+{ "cpop1", F3(2, 0x36, 0), F3(~2, ~0x36, ~1), "[1+2],d", F_ALIAS, v6notv9 },
+{ "cpop2", F3(2, 0x37, 0), F3(~2, ~0x37, ~1), "[1+2],d", F_ALIAS, v6notv9 },
+
+/* sparclet specific insns */
+
+COMMUTEOP ("umac", 0x3e, sparclet),
+COMMUTEOP ("smac", 0x3f, sparclet),
+COMMUTEOP ("umacd", 0x2e, sparclet),
+COMMUTEOP ("smacd", 0x2f, sparclet),
+COMMUTEOP ("umuld", 0x09, sparclet),
+COMMUTEOP ("smuld", 0x0d, sparclet),
+
+{ "shuffle", F3(2, 0x2d, 0), F3(~2, ~0x2d, ~0)|ASI(~0), "1,2,d", 0, sparclet },
+{ "shuffle", F3(2, 0x2d, 1), F3(~2, ~0x2d, ~1), "1,i,d", 0, sparclet },
+
+/* The manual isn't completely accurate on these insns. The `rs2' field is
+ treated as being 6 bits to account for 6 bit immediates to cpush. It is
+ assumed that it is intended that bit 5 is 0 when rs2 contains a reg. */
+#define BIT5 (1<<5)
+{ "crdcxt", F3(2, 0x36, 0)|SLCPOP(4), F3(~2, ~0x36, ~0)|SLCPOP(~4)|BIT5|RS2(~0), "U,d", 0, sparclet },
+{ "cwrcxt", F3(2, 0x36, 0)|SLCPOP(3), F3(~2, ~0x36, ~0)|SLCPOP(~3)|BIT5|RS2(~0), "1,u", 0, sparclet },
+{ "cpush", F3(2, 0x36, 0)|SLCPOP(0), F3(~2, ~0x36, ~0)|SLCPOP(~0)|BIT5|RD(~0), "1,2", 0, sparclet },
+{ "cpush", F3(2, 0x36, 1)|SLCPOP(0), F3(~2, ~0x36, ~1)|SLCPOP(~0)|RD(~0), "1,Y", 0, sparclet },
+{ "cpusha", F3(2, 0x36, 0)|SLCPOP(1), F3(~2, ~0x36, ~0)|SLCPOP(~1)|BIT5|RD(~0), "1,2", 0, sparclet },
+{ "cpusha", F3(2, 0x36, 1)|SLCPOP(1), F3(~2, ~0x36, ~1)|SLCPOP(~1)|RD(~0), "1,Y", 0, sparclet },
+{ "cpull", F3(2, 0x36, 0)|SLCPOP(2), F3(~2, ~0x36, ~0)|SLCPOP(~2)|BIT5|RS1(~0)|RS2(~0), "d", 0, sparclet },
+#undef BIT5
+
+/* sparclet coprocessor branch insns */
+#define SLCBCC2(opcode, mask, lose) \
+ { opcode, (mask), ANNUL|(lose), "l", F_DELAYED|F_CONDBR, sparclet }, \
+ { opcode, (mask)|ANNUL, (lose), ",a l", F_DELAYED|F_CONDBR, sparclet }
+#define SLCBCC(opcode, mask) \
+ SLCBCC2(opcode, F2(0, 7)|COND(mask), F2(~0, ~7)|COND(~(mask)))
+
+/* cbn,cba can't be defined here because they're defined elsewhere and GAS
+ requires all mnemonics of the same name to be consecutive. */
+/*SLCBCC("cbn", 0), - already defined */
+SLCBCC("cbe", 1),
+SLCBCC("cbf", 2),
+SLCBCC("cbef", 3),
+SLCBCC("cbr", 4),
+SLCBCC("cber", 5),
+SLCBCC("cbfr", 6),
+SLCBCC("cbefr", 7),
+/*SLCBCC("cba", 8), - already defined */
+SLCBCC("cbne", 9),
+SLCBCC("cbnf", 10),
+SLCBCC("cbnef", 11),
+SLCBCC("cbnr", 12),
+SLCBCC("cbner", 13),
+SLCBCC("cbnfr", 14),
+SLCBCC("cbnefr", 15),
+
+#undef SLCBCC2
+#undef SLCBCC
+
+{ "casa", F3(3, 0x3c, 0), F3(~3, ~0x3c, ~0), "[1]A,2,d", 0, v9 },
+{ "casa", F3(3, 0x3c, 1), F3(~3, ~0x3c, ~1), "[1]o,2,d", 0, v9 },
+{ "casxa", F3(3, 0x3e, 0), F3(~3, ~0x3e, ~0), "[1]A,2,d", 0, v9 },
+{ "casxa", F3(3, 0x3e, 1), F3(~3, ~0x3e, ~1), "[1]o,2,d", 0, v9 },
+
+/* v9 synthetic insns */
+{ "iprefetch", F2(0, 1)|(2<<20)|BPRED, F2(~0, ~1)|(1<<20)|ANNUL|COND(~0), "G", 0, v9 }, /* bn,a,pt %xcc,label */
+{ "signx", F3(2, 0x27, 0), F3(~2, ~0x27, ~0)|(1<<12)|ASI(~0)|RS2_G0, "1,d", F_ALIAS, v9 }, /* sra rs1,%g0,rd */
+{ "signx", F3(2, 0x27, 0), F3(~2, ~0x27, ~0)|(1<<12)|ASI(~0)|RS2_G0, "r", F_ALIAS, v9 }, /* sra rd,%g0,rd */
+{ "clruw", F3(2, 0x26, 0), F3(~2, ~0x26, ~0)|(1<<12)|ASI(~0)|RS2_G0, "1,d", F_ALIAS, v9 }, /* srl rs1,%g0,rd */
+{ "clruw", F3(2, 0x26, 0), F3(~2, ~0x26, ~0)|(1<<12)|ASI(~0)|RS2_G0, "r", F_ALIAS, v9 }, /* srl rd,%g0,rd */
+{ "cas", F3(3, 0x3c, 0)|ASI(0x80), F3(~3, ~0x3c, ~0)|ASI(~0x80), "[1],2,d", F_ALIAS, v9 }, /* casa [rs1]ASI_P,rs2,rd */
+{ "casl", F3(3, 0x3c, 0)|ASI(0x88), F3(~3, ~0x3c, ~0)|ASI(~0x88), "[1],2,d", F_ALIAS, v9 }, /* casa [rs1]ASI_P_L,rs2,rd */
+{ "casx", F3(3, 0x3e, 0)|ASI(0x80), F3(~3, ~0x3e, ~0)|ASI(~0x80), "[1],2,d", F_ALIAS, v9 }, /* casxa [rs1]ASI_P,rs2,rd */
+{ "casxl", F3(3, 0x3e, 0)|ASI(0x88), F3(~3, ~0x3e, ~0)|ASI(~0x88), "[1],2,d", F_ALIAS, v9 }, /* casxa [rs1]ASI_P_L,rs2,rd */
+
+/* Ultrasparc extensions */
+{ "shutdown", F3F(2, 0x36, 0x080), F3F(~2, ~0x36, ~0x080)|RD_G0|RS1_G0|RS2_G0, "", 0, v9a },
+
+/* FIXME: Do we want to mark these as F_FLOAT, or something similar? */
+{ "fpadd16", F3F(2, 0x36, 0x050), F3F(~2, ~0x36, ~0x050), "v,B,H", 0, v9a },
+{ "fpadd16s", F3F(2, 0x36, 0x051), F3F(~2, ~0x36, ~0x051), "e,f,g", 0, v9a },
+{ "fpadd32", F3F(2, 0x36, 0x052), F3F(~2, ~0x36, ~0x052), "v,B,H", 0, v9a },
+{ "fpadd32s", F3F(2, 0x36, 0x053), F3F(~2, ~0x36, ~0x053), "e,f,g", 0, v9a },
+{ "fpsub16", F3F(2, 0x36, 0x054), F3F(~2, ~0x36, ~0x054), "v,B,H", 0, v9a },
+{ "fpsub16s", F3F(2, 0x36, 0x055), F3F(~2, ~0x36, ~0x055), "e,f,g", 0, v9a },
+{ "fpsub32", F3F(2, 0x36, 0x056), F3F(~2, ~0x36, ~0x056), "v,B,H", 0, v9a },
+{ "fpsub32s", F3F(2, 0x36, 0x057), F3F(~2, ~0x36, ~0x057), "e,f,g", 0, v9a },
+
+{ "fpack32", F3F(2, 0x36, 0x03a), F3F(~2, ~0x36, ~0x03a), "v,B,H", 0, v9a },
+{ "fpack16", F3F(2, 0x36, 0x03b), F3F(~2, ~0x36, ~0x03b)|RS1_G0, "B,g", 0, v9a },
+{ "fpackfix", F3F(2, 0x36, 0x03d), F3F(~2, ~0x36, ~0x03d)|RS1_G0, "B,g", 0, v9a },
+{ "fexpand", F3F(2, 0x36, 0x04d), F3F(~2, ~0x36, ~0x04d)|RS1_G0, "f,H", 0, v9a },
+{ "fpmerge", F3F(2, 0x36, 0x04b), F3F(~2, ~0x36, ~0x04b), "e,f,H", 0, v9a },
+
+/* Note that the mixing of 32/64 bit regs is intentional. */
+{ "fmul8x16", F3F(2, 0x36, 0x031), F3F(~2, ~0x36, ~0x031), "e,B,H", 0, v9a },
+{ "fmul8x16au", F3F(2, 0x36, 0x033), F3F(~2, ~0x36, ~0x033), "e,f,H", 0, v9a },
+{ "fmul8x16al", F3F(2, 0x36, 0x035), F3F(~2, ~0x36, ~0x035), "e,f,H", 0, v9a },
+{ "fmul8sux16", F3F(2, 0x36, 0x036), F3F(~2, ~0x36, ~0x036), "v,B,H", 0, v9a },
+{ "fmul8ulx16", F3F(2, 0x36, 0x037), F3F(~2, ~0x36, ~0x037), "v,B,H", 0, v9a },
+{ "fmuld8sux16", F3F(2, 0x36, 0x038), F3F(~2, ~0x36, ~0x038), "e,f,H", 0, v9a },
+{ "fmuld8ulx16", F3F(2, 0x36, 0x039), F3F(~2, ~0x36, ~0x039), "e,f,H", 0, v9a },
+
+{ "alignaddr", F3F(2, 0x36, 0x018), F3F(~2, ~0x36, ~0x018), "1,2,d", 0, v9a },
+{ "alignaddrl", F3F(2, 0x36, 0x01a), F3F(~2, ~0x36, ~0x01a), "1,2,d", 0, v9a },
+{ "faligndata", F3F(2, 0x36, 0x048), F3F(~2, ~0x36, ~0x048), "v,B,H", 0, v9a },
+
+{ "fzero", F3F(2, 0x36, 0x060), F3F(~2, ~0x36, ~0x060), "H", 0, v9a },
+{ "fzeros", F3F(2, 0x36, 0x061), F3F(~2, ~0x36, ~0x061), "g", 0, v9a },
+{ "fone", F3F(2, 0x36, 0x07e), F3F(~2, ~0x36, ~0x07e), "H", 0, v9a },
+{ "fones", F3F(2, 0x36, 0x07f), F3F(~2, ~0x36, ~0x07f), "g", 0, v9a },
+{ "fsrc1", F3F(2, 0x36, 0x074), F3F(~2, ~0x36, ~0x074), "v,H", 0, v9a },
+{ "fsrc1s", F3F(2, 0x36, 0x075), F3F(~2, ~0x36, ~0x075), "e,g", 0, v9a },
+{ "fsrc2", F3F(2, 0x36, 0x078), F3F(~2, ~0x36, ~0x078), "B,H", 0, v9a },
+{ "fsrc2s", F3F(2, 0x36, 0x079), F3F(~2, ~0x36, ~0x079), "f,g", 0, v9a },
+{ "fnot1", F3F(2, 0x36, 0x06a), F3F(~2, ~0x36, ~0x06a), "v,H", 0, v9a },
+{ "fnot1s", F3F(2, 0x36, 0x06b), F3F(~2, ~0x36, ~0x06b), "e,g", 0, v9a },
+{ "fnot2", F3F(2, 0x36, 0x066), F3F(~2, ~0x36, ~0x066), "B,H", 0, v9a },
+{ "fnot2s", F3F(2, 0x36, 0x067), F3F(~2, ~0x36, ~0x067), "f,g", 0, v9a },
+{ "for", F3F(2, 0x36, 0x07c), F3F(~2, ~0x36, ~0x07c), "v,B,H", 0, v9a },
+{ "fors", F3F(2, 0x36, 0x07d), F3F(~2, ~0x36, ~0x07d), "e,f,g", 0, v9a },
+{ "fnor", F3F(2, 0x36, 0x062), F3F(~2, ~0x36, ~0x062), "v,B,H", 0, v9a },
+{ "fnors", F3F(2, 0x36, 0x063), F3F(~2, ~0x36, ~0x063), "e,f,g", 0, v9a },
+{ "fand", F3F(2, 0x36, 0x070), F3F(~2, ~0x36, ~0x070), "v,B,H", 0, v9a },
+{ "fands", F3F(2, 0x36, 0x071), F3F(~2, ~0x36, ~0x071), "e,f,g", 0, v9a },
+{ "fnand", F3F(2, 0x36, 0x06e), F3F(~2, ~0x36, ~0x06e), "v,B,H", 0, v9a },
+{ "fnands", F3F(2, 0x36, 0x06f), F3F(~2, ~0x36, ~0x06f), "e,f,g", 0, v9a },
+{ "fxor", F3F(2, 0x36, 0x06c), F3F(~2, ~0x36, ~0x06c), "v,B,H", 0, v9a },
+{ "fxors", F3F(2, 0x36, 0x06d), F3F(~2, ~0x36, ~0x06d), "e,f,g", 0, v9a },
+{ "fxnor", F3F(2, 0x36, 0x072), F3F(~2, ~0x36, ~0x072), "v,B,H", 0, v9a },
+{ "fxnors", F3F(2, 0x36, 0x073), F3F(~2, ~0x36, ~0x073), "e,f,g", 0, v9a },
+{ "fornot1", F3F(2, 0x36, 0x07a), F3F(~2, ~0x36, ~0x07a), "v,B,H", 0, v9a },
+{ "fornot1s", F3F(2, 0x36, 0x07b), F3F(~2, ~0x36, ~0x07b), "e,f,g", 0, v9a },
+{ "fornot2", F3F(2, 0x36, 0x076), F3F(~2, ~0x36, ~0x076), "v,B,H", 0, v9a },
+{ "fornot2s", F3F(2, 0x36, 0x077), F3F(~2, ~0x36, ~0x077), "e,f,g", 0, v9a },
+{ "fandnot1", F3F(2, 0x36, 0x068), F3F(~2, ~0x36, ~0x068), "v,B,H", 0, v9a },
+{ "fandnot1s", F3F(2, 0x36, 0x069), F3F(~2, ~0x36, ~0x069), "e,f,g", 0, v9a },
+{ "fandnot2", F3F(2, 0x36, 0x064), F3F(~2, ~0x36, ~0x064), "v,B,H", 0, v9a },
+{ "fandnot2s", F3F(2, 0x36, 0x065), F3F(~2, ~0x36, ~0x065), "e,f,g", 0, v9a },
+
+{ "fcmpgt16", F3F(2, 0x36, 0x028), F3F(~2, ~0x36, ~0x028), "v,B,d", 0, v9a },
+{ "fcmpgt32", F3F(2, 0x36, 0x02c), F3F(~2, ~0x36, ~0x02c), "v,B,d", 0, v9a },
+{ "fcmple16", F3F(2, 0x36, 0x020), F3F(~2, ~0x36, ~0x020), "v,B,d", 0, v9a },
+{ "fcmple32", F3F(2, 0x36, 0x024), F3F(~2, ~0x36, ~0x024), "v,B,d", 0, v9a },
+{ "fcmpne16", F3F(2, 0x36, 0x022), F3F(~2, ~0x36, ~0x022), "v,B,d", 0, v9a },
+{ "fcmpne32", F3F(2, 0x36, 0x026), F3F(~2, ~0x36, ~0x026), "v,B,d", 0, v9a },
+{ "fcmpeq16", F3F(2, 0x36, 0x02a), F3F(~2, ~0x36, ~0x02a), "v,B,d", 0, v9a },
+{ "fcmpeq32", F3F(2, 0x36, 0x02e), F3F(~2, ~0x36, ~0x02e), "v,B,d", 0, v9a },
+
+{ "edge8", F3F(2, 0x36, 0x000), F3F(~2, ~0x36, ~0x000), "1,2,d", 0, v9a },
+{ "edge8l", F3F(2, 0x36, 0x002), F3F(~2, ~0x36, ~0x002), "1,2,d", 0, v9a },
+{ "edge16", F3F(2, 0x36, 0x004), F3F(~2, ~0x36, ~0x004), "1,2,d", 0, v9a },
+{ "edge16l", F3F(2, 0x36, 0x006), F3F(~2, ~0x36, ~0x006), "1,2,d", 0, v9a },
+{ "edge32", F3F(2, 0x36, 0x008), F3F(~2, ~0x36, ~0x008), "1,2,d", 0, v9a },
+{ "edge32l", F3F(2, 0x36, 0x00a), F3F(~2, ~0x36, ~0x00a), "1,2,d", 0, v9a },
+
+{ "pdist", F3F(2, 0x36, 0x03e), F3F(~2, ~0x36, ~0x03e), "v,B,H", 0, v9a },
+
+{ "array8", F3F(2, 0x36, 0x010), F3F(~2, ~0x36, ~0x010), "1,2,d", 0, v9a },
+{ "array16", F3F(2, 0x36, 0x012), F3F(~2, ~0x36, ~0x012), "1,2,d", 0, v9a },
+{ "array32", F3F(2, 0x36, 0x014), F3F(~2, ~0x36, ~0x014), "1,2,d", 0, v9a },
+
+/* Cheetah instructions */
+{ "edge8n", F3F(2, 0x36, 0x001), F3F(~2, ~0x36, ~0x001), "1,2,d", 0, v9b },
+{ "edge8ln", F3F(2, 0x36, 0x003), F3F(~2, ~0x36, ~0x003), "1,2,d", 0, v9b },
+{ "edge16n", F3F(2, 0x36, 0x005), F3F(~2, ~0x36, ~0x005), "1,2,d", 0, v9b },
+{ "edge16ln", F3F(2, 0x36, 0x007), F3F(~2, ~0x36, ~0x007), "1,2,d", 0, v9b },
+{ "edge32n", F3F(2, 0x36, 0x009), F3F(~2, ~0x36, ~0x009), "1,2,d", 0, v9b },
+{ "edge32ln", F3F(2, 0x36, 0x00b), F3F(~2, ~0x36, ~0x00b), "1,2,d", 0, v9b },
+
+{ "bmask", F3F(2, 0x36, 0x019), F3F(~2, ~0x36, ~0x019), "1,2,d", 0, v9b },
+{ "bshuffle", F3F(2, 0x36, 0x04c), F3F(~2, ~0x36, ~0x04c), "v,B,H", 0, v9b },
+
+{ "siam", F3F(2, 0x36, 0x081), F3F(~2, ~0x36, ~0x081)|RD_G0|RS1_G0|RS2(~7), "3", 0, v9b },
+
+/* More v9 specific insns, these need to come last so they do not clash
+ with v9a instructions such as "edge8" which looks like impdep1. */