diff --git a/arch/s390/net/bpf_jit.S b/arch/s390/net/bpf_jit.S
index ba44c9f..a1c917d 100644
--- a/arch/s390/net/bpf_jit.S
+++ b/arch/s390/net/bpf_jit.S
@@ -1,134 +1,115 @@
 /*
  * BPF Jit compiler for s390, help functions.
  *
- * Copyright IBM Corp. 2012
+ * Copyright IBM Corp. 2012,2015
  *
  * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *	      Michael Holzheu <holzheu@linux.vnet.ibm.com>
  */
+
 #include <linux/linkage.h>
+#include "bpf_jit.h"
 
 /*
  * Calling convention:
- * registers %r2, %r6-%r8, %r10-%r11, %r13, %r15 are call saved
- *   %r2: skb pointer
- *   %r3: offset parameter
- *   %r5: BPF A accumulator
- *   %r8: return address
- *   %r9: save register for skb pointer
- *   %r10: skb->data
- *   %r11: skb->len - skb->data_len (headlen)
- *   %r12: BPF X accumulator
+ * registers %r7-%r10, %r11,%r13, and %r15 are call saved
+ *
+ * Input (64 bit):
+ *   %r3 (%b2) = offset into skb data
+ *   %r6 (%b5) = return address
+ *   %r7 (%b6) = skb pointer
+ *   %r12      = skb data pointer
+ *
+ * Output:
+ *   %r14= %b0 = return value (read skb value)
+ *
+ * Work registers: %r2,%r4,%r5,%r14
  *
  * skb_copy_bits takes 4 parameters:
  *   %r2 = skb pointer
  *   %r3 = offset into skb data
  *   %r4 = pointer to temp buffer
  *   %r5 = length to copy
+ *   Return value in %r2: 0 = ok
+ *
+ * bpf_internal_load_pointer_neg_helper takes 3 parameters:
+ *   %r2 = skb pointer
+ *   %r3 = offset into data
+ *   %r4 = length to copy
+ *   Return value in %r2: Pointer to data
  */
-#define SKBDATA	%r8
 
-	/* A = *(u32 *) (skb->data+K+X) */
-ENTRY(sk_load_word_ind)
-	ar	%r3,%r12		# offset += X
-	bmr	%r8			# < 0 -> return with cc
+#define SKF_MAX_NEG_OFF	-0x200000	/* SKF_LL_OFF from filter.h */
 
-	/* A = *(u32 *) (skb->data+K) */
-ENTRY(sk_load_word)
-	llgfr	%r1,%r3			# extend offset
-	ahi	%r3,4			# offset + 4
-	clr	%r11,%r3		# hlen <= offset + 4 ?
-	jl	sk_load_word_slow
-	l	%r5,0(%r1,%r10)		# get word from skb
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
+/*
+ * Load SIZE bytes from SKB
+ */
+#define sk_load_common(NAME, SIZE, LOAD)				\
+ENTRY(sk_load_##NAME);							\
+	ltgr	%r3,%r3;		/* Is offset negative? */	\
+	jl	sk_load_##NAME##_slow_neg;				\
+ENTRY(sk_load_##NAME##_pos);						\
+	aghi	%r3,SIZE;		/* Offset + SIZE */		\
+	clg	%r3,STK_OFF_HLEN(%r15);	/* Offset + SIZE > hlen? */	\
+	jh	sk_load_##NAME##_slow;					\
+	LOAD	%r14,-SIZE(%r3,%r12);	/* Get data from skb */		\
+	b	OFF_OK(%r6);		/* Return */			\
+									\
+sk_load_##NAME##_slow:;							\
+	lgr	%r2,%r7;		/* Arg1 = skb pointer */	\
+	aghi	%r3,-SIZE;		/* Arg2 = offset */		\
+	la	%r4,STK_OFF_TMP(%r15);	/* Arg3 = temp bufffer */	\
+	lghi	%r5,SIZE;		/* Arg4 = size */		\
+	brasl	%r14,skb_copy_bits;	/* Get data from skb */		\
+	LOAD	%r14,STK_OFF_TMP(%r15);	/* Load from temp bufffer */	\
+	ltgr	%r2,%r2;		/* Set cc to (%r2 != 0) */	\
+	br	%r6;			/* Return */
 
-sk_load_word_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,160(%r15)		# pointer to temp buffer
-	lghi	%r5,4			# 4 bytes
-	brasl	%r14,skb_copy_bits	# get data from skb
-	l	%r5,160(%r15)		# load result from temp buffer
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
+sk_load_common(word, 4, llgf)	/* r14 = *(u32 *) (skb->data+offset) */
+sk_load_common(half, 2, llgh)	/* r14 = *(u16 *) (skb->data+offset) */
 
-	/* A = *(u16 *) (skb->data+K+X) */
-ENTRY(sk_load_half_ind)
-	ar	%r3,%r12		# offset += X
-	bmr	%r8			# < 0 -> return with cc
-
-	/* A = *(u16 *) (skb->data+K) */
-ENTRY(sk_load_half)
-	llgfr	%r1,%r3			# extend offset
-	ahi	%r3,2			# offset + 2
-	clr	%r11,%r3		# hlen <= offset + 2 ?
-	jl	sk_load_half_slow
-	llgh	%r5,0(%r1,%r10)		# get half from skb
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
-
-sk_load_half_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,162(%r15)		# pointer to temp buffer
-	lghi	%r5,2			# 2 bytes
-	brasl	%r14,skb_copy_bits	# get data from skb
-	xc	160(2,%r15),160(%r15)
-	l	%r5,160(%r15)		# load result from temp buffer
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
-
-	/* A = *(u8 *) (skb->data+K+X) */
-ENTRY(sk_load_byte_ind)
-	ar	%r3,%r12		# offset += X
-	bmr	%r8			# < 0 -> return with cc
-
-	/* A = *(u8 *) (skb->data+K) */
+/*
+ * Load 1 byte from SKB (optimized version)
+ */
+	/* r14 = *(u8 *) (skb->data+offset) */
 ENTRY(sk_load_byte)
-	llgfr	%r1,%r3			# extend offset
-	clr	%r11,%r3		# hlen < offset ?
-	jle	sk_load_byte_slow
-	lhi	%r5,0
-	ic	%r5,0(%r1,%r10)		# get byte from skb
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
+	ltgr	%r3,%r3			# Is offset negative?
+	jl	sk_load_byte_slow_neg
+ENTRY(sk_load_byte_pos)
+	clg	%r3,STK_OFF_HLEN(%r15)	# Offset >= hlen?
+	jnl	sk_load_byte_slow
+	llgc	%r14,0(%r3,%r12)	# Get byte from skb
+	b	OFF_OK(%r6)		# Return OK
 
 sk_load_byte_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,163(%r15)		# pointer to temp buffer
-	lghi	%r5,1			# 1 byte
-	brasl	%r14,skb_copy_bits	# get data from skb
-	xc	160(3,%r15),160(%r15)
-	l	%r5,160(%r15)		# load result from temp buffer
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
+	lgr	%r2,%r7			# Arg1 = skb pointer
+					# Arg2 = offset
+	la	%r4,STK_OFF_TMP(%r15)	# Arg3 = pointer to temp buffer
+	lghi	%r5,1			# Arg4 = size (1 byte)
+	brasl	%r14,skb_copy_bits	# Get data from skb
+	llgc	%r14,STK_OFF_TMP(%r15)	# Load result from temp buffer
+	ltgr	%r2,%r2			# Set cc to (%r2 != 0)
+	br	%r6			# Return cc
 
-	/* X = (*(u8 *)(skb->data+K) & 0xf) << 2 */
-ENTRY(sk_load_byte_msh)
-	llgfr	%r1,%r3			# extend offset
-	clr	%r11,%r3		# hlen < offset ?
-	jle	sk_load_byte_msh_slow
-	lhi	%r12,0
-	ic	%r12,0(%r1,%r10)	# get byte from skb
-	nill	%r12,0x0f
-	sll	%r12,2
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
+#define sk_negative_common(NAME, SIZE, LOAD)				\
+sk_load_##NAME##_slow_neg:;						\
+	cgfi	%r3,SKF_MAX_NEG_OFF;					\
+	jl	bpf_error;						\
+	lgr	%r2,%r7;		/* Arg1 = skb pointer */	\
+					/* Arg2 = offset */		\
+	lghi	%r4,SIZE;		/* Arg3 = size */		\
+	brasl	%r14,bpf_internal_load_pointer_neg_helper;		\
+	ltgr	%r2,%r2;						\
+	jz	bpf_error;						\
+	LOAD	%r14,0(%r2);		/* Get data from pointer */	\
+	xr	%r3,%r3;		/* Set cc to zero */		\
+	br	%r6;			/* Return cc */
 
-sk_load_byte_msh_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,163(%r15)		# pointer to temp buffer
-	lghi	%r5,1			# 1 byte
-	brasl	%r14,skb_copy_bits	# get data from skb
-	xc	160(3,%r15),160(%r15)
-	l	%r12,160(%r15)		# load result from temp buffer
-	nill	%r12,0x0f
-	sll	%r12,2
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
+sk_negative_common(word, 4, llgf)
+sk_negative_common(half, 2, llgh)
+sk_negative_common(byte, 1, llgc)
+
+bpf_error:
+# force a return 0 from jit handler
+	ltgr	%r15,%r15	# Set condition code
+	br	%r6