binutils-gdb/gdb/testsuite/gdb.arch/riscv-unwind-long-insn.exp
Andrew Burgess 312617a3d0 gdb/riscv: Don't error when decoding a 6 or 8 byte instruction
If the RISC-V prologue scanner finds a 6 or 8 byte instruction we
currently throw an internal error, which is not great for the user.

A mechanism already exists in the prologue scanner to leave
instructions marked as unknown so that we can stop the prologue scan
without raising an error, this is used for all 2 and 4 byte
instructions that are not part of the small set the prologue scanner
actually understands.

This commit changes GDB so that all 6 and 8 byte instructions are
marked as unknown, rather than causing an error.

gdb/ChangeLog:

	* riscv-tdep.c (riscv_insn::decode): Gracefully ignore
	instructions of lengths 6 or 8 bytes.

gdb/testsuite/ChangeLog:

	* gdb.arch/riscv-unwind-long-insn-6.s: New file.
	* gdb.arch/riscv-unwind-long-insn-8.s: New file.
	* gdb.arch/riscv-unwind-long-insn.c: New file.
	* gdb.arch/riscv-unwind-long-insn.exp: New file.
2019-06-05 09:54:47 +01:00

62 lines
2.1 KiB
Plaintext

# Copyright 2019 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
# the Free Software Foundation; either version 3 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, see <http://www.gnu.org/licenses/>.
# This tests GDB's ability to handle 6 and 8 byte instructions in the
# RISC-V prologue scanner. These instruction should be ignored, but
# should not result in an error that interrupts the debug session.
#
# Each of the files riscv-unwind-long-insn-*.s include a function
# (func) that contains a fake long instruction (6 or 8 bytes) in the
# prologue. We trick GDB into parsing the fake instruction by tail
# calling from a different function, 'bar' to the middle of 'func'.
if {![istarget "riscv*-*-*"]} {
verbose "Skipping ${gdb_test_file_name}."
return
}
foreach_with_prefix {insn_size} {6 8} {
standard_testfile riscv-unwind-long-insn.c \
riscv-unwind-long-insn-${insn_size}.s
set testfile "${testfile}-${insn_size}"
if {[prepare_for_testing "failed to prepare" $testfile \
"$srcfile $srcfile2" debug]} {
return -1
}
if ![runto_main] then {
fail "can't run to main"
return 0
}
gdb_breakpoint "bar"
gdb_continue_to_breakpoint "bar"
# This next single instruction step takes us through a tail-call
# from 'bar' into 'func'.
gdb_test "si" "func \(\).*"
# Now check that we have a sane backtrace.
gdb_test "bt" \
[multi_line \
"#0\[ \t\]*func \\\(\\\) at .*$srcfile2:\[0-9\]+" \
"#1\[ \t\]*$hex in main \\\(\\\) at .*$srcfile:\[0-9\]+"] \
"Backtrace to the main frame"
# Finally finish, and we should end up back in main.
gdb_test "finish" "main \\\(\\\) at .*$srcfile:.*"
}