aboutsummaryrefslogtreecommitdiff
path: root/gdb
diff options
context:
space:
mode:
Diffstat (limited to 'gdb')
-rw-r--r--gdb/.Sanitize5
-rw-r--r--gdb/ChangeLog11
-rw-r--r--gdb/Makefile.in28
-rw-r--r--gdb/configure.in7
-rw-r--r--gdb/i860-opcode.h137
-rw-r--r--gdb/i860-pinsn.c660
-rw-r--r--gdb/i860-tdep.c1636
-rw-r--r--gdb/tm-stratus.h21
-rw-r--r--gdb/xm-stratus.h29
9 files changed, 2523 insertions, 11 deletions
diff --git a/gdb/.Sanitize b/gdb/.Sanitize
index 4150b36..bd5e157 100644
--- a/gdb/.Sanitize
+++ b/gdb/.Sanitize
@@ -96,6 +96,9 @@ i386-stub.c
i386-tdep.c
i386-xdep.c
i387-tdep.c
+i860-opcode.h
+i860-pinsn.c
+i860-tdep.c
i960-pinsn.c
i960-tdep.c
ieee-float.c
@@ -209,6 +212,7 @@ tm-pn.h
tm-pyr.h
tm-rs6000.h
tm-sparc.h
+tm-stratus.h
tm-sun2.h
tm-sun2os4.h
tm-sun3.h
@@ -264,6 +268,7 @@ xm-pyr.h
xm-rs6000.h
xm-rtbsd.h
xm-sparc.h
+xm-stratus.h
xm-sun2.h
xm-sun3.h
xm-sun386.h
diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index ad97ecb..786af21 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,14 @@
+Sat Jan 18 16:45:01 1992 Fred Fish (fnf at cygnus.com)
+
+ * config/mh-stratus, config/mt-stratus, i860-opcode.h,
+ i860-pinsn.c, i860-tdep.c, tm-stratus.h, xm-stratus.h:
+ New files for stratus.
+
+ * Makefile.in (USER_CFLAGS): Add macro specifically reserved
+ for getting custom flags into CFLAGS when doing makes.
+
+ * configure.in: Add config information for stratus.
+
Thu Jan 16 00:51:46 1992 Fred Fish (fnf at cygnus.com)
* dbxread.c (process_one_symbol): Pass address of structure
diff --git a/gdb/Makefile.in b/gdb/Makefile.in
index 6278428..bd47624 100644
--- a/gdb/Makefile.in
+++ b/gdb/Makefile.in
@@ -135,7 +135,10 @@ MINUS_G=-g
GLOBAL_CFLAGS = ${MINUS_G} ${TM_CFLAGS} ${XM_CFLAGS}
#PROFILE_CFLAGS = -pg
-CFLAGS = ${GLOBAL_CFLAGS} ${PROFILE_CFLAGS} ${MALLOC_CFLAGS} ${INCLUDE_CFLAGS}
+# CFLAGS is the aggregate of several individual *_CFLAGS macros.
+# USER_CFLAGS is specifically reserved for setting from the command line
+# when running make. I.E. "make USER_CFLAGS=-Wmissing-prototypes".
+CFLAGS = ${GLOBAL_CFLAGS} ${PROFILE_CFLAGS} ${MALLOC_CFLAGS} ${INCLUDE_CFLAGS} ${USER_CFLAGS}
# None of the things in CFLAGS will do any harm, and on some systems
# (e.g. SunOS4) it is important to use the M_CFLAGS.
LDFLAGS = $(CFLAGS)
@@ -221,20 +224,23 @@ HFILES= breakpoint.h buildsym.h command.h defs.h environ.h \
expression.h frame.h gdbcmd.h gdbcore.h \
ieee-float.h inferior.h minimon.h \
signals.h signame.h symfile.h symtab.h \
- target.h terminal.h tm-68k.h tm-i960.h tm-sunos.h tm-svr4.h \
- xm-m68k.h xm-svr4.h language.h parser-defs.h value.h
+ target.h terminal.h tm-68k.h tm-i960.h tm-sunos.h tm-sysv4.h \
+ xm-m68k.h xm-sysv4.h language.h parser-defs.h value.h
# Header files for machine opcode tables.
# when OPCODE_DIR changes to ../include, remove the whole $(OPCODES)
# business, since they will be put into the GDB release files by somebody else.
-OPCODE_DIR =
-OPCODES = $(OPCODE_DIR)pn-opcode.h $(OPCODE_DIR)np1-opcode.h \
- $(OPCODE_DIR)sparc-opcode.h $(OPCODE_DIR)vax-opcode.h \
- $(OPCODE_DIR)m68k-opcode.h $(OPCODE_DIR)ns32k-opcode.h \
- $(OPCODE_DIR)convx-opcode.h $(OPCODE_DIR)pyr-opcode.h \
- $(OPCODE_DIR)mips-opcode.h $(OPCODE_DIR)am29k-opcode.h \
- $(OPCODE_DIR)arm-opcode.h $(OPCODE_DIR)m88k-opcode.h \
- $(OPCODE_DIR)tahoe-opcode.h $(OPCODE_DIR)rs6k-opcode.h
+OPCODE_DIR1 = ../include/opcode/
+OPCODE_DIR2 = ./
+OPCODES = $(OPCODE_DIR1)pn.h $(OPCODE_DIR1)np1.h \
+ $(OPCODE_DIR1)sparc.h $(OPCODE_DIR1)vax.h \
+ $(OPCODE_DIR1)m68k.h $(OPCODE_DIR2)ns32k-opcode.h \
+ $(OPCODE_DIR2)convx-opcode.h $(OPCODE_DIR1)pyr.h \
+ $(OPCODE_DIR1)mips.h $(OPCODE_DIR2)am29k-opcode.h \
+ $(OPCODE_DIR1)arm.h $(OPCODE_DIR1)m88k.h \
+ $(OPCODE_DIR1)tahoe.h $(OPCODE_DIR2)rs6k-opcode.h \
+ $(OPCODE_DIR1)h8300.h $(OPCODE_DIR1)i960.h \
+ $(OPCODE_DIR1)i386.h $(OPCODE_DIR1)i860.h
REMOTE_EXAMPLES = m68k-stub.c i386-stub.c rem-multi.shar
diff --git a/gdb/configure.in b/gdb/configure.in
index 37d55a2..9ce8a30 100644
--- a/gdb/configure.in
+++ b/gdb/configure.in
@@ -79,6 +79,10 @@ mips)
esac
;;
+i860)
+ gdb_host=stratus
+ ;;
+
i386)
case "${host_vendor}" in
sun) gdb_host=sun386 ;;
@@ -239,6 +243,9 @@ h8300)
gdb_target=h8300hds
;;
+i860) gdb_target=stratus
+ ;;
+
i386)
case "${target_vendor}" in
sco) gdb_target=i386sco ;;
diff --git a/gdb/i860-opcode.h b/gdb/i860-opcode.h
new file mode 100644
index 0000000..bdf45af
--- /dev/null
+++ b/gdb/i860-opcode.h
@@ -0,0 +1,137 @@
+/* Intel I860 opcde list for GDB, the GNU debugger.
+ Copyright (C) 1992 Free Software Foundation, Inc.
+ Contributed by Peggy Fieland (pfieland@stratus.com)
+
+GDB is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY. No author or distributor accepts responsibility to anyone
+for the consequences of using it or for whether it serves any
+particular purpose or works at all, unless he says so in writing.
+Refer to the GDB General Public License for full details.
+
+Everyone is granted permission to copy, modify and redistribute GDB,
+but only under the conditions described in the GDB General Public
+License. A copy of this license is supposed to have been given to you
+along with GDB so you can know your rights and responsibilities. It
+should be in a file named COPYING. Among other things, the copyright
+notice and this notice must be preserved on all copies.
+
+In other words, go ahead and share GDB, but don't try to stop
+anyone else from sharing it farther. Help stamp out software hoarding!
+*/
+
+#ifdef BIG_ENDIAN
+struct gen_fmt
+{
+ unsigned op1 : 6;
+ unsigned src2 : 5;
+ unsigned dest : 5;
+ unsigned src1 : 5;
+ unsigned offset : 11;
+};
+
+struct geni_fmt
+{
+ unsigned op1 : 6;
+ unsigned src2 : 5;
+ unsigned dest : 5;
+ unsigned offset : 16;
+};
+
+struct esc_fmt
+{
+ unsigned op1 : 6;
+ unsigned res1 : 10;
+ unsigned src1 : 5;
+ unsigned res2 : 6;
+ unsigned op2 : 5;
+};
+struct ctrl_fmt
+{
+ unsigned op1 : 6;
+ unsigned int offset : 26;
+};
+
+struct fp_fmt
+{
+ unsigned op1 : 6;
+ unsigned src2 : 5;
+ unsigned dest : 5;
+ unsigned src1 : 5;
+ unsigned p : 1;
+ unsigned d : 1;
+ unsigned s : 1;
+ unsigned r : 1;
+ unsigned op2 : 7;
+};
+
+union insn_fmt
+{
+ struct gen_fmt gen;
+ struct geni_fmt geni;
+ struct esc_fmt esc;
+ struct ctrl_fmt ctrl;
+ struct fp_fmt fp;
+ long int_val;
+};
+#else
+struct gen_fmt
+{
+ unsigned offset : 11;
+ unsigned src1 : 5;
+ unsigned dest : 5;
+ unsigned src2 : 5;
+ unsigned op1 : 6;
+};
+
+struct geni_fmt
+{
+ unsigned offset : 16;
+ unsigned dest : 5;
+ unsigned src2 : 5;
+ unsigned op1 : 6;
+};
+
+struct esc_fmt
+{
+ unsigned op2 : 5;
+ unsigned res2 : 6;
+ unsigned src1 : 5;
+ unsigned res1 : 10;
+ unsigned op1 : 6;
+};
+struct ctrl_fmt
+{
+ unsigned int offset : 26;
+ unsigned op1 : 6;
+};
+
+struct fp_fmt
+{
+ unsigned op2 : 7;
+ unsigned r : 1;
+ unsigned s : 1;
+ unsigned d : 1;
+ unsigned p : 1;
+ unsigned src1 : 5;
+ unsigned dest : 5;
+ unsigned src2 : 5;
+ unsigned op1 : 6;
+};
+
+union insn_fmt
+{
+ struct gen_fmt gen;
+ struct geni_fmt geni;
+ struct esc_fmt esc;
+ struct ctrl_fmt ctrl;
+ struct fp_fmt fp;
+ long int_val;
+};
+#endif
+
+typedef enum
+{
+ Error, not_branch, uncond, uncond_d, cond, cond_d
+} branch_type;
+
+
diff --git a/gdb/i860-pinsn.c b/gdb/i860-pinsn.c
new file mode 100644
index 0000000..29b1855
--- /dev/null
+++ b/gdb/i860-pinsn.c
@@ -0,0 +1,660 @@
+/* Print i860 instructions for GDB, the GNU debugger.
+ Copyright (C) 1992 Free Software Foundation, Inc.
+ Contributed by Peggy Fieland (pfieland@stratus.com)
+
+GDB is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY. No author or distributor accepts responsibility to anyone
+for the consequences of using it or for whether it serves any
+particular purpose or works at all, unless he says so in writing.
+Refer to the GDB General Public License for full details.
+
+Everyone is granted permission to copy, modify and redistribute GDB,
+but only under the conditions described in the GDB General Public
+License. A copy of this license is supposed to have been given to you
+along with GDB so you can know your rights and responsibilities. It
+should be in a file named COPYING. Among other things, the copyright
+notice and this notice must be preserved on all copies.
+
+In other words, go ahead and share GDB, but don't try to stop
+anyone else from sharing it farther. Help stamp out software hoarding!
+*/
+
+#include <stdio.h>
+
+#include "defs.h"
+#include "tm-i860.h"
+#include "i860-opcode.h"
+
+/* i860 instructions are never longer than this many bytes. */
+#define MAXLEN 4
+
+static int fp_instr();
+static void fld_offset();
+static void gen_rrr();
+static void gen_irr();
+static void ctrl_a();
+
+/*
+ * integer registers names
+ */
+static char *ireg[32] =
+ {
+ "r0", "r1", "sp", "fp", "r4", "r5", "r6", "r7",
+ "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15",
+ "r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
+ "r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31",
+ };
+
+/*
+ * Control registers of the ld.c and st.c instructions
+ */
+static char *ctlreg[32] =
+ {
+ "fir", "psr", "dirbase", "db", "fsr", "?", "?", "?",
+ "?", "?", "?", "?", "?", "?", "?", "?",
+ "?", "?", "?", "?", "?", "?", "?", "?",
+ "?", "?", "?", "?", "?", "?", "?", "?"
+ };
+
+
+/***********************************************************************
+ * Print the i860 instruction at address MEMADDR in debugged memory,
+ * on STREAM. Returns length of the instruction, in bytes, which
+ * is always 4.
+ */
+
+int
+print_insn (memaddr, stream)
+ CORE_ADDR memaddr; /* address of the instruction */
+ FILE *stream; /* stream on which to write result */
+{
+ union insn_fmt insn; /* the instruction we're decoding */
+ long offset; /* the (decoded) offset from the instruction */
+ long split_offset; /* the value of a ld/st-style split offset */
+ int ai; /* autoincrement flag */
+ char suffix; /* length suffix */
+
+ adj_read_memory (memaddr, &insn, MAXLEN);
+
+ /* These offsets used in ld, st, bte, etc. instructions and are formed by
+ * combining 2 separate fields within the instruction and sign-extending
+ * the result
+ */
+
+ split_offset = (insn.gen.dest << 11) | insn.gen.offset;
+ split_offset = SIGN_EXT(16, split_offset);
+
+
+ switch (insn.gen.op1)
+ {
+ case 000:
+ fprintf (stream, "ld.b %s(%s),%s", ireg[insn.gen.src1],
+ ireg[insn.gen.src2], ireg[insn.gen.dest]);
+ break;
+ case 001:
+ offset = SIGN_EXT(16, insn.geni.offset);
+ fprintf (stream, "ld.b 0x%x(%s),%s", offset,
+ ireg[insn.geni.src2], ireg[insn.geni.dest]);
+ break;
+ case 002:
+ fprintf (stream, "ixfr %s,f%d", ireg[insn.gen.src1], insn.gen.dest);
+ break;
+ case 003:
+ fprintf (stream, "st.b %s,0x%x(%s)", ireg[insn.gen.src1], split_offset,
+ ireg[insn.geni.src2]);
+ break;
+ case 004:
+ fprintf (stream, "ld.%c %s(%s),%s", (insn.gen.offset & 1) ? 'l' : 's',
+ ireg[insn.gen.src1], ireg[insn.gen.src2], ireg[insn.gen.dest]);
+ break;
+ case 005:
+ offset = SIGN_EXT(16, insn.geni.offset);
+ fprintf (stream, "ld.%c 0x%x(%s),%s", (insn.geni.offset & 1) ? 'l' : 's',
+ (offset & ~1), ireg[insn.geni.src2], ireg[insn.geni.dest]);
+ break;
+
+ case 007:
+ fprintf (stream, "st.%c %s,0x%x(%s)", (insn.geni.offset & 1) ? 'l' : 's',
+ ireg[insn.gen.src1], (split_offset & ~1), ireg[insn.geni.src2]);
+ break;
+
+ case 010:
+ offset = insn.gen.offset;
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "fld.%c %s(%s)%s,f%d", suffix,
+ ireg[insn.gen.src1], ireg[insn.gen.src2], ai ? "++" : "",
+ insn.gen.dest);
+ break;
+ case 011:
+ offset = SIGN_EXT(16, insn.geni.offset);
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "fld.%c 0x%x(%s)%s,f%d", suffix,
+ offset, ireg[insn.gen.src2], ai ? "++" : "", insn.gen.dest);
+ break;
+ case 012:
+ offset = insn.gen.offset;
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "fst.%c f%d,%s(%s)%s", suffix,
+ insn.gen.dest, ireg[insn.gen.src1], ireg[insn.gen.src2],
+ ai ? "++" : "");
+ break;
+ case 013:
+ offset = SIGN_EXT(16, insn.geni.offset);
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "fst.%c f%d,0x%x(%s)%s", suffix,
+ insn.gen.dest, offset, ireg[insn.gen.src2], ai ? "++" : "");
+ break;
+ case 014:
+ fprintf (stream, "ld.c %s,%s", ctlreg[insn.gen.src2],
+ ireg[insn.gen.dest]);
+ break;
+ case 015:
+ offset = SIGN_EXT(16, insn.geni.offset);
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "flush 0x%x(%s)%s", offset, ireg[insn.gen.src2],
+ ai ? "++" : "");
+ break;
+ case 016:
+ fprintf (stream, "st.c %s,%s", ireg[insn.gen.src1],
+ ctlreg[insn.gen.src2]);
+ break;
+ case 017:
+ offset = SIGN_EXT(16, insn.geni.offset);
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "pst.d f%d,0x%x(%s)%s", insn.gen.dest,
+ offset, ireg[insn.gen.src2], ai ? "++" : "");
+ break;
+
+ case 020:
+ fprintf (stream, "bri %s", ireg[insn.gen.src1]);
+ break;
+ case 021:
+ gen_rrr("trap", insn, stream);
+ break;
+ case 022:
+ /*
+ * Floating-point Opcodes
+ */
+ if (!fp_instr(insn.fp, stream))
+ fprintf (stream, "0x%08x (invalid instruction)", insn.int_val);
+ break;
+ case 023:
+ /*
+ * Core Escape Opcodes
+ */
+ switch (insn.esc.op2)
+ {
+ case 1:
+ fprintf (stream, "lock");
+ break;
+ case 2:
+ fprintf (stream, "calli %s", ireg[insn.esc.src1]);
+ break;
+ case 4:
+ fprintf (stream, "intovr");
+ break;
+ case 7:
+ fprintf (stream, "unlock");
+ break;
+ default:
+ fprintf (stream, "0x%08x (invalid instruction)", insn.int_val);
+ break;
+ }
+ break;
+
+ case 024:
+ fprintf (stream, "btne %s,%s,", ireg[insn.gen.src1],
+ ireg[insn.gen.src2]);
+ offset = split_offset << 2;
+ print_address ((CORE_ADDR) (memaddr + 4 + offset), stream);
+ break;
+ case 025:
+ fprintf (stream, "btne 0x%x,%s,", insn.gen.src1, ireg[insn.gen.src2]);
+ offset = split_offset << 2;
+ print_address ((CORE_ADDR) (memaddr + 4 + offset), stream);
+ break;
+ case 026:
+ fprintf (stream, "bte %s,%s,", ireg[insn.gen.src1],
+ ireg[insn.gen.src2]);
+ offset = split_offset << 2;
+ print_address ((CORE_ADDR) (memaddr + 4 + offset), stream);
+ break;
+ case 027:
+ fprintf (stream, "bte 0x%x,%s,", insn.gen.src1, ireg[insn.gen.src2]);
+ offset = split_offset << 2;
+ print_address ((CORE_ADDR) (memaddr + 4 + offset), stream);
+ break;
+
+ case 030:
+ offset = insn.gen.offset;
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "pfld.%c %s(%s)%s,f%d", suffix,
+ ireg[insn.gen.src1], ireg[insn.gen.src2], ai ? "++" : "",
+ insn.gen.dest);
+ break;
+ case 031:
+ offset = SIGN_EXT(16, insn.geni.offset);
+ fld_offset(&offset, &suffix, &ai);
+
+ fprintf (stream, "pfld.%c 0x%x(%s)%s,f%d", suffix,
+ offset, ireg[insn.gen.src2], ai ? "++" : "", insn.gen.dest);
+ break;
+ case 032:
+ ctrl_a("br", insn, memaddr, stream);
+ break;
+ case 033:
+ ctrl_a("call", insn, memaddr, stream);
+ break;
+ case 034:
+ ctrl_a("bc", insn, memaddr, stream);
+ break;
+ case 035:
+ ctrl_a("bc.t", insn, memaddr, stream);
+ break;
+ case 036:
+ ctrl_a("bnc", insn, memaddr, stream);
+ break;
+ case 037:
+ ctrl_a("bnc.t", insn, memaddr, stream);
+ break;
+
+ case 040:
+ gen_rrr("addu", insn, stream);
+ break;
+ case 041:
+ gen_irr("addu", insn, SIGN_EXT(16, insn.geni.offset), stream);
+ break;
+ case 042:
+ gen_rrr("subu", insn, stream);
+ break;
+ case 043:
+ gen_irr("subu", insn, SIGN_EXT(16, insn.geni.offset), stream);
+ break;
+ case 044:
+ gen_rrr("adds", insn, stream);
+ break;
+ case 045:
+ gen_irr("adds", insn, SIGN_EXT(16, insn.geni.offset), stream);
+ break;
+ case 046:
+ gen_rrr("subs", insn, stream);
+ break;
+ case 047:
+ gen_irr("subs", insn, SIGN_EXT(16, insn.geni.offset), stream);
+ break;
+
+ case 050:
+ if (insn.gen.src1 == 0)
+ {
+ if (insn.gen.src2 == 0 && insn.gen.dest == 0)
+ fprintf (stream, "nop");
+ else
+ fprintf (stream, "mov %s,%s", ireg[insn.gen.src2],
+ ireg[insn.gen.dest]);
+ }
+ else
+ gen_rrr("shl", insn, stream);
+ break;
+ case 051:
+ gen_irr("shl", insn, insn.geni.offset, stream);
+ break;
+ case 052:
+ gen_rrr("shr", insn, stream);
+ break;
+ case 053:
+ gen_irr("shr", insn, insn.geni.offset, stream);
+ break;
+ case 054:
+ if (insn.gen.src1 == 0 && insn.gen.src2 == 0 && insn.gen.dest == 0)
+ {
+ if ((insn.int_val & (1 << 9)) != 0)
+ fprintf (stream, "d.");
+ fprintf (stream, "fnop");
+ }
+ else
+ gen_rrr("shrd", insn, stream);
+ break;
+ case 055:
+ fprintf (stream, "bla %s,%s,", ireg[insn.gen.src1],
+ ireg[insn.gen.src2]);
+ offset = split_offset << 2;
+ print_address ((CORE_ADDR) (memaddr + 4 + offset), stream);
+ break;
+ case 056:
+ gen_rrr("shra", insn, stream);
+ break;
+ case 057:
+ gen_irr("shra", insn, insn.geni.offset, stream);
+ break;
+
+ case 060:
+ gen_rrr("and", insn, stream);
+ break;
+ case 061:
+ gen_irr("and", insn, insn.geni.offset, stream);
+ break;
+
+ case 063:
+ gen_irr("andh", insn, insn.geni.offset, stream);
+ break;
+ case 064:
+ gen_rrr("andnot", insn, stream);
+ break;
+ case 065:
+ gen_irr("andnot", insn, insn.geni.offset, stream);
+ break;
+
+ case 067:
+ gen_irr("andnoth", insn, insn.geni.offset, stream);
+ break;
+
+ case 070:
+ gen_rrr("or", insn, stream);
+ break;
+ case 071:
+ gen_irr("or", insn, insn.geni.offset, stream);
+ break;
+
+ case 073:
+ gen_irr("orh", insn, insn.geni.offset, stream);
+ break;
+ case 074:
+ gen_rrr("xor", insn, stream);
+ break;
+ case 075:
+ gen_irr("xor", insn, insn.geni.offset, stream);
+ break;
+
+ case 077:
+ gen_irr("xorh", insn, insn.geni.offset, stream);
+ break;
+
+ default:
+ fprintf (stream, "0x%08x (invalid instruction)", insn.int_val);
+ break;
+ }
+
+ return(4);
+}
+
+/* A full list of floating point opcodes - if the entry is NULL, there is
+ * no corresponding instruction
+ */
+
+static char *fp_ops[] =
+ {
+ "r2p1", "r2pt", "r2ap1", "r2apt",
+ "i2p1", "i2pt", "i2ap1", "i2apt",
+ "rat1p2", "m12apm", "ra1p2", "m12ttpa",
+ "iat1p2", "m12tpm", "ia1p2", "m12tpa",
+
+ "r2s1", "r2st", "r2as1", "r2ast",
+ "i2s1", "i2st", "i2as1", "i2ast",
+ "rat1s2", "m12asm", "ra1s2", "m12ttsa",
+ "iat1s2", "m12tsm", "ia1s2", "m12tsa",
+
+ "fmul", "fmlow", "frcp", "frsqr",
+ "fmul3", NULL, NULL, NULL,
+ NULL, NULL, NULL, NULL,
+ NULL, NULL, NULL, NULL,
+
+ "fadd", "fsub", "fix", "famov",
+ "fgt", "feq", NULL, NULL,
+ NULL, NULL, "ftrunc", NULL,
+ NULL, NULL, NULL, NULL,
+
+ "fxfr", NULL, NULL, NULL,
+ NULL, NULL, NULL, NULL,
+ NULL, "fiadd", NULL, NULL,
+ NULL, "fisub", NULL, NULL,
+
+ "faddp", "faddz", NULL, NULL,
+ NULL, NULL, NULL, "fzchkl",
+ NULL, NULL, "form", NULL,
+ NULL, NULL, NULL, "fzchks",
+ };
+
+/* Alternate list of floating point opcodes for PFMAM/PFMSM instructions
+ */
+
+static char *alt_fp_ops[] =
+ {
+ "mr2p1", "mr2pt", "mr2mp1", "mr2mpt",
+ "mi2p1", "mi2pt", "mi2mp1", "mi2mpt",
+ "mrmt1p2", "mm12mpm", "mrm1p2", "mm12ttpm",
+ "mimt1p2", "mm12tpm", "mim1p2", "mm12tpm",
+
+ "mr2s1", "mr2st", "mr2ms1", "mr2mst",
+ "mi2s1", "mi2st", "mi2ms1", "mi2mst",
+ "mrmt1s2", "mm12msm", "mrm1s2", "mm12ttsm",
+ "mimt1s2", "mm12tsm", "mim1s2", "mm12tsm",
+ };
+
+
+/* Floating point precision suffix values - indexed by s and r bits of
+ * instructions.
+ */
+
+static char precision[2] =
+ {
+ 's', 'd',
+ };
+
+/***********************************************************************
+ * Print floating-point instruction 'insn' on the indicated stream
+ * Returns 1 if successful, 0 on failure (invalid instruction)
+ */
+
+static int
+fp_instr(insn, stream)
+ struct fp_fmt insn; /* instruction to decode */
+ FILE *stream; /* stream to print on */
+{
+ char *name; /* the opcode name */
+
+ name = fp_ops[insn.op2];
+ if (name && insn.d)
+ fprintf(stream, "d.");
+
+
+ if (insn.op2 < 0x20)
+ {
+ /*
+ * DPC Ops
+ */
+ if (insn.p == 0) /* use PFMAM/PFMSM ops if p=0 */
+ name = alt_fp_ops[insn.op2];
+
+ fprintf (stream, "%s.%c%c f%d,f%d,f%d", name,
+ precision[insn.s], precision[insn.r],
+ insn.src1, insn.src2, insn.dest);
+ }
+ else
+ {
+ switch (insn.op2)
+ {
+ case 0x21: /* fmlow (no pipeline allowed) */
+ fprintf (stream, "%s.%c%c f%d,f%d,f%d", name,
+ precision[insn.s], precision[insn.r],
+ insn.src1, insn.src2, insn.dest);
+ break;
+
+ case 0x22: /* frcp */
+ case 0x23: /* fsqrt */
+ fprintf (stream, "%s.%c%c f%d,f%d", name,
+ precision[insn.s], precision[insn.r],
+ insn.src2, insn.dest);
+ break;
+
+ case 0x24: /* pfmul3 */
+ fprintf (stream, "pfmul3.dd f%d,f%d,f%d",
+ insn.src1, insn.src2, insn.dest);
+ break;
+
+ case 0x30: /* fadd */
+ case 0x49: /* fiadd */
+ if (insn.src2 == 0)
+ {
+ /*
+ * Really fmov
+ */
+ fprintf (stream, "%sfmov.%c%c f%d,f%d", insn.p ? "p" : "",
+ precision[insn.s], precision[insn.r],
+ insn.src1, insn.dest);
+ }
+ else
+ {
+ fprintf (stream, "%s%s.%c%c f%d,f%d,f%d", insn.p ? "p" : "", name,
+ precision[insn.s], precision[insn.r],
+ insn.src1, insn.src2, insn.dest);
+ }
+ break;
+
+ case 0x32: /* fix */
+ case 0x3A: /* ftrunc */
+ fprintf (stream, "%s%s.%c%c f%d,f%d", insn.p ? "p" : "", name,
+ precision[insn.s], precision[insn.r],
+ insn.src1, insn.dest);
+ break;
+
+ case 0x34: /* pfgt/pfle */
+ if (insn.r)
+ name = "fle";
+ fprintf (stream, "p%s.%c%c f%d,f%d,f%d", name,
+ precision[insn.s], precision[insn.s],
+ insn.src1, insn.src2, insn.dest);
+ break;
+
+ case 0x35: /* pfeq */
+ fprintf (stream, "pfeq.%c%c f%d,f%d,f%d",
+ precision[insn.s], precision[insn.r],
+ insn.src1, insn.src2, insn.dest);
+ break;
+
+ case 0x40: /* fxfr */
+ fprintf (stream, "fxfr f%d,%s", insn.src1, ireg[insn.dest]);
+ break;
+
+ case 0x50: /* faddp */
+ case 0x51: /* faddz */
+ case 0x57: /* fzchkl */
+ case 0x5F: /* fzchks */
+ /*
+ * Graphics ops with no precision
+ */
+ fprintf (stream, "%s%s f%d,f%d,f%d", insn.p ? "p" : "", name,
+ insn.src1, insn.src2, insn.dest);
+ break;
+
+ case 0x5A: /* form */
+ fprintf (stream, "%sform f%d,f%d", insn.p ? "p" : "",
+ insn.src1, insn.dest);
+ break;
+
+ default:
+ /*
+ * All the rest are uniform 3-address, optionally pipelined, etc
+ */
+ if (name)
+ fprintf (stream, "%s%s.%c%c f%d,f%d,f%d", insn.p ? "p" : "", name,
+ precision[insn.s], precision[insn.r],
+ insn.src1, insn.src2, insn.dest);
+ else
+ return (0);
+ break;
+ }
+ }
+ return (1);
+}
+
+/***********************************************************************
+ * Decode fld/fst-style offset encodings into actual offset, precision suffix,
+ * and autoincrement flag
+ */
+
+static void
+fld_offset(offset, suffix, autoincrement)
+ long *offset; /* original and returned offset */
+ char *suffix; /* returned suffix character */
+ int *autoincrement; /* autoincrement flag (1 if ai) */
+{
+ long off = *offset; /* working copy of *offset */
+
+ *autoincrement = ((off & 1) != 0);
+
+ if (off & 2)
+ {
+ *suffix = 'l';
+ *offset = (off & ~3);
+ }
+ else if (off & 4)
+ {
+ *suffix = 'q';
+ *offset = (off & ~7);
+ }
+ else
+ {
+ *suffix = 'd';
+ *offset = (off & ~7);
+ }
+}
+
+/***********************************************************************
+ * Print a general format instruction of the three register form:
+ * op rx,ry,rz
+ */
+
+static void
+gen_rrr(name, insn, stream)
+ char *name;
+ union insn_fmt insn;
+ FILE *stream;
+{
+ fprintf (stream, "%s %s,%s,%s", name, ireg[insn.gen.src1],
+ ireg[insn.gen.src2], ireg[insn.gen.dest]);
+}
+
+/***********************************************************************
+ * Print a general format instruction of the immed + two register form:
+ * op i,ry,rz
+ */
+
+static void
+gen_irr(name, insn, immed, stream)
+ char *name;
+ union insn_fmt insn;
+ long immed;
+ FILE *stream;
+{
+ fprintf (stream, "%s 0x%x,%s,%s", name, immed,
+ ireg[insn.gen.src2], ireg[insn.gen.dest]);
+}
+
+/***********************************************************************
+ * Print a ctrl format instruction with a 26-bit displacement:
+ * op addr
+ */
+
+static void
+ctrl_a(name, insn, memaddr, stream)
+ char *name;
+ union insn_fmt insn;
+ CORE_ADDR memaddr;
+ FILE *stream;
+{
+ long offset;
+
+ fprintf (stream, "%s ", name);
+ offset = SIGN_EXT(28, insn.ctrl.offset << 2);
+
+ print_address ((CORE_ADDR) (memaddr + 4 + offset), stream);
+}
+
+
diff --git a/gdb/i860-tdep.c b/gdb/i860-tdep.c
new file mode 100644
index 0000000..2742fbd
--- /dev/null
+++ b/gdb/i860-tdep.c
@@ -0,0 +1,1636 @@
+/* Machine-dependent code which would otherwise be in inflow.c and core.c,
+ for GDB, the GNU debugger.
+ Copyright (C) 1992 Free Software Foundation, Inc.
+ This code is for the i860 cpu.
+ Contributed by Peggy Fieland (pfieland@stratus.com)
+
+ GDB is distributed in the hope that it will be useful, but WITHOUT ANY
+ WARRANTY. No author or distributor accepts responsibility to anyone
+ for the consequences of using it or for whether it serves any
+ particular purpose or works at all, unless he says so in writing.
+ Refer to the GDB General Public License for full details.
+
+ Everyone is granted permission to copy, modify and redistribute GDB,
+ but only under the conditions described in the GDB General Public
+ License. A copy of this license is supposed to have been given to you
+ along with GDB so you can know your rights and responsibilities. It
+ should be in a file named COPYING. Among other things, the copyright
+ notice and this notice must be preserved on all copies.
+
+ In other words, go ahead and share GDB, but don't try to stop
+ anyone else from sharing it farther. Help stamp out software hoarding!
+ */
+
+#include "defs.h"
+#include "tm-i860.h"
+#include "frame.h"
+#include "inferior.h"
+#include "obstack.h"
+#include "symtab.h"
+#include "value.h"
+
+#include "tm-i860.h"
+#include "i860-opcode.h"
+
+#include <stdio.h>
+#include "break.h"
+
+#ifdef notdef
+
+#include <sys/types.h>
+#include <sys/param.h>
+#include <sys/dir.h>
+
+#endif
+
+#include <signal.h>
+#include <sys/ioctl.h>
+#include <fcntl.h>
+
+/* #include <sys/reg.h> */
+#include "i860_reg.h"
+
+#include <a.out.h>
+#include <sys/file.h>
+#include <sys/stat.h>
+#include <core.h>
+
+#include <sys/user.h>
+
+#include <elf.h>
+#include <sys/elftypes.h>
+#include <sys/elf_860.h>
+#include <libelf.h>
+
+
+int btdebug = 0; /* change value to 1 to enable debugging code */
+
+#define BTDEBUG if (btdebug) btdebug_message
+
+#include <stdarg.h>
+
+int read_memory();
+int write_memory();
+
+btdebug_message(char *format, ...)
+{
+ va_list arglist;
+ va_start( arglist, format );
+
+ if( btdebug )
+ vfprintf (stderr, format, arglist );
+ va_end ( arglist );
+}
+
+extern int errno;
+extern int attach_flag;
+
+
+/* This is used when GDB is exiting. It gives less chance of error.*/
+
+
+/* Simulate single-step ptrace call for sun4. Code written by Gary
+ Beihl (beihl@mcc.com). */
+/* Modified for i860 by Jim Hanko (hanko@orc.olivetti.com) */
+
+
+static struct breakpoint brk;
+typedef char binsn_quantum[sizeof break_insn];
+static binsn_quantum break_mem[2];
+
+/* Non-zero if we just simulated a single-step ptrace call. This is
+ needed because we cannot remove the breakpoints in the inferior
+ process until after the `wait' in `wait_for_inferior'. Used for
+ i860. */
+
+int one_stepped;
+void
+ single_step (signal)
+int signal;
+{
+ CORE_ADDR pc;
+ branch_type place_brk();
+
+ pc = read_register (PC_REGNUM);
+
+ if (!one_stepped)
+ {
+ brk.address = pc;
+ place_brk (pc, SINGLE_STEP_MODE, &brk);
+ brk.shadow_contents[0] = brk.shadow_contents[1] = 0;
+ brk.shadow_contents[2] = brk.shadow_contents[3] = 0;
+
+ if (brk.mode == DIM)
+ {
+ if ( brk.address1 )
+ {
+ printf(" DIM1 ->");
+ printf(" %x : ", brk.act_addr[3]);
+ print_insn( brk.act_addr[3], stdout);
+ printf("\t -|- ");
+ printf(" %x : ", brk.act_addr[2]);
+ print_insn( brk.act_addr[2], stdout);
+ printf("\n");
+ fflush(stdout);
+
+ adj_read_memory (brk.act_addr[2], &brk.shadow_contents[2], 4);
+ adj_write_memory (brk.act_addr[2], break_insn, 4);
+ adj_read_memory (brk.act_addr[3], &brk.shadow_contents[3], 4);
+ /* adj_write_memory (brk.act_addr[3], float_insn, 4); */
+
+ }
+ if ( brk.address1)
+ printf(" DIM2 ->");
+ else
+ printf(" DIM1 ->");
+
+ printf(" %x : ", brk.act_addr[1]);
+ print_insn( brk.act_addr[1], stdout);
+ printf("\t -|- ");
+ printf(" %x : ", brk.act_addr[0]);
+ print_insn( brk.act_addr[0], stdout);
+ printf("\n");
+ fflush(stdout);
+
+ adj_read_memory (brk.act_addr[0], &brk.shadow_contents[0], 4);
+ adj_write_memory (brk.act_addr[0], break_insn, 4);
+ adj_read_memory (brk.act_addr[1], &brk.shadow_contents[1], 4);
+ /* adj_write_memory (brk.act_addr[1], float_insn, 4); */
+
+ }
+ else {
+ if (brk.address1)
+ {
+ if (btdebug)
+ {
+ printf(" SIM1 ->");
+ printf(" %x : ", brk.act_addr[2]);
+ print_insn( brk.act_addr[2], stdout);
+ printf("\n");
+ fflush(stdout);
+ }
+ adj_read_memory (brk.act_addr[2], &brk.shadow_contents[2], 4);
+ adj_write_memory (brk.act_addr[2], break_insn, 4);
+ }
+ if (btdebug)
+ {
+ if ( brk.address1)
+ printf(" SIM2 ->");
+ else
+ printf(" SIM1 ->");
+
+ printf(" %x : ", brk.act_addr[0]);
+ print_insn( brk.act_addr[0], stdout);
+ printf("\n");
+ fflush(stdout);
+ }
+ adj_read_memory (brk.act_addr[0], &brk.shadow_contents[0], 4);
+ adj_write_memory (brk.act_addr[0], break_insn, 4);
+ }
+
+ /* Let it go */
+ one_stepped = 1;
+ return;
+ }
+ else
+ {
+ /* Remove breakpoints */
+ if (brk.mode == DIM)
+ {
+ adj_write_memory (brk.act_addr[0], &brk.shadow_contents[0], 4);
+ adj_write_memory (brk.act_addr[1], &brk.shadow_contents[1], 4);
+ } else {
+ adj_write_memory (brk.act_addr[0], &brk.shadow_contents[0], 4);
+ }
+
+ if (brk.address1)
+ {
+ if (brk.mode == DIM)
+ {
+ adj_write_memory (brk.act_addr[2], &brk.shadow_contents[2], 4);
+ adj_write_memory (brk.act_addr[3], &brk.shadow_contents[3], 4);
+ } else {
+ adj_write_memory (brk.act_addr[2], &brk.shadow_contents[2], 4);
+ }
+ }
+ one_stepped = 0;
+ }
+}
+
+
+
+
+/* return nonzero if the routine containing pc has been
+ * compiled with -g. We assume -g if the first instruction is
+ * an addu|adds -X,sp and the second is st.l fp,XX(sp)
+ *
+ * based on skip_prologue();
+ */
+
+g_routine(pc)
+ CORE_ADDR pc;
+{
+ long instr;
+ int regno;
+ CORE_ADDR top_pc;
+
+ top_pc = get_pc_function_start(pc);
+ if (top_pc)
+ {
+ instr = adj_read_memory_integer (top_pc);
+ /* Recognize "addu|adds -X,sp,sp" insn. */
+
+ if ((instr & 0xEFFF0000) == 0x84420000)
+ {
+ top_pc += 4;
+ instr = adj_read_memory_integer (top_pc);
+
+ if( (instr & 0xFFE0F801) == 0x1C401801 ) /* st.l fp,X(sp) */
+ return(1);
+ }
+ }
+ return(0);
+}
+
+
+/* Written for i860 by Jim Hanko (hanko@orc.olivetti.com) */
+/* This code was based on SPARC code written by Gary Beihl (beihl@mcc.com),
+ by Michael Tiemann (tiemann@corto.inria.fr). */
+
+struct command_line *get_breakpoint_commands ();
+
+CORE_ADDR
+ skip_prologue (pc)
+CORE_ADDR pc;
+{
+ long instr;
+ int regno;
+
+ instr = adj_read_memory_integer (pc);
+
+ /* Recognize "addu|adds -X,sp,sp" insn. */
+ if ((instr & 0xEFFF0000) == 0x84420000)
+ {
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ else
+ return(pc); /* No frame! */
+
+ /* Recognize store of return addr and frame pointer into frame */
+ while (1)
+ {
+ if ((instr & 0xFFE0F801) == 0x1C400801 || /* st.l r1,X(sp) */
+ (instr & 0xFFE0F801) == 0x1C401801) /* st.l fp,X(sp) */
+ {
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ else
+ break;
+ }
+
+ /* Recognize "addu|adds X,sp,fp" insn. */
+ if ((instr & 0xEFFF0000) == 0x84430000)
+ {
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+
+ /* Now recognize stores into the frame from the registers. */
+
+ while (1)
+ {
+ if ((instr & 0xFFA00003) == 0x1C200001 || /* st.l rn,X(fp|sp) */
+ (instr & 0xFFA00001) == 0x4C200000) /* fst.y fn,X(fp|sp) */
+ {
+ regno = (instr >> 11) & 0x1f;
+ if (regno == 0) /* source reg == 0? quit */
+ break;
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ else
+ break;
+ }
+
+ return(pc);
+}
+
+
+/* Set *nextpc to branch target if we find a branch. If it is not a branch,
+ set it to the next instruction (addr + 4) */
+
+
+branch_type
+ isabranch (addr, nextpc)
+CORE_ADDR addr, *nextpc;
+{
+ long instr;
+ branch_type val = not_branch;
+ long offset; /* Must be signed for sign-extend */
+
+ printf(" isabranch\n");
+ *nextpc = addr;
+ instr = adj_read_memory_integer (addr);
+
+ if ((instr & 0xE0000000) == 0x60000000 && /* CTRL format */
+ (instr & 0xF8000000) != 0x60000000) /* not pfld.y */
+ {
+ if ((instr & 0xF8000000) == 0x68000000) /* br or call */
+ val = uncond_d;
+ else if ((instr & 0xF4000000) == 0x74000000) /* bc.t or bnc.t */
+ val = cond_d;
+ else if ((instr & 0xF4000000) == 0x70000000) /* bc or bnc */
+ val = cond;
+
+ offset = (instr & 0x03ffffff);
+ if (offset & 0x02000000) /* sign extend? */
+ offset |= 0xFC000000;
+ *nextpc = addr + 4 + (offset << 2);
+ }
+ else if ((instr & 0xFC00003F) == 0x4C000002 || /* calli */
+ (instr & 0xFC000000) == 0x40000000) /* bri */
+ {
+ val = uncond_d;
+ offset = ((instr & 0x0000F800) >> 11);
+ *nextpc = (read_register(offset) & 0xFFFFFFFC);
+ }
+ else if ((instr & 0xF0000000) == 0x50000000) /* bte or btne */
+ {
+ val = cond;
+
+ offset = SIGN_EXT16(((instr & 0x001F0000) >> 5) | (instr & 0x000007FF));
+ *nextpc = addr + 4 + (offset << 2);
+ }
+ else if ((instr & 0xFC000000) == 0xB4000000) /* bla */
+ {
+ val = cond_d;
+
+ offset = SIGN_EXT16(((instr & 0x001F0000) >> 5) | (instr & 0x000007FF));
+ *nextpc = addr + 4 + (offset << 2);
+ }
+
+ printf(" Final addr - %x\n", *nextpc);
+ /*printf("isabranch ret: %d\n",val); */
+ return val;
+}
+
+/* set in call_function() [valops.c] to the address of the "call dummy" code
+ so dummy frames can be easily recognized; also used in wait_for_inferior()
+ [infrun.c]. When not used, it points into the ABI's 'reserved area' */
+
+CORE_ADDR call_dummy_set = 0; /* true if dummy call being done */
+CORE_ADDR call_dummy_start; /* address of call dummy code */
+
+frame_find_saved_regs(frame_info, frame_saved_regs)
+ struct frame_info *frame_info;
+ struct frame_saved_regs *frame_saved_regs;
+{
+ register CORE_ADDR pc;
+ long instr, spdelta = 0, offset;
+ int i, size, reg;
+ int r1_off = -1, fp_off = -1;
+ int framesize;
+
+ bzero (frame_saved_regs, sizeof(*frame_saved_regs));
+
+ if (call_dummy_set && frame_info->pc >= call_dummy_start &&
+ frame_info->pc <= call_dummy_start + CALL_DUMMY_LENGTH)
+ {
+ /* DUMMY frame - all registers stored in order at fp; old sp is
+ at fp + NUM_REGS*4 */
+
+ for (i = 1; i < NUM_REGS; i++) /* skip reg 0 */
+ if (i != SP_REGNUM && i != FP0_REGNUM && i != FP0_REGNUM + 1)
+ frame_saved_regs->regs[i] = frame_info->frame + i*4;
+
+ frame_saved_regs->regs[SP_REGNUM] = frame_info->frame + NUM_REGS*4;
+
+ call_dummy_set = 0;
+ return;
+ }
+
+ pc = get_pc_function_start (frame_info->pc);
+ if (pc)
+ {
+ instr = adj_read_memory_integer (pc);
+ /* Recognize "addu|adds -X,sp,sp" insn. */
+ if ((instr & 0xEFFF0000) == 0x84420000)
+ {
+ framesize = -SIGN_EXT16(instr & 0x0000FFFF);
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ }
+ else
+ goto punt; /* No frame! */
+
+ /* Recognize store of return addr and frame pointer into frame */
+ while (1)
+ {
+ if ((instr & 0xFFE0F801) == 0x1C400801) /* st.l r1,X(sp) */
+ {
+ r1_off = SIGN_EXT16(((instr&0x001F0000) >> 5) | (instr&0x000007FE));
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ else if ((instr & 0xFFE0F801) == 0x1C401801) /* st.l fp,X(sp) */
+ {
+ fp_off = SIGN_EXT16(((instr&0x001F0000) >> 5) | (instr&0x000007FE));
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ else
+ break;
+ }
+
+ /* Recognize "addu|adds X,sp,fp" insn. */
+ if ((instr & 0xEFFF0000) == 0x84430000)
+ {
+ spdelta = SIGN_EXT16(instr & 0x0000FFFF);
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+
+ /* Now recognize stores into the frame from the registers. */
+
+ while (1)
+ {
+ if ((instr & 0xFFC00003) == 0x1C400001) /* st.l rn,X(fp|sp) */
+ {
+ offset = SIGN_EXT16(((instr&0x001F0000) >> 5) | (instr&0x000007FE));
+ reg = (instr >> 11) & 0x1F;
+ if (reg == 0)
+ break;
+ if ((instr & 0x00200000) == 0) /* was this using sp? */
+ if (spdelta) /* and we know sp-fp delta */
+ offset -= spdelta; /* if so, adjust the offset */
+ else
+ break; /* if not, give up */
+
+
+ /* Handle the case where the return address is stored after the fp
+ is adjusted */
+
+ if (reg == 1)
+ frame_saved_regs->regs[PC_REGNUM] = frame_info->frame + offset;
+ else
+ frame_saved_regs->regs[reg] = frame_info->frame + offset;
+
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ else if ((instr & 0xFFC00001) == 0x2C400000) /* fst.y fn,X(fp|sp) */
+ {
+ /*
+ * The number of words in a floating store based on 3 LSB of instr
+ */
+ static int fst_sizes[] = {2, 0, 1, 0, 4, 0, 1, 0};
+
+ size = fst_sizes[instr & 7];
+ reg = ((instr >> 16) & 0x1F) + FP0_REGNUM;
+ if (reg == 0)
+ break;
+
+ if (size > 1) /* align the offset */
+ offset = SIGN_EXT16(instr & 0x0000FFF8); /* drop 3 bits */
+ else
+ offset = SIGN_EXT16(instr & 0x0000FFFC); /* drop 2 bits */
+
+ if ((instr & 0x00200000) == 0) /* was this using sp? */
+ if (spdelta) /* and we know sp-fp delta */
+ offset -= spdelta; /* if so, adjust the offset */
+ else
+ break; /* if not, give up */
+
+ for (i = 0; i < size; i++)
+ {
+ frame_saved_regs->regs[reg] = frame_info->frame + offset;
+
+ offset += 4;
+ reg++;
+ }
+
+ pc += 4;
+ instr = adj_read_memory_integer (pc);
+ }
+ else
+ break;
+ }
+
+ punt: ;
+ if (framesize != 0 && spdelta != 0)
+ frame_saved_regs->regs[SP_REGNUM] = frame_info->frame+(framesize-spdelta);
+ else
+ frame_saved_regs->regs[SP_REGNUM] = frame_info->frame + 8;
+
+ if (spdelta && fp_off != -1)
+ frame_saved_regs->regs[FP_REGNUM] = frame_info->frame - spdelta + fp_off;
+ else
+ frame_saved_regs->regs[FP_REGNUM] = frame_info->frame;
+
+ if (spdelta && r1_off != -1)
+ frame_saved_regs->regs[PC_REGNUM] = frame_info->frame - spdelta + r1_off;
+ else
+ frame_saved_regs->regs[PC_REGNUM] = frame_info->frame + 4;
+}
+
+/* return the stack offset where the fp register is stored */
+find_fp_offset(pc)
+{
+ int fp_off,i;
+ long instr;
+
+ /* look for the instruction and examine the offset */
+
+ for(i=4; i<16; i+=4){
+ instr = adj_read_memory_integer(pc+i);
+ if( (instr & 0xFFE0F801) == 0x1C401801) { /* st.l fp,X(sp) */
+
+ fp_off = SIGN_EXT16(((instr&0x001F0000) >> 5) |
+ (instr&0x000007FE));
+ return(fp_off);
+ }
+ }
+ return(0);
+}
+
+/* return the stack offset where r1 (return linkage ) register is stored */
+
+find_r1_offset(pc)
+{
+ int r1_off,i;
+ long instr;
+
+ /* look for the instruction and examine the offset */
+
+ for(i=4; i<16; i+=4){
+
+ instr = adj_read_memory_integer(pc+i);
+ if ((instr & 0xFFE0F801) == 0x1C400801) { /* st.l r1,X(sp) */
+
+ r1_off = SIGN_EXT16(((instr&0x001F0000) >> 5) |
+ (instr&0x000007FE));
+ return(r1_off);
+ }
+ }
+ return(-1);
+}
+
+
+/* dose routine starting at pc build a stack frame of any kind?? */
+has_a_frame(pc)
+{
+ if( skip_prologue(pc) != pc )return(1);
+ else return(0);
+}
+
+/* get the frame pointer of the caller.
+ * note that only routines that have been compiled with
+ * -g have full (XX)fp style stack frames
+ * if we are not returning to a non -g caller then we
+ * return the sp at entry to us as it is the caller's
+ * frame reference.
+ */
+
+frame_chain(thisframe)
+ FRAME thisframe;
+{
+ unsigned long fp, sp, pc;
+ unsigned long func_start;
+ long instr;
+ int offset;
+ unsigned long thisfp = thisframe->frame;
+
+ /* get the frame pointer actually sp for a non -g
+ * for the routine that called us routine
+ */
+
+ BTDEBUG("FRAME_CHAIN(%x)\n",thisframe);
+
+ if ( !read_memory_integer (thisframe->frame,4))
+ {
+ return (0);
+ }
+
+ if( ! g_routine(thisframe->pc) ){
+ BTDEBUG( "non g at %x\n",thisframe->pc);
+ caller_pc(thisframe->pc,thisframe->sp,&pc,&fp);
+ BTDEBUG("caller_pc returned %x %x \n",pc,fp);
+ return(fp);
+
+ }/* else a -g routine */
+
+
+ fp = read_memory_integer (thisfp, 4);
+
+ if (fp < thisfp || fp > STACK_END_ADDR)
+ {
+ /* handle the Metaware-type pseudo-frame */
+
+ func_start = get_pc_function_start(thisframe->pc);
+
+ if (func_start)
+ {
+
+ instr = adj_read_memory_integer (func_start);
+ /* Recognize "addu|adds -X,sp,sp" insn. */
+ if ((instr & 0xEFFF0000) == 0x84420000)
+ offset = SIGN_EXT16(instr & 0x0000FFFF);
+
+ }
+
+ fp = 0;
+ if (offset < 0)
+ fp = thisfp - offset;
+ }
+ BTDEBUG("frame_chain returned %d\n",fp);
+ return(fp);
+}
+
+/* get the pc and frame pointer (or sp )
+ * for the routine that called us
+ * when we (this_pc) is not within a -g routine
+ * if caller is non g we return sp for fp
+ */
+
+/* note this is written for Metaware version R2.1d compiler */
+
+caller_pc(this_pc,this_sp,to_pc,to_fp)
+ int * to_pc, *to_fp;
+ unsigned long this_pc,this_sp;
+{
+ unsigned long func_start;
+ int sp_offset,offset;
+ unsigned long sp,pc,fp,instr;
+
+ BTDEBUG("caller_pc %x sp = %x\n",this_pc,this_sp);
+
+ func_start = get_pc_function_start(this_pc);
+
+ BTDEBUG("caller_pc func_start %x\n", func_start);
+
+ if (func_start)
+ {
+ if( has_a_frame(func_start) ){
+
+ BTDEBUG("has_a_frame\n");
+
+ /* if our caller has a preamble and
+ * declares space for a stack frame
+ * then we must work to find our return address
+ */
+ instr = adj_read_memory_integer (func_start);
+ /* Recognize "addu|adds -X,sp,sp" insn. */
+
+ if ((instr & 0xEFFF0000) == 0x84420000)
+ sp_offset=SIGN_EXT16(instr&0x0000FFFF);
+ }
+ else { printf("error frame_chain\n");return(0); }
+
+ BTDEBUG("sp_offset = %d %x\n",sp_offset,sp_offset);
+
+ offset = find_r1_offset(func_start);
+
+ if( offset < 0 ){
+ printf("cant find return address for routine at %x\n",
+ func_start);
+ return(0);
+ }
+ pc = read_memory_integer(this_sp+offset,4);
+ sp= this_sp - sp_offset;
+
+ BTDEBUG("callers pc = %x sp = %x\n",pc,sp);
+
+ /* our caller a -g routine ?
+ * if he is we have to find his real fp
+ * else provide the sp as his fp
+ */
+
+ if( g_routine(pc) ){
+
+ BTDEBUG("caller_a_g\n");
+
+ if( ! (offset = find_fp_offset(func_start)) ) {
+ printf("error fp_offset\n");
+ return(0);
+ }
+ BTDEBUG("offset = %x %d\n",offset,offset);
+
+ fp = read_memory_integer(this_sp+offset,4);
+ *to_pc = CLEAN_PC(pc);
+ *to_fp = fp;
+ return(1);
+ }else
+ *to_pc = CLEAN_PC(pc);
+ *to_fp = sp;
+ return(1);
+ } else {
+/* pc = read_register(RP_REGNUM); */
+ pc = 0;
+ BTDEBUG("read_register pc %x\n",pc);
+ if( g_routine(pc) ){
+
+ *to_pc = CLEAN_PC(pc);
+ *to_fp = read_register(FP_REGNUM);
+ return(1);
+ }else {
+ *to_pc = CLEAN_PC(pc);
+ *to_fp = this_sp;
+ return(1);
+ }
+ }
+}
+
+int outside_startup_file();
+
+/* get the PC of the caller */
+frame_saved_pc(frame_struct)
+FRAME frame_struct;
+{
+ unsigned long frame;
+ unsigned long pc;
+ unsigned long pc1;
+ unsigned long sp;
+
+ CORE_ADDR fp;
+
+ CORE_ADDR rp;
+
+ frame = frame_struct->frame;
+ pc = frame_struct->pc;
+ BTDEBUG("frame_saved_pc input: frame %x, pc %x, sp %x ",
+ frame, pc, sp);
+
+ /* First see if this is the current frame. If it is, return the value in r1,
+ as it may not have been stored */
+
+ fp = read_register(FP_REGNUM);
+
+ /* check to see if we are in an entry sequence, where the return pointer has not yet been stored */
+ if (fp == frame && no_stored_rp(pc))
+ {
+ pc = read_register(RP_REGNUM);
+ frame_struct->rp = pc;
+ }
+ else if( ! g_routine(pc) )
+ {
+ caller_pc(pc,sp,&pc,&frame);
+ }
+ else
+ {
+
+ pc = read_memory_integer (frame + 4, 4);
+
+ if (!outside_startup_file(pc))
+ {
+
+ BTDEBUG("pc %x outside startup file \n",pc);
+
+ pc1 = read_memory_integer (frame, 4);
+
+ if (outside_startup_file(pc1))
+ pc = pc1;
+ else
+ pc = 0;
+ }
+ }
+ BTDEBUG(" returning pc %x\n", CLEAN_PC(pc));
+ return(CLEAN_PC(pc));
+
+ }
+
+/* Pass arguments to a function in the inferior process - ABI compliant
+ */
+
+pass_function_arguments(args, nargs, struct_return)
+ value *args;
+ int nargs;
+ int struct_return;
+{
+ int ireg = (struct_return) ? 17 : 16;
+ int freg = FP0_REGNUM + 8;
+ int i;
+ struct type *type;
+ value arg;
+ long tmp;
+ value value_arg_coerce();
+
+
+ for (i = 0; i < nargs; i++)
+ {
+ arg = value_arg_coerce(args[i]);
+ type = VALUE_TYPE(arg);
+ if (type == builtin_type_double)
+ {
+ write_register_bytes(REGISTER_BYTE(freg), VALUE_CONTENTS(arg), 8);
+ freg += 2;
+ }
+ else
+ {
+ bcopy(VALUE_CONTENTS(arg), &tmp, 4);
+ write_register(ireg, tmp);
+ ireg++;
+ }
+ }
+ if (ireg >= 28 || freg >= FP0_REGNUM + 16)
+ error("Too many arguments to function");
+}
+
+
+#define SPACES " "
+#define P_SPACES " "
+#define BYTE 0xff
+
+int screen_lines=24;
+
+char *spec_reg[] = {
+ "fsr", "db", "dirbase", "fir", "psr", "epsr",
+};
+
+char *doro_reg[] = {
+ "scp", "cbsp", "pt_cs", "intmsk", "intack",
+};
+#define NREGS 32
+
+
+get_reg(regno)
+{
+ char raw_buffer[32];
+ int addr;
+ int virtual_buffer;
+
+ read_relative_register_raw_bytes (regno, raw_buffer);
+ REGISTER_CONVERT_TO_VIRTUAL (addr, raw_buffer, &virtual_buffer);
+ return(virtual_buffer);
+}
+
+int jhdebug = 0;
+/*
+ ** Figure out address to place next breakpoint. Avoid tricky spots,
+ ** ie. delayed instruction slots etc.
+ ** Need to upgrade this later to allow delayed instruction breakpoints
+ ** with fix-up work done AFTER breakpoint.
+ ** Note that this routine DOES deal with dual instruction mode
+ */
+#define BIM 0x8008
+
+branch_type
+ place_brk (addr, mode, brk)
+CORE_ADDR addr;
+int mode;
+struct breakpoint *brk;
+{
+ long nextadr, prevadr, instr;
+ int val = not_branch;
+ long offset; /* Must be signed for sign-extend */
+ extern char registers[];
+ prevadr = nextadr = 0;
+
+ brk->address1 = 0;
+
+ if (mode == SINGLE_STEP_MODE)
+ {
+ if (INDIM || ENDIM)
+ {
+ nextadr = brk->address = (addr + 8);
+ instr = adj_read_memory_integer ((addr + 4));
+ brk->mode = DIM;
+ }
+ else
+ {
+ nextadr = brk->address = (addr + 4);
+ instr = adj_read_memory_integer (addr);
+ if (STDIM)
+ brk->mode = DIM;
+ else
+ brk->mode = SIM;
+ }
+
+
+ /*
+ ** For br/call one more sequential instruction gets executed and then we
+ ** continue at the current addr + offset. We are definitely going to
+ ** the dest. We are NOT allowed to place a breakpoint in the "delay"
+ ** slot - (the next sequential instruction) so we only place 1 breakpoint
+ ** at the destination.
+ ** For the bc/bnc the next instruction executed is EITHER the next sequential
+ ** or the destination of the branch, we therefore place 2 breakpoints one
+ ** at each location.
+ ** For the bc.t/bnc.t either 1 more sequential instruction is performed
+ ** followed by a branch (like br/call) OR we skip the sequential
+ ** instruction and keep going. We therefore place a breakpoint at the
+ ** destination of the branch AND the second sequential instruction after
+ ** the branch. Again a breakpoint is NOT allowed in the "delay slot"
+ */
+ if ((instr & 0xE0000000) == 0x60000000 && /* CTRL format */
+ (instr & 0xF8000000) != 0x60000000) /* not pfld.y */
+ {
+ if ((instr & 0xF8000000) == 0x68000000) /* br or call */
+ val = uncond_d;
+ else if ((instr & 0xF4000000) == 0x74000000) /* bc.t/bnc.t */
+ val = cond_d;
+ else if ((instr & 0xF4000000) == 0x70000000) /* bc or bnc */
+ val = cond;
+ offset = (instr & 0x03ffffff);
+ if (offset & 0x02000000) /*?sign extend*/
+ offset |= 0xFC000000;
+ if (val == uncond_d) /* br/call*/
+ prevadr = 0;
+ else if (val == cond_d) /* bc.t/bnc.t */
+ {
+ if ((INDIM) && !(ENDIM))
+ prevadr = nextadr + 8;
+ else
+ prevadr = nextadr + 4;
+ } else { /* bc /bnc */
+ if ((INDIM) && !(ENDIM))
+ prevadr = nextadr;
+ else
+ prevadr = nextadr;
+ }
+ nextadr += (offset << 2);
+ }
+ /*
+ ** We treat the bri/calli the same way as the br/call case.
+ */
+ else if ((instr & 0xFC00003F) == 0x4C000002 || /* calli */
+ (instr & 0xFC000000) == 0x40000000) /* bri */
+ {
+ val = uncond_d;
+ offset = ((instr & 0x0000F800) >> 11);
+ nextadr = (read_register(offset + R0) & 0xFFFFFFFC);
+ prevadr = 0;
+ }
+ /*
+ ** We treat the bte/btne the same way as the bc/bnc case.
+ */
+ else if ((instr & 0xF0000000) == 0x50000000) /* bte/btne */
+ {
+ val = cond;
+ offset = SIGN_EXT16(((instr & 0x001F0000) >> 5) |
+ (instr & 0x000007FF));
+ if ((INDIM) && !(ENDIM))
+ prevadr = nextadr;
+ else
+ prevadr = nextadr;
+
+ nextadr += (offset << 2);
+ }
+ /*
+ ** We treat the bte/btne the same way as the bc/bnc case.
+ ** With the caveat that the 2 breakpoints may turn out to be at the same
+ ** address in which case we ignore one of them.
+ */
+ else if ((instr & 0xFC000000) == 0xB4000000) /* bla */
+ {
+ val = cond_d;
+ offset = SIGN_EXT16(((instr & 0x001F0000) >> 5) |
+ (instr & 0x000007FF));
+ if ((INDIM) && !(ENDIM))
+ {
+ prevadr = nextadr + 8;
+ } else {
+ prevadr = nextadr + 4;
+ }
+ nextadr += (offset << 2);
+ if (prevadr == nextadr) prevadr = 0;
+ }
+ } else {
+ int adjust = 0;
+
+ nextadr = addr;
+
+ if (ISDIM(FOPADR(addr)))
+ {
+ if (ISDIM(FOPADR(nextadr-8)))
+ {
+ instr = adj_read_memory_integer(CORADR(addr-8));
+ brk->mode = DIM;
+ } else {
+ instr = adj_read_memory_integer(addr-4);
+ brk->mode = RIM;
+ }
+ } else {
+ if (ISDIM(addr-4))
+ {
+ instr = adj_read_memory_integer(addr-4);
+ brk->mode = BIM;
+ } else {
+ instr = adj_read_memory_integer (addr-4);
+ brk->mode = SIM;
+ }
+ }
+
+ /* examine the PREVIOUS instruction to determine if we are in a branch delay
+ slot. If we are, dont set a break here -- set it on the previous instruction.
+ This code also accounts for dual instruction mode */
+ if ((instr & 0xE0000000) == 0x60000000 &&
+ (instr & 0xF8000000) != 0x60000000) /* not pfld.y */
+ {
+ adjust++;
+ /* br /call */
+ /* bc /bnc */
+ /* bc.t /bnc.t*/
+ if ((instr & 0xF8000000) == 0x68000000) /* br or call */
+ printf(" Breakpoint adjusted to avoid br/call delay slot and multiple breakpoints\n");
+
+ if ((instr & 0xF4000000) == 0x74000000) /* bc.t or bnc.t */
+ printf(" Breakpoint adjusted to avoid bc.t/bnc.t delay slot and multiple breakpoints\n");
+ /* it IS really OK to set a break on the instruction AFTER the conditional branch
+ -- it DOESN't have a delay slot */
+ if ((instr & 0xF4000000) == 0x70000000) /* bc / bnc */
+ /* printf(" Breakpoint adjusted to avoid bc/bnc delay slot and multiple breakpoints\n"); */
+ adjust = 0;
+ } else if
+ ((instr & 0xFC00003F) == 0x4C000002 || /* bri/ calli */
+ (instr & 0xFC000000) == 0x40000000)
+ {
+ adjust++;
+ printf(" Breakpoint adjusted to avoid calli/bri delay slot and multiple breakpoints\n");
+ } else if
+ ((instr & 0xF0000000) == 0x50000000) /* bte - btne */
+ {
+ /* it's OK to set a break here -- we are NOT in aa branch delay slot */
+ /*
+ adjust++;
+ printf(" Breakpoint adjusted to avoid bte/btne multiple breakpoints\n");
+ */
+ adjust = 0;
+ } else if
+ ((instr & 0xFC000000) == 0xB4000000)
+ {
+ adjust++;
+ printf(" Breakpoint adjusted to avoid bla delay slot and multiple breakpoints\n");
+ }
+ if (adjust)
+ {
+ if (brk->mode == DIM)
+ {
+ nextadr -= 8;
+ nextadr = CORADR(nextadr);
+ }
+ else
+ nextadr -=4;
+ }
+
+ }
+
+ if (brk->mode == RIM)
+ brk->mode = DIM;
+ if (brk->mode == BIM)
+ brk->mode = SIM;
+
+ if (nextadr)
+ {
+ if (brk->mode == DIM)
+ {
+ brk->act_addr[0] = CORADR(nextadr);
+ brk->act_addr[1] = FOPADR(nextadr);
+ } else {
+ brk->act_addr[0] = nextadr;
+ brk->act_addr[1] = 0;
+ }
+ }
+
+ if (prevadr)
+ {
+ brk->address1 = prevadr;
+ if (brk->mode == DIM)
+ {
+ brk->act_addr[2] = CORADR(prevadr);
+ brk->act_addr[3] = FOPADR(prevadr);
+ } else {
+ brk->act_addr[2] = prevadr;
+ brk->act_addr[3] = 0;
+ }
+ } else {
+ brk->act_addr[2] = brk->act_addr[3] = 0;
+ }
+ return val;
+}
+
+/*
+ ** Figure out whether we are in a delayed slot and if so then take necessary
+ ** action to resume properly - remember trap pre-empts instruction
+ */
+int
+ wasabranch (addr, nextpc, ss)
+CORE_ADDR addr, *nextpc;
+int ss;
+{
+ long nextadr, instr;
+ int val = not_branch;
+ long offset; /* Must be signed for sign-extend */
+
+ if (ss)
+ {
+ if (INDIM)
+ {
+ nextadr = CORADR((int)(addr + 8));
+ instr = adj_read_memory_integer (CORADR(addr));
+ }
+ else
+ {
+ nextadr = addr + 4;
+ instr = adj_read_memory_integer (addr);
+ }
+ } else {
+ if (ISDIM(addr))
+ {
+ nextadr = CORADR(addr);
+ instr = adj_read_memory_integer (nextadr);
+ }
+ else
+ {
+ nextadr = addr;
+ instr = adj_read_memory_integer (addr);
+ }
+ }
+
+
+ if ((instr & 0xE0000000) == 0x60000000 && /* CTRL format */
+ (instr & 0xF8000000) != 0x60000000) /* not pfld.y */
+ {
+ if ((instr & 0xF8000000) == 0x68000000) /* br or call */
+ val = uncond_d;
+ else if ((instr & 0xF4000000) == 0x74000000) /* bc.t or bnc.t */
+ val = cond_d;
+ else if ((instr & 0xF4000000) == 0x70000000) /* bc or bnc */
+ val = cond;
+
+ offset = (instr & 0x03ffffff);
+ if (offset & 0x02000000) /* sign extend? */
+ offset |= 0xFC000000;
+ nextadr += (offset << 2);
+ }
+ else if ((instr & 0xFC00003F) == 0x4C000002 || /* calli */
+ (instr & 0xFC000000) == 0x40000000) /* bri */
+ {
+ if (ss)
+ {
+ val = uncond_d;
+ offset = ((instr & 0x0000F800) >> 11);
+ nextadr = (read_register(offset) & 0xFFFFFFFC);
+ } else {
+ val = uncond_d;
+ }
+ }
+ else if ((instr & 0xF0000000) == 0x50000000) /* bte or btne */
+ {
+ val = cond;
+
+ offset = SIGN_EXT16(((instr & 0x001F0000) >> 5) | (instr & 0x000007FF));
+ nextadr += (offset << 2);
+ }
+ else if ((instr & 0xFC000000) == 0xB4000000) /* bla */
+ {
+ val = cond_d;
+
+ offset = SIGN_EXT16(((instr & 0x001F0000) >> 5) | (instr & 0x000007FF));
+ nextadr += (offset << 2);
+ }
+
+ *nextpc = nextadr;
+ return val;
+}
+
+extern char registers[];
+
+i860_do_registers_info(regnum,fpregs)
+ int regnum;
+ int fpregs;
+{
+ register int i;
+ unsigned int val;
+ unsigned int j,k;
+
+ if (regnum == -1)
+ printf_filtered (
+ "Register Contents (relative to selected stack frame)\n\n");
+
+ if (regnum != -1) /* print one register */
+ {
+ if ((regnum >=F0 ) && (regnum <= F31))
+ bcopy (&registers[ADJ_FREG(regnum)<<2], &val, sizeof (long));
+ else
+ bcopy (&registers[regnum<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[regnum], val);
+ printf("\n\t"); fflush(stdout);
+ }
+ else /* print all registers */
+ {
+
+ printf("\n Control/Status Registers :- \n\t");
+ for (j=0; j<=DB; j++)
+ {
+ bcopy (&registers[j<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[j], val);
+ }
+ printf("\n\t"); fflush(stdout);
+
+ /* EPSR */
+ bcopy (&registers[EPSR<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[EPSR], val);
+
+ /* FSR */
+ bcopy (&registers[FSR<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[FSR], val);
+
+ /* CCR */
+ bcopy (&registers[CCR<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[CCR], val);
+ /* BEAR*/
+ bcopy (&registers[BEAR<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[BEAR], val);
+
+
+#ifdef JIM_ADD_PRIV
+ for (j=P0; j<=P3; j++)
+ {
+ bcopy (&registers[j<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[j], val);
+ }
+#endif
+
+ printf("\n Integer Registers :- \n\t");
+ for (j=R0; j<=R31; j++)
+ {
+ if (j != IREGS && (j % 4 == 0))
+ {
+ printf("\n\t"); fflush(stdout);
+ }
+ bcopy (&registers[j<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[j], val);
+ }
+
+ printf("\n Floating Registers :- \n\t");
+ for (j=F0; j<=F31; j++)
+ {
+ if (j != FREGS && (j % 4 == 0))
+ {
+ printf("\n\t"); fflush(stdout);
+ }
+ bcopy (&registers[ADJ_FREG(j)<<2], &val, sizeof (long));
+ printf("%-4s 0x%08x\t", reg_names[j], val);
+ }
+
+ printf("\n Special Registers :- \n\t");
+ for (j=SPC_KI; j<=SPC_MERGE; j+=2)
+ {
+ unsigned int valh;
+ if (j == SPC_T)
+ {
+ printf("\n\t"); fflush(stdout);
+ }
+ bcopy (&registers[j<<2], &val, sizeof (long));
+ bcopy (&registers[(j+1)<<2], &valh, sizeof (long));
+ printf("%-6s 0x%08x %08x\t", reg_names[j], val,valh);
+ }
+
+ printf("\n Graphics Pipeline :- \n");
+ {
+ unsigned int valh, val2,val3;
+ j = PSV_I1;
+ bcopy (&registers[j<<2], &val, sizeof (long));
+ bcopy (&registers[(j+1)<<2], &valh, sizeof (long));
+ printf("\t\t\t%-8s 0x%08x %08x \n", reg_names[j], val,valh);
+ }
+
+ printf(" Memory Load Pipeline :- \n");
+ for (j=PSV_L1; j<=PSV_L3; j+=4)
+ {
+ unsigned int valh, val2,val3;
+ bcopy (&registers[j<<2], &val, sizeof (long));
+ bcopy (&registers[(j+1)<<2], &valh, sizeof (long));
+ bcopy (&registers[(j+2)<<2], &val2, sizeof (long));
+ bcopy (&registers[(j+3)<<2], &val3, sizeof (long));
+ printf("\t\t%-8s 0x%08x %08x %08x %08x\n", reg_names[j],
+ val,valh,val2,val3);
+ }
+
+ printf("\n Adder Pipeline :-\t\tMultiplier Pipeline :-\t\tFSR results :-\n");
+ for (i=PSV_FSR1,j=PSV_A1,k=PSV_M1; j<=PSV_A3; i++,j+=2,k+=2)
+ {
+ unsigned int valh,val2,val3,val4;
+ bcopy (&registers[i<<2], &val4, sizeof (long));
+ bcopy (&registers[j<<2], &val, sizeof (long));
+ bcopy (&registers[(j+1)<<2], &valh, sizeof (long));
+ bcopy (&registers[k<<2], &val2, sizeof (long));
+ bcopy (&registers[(k+1)<<2], &val3, sizeof (long));
+ printf(" %-4s 0x%08x %08x\t", reg_names[j], val,valh);
+ printf("%-4s 0x%08x %08x\t", reg_names[k], val2,val3);
+ printf("%-4s 0x%08x\n", reg_names[i], val4);
+ }
+
+ }
+
+
+}
+
+int has_stored_r1(CORE_ADDR prologue_start, CORE_ADDR prologue_end)
+{
+ long instr;
+ CORE_ADDR addr;
+
+ BTDEBUG("has_stored_r1, prologue_start %x, prologue_end %x\n",
+ prologue_start, prologue_end);
+
+ for (addr = prologue_start; addr <= prologue_end; addr += 4)
+ {
+
+ instr = adj_read_memory_integer (addr);
+ if ((instr & 0xFFE0F801) == 0x1C400801 /* st.l r1,X(sp) */
+ || (instr & 0xFFE0F801) == 0x1C600801) /* st.l r1,X(fp) */
+ return (1);
+ }
+ return 0;
+}
+
+/* This function returns 1 if there is no stored r1, 0 otherwise.
+ The function returns 1 if the pc is in a function prologue,
+ or the function prologue didn't save the return pointer in
+ the stack frame, 0 otherwise */
+
+int no_stored_rp(CORE_ADDR pc)
+{
+ CORE_ADDR func_start, prologue_end;
+
+ func_start = get_pc_function_start(pc);
+ if (func_start)
+ {
+ prologue_end = func_start;
+ SKIP_PROLOGUE(prologue_end);
+ if ( (pc >= func_start) && (pc <= prologue_end))
+ {
+ BTDEBUG("no_stored_rp: pc %x is in prologue \n",pc);
+ return 1;
+ }
+ /* otherwise, see if the entry sequence stored the return pointer.
+ If it didn't, return 1 */
+ /* Some procedures , at least, store the return pointer AFTER the prologue sequence! */
+ if (!has_stored_r1(func_start, pc))
+ {
+ BTDEBUG("no_stored_rp, for pc %x, prologue didn't store r1\n",pc);
+ return 1;
+ }
+ }
+ BTDEBUG("no_stored_rp for pc %x return pointer was stored \n", pc);
+
+ return 0;
+}
+
+/* The following set of routines was adapted from existing code previously
+ in an i860-specific version of breakpoint.c by Peggy Fieland
+ (Margaret_Fieland@vos.stratus.com) */
+unsigned int dbrkval, dbrkmod;
+void i860_dbrk_breakpoint()
+{
+ BTDEBUG("i860_dbrk_breakpoint was called , dbrkval %x\n", dbrkval);
+
+ if (dbrkval)
+ {
+ *(int *)&registers[DB<<2] = dbrkval;
+ }
+ else
+ {
+ *(int *)&registers[DB<<2] = 0;
+ }
+
+ *(int *)&registers[PSR<<2] &= ~3;
+ *(int *)&registers[PSR<<2] |= dbrkmod;
+
+ store_inferior_registers(DB);
+ store_inferior_registers(PSR);
+
+}
+
+void
+d_ro_break_command(arg)
+char *arg;
+{
+ dbrkval = strtoul(arg, NULL, 0);
+ dbrkmod = 0x01;
+printf(" ro_dbreak - %x %x\n", dbrkval, dbrkmod);
+}
+
+void
+d_wo_break_command(arg)
+char *arg;
+{
+ dbrkval = strtoul(arg, NULL, 0);
+ dbrkmod = 0x02;
+printf(" wo_dbreak - %x %x\n", dbrkval, dbrkmod);
+}
+
+void
+d_rw_break_command(arg)
+char *arg;
+{
+ dbrkval = strtoul(arg, NULL, 0);
+ dbrkmod = 0x03;
+printf(" rw_dbreak - %x %x\n", dbrkval, dbrkmod);
+}
+
+void
+clear_dbreak()
+{
+ dbrkval = 0;
+ dbrkmod = 0;
+}
+
+void
+i860_init_breakpoints()
+{
+ dbrkval = dbrkmod = 0;
+ add_com ("dbro", class_breakpoint, d_ro_break_command,
+ "Set a data breakpoint READ ONLY, 32-bit data element.");
+ add_com ("dbwo", class_breakpoint, d_wo_break_command,
+ "Set a data breakpoint WRITE ONLY, 32-bit data element.");
+ add_com ("dbrw", class_breakpoint, d_rw_break_command,
+ "Set a data breakpoint READ/WRITE, 32-bit data element.");
+ add_com ("dclear", class_breakpoint, clear_dbreak,
+ "clear the current data breakpoint.");
+ add_com_alias ("dc", "dclear", class_breakpoint, 1);
+
+}
+
+int i860_insert_breakpoint(b)
+struct breakpoint *b;
+{
+ int val;
+
+ place_brk( b->address, BREAK_MODE, b );
+ if (b->mode == DIM)
+ {
+
+ adj_read_memory (b->act_addr[0], &b->shadow_contents[0], 4);
+ val = adj_write_memory (b->act_addr[0], break_insn, 4);
+ if (val) return val;
+ adj_read_memory (b->act_addr[1], &b->shadow_contents[1], 4);
+ /* val = adj_write_memory (b->act_addr[1], float_insn, 4); */
+ if (val) return val;
+ }
+ else
+ {
+ adj_read_memory (b->act_addr[0], &b->shadow_contents[0], 4);
+ val = adj_write_memory (b->act_addr[0], break_insn, 4);
+ }
+ if (b->address1)
+ {
+ if (b->mode == DIM)
+ {
+
+ adj_read_memory (b->act_addr[2], &b->shadow_contents[2], 4);
+ val = adj_write_memory (b->act_addr[2], break_insn, 4);
+ if (val) return val;
+ adj_read_memory (b->act_addr[3], &b->shadow_contents[3], 4);
+ /* val = adj_write_memory (b->act_addr[3], float_insn, 4); */
+ if (val) return val;
+ }
+ else
+ {
+ adj_read_memory (b->act_addr[2], &b->shadow_contents[0], 4);
+ val = adj_write_memory (b->act_addr[2], break_insn, 4);
+ }
+ }
+ if (val)
+ return val;
+ BTDEBUG("Inserted breakpoint at 0x%x, shadow 0x%x, 0x%x.\n",
+ b->address, b->shadow_contents[0], b->shadow_contents[1]);
+ b->inserted = 1;
+ return 0;
+}
+
+int i860_remove_breakpoint(b)
+struct breakpoint *b;
+{
+ int val;
+
+ if (b->inserted)
+ {
+ if (b->mode == DIM)
+ {
+ val =adj_write_memory (b->act_addr[0], &(b->shadow_contents[0]), 4);
+ val =adj_write_memory (b->act_addr[1], &(b->shadow_contents[1]), 4);
+ if (b->address1)
+ {
+ val =adj_write_memory (b->act_addr[2], &(b->shadow_contents[2]), 4);
+ val =adj_write_memory (b->act_addr[3], &(b->shadow_contents[3]), 4);
+ }
+ }
+ else
+ {
+ val =adj_write_memory (b->act_addr[0], b->shadow_contents, 4);
+ if (b->address1)
+ {
+ val =adj_write_memory (b->act_addr[2], b->shadow_contents, 4);
+ }
+ }
+ if (val)
+ return val;
+ b->inserted = 0;
+ BTDEBUG( "Removed breakpoint at 0x%x, shadow 0x%x, 0x%x.\n",
+ b->address, b->shadow_contents[0], b->shadow_contents[1]);
+ }
+
+ return 0;
+
+
+}
+
+
+#ifdef USE_PROC_FS /* Target dependent support for /proc */
+
+#include <sys/procfs.h>
+
+/* The following routines were added by Peggy Fieland (Margaret_Fieland@vos.stratus.com)
+They were adapted from the m-68k versions of the routines .*/
+
+/* Given a pointer to a floating point register set in /proc format
+ (fpregset_t *), unpack the register contents and supply them as gdb's
+ idea of the current floating point register values. */
+
+void
+supply_fpregset (fpregsetp)
+fpregset_t *fpregsetp;
+{
+ register int regno;
+
+ BTDEBUG("supply_fregset called \n");
+
+ for (regno = F0 ; regno <= F31 ; regno++)
+ {
+ supply_register (regno, (char *) &(fpregsetp -> fpu.r_freg[regno-F0]));
+ }
+}
+
+/* Given a pointer to a floating point register set in /proc format
+ (fpregset_t *), update the register specified by REGNO from gdb's idea
+ of the current floating point register set. If REGNO is -1, update
+ them all. */
+
+void
+fill_fpregset (fpregsetp, regno)
+fpregset_t *fpregsetp;
+int regno;
+{
+ int regi;
+ char *to;
+ char *from;
+ extern char registers[];
+ BTDEBUG("fill_fregset regno %d\n",regno);
+
+ for (regi = F0 ; regi <= F31 ; regi++)
+ {
+ if ((regno == -1) || (regno == regi))
+ {
+ from = (char *) &registers[REGISTER_BYTE (regi)];
+ to = (char *) &(fpregsetp -> fpu.r_freg[regi-F0]);
+ bcopy (from, to, REGISTER_RAW_SIZE (regno));
+ }
+ }
+}
+
+
+/* Given a pointer to a general register set in /proc format (gregset_t *),
+ unpack the register contents and supply them as gdb's idea of the current
+ register values. */
+
+void
+supply_gregset (gregsetp)
+gregset_t *gregsetp;
+{
+ register int regno;
+ register greg_t *regp = (greg_t *) gregsetp;
+
+ BTDEBUG("supply_gregset called \n");
+
+ for (regno = 0 ; regno <= R31 ; regno++)
+ {
+ supply_register (regno, (char *) (regp + regno));
+ }
+}
+
+void
+fill_gregset (gregsetp, regno)
+gregset_t *gregsetp;
+int regno;
+{
+ int regi;
+ extern char registers[];
+ register greg_t *regp = (greg_t *) gregsetp;
+ BTDEBUG("fill_gregset regno %d \n",regno);
+
+ for (regi = 0 ; regi <= R31 ; regi++)
+ {
+ if ((regno == -1) || (regno == regi))
+ {
+ *(regp + regi) = *(int *) &registers[REGISTER_BYTE (regi)];
+ }
+
+ }
+}
+#endif
+
diff --git a/gdb/tm-stratus.h b/gdb/tm-stratus.h
new file mode 100644
index 0000000..f7df3d4
--- /dev/null
+++ b/gdb/tm-stratus.h
@@ -0,0 +1,21 @@
+/* Macro definitions for GDB on a Stratus machine.
+ Copyright (C) 1992, Free Software Foundation, Inc.
+
+This file is part of GDB.
+
+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
+the Free Software Foundation; either version 2 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+GNU General Public License for more details.
+
+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. */
+
+#include "tm-sysv4.h"
+#include "tm-i860.h"
diff --git a/gdb/xm-stratus.h b/gdb/xm-stratus.h
new file mode 100644
index 0000000..a225bbe
--- /dev/null
+++ b/gdb/xm-stratus.h
@@ -0,0 +1,29 @@
+/* Macro definitions for GDB on a Stratus machine.
+ Copyright (C) 1992, Free Software Foundation, Inc.
+ Contributed by Peggy Fieland (pfieland@stratus.com)
+
+This file is part of GDB.
+
+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
+the Free Software Foundation; either version 2 of the License, or
+(at your option) any later version.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+GNU General Public License for more details.
+
+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. */
+
+/* Pick up most of what we need from the generic mi860 host include file. */
+
+#include "xm-i860.h"
+
+/* Pick up more stuff from the generic SVR4 host include file. */
+
+#include "xm-svr4.h"
+#define NOVOLVOID 1 /* compiler doesn't like volatile void declaration on functions*/
+#define NO_JOB_CONTROL
generated by cgit v1.1 at 2024-12-19 19:48:20 +0000