From: Andrew Pinski <pinskia@gmail.com>
To: Kyle McMartin <kmcmarti@redhat.com>
Cc: "gdb-patches@sourceware.org" <gdb-patches@sourceware.org>,
Joel Brobecker <brobecker@adacore.com>
Subject: Re: [PATCHv4] aarch64: detect atomic sequences like other ll/sc architectures
Date: Thu, 24 Apr 2014 18:39:00 -0000 [thread overview]
Message-ID: <CA+=Sn1mWke-_8O0-rJb5186MLZMOqtq46hBfNueW_v0BS13fzw@mail.gmail.com> (raw)
In-Reply-To: <20140424183510.GI7588@redacted.bos.redhat.com>
On Thu, Apr 24, 2014 at 11:35 AM, Kyle McMartin <kmcmarti@redhat.com> wrote:
> Once more with feeling. :-) Fixed ChangeLog entries and whitespace.
>
> 2014-04-24 Kyle McMartin <kyle@redhat.com>
>
> gdb:
> * aarch64-tdep.c (aarch64_software_single_step): New function.
> (aarch64_gdbarch_init): Handle single stepping of atomic sequences
> with aarch64_software_single_step.
>
> gdb/testsuite:
> * gdb.arch/aarch64-atomic-inst.c: New file.
> * gdb.arch/aarch64-atomic-inst.exp: New file.
>
> ---
> gdb/aarch64-tdep.c | 77 ++++++++++++++++++++++++++
> gdb/testsuite/gdb.arch/aarch64-atomic-inst.c | 48 ++++++++++++++++
> gdb/testsuite/gdb.arch/aarch64-atomic-inst.exp | 48 ++++++++++++++++
> 3 files changed, 173 insertions(+)
> create mode 100644 gdb/testsuite/gdb.arch/aarch64-atomic-inst.c
> create mode 100644 gdb/testsuite/gdb.arch/aarch64-atomic-inst.exp
>
> diff --git a/gdb/aarch64-tdep.c b/gdb/aarch64-tdep.c
> index bba10d8..b64bbde 100644
> --- a/gdb/aarch64-tdep.c
> +++ b/gdb/aarch64-tdep.c
> @@ -2509,6 +2509,82 @@ value_of_aarch64_user_reg (struct frame_info *frame, const void *baton)
> }
>
>
> +/* Implement the "software_single_step" gdbarch method, needed to
> + single step through atomic sequences on AArch64. */
> +
> +static int
> +aarch64_software_single_step (struct frame_info *frame)
> +{
> + struct gdbarch *gdbarch = get_frame_arch (frame);
> + struct address_space *aspace = get_frame_address_space (frame);
> + enum bfd_endian byte_order = gdbarch_byte_order (gdbarch);
> + const int insn_size = 4;
> + const int atomic_sequence_length = 16; /* Instruction sequence length. */
> + CORE_ADDR pc = get_frame_pc (frame);
> + CORE_ADDR breaks[2] = { -1, -1 };
> + CORE_ADDR loc = pc;
> + CORE_ADDR closing_insn = 0;
> + uint32_t insn = read_memory_unsigned_integer (loc, insn_size, byte_order);
I think this is broken for big-endian aarch64. Instructions are
always in little-endian mode so we really should be reading them
always as little-endian.
Thanks,
Andrew Pinski
> + int index;
> + int insn_count;
> + int bc_insn_count = 0; /* Conditional branch instruction count. */
> + int last_breakpoint = 0; /* Defaults to 0 (no breakpoints placed). */
> +
> + /* Look for a Load Exclusive instruction which begins the sequence. */
> + if (!decode_masked_match (insn, 0x3fc00000, 0x08400000))
> + return 0;
> +
> + for (insn_count = 0; insn_count < atomic_sequence_length; ++insn_count)
> + {
> + int32_t offset;
> + unsigned cond;
> +
> + loc += insn_size;
> + insn = read_memory_unsigned_integer (loc, insn_size, byte_order);
> +
> + /* Check if the instruction is a conditional branch. */
> + if (decode_bcond (loc, insn, &cond, &offset))
> + {
> + if (bc_insn_count >= 1)
> + return 0;
> +
> + /* It is, so we'll try to set a breakpoint at the destination. */
> + breaks[1] = loc + offset;
> +
> + bc_insn_count++;
> + last_breakpoint++;
> + }
> +
> + /* Look for the Store Exclusive which closes the atomic sequence. */
> + if (decode_masked_match (insn, 0x3fc00000, 0x08000000))
> + {
> + closing_insn = loc;
> + break;
> + }
> + }
> +
> + /* We didn't find a closing Store Exclusive instruction, fall back. */
> + if (!closing_insn)
> + return 0;
> +
> + /* Insert breakpoint after the end of the atomic sequence. */
> + breaks[0] = loc + insn_size;
> +
> + /* Check for duplicated breakpoints, and also check that the second
> + breakpoint is not within the atomic sequence. */
> + if (last_breakpoint
> + && (breaks[1] == breaks[0]
> + || (breaks[1] >= pc && breaks[1] <= closing_insn)))
> + last_breakpoint = 0;
> +
> + /* Insert the breakpoint at the end of the sequence, and one at the
> + destination of the conditional branch, if it exists. */
> + for (index = 0; index <= last_breakpoint; index++)
> + insert_single_step_breakpoint (gdbarch, aspace, breaks[index]);
> +
> + return 1;
> +}
> +
> /* Initialize the current architecture based on INFO. If possible,
> re-use an architecture from ARCHES, which is a list of
> architectures already created during this debugging session.
> @@ -2624,6 +2700,7 @@ aarch64_gdbarch_init (struct gdbarch_info info, struct gdbarch_list *arches)
> set_gdbarch_breakpoint_from_pc (gdbarch, aarch64_breakpoint_from_pc);
> set_gdbarch_cannot_step_breakpoint (gdbarch, 1);
> set_gdbarch_have_nonsteppable_watchpoint (gdbarch, 1);
> + set_gdbarch_software_single_step (gdbarch, aarch64_software_single_step);
>
> /* Information about registers, etc. */
> set_gdbarch_sp_regnum (gdbarch, AARCH64_SP_REGNUM);
> diff --git a/gdb/testsuite/gdb.arch/aarch64-atomic-inst.c b/gdb/testsuite/gdb.arch/aarch64-atomic-inst.c
> new file mode 100644
> index 0000000..9a73c7a
> --- /dev/null
> +++ b/gdb/testsuite/gdb.arch/aarch64-atomic-inst.c
> @@ -0,0 +1,48 @@
> +/* This file is part of GDB, the GNU debugger.
> +
> + Copyright 2008-2014 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/>. */
> +
> +int main(void)
> +{
> + unsigned long tmp, cond;
> + unsigned long dword = 0;
> +
> + /* Test that we can step over ldxr/stxr. This sequence should step from
> + ldxr to the following __asm __volatile. */
> + __asm __volatile ("1: ldxr %0,%2\n" \
> + " cmp %0,#1\n" \
> + " b.eq out\n" \
> + " add %0,%0,1\n" \
> + " stxr %w1,%0,%2\n" \
> + " cbnz %w1,1b" \
> + : "=&r" (tmp), "=&r" (cond), "+Q" (dword) \
> + : : "memory");
> +
> + /* This sequence should take the conditional branch and step from ldxr
> + to the return dword line. */
> + __asm __volatile ("1: ldxr %0,%2\n" \
> + " cmp %0,#1\n" \
> + " b.eq out\n" \
> + " add %0,%0,1\n" \
> + " stxr %w1,%0,%2\n" \
> + " cbnz %w1,1b\n" \
> + : "=&r" (tmp), "=&r" (cond), "+Q" (dword) \
> + : : "memory");
> +
> + dword = -1;
> +__asm __volatile ("out:\n");
> + return dword;
> +}
> diff --git a/gdb/testsuite/gdb.arch/aarch64-atomic-inst.exp b/gdb/testsuite/gdb.arch/aarch64-atomic-inst.exp
> new file mode 100644
> index 0000000..377aebc
> --- /dev/null
> +++ b/gdb/testsuite/gdb.arch/aarch64-atomic-inst.exp
> @@ -0,0 +1,48 @@
> +# Copyright 2008-2014 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, write to the Free Software
> +# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
> +#
> +# This file is part of the gdb testsuite.
> +
> +# Test single stepping through atomic sequences beginning with
> +# a ldxr instruction and ending with a stxr instruction.
> +
> +if {![istarget "aarch64*"]} {
> + verbose "Skipping ${gdb_test_file_name}."
> + return
> +}
> +
> +standard_testfile
> +if { [prepare_for_testing ${testfile}.exp ${testfile} ${srcfile}] } {
> + return -1
> +}
> +
> +if ![runto_main] {
> + untested "could not run to main"
> + return -1
> +}
> +
> +gdb_breakpoint "[gdb_get_line_number "ldxr"]" \
> + "Breakpoint $decimal at $hex" \
> + "Set the breakpoint at the start of the sequence"
> +
> +gdb_test "continue" "Continuing.*Breakpoint $decimal.*" \
> + "Continue until breakpoint"
> +
> +gdb_test "next" ".*__asm __volatile.*" \
> + "Step through the ldxr/stxr sequence"
> +
> +gdb_test "next" ".*return dword.*" \
> + "Stepped through sequence through conditional branch"
> --
> 1.9.0
>
next prev parent reply other threads:[~2014-04-24 18:39 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-04-24 18:35 Kyle McMartin
2014-04-24 18:39 ` Andrew Pinski [this message]
2014-04-24 18:51 ` Kyle McMartin
2014-04-30 16:04 ` [PATCHv5] " Kyle McMartin
2014-05-07 13:52 ` Joel Brobecker
2014-05-07 15:10 ` Kyle McMartin
2014-05-07 16:41 ` Andrew Pinski
2014-05-07 16:46 ` Joel Brobecker
2014-05-07 18:12 ` Kyle McMartin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='CA+=Sn1mWke-_8O0-rJb5186MLZMOqtq46hBfNueW_v0BS13fzw@mail.gmail.com' \
--to=pinskia@gmail.com \
--cc=brobecker@adacore.com \
--cc=gdb-patches@sourceware.org \
--cc=kmcmarti@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox