Implement threads on Zephyr (#19156)

* pthreads setup for zephyr

- enable tweak stack size
- update lib/system/threads.nim
- Fix int/uint in casting pointer.

* add documentation and tweak flag names

* add documentation and tweak flag names

* fix configuration flag names

* fix configuration flag names

* cleanup

Co-authored-by: Jaremy Creechley <jaremy.creechley@panthalassa.com>

doc/nimc.rst

@@ -681,6 +681,15 @@ nimMemAlignTiny
 Sets `MemAlign` to `4` bytes which reduces the memory alignment
 to better match some embedded devices.
 
+Thread stack size 
+=================
+
+Nim's thread API provides a simple wrapper around more advanced
+RTOS task features. Customizing the stack size and stack guard size can
+be done by setting `-d:nimThreadStackSize=16384` or `-d:nimThreadStackGuard=32`.
+
+Currently only Zephyr and FreeRTOS support these configurations. 
+
 Nim for realtime systems
 ========================
 

lib/system/threadlocalstorage.nim

@@ -153,6 +153,8 @@ else:
 
   proc pthread_attr_init(a1: var Pthread_attr): cint {.
     importc, header: pthreadh.}
+  proc pthread_attr_setstack*(a1: ptr Pthread_attr, a2: pointer, a3: int): cint {.
+    importc, header: pthreadh.}
   proc pthread_attr_setstacksize(a1: var Pthread_attr, a2: int): cint {.
     importc, header: pthreadh.}
   proc pthread_attr_destroy(a1: var Pthread_attr): cint {.

lib/system/threads.nim

@@ -47,14 +47,23 @@
 when not declared(ThisIsSystem):
   {.error: "You must not import this module explicitly".}
 
-const
+when defined(zephyr) or defined(freertos):
-  StackGuardSize = 4096
+  const
-  ThreadStackMask =
+    nimThreadStackSize {.intdefine.} = 8192 
-    when defined(genode):
+    nimThreadStackGuard {.intdefine.} = 128
-      1024*64*sizeof(int)-1
+
-    else:
+    StackGuardSize = nimThreadStackGuard 
-      1024*256*sizeof(int)-1
+    ThreadStackSize = nimThreadStackSize - nimThreadStackGuard 
-  ThreadStackSize = ThreadStackMask+1 - StackGuardSize
+else:
+  const
+    StackGuardSize = 4096
+    ThreadStackMask =
+      when defined(genode):
+        1024*64*sizeof(int)-1
+      else:
+        1024*256*sizeof(int)-1
+
+    ThreadStackSize = ThreadStackMask+1 - StackGuardSize
 
 #const globalsSlot = ThreadVarSlot(0)
 #sysAssert checkSlot.int == globalsSlot.int
@@ -321,7 +330,14 @@ else:
     when hasSharedHeap: t.core.stackSize = ThreadStackSize
     var a {.noinit.}: Pthread_attr
     doAssert pthread_attr_init(a) == 0
-    let setstacksizeResult = pthread_attr_setstacksize(a, ThreadStackSize)
+    when defined(zephyr):
+      var
+        rawstk = allocShared0(ThreadStackSize + StackGuardSize)
+        stk = cast[pointer](cast[uint](rawstk) + StackGuardSize)
+      let setstacksizeResult = pthread_attr_setstack(addr a, stk, ThreadStackSize)
+    else:
+      let setstacksizeResult = pthread_attr_setstacksize(a, ThreadStackSize)
+
     when not defined(ios):
       # This fails on iOS
       doAssert(setstacksizeResult == 0)