diff --git a/demos/ARM7-LPC214x-GCC-minimal/Makefile b/demos/ARM7-LPC214x-GCC-minimal/Makefile
index f440f799d..e74d2f382 100644
--- a/demos/ARM7-LPC214x-GCC-minimal/Makefile
+++ b/demos/ARM7-LPC214x-GCC-minimal/Makefile
@@ -121,7 +121,7 @@ LIBS    = $(DLIBS) $(ULIBS)
 MCFLAGS = -mcpu=$(MCU)
 
 ASFLAGS = $(MCFLAGS) -Wa,-amhls=$(<:.s=.lst) $(ADEFS)
-CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-ahlms=$(<:.c=.lst) $(DEFS)
+CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-alms=$(<:.c=.lst) $(DEFS)
 LDFLAGS = $(MCFLAGS) -nostartfiles -T$(LDSCRIPT) -Wl,-Map=$(PROJECT).map,--cref,--no-warn-mismatch $(LIBDIR)
 ODFLAGS	= -x --syms
 
diff --git a/demos/ARM7-LPC214x-GCC-minimal/Makefile.thumb b/demos/ARM7-LPC214x-GCC-minimal/Makefile.thumb
index 03f2c7ba9..68f308449 100644
--- a/demos/ARM7-LPC214x-GCC-minimal/Makefile.thumb
+++ b/demos/ARM7-LPC214x-GCC-minimal/Makefile.thumb
@@ -120,7 +120,7 @@ LIBS    = $(DLIBS) $(ULIBS)
 MCFLAGS = -mcpu=$(MCU)
 
 ASFLAGS = $(MCFLAGS) -Wa,-amhls=$(<:.s=.lst) $(ADEFS)
-CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-ahlms=$(<:.c=.lst) $(DEFS)
+CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-alms=$(<:.c=.lst) $(DEFS)
 LDFLAGS = $(MCFLAGS) -nostartfiles -T$(LDSCRIPT) -Wl,-Map=$(PROJECT).map,--cref,--no-warn-mismatch $(LIBDIR)
 ODFLAGS	= -x --syms
 
diff --git a/demos/ARM7-LPC214x-GCC/Makefile b/demos/ARM7-LPC214x-GCC/Makefile
index 3698582e7..60c2df008 100644
--- a/demos/ARM7-LPC214x-GCC/Makefile
+++ b/demos/ARM7-LPC214x-GCC/Makefile
@@ -123,7 +123,7 @@ LIBS    = $(DLIBS) $(ULIBS)
 MCFLAGS = -mcpu=$(MCU)
 
 ASFLAGS = $(MCFLAGS) -Wa,-amhls=$(<:.s=.lst) $(ADEFS)
-CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-ahlms=$(<:.c=.lst) $(DEFS)
+CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-alms=$(<:.c=.lst) $(DEFS)
 LDFLAGS = $(MCFLAGS) -nostartfiles -T$(LDSCRIPT) -Wl,-Map=$(PROJECT).map,--cref,--no-warn-mismatch $(LIBDIR)
 ODFLAGS	= -x --syms
 
diff --git a/demos/ARM7-LPC214x-GCC/Makefile.thumb b/demos/ARM7-LPC214x-GCC/Makefile.thumb
index c08d632f9..111fa8c7b 100644
--- a/demos/ARM7-LPC214x-GCC/Makefile.thumb
+++ b/demos/ARM7-LPC214x-GCC/Makefile.thumb
@@ -122,7 +122,7 @@ LIBS    = $(DLIBS) $(ULIBS)
 MCFLAGS = -mcpu=$(MCU)
 
 ASFLAGS = $(MCFLAGS) -Wa,-amhls=$(<:.s=.lst) $(ADEFS)
-CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-ahlms=$(<:.c=.lst) $(DEFS)
+CPFLAGS = $(MCFLAGS) $(OPT) $(WARN) -Wa,-alms=$(<:.c=.lst) $(DEFS)
 LDFLAGS = $(MCFLAGS) -nostartfiles -T$(LDSCRIPT) -Wl,-Map=$(PROJECT).map,--cref,--no-warn-mismatch $(LIBDIR)
 ODFLAGS	= -x --syms
 
diff --git a/demos/ARM7-LPC214x-GCC/chcore2.s b/demos/ARM7-LPC214x-GCC/chcore2.s
index 3bbffe26f..92d9d6d24 100644
--- a/demos/ARM7-LPC214x-GCC/chcore2.s
+++ b/demos/ARM7-LPC214x-GCC/chcore2.s
@@ -115,10 +115,11 @@ chSysSwitchI:
  * interrupt handler:
  *
  * High +------------+
- *      |     R12    | -+
+ *      |   LR_USR   | -+
+ *      |     R12    |  |
  *      |     R3     |  |
- *      |     R2     |  |
- *      |     R1     |  | External context: IRQ handler frame
+ *      |     R2     |  | External context: IRQ handler frame
+ *      |     R1     |  |
  *      |     R0     |  |
  *      |   LR_IRQ   |  |   (user code return address)
  *      |    SPSR    | -+   (user code status)
@@ -136,7 +137,6 @@ chSysSwitchI:
  */
 .globl IrqHandler
 IrqHandler:
-        sub     lr, lr, #4
         stmfd   sp!, {r0-r3, r12, lr}
 #ifdef THUMB_NO_INTERWORKING
         add     r0, pc, #1
@@ -152,7 +152,6 @@ IrqHandler:
 
 .globl T0IrqHandler
 T0IrqHandler:
-        sub     lr, lr, #4
         stmfd   sp!, {r0-r3, r12, lr}
 #ifdef THUMB_NO_INTERWORKING
         add     r0, pc, #1
@@ -168,7 +167,6 @@ T0IrqHandler:
 
 .globl UART0IrqHandler
 UART0IrqHandler:
-        sub     lr, lr, #4
         stmfd   sp!, {r0-r3, r12, lr}
 #ifdef THUMB_NO_INTERWORKING
         add     r0, pc, #1
@@ -184,7 +182,6 @@ UART0IrqHandler:
 
 .globl UART1IrqHandler
 UART1IrqHandler:
-        sub     lr, lr, #4
         stmfd   sp!, {r0-r3, r12, lr}
 #ifdef THUMB_NO_INTERWORKING
         add     r0, pc, #1
@@ -216,12 +213,13 @@ IrqCommon:
         bl      chSchRescRequiredI
 #endif
         cmp     r0, #0                          // Simply returns if a
-        ldmeqfd sp!, {r0-r3, r12, pc}^          // reschedule is not required.
+        ldmeqfd sp!, {r0-r3, r12, lr}           // reschedule is not
+        subeqs  pc, lr, #4			// required.
 
         // Saves the IRQ mode registers in the system stack.
         ldmfd   sp!, {r0-r3, r12, lr}           // IRQ stack now empty.
         msr     CPSR_c, #MODE_SYS | I_BIT
-        stmfd   sp!, {r0-r3, r12}               // Registers on System Stack.
+        stmfd   sp!, {r0-r3, r12, lr}           // Registers on System Stack.
         msr     CPSR_c, #MODE_IRQ | I_BIT
         mrs     r0, SPSR
         mov     r1, lr
@@ -247,6 +245,6 @@ IrqCommon:
         msr     SPSR_fsxc, r0
         mov     lr, r1
         msr     CPSR_c, #MODE_SYS | I_BIT
-        ldmfd   sp!, {r0-r3, r12}
+        ldmfd   sp!, {r0-r3, r12, lr}
         msr     CPSR_c, #MODE_IRQ | I_BIT
-        subs    pc, lr, #0
+        subs    pc, lr, #4
diff --git a/readme.txt b/readme.txt
index c9d29eeb8..61b5347d1 100644
--- a/readme.txt
+++ b/readme.txt
@@ -40,9 +40,15 @@ AVR-AT90CANx-GCC       - Port on AVR AT90CAN128, not complete yet.
 *****************************************************************************
 
 *** 0.4.4 ***
+- Fixed a very important bug in the preemption ARM code, important enough to
+  make this update *mandatory*.
+  Note: This is not a kernel bug but something specific with the ARM port, the
+        other ports are not affected.
 - Fixed a nasty bug in the pure THUMB mode threads trampoline code (chcore2.s,
   threadstart), it failed on THUMB threads returning with a "bx" instruction.
   The bug did not affect ARM mode or THUMB with interworking mode.
+  Note: This is not a kernel bug but something specific with the ARM port, the
+        other ports are not affected.
 - Fixed a bug in chIQGetTimeout(), interrupts were not re-enabled when exiting
   the function because a timeout. The problem affected that API only.
 - Fixed a potential problem in chSysInit(), it should not affect any past
diff --git a/test/test.c b/test/test.c
index 289e74a3f..15b70b9d4 100644
--- a/test/test.c
+++ b/test/test.c
@@ -310,7 +310,7 @@ t_msg TestThread(void *p) {
   }
   print("Queues throughput = ");
   printn(i * 4);
-  print(" bytes/S");
+  println(" bytes/S");
 
   println("\r\nTest complete");
   return 0;