public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Simon Marchi <simon.marchi@polymtl.ca>
To: Omair Javaid <omair.javaid@linaro.org>
Cc: gdb-patches@sourceware.org, Tom Tromey <tom@tromey.com>
Subject: Re: [PATCH] [PR gdb/21695] Fix lost line info for symbol at addr zero
Date: Wed, 20 Jun 2018 16:12:00 -0000	[thread overview]
Message-ID: <ea8c34708df7c527edde40ac14a216a8@polymtl.ca> (raw)
In-Reply-To: <1526393061-18039-1-git-send-email-omair.javaid@linaro.org>

On 2018-05-15 10:04, Omair Javaid wrote:
> This patch fixes a unique condition where GDB fails to provide line
> information of symbol at address zero when code is compiled with text
> address zero but loaded at an offset > 0.
> 
> For example lets compile following code snippet:
> 
> int main() {
>   return 0;
> }
> 
> gcc -O0 -g3 -nostdlib -emain -Wl,-Ttext=0x00 -o file.out file.c
> 
> Start gdb and run:
> 
> add-symbol-file file.out 0xffff0000
> info line main
> 
> GDB will return error saying no line info is available for the symbol.
> 
> This is a direct consequence of the fix for PR 12528 where GDB tries to 
> ignore
> line table for a function which has been garbage collected by the 
> linker.
> 
> As the garbage collected symbols are sent to address zero GDB assumes a 
> symbol
> actually placed at address zero as garbage collected.
> 
> This was fixed with an additional check address < lowpc. But when 
> symbol is
> loaded at an offset lowpc becomes lowpc + offset while no offset is 
> added to
> address rather final symbol address is calculated based on baseaddr and 
> address
> added together. So in case where symbols are loaded at an offset the 
> condition
> address < lowpc will always return true.
> 
> This patch fixes this by comparing address against a non offset lowpc.
> This patch also adds a GDB test case to replicate this behavior.

Hi Omair,

Sorry for the wait.

I have tried your patch and see that it fixes the bug you report, but I 
am not sure I understand your explanation.  It sounds like we are 
comparing an unrelocated address with a relocated low_pc.  What your 
patch does is make sure we compare apples to apples, is that right?

I have added Tom Tromey in CC, he is currently playing with storing 
symbols relocated vs unrelocated, so maybe he knows better.  And maybe 
there's something in his pending patches that also touch this area.

> gdb:
> 
> 2018-05-15  Omair Javaid  <omair.javaid@linaro.org>
> 
> 	PR gdb/21695
> 	* dwarf2read.c (lnp_state_machine::check_line_address): Update 
> declaration.
> 	(dwarf_decode_lines_1): Adjust.
> 
> gdb/testsuite:
> 
> 2018-05-15  Omair Javaid  <omair.javaid@linaro.org>
> 
> 	PR gdb/21695
> 	* gdb.base/infoline-reloc-main-from-zero.exp: New test.
> 	* gdb.base/infoline-reloc-main-from-zero.c: New file.
> ---
>  gdb/dwarf2read.c                                   |  8 ++--
>  .../gdb.base/infoline-reloc-main-from-zero.c       | 24 ++++++++++
>  .../gdb.base/infoline-reloc-main-from-zero.exp     | 51 
> ++++++++++++++++++++++
>  3 files changed, 79 insertions(+), 4 deletions(-)
>  create mode 100644 
> gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.c
>  create mode 100644 
> gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.exp
> 
> diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
> index 575d316..b171bb2 100644
> --- a/gdb/dwarf2read.c
> +++ b/gdb/dwarf2read.c
> @@ -20411,7 +20411,7 @@ public:
>       sequence.  */
>    void check_line_address (struct dwarf2_cu *cu,
>  			   const gdb_byte *line_ptr,
> -			   CORE_ADDR lowpc, CORE_ADDR address);
> +			   CORE_ADDR lowpc_minus_base, CORE_ADDR address);
> 
>    void handle_set_discriminator (unsigned int discriminator)
>    {
> @@ -20755,14 +20755,14 @@ lnp_state_machine::lnp_state_machine
> (gdbarch *arch, line_header *lh,
>  void
>  lnp_state_machine::check_line_address (struct dwarf2_cu *cu,
>  				       const gdb_byte *line_ptr,
> -				       CORE_ADDR lowpc, CORE_ADDR address)
> +				       CORE_ADDR lowpc_minus_base, CORE_ADDR address)

I would suggest naming this "unrelocated_lowpc".  Can you update the 
documentation of the method to describe that parameter?

>  {
>    /* If address < lowpc then it's not a usable value, it's outside the
>       pc range of the CU.  However, we restrict the test to only 
> address
>       values of zero to preserve GDB's previous behaviour which is to
>       handle the specific case of a function being GC'd by the linker.  
> */
> 
> -  if (address == 0 && address < lowpc)
> +  if (address == 0 && address < lowpc_minus_base)
>      {
>        /* This line table is for a function which has been
>  	 GCd by the linker.  Ignore it.  PR gdb/12528 */
> @@ -20857,7 +20857,7 @@ dwarf_decode_lines_1 (struct line_header *lh,
> struct dwarf2_cu *cu,
>  		    line_ptr += bytes_read;
> 
>  		    state_machine.check_line_address (cu, line_ptr,
> -						      lowpc, address);
> +						      lowpc - baseaddr, address);
>  		    state_machine.handle_set_address (baseaddr, address);
>  		  }
>  		  break;
> diff --git a/gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.c
> b/gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.c
> new file mode 100644
> index 0000000..8902051
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.c
> @@ -0,0 +1,24 @@
> +/* This testcase is part of GDB, the GNU debugger.
> +
> +   Copyright 2011-2018 Free Software Foundation, Inc.

This should probably be just 2018.

> +
> +   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/>.  */
> +
> +// Test case for PR gdb/21695
> +
> +int
> +main ()
> +{
> +  return 0;
> +}
> diff --git a/gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.exp
> b/gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.exp
> new file mode 100644
> index 0000000..68564688
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/infoline-reloc-main-from-zero.exp
> @@ -0,0 +1,51 @@
> +# Copyright 2011-2018 Free Software Foundation, Inc.

Here too.

> +
> +# 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
> +
> +#
> +# test running programs
> +#

Can you update this comment to reflect what this test does?

> +
> +standard_testfile .c
> +
> +if [get_compiler_info] {
> +    return -1
> +}
> +
> +# Build executable with stripped startup code and text section 
> starting at zero
> +
> +set opts {debug "additional_flags=-nostdlib -emain -Wl,-Ttext=0x00"}
> +
> +if {[build_executable $testfile.exp $testfile $srcfile $opts] == -1} {
> +    untested "failed to compile"
> +    return -1
> +}
> +
> +gdb_exit
> +gdb_start
> +gdb_reinitialize_dir $srcdir/$subdir

You can use clean_restart instead of these last 3 lines.

> +
> +# Load symbols at an offset 0xffff0000 using add-symbol-file
> +
> +gdb_test "add-symbol-file [standard_output_file ${testfile}] 
> 0xffff000" \

Use ${binfile} instead of [standard_output_file ${testfile}].

> +    "Reading symbols from .*" \
> +    "add-symbol-file" \
> +    "add symbol table from file \".*\" at.*\\(y or n\\) " "y"
> +
> +# Check if we are able to read offset adjusted line information of 
> main
> +
> +gdb_test "info line main" \
> +	"Line.*starts at address.*and ends at.*"

Thanks,

Simon

  parent reply	other threads:[~2018-06-20 16:12 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-05-15 14:56 Omair Javaid
2018-05-28 12:36 ` Omair Javaid
2018-06-12  2:26   ` Omair Javaid
2018-06-20  7:53     ` Omair Javaid
2018-06-20 16:12 ` Simon Marchi [this message]
2018-06-21 18:27   ` Tom Tromey
2018-06-25 10:11     ` Omair Javaid

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=ea8c34708df7c527edde40ac14a216a8@polymtl.ca \
    --to=simon.marchi@polymtl.ca \
    --cc=gdb-patches@sourceware.org \
    --cc=omair.javaid@linaro.org \
    --cc=tom@tromey.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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).