From: Palmer Dabbelt <palmer@sifive.com>
To: andrew.burgess@embecosm.com
Cc: brobecker@adacore.com, gdb-patches@sourceware.org,
Jim Wilson <jimw@sifive.com>
Subject: Re: [PATCH] gdb/riscv: Improve non-dwarf stack unwinding
Date: Wed, 26 Sep 2018 17:13:00 -0000 [thread overview]
Message-ID: <mhng-06926dec-9f45-4801-a163-690cc865274b@palmer-si-x1c4> (raw)
In-Reply-To: <20180926151112.GL5952@embecosm.com>
On Wed, 26 Sep 2018 08:11:12 PDT (-0700), andrew.burgess@embecosm.com wrote:
> * Joel Brobecker <brobecker@adacore.com> [2018-09-26 06:49:16 -0700]:
>
>> Hi Andrew,
>>
>> > This commit improves the prologue scanning stack unwinder, to better
>> > support AUIPC, LUI, and more variants of ADD and ADDI.
>> >
>> > This allows unwinding over frames containing large local variables,
>> > where the frame size does not fit into a single instruction immediate,
>> > and is first loaded into a temporary register, before being added to
>> > the stack pointer.
>> >
>> > A new test is added that tests this behaviour. As there's nothing
>> > truely RiscV specific about this test I've added it into gdb.base, but
>> > as this depends on target specific code to perform the unwind it is
>> > possible that some targets might fail this new test.
>> >
>> > gdb/ChangeLog:
>> >
>> > * riscv-tdep.c (riscv_insn::decode): Decode c.lui.
>> > (riscv_scan_prologue): Split handling of AUIPC, LUI, ADD, ADDI,
>> > and NOP.
>> >
>> > gdb/testsuite/ChangeLog:
>> >
>> > * gdb.base/large-frame-1.c: New file.
>> > * gdb.base/large-frame-2.c: New file.
>> > * gdb.base/large-frame.exp: New file.
>> > * gdb.base/large-frame.h: New file.
>>
>> I am OK with the patch.
>>
>> Are you sure you want to assert on the register number? What if
>> the instruction was malformed and the register number ended up
>> being larger than RISCV_NUM_INTEGER_REGS; is that possible?
>
> Well, it shouldn't happen in the sense that, for a given instruction,
> if it operates on integer registers then the space in the encoding
> should only allow for up to RISCV_NUM_INTEGER_REGS registers (some
> encodings allow for less registers).
>
> Right now I can't imagine a situation where one of those asserts could
> fire and it's not either a bug in GDB, or a bug in the libopcode field
> extraction code - hence an assert rather than an error.
While I agree this is true for I-based ISAs, I think this might be able to fire
for E-based ISAs because those can actually encode invalid register indices.
That said, these should be decoded as invalid instructions so I think we're
safe here. I'm OK either way (ie, abort or warn).
>
> Thanks,
> Andrew
>
>
>
>
>>
>>
>> > ---
>> > gdb/ChangeLog | 6 ++++
>> > gdb/riscv-tdep.c | 59 ++++++++++++++++++++++------------
>> > gdb/testsuite/ChangeLog | 7 ++++
>> > gdb/testsuite/gdb.base/large-frame-1.c | 32 ++++++++++++++++++
>> > gdb/testsuite/gdb.base/large-frame-2.c | 25 ++++++++++++++
>> > gdb/testsuite/gdb.base/large-frame.exp | 57 ++++++++++++++++++++++++++++++++
>> > gdb/testsuite/gdb.base/large-frame.h | 24 ++++++++++++++
>> > 7 files changed, 189 insertions(+), 21 deletions(-)
>> > create mode 100644 gdb/testsuite/gdb.base/large-frame-1.c
>> > create mode 100644 gdb/testsuite/gdb.base/large-frame-2.c
>> > create mode 100644 gdb/testsuite/gdb.base/large-frame.exp
>> > create mode 100644 gdb/testsuite/gdb.base/large-frame.h
>> >
>> > diff --git a/gdb/riscv-tdep.c b/gdb/riscv-tdep.c
>> > index 254914c9c7e..163c8ec231d 100644
>> > --- a/gdb/riscv-tdep.c
>> > +++ b/gdb/riscv-tdep.c
>> > @@ -1227,7 +1227,11 @@ riscv_insn::decode (struct gdbarch *gdbarch, CORE_ADDR pc)
>> > m_imm.s = EXTRACT_RVC_ADDI4SPN_IMM (ival);
>> > }
>> > else if (is_c_lui_insn (ival))
>> > - m_opcode = OTHER;
>> > + {
>> > + m_opcode = LUI;
>> > + m_rd = decode_register_index (ival, OP_SH_CRS1S);
>> > + m_imm.s = EXTRACT_RVC_LUI_IMM (ival);
>> > + }
>> > /* C_SD and C_FSW have the same opcode. C_SD is RV64 and RV128 only,
>> > and C_FSW is RV32 only. */
>> > else if (xlen != 4 && is_c_sd_insn (ival))
>> > @@ -1359,28 +1363,41 @@ riscv_scan_prologue (struct gdbarch *gdbarch,
>> > gdb_assert (insn.rs1 () < RISCV_NUM_INTEGER_REGS);
>> > regs[insn.rd ()] = pv_add_constant (regs[insn.rs1 ()], 0);
>> > }
>> > - else if ((insn.rd () == RISCV_GP_REGNUM
>> > - && (insn.opcode () == riscv_insn::AUIPC
>> > - || insn.opcode () == riscv_insn::LUI
>> > - || (insn.opcode () == riscv_insn::ADDI
>> > - && insn.rs1 () == RISCV_GP_REGNUM)
>> > - || (insn.opcode () == riscv_insn::ADD
>> > - && (insn.rs1 () == RISCV_GP_REGNUM
>> > - || insn.rs2 () == RISCV_GP_REGNUM))))
>> > - || (insn.opcode () == riscv_insn::ADDI
>> > - && insn.rd () == RISCV_ZERO_REGNUM
>> > - && insn.rs1 () == RISCV_ZERO_REGNUM
>> > - && insn.imm_signed () == 0))
>> > + else if ((insn.opcode () == riscv_insn::ADDI
>> > + && insn.rd () == RISCV_ZERO_REGNUM
>> > + && insn.rs1 () == RISCV_ZERO_REGNUM
>> > + && insn.imm_signed () == 0))
>> > {
>> > - /* Handle: auipc gp, n
>> > - or: addi gp, gp, n
>> > - or: add gp, gp, reg
>> > - or: add gp, reg, gp
>> > - or: lui gp, n
>> > - or: add x0, x0, 0 (NOP) */
>> > - /* These instructions are part of the prologue, but we don't need
>> > - to do anything special to handle them. */
>> > + /* Handle: add x0, x0, 0 (NOP) */
>> > }
>> > + else if (insn.opcode () == riscv_insn::AUIPC)
>> > + {
>> > + gdb_assert (insn.rd () < RISCV_NUM_INTEGER_REGS);
>> > + regs[insn.rd ()] = pv_constant (cur_pc + insn.imm_signed ());
>> > + }
>> > + else if (insn.opcode () == riscv_insn::LUI)
>> > + {
>> > + /* Handle: lui REG, n
>> > + Where REG is not gp register. */
>> > + gdb_assert (insn.rd () < RISCV_NUM_INTEGER_REGS);
>> > + regs[insn.rd ()] = pv_constant (insn.imm_signed ());
>> > + }
>> > + else if (insn.opcode () == riscv_insn::ADDI)
>> > + {
>> > + /* Handle: addi REG1, REG2, IMM */
>> > + gdb_assert (insn.rd () < RISCV_NUM_INTEGER_REGS);
>> > + gdb_assert (insn.rs1 () < RISCV_NUM_INTEGER_REGS);
>> > + regs[insn.rd ()]
>> > + = pv_add_constant (regs[insn.rs1 ()], insn.imm_signed ());
>> > + }
>> > + else if (insn.opcode () == riscv_insn::ADD)
>> > + {
>> > + /* Handle: addi REG1, REG2, IMM */
>> > + gdb_assert (insn.rd () < RISCV_NUM_INTEGER_REGS);
>> > + gdb_assert (insn.rs1 () < RISCV_NUM_INTEGER_REGS);
>> > + gdb_assert (insn.rs2 () < RISCV_NUM_INTEGER_REGS);
>> > + regs[insn.rd ()] = pv_add (regs[insn.rs1 ()], regs[insn.rs2 ()]);
>> > + }
>> > else
>> > {
>> > end_prologue_addr = cur_pc;
>> > diff --git a/gdb/testsuite/gdb.base/large-frame-1.c b/gdb/testsuite/gdb.base/large-frame-1.c
>> > new file mode 100644
>> > index 00000000000..cd9c5ccdd0a
>> > --- /dev/null
>> > +++ b/gdb/testsuite/gdb.base/large-frame-1.c
>> > @@ -0,0 +1,32 @@
>> > +/* This file is part of GDB, the GNU debugger.
>> > +
>> > + Copyright 2018 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/>. */
>> > +
>> > +#include "large-frame.h"
>> > +
>> > +__attribute__ ((noinline)) void
>> > +blah (int *a)
>> > +{
>> > + asm ("" :: "m" (a) : "memory");
>> > +}
>> > +
>> > +int
>> > +main ()
>> > +{
>> > + func ();
>> > +
>> > + return 0;
>> > +}
>> > diff --git a/gdb/testsuite/gdb.base/large-frame-2.c b/gdb/testsuite/gdb.base/large-frame-2.c
>> > new file mode 100644
>> > index 00000000000..2491c347292
>> > --- /dev/null
>> > +++ b/gdb/testsuite/gdb.base/large-frame-2.c
>> > @@ -0,0 +1,25 @@
>> > +/* This file is part of GDB, the GNU debugger.
>> > +
>> > + Copyright 2018 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/>. */
>> > +
>> > +#include "large-frame.h"
>> > +
>> > +__attribute__ ((noinline)) int
>> > +func (void)
>> > +{
>> > + int a[4096];
>> > + blah (a);
>> > +}
>> > diff --git a/gdb/testsuite/gdb.base/large-frame.exp b/gdb/testsuite/gdb.base/large-frame.exp
>> > new file mode 100644
>> > index 00000000000..00090da795e
>> > --- /dev/null
>> > +++ b/gdb/testsuite/gdb.base/large-frame.exp
>> > @@ -0,0 +1,57 @@
>> > +# Copyright 2018 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 file is part of the gdb testsuite.
>> > +
>> > +# This test was added to test GDB's ability to backtrace over a large
>> > +# stack frame for which there is no debug information. This should
>> > +# test the non-DWARF stack unwinder.
>> > +#
>> > +# The test was originally added for Risc-V where this case caused a
>> > +# problem at one point, however, there's nothing Risc-V specific in
>> > +# the test.
>> > +
>> > +proc run_test { opt_level } {
>> > + global srcfile srcfile2 binfile hex
>> > +
>> > + standard_testfile large-frame-1.c large-frame-2.c
>> > +
>> > + if {[prepare_for_testing_full "failed to prepare" \
>> > + [list ${binfile}-${opt_level} debug \
>> > + $srcfile [list debug] \
>> > + $srcfile2 [list nodebug optimize=-$opt_level]]]} {
>> > + return
>> > + }
>> > +
>> > + if { ![runto_main] } then {
>> > + unsupported "runto main"
>> > + return
>> > + }
>> > +
>> > + gdb_breakpoint "blah"
>> > + gdb_continue_to_breakpoint "blah"
>> > +
>> > + gdb_test "backtrace" [multi_line \
>> > + "#0 blah \[^\n\r\]+" \
>> > + "#1 $hex in func \[^\n\r\]+" \
>> > + "#2 $hex in main \[^\n\r\]+"]
>> > +}
>> > +
>> > +foreach opt { O0 O1 O2 } {
>> > + with_test_prefix "optimize=-$opt" {
>> > + run_test $opt
>> > + }
>> > +}
>> > +
>> > diff --git a/gdb/testsuite/gdb.base/large-frame.h b/gdb/testsuite/gdb.base/large-frame.h
>> > new file mode 100644
>> > index 00000000000..17058a5efd2
>> > --- /dev/null
>> > +++ b/gdb/testsuite/gdb.base/large-frame.h
>> > @@ -0,0 +1,24 @@
>> > +/* This file is part of GDB, the GNU debugger.
>> > +
>> > + Copyright 2018 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/>. */
>> > +
>> > +#ifndef LARGE_FRAME_H
>> > +#define LARGE_FRAME_H
>> > +
>> > +extern void blah (int *);
>> > +extern int func (void);
>> > +
>> > +#endif /* LARGE_FRAME_H */
>> > --
>> > 2.14.4
>>
>> --
>> Joel
next prev parent reply other threads:[~2018-09-26 17:13 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-19 16:41 Andrew Burgess
2018-09-21 16:20 ` Palmer Dabbelt
2018-09-26 13:13 ` PING: " Andrew Burgess
2018-09-26 13:49 ` Joel Brobecker
2018-09-26 15:11 ` Andrew Burgess
2018-09-26 16:53 ` Joel Brobecker
2018-09-26 17:13 ` Palmer Dabbelt [this message]
2018-09-26 19:15 ` Joel Brobecker
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=mhng-06926dec-9f45-4801-a163-690cc865274b@palmer-si-x1c4 \
--to=palmer@sifive.com \
--cc=andrew.burgess@embecosm.com \
--cc=brobecker@adacore.com \
--cc=gdb-patches@sourceware.org \
--cc=jimw@sifive.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