/* Print SPARC instructions.
- Copyright 1989, 1991, 1992, 1993 Free Software Foundation, Inc.
+ Copyright 1989, 1991, 1992, 1993, 1995 Free Software Foundation, Inc.
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
-Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
+Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
+#include "ansidecl.h"
#include "opcode/sparc.h"
#include "dis-asm.h"
+#include "libiberty.h"
#include <string.h>
+/* For faster lookup, after insns are sorted they are hashed. */
+/* ??? I think there is room for even more improvement. */
+
+#define HASH_SIZE 256
+/* It is important that we only look at insn code bits as that is how the
+ opcode table is hashed. OPCODE_BITS is a table of valid bits for each
+ of the main types (0,1,2,3). */
+static int opcode_bits[4] = { 0x01c00000, 0x0, 0x01f80000, 0x01f80000 };
+#define HASH_INSN(INSN) \
+ ((((INSN) >> 24) & 0xc0) | (((INSN) & opcode_bits[((INSN) >> 30) & 3]) >> 19))
+struct opcode_hash {
+ struct opcode_hash *next;
+ struct sparc_opcode *opcode;
+};
+static struct opcode_hash *opcode_hash_table[HASH_SIZE];
+static void build_hash_table ();
+
+/* Sign-extend a value which is N bits long. */
+#define SEX(value, bits) \
+ ((((int)(value)) << ((8 * sizeof (int)) - bits)) \
+ >> ((8 * sizeof (int)) - bits) )
+
static char *reg_names[] =
- { "g0", "g1", "g2", "g3", "g4", "g5", "g6", "g7",
+{ "g0", "g1", "g2", "g3", "g4", "g5", "g6", "g7",
"o0", "o1", "o2", "o3", "o4", "o5", "sp", "o7",
"l0", "l1", "l2", "l3", "l4", "l5", "l6", "l7",
"i0", "i1", "i2", "i3", "i4", "i5", "fp", "i7",
"f8", "f9", "f10", "f11", "f12", "f13", "f14", "f15",
"f16", "f17", "f18", "f19", "f20", "f21", "f22", "f23",
"f24", "f25", "f26", "f27", "f28", "f29", "f30", "f31",
- "y", "psr", "wim", "tbr", "pc", "npc", "fpsr", "cpsr" };
+#ifndef NO_V9
+ "f32", "f33", "f34", "f35", "f36", "f37", "f38", "f39",
+ "f40", "f41", "f42", "f43", "f44", "f45", "f46", "f47",
+ "f48", "f49", "f50", "f51", "f52", "f53", "f54", "f55",
+ "f56", "f57", "f58", "f59", "f60", "f61", "f62", "f63",
+/* psr, wim, tbr, fpsr, cpsr are v8 only. */
+#endif
+ "y", "psr", "wim", "tbr", "pc", "npc", "fpsr", "cpsr"
+};
#define freg_names (®_names[4 * 8])
-/* FIXME--need to deal with byte order (probably using masking and
- shifting rather than bitfields is easiest). */
+#ifndef NO_V9
+/* These are ordered according to there register number in
+ rdpr and wrpr insns. */
+static char *v9_priv_reg_names[] =
+{
+ "tpc", "tnpc", "tstate", "tt", "tick", "tba", "pstate", "tl",
+ "pil", "cwp", "cansave", "canrestore", "cleanwin", "otherwin",
+ "wstate", "fq"
+ /* "ver" - special cased */
+};
+#endif
+
+/* Macros used to extract instruction fields. Not all fields have
+ macros defined here, only those which are actually used. */
+
+#define X_RD(i) (((i) >> 25) & 0x1f)
+#define X_RS1(i) (((i) >> 14) & 0x1f)
+#define X_LDST_I(i) (((i) >> 13) & 1)
+#define X_ASI(i) (((i) >> 5) & 0xff)
+#define X_RS2(i) (((i) >> 0) & 0x1f)
+#define X_IMM13(i) (((i) >> 0) & 0x1fff)
+#define X_DISP22(i) (((i) >> 0) & 0x3fffff)
+#define X_IMM22(i) X_DISP22 (i)
+#define X_DISP30(i) (((i) >> 0) & 0x3fffffff)
-union sparc_insn
- {
- unsigned long int code;
- struct
- {
- unsigned int anop:2;
-#define op ldst.anop
- unsigned int anrd:5;
-#define rd ldst.anrd
- unsigned int op3:6;
- unsigned int anrs1:5;
-#define rs1 ldst.anrs1
- unsigned int i:1;
- unsigned int anasi:8;
-#define asi ldst.anasi
- unsigned int anrs2:5;
-#define rs2 ldst.anrs2
-#define shcnt rs2
- } ldst;
- struct
- {
- unsigned int anop:2, anrd:5, op3:6, anrs1:5, i:1;
- unsigned int IMM13:13;
-#define imm13 IMM13.IMM13
- } IMM13;
- struct
- {
- unsigned int anop:2;
- unsigned int a:1;
- unsigned int cond:4;
- unsigned int op2:3;
- unsigned int DISP22:22;
-#define disp22 branch.DISP22
- } branch;
#ifndef NO_V9
- struct
- {
- unsigned int _OP:2, _RD:5, op3:6, _RS1:5;
- unsigned int DISP14:14;
-#define disp14 DISP14.DISP14
- } DISP14;
- struct
- {
- unsigned int _OP:2;
- unsigned int a:1;
- unsigned int cond:4;
- unsigned int op2:3;
- unsigned int p:1;
- unsigned int DISP21:21;
-#define disp21 branch2.DISP21
- } branch2;
-#endif /* NO_V9 */
-
-#define imm22 disp22
- struct
- {
- unsigned int anop:2;
- unsigned int adisp30:30;
-#define disp30 call.adisp30
- } call;
- };
+#define X_DISP16(i) (((((i) >> 20) & 3) << 14) | (((i) >> 0) & 0x3fff))
+#endif
+
+/* Here is the union which was used to extract instruction fields
+ before the shift and mask macros were written.
+
+ union sparc_insn
+ {
+ unsigned long int code;
+ struct
+ {
+ unsigned int anop:2;
+ #define op ldst.anop
+ unsigned int anrd:5;
+ #define rd ldst.anrd
+ unsigned int op3:6;
+ unsigned int anrs1:5;
+ #define rs1 ldst.anrs1
+ unsigned int i:1;
+ unsigned int anasi:8;
+ #define asi ldst.anasi
+ unsigned int anrs2:5;
+ #define rs2 ldst.anrs2
+ #define shcnt rs2
+ } ldst;
+ struct
+ {
+ unsigned int anop:2, anrd:5, op3:6, anrs1:5, i:1;
+ unsigned int IMM13:13;
+ #define imm13 IMM13.IMM13
+ } IMM13;
+ struct
+ {
+ unsigned int anop:2;
+ unsigned int a:1;
+ unsigned int cond:4;
+ unsigned int op2:3;
+ unsigned int DISP22:22;
+ #define disp22 branch.DISP22
+ #define imm22 disp22
+ } branch;
+ #ifndef NO_V9
+ struct
+ {
+ unsigned int anop:2;
+ unsigned int a:1;
+ unsigned int z:1;
+ unsigned int rcond:3;
+ unsigned int op2:3;
+ unsigned int DISP16HI:2;
+ unsigned int p:1;
+ unsigned int _rs1:5;
+ unsigned int DISP16LO:14;
+ } branch16;
+ #endif
+ struct
+ {
+ unsigned int anop:2;
+ unsigned int adisp30:30;
+ #define disp30 call.adisp30
+ } call;
+ };
+
+ */
/* Nonzero if INSN is the opcode for a delayed branch. */
static int
is_delayed_branch (insn)
- union sparc_insn insn;
+ unsigned long insn;
{
- unsigned int i;
+ struct opcode_hash *op;
- for (i = 0; i < NUMOPCODES; ++i)
+ for (op = opcode_hash_table[HASH_INSN (insn)]; op; op = op->next)
{
- const struct sparc_opcode *opcode = &sparc_opcodes[i];
- if ((opcode->match & insn.code) == opcode->match
- && (opcode->lose & insn.code) == 0)
+ CONST struct sparc_opcode *opcode = op->opcode;
+ if ((opcode->match & insn) == opcode->match
+ && (opcode->lose & insn) == 0)
return (opcode->flags & F_DELAYED);
}
return 0;
}
-static int opcodes_sorted = 0;
-extern void qsort ();
+/* Nonzero of opcode table has been initialized. */
+static int opcodes_initialized = 0;
-/* Print one instruction from MEMADDR on STREAM.
+/* Nonzero of the current architecture is sparc64.
+ This is kept in a global because compare_opcodes uses it. */
+static int sparc64_p;
+
+/* extern void qsort (); */
+static int compare_opcodes ();
+
+/* Print one instruction from MEMADDR on INFO->STREAM.
We suffix the instruction with a comment that gives the absolute
address involved, as well as its symbolic form, if the instruction
is preceded by a findable `sethi' and it either adds an immediate
displacement to that register, or it is an `add' or `or' instruction
on that register. */
-int
-print_insn_sparc (memaddr, info)
+
+static int
+print_insn (memaddr, info)
bfd_vma memaddr;
disassemble_info *info;
{
FILE *stream = info->stream;
- union sparc_insn insn;
-
+ bfd_byte buffer[4];
+ unsigned long insn;
register unsigned int i;
+ register struct opcode_hash *op;
- if (!opcodes_sorted)
+ if (!opcodes_initialized)
{
- static int compare_opcodes ();
qsort ((char *) sparc_opcodes, NUMOPCODES,
sizeof (sparc_opcodes[0]), compare_opcodes);
- opcodes_sorted = 1;
+ build_hash_table (sparc_opcodes, opcode_hash_table, NUMOPCODES);
+ opcodes_initialized = 1;
}
{
int status =
- (*info->read_memory_func) (memaddr, (char *) &insn, sizeof (insn), info);
+ (*info->read_memory_func) (memaddr, buffer, sizeof (buffer), info);
if (status != 0)
{
(*info->memory_error_func) (status, memaddr, info);
}
}
- for (i = 0; i < NUMOPCODES; ++i)
+ insn = bfd_getb32 (buffer);
+
+ info->insn_info_valid = 1; /* We do return this info */
+ info->insn_type = dis_nonbranch; /* Assume non branch insn */
+ info->branch_delay_insns = 0; /* Assume no delay */
+ info->target = 0; /* Assume no target known */
+
+ for (op = opcode_hash_table[HASH_INSN (insn)]; op; op = op->next)
{
- const struct sparc_opcode *opcode = &sparc_opcodes[i];
- if ((opcode->match & insn.code) == opcode->match
- && (opcode->lose & insn.code) == 0)
+ CONST struct sparc_opcode *opcode = op->opcode;
+
+ if ((opcode->match & insn) == opcode->match
+ && (opcode->lose & insn) == 0)
{
/* Nonzero means that we have found an instruction which has
the effect of adding or or'ing the imm13 field to rs1. */
field of the opcode table. */
int found_plus = 0;
+ /* Nonzero means we have an annulled branch. */
+ int is_annulled = 0;
+
/* Do we have an `add' or `or' instruction where rs1 is the same
as rsd, and which has the i bit set? */
if ((opcode->match == 0x80102000 || opcode->match == 0x80002000)
/* (or) (add) */
- && insn.rs1 == insn.rd)
+ && X_RS1 (insn) == X_RD (insn))
imm_added_to_rs1 = 1;
- if (insn.rs1 != insn.rd
+ if (X_RS1 (insn) != X_RD (insn)
&& strchr (opcode->args, 'r') != 0)
/* Can't do simple format if source and dest are different. */
continue;
(*info->fprintf_func) (stream, opcode->name);
{
- register const char *s;
+ register CONST char *s;
if (opcode->args[0] != ',')
(*info->fprintf_func) (stream, " ");
switch (*s) {
case 'a':
(*info->fprintf_func) (stream, "a");
+ is_annulled = 1;
++s;
continue;
#ifndef NO_V9
(*info->fprintf_func) (stream, "pt");
++s;
continue;
-#endif /* NO_V9 */
+#endif /* NO_V9 */
default:
break;
#define reg(n) (*info->fprintf_func) (stream, "%%%s", reg_names[n])
case '1':
case 'r':
- reg (insn.rs1);
+ reg (X_RS1 (insn));
break;
case '2':
- reg (insn.rs2);
+ reg (X_RS2 (insn));
break;
case 'd':
- reg (insn.rd);
+ reg (X_RD (insn));
break;
#undef reg
-#define freg(n) (*info->fprintf_func) (stream, "%%%s", freg_names[n])
+#define freg(n) (*info->fprintf_func) (stream, "%%%s", freg_names[n])
+#define fregx(n) (*info->fprintf_func) (stream, "%%%s", freg_names[((n) & ~1) | (((n) & 1) << 5)])
case 'e':
+ freg (X_RS1 (insn));
+ break;
case 'v': /* double/even */
case 'V': /* quad/multiple of 4 */
- freg (insn.rs1);
+ fregx (X_RS1 (insn));
break;
case 'f':
+ freg (X_RS2 (insn));
+ break;
case 'B': /* double/even */
case 'R': /* quad/multiple of 4 */
- freg (insn.rs2);
+ fregx (X_RS2 (insn));
break;
case 'g':
+ freg (X_RD (insn));
+ break;
case 'H': /* double/even */
case 'J': /* quad/multiple of 4 */
- freg (insn.rd);
+ fregx (X_RD (insn));
break;
#undef freg
+#undef fregx
#define creg(n) (*info->fprintf_func) (stream, "%%c%u", (unsigned int) (n))
case 'b':
- creg (insn.rs1);
+ creg (X_RS1 (insn));
break;
case 'c':
- creg (insn.rs2);
+ creg (X_RS2 (insn));
break;
case 'D':
- creg (insn.rd);
+ creg (X_RD (insn));
break;
#undef creg
case 'h':
(*info->fprintf_func) (stream, "%%hi(%#x)",
- (int) insn.imm22 << 10);
+ (0xFFFFFFFF
+ & ((int) X_IMM22 (insn) << 10)));
break;
case 'i':
{
- /* We cannot trust the compiler to sign-extend
- when extracting the bitfield, hence the shifts. */
- int imm = ((int) insn.imm13 << 19) >> 19;
+ int imm = SEX (X_IMM13 (insn), 13);
/* Check to see whether we have a 1+i, and take
note of that fact.
case 'I': /* 11 bit immediate. */
case 'j': /* 10 bit immediate. */
{
- /* We cannot trust the compiler to sign-extend
- when extracting the bitfield, hence the shifts. */
int imm;
if (*s == 'I')
- imm = ((int) insn.imm13 << 21) >> 21;
+ imm = SEX (X_IMM13 (insn), 11);
else
- imm = ((int) insn.imm13 << 22) >> 22;
+ imm = SEX (X_IMM13 (insn), 10);
/* Check to see whether we have a 1+i, and take
note of that fact.
}
break;
-
-
-
case 'k':
- (*info->print_address_func)
- ((bfd_vma) (memaddr
- + (((int) insn.disp14 << 18) >> 18) * 4),
- info);
+ info->target = memaddr + (SEX (X_DISP16 (insn), 16)) * 4;
+ (*info->print_address_func) (info->target, info);
break;
case 'G':
- (*info->print_address_func)
- ((bfd_vma) (memaddr
- /* We use only 19 of the 21 bits. */
- + (((int) insn.disp21 << 13) >> 13) * 4),
- info);
+ info->target = memaddr + (SEX (X_DISP22 (insn), 19)) * 4;
+ (*info->print_address_func) (info->target, info);
break;
case '6':
case '7':
case '8':
case '9':
- (*info->fprintf_func) (stream, "fcc%c", *s - '6' + '0');
+ (*info->fprintf_func) (stream, "%%fcc%c", *s - '6' + '0');
break;
case 'z':
- (*info->fprintf_func) (stream, "icc");
+ (*info->fprintf_func) (stream, "%%icc");
break;
case 'Z':
- (*info->fprintf_func) (stream, "xcc");
+ (*info->fprintf_func) (stream, "%%xcc");
break;
case 'E':
case 's':
(*info->fprintf_func) (stream, "%%fprs");
break;
-#endif /* NO_V9 */
+
+ case 'o':
+ (*info->fprintf_func) (stream, "%%asi");
+ break;
+
+ case 'W':
+ (*info->fprintf_func) (stream, "%%tick");
+ break;
+
+ case 'P':
+ (*info->fprintf_func) (stream, "%%pc");
+ break;
+
+ case '?':
+ if (X_RS1 (insn) == 31)
+ (*info->fprintf_func) (stream, "%%ver");
+ else if ((unsigned) X_RS1 (insn) < 16)
+ (*info->fprintf_func) (stream, "%%%s",
+ v9_priv_reg_names[X_RS1 (insn)]);
+ else
+ (*info->fprintf_func) (stream, "%%reserved");
+ break;
+
+ case '!':
+ if ((unsigned) X_RD (insn) < 15)
+ (*info->fprintf_func) (stream, "%%%s",
+ v9_priv_reg_names[X_RD (insn)]);
+ else
+ (*info->fprintf_func) (stream, "%%reserved");
+ break;
+ break;
+#endif /* NO_V9 */
case 'M':
- (*info->fprintf_func) (stream, "%%asr%d", insn.rs1);
+ (*info->fprintf_func) (stream, "%%asr%d", X_RS1 (insn));
break;
case 'm':
- (*info->fprintf_func) (stream, "%%asr%d", insn.rd);
+ (*info->fprintf_func) (stream, "%%asr%d", X_RD (insn));
break;
case 'L':
- (*info->print_address_func)
- ((bfd_vma) memaddr + insn.disp30 * 4,
- info);
+ info->target = memaddr + X_DISP30 (insn) * 4;
+ (*info->print_address_func) (info->target, info);
+ break;
+
+ case 'n':
+ (*info->fprintf_func)
+ (stream, "%#x", (SEX (X_DISP22 (insn), 22)));
break;
case 'l':
- if ((insn.code >> 22) == 0)
- /* Special case for `unimp'. Don't try to turn
- it's operand into a function offset. */
- (*info->fprintf_func)
- (stream, "%#x",
- (int) (((int) insn.disp22 << 10) >> 10));
- else
- /* We cannot trust the compiler to sign-extend
- when extracting the bitfield, hence the shifts. */
- (*info->print_address_func)
- ((bfd_vma) (memaddr
- + (((int) insn.disp22 << 10) >> 10) * 4),
- info);
+ info->target = memaddr + (SEX (X_DISP22 (insn), 22)) * 4;
+ (*info->print_address_func) (info->target, info);
break;
case 'A':
- (*info->fprintf_func) (stream, "(%d)", (int) insn.asi);
+ (*info->fprintf_func) (stream, "(%d)", X_ASI (insn));
break;
case 'C':
(*info->fprintf_func) (stream, "%%wim");
break;
+ case 'x':
+ (*info->fprintf_func) (stream, "%d",
+ ((X_LDST_I (insn) << 8)
+ + X_ASI (insn)));
+ break;
+
case 'y':
(*info->fprintf_func) (stream, "%%y");
break;
and its symbolic value. */
if (imm_added_to_rs1)
{
- union sparc_insn prev_insn;
+ unsigned long prev_insn;
int errcode;
errcode =
(*info->read_memory_func)
- (memaddr - 4,
- (char *)&prev_insn, sizeof (prev_insn), info);
+ (memaddr - 4, buffer, sizeof (buffer), info);
+ prev_insn = bfd_getb32 (buffer);
if (errcode == 0)
{
*/
if (is_delayed_branch (prev_insn))
- errcode = (*info->read_memory_func)
- (memaddr - 8, (char *)&prev_insn, sizeof (prev_insn),
- info);
+ {
+ errcode = (*info->read_memory_func)
+ (memaddr - 8, buffer, sizeof (buffer), info);
+ prev_insn = bfd_getb32 (buffer);
+ }
}
/* If there was a problem reading memory, then assume
if (errcode == 0)
{
/* Is it sethi to the same register? */
- if ((prev_insn.code & 0xc1c00000) == 0x01000000
- && prev_insn.rd == insn.rs1)
+ if ((prev_insn & 0xc1c00000) == 0x01000000
+ && X_RD (prev_insn) == X_RS1 (insn))
{
(*info->fprintf_func) (stream, "\t! ");
- /* We cannot trust the compiler to sign-extend
- when extracting the bitfield, hence the shifts. */
- (*info->print_address_func)
- (((int) prev_insn.imm22 << 10)
- | (insn.imm13 << 19) >> 19,
- info);
+ info->target =
+ (0xFFFFFFFF & (int) X_IMM22 (prev_insn) << 10)
+ | SEX (X_IMM13 (insn), 13);
+ (*info->print_address_func) (info->target, info);
+ info->insn_type = dis_dref;
+ info->data_size = 4; /* FIXME!!! */
}
}
}
- return sizeof (insn);
+ if (opcode->flags & (F_UNBR|F_CONDBR|F_JSR))
+ {
+ /* FIXME -- check is_annulled flag */
+ if (opcode->flags & F_UNBR)
+ info->insn_type = dis_branch;
+ if (opcode->flags & F_CONDBR)
+ info->insn_type = dis_condbranch;
+ if (opcode->flags & F_JSR)
+ info->insn_type = dis_jsr;
+ if (opcode->flags & F_DELAYED)
+ info->branch_delay_insns = 1;
+ }
+
+ return sizeof (buffer);
}
}
- (*info->fprintf_func) (stream, "%#8x", insn.code);
- return sizeof (insn);
+ info->insn_type = dis_noninsn; /* Mark as non-valid instruction */
+ (*info->fprintf_func) (stream, "%#8x", insn);
+ return sizeof (buffer);
}
/* Compare opcodes A and B. */
lose1 = op1->lose;
}
+ /* If the current architecture isn't sparc64, move v9 insns to the end.
+ Only do this when one isn't v9 and one is. If both are v9 we still
+ need to properly sort them.
+ This must be done before checking match and lose. */
+ if (!sparc64_p
+ && (op0->architecture == v9) != (op1->architecture == v9))
+ return (op0->architecture == v9) - (op1->architecture == v9);
+
+ /* If the current architecture is sparc64, move non-v9 insns to the end.
+ This must be done before checking match and lose. */
+ if (sparc64_p
+ && (op0->flags & F_NOTV9) != (op1->flags & F_NOTV9))
+ return (op0->flags & F_NOTV9) - (op1->flags & F_NOTV9);
+
/* Because the bits that are variable in one opcode are constant in
another, it is important to order the opcodes in the right order. */
for (i = 0; i < 32; ++i)
written, so just say there are equal. */
return 0;
}
+
+/* Build a hash table from the opcode table. */
+
+static void
+build_hash_table (table, hash_table, num_opcodes)
+ struct sparc_opcode *table;
+ struct opcode_hash **hash_table;
+ int num_opcodes;
+{
+ int i;
+ int hash_count[HASH_SIZE];
+
+ /* Start at the end of the table and work backwards so that each
+ chain is sorted. */
+ /* ??? Do we really need to sort them now? */
+
+ memset (hash_table, 0, HASH_SIZE * sizeof (hash_table[0]));
+ memset (hash_count, 0, HASH_SIZE * sizeof (hash_count[0]));
+ for (i = num_opcodes - 1; i >= 0; --i)
+ {
+ int hash = HASH_INSN (sparc_opcodes[i].match);
+ struct opcode_hash *h = (struct opcode_hash *) xmalloc (sizeof (struct opcode_hash));
+ h->next = hash_table[hash];
+ h->opcode = &sparc_opcodes[i];
+ hash_table[hash] = h;
+ ++hash_count[hash];
+ }
+
+#if 0 /* for debugging */
+ {
+ int min_count = num_opcodes, max_count = 0;
+ int total;
+
+ for (i = 0; i < HASH_SIZE; ++i)
+ {
+ if (hash_count[i] < min_count)
+ min_count = hash_count[i];
+ if (hash_count[i] > max_count)
+ max_count = hash_count[i];
+ total += hash_count[i];
+ }
+
+ printf ("Opcode hash table stats: min %d, max %d, ave %f\n",
+ min_count, max_count, (double) total / HASH_SIZE);
+ }
+#endif
+}
+
+int
+print_insn_sparc (memaddr, info)
+ bfd_vma memaddr;
+ disassemble_info *info;
+{
+ /* It's not clear that we'll ever switch cpus in a running program.
+ It could theoretically happen in gdb so we handle it.
+ ??? There is currently a memory leak but it's not worth the trouble. */
+ if (sparc64_p)
+ opcodes_initialized = 0;
+ sparc64_p = 0;
+ return print_insn (memaddr, info);
+}
+
+int
+print_insn_sparc64 (memaddr, info)
+ bfd_vma memaddr;
+ disassemble_info *info;
+{
+ /* It's not clear that we'll ever switch cpus in a running program.
+ It could theoretically happen in gdb so we handle it.
+ ??? There is currently a memory leak but it's not worth the trouble. */
+ if (!sparc64_p)
+ opcodes_initialized = 0;
+ sparc64_p = 1;
+ return print_insn (memaddr, info);
+}
-/* to sanitize this file, grep -v v9 < sparc.h > clean-sparc.h */
-
/* Table of opcodes for the sparc.
- Copyright 1989, 1991, 1992 Free Software Foundation, Inc.
+ Copyright 1989, 1991, 1992, 1995 Free Software Foundation, Inc.
This file is part of the BFD library.
You should have received a copy of the GNU General Public License
along with this software; see the file COPYING. If not, write to
-the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA. */
/* FIXME-someday: perhaps the ,a's and such should be embedded in the
instruction's name rather than the args. This would make gas faster, pinsn
#include "ansidecl.h"
#include "opcode/sparc.h"
-CONST char *architecture_pname[] = {
+const char *architecture_pname[] = {
"v6",
"v7",
"v8",
{ "ld", F3(3, 0x21, 1), F3(~3, ~0x21, ~1)|RS1_G0|RD(~0),"[i],F", 0, v6 },
{ "ld", F3(3, 0x21, 1), F3(~3, ~0x21, ~1)|SIMM13(~0)|RD(~0),"[1],F", 0, v6 }, /* ld [rs1+0],d */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "ld", F3(3, 0x30, 0), F3(~3, ~0x30, ~0), "[1+2],D", F_ALIAS, v6 },
-{ "ld", F3(3, 0x30, 0), F3(~3, ~0x30, ~0)|RS2_G0, "[1],D", F_ALIAS, v6 }, /* ld [rs1+%g0],d */
-{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1), "[1+i],D", F_ALIAS, v6 },
-{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1), "[i+1],D", F_ALIAS, v6 },
-{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1)|RS1_G0, "[i],D", F_ALIAS, v6 },
-{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1)|SIMM13(~0), "[1],D", F_ALIAS, v6 }, /* ld [rs1+0],d */
-{ "ld", F3(3, 0x31, 0), F3(~3, ~0x31, ~0), "[1+2],C", 0, v6 },
-{ "ld", F3(3, 0x31, 0), F3(~3, ~0x31, ~0)|RS2_G0, "[1],C", 0, v6 }, /* ld [rs1+%g0],d */
-{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1), "[1+i],C", 0, v6 },
-{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1), "[i+1],C", 0, v6 },
-{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1)|RS1_G0, "[i],C", 0, v6 },
-{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1)|SIMM13(~0), "[1],C", 0, v6 }, /* ld [rs1+0],d */
+{ "ld", F3(3, 0x30, 0), F3(~3, ~0x30, ~0), "[1+2],D", F_NOTV9, v6 },
+{ "ld", F3(3, 0x30, 0), F3(~3, ~0x30, ~0)|RS2_G0, "[1],D", F_NOTV9, v6 }, /* ld [rs1+%g0],d */
+{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1), "[1+i],D", F_NOTV9, v6 },
+{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1), "[i+1],D", F_NOTV9, v6 },
+{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1)|RS1_G0, "[i],D", F_NOTV9, v6 },
+{ "ld", F3(3, 0x30, 1), F3(~3, ~0x30, ~1)|SIMM13(~0), "[1],D", F_NOTV9, v6 }, /* ld [rs1+0],d */
+{ "ld", F3(3, 0x31, 0), F3(~3, ~0x31, ~0), "[1+2],C", F_NOTV9, v6 },
+{ "ld", F3(3, 0x31, 0), F3(~3, ~0x31, ~0)|RS2_G0, "[1],C", F_NOTV9, v6 }, /* ld [rs1+%g0],d */
+{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1), "[1+i],C", F_NOTV9, v6 },
+{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1), "[i+1],C", F_NOTV9, v6 },
+{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1)|RS1_G0, "[i],C", F_NOTV9, v6 },
+{ "ld", F3(3, 0x31, 1), F3(~3, ~0x31, ~1)|SIMM13(~0), "[1],C", F_NOTV9, v6 }, /* ld [rs1+0],d */
/* The v9 LDUW is the same as the old 'ld' opcode, it is not the same as the
'ld' pseudo-op in v9. */
{ "ldd", F3(3, 0x23, 1), F3(~3, ~0x23, ~1), "[i+1],H", 0, v6 },
{ "ldd", F3(3, 0x23, 1), F3(~3, ~0x23, ~1)|RS1_G0, "[i],H", 0, v6 },
{ "ldd", F3(3, 0x23, 1), F3(~3, ~0x23, ~1)|SIMM13(~0), "[1],H", 0, v6 }, /* ldd [rs1+0],d */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "ldd", F3(3, 0x33, 0), F3(~3, ~0x33, ~0)|ASI(~0), "[1+2],D", F_ALIAS, v6 },
-{ "ldd", F3(3, 0x33, 0), F3(~3, ~0x33, ~0)|ASI_RS2(~0), "[1],D", F_ALIAS, v6 }, /* ldd [rs1+%g0],d */
-{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1), "[1+i],D", F_ALIAS, v6 },
-{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1), "[i+1],D", F_ALIAS, v6 },
-{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1)|RS1_G0, "[i],D", F_ALIAS, v6 },
-{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1)|SIMM13(~0), "[1],D", F_ALIAS, v6 }, /* ldd [rs1+0],d */
+
+{ "ldd", F3(3, 0x33, 0), F3(~3, ~0x33, ~0)|ASI(~0), "[1+2],D", F_NOTV9, v6 },
+{ "ldd", F3(3, 0x33, 0), F3(~3, ~0x33, ~0)|ASI_RS2(~0), "[1],D", F_NOTV9, v6 }, /* ldd [rs1+%g0],d */
+{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1), "[1+i],D", F_NOTV9, v6 },
+{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1), "[i+1],D", F_NOTV9, v6 },
+{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1)|RS1_G0, "[i],D", F_NOTV9, v6 },
+{ "ldd", F3(3, 0x33, 1), F3(~3, ~0x33, ~1)|SIMM13(~0), "[1],D", F_NOTV9, v6 }, /* ldd [rs1+0],d */
{ "ldq", F3(3, 0x22, 0), F3(~3, ~0x22, ~0)|ASI(~0), "[1+2],J", 0, v9 },
{ "ldq", F3(3, 0x22, 0), F3(~3, ~0x22, ~0)|ASI_RS2(~0), "[1],J", 0, v9 }, /* ldd [rs1+%g0],d */
{ "st", F3(3, 0x24, 1), F3(~3, ~0x24, ~1), "g,[i+1]", 0, v6 },
{ "st", F3(3, 0x24, 1), F3(~3, ~0x24, ~1)|RS1_G0, "g,[i]", 0, v6 },
{ "st", F3(3, 0x24, 1), F3(~3, ~0x24, ~1)|SIMM13(~0), "g,[1]", 0, v6 }, /* st d,[rs1+0] */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "st", F3(3, 0x34, 0), F3(~3, ~0x34, ~0)|ASI(~0), "D,[1+2]", F_ALIAS, v6 },
-{ "st", F3(3, 0x34, 0), F3(~3, ~0x34, ~0)|ASI_RS2(~0), "D,[1]", F_ALIAS, v6 }, /* st d,[rs1+%g0] */
-{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1), "D,[1+i]", F_ALIAS, v6 },
-{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1), "D,[i+1]", F_ALIAS, v6 },
-{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1)|RS1_G0, "D,[i]", F_ALIAS, v6 },
-{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1)|SIMM13(~0), "D,[1]", F_ALIAS, v6 }, /* st d,[rs1+0] */
-{ "st", F3(3, 0x35, 0), F3(~3, ~0x35, ~0)|ASI(~0), "C,[1+2]", 0, v6 },
-{ "st", F3(3, 0x35, 0), F3(~3, ~0x35, ~0)|ASI_RS2(~0), "C,[1]", 0, v6 }, /* st d,[rs1+%g0] */
-{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1), "C,[1+i]", 0, v6 },
-{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1), "C,[i+1]", 0, v6 },
-{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1)|RS1_G0, "C,[i]", 0, v6 },
-{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1)|SIMM13(~0), "C,[1]", 0, v6 }, /* st d,[rs1+0] */
+
+{ "st", F3(3, 0x34, 0), F3(~3, ~0x34, ~0)|ASI(~0), "D,[1+2]", F_NOTV9, v6 },
+{ "st", F3(3, 0x34, 0), F3(~3, ~0x34, ~0)|ASI_RS2(~0), "D,[1]", F_NOTV9, v6 }, /* st d,[rs1+%g0] */
+{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1), "D,[1+i]", F_NOTV9, v6 },
+{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1), "D,[i+1]", F_NOTV9, v6 },
+{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1)|RS1_G0, "D,[i]", F_NOTV9, v6 },
+{ "st", F3(3, 0x34, 1), F3(~3, ~0x34, ~1)|SIMM13(~0), "D,[1]", F_NOTV9, v6 }, /* st d,[rs1+0] */
+{ "st", F3(3, 0x35, 0), F3(~3, ~0x35, ~0)|ASI(~0), "C,[1+2]", F_NOTV9, v6 },
+{ "st", F3(3, 0x35, 0), F3(~3, ~0x35, ~0)|ASI_RS2(~0), "C,[1]", F_NOTV9, v6 }, /* st d,[rs1+%g0] */
+{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1), "C,[1+i]", F_NOTV9, v6 },
+{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1), "C,[i+1]", F_NOTV9, v6 },
+{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1)|RS1_G0, "C,[i]", F_NOTV9, v6 },
+{ "st", F3(3, 0x35, 1), F3(~3, ~0x35, ~1)|SIMM13(~0), "C,[1]", F_NOTV9, v6 }, /* st d,[rs1+0] */
{ "st", F3(3, 0x25, 0), F3(~3, ~0x25, ~0)|RD_G0|ASI(~0), "F,[1+2]", 0, v6 },
{ "st", F3(3, 0x25, 0), F3(~3, ~0x25, ~0)|RD_G0|ASI_RS2(~0), "F,[1]", 0, v6 }, /* st d,[rs1+%g0] */
{ "std", F3(3, 0x07, 1), F3(~3, ~0x07, ~1), "d,[i+1]", 0, v6 },
{ "std", F3(3, 0x07, 1), F3(~3, ~0x07, ~1)|RS1_G0, "d,[i]", 0, v6 },
{ "std", F3(3, 0x07, 1), F3(~3, ~0x07, ~1)|SIMM13(~0), "d,[1]", 0, v6 }, /* std d,[rs1+0] */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "std", F3(3, 0x26, 0), F3(~3, ~0x26, ~0)|ASI(~0), "q,[1+2]", F_ALIAS, v6 },
-{ "std", F3(3, 0x26, 0), F3(~3, ~0x26, ~0)|ASI_RS2(~0), "q,[1]", F_ALIAS, v6 }, /* std d,[rs1+%g0] */
-{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1), "q,[1+i]", F_ALIAS, v6 },
-{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1), "q,[i+1]", F_ALIAS, v6 },
-{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1)|RS1_G0, "q,[i]", F_ALIAS, v6 },
-{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1)|SIMM13(~0), "q,[1]", F_ALIAS, v6 }, /* std d,[rs1+0] */
+
+{ "std", F3(3, 0x26, 0), F3(~3, ~0x26, ~0)|ASI(~0), "q,[1+2]", F_NOTV9, v6 },
+{ "std", F3(3, 0x26, 0), F3(~3, ~0x26, ~0)|ASI_RS2(~0), "q,[1]", F_NOTV9, v6 }, /* std d,[rs1+%g0] */
+{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1), "q,[1+i]", F_NOTV9, v6 },
+{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1), "q,[i+1]", F_NOTV9, v6 },
+{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1)|RS1_G0, "q,[i]", F_NOTV9, v6 },
+{ "std", F3(3, 0x26, 1), F3(~3, ~0x26, ~1)|SIMM13(~0), "q,[1]", F_NOTV9, v6 }, /* std d,[rs1+0] */
{ "std", F3(3, 0x27, 0), F3(~3, ~0x27, ~0)|ASI(~0), "H,[1+2]", 0, v6 },
{ "std", F3(3, 0x27, 0), F3(~3, ~0x27, ~0)|ASI_RS2(~0), "H,[1]", 0, v6 }, /* std d,[rs1+%g0] */
{ "std", F3(3, 0x27, 1), F3(~3, ~0x27, ~1), "H,[1+i]", 0, v6 },
{ "std", F3(3, 0x27, 1), F3(~3, ~0x27, ~1), "H,[i+1]", 0, v6 },
{ "std", F3(3, 0x27, 1), F3(~3, ~0x27, ~1)|RS1_G0, "H,[i]", 0, v6 },
{ "std", F3(3, 0x27, 1), F3(~3, ~0x27, ~1)|SIMM13(~0), "H,[1]", 0, v6 }, /* std d,[rs1+0] */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "std", F3(3, 0x36, 0), F3(~3, ~0x36, ~0)|ASI(~0), "Q,[1+2]", F_ALIAS, v6 },
-{ "std", F3(3, 0x36, 0), F3(~3, ~0x36, ~0)|ASI_RS2(~0), "Q,[1]", F_ALIAS, v6 }, /* std d,[rs1+%g0] */
-{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1), "Q,[1+i]", F_ALIAS, v6 },
-{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1), "Q,[i+1]", F_ALIAS, v6 },
-{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1)|RS1_G0, "Q,[i]", F_ALIAS, v6 },
-{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1)|SIMM13(~0), "Q,[1]", F_ALIAS, v6 }, /* std d,[rs1+0] */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "std", F3(3, 0x37, 0), F3(~3, ~0x37, ~0)|ASI(~0), "D,[1+2]", F_ALIAS, v6 },
-{ "std", F3(3, 0x37, 0), F3(~3, ~0x37, ~0)|ASI_RS2(~0), "D,[1]", F_ALIAS, v6 }, /* std d,[rs1+%g0] */
-{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1), "D,[1+i]", F_ALIAS, v6 },
-{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1), "D,[i+1]", F_ALIAS, v6 },
-{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1)|RS1_G0, "D,[i]", F_ALIAS, v6 },
-{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1)|SIMM13(~0), "D,[1]", F_ALIAS, v6 }, /* std d,[rs1+0] */
+
+{ "std", F3(3, 0x36, 0), F3(~3, ~0x36, ~0)|ASI(~0), "Q,[1+2]", F_NOTV9, v6 },
+{ "std", F3(3, 0x36, 0), F3(~3, ~0x36, ~0)|ASI_RS2(~0), "Q,[1]", F_NOTV9, v6 }, /* std d,[rs1+%g0] */
+{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1), "Q,[1+i]", F_NOTV9, v6 },
+{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1), "Q,[i+1]", F_NOTV9, v6 },
+{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1)|RS1_G0, "Q,[i]", F_NOTV9, v6 },
+{ "std", F3(3, 0x36, 1), F3(~3, ~0x36, ~1)|SIMM13(~0), "Q,[1]", F_NOTV9, v6 }, /* std d,[rs1+0] */
+{ "std", F3(3, 0x37, 0), F3(~3, ~0x37, ~0)|ASI(~0), "D,[1+2]", F_NOTV9, v6 },
+{ "std", F3(3, 0x37, 0), F3(~3, ~0x37, ~0)|ASI_RS2(~0), "D,[1]", F_NOTV9, v6 }, /* std d,[rs1+%g0] */
+{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1), "D,[1+i]", F_NOTV9, v6 },
+{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1), "D,[i+1]", F_NOTV9, v6 },
+{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1)|RS1_G0, "D,[i]", F_NOTV9, v6 },
+{ "std", F3(3, 0x37, 1), F3(~3, ~0x37, ~1)|SIMM13(~0), "D,[1]", F_NOTV9, v6 }, /* std d,[rs1+0] */
{ "stda", F3(3, 0x17, 0), F3(~3, ~0x17, ~0), "d,[1+2]A", 0, v6 },
{ "stda", F3(3, 0x17, 0), F3(~3, ~0x17, ~0)|RS2(~0), "d,[1]A", 0, v6 }, /* stda d,[rs1+%g0] */
{ "jmpl", F3(2, 0x38, 1), F3(~2, ~0x38, ~1), "1+i,d", F_JSR|F_DELAYED, v6 },
{ "jmpl", F3(2, 0x38, 1), F3(~2, ~0x38, ~1), "i+1,d", F_JSR|F_DELAYED, v6 },
-{ "done", F3(2, 0x3e, 1)|RD(0), F3(~2, ~0x3e, ~1)|RD(~0)|RS1_G0|SIMM13(~0), "", 0, v9 },
-{ "retry", F3(2, 0x3e, 1)|RD(1), F3(~2, ~0x3e, ~1)|RD(~1)|RS1_G0|SIMM13(~0), "", 0, v9 },
-{ "saved", F3(2, 0x31, 1)|RD(0), F3(~2, ~0x31, ~1)|RD(~0)|RS1_G0|SIMM13(~0), "", 0, v9 },
-{ "restored", F3(2, 0x31, 1)|RD(1), F3(~2, ~0x31, ~1)|RD(~1)|RS1_G0|SIMM13(~0), "", 0, v9 },
-{ "sir", F3(2, 0x30, 0)|RD(0xf), F3(~2, ~0x30, ~0)|RD(~0xf)|RS1_G0|SIMM13(~0), "", 0, v9 },
+{ "done", F3(2, 0x3e, 0)|RD(0), F3(~2, ~0x3e, ~0)|RD(~0)|RS1_G0|SIMM13(~0), "", 0, v9 },
+{ "retry", F3(2, 0x3e, 0)|RD(1), F3(~2, ~0x3e, ~0)|RD(~1)|RS1_G0|SIMM13(~0), "", 0, v9 },
+{ "saved", F3(2, 0x31, 0)|RD(0), F3(~2, ~0x31, ~0)|RD(~0)|RS1_G0|SIMM13(~0), "", 0, v9 },
+{ "restored", F3(2, 0x31, 0)|RD(1), F3(~2, ~0x31, ~0)|RD(~1)|RS1_G0|SIMM13(~0), "", 0, v9 },
+{ "sir", F3(2, 0x30, 1)|RD(0xf), F3(~2, ~0x30, ~1)|RD(~0xf)|RS1_G0, "i", 0, v9 },
{ "flush", F3(2, 0x3b, 0), F3(~2, ~0x3b, ~0)|ASI(~0), "1+2", 0, v8 },
{ "flush", F3(2, 0x3b, 0), F3(~2, ~0x3b, ~0)|ASI_RS2(~0), "1", 0, v8 }, /* flush rs1+%g0 */
{ "flushw", F3(2, 0x2b, 0), F3(~2, ~0x2b, ~0)|RD_G0|RS1_G0|ASI_RS2(~0), "", 0, v9 },
-{ "membar", F3(2, 0x28, 1)|RS1(0xf), F3(~2, ~0x28, ~1)|RD_G0|RS1(~0xf)|ASI(~0), "K", 0, v9 },
+{ "membar", F3(2, 0x28, 1)|RS1(0xf), F3(~2, ~0x28, ~1)|RD_G0|RS1(~0xf)|SIMM13(~0), "K", 0, v9 },
{ "stbar", F3(2, 0x28, 0)|RS1(0xf), F3(~2, ~0x28, ~0)|RD_G0|RS1(~0xf)|SIMM13(~0), "", 0, v8 },
{ "prefetch", F3(3, 0x2d, 0), F3(~3, ~0x2d, ~0), "[1+2],*", 0, v9 },
{ "prefetcha", F3(3, 0x3d, 1), F3(~3, ~0x3d, ~1)|SIMM13(~0), "[1]o,d", 0, v9 }, /* prefetcha [rs1+0],d */
/* The 1<<12 is a long story. It is necessary. For more info, please contact rich@cygnus.com */
+ /* FIXME: 'i' is wrong, need new letter for 5 bit unsigned constants. */
{ "sll", F3(2, 0x25, 0), F3(~2, ~0x25, ~0)|(1<<12)|ASI(~0), "1,2,d", 0, v6 },
{ "sll", F3(2, 0x25, 1), F3(~2, ~0x25, ~1)|(1<<12), "1,i,d", 0, v6 },
{ "sra", F3(2, 0x27, 0), F3(~2, ~0x27, ~0)|(1<<12)|ASI(~0), "1,2,d", 0, v6 },
{ "srl", F3(2, 0x26, 0), F3(~2, ~0x26, ~0)|(1<<12)|ASI(~0), "1,2,d", 0, v6 },
{ "srl", F3(2, 0x26, 1), F3(~2, ~0x26, ~1)|(1<<12), "1,i,d", 0, v6 },
+ /* FIXME: 'j' is wrong, need new letter for 6 bit unsigned constants. */
{ "sllx", F3(2, 0x25, 0)|(1<<12), F3(~2, ~0x25, ~0)|(ASI(~0)^(1<<12)), "1,2,d", 0, v9 },
-{ "sllx", F3(2, 0x25, 1)|(1<<12), F3(~2, ~0x25, ~1), "1,i,d", 0, v9 },
+{ "sllx", F3(2, 0x25, 1)|(1<<12), F3(~2, ~0x25, ~1)|(0x3f<<6), "1,j,d", 0, v9 },
{ "srax", F3(2, 0x27, 0)|(1<<12), F3(~2, ~0x27, ~0)|(ASI(~0)^(1<<12)), "1,2,d", 0, v9 },
-{ "srax", F3(2, 0x27, 1)|(1<<12), F3(~2, ~0x27, ~1), "1,i,d", 0, v9 },
+{ "srax", F3(2, 0x27, 1)|(1<<12), F3(~2, ~0x27, ~1)|(0x3f<<6), "1,j,d", 0, v9 },
{ "srlx", F3(2, 0x26, 0)|(1<<12), F3(~2, ~0x26, ~0)|(ASI(~0)^(1<<12)), "1,2,d", 0, v9 },
-{ "srlx", F3(2, 0x26, 1)|(1<<12), F3(~2, ~0x26, ~1), "1,i,d", 0, v9 },
+{ "srlx", F3(2, 0x26, 1)|(1<<12), F3(~2, ~0x26, ~1)|(0x3f<<6), "1,j,d", 0, v9 },
{ "mulscc", F3(2, 0x24, 0), F3(~2, ~0x24, ~0)|ASI(~0), "1,2,d", 0, v6 },
{ "mulscc", F3(2, 0x24, 1), F3(~2, ~0x24, ~1), "1,i,d", 0, v6 },
{ "tst", F3(2, 0x12, 0), F3(~2, ~0x12, ~0)|RD_G0|RS1_G0|ASI(~0), "2", 0, v6 }, /* orcc %g0, rs2, %g0 */
{ "tst", F3(2, 0x12, 1), F3(~2, ~0x12, ~1)|RD_G0|SIMM13(~0), "1", 0, v6 }, /* orcc rs1, 0, %g0 */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
{ "wr", F3(2, 0x30, 0), F3(~2, ~0x30, ~0)|ASI(~0), "1,2,m", 0, v8 }, /* wr r,r,%asrX */
{ "wr", F3(2, 0x30, 0), F3(~2, ~0x30, ~0)|RD_G0|ASI(~0), "1,2,y", 0, v6 }, /* wr r,r,%y */
{ "wr", F3(2, 0x30, 1), F3(~2, ~0x30, ~1), "1,i,m", 0, v8 }, /* wr r,i,%asrX */
{ "wr", F3(2, 0x30, 1), F3(~2, ~0x30, ~1)|RD_G0, "1,i,y", 0, v6 }, /* wr r,i,%y */
-{ "wr", F3(2, 0x31, 0), F3(~2, ~0x31, ~0)|RD_G0|ASI(~0), "1,2,p", F_ALIAS, v6 }, /* wr r,r,%psr */
-{ "wr", F3(2, 0x31, 1), F3(~2, ~0x31, ~1)|RD_G0, "1,i,p", F_ALIAS, v6 }, /* wr r,i,%psr */
-{ "wr", F3(2, 0x32, 0), F3(~2, ~0x32, ~0)|RD_G0|ASI(~0), "1,2,w", F_ALIAS, v6 }, /* wr r,r,%wim */
-{ "wr", F3(2, 0x32, 1), F3(~2, ~0x32, ~1)|RD_G0, "1,i,w", F_ALIAS, v6 }, /* wr r,i,%wim */
-{ "wr", F3(2, 0x33, 0), F3(~2, ~0x33, ~0)|RD_G0|ASI(~0), "1,2,t", F_ALIAS, v6 }, /* wr r,r,%tbr */
-{ "wr", F3(2, 0x33, 1), F3(~2, ~0x33, ~1)|RD_G0, "1,i,t", F_ALIAS, v6 }, /* wr r,i,%tbr */
+{ "wr", F3(2, 0x31, 0), F3(~2, ~0x31, ~0)|RD_G0|ASI(~0), "1,2,p", F_NOTV9, v6 }, /* wr r,r,%psr */
+{ "wr", F3(2, 0x31, 1), F3(~2, ~0x31, ~1)|RD_G0, "1,i,p", F_NOTV9, v6 }, /* wr r,i,%psr */
+{ "wr", F3(2, 0x32, 0), F3(~2, ~0x32, ~0)|RD_G0|ASI(~0), "1,2,w", F_NOTV9, v6 }, /* wr r,r,%wim */
+{ "wr", F3(2, 0x32, 1), F3(~2, ~0x32, ~1)|RD_G0, "1,i,w", F_NOTV9, v6 }, /* wr r,i,%wim */
+{ "wr", F3(2, 0x33, 0), F3(~2, ~0x33, ~0)|RD_G0|ASI(~0), "1,2,t", F_NOTV9, v6 }, /* wr r,r,%tbr */
+{ "wr", F3(2, 0x33, 1), F3(~2, ~0x33, ~1)|RD_G0, "1,i,t", F_NOTV9, v6 }, /* wr r,i,%tbr */
{ "wr", F3(2, 0x30, 0)|RD(0), F3(~2, ~0x30, ~0)|RD(~0)|ASI(~0), "1,2,y", 0, v9 }, /* wr r,r,%y */
{ "wr", F3(2, 0x30, 1)|RD(0), F3(~2, ~0x30, ~1)|RD(~0), "1,i,y", 0, v9 }, /* wr r,i,%y */
{ "wr", F3(2, 0x30, 0)|RD(6), F3(~2, ~0x30, ~0)|RD(~6)|ASI(~0), "1,2,s", 0, v9 }, /* wr r,i,%fprs */
{ "wr", F3(2, 0x30, 1)|RD(6), F3(~2, ~0x30, ~1)|RD(~6), "1,i,s", 0, v9 }, /* wr r,i,%fprs */
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
{ "rd", F3(2, 0x28, 0), F3(~2, ~0x28, ~0)|SIMM13(~0), "M,d", 0, v8 }, /* rd %asrX,r */
{ "rd", F3(2, 0x28, 0), F3(~2, ~0x28, ~0)|RS1_G0|SIMM13(~0), "y,d", 0, v6 }, /* rd %y,r */
-{ "rd", F3(2, 0x29, 0), F3(~2, ~0x29, ~0)|RS1_G0|SIMM13(~0), "p,d", F_ALIAS, v6 }, /* rd %psr,r */
-{ "rd", F3(2, 0x2a, 0), F3(~2, ~0x2a, ~0)|RS1_G0|SIMM13(~0), "w,d", F_ALIAS, v6 }, /* rd %wim,r */
-{ "rd", F3(2, 0x2b, 0), F3(~2, ~0x2b, ~0)|RS1_G0|SIMM13(~0), "t,d", F_ALIAS, v6 }, /* rd %tbr,r */
+{ "rd", F3(2, 0x29, 0), F3(~2, ~0x29, ~0)|RS1_G0|SIMM13(~0), "p,d", F_NOTV9, v6 }, /* rd %psr,r */
+{ "rd", F3(2, 0x2a, 0), F3(~2, ~0x2a, ~0)|RS1_G0|SIMM13(~0), "w,d", F_NOTV9, v6 }, /* rd %wim,r */
+{ "rd", F3(2, 0x2b, 0), F3(~2, ~0x2b, ~0)|RS1_G0|SIMM13(~0), "t,d", F_NOTV9, v6 }, /* rd %tbr,r */
{ "rd", F3(2, 0x28, 0)|RS1(2), F3(~2, ~0x28, ~0)|RS1(~2)|SIMM13(~0), "E,d", 0, v9 }, /* rd %ccr,r */
{ "rd", F3(2, 0x28, 0)|RS1(3), F3(~2, ~0x28, ~0)|RS1(~3)|SIMM13(~0), "o,d", 0, v9 }, /* rd %asi,r */
{ "rd", F3(2, 0x28, 0)|RS1(4), F3(~2, ~0x28, ~0)|RS1(~4)|SIMM13(~0), "W,d", 0, v9 }, /* rd %tick,r */
{ "rd", F3(2, 0x28, 0)|RS1(5), F3(~2, ~0x28, ~0)|RS1(~5)|SIMM13(~0), "P,d", 0, v9 }, /* rd %pc,r */
-{ "rd", F3(2, 0x2b, 0)|RS1(6), F3(~2, ~0x2b, ~0)|RS1(~6)|SIMM13(~0), "s,d", 0, v9 }, /* rd %fprs,r */
+{ "rd", F3(2, 0x28, 0)|RS1(6), F3(~2, ~0x28, ~0)|RS1(~6)|SIMM13(~0), "s,d", 0, v9 }, /* rd %fprs,r */
{ "rdpr", F3(2, 0x2a, 0), F3(~2, ~0x2a, ~0)|SIMM13(~0), "?,d", 0, v9 }, /* rdpr %priv,r */
{ "wrpr", F3(2, 0x32, 0), F3(~2, ~0x32, ~0), "1,2,!", 0, v9 }, /* wrpr r1,r2,%priv */
v9: Methinks we will need some sort of F_NOTFORV9 flag. */
{ "mov", F3(2, 0x28, 0), F3(~2, ~0x28, ~0)|SIMM13(~0), "M,d", F_ALIAS, v8 }, /* rd %asr1,r */
{ "mov", F3(2, 0x28, 0), F3(~2, ~0x28, ~0)|RS1_G0|SIMM13(~0), "y,d", F_ALIAS, v6 }, /* rd %y,r */
-{ "mov", F3(2, 0x29, 0), F3(~2, ~0x29, ~0)|RS1_G0|SIMM13(~0), "p,d", F_ALIAS, v6 }, /* rd %psr,r */
-{ "mov", F3(2, 0x2a, 0), F3(~2, ~0x2a, ~0)|RS1_G0|SIMM13(~0), "w,d", F_ALIAS, v6 }, /* rd %wim,r */
-{ "mov", F3(2, 0x2b, 0), F3(~2, ~0x2b, ~0)|RS1_G0|SIMM13(~0), "t,d", F_ALIAS, v6 }, /* rd %tbr,r */
+{ "mov", F3(2, 0x29, 0), F3(~2, ~0x29, ~0)|RS1_G0|SIMM13(~0), "p,d", F_ALIAS|F_NOTV9, v6 }, /* rd %psr,r */
+{ "mov", F3(2, 0x2a, 0), F3(~2, ~0x2a, ~0)|RS1_G0|SIMM13(~0), "w,d", F_ALIAS|F_NOTV9, v6 }, /* rd %wim,r */
+{ "mov", F3(2, 0x2b, 0), F3(~2, ~0x2b, ~0)|RS1_G0|SIMM13(~0), "t,d", F_ALIAS|F_NOTV9, v6 }, /* rd %tbr,r */
{ "mov", F3(2, 0x30, 0), F3(~2, ~0x30, ~0)|ASI_RS2(~0), "1,y", F_ALIAS, v6 }, /* wr rs1,%g0,%y */
{ "mov", F3(2, 0x30, 1), F3(~2, ~0x30, ~1), "i,y", F_ALIAS, v6 },
{ "subcc", F3(2, 0x14, 0), F3(~2, ~0x14, ~0)|ASI(~0), "1,2,d", 0, v6 },
{ "subcc", F3(2, 0x14, 1), F3(~2, ~0x14, ~1), "1,i,d", 0, v6 },
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "subx", F3(2, 0x0c, 0), F3(~2, ~0x0c, ~0)|ASI(~0), "1,2,d", F_ALIAS, v6 },
-{ "subx", F3(2, 0x0c, 1), F3(~2, ~0x0c, ~1), "1,i,d", F_ALIAS, v6 },
+{ "subx", F3(2, 0x0c, 0), F3(~2, ~0x0c, ~0)|ASI(~0), "1,2,d", F_NOTV9, v6 },
+{ "subx", F3(2, 0x0c, 1), F3(~2, ~0x0c, ~1), "1,i,d", F_NOTV9, v6 },
{ "subc", F3(2, 0x0c, 0), F3(~2, ~0x0c, ~0)|ASI(~0), "1,2,d", 0, v9 },
{ "subc", F3(2, 0x0c, 1), F3(~2, ~0x0c, ~1), "1,i,d", 0, v9 },
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "subxcc", F3(2, 0x1c, 0), F3(~2, ~0x1c, ~0)|ASI(~0), "1,2,d", F_ALIAS, v6 },
-{ "subxcc", F3(2, 0x1c, 1), F3(~2, ~0x1c, ~1), "1,i,d", F_ALIAS, v6 },
+{ "subxcc", F3(2, 0x1c, 0), F3(~2, ~0x1c, ~0)|ASI(~0), "1,2,d", F_NOTV9, v6 },
+{ "subxcc", F3(2, 0x1c, 1), F3(~2, ~0x1c, ~1), "1,i,d", F_NOTV9, v6 },
{ "subccc", F3(2, 0x1c, 0), F3(~2, ~0x1c, ~0)|ASI(~0), "1,2,d", 0, v9 },
{ "subccc", F3(2, 0x1c, 1), F3(~2, ~0x1c, ~1), "1,i,d", 0, v9 },
{ "addcc", F3(2, 0x10, 0), F3(~2, ~0x10, ~0)|ASI(~0), "1,2,d", 0, v6 },
{ "addcc", F3(2, 0x10, 1), F3(~2, ~0x10, ~1), "1,i,d", 0, v6 },
{ "addcc", F3(2, 0x10, 1), F3(~2, ~0x10, ~1), "i,1,d", 0, v6 },
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "addx", F3(2, 0x08, 0), F3(~2, ~0x08, ~0)|ASI(~0), "1,2,d", F_ALIAS, v6 },
-{ "addx", F3(2, 0x08, 1), F3(~2, ~0x08, ~1), "1,i,d", F_ALIAS, v6 },
-{ "addx", F3(2, 0x08, 1), F3(~2, ~0x08, ~1), "i,1,d", F_ALIAS, v6 },
+
+{ "addx", F3(2, 0x08, 0), F3(~2, ~0x08, ~0)|ASI(~0), "1,2,d", F_NOTV9, v6 },
+{ "addx", F3(2, 0x08, 1), F3(~2, ~0x08, ~1), "1,i,d", F_NOTV9, v6 },
+{ "addx", F3(2, 0x08, 1), F3(~2, ~0x08, ~1), "i,1,d", F_NOTV9, v6 },
{ "addc", F3(2, 0x08, 0), F3(~2, ~0x08, ~0)|ASI(~0), "1,2,d", 0, v9 },
{ "addc", F3(2, 0x08, 1), F3(~2, ~0x08, ~1), "1,i,d", 0, v9 },
{ "addc", F3(2, 0x08, 1), F3(~2, ~0x08, ~1), "i,1,d", 0, v9 },
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "addxcc", F3(2, 0x18, 0), F3(~2, ~0x18, ~0)|ASI(~0), "1,2,d", F_ALIAS, v6 },
-{ "addxcc", F3(2, 0x18, 1), F3(~2, ~0x18, ~1), "1,i,d", F_ALIAS, v6 },
-{ "addxcc", F3(2, 0x18, 1), F3(~2, ~0x18, ~1), "i,1,d", F_ALIAS, v6 },
+
+{ "addxcc", F3(2, 0x18, 0), F3(~2, ~0x18, ~0)|ASI(~0), "1,2,d", F_NOTV9, v6 },
+{ "addxcc", F3(2, 0x18, 1), F3(~2, ~0x18, ~1), "1,i,d", F_NOTV9, v6 },
+{ "addxcc", F3(2, 0x18, 1), F3(~2, ~0x18, ~1), "i,1,d", F_NOTV9, v6 },
{ "addccc", F3(2, 0x18, 0), F3(~2, ~0x18, ~0)|ASI(~0), "1,2,d", 0, v9 },
{ "addccc", F3(2, 0x18, 1), F3(~2, ~0x18, ~1), "1,i,d", 0, v9 },
{ "addccc", F3(2, 0x18, 1), F3(~2, ~0x18, ~1), "i,1,d", 0, v9 },
{ "tsubcctv", F3(2, 0x23, 0), F3(~2, ~0x23, ~0)|ASI(~0), "1,2,d", 0, v6 },
{ "tsubcctv", F3(2, 0x23, 1), F3(~2, ~0x23, ~1), "1,i,d", 0, v6 },
-/* FIXME This is marked F_ALIAS, so that it won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "unimp", F2(0x0, 0x0), 0xffc00000, "n", F_ALIAS, v6 },
+{ "unimp", F2(0x0, 0x0), 0xffc00000, "n", F_NOTV9, v6 },
{ "illtrap", F2(0, 0), F2(~0, ~0)|RD_G0, "n", 0, v9 },
/* This *is* a commutative instruction. */
{ "fcmpes", CMPFCC(2)|F3F(2, 0x35, 0x055), CMPFCC(~2)|F3F(~2, ~0x35, ~0x055), "8,e,f", 0, v9 },
{ "fcmpes", CMPFCC(3)|F3F(2, 0x35, 0x055), CMPFCC(~3)|F3F(~2, ~0x35, ~0x055), "9,e,f", 0, v9 },
-/* FIXME These are marked F_ALIAS, so that they won't conflict with new v9
- insns when v9 is present. Otherwise, the F_ALIAS flag is ignored. */
-{ "cpop1", F3(2, 0x36, 0), F3(~2, ~0x36, ~1), "[1+2],d", F_ALIAS, v6 },
-{ "cpop2", F3(2, 0x37, 0), F3(~2, ~0x37, ~1), "[1+2],d", F_ALIAS, v6 },
+/* These Extended FPop (FIFO) instructions are new in the Fujitsu
+ MB86934, replacing the CPop instructions from v6 and later
+ processors. */
+
+#define EFPOP1_2(name, op, args) { name, F3F(2, 0x36, op), F3F(~2, ~0x36, ~op)|RS1_G0, args, 0, sparclite }
+#define EFPOP1_3(name, op, args) { name, F3F(2, 0x36, op), F3F(~2, ~0x36, ~op), args, 0, sparclite }
+#define EFPOP2_2(name, op, args) { name, F3F(2, 0x37, op), F3F(~2, ~0x37, ~op)|RD_G0, args, 0, sparclite }
+
+EFPOP1_2 ("efitod", 0x0c8, "f,H"),
+EFPOP1_2 ("efitos", 0x0c4, "f,g"),
+EFPOP1_2 ("efdtoi", 0x0d2, "B,g"),
+EFPOP1_2 ("efstoi", 0x0d1, "f,g"),
+EFPOP1_2 ("efstod", 0x0c9, "f,H"),
+EFPOP1_2 ("efdtos", 0x0c6, "B,g"),
+EFPOP1_2 ("efmovs", 0x001, "f,g"),
+EFPOP1_2 ("efnegs", 0x005, "f,g"),
+EFPOP1_2 ("efabss", 0x009, "f,g"),
+EFPOP1_2 ("efsqrtd", 0x02a, "B,H"),
+EFPOP1_2 ("efsqrts", 0x029, "f,g"),
+EFPOP1_3 ("efaddd", 0x042, "v,B,H"),
+EFPOP1_3 ("efadds", 0x041, "e,f,g"),
+EFPOP1_3 ("efsubd", 0x046, "v,B,H"),
+EFPOP1_3 ("efsubs", 0x045, "e,f,g"),
+EFPOP1_3 ("efdivd", 0x04e, "v,B,H"),
+EFPOP1_3 ("efdivs", 0x04d, "e,f,g"),
+EFPOP1_3 ("efmuld", 0x04a, "v,B,H"),
+EFPOP1_3 ("efmuls", 0x049, "e,f,g"),
+EFPOP1_3 ("efsmuld", 0x069, "e,f,H"),
+EFPOP2_2 ("efcmpd", 0x052, "v,B"),
+EFPOP2_2 ("efcmped", 0x056, "v,B"),
+EFPOP2_2 ("efcmps", 0x051, "e,f"),
+EFPOP2_2 ("efcmpes", 0x055, "e,f"),
+
+#undef EFPOP1_2
+#undef EFPOP1_3
+#undef EFPOP2_2
+
+/* These are marked F_ALIAS, so that they won't conflict with sparclite insns
+ present. Otherwise, the F_ALIAS flag is ignored. */
+{ "cpop1", F3(2, 0x36, 0), F3(~2, ~0x36, ~1), "[1+2],d", F_ALIAS|F_NOTV9, v6 },
+{ "cpop2", F3(2, 0x37, 0), F3(~2, ~0x37, ~1), "[1+2],d", F_ALIAS|F_NOTV9, v6 },
+
+#define IMPDEP(name, code) \
+{ name, F3(2, code, 0), F3(~2, ~code, ~0)|ASI(~0), "1,2,d", 0, v9 }, \
+{ name, F3(2, code, 1), F3(~2, ~code, ~1), "1,i,d", 0, v9 }, \
+{ name, F3(2, code, 0), F3(~2, ~code, ~0), "x,1,2,d", 0, v9 }, \
+{ name, F3(2, code, 0), F3(~2, ~code, ~0), "x,e,f,g", 0, v9 }
+
+IMPDEP ("impdep1", 0x36),
+IMPDEP ("impdep2", 0x37),
+
+#undef IMPDEP
-{ "impdep1", F3(2, 0x36, 0), F3(~2, ~0x36, ~0)|ASI(~0), "1,2,d", 0, v9 },
-{ "impdep1", F3(2, 0x36, 1), F3(~2, ~0x36, ~1), "1,i,d", 0, v9 },
-{ "impdep2", F3(2, 0x37, 0), F3(~2, ~0x37, ~0)|ASI(~0), "1,2,d", 0, v9 },
-{ "impdep2", F3(2, 0x37, 1), F3(~2, ~0x37, ~1), "1,i,d", 0, v9 },
-
{ "casa", F3(3, 0x3c, 0), F3(~3, ~0x3c, ~0), "[1]A,2,d", 0, v9 },
{ "casa", F3(3, 0x3c, 1), F3(~3, ~0x3c, ~1), "[1]o,2,d", 0, v9 },
{ "casxa", F3(3, 0x3e, 0), F3(~3, ~0x3e, ~0), "[1]A,2,d", 0, v9 },
};
-CONST int bfd_sparc_num_opcodes = ((sizeof sparc_opcodes)/(sizeof sparc_opcodes[0]));
+const int bfd_sparc_num_opcodes = ((sizeof sparc_opcodes)/(sizeof sparc_opcodes[0]));
projects / binutils-gdb.git / commitdiff