6 files changed, 9 insertions, 11 deletions
diff --git a/arch/m68k/fpsp040/res_func.S b/arch/m68k/fpsp040/res_func.S
index 9a81bde1d..66e42d749 100644
--- a/arch/m68k/fpsp040/res_func.S
+++ b/arch/m68k/fpsp040/res_func.S
@@ -1889,7 +1889,7 @@ dp_over:
 |
 | Input: a0  points to a normalized number in internal extended format
 |	 d0  is the round precision (=1 for sgl; =2 for dbl)
-|	 d1  is the the single precision or double precision
+|	 d1  is the single precision or double precision
 |	     denorm threshold
 |
 | Output: (In the format for dest_sgl or dest_dbl)
diff --git a/arch/m68k/fpsp040/skeleton.S b/arch/m68k/fpsp040/skeleton.S
index 719bcc9ce..1e4d0cf72 100644
--- a/arch/m68k/fpsp040/skeleton.S
+++ b/arch/m68k/fpsp040/skeleton.S
@@ -378,14 +378,12 @@ fpsp_done:
 	beq	Lnotkern
 	rte
 Lnotkern:
-	tstl	SYMBOL_NAME(need_resched)
-	bne	Lmustsched
-	rte
-Lmustsched:
 	SAVE_ALL_INT
 	GET_CURRENT(%d0)
-	bral	SYMBOL_NAME(ret_from_exception)	| deliver signals, reschedule etc..
-
+	tstl	%curptr@(LTASK_NEEDRESCHED)
+	jne	SYMBOL_NAME(ret_from_exception)	| deliver signals,
+						| reschedule etc..
+	RESTORE_ALL
 
 |
 |	mem_write --- write to user or supervisor address space
diff --git a/arch/m68k/fpsp040/slogn.S b/arch/m68k/fpsp040/slogn.S
index b660acd0d..20c5bae9f 100644
--- a/arch/m68k/fpsp040/slogn.S
+++ b/arch/m68k/fpsp040/slogn.S
@@ -17,7 +17,7 @@
 |		result is provably monotonic in double precision.
 |
 |	Speed: The program slogn takes approximately 190 cycles for input 
-|		argument X such that |X-1| >= 1/16, which is the the usual 
+|		argument X such that |X-1| >= 1/16, which is the usual 
 |		situation. For those arguments, slognp1 takes approximately
 |		 210 cycles. For the less common arguments, the program will
 |		 run no worse than 10% slower.
diff --git a/arch/m68k/fpsp040/ssin.S b/arch/m68k/fpsp040/ssin.S
index 978941f4e..96e0e7141 100644
--- a/arch/m68k/fpsp040/ssin.S
+++ b/arch/m68k/fpsp040/ssin.S
@@ -21,7 +21,7 @@
 |		result is provably monotonic in double precision.
 |
 |	Speed: The programs sSIN and sCOS take approximately 150 cycles for
-|		input argument X such that |X| < 15Pi, which is the the usual
+|		input argument X such that |X| < 15Pi, which is the usual
 |		situation. The speed for sSINCOS is approximately 190 cycles.
 |
 |	Algorithm:
diff --git a/arch/m68k/fpsp040/stan.S b/arch/m68k/fpsp040/stan.S
index b7a880a2a..6bd2dda60 100644
--- a/arch/m68k/fpsp040/stan.S
+++ b/arch/m68k/fpsp040/stan.S
@@ -16,7 +16,7 @@
 |		result is provably monotonic in double precision.
 |
 |	Speed: The program sTAN takes approximately 170 cycles for
-|		input argument X such that |X| < 15Pi, which is the the usual
+|		input argument X such that |X| < 15Pi, which is the usual
 |		situation.
 |
 |	Algorithm:
diff --git a/arch/m68k/fpsp040/sto_res.S b/arch/m68k/fpsp040/sto_res.S
index 634b10855..6be6dd5ac 100644
--- a/arch/m68k/fpsp040/sto_res.S
+++ b/arch/m68k/fpsp040/sto_res.S
@@ -3,7 +3,7 @@
 |
 |	Takes the result and puts it in where the user expects it.
 |	Library functions return result in fp0.	If fp0 is not the
-|	users destination register then fp0 is moved to the the
+|	users destination register then fp0 is moved to the
 |	correct floating-point destination register.  fp0 and fp1
 |	are then restored to the original contents. 
 |