Browse Source

MIPS/opcodes: Disassemble the RFE instruction

Fix a commit b015e599c7 ("[MIPS] Add new virtualization instructions"),
<https://sourceware.org/ml/binutils/2013-05/msg00118.html>, regression
and bring the disassembly of the RFE instruction back for the relevant
ISA levels.

It is because the "rfe" opcode table entry was incorrectly moved behind
the catch-all generic "c0" entry for CP0 instructions, causing output
like:

  00:	42000010 	c0	0x10

to be produced rather than:

  00:	42000010 	rfe

even for ISA levels that do include the RFE instruction.

Move the "rfe" entry ahead of "c0" then, correcting the problem.  Add a
suitable test case.

	opcodes/
	* mips-opc.c (mips_builtin_opcodes): Move the "rfe" entry ahead
	of "c0".

	gas/
	* testsuite/gas/mips/rfe.d: New test.
	* testsuite/gas/mips/rfe.s: New test source.
	* testsuite/gas/mips/mips.exp: Run the new test.
binutils-2_37-branch
Maciej W. Rozycki 5 years ago
parent
commit
b930964c42
  1. 6
      gas/ChangeLog
  2. 3
      gas/testsuite/gas/mips/mips.exp
  3. 9
      gas/testsuite/gas/mips/rfe.d
  4. 8
      gas/testsuite/gas/mips/rfe.s
  5. 5
      opcodes/ChangeLog
  6. 5
      opcodes/mips-opc.c

6
gas/ChangeLog

@ -1,3 +1,9 @@
2021-05-29 Maciej W. Rozycki <macro@orcam.me.uk>
* testsuite/gas/mips/rfe.d: New test.
* testsuite/gas/mips/rfe.s: New test source.
* testsuite/gas/mips/mips.exp: Run the new test.
2021-05-29 Maciej W. Rozycki <macro@orcam.me.uk>
* testsuite/gas/mips/cp1-names-r3900.d: New test.

3
gas/testsuite/gas/mips/mips.exp

@ -1338,6 +1338,9 @@ if { [istarget mips*-*-vxworks*] } {
run_dump_test_arches "cp0m" [mips_arch_list_matching mips1 \
!mips2 !micromips]
run_dump_test_arches "rfe" [mips_arch_list_matching mips1 \
!mips3 !mips32 !micromips]
run_dump_test "cp1-names-numeric"
run_dump_test "cp1-names-r3000"
run_dump_test "cp1-names-r3900"

9
gas/testsuite/gas/mips/rfe.d

@ -0,0 +1,9 @@
#objdump: -d --prefix-addresses --show-raw-insn
#name: MIPS RFE instruction
#as: -32
.*: +file format .*mips.*
Disassembly of section \.text:
[0-9a-f]+ <[^>]*> 42000010 rfe
\.\.\.

8
gas/testsuite/gas/mips/rfe.s

@ -0,0 +1,8 @@
.text
.set noreorder
foo:
rfe
# Force some (non-delay-slot) zero bytes, to make 'objdump' print ...
.align 4, 0
.space 16

5
opcodes/ChangeLog

@ -1,3 +1,8 @@
2021-05-29 Maciej W. Rozycki <macro@orcam.me.uk>
* mips-opc.c (mips_builtin_opcodes): Move the "rfe" entry ahead
of "c0".
2021-05-29 Maciej W. Rozycki <macro@orcam.me.uk>
* mips-dis.c (mips_cp1_names_mips): New variable.

5
opcodes/mips-opc.c

@ -3399,6 +3399,9 @@ const struct mips_opcode mips_builtin_opcodes[] =
{"ginvi", "s", 0x7c00003d, 0xfc1fffff, RD_1, 0, 0, GINV, 0 },
{"ginvt", "s,+\\", 0x7c0000bd, 0xfc1ffcff, RD_1, 0, 0, GINV, 0 },
/* RFE conflicts with the new Virt spec instruction tlbgp. */
{"rfe", "", 0x42000010, 0xffffffff, 0, 0, I1|T3, 0, 0 },
/* No hazard protection on coprocessor instructions--they shouldn't
change the state of the processor and if they do it's up to the
user to put in nops as necessary. These are at the end so that the
@ -3411,8 +3414,6 @@ const struct mips_opcode mips_builtin_opcodes[] =
{"cop1", "C", 0, (int) M_COP1, INSN_MACRO, INSN2_M_FP_S, I1, 0, 0 },
{"cop2", "C", 0, (int) M_COP2, INSN_MACRO, 0, I1, 0, IOCT|IOCTP|IOCT2 },
{"cop3", "C", 0, (int) M_COP3, INSN_MACRO, 0, I1, 0, IOCT|IOCTP|IOCT2 },
/* RFE conflicts with the new Virt spec instruction tlbgp. */
{"rfe", "", 0x42000010, 0xffffffff, 0, 0, I1|T3, 0, 0 },
};
#define MIPS_NUM_OPCODES \

Loading…
Cancel
Save