diff -urp lame-398-orig/libmp3lame/i386/choose_table.nas lame-398/libmp3lame/i386/choose_table.nas
--- lame-398-orig/libmp3lame/i386/choose_table.nas	2008-07-16 21:47:19.000000000 +0200
+++ lame-398/libmp3lame/i386/choose_table.nas	2008-07-16 21:47:30.000000000 +0200
@@ -111,33 +111,42 @@ choose_table_H
 	dw	0x1d16, 0x1e16, 0x1e17, 0x1f17, 0x1f17
 
 choose_jump_table_L:
-	dd	table_MMX.L_case_0
-	dd	table_MMX.L_case_1
-	dd	table_MMX.L_case_2
-	dd	table_MMX.L_case_3
-	dd	table_MMX.L_case_45
-	dd	table_MMX.L_case_45
-	dd	table_MMX.L_case_67
-	dd	table_MMX.L_case_67
-	dd	table_MMX.L_case_8_15
-	dd	table_MMX.L_case_8_15
-	dd	table_MMX.L_case_8_15
-	dd	table_MMX.L_case_8_15
-	dd	table_MMX.L_case_8_15
-	dd	table_MMX.L_case_8_15
-	dd	table_MMX.L_case_8_15
-	dd	table_MMX.L_case_8_15
+	dd	table_MMX.L_case_0    - choose_table_MMX
+	dd	table_MMX.L_case_1    - choose_table_MMX
+	dd	table_MMX.L_case_2    - choose_table_MMX
+	dd	table_MMX.L_case_3    - choose_table_MMX
+	dd	table_MMX.L_case_45   - choose_table_MMX
+	dd	table_MMX.L_case_45   - choose_table_MMX
+	dd	table_MMX.L_case_67   - choose_table_MMX
+	dd	table_MMX.L_case_67   - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
+	dd	table_MMX.L_case_8_15 - choose_table_MMX
 
 	segment_code
 ;
 ; use MMX
 ;
 
+extern  _GLOBAL_OFFSET_TABLE_
+get_pc.bp:
+	mov ebp, [esp]
+	retn
+
 	align	16
 ; int choose_table(int *ix, int *end, int *s)
 choose_table_MMX:
-	mov	ecx,[esp+4]	;ecx = begin
-	mov	edx,[esp+8]	;edx = end
+	push	ebp
+	call	get_pc.bp
+	add	ebp, _GLOBAL_OFFSET_TABLE_ + $$ - $ wrt ..gotpc
+
+	mov	ecx,[esp+8]	;ecx = begin
+	mov	edx,[esp+12]	;edx = end
 	sub	ecx,edx		;ecx = begin-end(should be minus)
 	test	ecx,8
  	pxor	mm0,mm0		;mm0=[0:0]
@@ -169,13 +178,16 @@ choose_table_MMX:
 
 	cmp	eax,15
 	ja	.with_ESC
-	jmp	[choose_jump_table_L+eax*4]
+	lea	ecx,[ebp + choose_table_MMX wrt ..gotoff]
+	add	ecx,[ebp + choose_jump_table_L+eax*4 wrt ..gotoff]
+	jmp 	ecx
 
 .with_ESC1:
 	emms
-	mov	ecx, [esp+12]	; *s
+	mov	ecx, [esp+16]	; *s
 	mov	[ecx], eax
 	or	eax,-1
+	pop	ebp
 	ret
 
 .with_ESC:
@@ -187,12 +199,12 @@ choose_table_MMX:
 	push	esi
 	bsr	eax, eax
 %assign _P 4*2
-	movq    mm5, [D15_15_15_15]
-	movq	mm6, [D14_14_14_14]
-	movq	mm3, [mul_add]
+	movq    mm5, [ebp + D15_15_15_15 wrt ..gotoff]
+	movq	mm6, [ebp + D14_14_14_14 wrt ..gotoff]
+	movq	mm3, [ebp + mul_add wrt ..gotoff]
 
-	mov	ecx, [esp+_P+4]		; = ix
-;	mov	edx, [esp+_P+8]		; = end
+	mov	ecx, [esp+_P+8]		; = ix
+;	mov	edx, [esp+_P+12]	; = end
 	sub	ecx, edx
 
 	xor	esi, esi	; sum = 0
@@ -209,7 +221,7 @@ choose_table_MMX:
 	psubw	mm7, mm2	; 14$B$h$jBg$-$$$H$-(B linbits_sum++;
 	pmaddwd	mm0, mm3	; {0, 0, y, x}*{1, 16, 1, 16}
 	movd	ebx, mm0
-	mov	esi, [largetbl+ebx*4+(16*16+16)*4]
+	mov	esi, [ebp + largetbl+ebx*4+(16*16+16)*4 wrt ..gotoff]
 
 	jz	.H_dual_exit
 
@@ -224,9 +236,9 @@ choose_table_MMX:
 	pmaddwd	mm0, mm3	; {y, x, y, x}*{1, 16, 1, 16}
 	movd	ebx, mm0
 	punpckhdq	mm0,mm0
-	add	esi, [largetbl+ebx*4+(16*16+16)*4]
+	add	esi, [ebp + largetbl+ebx*4+(16*16+16)*4 wrt ..gotoff]
 	movd	ebx, mm0
-	add	esi, [largetbl+ebx*4+(16*16+16)*4]
+	add	esi, [ebp + largetbl+ebx*4+(16*16+16)*4 wrt ..gotoff]
 	add	ecx, 16
 	psubw	mm7, mm2	; 14$B$h$jBg$-$$$H$-(B linbits_sum++;
 	jnz	.H_dual_lp1
@@ -237,8 +249,8 @@ choose_table_MMX:
 	paddd	mm7,mm1
 	punpckldq	mm7,mm7
 
-	pmaddwd	mm7, [linbits32+eax*8]	; linbits
-	mov	ax, [choose_table_H+eax*2]
+	pmaddwd	mm7, [ebp + linbits32+eax*8 wrt ..gotoff]	; linbits
+	mov	ax, [ebp + choose_table_H+eax*2 wrt ..gotoff]
 
 	movd	ecx, mm7
 	punpckhdq	mm7,mm7
@@ -261,54 +273,57 @@ choose_table_MMX:
 	mov	edx, ecx
 	shr	eax, 8
 .chooseE_s1:
-	mov	ecx, [esp+12] ; *s
+	mov	ecx, [esp+16] ; *s
 	and	eax, 0xff
 	add	[ecx], edx
+	pop	ebp
 	ret
 
 table_MMX.L_case_0:
 	emms
+	pop	ebp
 	ret
 
 table_MMX.L_case_1:
 	emms
-	mov	eax, [esp+12] ; *s
-	mov	ecx, [esp+4] ; *ix
+	mov	eax, [esp+16] ; *s
+	mov	ecx, [esp+8] ; *ix
 	sub	ecx, edx
 	push	ebx
 .lp:
 	mov	ebx, [edx+ecx]
 	add	ebx, ebx
 	add	ebx, [edx+ecx+4]
-	movzx	ebx, byte [ebx+t1l]
+	movzx	ebx, byte [ebp + ebx+t1l wrt ..gotoff]
 	add	[eax], ebx
 	add	ecx, 8
 	jnz	.lp
 	pop	ebx
 	mov	eax, 1
+	pop	ebp
 	ret
 
 table_MMX.L_case_45:
 	push	dword 7
-	mov	ecx, tableABC+9*8
+	lea	ecx, [ebp + tableABC+9*8 wrt ..gotoff]
 	jmp	from3
 
 table_MMX.L_case_67:
 	push	dword 10
-	mov	ecx, tableABC
+	lea	ecx, [ebp + tableABC wrt ..gotoff]
 	jmp	from3
 
 table_MMX.L_case_8_15:
 	push	dword 13
-	mov	ecx, tableDEF
+	lea	ecx, [ebp + tableDEF wrt ..gotoff]
 from3:
-	mov	eax,[esp+8]	;eax = *begin
-;	mov	edx,[esp+12]	;edx = *end
+	mov	eax,[esp+12]	;eax = *begin
+;	mov	edx,[esp+16]	;edx = *end
 
 	push	ebx
 	sub	eax, edx
 
-	movq	mm5,[mul_add]
+	movq	mm5,[ebp + mul_add wrt ..gotoff]
 	pxor	mm2,mm2	;mm2 = sum
 
 	test	eax, 8
@@ -361,22 +376,23 @@ from3:
 .choose3_s2:
 	pop	ecx
 	add	eax, ecx
-	mov	ecx, [esp+12] ; *s
+	mov	ecx, [esp+16] ; *s
 	add	[ecx], edx
+	pop	ebp
 	ret
 
 table_MMX.L_case_2:
 	push	dword 2
-	mov	ecx,table23
-	pmov	mm5,[mul_add23]
+	lea	ecx,[ebp + table23 wrt ..gotoff]
+	pmov	mm5,[ebp + mul_add23 wrt ..gotoff]
 	jmp	from2
 table_MMX.L_case_3:
 	push	dword 5
-	mov	ecx,table56
-	pmov	mm5,[mul_add56]
+	lea	ecx,[ebp + table56 wrt ..gotoff]
+	pmov	mm5,[ebp + mul_add56 wrt ..gotoff]
 from2:
-	mov	eax,[esp+8]	;eax = *begin
-;	mov	edx,[esp+12]	;edx = *end
+	mov	eax,[esp+12]	;eax = *begin
+;	mov	edx,[esp+16]	;edx = *end
 	push	ebx
 	push	edi
 
@@ -426,8 +442,9 @@ from2:
 	mov	edx, ecx
 	inc	eax
 .choose2_s1:
-	mov	ecx, [esp+12] ; *s
+	mov	ecx, [esp+16] ; *s
 	add	[ecx], edx
+	pop	ebp
 	ret
 
 	end
diff -urp lame-398-orig/libmp3lame/i386/fft3dn.nas lame-398/libmp3lame/i386/fft3dn.nas
--- lame-398-orig/libmp3lame/i386/fft3dn.nas	2008-07-16 21:47:19.000000000 +0200
+++ lame-398/libmp3lame/i386/fft3dn.nas	2008-07-16 21:47:30.000000000 +0200
@@ -24,26 +24,35 @@ D_1_0_0_0	dd	0.0		, 1.0
 
 	segment_code
 
+extern  _GLOBAL_OFFSET_TABLE_
+get_pc.bp:
+	mov ebp, [esp]
+	retn
+
 ;void fht_3DN(float *fz, int nn);
 
 proc	fht_3DN
 
 	pushd	ebp, ebx, esi, edi
 
-	mov	r0, [esp+20]		;fi
-	mov	r1, [esp+24]		;r1 = nn
-	sub	esp, 16
+	sub	esp, 20
+
+	call	get_pc.bp
+	add	ebp, _GLOBAL_OFFSET_TABLE_ + $$ - $ wrt ..gotpc
 
+	mov	r0, [esp+40]		;fi
+	mov	r1, [esp+44]		;r1 = nn
+	lea	r3, [ebp + costab wrt ..gotoff]		;tri = costab
+	lea	r4, [r0+r1*8]		;r4 = fn = &fz[n]
+	mov	[esp+16], r4
 	mov	r4, 8			;kx = k1/2
-	mov	r3, costab		;tri = costab
-	lea	r6, [r0+r1*8]		;r6 = fn = &fz[n]
 
 	pmov	mm7, [r3]
 
 	loopalign 16
 .do1
 	lea	r3, [r3+16]	;tri += 2;
-	pmov	mm6, [costab+8]
+	pmov	mm6, [ebp + costab+8 wrt ..gotoff]
 	lea	r2, [r4+r4*2]		;k3*fsize/2
 	mov	r5, 4		;i = 1*fsize
 
@@ -104,7 +113,7 @@ proc	fht_3DN
 	pmovd	[r1+r4*4], mm4	;gi[k2]
 	puphdq	mm4, mm4
 
-	cmp	r0, r6
+	cmp	r0, [esp + 16]
 	pmovd	[r1+r4*2], mm0	;gi[k1]
 	pmovd	[r1+r2*2], mm4	;gi[k3]
 
@@ -119,12 +128,12 @@ proc	fht_3DN
 ; mm7 = 0x800000000 | 0
 ;
 	pmov	mm1, mm6
-	mov	r0, [esp+36]	; fz
+	mov	r0, [esp+40]	; fz
 	puphdq	mm1, mm1	; c1 | c1
 	lea	r1, [r0+r4*2]
 	pfadd	mm1, mm1	; c1+c1 | c1+c1
 	pfmul	mm1, mm6	; 2*c1*c1 | 2*c1*s1
-	pfsub	mm1, [D_1_0_0_0] ; 2*c1*c1-1.0 | 2*c1*s1 = -c2 | s2
+	pfsub	mm1, [ebp + D_1_0_0_0 wrt ..gotoff] ; 2*c1*c1-1.0 | 2*c1*s1 = -c2 | s2
 
 	pmov	mm0, mm1
 	pxor	mm7, mm6	; c1 | -s1
@@ -134,7 +143,7 @@ proc	fht_3DN
 	puphdq	mm0, mm2	; s2 | c2
 	puphdq	mm6, mm3	;-s1 | c1
 
-	pxor	mm0, [costab]	; c2 | -s2
+	pxor	mm0, [ebp + costab wrt ..gotoff]	; c2 | -s2
 
 ; mm0 =  s2| c2
 ; mm1 = -c2| s2
@@ -233,7 +242,7 @@ proc	fht_3DN
 
 	lea	r0, [r0+r4*8]
 	lea	r1, [r1+r4*8]
-	cmp	r0, r6
+	cmp	r0, [esp + 16]
 	pmov	mm0, [esp]
 	pmov	mm1, [esp+8]
 
@@ -249,17 +258,17 @@ proc	fht_3DN
 	pfsub	mm6, mm7	; c1*a-s1*b | s1*a+c1*b
 	pupldq	mm7,mm6
 	puphdq	mm6,mm7
-	pmov	mm7, [costab]
+	pmov	mm7, [ebp + costab wrt ..gotoff]
 	jb near	.for
 
-	mov	r0, [esp+36]	;fi
-	cmp	r4, [esp+36+4]
+	mov	r0, [esp+40]	;fi
+	cmp	r4, [esp+40+4]
 	lea	r4, [r4*4]	;kx *= 4
 
 	jb near	.do1
 .exitttt
 	femms
-	add	esp,16
+	add	esp,20
 	popd	ebp, ebx, esi, edi
 endproc
 
@@ -270,20 +279,24 @@ proc	fht_E3DN
 
 	pushd	ebp, ebx, esi, edi
 
-	mov	r0, [esp+20]		;fi
-	mov	r1, [esp+24]		;r1 = nn
-	sub	esp, 16
+	sub	esp, 20
+
+	call	get_pc.bp
+	add	ebp, _GLOBAL_OFFSET_TABLE_ + $$ - $ wrt ..gotpc
 
+	mov	r0, [esp+40]		;fi
+	mov	r1, [esp+44]		;r1 = nn
+	lea	r3, [ebp + costab wrt ..gotoff]		;tri = costab
+	lea	r4, [r0+r1*8]		;r4 = fn = &fz[n]
+	mov	[esp+16], r4
 	mov	r4, 8			;kx = k1/2
-	mov	r3, costab		;tri = costab
-	lea	r6, [r0+r1*8]		;r6 = fn = &fz[n]
 
 	pmov	mm7, [r3]
 
 	loopalign 16
 .do1
 	lea	r3, [r3+16]	;tri += 2;
-	pmov	mm6, [costab+8]
+	pmov	mm6, [ebp + costab+8 wrt ..gotoff]
 	lea	r2, [r4+r4*2]		;k3*fsize/2
 	mov	r5, 4		;i = 1*fsize
 
@@ -324,7 +337,7 @@ proc	fht_E3DN
 	pfadd	mm3, mm4	;f0+f2|f1+f3
 	pfsub	mm5, mm4	;f0-f2|f1-f3
 
-	cmp	r0, r6
+	cmp	r0, [esp + 16]
 	pmovd	[r1+r4*2], mm3	;gi[k1]
 	pmovd	[r1+r2*2], mm5	;gi[k3]
 	puphdq	mm3, mm3
@@ -343,12 +356,12 @@ proc	fht_E3DN
 ; mm7 = 0x800000000 | 0
 ;
 	pmov	mm5, mm6
-	mov	r0, [esp+36]	; fz
+	mov	r0, [esp+40]	; fz
 	puphdq	mm5, mm5	; c1 | c1
 	lea	r1, [r0+r4*2]
 	pfadd	mm5, mm5	; c1+c1 | c1+c1
 	pfmul	mm5, mm6	; 2*c1*c1 | 2*c1*s1
-	pfsub	mm5, [D_1_0_0_0] ; 2*c1*c1-1.0 | 2*c1*s1 = -c2 | s2
+	pfsub	mm5, [ebp + D_1_0_0_0 wrt ..gotoff] ; 2*c1*c1-1.0 | 2*c1*s1 = -c2 | s2
 
 	pswapd	mm4, mm5	; s2 |-c2
 	pxor	mm4, mm7	; s2 | c2
@@ -447,7 +460,7 @@ proc	fht_E3DN
 
 	lea	r0, [r0+r4*8]
 	lea	r1, [r1+r4*8]
-	cmp	r0, r6
+	cmp	r0, [esp + 16]
 	pmov	mm4, [esp]
 	pmov	mm5, [esp+8]
 
@@ -462,16 +475,16 @@ proc	fht_E3DN
 
 	pfsub	mm6, mm7	; c1*a-s1*b | s1*a+c1*b
 	pswapd	mm6, mm6 ; ???	; s1*a+c1*b | c1*a-s1*b
-	pmov	mm7, [costab]
+	pmov	mm7, [ebp + costab wrt ..gotoff]
 	jb near	.for
 
-	mov	r0, [esp+36]	;fi
-	cmp	r4, [esp+36+4]
+	mov	r0, [esp+40]	;fi
+	cmp	r4, [esp+40+4]
 	lea	r4, [r4*4]	;kx *= 4
 
 	jb near	.do1
 .exitttt
 	femms
-	add	esp,16
+	add	esp,20
 	popd	ebp, ebx, esi, edi
 endproc
diff -urp lame-398-orig/libmp3lame/i386/fftsse.nas lame-398/libmp3lame/i386/fftsse.nas
--- lame-398-orig/libmp3lame/i386/fftsse.nas	2008-07-16 21:47:19.000000000 +0200
+++ lame-398/libmp3lame/i386/fftsse.nas	2008-07-16 21:48:10.000000000 +0200
@@ -25,6 +25,12 @@ costab_fft:
 S_SQRT2	dd	1.414213562
 
 	segment_code
+
+extern  _GLOBAL_OFFSET_TABLE_
+get_pc.bp:
+	mov ebp, [esp]
+	retn
+
 ;------------------------------------------------------------------------
 ;	by K. SAKAI
 ;	99/08/18	PIII 23k[clk]
@@ -40,15 +46,20 @@ fht_SSE:
 	push	esi
 	push	edi
 	push	ebp
-%assign _P 4*4
+
+%assign _P 4*5
 
 	;2つ目のループ
-	mov	eax,[esp+_P+4]	;eax=fz
-	mov	ebp,[esp+_P+8]	;=n
+	mov	eax,[esp+_P+0]	;eax=fz
+	mov	ebp,[esp+_P+4]	;=n
 	shl	ebp,3
 	add	ebp,eax		; fn  = fz + n, この関数終了まで不変
+	push	ebp
+
+	call	get_pc.bp
+	add	ebp, _GLOBAL_OFFSET_TABLE_ + $$ - $ wrt ..gotpc
 
-	lea	ecx,[costab_fft]
+	lea	ecx,[ebp + costab_fft wrt ..gotoff]
 	xor	eax,eax
 	mov	al,8		; =k1=1*(sizeof float)	// 4, 16, 64, 256,...
 .lp2:				; do{
@@ -101,12 +112,12 @@ fht_SSE:
 ;                       gi[k3] = g1     - g3;
 	fld	dword [edi]
 	fadd	dword [edi+eax*2]
-	fld	dword [S_SQRT2]
+	fld	dword [ebp + S_SQRT2 wrt ..gotoff]
 	fmul	dword [edi+eax*4]
 
 	fld	dword [edi]
 	fsub	dword [edi+eax*2]
-	fld	dword [S_SQRT2]
+	fld	dword [ebp + S_SQRT2 wrt ..gotoff]
 	fmul	dword [edi+edx*2]
 
 	fld	st1
@@ -121,7 +132,7 @@ fht_SSE:
 	fsubp	st1,st0
 	fstp	dword [edi+eax*4]
 
-	cmp	ebx,ebp
+	cmp	ebx,[esp]
 	jl	near .lp20		; while (fi<fn);
 
 
@@ -136,17 +147,17 @@ fht_SSE:
 ;                       s2 = c1*s1 + s1*c1 = 2*s1*c1;
 	shufps	xmm7,xmm7,R4(1,0,0,1)
 	movss	xmm5,xmm7		; = { --,  --,  --, s1}
-	xorps	xmm7,[Q_MMPP]	; = {-s1, -c1, +c1, +s1} -> 必要
+	xorps	xmm7,[ebp + Q_MMPP wrt ..gotoff]	; = {-s1, -c1, +c1, +s1} -> 必要
 
 	addss	xmm5,xmm5		; = (--, --,  --, 2*s1)
 	add	esi,4		; esi = fi = fz + i
 	shufps	xmm5,xmm5,R4(0,0,0,0)	; = (2*s1, 2*s1, 2*s1, 2*s1)
 	mulps	xmm5,xmm6		; = (2*s1*c1, 2*s1*s1, 2*s1*s1, 2*s1*c1)
-	subps	xmm5,[D_1100]		; = (--, 2*s1*s1-1, --, 2*s1*c1) = {-- -c2 -- s2}
+	subps	xmm5,[ebp + D_1100 wrt ..gotoff]		; = (--, 2*s1*s1-1, --, 2*s1*c1) = {-- -c2 -- s2}
 	movaps	xmm4,xmm5
 	shufps	xmm5,xmm5,R4(2,0,2,0)	; = {-c2, s2, -c2, s2} -> 必要
 
-	xorps	xmm4,[Q_MMPP]		; = {--, c2, --, s2}
+	xorps	xmm4,[ebp + Q_MMPP wrt ..gotoff]		; = {--, c2, --, s2}
 	shufps	xmm4,xmm4,R4(0,2,0,2)	; = {s2, c2, s2, c2} -> 必要
 
 	loopalign	16
@@ -222,7 +233,7 @@ fht_SSE:
 	movss	[edi+eax*4],xmm2
 	movss	[esi+edx*2],xmm0
 	lea	esi,[esi + eax*8] ; fi += (k1 * 4);
-	cmp	esi,ebp
+	cmp	esi,[esp]
 	jl	near .lp21		; while (fi<fn);
 
 
@@ -247,7 +258,7 @@ fht_SSE:
 	shufps	xmm0,xmm0,R4(1,1,0,0)	; = {t_s, t_s, t_c, t_c}
 	mulps	xmm6,xmm0	; = {c3*ts, s3*ts, s3*tc, c3*tc}
 	movhlps	xmm4,xmm6	; = {--,    --,    c3*ts, s3*ts}
-	xorps	xmm4,[Q_MPMP]	; = {--,    --,   -c3*ts, s3*ts}
+	xorps	xmm4,[ebp + Q_MPMP wrt ..gotoff]	; = {--,    --,   -c3*ts, s3*ts}
 	subps	xmm6,xmm4	; = {-,-, c3*ts+s3*tc, c3*tc-s3*ts}={-,-,s1,c1}
 
 ;                       c3 = c1*t_c - s1*t_s;
@@ -255,7 +266,7 @@ fht_SSE:
 	shufps	xmm6,xmm6,0x14	; = {c1, s1, s1, c1}
 	mulps	xmm0,xmm6	; = {ts*c1 ts*s1 tc*s1 tc*c1}
 	movhlps	xmm3,xmm0
-	xorps	xmm3,[Q_MPMP]
+	xorps	xmm3,[ebp + Q_MPMP wrt ..gotoff]
 	subps	xmm0,xmm3	; = {--, --, s3, c3}
 
 ; {s2 s4 c4 c2} = {2*s1*c1 2*s3*c3 1-2*s3*s3 1-2*s1*s1}
@@ -268,7 +279,7 @@ fht_SSE:
 	sub	edi,ebx			; edi = fz - i/2
 	mulps	xmm7, xmm6		; {s1*s1*2, s3*s3*2, s3*c3*2, s1*c1*2}
 	lea	esi,[edi + ebx*2]	; esi = fi = fz +i/2
-	subps	xmm7, [D_1100]		; {-c2, -c4, s4, s2}
+	subps	xmm7, [ebp + D_1100 wrt ..gotoff]		; {-c2, -c4, s4, s2}
 	lea	edi,[edi + eax*2-4]	; edi = gi = fz +k1-i/2
 
 ;                       fi = fz +i;
@@ -286,7 +297,7 @@ fht_SSE:
 ;                               d       = s2*fi[k3  ] - c2*gi[k3  ];
 
 	movaps	xmm4,xmm7	; = {-c2 -c4  s4  s2}
-	xorps	xmm4,[Q_MMPP]	; = { c2  c4  s4  s2}
+	xorps	xmm4,[ebp + Q_MMPP wrt ..gotoff]	; = { c2  c4  s4  s2}
 	shufps	xmm4,xmm4,0x1B	; = { s2  s4  c4  c2}
 	movlps	xmm0,[esi+eax*2]
 	movlps	xmm1,[edi+eax*2]
@@ -390,7 +401,7 @@ fht_SSE:
 ;                               fi     += k4;
 	lea	edi,[edi + eax*8] ; gi += (k1 * 4);
 	lea	esi,[esi + eax*8] ; fi += (k1 * 4);
-	cmp	esi,ebp
+	cmp	esi,[esp]
 	jl	near .lp220		; while (fi<fn);
 ;                       } while (fi<fn);
 
@@ -405,6 +416,7 @@ fht_SSE:
 	cmp	eax,[esp+_P+8]	; while ((k1 * 4)<n);
 	jle	near .lp2
 	pop	ebp
+	pop	ebp
 	pop	edi
 	pop	esi
 	pop	ebx