Fix Cortex m0 Raw32 scheduler.

Ensure 8 byte alignment of the stack for all ARM specific scheduler code.

3 files changed, 25 insertions, 18 deletions

diff --git a/src/gos/gos_x_threads_cortexm01.h b/src/gos/gos_x_threads_cortexm01.h
index 659d7af7..15a90b2a 100644
--- a/src/gos/gos_x_threads_cortexm01.h
+++ b/src/gos/gos_x_threads_cortexm01.h
@@ -27,8 +27,10 @@
 							"mov     r6, r10                                \n\t"
 							"mov     r7, r11                                \n\t"
 							"push    {r4, r5, r6, r7}						\n\t"
-							"str	sp, %[oldtcxt]							\n\t"
-							"ldr	sp, %[newtcxt]							\n\t"
+							"mov     r4, sp                                 \n\t"
+							"str     r4, %[oldtcxt]							\n\t"
+							"ldr     r4, %[newtcxt]							\n\t"
+							"mov     sp, r4                                 \n\t"
 							"pop     {r4, r5, r6, r7}                   	\n\t"
 							"mov     r8, r4                                 \n\t"
 							"mov     r9, r5                                 \n\t"
@@ -41,15 +43,17 @@
 	}
 
 	static __attribute__((pcs("aapcs"),naked)) void _gfxStartThread(thread *oldt, thread *newt) {
-		newt->cxt = (char *)newt + newt->size;
+		newt->cxt = (void *)(((unsigned)newt + newt->size) & ~7);
 		__asm__ volatile (	"push    {r4, r5, r6, r7, lr}                   \n\t"
 							"mov     r4, r8                                 \n\t"
 							"mov     r5, r9                                 \n\t"
 							"mov     r6, r10                                \n\t"
 							"mov     r7, r11                                \n\t"
 							"push    {r4, r5, r6, r7}						\n\t"
-							"str	sp, %[oldtcxt]							\n\t"
-							"ldr	sp, %[newtcxt]							\n\t"
+							"mov     r4, sp                                 \n\t"
+							"str     r4, %[oldtcxt]							\n\t"
+							"ldr     r4, %[newtcxt]							\n\t"
+							"mov     sp, r4                                 \n\t"
 							: [newtcxt] "=m" (newt->cxt)
 							: [oldtcxt] "m" (oldt->cxt)
 							: "memory");
@@ -72,10 +76,12 @@
 		mov     r6, r10
 		mov     r7, r11
 		push    {r4, r5, r6, r7}
-		str	sp, [r0,#__cpp(offsetof(thread,cxt))]		// oldt->cxt
+		mov		r4, sp
+		str		r4, [r0,#__cpp(offsetof(thread,cxt))]		// oldt->cxt
 		
 		// Load the new context
-		ldr	sp, [r1,#__cpp(offsetof(thread,cxt))]		// newt->cxt
+		ldr		r4, [r1,#__cpp(offsetof(thread,cxt))]		// newt->cxt
+		mov		sp, r4
 		pop     {r4, r5, r6, r7}
 		mov     r8, r4
 		mov     r9, r5
@@ -88,9 +94,10 @@
 		PRESERVE8
 
 		// Calculate where to generate the new context
-		//		newt->cxt = (char *)newt + newt->size;
+		//		newt->cxt = (void *)(((unsigned)newt + newt->size) & ~7);
 		ldr      r2,[r1,#__cpp(offsetof(thread,size))]
 		add      r2,r2,r1
+		and      r2, r2, #0xFFFFFFF8
 		str      r2,[r1,#__cpp(offsetof(thread,cxt))]
 		
 		// Save the old context
@@ -100,10 +107,12 @@
 		mov     r6, r10
 		mov     r7, r11
 		push    {r4, r5, r6, r7}
-		str		sp, [r0,#__cpp(offsetof(thread,cxt))]	// oldt->cxt
+		mov		r4, sp
+		str		r4, [r0,#__cpp(offsetof(thread,cxt))]	// oldt->cxt
 		
 		// Load the new (imcomplete) context
-		ldr		sp, [r1,#__cpp(offsetof(thread,cxt))]	// newt->cxt
+		ldr		r4, [r1,#__cpp(offsetof(thread,cxt))]	// newt->cxt
+		mov		sp, r4
 		
 		// Run the users function - we save some code because gfxThreadExit() never returns
 		//		gfxThreadExit(_gfxCurrentThread->fn(_gfxCurrentThread->param));
diff --git a/src/gos/gos_x_threads_cortexm347.h b/src/gos/gos_x_threads_cortexm347.h
index 5248508e..40d30138 100644
--- a/src/gos/gos_x_threads_cortexm347.h
+++ b/src/gos/gos_x_threads_cortexm347.h
@@ -32,7 +32,7 @@
 	}
 
 	static __attribute__((pcs("aapcs"),naked)) void _gfxStartThread(thread *oldt, thread *newt) {
-		newt->cxt = (char *)newt + newt->size;
+		newt->cxt = (void *)(((unsigned)newt + newt->size) & ~7);
 		__asm__ volatile (	"push	{r4, r5, r6, r7, r8, r9, r10, r11, lr}	\n\t"
 							"str	sp, %[oldtcxt]							\n\t"
 							"ldr	sp, %[newtcxt]							\n\t"
@@ -64,9 +64,10 @@
 		PRESERVE8
 
 		// Calculate where to generate the new context
-		//		newt->cxt = (char *)newt + newt->size;
+		//		newt->cxt = (void *)(((unsigned)newt + newt->size) & ~7);
 		ldr      r2,[r1,#__cpp(offsetof(thread,size))]
 		add      r2,r2,r1
+		and      r2, r2, #0xFFFFFFF8
 		str      r2,[r1,#__cpp(offsetof(thread,cxt))]
 		
 		// Save the old context
diff --git a/src/gos/gos_x_threads_cortexm47fp.h b/src/gos/gos_x_threads_cortexm47fp.h
index 8134a05d..f5eeff80 100644
--- a/src/gos/gos_x_threads_cortexm47fp.h
+++ b/src/gos/gos_x_threads_cortexm47fp.h
@@ -12,10 +12,6 @@
  * The context is saved at the current stack location and a pointer is maintained in the thread structure.
  */
 
-#if !CORTEX_USE_FPU
-	#warning "GOS Threads: You have specified GFX_CPU=GFX_CPU_CORTX_M?_FP with hardware floating point support but CORTEX_USE_FPU is FALSE. Try using GFX_CPU_GFX_CPU_CORTEX_M? instead"
-#endif
-
 #if GFX_COMPILER == GFX_COMPILER_GCC || GFX_COMPILER == GFX_COMPILER_CYGWIN || GFX_COMPILER == GFX_COMPILER_MINGW32 || GFX_COMPILER == GFX_COMPILER_MINGW64
 	#define GFX_THREADS_DONE
 	#define _gfxThreadsInit()
@@ -33,7 +29,7 @@
 	}
 
 	static __attribute__((pcs("aapcs-vfp"),naked)) void _gfxStartThread(thread *oldt, thread *newt) {
-		newt->cxt = (char *)newt + newt->size;
+		newt->cxt = (void *)(((unsigned)newt + newt->size) & ~7);
 		__asm__ volatile (	"push	{r4, r5, r6, r7, r8, r9, r10, r11, lr}	\n\t"
 							"vpush	{s16-s31}								\n\t"
 							"str	sp, %[oldtcxt]							\n\t"
@@ -68,9 +64,10 @@
 		PRESERVE8
 
 		// Calculate where to generate the new context
-		//		newt->cxt = (char *)newt + newt->size;
+		//		newt->cxt = (void *)(((unsigned)newt + newt->size) & ~7);
 		ldr      r2,[r1,#__cpp(offsetof(thread,size))]
 		add      r2,r2,r1
+		and      r2, r2, #0xFFFFFFF8
 		str      r2,[r1,#__cpp(offsetof(thread,cxt))]
 		
 		// Save the old context