Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Simon Marchi <simark@simark.ca>
To: Gregori Mignerot <gregori.mignerot@laposte.net>,
	gdb-patches@sourceware.org
Subject: Re: [PATCH v2] [gdb] Support default locations in DWARF5 loclists
Date: Sun, 26 Apr 2026 23:12:56 -0400	[thread overview]
Message-ID: <ac041e7a-5fe3-4342-b97e-90cd0408b7cb@simark.ca> (raw)
In-Reply-To: <20260426115337.1245133-2-gregori.mignerot@laposte.net>



On 2026-04-26 07:40, Gregori Mignerot wrote:
> Add support and a test for default location entries in DWARF5 loclists
> (DW_LLE_default_location, section 2.6.2 in the DWARF5 standard)
> ---
> Applied the review comments and added the test for `info address`,
> thanks for the quick feedback !
> 
>> Just wondering, is this valid DWARF?  The spec implies that the default
>> location has the be last (if present):
>>
>>     A location list consists of a sequence of zero or more bounded
>>     location description or base address entries, optionally followed by
>>     a default location entry, and terminated by an end-of-list entry.
>>
>> I'm not sure why DWARF 5 requires that, it seems to me like it would
>> work just as well if the default location was allowed to be anywhere in
>> the list.  Do you know about producers that do this?
> 
> I'm not sure either, there isn't a lot of producers or consumers for
> this. LLVM and libdwarf don't seem to care about the position of the
> default location entry.
> I wasn't sure either about that test, it's probably superfluous, removed
> it in v2.

Are you able to mention some producers that produce default locations,
so we can look?

If we don't need to support non-conforming implementations, I think we
could stick to what the spec says, it would make the code a little bit
simpler: when encountering a DW_LLE_default_location, you can return
this that expression right away, instead of saving it to a local
variable to be returned later.

> diff --git a/gdb/dwarf2/loc.c b/gdb/dwarf2/loc.c
> index a493a46dcfc..c728fa9c4b1 100644
> --- a/gdb/dwarf2/loc.c
> +++ b/gdb/dwarf2/loc.c
> @@ -83,6 +83,10 @@ enum debug_loc_kind
>       the applicable base address.  */
>    DEBUG_LOC_OFFSET_PAIR = 4,
>  
> +  /* This is followed by a normal location expression as in .debug_loc,
> +     to use when no other entry matches the current PC */

Finish the sentence with <dot><space><space>:

  /* This is followed by a normal location expression as in .debug_loc,
     to use when no other entry matches the current PC.  */

The concept of "default location" appeared in DWARF 5, right?  If so, I
think this comment should say ".debug_loclists" instead of ".debug_loc".

> @@ -379,6 +386,11 @@ dwarf2_find_location_expression (const dwarf2_loclist_baton *baton,
>    unrelocated_addr base_address = baton->base_address;
>    const gdb_byte *loc_ptr, *buf_end;
>  
> +  /* When no DEBUG_LOC_DEFAULT_LOCATION is found
> +     return NULL with length 0 (= no location found) */

Finish sentence with <dot><space><space>.

> +  const gdb_byte *default_loc_ptr = NULL;

nullptr (instead of NULL) for new (or modified) code.

> @@ -4042,10 +4063,11 @@ loclist_describe_location (struct symbol *symbol, CORE_ADDR addr,
>  	  gdb_printf (stream, _("  Base address %s"),
>  		      paddress (gdbarch, (CORE_ADDR) base_address));
>  	  continue;
> -
> +	

Spurious tab added.

> @@ -4057,6 +4079,29 @@ loclist_describe_location (struct symbol *symbol, CORE_ADDR addr,
>  	  gdb_assert_not_reached ("bad debug_loc_kind");
>  	}
>  
> +      if (dlbaton->dwarf_version < 5)
> +	{
> +	  length = extract_unsigned_integer (loc_ptr, 2, byte_order);
> +	  loc_ptr += 2;
> +	}
> +      else
> +	{
> +	  unsigned int bytes_read;
> +	  length = read_unsigned_leb128 (NULL, loc_ptr, &bytes_read);
> +	  loc_ptr += bytes_read;
> +	}
> +
> +      if (kind == DEBUG_LOC_DEFAULT_LOCATION)
> +	{
> +	  gdb_printf (stream, _("  Default location: "));
> +	  locexpr_describe_location_1 (symbol, addr, stream, loc_ptr, length,
> +				       addr_size, offset_size,
> +				dlbaton->per_cu, per_objfile);

Alight the last line with the one before:

	  locexpr_describe_location_1 (symbol, addr, stream, loc_ptr, length,
				       addr_size, offset_size,
				       dlbaton->per_cu, per_objfile);

> +	  gdb_printf(stream, "\n");

Space before parenthesis.

> diff --git a/gdb/testsuite/gdb.dwarf2/loclists-default-location.exp b/gdb/testsuite/gdb.dwarf2/loclists-default-location.exp
> new file mode 100644
> index 00000000000..091cc0e01ed
> --- /dev/null
> +++ b/gdb/testsuite/gdb.dwarf2/loclists-default-location.exp
> @@ -0,0 +1,130 @@
> +# Copyright 2026 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/>.
> +
> +# Test support for DW_LLE_default_location
> +
> +load_lib dwarf.exp
> +
> +require dwarf2_support
> +
> +# Test with 32-bit and 64-bit DWARF.
> +foreach_with_prefix is_64 {false true} {
> +    if { $is_64 } {
> +	standard_testfile .c -dw64.S
> +	set testfile ${testfile}-dw64
> +    } else {
> +	standard_testfile .c -dw32.S
> +	set testfile ${testfile}-dw32
> +    }
> +
> +    # Get the addresses / lengths of func1 and func2.
> +    lassign [function_range func1 $srcdir/$subdir/$srcfile] func1_addr func1_len
> +    lassign [function_range func2 $srcdir/$subdir/$srcfile] func2_addr func2_len
> +
> +    set asm_file [standard_output_file $srcfile2]
> +    Dwarf::assemble $asm_file {
> +	global func1_addr func1_len
> +	global func2_addr func2_len
> +	global is_64
> +
> +	# The CU uses the DW_FORM_loclistx form to refer to the .debug_loclists
> +	# section.
> +	cu {
> +	    version 5
> +	    is_64 $is_64
> +	} {
> +	    declare_labels int_type
> +
> +	    DW_TAG_compile_unit {
> +		DW_AT_loclists_base cu_table DW_FORM_sec_offset
> +	    } {
> +		int_type: DW_TAG_base_type {
> +		    DW_AT_byte_size 4 DW_FORM_data1
> +		    DW_AT_encoding @DW_ATE_signed
> +		    DW_AT_name "int"
> +		}
> +
> +		DW_TAG_variable {
> +		    DW_AT_name "x"
> +		    DW_AT_location 0 DW_FORM_loclistx
> +		    DW_AT_type :$int_type
> +		}
> +
> +		DW_TAG_subprogram {
> +		    DW_AT_name "func1"
> +		    DW_AT_low_pc $func1_addr
> +		    DW_AT_high_pc $func1_len DW_FORM_udata
> +		}
> +
> +		DW_TAG_subprogram {
> +		    DW_AT_name "func2"
> +		    DW_AT_low_pc $func2_addr
> +		    DW_AT_high_pc $func2_len DW_FORM_udata
> +		}
> +	    }
> +	}
> +
> +	loclists {is-64 $is_64} {
> +	    # The lists in this table are accessed by index (DW_FORM_loclistx).
> +	    table {post-header-label cu_table} {
> +		# With the default location last in the list (default_last)

This comment is stale.

Simon

      reply	other threads:[~2026-04-27  3:13 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-26 11:40 Gregori Mignerot
2026-04-27  3:12 ` Simon Marchi [this message]

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=ac041e7a-5fe3-4342-b97e-90cd0408b7cb@simark.ca \
    --to=simark@simark.ca \
    --cc=gdb-patches@sourceware.org \
    --cc=gregori.mignerot@laposte.net \
    /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