###############################################################################
# Copyright (c) 2010 Linaro Limited
# All rights reserved. This program and the accompanying materials
# are made available under the terms of the Eclipse Public License v1.0
# which accompanies this distribution, and is available at
# http://www.eclipse.org/legal/epl-v10.html
#
# Contributors:
#     Peter Maydell (Linaro) - initial implementation
###############################################################################

# Input file for risugen defining ARM instructions
.mode arm

# Some random patterns
#ADD A1 cond:4 0000 100 s rn:4 rd:4 imm:5 type:2 0 rm:4
#RBIT A1 cond:4 0110 1111 1111 rd:4 1111 0011 rm:4

# various 32x32->64 multiplies
# we omit the v5-and-below constraint that rn must not be rdhi or rdlo
UMAAL A1 cond:4 0000 0100 rdhi:4 rdlo:4 rm:4 1001 rn:4 { $rdhi != $rdlo; }
UMLAL A1 cond:4 0000 101 s rdhi:4 rdlo:4 rm:4 1001 rn:4 { $rdhi != $rdlo; }
UMULL A1 cond:4 0000 100 s rdhi:4 rdlo:4 rm:4 1001 rn:4 { $rdhi != $rdlo; }
SMLAL A1 cond:4 0000 111 s rdhi:4 rdlo:4 rm:4 1001 rn:4 { $rdhi != $rdlo; }
SMULL A1 cond:4 0000 110 s rdhi:4 rdlo:4 rm:4 1001 rn:4 { $rdhi != $rdlo; }

# 32x32->64 but result is high word only
SMMLA A1 cond:4 01110101 rd:4 ra:4 rm:4 00 r 1 rn:4
SMMLS A1 cond:4 01110101 rd:4 ra:4 rm:4 11 r 1 rn:4
# Note that this doesn't overlap with SMMLA because of the implicit
# constraints on registers fields (ie not 13 or 15)
SMMUL A1 cond:4 01110101 rd:4 1111 rm:4 00 r 1 rn:4

# dual multiplies
SMLAD A1 cond:4 0111 0000 rd:4 ra:4 rm:4 00 m 1 rn:4
SMUAD A1 cond:4 0111 0000 rd:4 1111 rm:4 00 m 1 rn:4
SMLSD A1 cond:4 0111 0000 rd:4 ra:4 rm:4 01 m 1 rn:4
SMUSD A1 cond:4 0111 0000 rd:4 1111 rm:4 01 m 1 rn:4
SMLALD A1 cond:4 0111 0100 rdhi:4 rdlo:4 rm:4 00 m 1 rn:4 { $rdhi != $rdlo; }
SMLSLD A1 cond:4 0111 0100 rdhi:4 rdlo:4 rm:4 01 m 1 rn:4 { $rdhi != $rdlo; }

# divide (A15 and above only!)
SDIV A1 cond:4 01110 001 rd:4 1111 rm:4 000 1 rn:4
UDIV A1 cond:4 01110 011 rd:4 1111 rm:4 000 1 rn:4


USAT A1   cond:4 0110111 satimm:5 rd:4 imm:5 sh 0 1 rn:4
SSAT A1   cond:4 0110101 satimm:5 rd:4 imm:5 sh 0 1 rn:4
SSAT16 A1 cond:4 01101010 satimm:4 rd:4 1111 0011 rn:4
USAT16 A1 cond:4 01101110 satimm:4 rd:4 1111 0011 rn:4

# various preload and hint instructions
# see table A5-24 for this unallocated hint insn block (must NOP on v7MP)
UNALLOC_HINT A1 11110 100 x 001 anything:20
UNALLOC_HINT_b A1 11110 110 x 001 anything:15 0 any:4
PLI_imm A1 1111 0100 u 101 rn:4 1111 imm:12
PLI_reg A1 1111 0110 u 101 rn:4 1111 imm:5 type:2 0 rm:4
PLD_imm A1 1111 0101 u 101 rn:4 1111 imm:12
PLD_reg A1 1111 0111 u 101 rn:4 1111 imm:5 type:2 0 rm:4
PLDW_imm A1 1111 0101 u 001 rn:4 1111 imm:12
PLDW_reg A1 1111 0111 u 001 rn:4 1111 imm:5 type:2 0 rm:4
# no overlap with PLD_imm because rn can't be 15
PLD_lit A1 1111 0101 u 101 1111 1111 imm:12

# Unsigned saturating add/subtract
# UQADD16, UQSUB16, UQADD8, UQSUB8
UQADD16 A1 cond:4 01100110 rn:4 rd:4 1111 0001 rm:4
UQADD8 A1 cond:4 01100110 rn:4 rd:4 1111 1001 rm:4
UQSUB16 A1 cond:4 01100110 rn:4 rd:4 1111 0111 rm:4
UQSUB8 A1 cond:4 01100110 rn:4 rd:4 1111 1111 rm:4
# Signed ditto
QADD16 A1 cond:4 01100010 rn:4 rd:4 1111 0001 rm:4
QADD8 A1 cond:4 01100010 rn:4 rd:4 1111 1001 rm:4
QSUB16 A1 cond:4 01100010 rn:4 rd:4 1111 0111 rm:4
QSUB8 A1 cond:4 01100010 rn:4 rd:4 1111 1111 rm:4

# Signed parallel add/subtract
SADD8 A1 cond:4 0110 0001 rn:4 rd:4 1111 1001 rm:4
SADD16 A1 cond:4 0110 0001 rn:4 rd:4 1111 0001 rm:4
SSUB8 A1 cond:4 0110 0001 rn:4 rd:4 1111 1111 rm:4
SSUB16 A1 cond:4 0110 0001 rn:4 rd:4 1111 0111 rm:4
# unsigned ditto
UADD8 A1 cond:4 0110 0101 rn:4 rd:4 1111 1001 rm:4
UADD16 A1 cond:4 0110 0101 rn:4 rd:4 1111 0001 rm:4
USUB8 A1 cond:4 0110 0101 rn:4 rd:4 1111 1111 rm:4
USUB16 A1 cond:4 0110 0101 rn:4 rd:4 1111 0111 rm:4

# UNDEF cases: op1 == 0 or op2 == 101 or 110
SADD_UNDEF A1a cond:4 011000 00 any:12 op2:3 1 any2:4
SADD_UNDEF A1b cond:4 011000 op1:2 any:12 101 1 any2:4
SADD_UNDEF A1c cond:4 011000 op1:2 any:12 110 1 any2:4

SASX A1 cond:4 0110 0001 rn:4 rd:4 1111 0011 rm:4
SSAX A1 cond:4 0110 0001 rn:4 rd:4 1111 0101 rm:4

# SUBS PC, LR: these are actually unpredictable...
#SUBS_PC_LR A1 cond:4 001 0010 1 rn:4 1111 imm:12
#MOVS_PC_LR A1 cond:4 001 1101 1 rn:4 1111 imm:12

# MLS - v6T2 and later only
MLS A1 cond:4 00000110 rd:4 ra:4 rm:4 1001 rn:4

REVSH A1 cond:4 01101 111 1111 rd:4 1111 1011 rm:4

# ADC/SBC various forms
ADC_imm A1 cond:4 0010101 s:1 rn:4 rd:4 imm:12
ADC_reg A1 cond:4 0000101 s:1 rn:4 rd:4 imm:5 type:2 0 rm:4
ADC_rsr A1 cond:4 0000101 s:1 rn:4 rd:4 rs:4 0 type:2 1 rm:4

SBC_imm A1 cond:4 0010110 s:1 rn:4 rd:4 imm:12
SBC_reg A1 cond:4 0000110 s:1 rn:4 rd:4 imm:5 type:2 0 rm:4
SBC_rsr A1 cond:4 0000110 s:1 rn:4 rd:4 rs:4 0 type:2 1 rm:4

########### Neon loads and stores #########################
# These patterns cover all the Neon element/structure
# load store insns, ie the whole of the space in section
# A7.7 of the ARM ARM DDI0406B, including the UNDEF space.
# This is all of VLD1,VLD2,VLD3,VLD4,VST1,VST2,VST3,VST4.
###########################################################

# VLD*, single element to all lanes

# All addressing modes (reg, reg postindex reg, reg postindex eltsz)
# Note use of 'xm' for 'rm' to avoid the implicit "not 13 or 15"
# constraint -- 13 and 15 encode the other two addr modes so are OK here.
# The constraints are avoiding: d+regs > 32 (UNPREDICTABLE);
# also the UNDEF sz/a combinations; and the risugen restriction
# that the two regs in reg postindex reg must be different.
# Max alignment requirement for VLD1 is 4 bytes.
VLD1_stoa A1a 1111 0100 1 d 10 rn:4 vd:4 11 00 sz:2 t a xm:4 \
    !constraints { ($d == 0 || $t == 0 || $vd != 0xf) && $sz != 3 && ($sz != 0 || $a != 1) && ($rn != $xm); } \
    !memory { reg($rn); }

# As usual we need to separate out the UNDEF cases as they
# must not have !memory blocks
# sz 11: UNDEF
VLD1_stoa A1b 1111 0100 1 d 10 rn:4 vd:4 11 00 11 t a rm:4
# sz 00, a 1 : UNDEF
VLD1_stoa A1c 1111 0100 1 d 10 rn:4 vd:4 11 00 00 t 1 rm:4

# VLD2: d+t+1 > 31 is unpredictable
VLD2_stoa A1a 1111 0100 1 d 10 rn:4 vd:4 11 01 sz:2 t a xm:4 \
    !constraints { (((($d << 4)|$vd) + $t + 1) < 32) && $sz != 3 && ($rn != $xm); } \
    !memory { align(8); reg($rn); }

# UNDEF case : sz 11
VLD2_stoa A1b 1111 0100 1 d 10 rn:4 vd:4 11 01 11 t a xm:4

# VLD3: d+(t+1)*2 > 31 is unpredictable
VLD3_stoa A1a 1111 0100 1 d 10 rn:4 vd:4 11 01 sz:2 t 0 xm:4 \
    !constraints { (((($d << 4)|$vd) + ($t + 1)*2) < 32) && $sz != 3 && ($rn != $xm); } \
    !memory { reg($rn); }

# UNDEF case : sz 11 or a 1
VLD3_stoa A1b 1111 0100 1 d 10 rn:4 vd:4 11 10 11 t a xm:4
VLD3_stoa A1c 1111 0100 1 d 10 rn:4 vd:4 11 10 sz:2 t 1 xm:4

# VLD4: d+(t+1)*3 > 31 is unpredictable
VLD4_stoa A1a 1111 0100 1 d 10 rn:4 vd:4 11 11 sz:2 t a xm:4 \
    !constraints { (((($d << 4)|$vd) + ($t + 1)*3) < 32) && ($sz != 3 || $a != 0) && ($rn != $xm); } \
    !memory { align(16); reg($rn); }

# UNDEF case : sz 11 and a 0
VLD4_stoa A1b 1111 0100 1 d 10 rn:4 vd:4 11 11 11 t 0 xm:4

# VLD*, single element to one lane. We split the sz cases out
# for convenience of filtering the UNDEF cases and the VLD1-to-all-lanes
# sz == 00
VLD1_s A1a 1111 0100 1 d 10 rn:4 vd:4 00 00 idx:3 0 xm:4 \
   !constraints { ($rn != $xm); } \
   !memory { reg($rn); }

# sz == 01
VLD1_s A1b 1111 0100 1 d 10 rn:4 vd:4 01 00 idx:2 0 idx0 xm:4 \
   !constraints { ($rn != $xm); } \
   !memory { reg($rn); }

# sz == 10
VLD1_s A1c 1111 0100 1 d 10 rn:4 vd:4 10 00 idx3 0 idx:2 xm:4 \
   !constraints { ($rn != $xm) && ($idx == 0 || $idx == 3); } \
   !memory { reg($rn); }

# UNDEF cases: bad index fields for each size
VLD1_s A1d 1111 0100 1 d 10 rn:4 vd:4 00 00 idx:3 1 xm:4
VLD1_s A1e 1111 0100 1 d 10 rn:4 vd:4 01 00 idx:2 1 idx0 xm:4
VLD1_s A1f 1111 0100 1 d 10 rn:4 vd:4 10 00 idx:4 xm:4 \
   !constraints { ($idx & 4) == 1 || ($idx & 3) == 1 || ($idx & 3) == 2; }

# VLD2 has an UNPREDICTABLE case for d+inc > 31.
# sz == 00, 01 (no UNDEF cases)
VLD2_s A1a 1111 0100 1 d 10 rn:4 vd:4 0 sz 01 idx:4 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 1 + (($idx >> $sz) & 1)) < 32; } \
   !memory { reg($rn); }

# sz == 10
VLD2_s A1b 1111 0100 1 d 10 rn:4 vd:4 10 01 idx:2 0 idx0 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 1 + ($idx & 1)) < 32; } \
   !memory { align(8); reg($rn); }

# only UNDEF case is sz=10, idx<1>=1
VLD2_s A1c 1111 0100 1 d 10 rn:4 vd:4 10 01 idx:2 1 idx0 xm:4

# UNPREDICTABLE here is for d+inc+inc > 31
# sz == 00, 01
VLD3_s A1a 1111 0100 1 d 10 rn:4 vd:4 0 sz 10 idx:3 0 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 2 * (1 + ((($idx << 1) >> $sz) & 1))) < 32; } \
   !memory { reg($rn); }

# sz == 10
VLD3_s A1b 1111 0100 1 d 10 rn:4 vd:4 10 10 idx:2 00 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 2 * (1 + ($idx & 1))) < 32; } \
   !memory { reg($rn); }

# UNDEF: sz == 00, 01, idx<0> != 0
VLD3_s A1c 1111 0100 1 d 10 rn:4 vd:4 0 sz 10 idx:3 1 xm:4
# UNDEF: sz == 10, idx<1:0> != 00
VLD3_s A1d 1111 0100 1 d 10 rn:4 vd:4 10 10 idx:4 xm:4 \
   !constraints { ($idx & 3) != 0; }

# VLD4 has an UNPREDICTABLE case for d+3*inc > 31.
# sz == 00, 01 (no UNDEF cases)
VLD4_s A1a 1111 0100 1 d 10 rn:4 vd:4 0 sz 11 idx:4 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 3 * (1 + (($idx & (1 << $sz)) >> 1))) < 32; } \
   !memory { align(8); reg($rn); }

# sz == 10
VLD4_s A1b 1111 0100 1 d 10 rn:4 vd:4 10 01 idx:2 0 idx0 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 3 * (1 + ($idx & 1))) < 32; } \
   !memory { align(16); reg($rn); }

# only UNDEF case is sz=10, idx<1:0>==11
VLD4_s A1c 1111 0100 1 d 10 rn:4 vd:4 10 01 idx:2 11 xm:4


# VST* single element from one lane
# These are actually identical to the VLD* patterns except that
# bit 21 is clear to indicate store rather than load.

# sz == 00
VST1_s A1a 1111 0100 1 d 00 rn:4 vd:4 00 00 idx:3 0 xm:4 \
   !constraints { ($rn != $xm); } \
   !memory { reg($rn); }

# sz == 01
VST1_s A1b 1111 0100 1 d 00 rn:4 vd:4 01 00 idx:2 0 idx0 xm:4 \
   !constraints { ($rn != $xm); } \
   !memory { reg($rn); }

# sz == 10
VST1_s A1c 1111 0100 1 d 00 rn:4 vd:4 10 00 idx3 0 idx:2 xm:4 \
   !constraints { ($rn != $xm) && ($idx == 0 || $idx == 3); } \
   !memory { reg($rn); }

# UNDEF cases: bad index fields for each size
VST1_s A1d 1111 0100 1 d 00 rn:4 vd:4 00 00 idx:3 1 xm:4
VST1_s A1e 1111 0100 1 d 00 rn:4 vd:4 01 00 idx:2 1 idx0 xm:4
VST1_s A1f 1111 0100 1 d 00 rn:4 vd:4 10 00 idx:4 xm:4 \
   !constraints { ($idx & 4) == 1 || ($idx & 3) == 1 || ($idx & 3) == 2; }

# sz == 00, 01 (no UNDEF cases)
VST2_s A1a 1111 0100 1 d 00 rn:4 vd:4 0 sz 01 idx:4 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 1 + (($idx >> $sz) & 1)) < 32; } \
   !memory { reg($rn); }

# sz == 10
VST2_s A1b 1111 0100 1 d 00 rn:4 vd:4 10 01 idx:2 0 idx0 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 1 + ($idx & 1)) < 32; } \
   !memory { align(8); reg($rn); }

# only UNDEF case is sz=10, idx<1>=1
VST2_s A1c 1111 0100 1 d 00 rn:4 vd:4 10 01 idx:2 1 idx0 xm:4

# UNPREDICTABLE here is for d+inc+inc > 31
# sz == 00, 01
VST3_s A1a 1111 0100 1 d 00 rn:4 vd:4 0 sz 10 idx:3 0 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 2 * (1 + ((($idx << 1) >> $sz) & 1))) < 32; } \
   !memory { reg($rn); }

# sz == 10
VST3_s A1b 1111 0100 1 d 00 rn:4 vd:4 10 10 idx:2 00 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 2 * (1 + ($idx & 1))) < 32; } \
   !memory { reg($rn); }

# UNDEF: sz == 00, 01, idx<0> != 0
VST3_s A1c 1111 0100 1 d 00 rn:4 vd:4 0 sz 10 idx:3 1 xm:4
# UNDEF: sz == 10, idx<1:0> != 00
VST3_s A1d 1111 0100 1 d 00 rn:4 vd:4 10 10 idx:4 xm:4 \
   !constraints { ($idx & 3) != 0; }

# VST4 has an UNPREDICTABLE case for d+3*inc > 31.
# sz == 00, 01 (no UNDEF cases)
VST4_s A1a 1111 0100 1 d 00 rn:4 vd:4 0 sz 11 idx:4 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 3 * (1 + (($idx & (1 << $sz)) >> 1))) < 32; } \
   !memory { align(8); reg($rn); }

# sz == 10
VST4_s A1b 1111 0100 1 d 00 rn:4 vd:4 10 01 idx:2 0 idx0 xm:4 \
   !constraints { ($rn != $xm) && ((($d << 4)|$vd) + 3 * (1 + ($idx & 1))) < 32; } \
   !memory { align(16); reg($rn); }

# only UNDEF case is sz=10, idx<1:0>==11
VST4_s A1c 1111 0100 1 d 00 rn:4 vd:4 10 01 idx:2 11 xm:4


# VLD*, multiple single elements

# We separate these out by 'type' field
# type 0111: VLD1 regs=1
VLD1_m A1a 1111 0100 0 d 10 rn:4 vd:4 0111 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 1) <= 32); } \
   !memory { align(8); reg($rn); }
# type 1010: VLD1 regs=2
VLD1_m A1b 1111 0100 0 d 10 rn:4 vd:4 1010 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2) <= 32) && ($align != 3); } \
   !memory { align(16); reg($rn); }
# type 0110: VLD1 regs=3
VLD1_m A1c 1111 0100 0 d 10 rn:4 vd:4 0110 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 3) <= 32); } \
   !memory { align(8); reg($rn); }
# type 0010: VLD1 regs=4
VLD1_m A1d 1111 0100 0 d 10 rn:4 vd:4 0010 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 4) <= 32); } \
   !memory { align(32); reg($rn); }

# type 1000: VLD2 regs=1 inc=1
VLD2_m A1a 1111 0100 0 d 10 rn:4 vd:4 1000 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 1 + 1) <= 32) && ($align != 3) && ($sz != 3); } \
   !memory { align(16); reg($rn); }
# type 1001: VLD2 regs=1 inc=2
VLD2_m A1b 1111 0100 0 d 10 rn:4 vd:4 1001 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2 + 1) <= 32) && ($align != 3) && ($sz != 3); } \
   !memory { align(16); reg($rn); }
# type 0011: VLD2 regs=2 inc=2
VLD2_m A1c 1111 0100 0 d 10 rn:4 vd:4 0011 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2 + 2) <= 32) && ($sz != 3); } \
   !memory { align(32); reg($rn); }

# type 0100: VLD3 inc=1
VLD3_m A1a 1111 0100 0 d 10 rn:4 vd:4 0100 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2) <= 31) && ($sz != 3); } \
   !memory { align(8); reg($rn); }
# type 0101: VLD3 inc=2
VLD3_m A1b 1111 0100 0 d 10 rn:4 vd:4 0101 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 4) <= 31) && ($sz != 3); } \
   !memory { align(8); reg($rn); }

# type 0000: VLD4 inc=1
VLD4_m A1a 1111 0100 0 d 10 rn:4 vd:4 0000 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 3) <= 31) && ($sz != 3); } \
   !memory { align(32); reg($rn); }
# type 0001: VLD4 inc=2
VLD4_m A1b 1111 0100 0 d 10 rn:4 vd:4 0000 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 6) <= 31) && ($sz != 3); } \
   !memory { align(32); reg($rn); }

# UNDEF cases for all VLD*_m:

# These are the patterns for VLD*_m UNDEFs on align bits being wrong
# Conveniently the conditions for all the VLD* line up with the top
# bits of the type field.
# type = 01xx align = 1x
VLDn_m A1a 1111 0100 0 d 10 rn:4 vd:4 01 type:2 sz:2 1 align xm:4
# type = 10xx align = 11
VLDn_m A1b 1111 0100 0 d 10 rn:4 vd:4 10 type:2 sz:2 11 xm:4

# VLD2/3/4 UNDEF on SZ=1; slightly tedious to pick these out
# but they are type = xx0x or type = 0011
VLDn_m A1c 1111 0100 0 d 10 rn:4 vd:4 type:2 0 type0 11 align:2 xm:4
VLDn_m A1d 1111 0100 0 d 10 rn:4 vd:4 0011 11 align:2 xm:4

# VST*, multiple elements

# We separate these out by 'type' field
# type 0111: VST1 regs=1
VST1_m A1a 1111 0100 0 d 00 rn:4 vd:4 0111 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 1) <= 32); } \
   !memory { align(8); reg($rn); }
# type 1010: VST1 regs=2
VST1_m A1b 1111 0100 0 d 00 rn:4 vd:4 1010 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2) <= 32) && ($align != 3); } \
   !memory { align(16); reg($rn); }
# type 0110: VST1 regs=3
VST1_m A1c 1111 0100 0 d 00 rn:4 vd:4 0110 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 3) <= 32); } \
   !memory { align(8); reg($rn); }
# type 0010: VST1 regs=4
VST1_m A1d 1111 0100 0 d 00 rn:4 vd:4 0010 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 4) <= 32); } \
   !memory { align(32); reg($rn); }

# type 1000: VST2 regs=1 inc=1
VST2_m A1a 1111 0100 0 d 00 rn:4 vd:4 1000 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 1 + 1) <= 32) && ($align != 3) && ($sz != 3); } \
   !memory { align(16); reg($rn); }
# type 1001: VST2 regs=1 inc=2
VST2_m A1b 1111 0100 0 d 00 rn:4 vd:4 1001 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2 + 1) <= 32) && ($align != 3) && ($sz != 3); } \
   !memory { align(16); reg($rn); }
# type 0011: VST2 regs=2 inc=2
VST2_m A1c 1111 0100 0 d 00 rn:4 vd:4 0011 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2 + 2) <= 32) && ($sz != 3); } \
   !memory { align(32); reg($rn); }

# type 0100: VST3 inc=1
VST3_m A1a 1111 0100 0 d 00 rn:4 vd:4 0100 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 2) <= 31) && ($sz != 3); } \
   !memory { align(8); reg($rn); }
# type 0101: VST3 inc=2
VST3_m A1b 1111 0100 0 d 00 rn:4 vd:4 0101 sz:2 0 align xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 4) <= 31) && ($sz != 3); } \
   !memory { align(8); reg($rn); }

# type 0000: VST4 inc=1
VST4_m A1a 1111 0100 0 d 00 rn:4 vd:4 0000 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 3) <= 31) && ($sz != 3); } \
   !memory { align(32); reg($rn); }
# type 0001: VST4 inc=2
VST4_m A1b 1111 0100 0 d 00 rn:4 vd:4 0000 sz:2 align:2 xm:4 \
   !constraints { ($rn != $xm) && (((($d << 4)|$vd) + 6) <= 31) && ($sz != 3); } \
   !memory { align(32); reg($rn); }

# UNDEF cases for all VST*_m:

# These are the patterns for VST*_m UNDEFs on align bits being wrong
# Conveniently the conditions for all the VST* line up with the top
# bits of the type field.
# type = 01xx align = 1x
VSTn_m A1a 1111 0100 0 d 10 rn:4 vd:4 01 type:2 sz:2 1 align xm:4
# type = 10xx align = 11
VSTn_m A1b 1111 0100 0 d 10 rn:4 vd:4 10 type:2 sz:2 11 xm:4

# VST2/3/4 UNDEF on SZ=1; slightly tedious to pick these out
# but they are type = xx0x or type = 0011
VSTn_m A1c 1111 0100 0 d 10 rn:4 vd:4 type:2 0 type0 11 align:2 xm:4
VSTn_m A1d 1111 0100 0 d 10 rn:4 vd:4 0011 11 align:2 xm:4

# Unused spaces in the load/store instruction encoding space
# (everything else is some sort of VLDn/VSTn)
# A = 0 cases:
# type = 11xx : always UNDEF
VLDST_UNDEF A1a 1111 0100 0 any:2 0 any2:8 11 type:2 any3:8
# type = 1011 : always UNDEF
VLDST_UNDEF A1b 1111 0100 0 any:2 0 any2:8 1011 any3:8
# A = 1 cases: only stores with B=11xx, all else is allocated
VLDST_UNDEF A1c 1111 0100 1 x 0 0 any:8 11 any2:10

########### Neon Data Processing ##########################
# The following sets of patterns cover the whole of the
# "Advanced SIMD data-processing instructions" space
# as described in DDI0406B table A7-8 and the subtables
# it refers to.
###########################################################

########### Neon 3 reg same length ########################
# Instructions from the Neon "3 register same length"
# space (table A7-9 in DDI0406B)
# We include UNDEF combinations here; there are no
# UNPREDICTABLE encodings we need to avoid.
# We avoid size encodings that are fp16 under the v8.2-FP16 extension:
# sz=1 for the float insns which have a 1-bit size field in bit 20,
# and instead hard-wire that bit to 0.
###########################################################

VHADD A1 1111 001 u 0 d sz:2 vn:4 vd:4 0000 n q m 0 vm:4
VHSUB A1 1111 001 u 0 d sz:2 vn:4 vd:4 0010 n q m 0 vm:4
VQADD A1 1111 001 u 0 d sz:2 vn:4 vd:4 0000 n q m 1 vm:4
VRHADD A1 1111 001 u 0 d sz:2 vn:4 vd:4 0001 n q m 0 vm:4
VAND A1 1111 0010 0 d 00 vn:4 vd:4 0001 n q m 1 vm:4
VBIC A1 1111 0010 0 d 01 vn:4 vd:4 0001 n q m 1 vm:4
# includes what the ARM ARM calls VMOV T1/A1, which is just VORR with vn=vm
VORR A1 1111 0010 0 d 10 vn:4 vd:4 0001 n q m 1 vm:4
VORN A1 1111 0010 0 d 11 vn:4 vd:4 0001 n q m 1 vm:4
# VEOR, VBIF, VBIT and VBSL:
VEOR_VBIT A1 1111 0011 0 d op:2 vn:4 vd:4 0001 n q m 1 vm:4
VQSUB A1 1111 001 u 0 d sz:2 vn:4 vd:4 0010 n q m 1 vm:4
VCGT A1 1111 001 u 0 d sz:2 vn:4 vd:4 0011 n q m 0 vm:4
VCGT A2 1111 0011 0 d 1 0 vn:4 vd:4 1110 n q m 0 vm:4
VCGE A1 1111 001 u 0 d sz:2 vn:4 vd:4 0011 n q m 1 vm:4
VCGE A2 1111 0011 0 d 0 0 vn:4 vd:4 1110 n q m 0 vm:4
VSHL A1 1111 001 u 0 d sz:2 vn:4 vd:4 0100 n q m 0 vm:4
VQSHL A1 1111 001 u 0 d sz:2 vn:4 vd:4 0100 n q m 1 vm:4
VRSHL A1 1111 001 u 0 d sz:2 vn:4 vd:4 0101 n q m 0 vm:4
VQRSHL A1 1111 001 u 0 d sz:2 vn:4 vd:4 0101 n q m 1 vm:4
VMAX A1 1111 001 u 0 d sz:2 vn:4 vd:4 0110 n q m 0 vm:4
VMIN A1 1111 001 u 0 d sz:2 vn:4 vd:4 0110 n q m 1 vm:4
VABD A1 1111 001 u 0 d sz:2 vn:4 vd:4 0111 n q m 0 vm:4
VABA A1 1111 001 u 0 d sz:2 vn:4 vd:4 0111 n q m 1 vm:4
VADD A1 1111 0010 0 d sz:2 vn:4 vd:4 1000 n q m 0 vm:4
VSUB A1 1111 0011 0 d sz:2 vn:4 vd:4 1000 n q m 0 vm:4
VTST A1 1111 0010 0 d sz:2 vn:4 vd:4 1000 n q m 1 vm:4
VCEQ A1 1111 0011 0 d sz:2 vn:4 vd:4 1000 n q m 1 vm:4
VCEQ A2 1111 0010 0 d 0 0 vn:4 vd:4 1110 n q m 0 vm:4
VMLA A1 1111 001 op 0 d sz:2 vn:4 vd:4 1001 n q m 0 vm:4
VMUL A1 1111 001 op 0 d sz:2 vn:4 vd:4 1001 n q m 1 vm:4
VPMAX A1 1111 001 u 0 d sz:2 vn:4 vd:4 1010 n q m 0 vm:4
VPMIN A1 1111 001 u 0 d sz:2 vn:4 vd:4 1010 n q m 1 vm:4
VQDMULH A1 1111 0010 0 d sz:2 vn:4 vd:4 1011 n q m 0 vm:4
VQRDMULH A1 1111 0011 0 d sz:2 vn:4 vd:4 1011 n q m 0 vm:4
VPADD A1 1111 0010 0 d sz:2 vn:4 vd:4 1011 n q m 1 vm:4
# NB: VFM is VFPv4 only. There is no Neon encoding for VFNM.
VFM A1 1111 0010 0 d op 0 vn:4 vd:4 1100 n q m 1 vm:4
VADD_float A1 1111 0010 0 d 0 0 vn:4 vd:4 1101 n q m 0 vm:4
VSUB_float A1 1111 0010 0 d 1 0 vn:4 vd:4 1101 n q m 0 vm:4
VPADD_float A1 1111 0011 0 d 0 0 vn:4 vd:4 1101 n q m 0 vm:4
VABD_float A1 1111 0011 0 d 1 0 vn:4 vd:4 1101 n q m 0 vm:4
VMLA_float A1 1111 0010 0 d 0 0 vn:4 vd:4 1101 n q m 1 vm:4
VMLS_float A1 1111 0010 0 d 1 0 vn:4 vd:4 1101 n q m 1 vm:4
VMUL_float A1 1111 0011 0 d 0 0 vn:4 vd:4 1101 n q m 1 vm:4
VACGE A1 1111 0011 0 d 0 0 vn:4 vd:4 1110 n q m 1 vm:4
VACGT A1 1111 0011 0 d 1 0 vn:4 vd:4 1110 n q m 1 vm:4
VMAX_float A1 1111 0010 0 d 0 0 vn:4 vd:4 1111 n q m 0 vm:4
VMIN_float A1 1111 0010 0 d 1 0 vn:4 vd:4 1111 n q m 0 vm:4
VPMAX_float A1 1111 0011 0 d 0 0 vn:4 vd:4 1111 n q m 0 vm:4
VPMIN_float A1 1111 0011 0 d 1 0 vn:4 vd:4 1111 n q m 0 vm:4
VRECPS A1 1111 0010 0 d 0 0 vn:4 vd:4 1111 n q m 1 vm:4
VRSQRTS A1 1111 0010 0 d 1 0 vn:4 vd:4 1111 n q m 1 vm:4

########### Neon 1 reg + modified immediate ###############
# Instructions from the Neon "1 reg + modified immediate"
# space (table A7-14 in DDI0406B)
# We include UNDEF combinations here. There are some
# UNPREDICTABLEs in the constant encoding:
# abcdefgh == 0 and cmode == 001x 010x 011x 101x 110x
###########################################################
# We don't try to break these down into the separate
# VMOV, VORR, VMVN, VBIC ops (which are encoded in 'op'
# and 'cmode'); we just have a single pattern with a constraint
# which avoids the UNPREDICTABLE space.
Vimm A1 1111 001 imm1 1 d 000 imm3:3 vd:4 cmode:4 0 q op 1 imm4:4 \
   !constraints { $imm1 != 0 || $imm3 != 0 || $imm4 != 0 || ($cmode & 0xe) == 0 || ($cmode & 0xe) == 8 || ($cmode & 0xe == 0xe); }

########### Neon 2 regs + shift ###########################
# Instructions from the Neon "2 regs + shift" space
# (table A7-12 in DDI0406B)
# UNDEF cases included (but not generally the ones that
# fall in gaps in the table).
# NB L:imm == 0000xxx is in one-reg+modified-imm space.
###########################################################
VSHR A1 1111 001 u 1 d imm:6 vd:4 0000 l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
VSRA A1 1111 001 u 1 d imm:6 vd:4 0001 l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
VRSHR A1 1111 001 u 1 d imm:6 vd:4 0010 l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
VRSRA A1 1111 001 u 1 d imm:6 vd:4 0011 l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
VSRI A1 1111 0011 1 d imm:6 vd:4 0100 l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
VSHL_imm A1 1111 0010 1 d imm:6 vd:4 0101 l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
VSLI A1 1111 0011 1 d imm:6 vd:4 0101 l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
VQSHL_imm A1 1111 001 u 1 d imm:6 vd:4 011 op l q m 1 vm:4 { $l != 0 || ($imm & 0x38) != 0; }
# this includes VSHRN (if U=0 and op=0)
VQSHRN A1 1111 001 u 1 d imm:6 vd:4 100 op 0 0 m 1 vm:4 { ($imm & 0x38) != 0; }
# this includes VRSHRN (if U=0 and op=0)
VQRSHRN A1 1111 001 u 1 d imm:6 vd:4 100 op 0 1 m 1 vm:4  { ($imm & 0x38) != 0; }
# includes VMOVL where the shift amount is zero
VSHLL A1 1111 001 u 1 d imm:6 vd:4 1010 0 0 m 1 vm:4 { ($imm & 0x38) != 0; }
VCVT A1 1111 001 u 1 d imm:6 vd:4 111 op 0 q m 1 vm:4 { ($imm & 0x38) != 0; }

########### Neon 3 regs different lengths #################
# Instructions from the Neon "3 regs different lengths"
# space (table A7-10 in DDI0406B)
# UNDEF cases included.
# sz = 11 is in vext/vtbl/vtbx/vdup/2reg-misc space.
###########################################################
VADDL A1 1111 001 u 1 d sz:2 vn:4 vd:4 000 0 n 0 m 0 vm:4 { $sz != 3; }
VADDW A1 1111 001 u 1 d sz:2 vn:4 vd:4 000 1 n 0 m 0 vm:4 { $sz != 3; }
VSUBL A1 1111 001 u 1 d sz:2 vn:4 vd:4 001 0 n 0 m 0 vm:4 { $sz != 3; }
VSUBW A1 1111 001 u 1 d sz:2 vn:4 vd:4 001 1 n 0 m 0 vm:4 { $sz != 3; }
VADDHN A1 1111 0010 1 d sz:2 vn:4 vd:4 0100 n 0 m 0 vm:4 { $sz != 3; }
VRADDHN A1 1111 0011 1 d sz:2 vn:4 vd:4 0100 n 0 m 0 vm:4 { $sz != 3; }
VABAL A2 1111 001 u 1 d sz:2 vn:4 vd:4 0101 n 0 m 0 vm:4 { $sz != 3; }
VSUBHN A1 1111 0010 1 d sz:2 vn:4 vd:4 0110 n 0 m 0 vm:4 { $sz != 3; }
VRSUBHN A1 1111 0011 1 d sz:2 vn:4 vd:4 0110 n 0 m 0 vm:4 { $sz != 3; }
VABDL A2 1111 001 u 1 d sz:2 vn:4 vd:4 0111 n 0 m 0 vm:4 { $sz != 3; }
VMLAL A2 1111 001 u 1 d sz:2 vn:4 vd:4 10 0 0 n 0 m 0 vm:4 { $sz != 3; }
VMLSL A2 1111 001 u 1 d sz:2 vn:4 vd:4 10 1 0 n 0 m 0 vm:4 { $sz != 3; }
VQDMLAL A1 1111 0010 1 d sz:2 vn:4 vd:4 10 0 1 n 0 m 0 vm:4 { $sz != 3; }
VQDMLSL A1 1111 0010 1 d sz:2 vn:4 vd:4 10 1 1 n 0 m 0 vm:4 { $sz != 3; }
VMULL A2 1111 001 u 1 d sz:2 vn:4 vd:4 11 op 0 n 0 m 0 vm:4 { $sz != 3; }
VQDMULL A1 1111 0010 1 d sz:2 vn:4 vd:4 1101 n 0 m 0 vm:4 { $sz != 3; }

########### Neon 2 regs + scalar ##########################
# Instructions from the Neon "2 regs + scalar" space
# (table A7-11 in DDI0406B)
# UNDEF cases included.
# sz = 11 is in vext/vtbl/vtbx/vdup/2reg-misc space.
# We avoid f=1 sz=01 which is v8.2-FP16
###########################################################
# includes float variants
VMLA_scalar A1 1111 001 q 1  d sz:2 vn:4 vd:4 0 0 0 f n 1 m 0 vm:4 \
  { $sz != 3 && ($f == 0 || $sz != 1); }
VMLS_scalar A1 1111 001 q 1  d sz:2 vn:4 vd:4 0 1 0 f n 1 m 0 vm:4 \
  { $sz != 3 && ($f == 0 || $sz != 1); }
VMLAL_scalar A2 1111 001 u 1 d sz:2 vn:4 vd:4 0 0 1 0 n 1 m 0 vm:4 { $sz != 3; }
VMLSL_scalar A2 1111 001 u 1 d sz:2 vn:4 vd:4 0 1 1 0 n 1 m 0 vm:4 { $sz != 3; }
VQDMLAL_scalar A2 1111 0010 1 d sz:2 vn:4 vd:4 0 0 11 n 1 m 0 vm:4 { $sz != 3; }
VQDMLSL_scalar A2 1111 0010 1 d sz:2 vn:4 vd:4 0 1 11 n 1 m 0 vm:4 { $sz != 3; }
VMUL_scalar A1 1111 001 q 1 d sz:2 vn:4 vd:4 100 f n 1 m 0 vm:4 \
  { $sz != 3 && ($f == 0 || $sz != 1); }
VMULL_scalar A2 1111 001 u 1 d sz:2 vn:4 vd:4 1010 n 1 m 0 vm:4 { $sz != 3; }
VQDMULL_scalar A2 1111 0010 1 d sz:2 vn:4 vd:4 1011 n 1 m 0 vm:4 { $sz != 3; }
VQDMULH_scalar A2 1111 001 q 1 d sz:2 vn:4 vd:4 1100 n 1 m 0 vm:4 { $sz != 3; }
VQRDMULH_scalar A2 1111 001 q 1 d sz:2 vn:4 vd:4 1101 n 1 m 0 vm:4 { $sz != 3; }

########### Neon 2 regs miscellaneous #####################
# Instructions from the Neon "2 regs miscellaneous" space
# (table A7-13 in DDI0406B)
# UNDEF cases included.
# We avoid f=1 sz=01 which is v8.2-FP16
###########################################################
VREV A1 1111 0011 1 d 11 sz:2 00 vd:4 000 op:2 q m 0 vm:4
VPADDL A1 1111 0011 1 d 11 sz:2 00 vd:4 0010 op q m 0 vm:4
VCLS A1 1111 0011 1 d 11 sz:2 00 vd:4 0 1000 q m 0 vm:4
VCLZ A1 1111 0011 1 d 11 sz:2 00 vd:4 0 1001 q m 0 vm:4
VCNT A1 1111 0011 1 d 11 sz:2 00 vd:4 0 1010 q m 0 vm:4
VMVN A1 1111 0011 1 d 11 sz:2 00 vd:4 0 1011 q m 0 vm:4
VPADAL A1 1111 0011 1 d 11 sz:2 00 vd:4 0110 op q m 0 vm:4
VQABS A1 1111 0011 1 d 11 sz:2 00 vd:4 0111 0 q m 0 vm:4
VQNEG A1 1111 0011 1 d 11 sz:2 00 vd:4 0111 1 q m 0 vm:4
VCGT0 A1 1111 0011 1 d 11 sz:2 01 vd:4 0 f 000 q m 0 vm:4 { $f == 0 || $sz != 1; }
VCGE0 A1 1111 0011 1 d 11 sz:2 01 vd:4 0 f 001 q m 0 vm:4 { $f == 0 || $sz != 1; }
VCEQ0 A1 1111 0011 1 d 11 sz:2 01 vd:4 0 f 010 q m 0 vm:4 { $f == 0 || $sz != 1; }
VCLE0 A1 1111 0011 1 d 11 sz:2 01 vd:4 0 f 011 q m 0 vm:4 { $f == 0 || $sz != 1; }
VCLT0 A1 1111 0011 1 d 11 sz:2 01 vd:4 0 f 100 q m 0 vm:4 { $f == 0 || $sz != 1; }
VABS A1 1111 0011 1 d 11 sz:2 01 vd:4 0 f 110 q m 0 vm:4 { $f == 0 || $sz != 1; }
VNEG A1 1111 0011 1 d 11 sz:2 01 vd:4 0 f 111 q m 0 vm:4 { $f == 0 || $sz != 1; }
VSWP A1 1111 0011 1 d 11 sz:2 10 vd:4 00000 q m 0 vm:4
# d == m gives UNKNOWN results, so avoid it
VTRN A1 1111 0011 1 d 11 sz:2 10 vd:4 00001 q m 0 vm:4 { ($d != $m) || ($vd != $vm); }
# d == m gives UNKNOWN results, so avoid it
VUZP A1 1111 0011 1 d 11 sz:2 10 vd:4 00010 q m 0 vm:4 { ($d != $m) || ($vd != $vm); }
VZIP A1 1111 0011 1 d 11 sz:2 10 vd:4 00011 q m 0 vm:4 { ($d != $m) || ($vd != $vm); }
# includes VMOVN, VQMOVUN
VQMOVN A1 1111 0011 1 d 11 sz:2 10 vd:4 0010 op:2 m 0 vm:4
VSHLL A2 1111 0011 1 d 11 sz:2 10 vd:4 0011 0 0 m 0 vm:4
# float-halfprec (A8.6.299)
# NB that half-precision needs at least an A9; A8 doesn't have it
VCVT_half A1 1111 0011 1 d 11 sz:2 10 vd:4 011 op 0 0 m 0 vm:4
VRECPE A1 1111 0011 1 d 11 sz:2 11 vd:4 010 f 0 q m 0 vm:4 { $f == 0 || $sz != 1; }
VRSQRTE A1 1111 0011 1 d 11 sz:2 11 vd:4 010 f 1 q m 0 vm:4 { $f == 0 || $sz != 1; }
# float to int, neon versions (A8.6.294); avoid sz=01 which is FP16
VCVT_neon A1 1111 0011 1 d 11 sz:2 11 vd:4 0 11 op:2 q m 0 vm:4 { $sz != 1; }

########### Neon other ####################################
# Instructions which have their own entry in the top level
# Neon data processing instructions decode table A7-8:
# VEXT, VTBL, VTBX, VDUP (scalar)
###########################################################
VEXT A1 1111 0010 1 d 11 vn:4 vd:4 imm:4 n q m 0 vm:4
# VTBL and VTBX; n + length > 32 is UNPREDICTABLE
VTBL A1 1111 0011 1 d 11 vn:4 vd:4 1 0 len:2 n op m 0 vm:4 \
    !constraints { (($n << 4) | $vn) + $len + 1 <= 32; }
VDUP_scalar A1 1111 0011 1 d 11 imm:4 vd:4 11000 q m 0 vm:4

###########################################################
# End of Neon Data Processing instruction patterns.
###########################################################

########### VFP Data Processing ###########################
# The following sets of patterns cover the whole of the
# "VFP data-processing instructions" space
# as described in DDI0406B table A7-16 and the subtables
# it refers to.
# These don't include fp16, which has [11:9] 0b100
# (described in the Arm ARM as [11:9] 0b10 and a 2-bit size field)
###########################################################

# VMLA, VMLS
VMLA A2 cond:4 11100 d 00 vn:4 vd:4 101 sz n op m 0 vm:4
# VNMLA, VNMLS
VNMLA A1 cond:4 11100 d 01 vn:4 vd:4 101 sz n op m 0 vm:4
VNMUL A2 cond:4 11100 d 10 vn:4 vd:4 101 sz n 1 m 0 vm:4
VMUL A2 cond:4 11100 d 10 vn:4 vd:4 101 sz n 0 m 0 vm:4
VADD A2 cond:4 11100 d 11 vn:4 vd:4 101 sz n 0 m 0 vm:4
VSUB A2 cond:4 11100 d 11 vn:4 vd:4 101 sz n 1 m 0 vm:4
VDIV A1 cond:4 11101 d 00 vn:4 vd:4 101 sz n 0 m 0 vm:4

# Other VFP data processing instructions (opc1 1x11)
# We don't check that SBO/SBZ bits here UNDEF if wrong.
VMOV_imm A2 cond:4 11101 d 11 immh:4 vd:4 101 sz 0000 imml:4
VMOV A2 cond:4 11101 d 11 0000 vd:4 101 sz 0 1 m 0 vm:4
VABS A2 cond:4 11101 d 11 0000 vd:4 101 sz 1 1 m 0 vm:4
VNEG A2 cond:4 11101 d 11 0001 vd:4 101 sz 0 1 m 0 vm:4
VSQRT A1 cond:4 11101 d 11 0001 vd:4 101 sz 1 1 m 0 vm:4
# VCVTB, VCVTT (A8.6.300) [requires half-precision extension]
VCVT_B_TT A1 cond:4 1110 1 d 11 001 op vd:4 101 0 t 1 m 0 vm:4
VCMP A1 cond:4 11101 d 11 0100 vd:4 101 sz e 1 m 0 vm:4
VCMP A2 cond:4 11101 d 11 0101 vd:4 101 sz e 1 0 0 0000

# VCVT between double and single (A8.6.298)
VCVT_298 A1 cond:4 1110 1 d 11 0111 vd:4 101 sz 1 1 m 0 vm:4
# VCVT between fp and int: split in two because opc2 must be 000 or 10x (A8.6.295)
VCVT_a A1 cond:4 11101 d 111 000 vd:4 101 sz op 1 m 0 vm:4
VCVT_b A1 cond:4 11101 d 111 10 x vd:4 101 sz op 1 m 0 vm:4

# VCVT between fp and fixed point (A.8.6.297)
# Ugh. UNPREDICTABLE unless the 32 bit int formed by imm4:i is at least
# 16 (if sx is 0) or 32 (if sx is 1). That is, if sx==0 then either
# bit 3 must be 0 or bits 2..0 and 5 must be 0.
# sx==1 case first:
VCVT_c A1 cond:4 11101 d 111 op 1 u vd:4 101 sf 1 1 i 0 imm:4
# sx==0, bit 3 == 0
VCVT_d A1 cond:4 11101 d 111 op 1 u vd:4 101 sf 0 1 i 0 0 imm:3
# sx==0, bit 3 == 1, bits 2..0 and 5 0
VCVT_e A1 cond:4 11101 d 111 op 1 u vd:4 101 sf 0 1 0 0 1000

# VFPv4 fused multiply-add
VFM A2 cond:4 11101 d 10 vn:4 vd:4 101 sz n op m 0 vm:4
VFNM A1 cond:4 11101 d 01 vn:4 vd:4 101 sz n op m 0 vm:4

# UNDEF patterns in VFP data processing space (not currently checked):
# opc1 1x00 opc3 x1
# opc1 1x11 opc2 0110 opc3 x1
# opc1 1x11 opc2 0111 opc3 01
# opc1 1x11 opc2 1001 opc3 x1

########### Extension register load/store #################
# The following sets of patterns cover:
#  'extension register load/store insns' (A7.6)
# as described in DDI0406B
###########################################################

# Note that the ARM ARM treats VPUSH/VPOP as special cases
# but in fact they are the same as VLDM/VSTM
# VSTM of 64 bit regs
# NB that the constraints on imm are to avoid UNPREDICTABLEs.
# We force the high bits of imm to 0 in the pattern to avoid
# pointlessly generating things that fail the constraint anyway.
# postincrement (U == 1)
VSTM A1a cond:4 110 p 1 d w 0 rn:4 vd:4 1011 00 imm:5 x \
 !constraints { $p != 1 && $imm != 0 && $imm <= 16 && ((($d << 4) | $vd) + $imm) <= 32; } \
 !memory { reg($rn); }
# predecrement (U == 0)
VSTM A1b cond:4 110 p 0 d w 0 rn:4 vd:4 1011 00 imm:5 x \
 !constraints { $p == 1 && $w == 1 && $imm != 0 && $imm <= 16 && ((($d << 4) | $vd) + $imm) <= 32; } \
 !memory { reg_minus_imm($rn, $imm * 8); }
# VSTM of 32 bit regs
# postincrement (U == 1)
VSTM A2a cond:4 110 p 1 d w 0 rn:4 vd:4 1010 00 imm:6  \
 !constraints { $p != 1 && $imm != 0 &&  ((($vd << 1) | $d) + $imm) <= 32; } \
 !memory { reg($rn); }
# predecrement (U == 0)
VSTM A2b cond:4 110 p 0 d w 0 rn:4 vd:4 1010 00 imm:6  \
 !constraints { $p == 1 && $w == 1 && $imm != 0 &&  ((($vd << 1) | $d) + $imm) <= 32; } \
 !memory { reg_minus_imm($rn, $imm * 8); }
# UNDEF cases for both A1 and A2:  P==U && W==1
VSTM A1c cond:4 110 p u d 1 0 rn:4 vd:4 101 x imm:8 !constraints { $p == $u; }

# VLDM
VLDM A1a cond:4 110 p 1 d w 1 rn:4 vd:4 1011 00 imm:5 x \
 !constraints { $p != 1 && $imm != 0 && $imm <= 16 && ((($d << 4) | $vd) + $imm) <= 32; } \
 !memory { reg($rn); }
# predecrement (U == 0)
VLDM A1b cond:4 110 p 0 d w 1 rn:4 vd:4 1011 00 imm:5 x \
 !constraints { $p == 1 && $w == 1 && $imm != 0 && $imm <= 16 && ((($d << 4) | $vd) + $imm) <= 32; } \
 !memory { reg_minus_imm($rn, $imm * 8); }
# VLDM of 32 bit regs
# postincrement (U == 1)
VLDM A2a cond:4 110 p 1 d w 1 rn:4 vd:4 1010 00 imm:6  \
 !constraints { $p != 1 && $imm != 0 &&  ((($vd << 1) | $d) + $imm) <= 32; } \
 !memory { reg($rn); }
# predecrement (U == 0)
VLDM A2b cond:4 110 p 0 d w 1 rn:4 vd:4 1010 00 imm:6  \
 !constraints { $p == 1 && $w == 1 && $imm != 0 &&  ((($vd << 1) | $d) + $imm) <= 32; } \
 !memory { reg_minus_imm($rn, $imm * 8); }
# UNDEF cases for both A1 and A2:  P==U && W==1
VLDM A1c cond:4 110 p u d 1 1 rn:4 vd:4 101 x imm:8 !constraints { $p == $u; }

# VSTR (no overlap with VSTR_f16)
# both A1 and A2 encodings, U = 1
VSTR A1a cond:4 1101 1 d 00 rn:4 vd:4 101 x imm:8 \
 !memory { reg_plus_imm($rn, $imm * 4); }
# both A1 and A2 encodings, U = 0
VSTR A1b cond:4 1101 0 d 00 rn:4 vd:4 101 x imm:8 \
 !memory { reg_minus_imm($rn, $imm * 4); }

# VLDR (no overlap with VLDR_f16)
# both A1 and A2 encodings, U = 1
VLDR A1a cond:4 1101 1 d 01 rn:4 vd:4 101 x imm:8 \
 !memory { reg_plus_imm($rn, $imm * 4); }
# both A1 and A2 encodings, U = 0
VLDR A1b cond:4 1101 0 d 01 rn:4 vd:4 101 x imm:8 \
 !memory { reg_minus_imm($rn, $imm * 4); }

########### Extension register transfer ###################
# The following sets of patterns cover:
#  '8, 16 and 32-bit transfer between ARM core and
#  extension registers' (A7.8)
# as described in DDI0406C
# with the exception of VMSR/VMRS.
###########################################################

VMOV_core_single A1 cond:4 1110 000 op:1 vd:4 rt:4 1010 n:1 0010000
VMOV_core_scalar A1 cond:4 1110 0 opc:2 0 vd:4 rt:4 1011 d:1 opc2:2 10000
VMOV_scalar_core A1 cond:4 1110 u:1 opc:2 1 vn:4 rt:4 1011 n:1 opc2:2 10000

# vector duplicate (reg)
# b:e == 11 UNDEF
VDUP A1a cond:4 1110 1 b 1 0 vd:3 0 rt:4 1011 d 0 e 1 0000 { ($b == 0) || ($e == 0); }
VDUP A1b cond:4 1110 1 b 0 0 vd:4 rt:4 1011 d 0 e 1 0000 { ($b == 0) || ($e == 0); }

########### Extension register transfer ###################
# The following sets of patterns cover:
#  '64-bit transfers between ARM core and extension
#  registers' (A7.8)
# as described in DDI0406C
###########################################################
VMOV_core_2single A1 cond:4 1100 010 op:1 rt2:4 rt:4 1010 00 m:1 1 vm:4 { ($vm != 0xf || $m != 1) && ($op == 0 || $rt2 != $rt); }
VMOV_core_double A1 cond:4 1100 010 op:1 rt2:4 rt:4 1011 00 m:1 1 vm:4 { $op == 0 || $rt2 != $rt; }

#####
# v8 only insns
# VSEL
VSEL A1 1111 11100 d cc:2 vn:4 vd:4 101 sz n 0 m 0 vm:4
# VMINNM and VMAXNM
# neon: sz=0 (avoiding sz=1 which is FP16)
VMINMAXNM A1 1111 00110 d op 0 vn:4 vd:4 1111 n q m 1 vm:4
# vfp (does not overlap with FP16)
VMINMAXNM A2 1111 11101 d 00 vn:4 vd:4 101 sz n op m 0 vm:4

# Crypto
# AESD, AESE, AESIMC, AESMC
AESD A1 1111 0011 1 d 11 sz:2 00 vd:4 0011 01 m 0 vm:4
AESE A1 1111 0011 1 d 11 sz:2 00 vd:4 0011 00 m 0 vm:4
AESIMC A1 1111 0011 1 d 11 sz:2 00 vd:4 0011 11 m 0 vm:4
AESMC A1 1111 0011 1 d 11 sz:2 00 vd:4 0011 10 m 0 vm:4

# SHA1, SHA256
SHA1C    A1 1111 00100 d 00 vn:4 vd:4 1100 n q m 0 vm:4
SHA1M    A1 1111 00100 d 10 vn:4 vd:4 1100 n q m 0 vm:4
SHA1P    A1 1111 00100 d 01 vn:4 vd:4 1100 n q m 0 vm:4
SHA1SU0  A1 1111 00100 d 11 vn:4 vd:4 1100 n q m 0 vm:4

SHA256H   A1 1111 00110 d 00 vn:4 vd:4 1100 n q m 0 vm:4
SHA256H2  A1 1111 00110 d 01 vn:4 vd:4 1100 n q m 0 vm:4
SHA256SU1 A1 1111 00110 d 10 vn:4 vd:4 1100 n q m 0 vm:4

SHA1H     A1 1111 00111 d 11 sz:2 01 vd:4 0 01011 m 0 vm:4
SHA1SU1   A1 1111 00111 d 11 sz:2 10 vd:4 0 01110 m 0 vm:4
SHA256SU0 A1 1111 00111 d 11 sz:2 10 vd:4 0 01111 m 0 vm:4

# long polynomial multiply: op == 1, sz == 10
VMULL_P64 A2 1111 001 u 1 d 10 vn:4 vd:4 11 1 0 n 0 m 0 vm:4

# rounding, VFP: VRINTX, VRINTZ, VRINTR, VRINT[ANPM]
VRINTX A1 cond:4 11101 d 110111 vd:4 101 sz 0 1 m 0 vm:4
VRINTZR A1 cond:4 11101 d 110110 vd:4 101 sz op 1 m 0 vm:4
VRINTANPM A1 1111 11101 d 1110 rmode:2 vd:4 101 sz 0 1 m 0 vm:4
# rounding, Neon:
VRINTX_neon A1 1111 00111 d 11 sz:2 10 vd:4 01001 q m 0 vm:4
VRINTZ_neon A1 1111 00111 d 11 sz:2 10 vd:4 01011 q m 0 vm:4
VRINTANPM_neon A1 1111 00111 d 11 sz:2 10 vd:4 01 op:3 q m 0 vm:4

# VCVT with rounding mode specified
VCVT_rm A1 1111 11101 d 1111 rm:2 vd:4 101 sz op 1 m 0 vm:4
# 64<->16 conversions (see also pattern earlier which is the sz==0 case)
VCVT_B_TT_64 A1 cond:4 1110 1 d 11 001 op vd:4 101 1 t 1 m 0 vm:4

# VCVT with rounding mode specified, neon; avoid sz=0b01 which is FP16
VCVT_rm_neon A1 1111 00111 d 11 size:2 11 vd:4 00 rm:2 op q m 0 vm:4 { $size != 1; }

# CRC
# Note that sz == 0b11 is UNPREDICTABLE (either UNDEF, NOP or as if == 0b10)
# as is cond != 1110 (either UNDEF, NOP, cond-exec or unconditional exec)
CRC32 A1 1110 00010 sz:2 0 rn:4 rd:4 00 c 0 0100 rm:4 !constraints { $sz != 3; }

#
# ARMv8.1 extensions
#
@v8_1_simd

VQRDMLAH    A1 111100110 d:1 size:2 vn:4 vd:4 1011 n:1 q:1 m:1 1 vm:4
VQRDMLAH_s  A1 1111001 q:1 1 d:1 size:2 vn:4 vd:4 1110 n:1 1 m:1 0 vm:4

VQRDMLSH    A1 111100110 d:1 size:2 vn:4 vd:4 1100 n:1 q:1 m:1 1 vm:4
VQRDMLSH_s  A1 1111001 q:1 1 d:1 size:2 vn:4 vd:4 1111 n:1 1 m:1 0 vm:4

#
# ARMv8.2 extensions
#
@v8_2_dp

VUDOT       A1 1111110 00 d:1 10 vn:4 vd:4 1101 n:1 q:1 m:1 1 vm:4
VSDOT       A1 1111110 00 d:1 10 vn:4 vd:4 1101 n:1 q:1 m:1 0 vm:4

VUDOT_s     A1 11111110 0 d:1 10 vn:4 vd:4 1101 n:1 q:1 m:1 1 vm:4
VSDOT_s     A1 11111110 0 d:1 10 vn:4 vd:4 1101 n:1 q:1 m:1 0 vm:4

@v8_2_fhm

VFMAL       A1 1111110 00 d:1 10 vn:4 vd:4 1000 n:1 q:1 m:1 1 vm:4
VFMSL       A1 1111110 01 d:1 10 vn:4 vd:4 1000 n:1 q:1 m:1 1 vm:4
VFMAL_s     A1 11111110 0 d:1 00 vn:4 vd:4 1000 n:1 q:1 m:1 1 vm:4
VFMSL_s     A1 11111110 1 d:1 00 vn:4 vd:4 1000 n:1 q:1 m:1 1 vm:4

@v8_2_fp16

# v8.2-FP16 adds a lot of "and 16-bit flavour" to existing insn encodings;
# these patterns are arranged in the same order as the earlier v7 patterns.
# Pattern names follow the non-fp16 names with a _f16 suffix.

# FP16: neon 3-reg-same: bit 20 sz=1 for f16
VCGT_f16          A2 1111 0011 0 d 1 1 vn:4 vd:4 1110 n q m 0 vm:4
VCGE_f16          A2 1111 0011 0 d 0 1 vn:4 vd:4 1110 n q m 0 vm:4
VCEQ_f16          A2 1111 0010 0 d 0 1 vn:4 vd:4 1110 n q m 0 vm:4
VFM_f16           A1 1111 0010 0 d op 1 vn:4 vd:4 1100 n q m 1 vm:4
VADD_float_f16    A1 1111 0010 0 d 0 1 vn:4 vd:4 1101 n q m 0 vm:4
VSUB_float_f16    A1 1111 0010 0 d 1 1 vn:4 vd:4 1101 n q m 0 vm:4
VPADD_float_f16   A1 1111 0011 0 d 0 1 vn:4 vd:4 1101 n q m 0 vm:4
VABD_float_f16    A1 1111 0011 0 d 1 1 vn:4 vd:4 1101 n q m 0 vm:4
VMLA_float_f16    A1 1111 0010 0 d 0 1 vn:4 vd:4 1101 n q m 1 vm:4
VMLS_float_f16    A1 1111 0010 0 d 1 1 vn:4 vd:4 1101 n q m 1 vm:4
VMUL_float_f16    A1 1111 0011 0 d 0 1 vn:4 vd:4 1101 n q m 1 vm:4
VACGE_f16         A1 1111 0011 0 d 0 1 vn:4 vd:4 1110 n q m 1 vm:4
VACGT_f16         A1 1111 0011 0 d 1 1 vn:4 vd:4 1110 n q m 1 vm:4
VMAX_float_f16    A1 1111 0010 0 d 0 1 vn:4 vd:4 1111 n q m 0 vm:4
VMIN_float_f16    A1 1111 0010 0 d 1 1 vn:4 vd:4 1111 n q m 0 vm:4
VPMAX_float_f16   A1 1111 0011 0 d 0 1 vn:4 vd:4 1111 n q m 0 vm:4
VPMIN_float_f16   A1 1111 0011 0 d 1 1 vn:4 vd:4 1111 n q m 0 vm:4
VRECPS_f16        A1 1111 0010 0 d 0 1 vn:4 vd:4 1111 n q m 1 vm:4
VRSQRTS_f16       A1 1111 0010 0 d 1 1 vn:4 vd:4 1111 n q m 1 vm:4

# FP16: neon 2-reg-scalar : f=1 sz=01
VMLA_scalar_f16   A1 1111 001 q 1  d 01 vn:4 vd:4 0 0 0 f n 1 m 0 vm:4
VMLS_scalar_f16   A1 1111 001 q 1  d 01 vn:4 vd:4 0 1 0 f n 1 m 0 vm:4
VMUL_scalar_f16   A1 1111 001 q 1 d 01 vn:4 vd:4 100 f n 1 m 0 vm:4

# FP16: Neon 2-reg-shift
# this doesn't overlap with the non-fp16 insn, which has 111 in [11:9]
VCVT_f16 A1 1111 001 u 1 d imm:6 vd:4 110 op 0 q m 1 vm:4 { ($imm & 0x38) != 0; }

# FP16: neon 2-reg-misc: f=1 sz=01
VCGT0_f16        A1 1111 0011 1 d 11 01 01 vd:4 0 f 000 q m 0 vm:4
VCGE0_f16        A1 1111 0011 1 d 11 01 01 vd:4 0 f 001 q m 0 vm:4
VCEQ0_f16        A1 1111 0011 1 d 11 01 01 vd:4 0 f 010 q m 0 vm:4
VCLE0_f16        A1 1111 0011 1 d 11 01 01 vd:4 0 f 011 q m 0 vm:4
VCLT0_f16        A1 1111 0011 1 d 11 01 01 vd:4 0 f 100 q m 0 vm:4
VABS_f16         A1 1111 0011 1 d 11 01 01 vd:4 0 f 110 q m 0 vm:4
VNEG_f16         A1 1111 0011 1 d 11 01 01 vd:4 0 f 111 q m 0 vm:4
VRECPE_f16       A1 1111 0011 1 d 11 01 11 vd:4 010 f 0 q m 0 vm:4
VRSQRTE_f16      A1 1111 0011 1 d 11 01 11 vd:4 010 f 1 q m 0 vm:4
VCVT_neon_f16    A1 1111 0011 1 d 11 01 11 vd:4 0 11 op:2 q m 0 vm:4

# FP16: vfp: these have no overlap with non-fp16 patterns, where [11:9] is 101
VMLA_f16         A2 cond:4 11100 d 00 vn:4 vd:4 1001 n op m 0 vm:4
VNMLA_f16        A1 cond:4 11100 d 01 vn:4 vd:4 1001 n op m 0 vm:4
VNMUL_f16        A2 cond:4 11100 d 10 vn:4 vd:4 1001 n 1 m 0 vm:4
VMUL_f16         A2 cond:4 11100 d 10 vn:4 vd:4 1001 n 0 m 0 vm:4
VADD_f16         A2 cond:4 11100 d 11 vn:4 vd:4 1001 n 0 m 0 vm:4
VSUB_f16         A2 cond:4 11100 d 11 vn:4 vd:4 1001 n 1 m 0 vm:4
VDIV_f16         A1 cond:4 11101 d 00 vn:4 vd:4 1001 n 0 m 0 vm:4
VMOV_imm_f16     A2 cond:4 11101 d 11 immh:4 vd:4 1001 0000 imml:4
VABS_f16         A2 cond:4 11101 d 11 0000 vd:4 1001 1 1 m 0 vm:4
VNEG_f16         A2 cond:4 11101 d 11 0001 vd:4 1001 0 1 m 0 vm:4
VSQRT_f16        A1 cond:4 11101 d 11 0001 vd:4 1001 1 1 m 0 vm:4
VCMP_f16         A1 cond:4 11101 d 11 0100 vd:4 1001 e 1 m 0 vm:4
VCMP_f16         A2 cond:4 11101 d 11 0101 vd:4 1001 e 1 0 0 0000
VCVT_a_f16       A1 cond:4 11101 d 111 000 vd:4 1001 op 1 m 0 vm:4
VCVT_b_f16       A1 cond:4 11101 d 111 10 x vd:4 1001 op 1 m 0 vm:4
# VCVT between fp and fixed point (A.8.6.297); same UNPREDICTABLE as non-fp16
# sx==1 case first:
VCVT_c_f16       A1 cond:4 11101 d 111 op 1 u vd:4 1001 1 1 i 0 imm:4
# sx==0, bit 3 == 0
VCVT_d_f16       A1 cond:4 11101 d 111 op 1 u vd:4 1001 0 1 i 0 0 imm:3
# sx==0, bit 3 == 1, bits 2..0 and 5 0
VCVT_e_f16       A1 cond:4 11101 d 111 op 1 u vd:4 1001 0 1 0 0 1000
VFM_f16          A2 cond:4 11101 d 10 vn:4 vd:4 1001 n op m 0 vm:4
VFNM_f16         A1 cond:4 11101 d 01 vn:4 vd:4 1001 n op m 0 vm:4
# both A1 and A2 encodings, U = 1
VSTR_f16         A1a cond:4 1101 1 d 00 rn:4 vd:4 1001 imm:8 \
                 !memory { reg_plus_imm($rn, $imm * 2); }
# both A1 and A2 encodings, U = 0
VSTR_f16         A1b cond:4 1101 0 d 00 rn:4 vd:4 1001 imm:8 \
                 !memory { reg_minus_imm($rn, $imm * 2); }
# both A1 and A2 encodings, U = 1
VLDR_f16         A1a cond:4 1101 1 d 01 rn:4 vd:4 1001 imm:8 \
                 !memory { reg_plus_imm($rn, $imm * 2); }
# both A1 and A2 encodings, U = 0
VLDR_f16         A1b cond:4 1101 0 d 01 rn:4 vd:4 1001 imm:8 \
                 !memory { reg_minus_imm($rn, $imm * 2); }

# FP16: v8-only insns
# Neon insns with sz=0b01
VCVT_rm_neon_f16   A1 1111 00111 d 11 01 11 vd:4 00 rm:2 op q m 0 vm:4
VRINTX_neon_f16    A1 1111 00111 d 11 01 10 vd:4 01001 q m 0 vm:4
VRINTZ_neon_f16    A1 1111 00111 d 11 01 10 vd:4 01011 q m 0 vm:4
VRINTANPM_neon_f16 A1 1111 00111 d 11 01 10 vd:4 01 op:3 q m 0 vm:4
# Neon insn with sz=1
VMINMAXNM_f16    A1 1111 00110 d op 1 vn:4 vd:4 1111 n q m 1 vm:4
# VFP insns which don't overlap non-fp16 rules (which have 101 in [11:9])
VCVT_rm_f16      A1 1111 11101 d 1111 rm:2 vd:4 1001 op 1 m 0 vm:4
VSEL_f16         A1 1111 11100 d cc:2 vn:4 vd:4 1001 n 0 m 0 vm:4
VMINMAXNM_f16    A2 1111 11101 d 00 vn:4 vd:4 1001 n op m 0 vm:4
VRINTX_f16       A1 cond:4 11101 d 110111 vd:4 1001 0 1 m 0 vm:4
VRINTZR_f16      A1 cond:4 11101 d 110110 vd:4 1001 op 1 m 0 vm:4
VRINTANPM_f16    A1 1111 11101 d 1110 rmode:2 vd:4 1001 0 1 m 0 vm:4

# FP16: Insns which are new for v8.2 FP16:
VINS_f16         A1 1111 11101 d 110000 vd:4 101011 m 0 vm:4
VMOVX_f16        A1 1111 11101 d 110000 vd:4 101001 m 0 vm:4
# VMOV between general-purpose register and half-precision
VMOV_core_f16    A1 cond:4 1110000 op:1 vn:4 rt:4 1001 n 0010000

@v8_2_fp16_v8_3_compnum
# These are only present if both v8.2-FP16 and v8.3-CompNum are implemented
# sz=0 for FP16
VCADD_f16        A1 1111110 rot:1 1 d:1 0 0 vn:4 vd:4 1000 n:1 q:1 m:1 0 vm:4
VCMLA_f16        A1 1111110 rot:2 d:1 1 0 vn:4 vd:4 1000 n:1 q:1 m:1 0 vm:4
VCMLA_s_f16      A1 11111110 0 d:1 rot:2 vn:4 vd:4 1000 n:1 q:1 m:1 0 vm:4

#
# ARMv8.3 extensions
#
@v8_3_compnum

# We avoid the FP16 parts of this, which are in @v8_2_fp16_v8_3_compnum,
# so here s=1
VCADD       A1 1111110 rot:1 1 d:1 0 1 vn:4 vd:4 1000 n:1 q:1 m:1 0 vm:4
VCMLA       A1 1111110 rot:2 d:1 1 1 vn:4 vd:4 1000 n:1 q:1 m:1 0 vm:4
VCMLA_s     A1 11111110 1 d:1 rot:2 vn:4 vd:4 1000 n:1 q:1 m:1 0 vm:4