/* tc-arm.c -- Assemble for the ARM
- Copyright (C) 1994-2019 Free Software Foundation, Inc.
+ Copyright (C) 1994-2020 Free Software Foundation, Inc.
Contributed by Richard Earnshaw (rwe@pegasus.esprit.ec.org)
Modified by David Taylor (dtaylor@armltd.co.uk)
Cirrus coprocessor mods by Aldy Hernandez (aldyh@redhat.com)
#include "obstack.h"
#include "libiberty.h"
#include "opcode/arm.h"
+#include "cpu-arm.h"
#ifdef OBJ_ELF
#include "elf/arm.h"
should define CPU_DEFAULT here. */
#endif
+/* Perform range checks on positive and negative overflows by checking if the
+ VALUE given fits within the range of an BITS sized immediate. */
+static bfd_boolean out_of_range_p (offsetT value, offsetT bits)
+ {
+ gas_assert (bits < (offsetT)(sizeof (value) * 8));
+ return (value & ~((1 << bits)-1))
+ && ((value & ~((1 << bits)-1)) != ~((1 << bits)-1));
+}
+
#ifndef FPU_DEFAULT
# ifdef TE_LINUX
# define FPU_DEFAULT FPU_ARCH_FPA
static int fix_v4bx = FALSE;
/* Warn on using deprecated features. */
static int warn_on_deprecated = TRUE;
+static int warn_on_restrict_it = FALSE;
/* Understand CodeComposer Studio assembly syntax. */
bfd_boolean codecomposer_syntax = FALSE;
ARM_FEATURE_CORE_HIGH (ARM_EXT2_SB);
static const arm_feature_set arm_ext_predres =
ARM_FEATURE_CORE_HIGH (ARM_EXT2_PREDRES);
+static const arm_feature_set arm_ext_bf16 =
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_BF16);
+static const arm_feature_set arm_ext_i8mm =
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_I8MM);
+static const arm_feature_set arm_ext_crc =
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC);
static const arm_feature_set arm_arch_any = ARM_ANY;
-#ifdef OBJ_ELF
static const arm_feature_set fpu_any = FPU_ANY;
-#endif
static const arm_feature_set arm_arch_full ATTRIBUTE_UNUSED = ARM_FEATURE (-1, -1, -1);
static const arm_feature_set arm_arch_t2 = ARM_ARCH_THUMB2;
static const arm_feature_set arm_arch_none = ARM_ARCH_NONE;
static const arm_feature_set fpu_vfp_v3_or_neon_ext =
ARM_FEATURE_COPROC (FPU_NEON_EXT_V1 | FPU_VFP_EXT_V3);
static const arm_feature_set mve_ext =
- ARM_FEATURE_COPROC (FPU_MVE);
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_MVE);
static const arm_feature_set mve_fp_ext =
- ARM_FEATURE_COPROC (FPU_MVE_FP);
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_MVE_FP);
#ifdef OBJ_ELF
static const arm_feature_set fpu_vfp_fp16 =
ARM_FEATURE_COPROC (FPU_VFP_EXT_FP16);
ARM_FEATURE_COPROC (FPU_NEON_EXT_ARMV8);
static const arm_feature_set fpu_crypto_ext_armv8 =
ARM_FEATURE_COPROC (FPU_CRYPTO_EXT_ARMV8);
-static const arm_feature_set crc_ext_armv8 =
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8);
static const arm_feature_set fpu_neon_ext_v8_1 =
ARM_FEATURE_COPROC (FPU_NEON_EXT_RDMA);
static const arm_feature_set fpu_neon_ext_dotprod =
/* Feature bits selected by the last .object_arch directive. */
static arm_feature_set selected_object_arch = ARM_ARCH_NONE;
/* Must be long enough to hold any of the names in arm_cpus. */
+static const struct arm_ext_table * selected_ctx_ext_table = NULL;
static char selected_cpu_name[20];
extern FLONUM_TYPE generic_floating_point_number;
NT_float,
NT_poly,
NT_signed,
+ NT_bfloat,
NT_unsigned
};
_("cannot use writeback with PC-relative addressing")
#define BAD_RANGE _("branch out of range")
#define BAD_FP16 _("selected processor does not support fp16 instruction")
+#define BAD_BF16 _("selected processor does not support bf16 instruction")
#define UNPRED_REG(R) _("using " R " results in unpredictable behaviour")
#define THUMB1_RELOC_ONLY _("relocation valid in thumb1 code only")
#define MVE_NOT_IT _("Warning: instruction is UNPREDICTABLE in an IT " \
/* As in 0f12.456 */
/* or 0d1.2345e12 */
-const char FLT_CHARS[] = "rRsSfFdDxXeEpP";
+const char FLT_CHARS[] = "rRsSfFdDxXeEpPHh";
/* Prefix characters that indicate the start of an immediate
value. */
#define skip_whitespace(str) do { if (*(str) == ' ') ++(str); } while (0)
+enum fp_16bit_format
+{
+ ARM_FP16_FORMAT_IEEE = 0x1,
+ ARM_FP16_FORMAT_ALTERNATIVE = 0x2,
+ ARM_FP16_FORMAT_DEFAULT = 0x3
+};
+
+static enum fp_16bit_format fp16_format = ARM_FP16_FORMAT_DEFAULT;
+
+
static inline int
skip_past_char (char ** str, char c)
{
switch (type)
{
+ case 'H':
+ case 'h':
+ prec = 1;
+ break;
+
+ /* If this is a bfloat16, then parse it slightly differently, as it
+ does not follow the IEEE specification for floating point numbers
+ exactly. */
+ case 'b':
+ {
+ FLONUM_TYPE generic_float;
+
+ t = atof_ieee_detail (input_line_pointer, 1, 8, words, &generic_float);
+
+ if (t)
+ input_line_pointer = t;
+ else
+ return _("invalid floating point number");
+
+ switch (generic_float.sign)
+ {
+ /* Is +Inf. */
+ case 'P':
+ words[0] = 0x7f80;
+ break;
+
+ /* Is -Inf. */
+ case 'N':
+ words[0] = 0xff80;
+ break;
+
+ /* Is NaN. */
+ /* bfloat16 has two types of NaN - quiet and signalling.
+ Quiet NaN has bit[6] == 1 && faction != 0, whereas
+ signalling NaN's have bit[0] == 0 && fraction != 0.
+ Chosen this specific encoding as it is the same form
+ as used by other IEEE 754 encodings in GAS. */
+ case 0:
+ words[0] = 0x7fff;
+ break;
+
+ default:
+ break;
+ }
+
+ *sizeP = 2;
+
+ md_number_to_chars (litP, (valueT) words[0], sizeof (LITTLENUM_TYPE));
+
+ return NULL;
+ }
case 'f':
case 'F':
case 's':
input_line_pointer = t;
*sizeP = prec * sizeof (LITTLENUM_TYPE);
- if (target_big_endian)
- {
- for (i = 0; i < prec; i++)
- {
- md_number_to_chars (litP, (valueT) words[i], sizeof (LITTLENUM_TYPE));
- litP += sizeof (LITTLENUM_TYPE);
- }
- }
+ if (target_big_endian || prec == 1)
+ for (i = 0; i < prec; i++)
+ {
+ md_number_to_chars (litP, (valueT) words[i], sizeof (LITTLENUM_TYPE));
+ litP += sizeof (LITTLENUM_TYPE);
+ }
+ else if (ARM_CPU_HAS_FEATURE (cpu_variant, fpu_endian_pure))
+ for (i = prec - 1; i >= 0; i--)
+ {
+ md_number_to_chars (litP, (valueT) words[i], sizeof (LITTLENUM_TYPE));
+ litP += sizeof (LITTLENUM_TYPE);
+ }
else
- {
- if (ARM_CPU_HAS_FEATURE (cpu_variant, fpu_endian_pure))
- for (i = prec - 1; i >= 0; i--)
- {
- md_number_to_chars (litP, (valueT) words[i], sizeof (LITTLENUM_TYPE));
- litP += sizeof (LITTLENUM_TYPE);
- }
- else
- /* For a 4 byte float the order of elements in `words' is 1 0.
- For an 8 byte float the order is 1 0 3 2. */
- for (i = 0; i < prec; i += 2)
- {
- md_number_to_chars (litP, (valueT) words[i + 1],
- sizeof (LITTLENUM_TYPE));
- md_number_to_chars (litP + sizeof (LITTLENUM_TYPE),
- (valueT) words[i], sizeof (LITTLENUM_TYPE));
- litP += 2 * sizeof (LITTLENUM_TYPE);
- }
- }
+ /* For a 4 byte float the order of elements in `words' is 1 0.
+ For an 8 byte float the order is 1 0 3 2. */
+ for (i = 0; i < prec; i += 2)
+ {
+ md_number_to_chars (litP, (valueT) words[i + 1],
+ sizeof (LITTLENUM_TYPE));
+ md_number_to_chars (litP + sizeof (LITTLENUM_TYPE),
+ (valueT) words[i], sizeof (LITTLENUM_TYPE));
+ litP += 2 * sizeof (LITTLENUM_TYPE);
+ }
return NULL;
}
thissize = 64;
ptr++;
goto done;
+ case 'b':
+ thistype = NT_bfloat;
+ switch (TOLOWER (*(++ptr)))
+ {
+ case 'f':
+ ptr += 1;
+ thissize = strtoul (ptr, &ptr, 10);
+ if (thissize != 16)
+ {
+ as_bad (_("bad size %d in type specifier"), thissize);
+ return FAIL;
+ }
+ goto done;
+ case '0': case '1': case '2': case '3': case '4':
+ case '5': case '6': case '7': case '8': case '9':
+ case ' ': case '.':
+ as_bad (_("unexpected type character `b' -- did you mean `bf'?"));
+ return FAIL;
+ default:
+ break;
+ }
+ break;
default:
as_bad (_("unexpected character `%c' in type specifier"), *ptr);
return FAIL;
const char apsr_str[] = "apsr";
int apsr_str_len = strlen (apsr_str);
- reg = arm_reg_parse (&str, REGLIST_RN);
+ reg = arm_reg_parse (&str, REG_TYPE_RN);
if (etype == REGLIST_CLRM)
{
if (reg == REG_SP || reg == REG_PC)
}
#endif /* TE_PE */
+int
+arm_is_largest_exponent_ok (int precision)
+{
+ /* precision == 1 ensures that this will only return
+ true for 16 bit floats. */
+ return (precision == 1) && (fp16_format == ARM_FP16_FORMAT_ALTERNATIVE);
+}
+
+static void
+set_fp16_format (int dummy ATTRIBUTE_UNUSED)
+{
+ char saved_char;
+ char* name;
+ enum fp_16bit_format new_format;
+
+ new_format = ARM_FP16_FORMAT_DEFAULT;
+
+ name = input_line_pointer;
+ while (*input_line_pointer && !ISSPACE (*input_line_pointer))
+ input_line_pointer++;
+
+ saved_char = *input_line_pointer;
+ *input_line_pointer = 0;
+
+ if (strcasecmp (name, "ieee") == 0)
+ new_format = ARM_FP16_FORMAT_IEEE;
+ else if (strcasecmp (name, "alternative") == 0)
+ new_format = ARM_FP16_FORMAT_ALTERNATIVE;
+ else
+ {
+ as_bad (_("unrecognised float16 format \"%s\""), name);
+ goto cleanup;
+ }
+
+ /* Only set fp16_format if it is still the default (aka not already
+ been set yet). */
+ if (fp16_format == ARM_FP16_FORMAT_DEFAULT)
+ fp16_format = new_format;
+ else
+ {
+ if (new_format != fp16_format)
+ as_warn (_("float16 format cannot be set more than once, ignoring."));
+ }
+
+cleanup:
+ *input_line_pointer = saved_char;
+ ignore_rest_of_line ();
+}
+
/* This table describes all the machine specific pseudo-ops the assembler
has to support. The fields are:
pseudo-op name without dot
{ "extend", float_cons, 'x' },
{ "ldouble", float_cons, 'x' },
{ "packed", float_cons, 'p' },
+ { "bfloat16", float_cons, 'b' },
#ifdef TE_PE
{"secrel32", pe_directive_secrel, 0},
#endif
{"asmfunc", s_ccs_asmfunc, 0},
{"endasmfunc", s_ccs_endasmfunc, 0},
+ {"float16", float_cons, 'h' },
+ {"float16_format", set_fp16_format, 0 },
+
{ 0, 0, 0 }
};
-\f
+
/* Parser functions used exclusively in instruction operands. */
/* Generic immediate-value read function for use in insn parsing.
inst.operands[i].present = 1;
}
}
- else if ((val = arm_typed_reg_parse (&ptr, REG_TYPE_NSDQ, &rtype,
- &optype)) != FAIL)
+ else if (((val = arm_typed_reg_parse (&ptr, REG_TYPE_NSDQ, &rtype,
+ &optype)) != FAIL)
+ || ((val = arm_typed_reg_parse (&ptr, REG_TYPE_MQ, &rtype,
+ &optype)) != FAIL))
{
/* Case 0: VMOV<c><q> <Qd>, <Qm>
Case 1: VMOV<c><q> <Dd>, <Dm>
OP_I31w, /* 0 .. 31, optional trailing ! */
OP_I32, /* 1 .. 32 */
OP_I32z, /* 0 .. 32 */
+ OP_I48_I64, /* 48 or 64 */
OP_I63, /* 0 .. 63 */
OP_I63s, /* -64 .. 63 */
OP_I64, /* 1 .. 64 */
} \
while (0)
+#define po_imm1_or_imm2_or_fail(imm1, imm2, popt) \
+ do \
+ { \
+ expressionS exp; \
+ my_get_expression (&exp, &str, popt); \
+ if (exp.X_op != O_constant) \
+ { \
+ inst.error = _("constant expression required"); \
+ goto failure; \
+ } \
+ if (exp.X_add_number != imm1 && exp.X_add_number != imm2) \
+ { \
+ inst.error = _("immediate value 48 or 64 expected"); \
+ goto failure; \
+ } \
+ inst.operands[i].imm = exp.X_add_number; \
+ } \
+ while (0)
+
#define po_scalar_or_goto(elsz, label, reg_type) \
do \
{ \
case OP_I31: po_imm_or_fail ( 0, 31, FALSE); break;
case OP_I32: po_imm_or_fail ( 1, 32, FALSE); break;
case OP_I32z: po_imm_or_fail ( 0, 32, FALSE); break;
+ case OP_I48_I64: po_imm1_or_imm2_or_fail (48, 64, FALSE); break;
case OP_I63s: po_imm_or_fail (-64, 63, FALSE); break;
case OP_I63: po_imm_or_fail ( 0, 63, FALSE); break;
case OP_I64: po_imm_or_fail ( 1, 64, FALSE); break;
}
}
+/* For shifts with four operands in MVE. */
+static void
+do_mve_scalar_shift1 (void)
+{
+ unsigned int value = inst.operands[2].imm;
+
+ inst.instruction |= inst.operands[0].reg << 16;
+ inst.instruction |= inst.operands[1].reg << 8;
+
+ /* Setting the bit for saturation. */
+ inst.instruction |= ((value == 64) ? 0: 1) << 7;
+
+ /* Assuming Rm is already checked not to be 11x1. */
+ constraint (inst.operands[3].reg == inst.operands[0].reg, BAD_OVERLAP);
+ constraint (inst.operands[3].reg == inst.operands[1].reg, BAD_OVERLAP);
+ inst.instruction |= inst.operands[3].reg << 12;
+}
+
/* For shifts in MVE. */
static void
do_mve_scalar_shift (void)
#define M_MNEM_vmlsdavax 0xeef01e21
#define M_MNEM_vmullt 0xee011e00
#define M_MNEM_vmullb 0xee010e00
+#define M_MNEM_vctp 0xf000e801
#define M_MNEM_vst20 0xfc801e00
#define M_MNEM_vst21 0xfc801e20
#define M_MNEM_vst40 0xfc801e01
#define M_MNEM_vqrshrunt 0xfe801fc0
#define M_MNEM_vqrshrunb 0xfe800fc0
+/* Bfloat16 instruction encoder helpers. */
+#define B_MNEM_vfmat 0xfc300850
+#define B_MNEM_vfmab 0xfc300810
+
/* Neon instruction encoder helpers. */
/* Encodings for the different types for various Neon opcodes. */
N_F32 = 0x0080000,
N_F64 = 0x0100000,
N_P64 = 0x0200000,
+ N_BF16 = 0x0400000,
N_KEY = 0x1000000, /* Key element (main type specifier). */
N_EQK = 0x2000000, /* Given operand has the same type & size as the key. */
N_VFP = 0x4000000, /* VFP mode: operand size must match register width. */
}
break;
+ case NT_bfloat:
+ if (size == 16) return N_BF16;
+ break;
+
default: ;
}
if ((mask & (N_S8 | N_U8 | N_I8 | N_8 | N_P8)) != 0)
*size = 8;
- else if ((mask & (N_S16 | N_U16 | N_I16 | N_16 | N_F16 | N_P16)) != 0)
+ else if ((mask & (N_S16 | N_U16 | N_I16 | N_16 | N_F16 | N_P16 | N_BF16))
+ != 0)
*size = 16;
else if ((mask & (N_S32 | N_U32 | N_I32 | N_32 | N_F32)) != 0)
*size = 32;
*type = NT_poly;
else if ((mask & (N_F_ALL)) != 0)
*type = NT_float;
+ else if ((mask & (N_BF16)) != 0)
+ *type = NT_bfloat;
else
return FAIL;
abort ();
}
+/* For VCTP (create vector tail predicate) in MVE. */
+static void
+do_mve_vctp (void)
+{
+ int dt = 0;
+ unsigned size = 0x0;
+
+ if (inst.cond > COND_ALWAYS)
+ inst.pred_insn_type = INSIDE_VPT_INSN;
+ else
+ inst.pred_insn_type = MVE_OUTSIDE_PRED_INSN;
+
+ /* This is a typical MVE instruction which has no type but have size 8, 16,
+ 32 and 64. For instructions with no type, inst.vectype.el[j].type is set
+ to NT_untyped and size is updated in inst.vectype.el[j].size. */
+ if ((inst.operands[0].present) && (inst.vectype.el[0].type == NT_untyped))
+ dt = inst.vectype.el[0].size;
+
+ /* Setting this does not indicate an actual NEON instruction, but only
+ indicates that the mnemonic accepts neon-style type suffixes. */
+ inst.is_neon = 1;
+
+ switch (dt)
+ {
+ case 8:
+ break;
+ case 16:
+ size = 0x1; break;
+ case 32:
+ size = 0x2; break;
+ case 64:
+ size = 0x3; break;
+ default:
+ first_error (_("Type is not allowed for this instruction"));
+ }
+ inst.instruction |= size << 20;
+ inst.instruction |= inst.operands[0].reg << 16;
+}
+
static void
do_mve_vpt (void)
{
inst.operands[0].present = 1;
}
-static void
-do_vfp_nsyn_push (void)
-{
- nsyn_insert_sp ();
-
- constraint (inst.operands[1].imm < 1 || inst.operands[1].imm > 16,
- _("register list must contain at least 1 and at most 16 "
- "registers"));
-
- if (inst.operands[1].issingle)
- do_vfp_nsyn_opcode ("fstmdbs");
- else
- do_vfp_nsyn_opcode ("fstmdbd");
-}
-
-static void
-do_vfp_nsyn_pop (void)
-{
- nsyn_insert_sp ();
-
- constraint (inst.operands[1].imm < 1 || inst.operands[1].imm > 16,
- _("register list must contain at least 1 and at most 16 "
- "registers"));
-
- if (inst.operands[1].issingle)
- do_vfp_nsyn_opcode ("fldmias");
- else
- do_vfp_nsyn_opcode ("fldmiad");
-}
-
/* Fix up Neon data-processing instructions, ORing in the correct bits for
ARM mode or Thumb mode and moving the encoded bit 24 to bit 28. */
inst.is_neon = 1;
}
+/* Helper function for neon_three_same handling the operands. */
+static void
+neon_three_args (int isquad)
+{
+ inst.instruction |= LOW4 (inst.operands[0].reg) << 12;
+ inst.instruction |= HI1 (inst.operands[0].reg) << 22;
+ inst.instruction |= LOW4 (inst.operands[1].reg) << 16;
+ inst.instruction |= HI1 (inst.operands[1].reg) << 7;
+ inst.instruction |= LOW4 (inst.operands[2].reg);
+ inst.instruction |= HI1 (inst.operands[2].reg) << 5;
+ inst.instruction |= (isquad != 0) << 6;
+ inst.is_neon = 1;
+}
+
/* Encode insns with bit pattern:
|28/24|23|22 |21 20|19 16|15 12|11 8|7|6|5|4|3 0|
static void
neon_three_same (int isquad, int ubit, int size)
{
- inst.instruction |= LOW4 (inst.operands[0].reg) << 12;
- inst.instruction |= HI1 (inst.operands[0].reg) << 22;
- inst.instruction |= LOW4 (inst.operands[1].reg) << 16;
- inst.instruction |= HI1 (inst.operands[1].reg) << 7;
- inst.instruction |= LOW4 (inst.operands[2].reg);
- inst.instruction |= HI1 (inst.operands[2].reg) << 5;
- inst.instruction |= (isquad != 0) << 6;
+ neon_three_args (isquad);
inst.instruction |= (ubit != 0) << 24;
if (size != -1)
inst.instruction |= neon_logbits (size) << 20;
do_mve_vstr_vldr_RQ (int size, int elsize, int load)
{
unsigned os = inst.operands[1].imm >> 5;
+ unsigned type = inst.vectype.el[0].type;
constraint (os != 0 && size == 8,
_("can not shift offsets when accessing less than half-word"));
constraint (os && os != neon_logbits (size),
constraint (inst.operands[0].reg == (inst.operands[1].imm & 0x1f),
_("destination register and offset register may not be"
" the same"));
- constraint (size == elsize && inst.vectype.el[0].type != NT_unsigned,
+ constraint (size == elsize && type == NT_signed, BAD_EL_TYPE);
+ constraint (size != elsize && type != NT_unsigned && type != NT_signed,
BAD_EL_TYPE);
- constraint (inst.vectype.el[0].type != NT_unsigned
- && inst.vectype.el[0].type != NT_signed, BAD_EL_TYPE);
- inst.instruction |= (inst.vectype.el[0].type == NT_unsigned) << 28;
+ inst.instruction |= ((size == elsize) || (type == NT_unsigned)) << 28;
}
else
{
- constraint (inst.vectype.el[0].type != NT_untyped, BAD_EL_TYPE);
+ constraint (type != NT_untyped, BAD_EL_TYPE);
}
inst.instruction |= 1 << 23;
}
}
+static void
+do_bfloat_vfma (void)
+{
+ constraint (!mark_feature_used (&fpu_neon_ext_armv8), _(BAD_FPU));
+ constraint (!mark_feature_used (&arm_ext_bf16), _(BAD_BF16));
+ enum neon_shape rs;
+ int t_bit = 0;
+
+ if (inst.instruction != B_MNEM_vfmab)
+ {
+ t_bit = 1;
+ inst.instruction = B_MNEM_vfmat;
+ }
+
+ if (inst.operands[2].isscalar)
+ {
+ rs = neon_select_shape (NS_QQS, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_BF16 | N_KEY);
+
+ inst.instruction |= (1 << 25);
+ int index = inst.operands[2].reg & 0xf;
+ constraint (!(index < 4), _("index must be in the range 0 to 3"));
+ inst.operands[2].reg >>= 4;
+ constraint (!(inst.operands[2].reg < 8),
+ _("indexed register must be less than 8"));
+ neon_three_args (t_bit);
+ inst.instruction |= ((index & 1) << 3);
+ inst.instruction |= ((index & 2) << 4);
+ }
+ else
+ {
+ rs = neon_select_shape (NS_QQQ, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_BF16 | N_KEY);
+ neon_three_args (t_bit);
+ }
+
+}
+
static void
do_neon_fmac (void)
{
if (rs == NS_QQR)
{
+
if (inst.operands[2].reg == REG_SP)
as_tsktsk (MVE_BAD_SP);
else if (inst.operands[2].reg == REG_PC)
neon_dyadic_misc (NT_untyped, N_IF_32, 0);
}
+static void
+do_mve_vfma (void)
+{
+ if (!ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_bf16) &&
+ inst.cond == COND_ALWAYS)
+ {
+ constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, mve_ext), BAD_FPU);
+ inst.instruction = N_MNEM_vfma;
+ inst.pred_insn_type = INSIDE_VPT_INSN;
+ inst.cond = 0xf;
+ return do_neon_fmac();
+ }
+ else
+ {
+ do_bfloat_vfma();
+ }
+}
+
static void
do_neon_tst (void)
{
{
enum neon_shape rs = neon_select_shape (NS_QQR, NS_NULL);
struct neon_type_el et
- = neon_check_type (3, rs, N_EQK, N_EQK, N_SU_MVE | N_KEY);
+ = neon_check_type (3, rs, N_EQK, N_EQK, N_S_32 | N_KEY);
if (inst.cond > COND_ALWAYS)
inst.pred_insn_type = INSIDE_VPT_INSN;
else
inst.pred_insn_type = MVE_OUTSIDE_PRED_INSN;
- if (et.size == 32
- && (inst.operands[0].reg == inst.operands[1].reg
- || inst.operands[0].reg == inst.operands[2].reg))
- as_tsktsk (BAD_MVE_SRCDEST);
-
mve_encode_qqq (0, et.size);
}
{
enum neon_shape rs = neon_select_shape (NS_QQR, NS_NULL);
struct neon_type_el et
- = neon_check_type (3, rs, N_EQK, N_EQK, N_SU_MVE | N_KEY);
+ = neon_check_type (3, rs, N_EQK, N_EQK, N_S_32 | N_KEY);
NEON_ENCODE (INTEGER, inst);
mve_encode_qqr (et.size, et.type == NT_unsigned, 0);
CVT_VAR (f16_u32, N_F16 | N_KEY, N_U32, N_VFP, "fultos", "fuitos", NULL) \
CVT_VAR (u32_f16, N_U32, N_F16 | N_KEY, N_VFP, "ftouls", "ftouis", "ftouizs")\
CVT_VAR (s32_f16, N_S32, N_F16 | N_KEY, N_VFP, "ftosls", "ftosis", "ftosizs")\
+ CVT_VAR (bf16_f32, N_BF16, N_F32, whole_reg, NULL, NULL, NULL) \
/* VFP instructions. */ \
CVT_VAR (f32_f64, N_F32, N_F64, N_VFP, NULL, "fcvtsd", NULL) \
CVT_VAR (f64_f32, N_F64, N_F32, N_VFP, NULL, "fcvtds", NULL) \
}
if (rs == NS_DQ)
- inst.instruction = 0x3b60600;
+ {
+ if (flavour == neon_cvt_flavour_bf16_f32)
+ {
+ if (vfp_or_neon_is_neon (NEON_CHECK_ARCH8) == FAIL)
+ return;
+ constraint (!mark_feature_used (&arm_ext_bf16), _(BAD_BF16));
+ /* VCVT.bf16.f32. */
+ inst.instruction = 0x11b60640;
+ }
+ else
+ /* VCVT.f16.f32. */
+ inst.instruction = 0x3b60600;
+ }
else
+ /* VCVT.f32.f16. */
inst.instruction = 0x3b60700;
inst.instruction |= LOW4 (inst.operands[0].reg) << 12;
inst.error = NULL;
do_neon_cvttb_2 (t, /*to=*/FALSE, /*is_double=*/TRUE);
}
+ else if (neon_check_type (2, rs, N_BF16 | N_VFP, N_F32).type != NT_invtype)
+ {
+ constraint (!mark_feature_used (&arm_ext_bf16), _(BAD_BF16));
+ inst.error = NULL;
+ inst.instruction |= (1 << 8);
+ inst.instruction &= ~(1 << 9);
+ do_neon_cvttb_2 (t, /*to=*/TRUE, /*is_double=*/FALSE);
+ }
else
return;
}
if (ARM_CPU_HAS_FEATURE (cpu_variant, mve_ext))
{
constraint (!inst.operands[1].isreg && !inst.operands[0].isquad, BAD_FPU);
- constraint ((inst.instruction & 0xd00) == 0xd00,
- _("immediate value out of range"));
}
}
0x2. */
int size = -1;
- if (inst.cond != COND_ALWAYS)
- as_warn (_("vfmal/vfmsl with FP16 type cannot be conditional, the "
- "behaviour is UNPREDICTABLE"));
-
- constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_fp16_fml),
- _(BAD_FP16));
-
- constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, fpu_neon_ext_armv8),
- _(BAD_FPU));
-
/* vfmal/vfmsl are in three-same D/Q register format or the third operand can
be a scalar index register. */
if (inst.operands[2].isscalar)
rs = neon_select_shape (NS_DHH, NS_QDD, NS_NULL);
}
- neon_check_type (3, rs, N_EQK, N_EQK, N_KEY | N_F16);
+
+ if (inst.cond != COND_ALWAYS)
+ as_warn (_("vfmal/vfmsl with FP16 type cannot be conditional, the "
+ "behaviour is UNPREDICTABLE"));
+
+ constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_fp16_fml),
+ _(BAD_FP16));
+
+ constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, fpu_neon_ext_armv8),
+ _(BAD_FPU));
/* "opcode" from template has included "ubit", so simply pass 0 here. Also,
the "S" bit in size field has been reused to differentiate vfmal and vfmsl,
et = neon_check_type (2, rs, N_EQK, N_F64 | N_KEY);
/* It is not an error here if no type is given. */
inst.error = NULL;
- if (et.type == NT_float && et.size == 64)
+
+ /* In MVE we interpret the following instructions as same, so ignoring
+ the following type (float) and size (64) checks.
+ a: VMOV<c><q> <Dd>, <Dm>
+ b: VMOV<c><q>.F64 <Dd>, <Dm>. */
+ if ((et.type == NT_float && et.size == 64)
+ || (ARM_CPU_HAS_FEATURE (cpu_variant, mve_ext)))
{
do_vfp_nsyn_opcode ("fcpyd");
break;
static void
do_neon_ldm_stm (void)
{
+ constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, fpu_vfp_ext_v1xd)
+ && !ARM_CPU_HAS_FEATURE (cpu_variant, mve_ext),
+ _(BAD_FPU));
/* P, U and L bits are part of bitmask. */
int is_dbmode = (inst.instruction & (1 << 24)) != 0;
unsigned offsetbits = inst.operands[1].imm * 2;
do_vfp_cond_or_thumb ();
}
+static void
+do_vfp_nsyn_pop (void)
+{
+ nsyn_insert_sp ();
+ if (ARM_CPU_HAS_FEATURE (cpu_variant, mve_ext)) {
+ return do_vfp_nsyn_opcode ("vldm");
+ }
+
+ constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, fpu_vfp_ext_v1xd),
+ _(BAD_FPU));
+
+ constraint (inst.operands[1].imm < 1 || inst.operands[1].imm > 16,
+ _("register list must contain at least 1 and at most 16 "
+ "registers"));
+
+ if (inst.operands[1].issingle)
+ do_vfp_nsyn_opcode ("fldmias");
+ else
+ do_vfp_nsyn_opcode ("fldmiad");
+}
+
+static void
+do_vfp_nsyn_push (void)
+{
+ nsyn_insert_sp ();
+ if (ARM_CPU_HAS_FEATURE (cpu_variant, mve_ext)) {
+ return do_vfp_nsyn_opcode ("vstmdb");
+ }
+
+ constraint (!ARM_CPU_HAS_FEATURE (cpu_variant, fpu_vfp_ext_v1xd),
+ _(BAD_FPU));
+
+ constraint (inst.operands[1].imm < 1 || inst.operands[1].imm > 16,
+ _("register list must contain at least 1 and at most 16 "
+ "registers"));
+
+ if (inst.operands[1].issingle)
+ do_vfp_nsyn_opcode ("fstmdbs");
+ else
+ do_vfp_nsyn_opcode ("fstmdbd");
+}
+
+
static void
do_neon_ldr_str (void)
{
/* VLDR/VSTR. */
else
{
- if (!mark_feature_used (&fpu_vfp_ext_v1xd))
+ if (!mark_feature_used (&fpu_vfp_ext_v1xd)
+ && !ARM_CPU_HAS_FEATURE (cpu_variant, mve_ext))
as_bad (_("Instruction not permitted on this architecture"));
do_neon_ldr_str ();
}
return do_neon_dotproduct (1);
}
+static void
+do_vusdot (void)
+{
+ enum neon_shape rs;
+ set_pred_insn_type (OUTSIDE_PRED_INSN);
+ if (inst.operands[2].isscalar)
+ {
+ rs = neon_select_shape (NS_DDS, NS_QQS, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_S8 | N_KEY);
+
+ inst.instruction |= (1 << 25);
+ int index = inst.operands[2].reg & 0xf;
+ constraint ((index != 1 && index != 0), _("index must be 0 or 1"));
+ inst.operands[2].reg >>= 4;
+ constraint (!(inst.operands[2].reg < 16),
+ _("indexed register must be less than 16"));
+ neon_three_args (rs == NS_QQS);
+ inst.instruction |= (index << 5);
+ }
+ else
+ {
+ inst.instruction |= (1 << 21);
+ rs = neon_select_shape (NS_DDD, NS_QQQ, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_S8 | N_KEY);
+ neon_three_args (rs == NS_QQQ);
+ }
+}
+
+static void
+do_vsudot (void)
+{
+ enum neon_shape rs;
+ set_pred_insn_type (OUTSIDE_PRED_INSN);
+ if (inst.operands[2].isscalar)
+ {
+ rs = neon_select_shape (NS_DDS, NS_QQS, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_U8 | N_KEY);
+
+ inst.instruction |= (1 << 25);
+ int index = inst.operands[2].reg & 0xf;
+ constraint ((index != 1 && index != 0), _("index must be 0 or 1"));
+ inst.operands[2].reg >>= 4;
+ constraint (!(inst.operands[2].reg < 16),
+ _("indexed register must be less than 16"));
+ neon_three_args (rs == NS_QQS);
+ inst.instruction |= (index << 5);
+ }
+}
+
+static void
+do_vsmmla (void)
+{
+ enum neon_shape rs = neon_select_shape (NS_QQQ, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_S8 | N_KEY);
+
+ set_pred_insn_type (OUTSIDE_PRED_INSN);
+
+ neon_three_args (1);
+
+}
+
+static void
+do_vummla (void)
+{
+ enum neon_shape rs = neon_select_shape (NS_QQQ, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_U8 | N_KEY);
+
+ set_pred_insn_type (OUTSIDE_PRED_INSN);
+
+ neon_three_args (1);
+
+}
+
/* Crypto v1 instructions. */
static void
do_crypto_2op_1 (unsigned elttype, int op)
do_vfp_cond_or_thumb ();
}
+static void
+do_vdot (void)
+{
+ enum neon_shape rs;
+ constraint (!mark_feature_used (&fpu_neon_ext_armv8), _(BAD_FPU));
+ set_pred_insn_type (OUTSIDE_PRED_INSN);
+ if (inst.operands[2].isscalar)
+ {
+ rs = neon_select_shape (NS_DDS, NS_QQS, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_BF16 | N_KEY);
+
+ inst.instruction |= (1 << 25);
+ int index = inst.operands[2].reg & 0xf;
+ constraint ((index != 1 && index != 0), _("index must be 0 or 1"));
+ inst.operands[2].reg >>= 4;
+ constraint (!(inst.operands[2].reg < 16),
+ _("indexed register must be less than 16"));
+ neon_three_args (rs == NS_QQS);
+ inst.instruction |= (index << 5);
+ }
+ else
+ {
+ rs = neon_select_shape (NS_DDD, NS_QQQ, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_BF16 | N_KEY);
+ neon_three_args (rs == NS_QQQ);
+ }
+}
+
+static void
+do_vmmla (void)
+{
+ enum neon_shape rs = neon_select_shape (NS_QQQ, NS_NULL);
+ neon_check_type (3, rs, N_EQK, N_EQK, N_BF16 | N_KEY);
+
+ constraint (!mark_feature_used (&fpu_neon_ext_armv8), _(BAD_FPU));
+ set_pred_insn_type (OUTSIDE_PRED_INSN);
+
+ neon_three_args (1);
+}
+
\f
/* Overall per-instruction processing. */
handle_pred_state ();
if (now_pred.insn_cond
+ && warn_on_restrict_it
&& !now_pred.warn_deprecated
&& warn_on_deprecated
&& ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_v8)
out of the jump table, and chaos would ensue. */
if (label_is_thumb_function_name
&& (S_GET_NAME (sym)[0] != '.' || S_GET_NAME (sym)[1] != 'L')
- && (bfd_get_section_flags (stdoutput, now_seg) & SEC_CODE) != 0)
+ && (bfd_section_flags (now_seg) & SEC_CODE) != 0)
{
/* When the address of a Thumb function is taken the bottom
bit of that address should be set. This will allow
nUF(sha256su0, _sha2op, 2, (RNQ, RNQ), sha256su0),
#undef ARM_VARIANT
-#define ARM_VARIANT & crc_ext_armv8
+#define ARM_VARIANT & arm_ext_crc
#undef THUMB_VARIANT
-#define THUMB_VARIANT & crc_ext_armv8
+#define THUMB_VARIANT & arm_ext_crc
TUEc("crc32b", 1000040, fac0f080, 3, (RR, oRR, RR), crc32b),
TUEc("crc32h", 1200040, fac0f090, 3, (RR, oRR, RR), crc32h),
TUEc("crc32w", 1400040, fac0f0a0, 3, (RR, oRR, RR), crc32w),
#define THUMB_VARIANT & arm_ext_v6t2
mcCE(vmrs, ef00a10, 2, (APSR_RR, RVC), vmrs),
mcCE(vmsr, ee00a10, 2, (RVC, RR), vmsr),
+ mcCE(fldd, d100b00, 2, (RVD, ADDRGLDC), vfp_dp_ldst),
+ mcCE(fstd, d000b00, 2, (RVD, ADDRGLDC), vfp_dp_ldst),
+ mcCE(flds, d100a00, 2, (RVS, ADDRGLDC), vfp_sp_ldst),
+ mcCE(fsts, d000a00, 2, (RVS, ADDRGLDC), vfp_sp_ldst),
+
+ /* Memory operations. */
+ mcCE(fldmias, c900a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmia),
+ mcCE(fldmdbs, d300a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmdb),
+ mcCE(fstmias, c800a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmia),
+ mcCE(fstmdbs, d200a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmdb),
#undef THUMB_VARIANT
/* Moves and type conversions. */
cCE("fmxr", ee00a10, 2, (RVC, RR), rn_rd),
/* Memory operations. */
- cCE("flds", d100a00, 2, (RVS, ADDRGLDC), vfp_sp_ldst),
- cCE("fsts", d000a00, 2, (RVS, ADDRGLDC), vfp_sp_ldst),
- cCE("fldmias", c900a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmia),
cCE("fldmfds", c900a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmia),
- cCE("fldmdbs", d300a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmdb),
cCE("fldmeas", d300a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmdb),
cCE("fldmiax", c900b00, 2, (RRnpctw, VRDLST), vfp_xp_ldstmia),
cCE("fldmfdx", c900b00, 2, (RRnpctw, VRDLST), vfp_xp_ldstmia),
cCE("fldmdbx", d300b00, 2, (RRnpctw, VRDLST), vfp_xp_ldstmdb),
cCE("fldmeax", d300b00, 2, (RRnpctw, VRDLST), vfp_xp_ldstmdb),
- cCE("fstmias", c800a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmia),
cCE("fstmeas", c800a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmia),
- cCE("fstmdbs", d200a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmdb),
cCE("fstmfds", d200a00, 2, (RRnpctw, VRSLST), vfp_sp_ldstmdb),
cCE("fstmiax", c800b00, 2, (RRnpctw, VRDLST), vfp_xp_ldstmia),
cCE("fstmeax", c800b00, 2, (RRnpctw, VRDLST), vfp_xp_ldstmia),
/* Double precision load/store are still present on single precision
implementations. */
- cCE("fldd", d100b00, 2, (RVD, ADDRGLDC), vfp_dp_ldst),
- cCE("fstd", d000b00, 2, (RVD, ADDRGLDC), vfp_dp_ldst),
cCE("fldmiad", c900b00, 2, (RRnpctw, VRDLST), vfp_dp_ldstmia),
cCE("fldmfdd", c900b00, 2, (RRnpctw, VRDLST), vfp_dp_ldstmia),
cCE("fldmdbd", d300b00, 2, (RRnpctw, VRDLST), vfp_dp_ldstmdb),
Individual encoder functions perform additional architecture checks. */
#undef ARM_VARIANT
#define ARM_VARIANT & fpu_vfp_ext_v1xd
+#undef THUMB_VARIANT
+#define THUMB_VARIANT & arm_ext_v6t2
+
+ NCE(vldm, c900b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
+ NCE(vldmia, c900b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
+ NCE(vldmdb, d100b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
+ NCE(vstm, c800b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
+ NCE(vstmia, c800b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
+ NCE(vstmdb, d000b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
+
+ NCE(vpop, 0, 1, (VRSDLST), vfp_nsyn_pop),
+ NCE(vpush, 0, 1, (VRSDLST), vfp_nsyn_push),
+
#undef THUMB_VARIANT
#define THUMB_VARIANT & fpu_vfp_ext_v1xd
nCE(vnmul, _vnmul, 3, (RVSD, RVSD, RVSD), vfp_nsyn_nmul),
nCE(vnmla, _vnmla, 3, (RVSD, RVSD, RVSD), vfp_nsyn_nmul),
nCE(vnmls, _vnmls, 3, (RVSD, RVSD, RVSD), vfp_nsyn_nmul),
- NCE(vpush, 0, 1, (VRSDLST), vfp_nsyn_push),
- NCE(vpop, 0, 1, (VRSDLST), vfp_nsyn_pop),
NCE(vcvtz, 0, 2, (RVSD, RVSD), vfp_nsyn_cvtz),
/* Mnemonics shared by Neon and VFP. */
nCEF(vmls, _vmls, 3, (RNSDQ, oRNSDQ, RNSDQ_RNSC), neon_mac_maybe_scalar),
- NCE(vldm, c900b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
- NCE(vldmia, c900b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
- NCE(vldmdb, d100b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
- NCE(vstm, c800b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
- NCE(vstmia, c800b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
- NCE(vstmdb, d000b00, 2, (RRnpctw, VRSDLST), neon_ldm_stm),
-
mnCEF(vcvt, _vcvt, 3, (RNSDQMQ, RNSDQMQ, oI32z), neon_cvt),
nCEF(vcvtr, _vcvt, 2, (RNSDQ, RNSDQ), neon_cvtr),
MNCEF(vcvtb, eb20a40, 3, (RVSDMQ, RVSDMQ, oI32b), neon_cvtb),
NCE (vins, eb00ac0, 2, (RVS, RVS), neon_movhf),
/* New backported fma/fms instructions optional in v8.2. */
- NCE (vfmal, 810, 3, (RNDQ, RNSD, RNSD_RNSC), neon_vfmal),
- NCE (vfmsl, 810, 3, (RNDQ, RNSD, RNSD_RNSC), neon_vfmsl),
+ NUF (vfmsl, 810, 3, (RNDQ, RNSD, RNSD_RNSC), neon_vfmsl),
+ NUF (vfmal, 810, 3, (RNDQ, RNSD, RNSD_RNSC), neon_vfmal),
#undef THUMB_VARIANT
#define THUMB_VARIANT & fpu_neon_ext_v1
#define ARM_VARIANT & fpu_vfp_ext_fma
#undef THUMB_VARIANT
#define THUMB_VARIANT & fpu_vfp_ext_fma
- /* Mnemonics shared by Neon, VFP and MVE. These are included in the
+ /* Mnemonics shared by Neon, VFP, MVE and BF16. These are included in the
VFP FMA variant; NEON and VFP FMA always includes the NEON
FMA instructions. */
mnCEF(vfma, _vfma, 3, (RNSDQMQ, oRNSDQMQ, RNSDQMQR), neon_fmac),
+ TUF ("vfmat", c300850, fc300850, 3, (RNSDQMQ, oRNSDQMQ, RNSDQ_RNSC_MQ_RR), mve_vfma, mve_vfma),
mnCEF(vfms, _vfms, 3, (RNSDQMQ, oRNSDQMQ, RNSDQMQ), neon_fmac),
/* ffmas/ffmad/ffmss/ffmsd are dummy mnemonics to satisfy gas;
ToC("lsll", ea50010d, 3, (RRe, RRo, RRnpcsp_I32), mve_scalar_shift),
ToC("lsrl", ea50011f, 3, (RRe, RRo, I32), mve_scalar_shift),
ToC("asrl", ea50012d, 3, (RRe, RRo, RRnpcsp_I32), mve_scalar_shift),
- ToC("uqrshll", ea51010d, 3, (RRe, RRo, RRnpcsp), mve_scalar_shift),
- ToC("sqrshrl", ea51012d, 3, (RRe, RRo, RRnpcsp), mve_scalar_shift),
+ ToC("uqrshll", ea51010d, 4, (RRe, RRo, I48_I64, RRnpcsp), mve_scalar_shift1),
+ ToC("sqrshrl", ea51012d, 4, (RRe, RRo, I48_I64, RRnpcsp), mve_scalar_shift1),
ToC("uqshll", ea51010f, 3, (RRe, RRo, I32), mve_scalar_shift),
ToC("urshrl", ea51011f, 3, (RRe, RRo, I32), mve_scalar_shift),
ToC("srshrl", ea51012f, 3, (RRe, RRo, I32), mve_scalar_shift),
/* MVE and MVE FP only. */
mToC("vhcadd", ee000f00, 4, (RMQ, RMQ, RMQ, EXPi), mve_vhcadd),
+ mCEF(vctp, _vctp, 1, (RRnpc), mve_vctp),
mCEF(vadc, _vadc, 3, (RMQ, RMQ, RMQ), mve_vadc),
mCEF(vadci, _vadci, 3, (RMQ, RMQ, RMQ), mve_vadc),
mToC("vsbc", fe300f00, 3, (RMQ, RMQ, RMQ), mve_vsbc),
#define THUMB_VARIANT & arm_ext_v6t2_v8m
MNUF (vcadd, 0, 4, (RNDQMQ, RNDQMQ, RNDQMQ, EXPi), vcadd),
MNUF (vcmla, 0, 4, (RNDQMQ, RNDQMQ, RNDQMQ_RNSC, EXPi), vcmla),
+
+#undef ARM_VARIANT
+#define ARM_VARIANT &arm_ext_bf16
+#undef THUMB_VARIANT
+#define THUMB_VARIANT &arm_ext_bf16
+ TUF ("vdot", c000d00, fc000d00, 3, (RNDQ, RNDQ, RNDQ_RNSC), vdot, vdot),
+ TUF ("vmmla", c000c40, fc000c40, 3, (RNQ, RNQ, RNQ), vmmla, vmmla),
+ TUF ("vfmab", c300810, fc300810, 3, (RNDQ, RNDQ, RNDQ_RNSC), bfloat_vfma, bfloat_vfma),
+
+#undef ARM_VARIANT
+#define ARM_VARIANT &arm_ext_i8mm
+#undef THUMB_VARIANT
+#define THUMB_VARIANT &arm_ext_i8mm
+ TUF ("vsmmla", c200c40, fc200c40, 3, (RNQ, RNQ, RNQ), vsmmla, vsmmla),
+ TUF ("vummla", c200c50, fc200c50, 3, (RNQ, RNQ, RNQ), vummla, vummla),
+ TUF ("vusmmla", ca00c40, fca00c40, 3, (RNQ, RNQ, RNQ), vsmmla, vsmmla),
+ TUF ("vusdot", c800d00, fc800d00, 3, (RNDQ, RNDQ, RNDQ_RNSC), vusdot, vusdot),
+ TUF ("vsudot", c800d10, fc800d10, 3, (RNDQ, RNDQ, RNSC), vsudot, vsudot),
};
#undef ARM_VARIANT
#undef THUMB_VARIANT
/* PR 21809: Do not set a mapping state for debug sections
- it just confuses other tools. */
- if (bfd_get_section_flags (NULL, now_seg) & SEC_DEBUGGING)
+ if (bfd_section_flags (now_seg) & SEC_DEBUGGING)
return;
frag_thumb_mode = fragP->tc_frag_data.thumb_mode ^ MODE_RECORDED;
break;
case BFD_RELOC_THUMB_PCREL_BRANCH9: /* Conditional branch. */
- if ((value & ~0xff) && ((value & ~0xff) != ~0xff))
+ if (out_of_range_p (value, 8))
as_bad_where (fixP->fx_file, fixP->fx_line, BAD_RANGE);
if (fixP->fx_done || !seg->use_rela_p)
break;
case BFD_RELOC_THUMB_PCREL_BRANCH12: /* Unconditional branch. */
- if ((value & ~0x7ff) && ((value & ~0x7ff) != ~0x7ff))
+ if (out_of_range_p (value, 11))
as_bad_where (fixP->fx_file, fixP->fx_line, BAD_RANGE);
if (fixP->fx_done || !seg->use_rela_p)
}
break;
+ /* This relocation is misnamed, it should be BRANCH21. */
case BFD_RELOC_THUMB_PCREL_BRANCH20:
if (fixP->fx_addsy
&& (S_GET_SEGMENT (fixP->fx_addsy) == seg)
/* Force a relocation for a branch 20 bits wide. */
fixP->fx_done = 0;
}
- if ((value & ~0x1fffff) && ((value & ~0x0fffff) != ~0x0fffff))
+ if (out_of_range_p (value, 20))
as_bad_where (fixP->fx_file, fixP->fx_line,
_("conditional branch out of range"));
fixP->fx_r_type = BFD_RELOC_THUMB_PCREL_BRANCH23;
#endif
- if ((value & ~0x3fffff) && ((value & ~0x3fffff) != ~0x3fffff))
+ if (out_of_range_p (value, 22))
{
if (!(ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_v6t2)))
as_bad_where (fixP->fx_file, fixP->fx_line, BAD_RANGE);
- else if ((value & ~0x1ffffff)
- && ((value & ~0x1ffffff) != ~0x1ffffff))
+ else if (out_of_range_p (value, 24))
as_bad_where (fixP->fx_file, fixP->fx_line,
_("Thumb2 branch out of range"));
}
break;
case BFD_RELOC_THUMB_PCREL_BRANCH25:
- if ((value & ~0x0ffffff) && ((value & ~0x0ffffff) != ~0x0ffffff))
+ if (out_of_range_p (value, 24))
as_bad_where (fixP->fx_file, fixP->fx_line, BAD_RANGE);
if (fixP->fx_done || !seg->use_rela_p)
if (sec != NULL)
{
- bfd_set_section_flags
- (stdoutput, sec, SEC_READONLY | SEC_DEBUGGING /* | SEC_HAS_CONTENTS */);
- bfd_set_section_size (stdoutput, sec, 0);
+ bfd_set_section_flags (sec, SEC_READONLY | SEC_DEBUGGING);
+ bfd_set_section_size (sec, 0);
bfd_set_section_contents (stdoutput, sec, NULL, 0, 0);
}
}
{"mwarn-deprecated", NULL, &warn_on_deprecated, 1, NULL},
{"mno-warn-deprecated", N_("do not warn on use of deprecated feature"),
&warn_on_deprecated, 0, NULL},
+
+ {"mwarn-restrict-it", N_("warn about performance deprecated IT instructions"
+ " in ARMv8-A and ARMv8-R"), &warn_on_restrict_it, 1, NULL},
+ {"mno-warn-restrict-it", NULL, &warn_on_restrict_it, 0, NULL},
+
{"mwarn-syms", N_("warn about symbols that match instruction names [default]"), (int *) (& flag_warn_syms), TRUE, NULL},
{"mno-warn-syms", N_("disable warnings about symobls that match instructions"), (int *) (& flag_warn_syms), FALSE, NULL},
{NULL, NULL, NULL, 0, NULL}
ARM_ARCH_NONE,
FPU_ARCH_NEON_VFP_V4),
ARM_CPU_OPT ("cortex-a32", "Cortex-A32", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("cortex-a35", "Cortex-A35", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("cortex-a53", "Cortex-A53", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("cortex-a55", "Cortex-A55", ARM_ARCH_V8_2A,
ARM_FEATURE_CORE_HIGH (ARM_EXT2_FP16_INST),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_DOTPROD),
ARM_CPU_OPT ("cortex-a57", "Cortex-A57", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("cortex-a72", "Cortex-A72", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("cortex-a73", "Cortex-A73", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("cortex-a75", "Cortex-A75", ARM_ARCH_V8_2A,
ARM_FEATURE_CORE_HIGH (ARM_EXT2_FP16_INST),
ARM_CPU_OPT ("cortex-a76", "Cortex-A76", ARM_ARCH_V8_2A,
ARM_FEATURE_CORE_HIGH (ARM_EXT2_FP16_INST),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_DOTPROD),
+ ARM_CPU_OPT ("cortex-a76ae", "Cortex-A76AE", ARM_ARCH_V8_2A,
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_FP16_INST),
+ FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_DOTPROD),
+ ARM_CPU_OPT ("cortex-a77", "Cortex-A77", ARM_ARCH_V8_2A,
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_FP16_INST),
+ FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_DOTPROD),
ARM_CPU_OPT ("ares", "Ares", ARM_ARCH_V8_2A,
ARM_FEATURE_CORE_HIGH (ARM_EXT2_FP16_INST),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_DOTPROD),
ARM_FEATURE_CORE_LOW (ARM_EXT_ADIV),
FPU_ARCH_VFP_V3D16),
ARM_CPU_OPT ("cortex-r52", "Cortex-R52", ARM_ARCH_V8R,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_NEON_VFP_ARMV8),
+ ARM_CPU_OPT ("cortex-m35p", "Cortex-M35P", ARM_ARCH_V8M_MAIN,
+ ARM_FEATURE_CORE_LOW (ARM_EXT_V5ExP | ARM_EXT_V6_DSP),
+ FPU_NONE),
ARM_CPU_OPT ("cortex-m33", "Cortex-M33", ARM_ARCH_V8M_MAIN,
ARM_FEATURE_CORE_LOW (ARM_EXT_V5ExP | ARM_EXT_V6_DSP),
FPU_NONE),
ARM_ARCH_NONE,
FPU_NONE),
ARM_CPU_OPT ("exynos-m1", "Samsung Exynos M1", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("neoverse-n1", "Neoverse N1", ARM_ARCH_V8_2A,
ARM_FEATURE_CORE_HIGH (ARM_EXT2_FP16_INST),
ARM_ARCH_NONE,
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
ARM_CPU_OPT ("xgene2", "APM X-Gene 2", ARM_ARCH_V8A,
- ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC),
FPU_ARCH_CRYPTO_NEON_VFP_ARMV8),
{ NULL, 0, ARM_ARCH_NONE, ARM_ARCH_NONE, ARM_ARCH_NONE, NULL }
static const struct arm_ext_table armv8a_ext_table[] =
{
- ARM_ADD ("crc", ARCH_CRC_ARMV8),
+ ARM_ADD ("crc", ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC)),
ARM_ADD ("simd", FPU_ARCH_NEON_VFP_ARMV8),
ARM_EXT ("crypto", FPU_ARCH_CRYPTO_NEON_VFP_ARMV8,
ARM_FEATURE_COPROC (FPU_CRYPTO_ARMV8)),
ARM_ADD ("simd", FPU_ARCH_NEON_VFP_ARMV8_1),
ARM_ADD ("fp16", FPU_ARCH_NEON_VFP_ARMV8_2_FP16),
ARM_ADD ("fp16fml", FPU_ARCH_NEON_VFP_ARMV8_2_FP16FML),
+ ARM_ADD ("bf16", ARM_FEATURE_CORE_HIGH (ARM_EXT2_BF16)),
+ ARM_ADD ("i8mm", ARM_FEATURE_CORE_HIGH (ARM_EXT2_I8MM)),
ARM_EXT ("crypto", FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_1,
ARM_FEATURE_COPROC (FPU_CRYPTO_ARMV8)),
ARM_ADD ("dotprod", FPU_ARCH_DOTPROD_NEON_VFP_ARMV8),
{
ARM_ADD ("simd", FPU_ARCH_DOTPROD_NEON_VFP_ARMV8),
ARM_ADD ("fp16", FPU_ARCH_NEON_VFP_ARMV8_4_FP16FML),
+ ARM_ADD ("bf16", ARM_FEATURE_CORE_HIGH (ARM_EXT2_BF16)),
+ ARM_ADD ("i8mm", ARM_FEATURE_CORE_HIGH (ARM_EXT2_I8MM)),
ARM_EXT ("crypto", FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_4,
ARM_FEATURE_COPROC (FPU_CRYPTO_ARMV8)),
{
ARM_ADD ("simd", FPU_ARCH_DOTPROD_NEON_VFP_ARMV8),
ARM_ADD ("fp16", FPU_ARCH_NEON_VFP_ARMV8_4_FP16FML),
+ ARM_ADD ("bf16", ARM_FEATURE_CORE_HIGH (ARM_EXT2_BF16)),
+ ARM_ADD ("i8mm", ARM_FEATURE_CORE_HIGH (ARM_EXT2_I8MM)),
ARM_EXT ("crypto", FPU_ARCH_CRYPTO_NEON_VFP_ARMV8_4,
ARM_FEATURE_COPROC (FPU_CRYPTO_ARMV8)),
{ NULL, 0, ARM_ARCH_NONE, ARM_ARCH_NONE }
};
+static const struct arm_ext_table armv86a_ext_table[] =
+{
+ ARM_ADD ("i8mm", ARM_FEATURE_CORE_HIGH (ARM_EXT2_I8MM)),
+ { NULL, 0, ARM_ARCH_NONE, ARM_ARCH_NONE }
+};
+
static const struct arm_ext_table armv8m_main_ext_table[] =
{
- ARM_EXT ("dsp", ARM_FEATURE_CORE_LOW (ARM_EXT_V5ExP | ARM_EXT_V6_DSP),
- ARM_FEATURE_CORE_LOW (ARM_EXT_V5ExP | ARM_EXT_V6_DSP)),
+ ARM_EXT ("dsp", ARM_FEATURE_CORE_LOW (ARM_AEXT_V8M_MAIN_DSP),
+ ARM_FEATURE_CORE_LOW (ARM_AEXT_V8M_MAIN_DSP)),
ARM_EXT ("fp", FPU_ARCH_VFP_V5_SP_D16, ALL_FP),
ARM_ADD ("fp.dp", FPU_ARCH_VFP_V5D16),
{ NULL, 0, ARM_ARCH_NONE, ARM_ARCH_NONE }
};
+
static const struct arm_ext_table armv8_1m_main_ext_table[] =
{
- ARM_EXT ("dsp", ARM_FEATURE_CORE_LOW (ARM_EXT_V5ExP | ARM_EXT_V6_DSP),
- ARM_FEATURE_CORE_LOW (ARM_EXT_V5ExP | ARM_EXT_V6_DSP)),
+ ARM_EXT ("dsp", ARM_FEATURE_CORE_LOW (ARM_AEXT_V8M_MAIN_DSP),
+ ARM_FEATURE_CORE_LOW (ARM_AEXT_V8M_MAIN_DSP)),
ARM_EXT ("fp",
ARM_FEATURE (0, ARM_EXT2_FP16_INST,
FPU_VFP_V5_SP_D16 | FPU_VFP_EXT_FP16 | FPU_VFP_EXT_FMA),
ARM_ADD ("fp.dp",
ARM_FEATURE (0, ARM_EXT2_FP16_INST,
FPU_VFP_V5D16 | FPU_VFP_EXT_FP16 | FPU_VFP_EXT_FMA)),
- ARM_EXT ("mve", ARM_FEATURE_COPROC (FPU_MVE),
- ARM_FEATURE_COPROC (FPU_MVE | FPU_MVE_FP)),
+ ARM_EXT ("mve", ARM_FEATURE (ARM_AEXT_V8M_MAIN_DSP, ARM_EXT2_MVE, 0),
+ ARM_FEATURE_CORE_HIGH (ARM_EXT2_MVE | ARM_EXT2_MVE_FP)),
ARM_ADD ("mve.fp",
- ARM_FEATURE (0, ARM_EXT2_FP16_INST,
- FPU_MVE | FPU_MVE_FP | FPU_VFP_V5_SP_D16 |
- FPU_VFP_EXT_FP16 | FPU_VFP_EXT_FMA)),
+ ARM_FEATURE (ARM_AEXT_V8M_MAIN_DSP,
+ ARM_EXT2_FP16_INST | ARM_EXT2_MVE | ARM_EXT2_MVE_FP,
+ FPU_VFP_V5_SP_D16 | FPU_VFP_EXT_FP16 | FPU_VFP_EXT_FMA)),
{ NULL, 0, ARM_ARCH_NONE, ARM_ARCH_NONE }
};
static const struct arm_ext_table armv8r_ext_table[] =
{
- ARM_ADD ("crc", ARCH_CRC_ARMV8),
+ ARM_ADD ("crc", ARM_FEATURE_CORE_HIGH (ARM_EXT2_CRC)),
ARM_ADD ("simd", FPU_ARCH_NEON_VFP_ARMV8),
ARM_EXT ("crypto", FPU_ARCH_CRYPTO_NEON_VFP_ARMV8,
ARM_FEATURE_COPROC (FPU_CRYPTO_ARMV8)),
ARM_ARCH_OPT2 ("armv8-r", ARM_ARCH_V8R, FPU_ARCH_VFP, armv8r),
ARM_ARCH_OPT2 ("armv8.4-a", ARM_ARCH_V8_4A, FPU_ARCH_VFP, armv84a),
ARM_ARCH_OPT2 ("armv8.5-a", ARM_ARCH_V8_5A, FPU_ARCH_VFP, armv85a),
+ ARM_ARCH_OPT2 ("armv8.6-a", ARM_ARCH_V8_6A, FPU_ARCH_VFP, armv86a),
ARM_ARCH_OPT ("xscale", ARM_ARCH_XSCALE, FPU_ARCH_VFP),
ARM_ARCH_OPT ("iwmmxt", ARM_ARCH_IWMMXT, FPU_ARCH_VFP),
ARM_ARCH_OPT ("iwmmxt2", ARM_ARCH_IWMMXT2, FPU_ARCH_VFP),
use the context sensitive approach using arm_ext_table's. */
static const struct arm_option_extension_value_table arm_extensions[] =
{
- ARM_EXT_OPT ("crc", ARCH_CRC_ARMV8, ARM_FEATURE_COPROC (CRC_EXT_ARMV8),
+ ARM_EXT_OPT ("crc", ARM_FEATURE_CORE_HIGH(ARM_EXT2_CRC),
+ ARM_FEATURE_CORE_HIGH(ARM_EXT2_CRC),
ARM_FEATURE_CORE_LOW (ARM_EXT_V8)),
ARM_EXT_OPT ("crypto", FPU_ARCH_CRYPTO_NEON_VFP_ARMV8,
ARM_FEATURE_COPROC (FPU_CRYPTO_ARMV8),
return TRUE;
}
+static bfd_boolean
+arm_parse_fp16_opt (const char *str)
+{
+ if (strcasecmp (str, "ieee") == 0)
+ fp16_format = ARM_FP16_FORMAT_IEEE;
+ else if (strcasecmp (str, "alternative") == 0)
+ fp16_format = ARM_FP16_FORMAT_ALTERNATIVE;
+ else
+ {
+ as_bad (_("unrecognised float16 format \"%s\""), str);
+ return FALSE;
+ }
+
+ return TRUE;
+}
+
static bfd_boolean
arm_parse_cpu (const char *str)
{
march_ext_opt = XNEW (arm_feature_set);
*march_ext_opt = arm_arch_none;
march_fpu_opt = &opt->default_fpu;
+ selected_ctx_ext_table = opt->ext_table;
strcpy (selected_cpu_name, opt->name);
if (ext != NULL)
arm_parse_it_mode, NULL},
{"mccs", N_("\t\t\t TI CodeComposer Studio syntax compatibility mode"),
arm_ccs_mode, NULL},
+ {"mfp16-format=",
+ N_("[ieee|alternative]\n\
+ set the encoding for half precision floating point "
+ "numbers to IEEE\n\
+ or Arm alternative format."),
+ arm_parse_fp16_opt, NULL },
{NULL, NULL, 0, NULL}
};
{TAG_CPU_ARCH_V8, ARM_ARCH_V8_4A},
{TAG_CPU_ARCH_V8, ARM_ARCH_V8_5A},
{TAG_CPU_ARCH_V8_1M_MAIN, ARM_ARCH_V8_1M_MAIN},
- {-1, ARM_ARCH_NONE}
+ {TAG_CPU_ARCH_V8, ARM_ARCH_V8_6A},
+ {-1, ARM_ARCH_NONE}
};
/* Set an attribute if it has not already been set by the user. */
virt_sec |= 2;
if (virt_sec != 0)
aeabi_set_attribute_int (Tag_Virtualization_use, virt_sec);
+
+ if (fp16_format != ARM_FP16_FORMAT_DEFAULT)
+ aeabi_set_attribute_int (Tag_ABI_FP_16bit_format, fp16_format);
}
/* Post relaxation hook. Recompute ARM attributes now that relaxation is
if (streq (opt->name, name))
{
selected_arch = opt->value;
+ selected_ctx_ext_table = opt->ext_table;
selected_ext = arm_arch_none;
selected_cpu = selected_arch;
strcpy (selected_cpu_name, opt->name);
name += 2;
}
+ /* Check the context specific extension table */
+ if (selected_ctx_ext_table)
+ {
+ const struct arm_ext_table * ext_opt;
+ for (ext_opt = selected_ctx_ext_table; ext_opt->name != NULL; ext_opt++)
+ {
+ if (streq (ext_opt->name, name))
+ {
+ if (adding_value)
+ {
+ if (ARM_FEATURE_ZERO (ext_opt->merge))
+ /* TODO: Option not supported. When we remove the
+ legacy table this case should error out. */
+ continue;
+ ARM_MERGE_FEATURE_SETS (selected_ext, selected_ext,
+ ext_opt->merge);
+ }
+ else
+ ARM_CLEAR_FEATURE (selected_ext, selected_ext, ext_opt->clear);
+
+ ARM_MERGE_FEATURE_SETS (selected_cpu, selected_arch, selected_ext);
+ ARM_MERGE_FEATURE_SETS (cpu_variant, selected_cpu, selected_fpu);
+ *input_line_pointer = saved_char;
+ demand_empty_rest_of_line ();
+ return;
+ }
+ }
+ }
+
for (opt = arm_extensions; opt->name != NULL; opt++)
if (streq (opt->name, name))
{
if (streq (opt->name, name))
{
selected_fpu = opt->value;
+ ARM_CLEAR_FEATURE (selected_cpu, selected_cpu, fpu_any);
#ifndef CPU_DEFAULT
if (no_cpu_selected ())
ARM_MERGE_FEATURE_SETS (cpu_variant, arm_arch_any, selected_fpu);
projects / thirdparty / binutils-gdb.git / blobdiff