VADDPD (AX), X10, X28       // 6261ad085820
	VADDPD (CX)(AX*1), X29, X0  // 62f19500580401
	VADDPD (CX)(AX*1), X1, X28  // 6261f508582401
	VADDPD (CX)(AX*1), X29, X28 // 62619500582401
	VADDPD (CX)(AX*1), X29, X10 // 62719500581401
	VADDPD (CX)(AX*1), X10, X28 // 6261ad08582401
	VADDPD (CX)(AX*2), X29, X0  // 62f19500580441
	VADDPD (CX)(AX*2), X1, X28  // 6261f508582441
	VADDPD (CX)(AX*2), X29, X28 // 62619500582441

	{"$~15", "$-16"},
	{"(((8)&0xf)*4)(SP)", "32(SP)"},
	{"(((8-14)&0xf)*4)(SP)", "40(SP)"},
	{"(6+8)(AX)", "14(AX)"},
	{"(8*4)(BP)", "32(BP)"},
	{"(AX)", "(AX)"},
	{"(AX)(CX*8)", "(AX)(CX*8)"},
	{"(BP)(CX*4)", "(BP)(CX*4)"},
	{"(BP)(DX*4)", "(BP)(DX*4)"},
	{"(BP)(R8*4)", "(BP)(R8*4)"},
	{"(BX)", "(BX)"},
	{"(DI)", "(DI)"},
	{"(DI)(BX*1)", "(DI)(BX*1)"},
	{"(DX)", "(DX)"},
	{"(R9)", "(R9)"},

<code>g</code> pointer into.
</p>

<p>
For example, the sequence to load <code>g</code> and <code>m</code>
using <code>CX</code> looks like this:
</p>

<pre>
#include "go_tls.h"
#include "go_asm.h"
...
get_tls(CX)
MOVL	g(CX), AX     // Move g into AX.
MOVL	g_m(AX), BX   // Move g.m into BX.
</pre>

<p>

	CMPL runtime·writeBarrier(SB), $0
	MOVQ (CX)(R15*1), AX // ERROR "when dynamic linking, R15 is clobbered by a global variable access and is used here"
	RET
TEXT ·a19(SB), 0, $0-0
	CMPL runtime·writeBarrier(SB), $0
	MOVQ AX, (R15) // ERROR "when dynamic linking, R15 is clobbered by a global variable access and is used here"
	RET
TEXT ·a20(SB), 0, $0-0
	CMPL runtime·writeBarrier(SB), $0

	{"", "", -1},
	{"", "a", -1},
	{"", "abc", -1},
	{"a", "", -1},
	{"a", "a", 0},
	{"\x80", "\xffb", 0},
	{"aaa", "a", 0},
	{"abc", "xyz", -1},
	{"abc", "xcz", 2},
	{"ab☺c", "x☺yz", 2},
	{"a☺b☻c☹d", "cx", len("a☺b☻")},
	{"a☺b☻c☹d", "uvw☻xyz", len("a☺b")},
	{"aRegExp*", ".(|)*+?^$[]", 7},
	{dots + dots + dots, " ", -1},
	{"012abcba210", "\xffb", 4},
	{"012\x80bcb\x80210", "\xffb", 3},