liblzma: Fix x86-64 movzw compatibility in range_decoder.h

Support for instruction "movzw" without suffix in "GNU as" was
added in commit [1] and stabilized in binutils 2.27, released
in August 2016. Earlier systems don't accept this instruction
without a suffix, making range_decoder.h's inline assembly
unable to build on old systems such as Ubuntu 16.04, creating
error messages like:

    lzma_decoder.c: Assembler messages:
    lzma_decoder.c:371: Error: no such instruction: `movzw 2(%r11),%esi'
    lzma_decoder.c:373: Error: no such instruction: `movzw 4(%r11),%edi'
    lzma_decoder.c:388: Error: no such instruction: `movzw 6(%r11),%edx'
    lzma_decoder.c:398: Error: no such instruction: `movzw (%r11,%r14,4),%esi'

Change "movzw" to "movzwl" for compatibility.

[1] https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=c07315e0c610e0e3317b4c02266f81793df253d2

Suggested-by: Lasse Collin <[email protected]>
Tested-by: Yifeng Li <[email protected]>
Signed-off-by: Yifeng Li <[email protected]>
Fixes: 3182a33
Fixes: #121
Closes: #136

src/liblzma/rangecoder/range_decoder.h

@@ -592,13 +592,13 @@ do { \
 // *_only = rc_asm_y or _n to include or exclude code marked with them
 #define rc_asm_bittree(a, b, first_only, middle_only, last_only) \
 	first_only( \
-		"movzw	2(%[probs_base]), %[prob" #a "]\n\t" \
+		"movzwl	2(%[probs_base]), %[prob" #a "]\n\t" \
 		"mov	$2, %[symbol]\n\t" \
-		"movzw	4(%[probs_base]), %[prob" #b "]\n\t" \
+		"movzwl	4(%[probs_base]), %[prob" #b "]\n\t" \
 	) \
 	middle_only( \
 		/* Note the scaling of 4 instead of 2: */ \
-		"movzw	(%[probs_base], %q[symbol], 4), %[prob" #b "]\n\t" \
+		"movzwl	(%[probs_base], %q[symbol], 4), %[prob" #b "]\n\t" \
 	) \
 	last_only( \
 		"add	%[symbol], %[symbol]\n\t" \
@@ -610,11 +610,11 @@ do { \
 		"cmovae	%[t0], %[range]\n\t" \
 		\
 	first_only( \
-		"movzw	6(%[probs_base]), %[t0]\n\t" \
+		"movzwl	6(%[probs_base]), %[t0]\n\t" \
 		"cmovae	%[t0], %[prob" #b "]\n\t" \
 	) \
 	middle_only( \
-		"movzw	2(%[probs_base], %q[symbol], 4), %[t0]\n\t" \
+		"movzwl	2(%[probs_base], %q[symbol], 4), %[t0]\n\t" \
 		"lea	(%q[symbol], %q[symbol]), %[symbol]\n\t" \
 		"cmovae	%[t0], %[prob" #b "]\n\t" \
 	) \
@@ -716,12 +716,12 @@ do { \
 #define rc_asm_bittree_rev(a, b, add, dcur, dnext0, dnext1, \
 		first_only, middle_only, last_only) \
 	first_only( \
-		"movzw	2(%[probs_base]), %[prob" #a "]\n\t" \
+		"movzwl	2(%[probs_base]), %[prob" #a "]\n\t" \
 		"xor	%[symbol], %[symbol]\n\t" \
-		"movzw	4(%[probs_base]), %[prob" #b "]\n\t" \
+		"movzwl	4(%[probs_base]), %[prob" #b "]\n\t" \
 	) \
 	middle_only( \
-		"movzw	" #dnext0 "(%[probs_base], %q[symbol], 2), " \
+		"movzwl	" #dnext0 "(%[probs_base], %q[symbol], 2), " \
 			"%[prob" #b "]\n\t" \
 	) \
 		\
@@ -731,11 +731,11 @@ do { \
 		"cmovae	%[t0], %[range]\n\t" \
 		\
 	first_only( \
-		"movzw	6(%[probs_base]), %[t0]\n\t" \
+		"movzwl	6(%[probs_base]), %[t0]\n\t" \
 		"cmovae	%[t0], %[prob" #b "]\n\t" \
 	) \
 	middle_only( \
-		"movzw	" #dnext1 "(%[probs_base], %q[symbol], 2), %[t0]\n\t" \
+		"movzwl	" #dnext1 "(%[probs_base], %q[symbol], 2), %[t0]\n\t" \
 		"cmovae	%[t0], %[prob" #b "]\n\t" \
 	) \
 		\
@@ -788,7 +788,7 @@ do { \
 	uint32_t t_index; \
 	\
 	__asm__( \
-		"movzw	(%[probs_base], %q[symbol], 2), %[prob]\n\t" \
+		"movzwl	(%[probs_base], %q[symbol], 2), %[prob]\n\t" \
 		"mov	%[symbol], %[index]\n\t" \
 		\
 		"add	%[dest], %[t2]\n\t" \
@@ -844,7 +844,7 @@ do { \
 		"and	%[offset], %[match_bit]\n\t" \
 		"add	%[match_bit], %[symbol]\n\t" \
 		\
-		"movzw	(%[probs_base], %q[symbol], 2), %[prob]\n\t" \
+		"movzwl	(%[probs_base], %q[symbol], 2), %[prob]\n\t" \
 		\
 		"add	%[symbol], %[symbol]\n\t" \
 		\