly applied and cleanup
State callbacks assumed they were invoked from the thread that changed
its state, but this wasn't true for FuriThreadStateStarting in the past,
and now it's not true for FuriThreadStateStopped either.
Now it is safe to release the thread memory form the state callback
once it switches to FuriThreadStateStopped. Therefore, pending deletion
calls can be removed.
Co-authored-by: Aleksandr Kutuzov <alleteam@gmail.com>
* FuriTimer: Use a local variable to wait for deletion
This combines the current synchronous behaviour
(as we could have deferred the free call too) with
a smaller FuriTimer - it's safe to pass a pointer to
a local variable to this pending timer call, because we
know it'll be finished before the caller returns
* Tighten the use of FuriThread* vs FuriThreadId
Event loop and Loader mixed those two,
but the fact those are aliases should be an implementation detail.
For this reason, thread.c is still allowed to mix them freely.
strlcpy doesn't zero the buffer and ensures null termination,
just like snprintf
strlcat is already used by mjs and it's a safe alternative to strcat,
so it should be OK to expose to apps
