From: Andreas Arnez <arnez@linux.vnet.ibm.com>
To: gdb-patches@sourceware.org
Subject: [PATCH] PR gdb/21226: Take DWARF stack value pieces from LSB end
Date: Wed, 08 Mar 2017 18:26:00 -0000 [thread overview]
Message-ID: <m3o9xboca5.fsf@oc1027705133.ibm.com> (raw)
When taking a DW_OP_piece or DW_OP_bit_piece from a DW_OP_stack_value, the
existing logic always takes the piece from the lowest-addressed end, which
is wrong on big-endian targets. The DWARF standard states that the
"DW_OP_bit_piece operation describes a sequence of bits using the least
significant bits of that value", and this also matches the current logic
in GCC. For instance, the GCC guality test case pr54970.c fails on s390x
because of this.
This fix adjusts the piece accordingly on big-endian targets. It is
assumed that:
* DW_OP_piece shall take the piece from the LSB end as well;
* pieces reaching outside the stack value bits are considered undefined,
and a zero value can be used instead.
The new logic also respects the DW_OP_bit_piece offset for
DW_OP_stack_values. Previously the offset was ignored. Note that it is
still ignored for all other types of DWARF pieces.
gdb/ChangeLog:
PR gdb/21226
* dwarf2loc.c (read_pieced_value): Anchor stack value pieces at
the LSB end, independent of endianness.
gdb/testsuite/ChangeLog:
PR gdb/21226
* gdb.dwarf2/nonvar-access.exp: Add checks for verifying that
stack value pieces are taken from the LSB end.
---
gdb/dwarf2loc.c | 38 +++++++++++++-----------------
gdb/testsuite/gdb.dwarf2/nonvar-access.exp | 21 ++++++++++++++---
2 files changed, 35 insertions(+), 24 deletions(-)
diff --git a/gdb/dwarf2loc.c b/gdb/dwarf2loc.c
index 4393c1f..764b3fb 100644
--- a/gdb/dwarf2loc.c
+++ b/gdb/dwarf2loc.c
@@ -1868,22 +1868,19 @@ read_pieced_value (struct value *v)
case DWARF_VALUE_STACK:
{
- size_t n = this_size;
+ ULONGEST obj_size = 8 * TYPE_LENGTH (value_type (p->v.value));
- if (n > c->addr_size - source_offset)
- n = (c->addr_size >= source_offset
- ? c->addr_size - source_offset
- : 0);
- if (n == 0)
- {
- /* Nothing. */
- }
- else
- {
- const gdb_byte *val_bytes = value_contents_all (p->v.value);
+ /* Use zeroes if piece reaches beyond stack value. */
+ if (p->offset + p->size > obj_size)
+ goto skip_copy;
- intermediate_buffer = val_bytes + source_offset;
- }
+ /* Piece offset is from least significant bit end. */
+ if (bits_big_endian)
+ source_offset_bits += obj_size - (p->offset + p->size);
+ else
+ source_offset_bits += p->offset;
+ intermediate_buffer = value_contents_all (p->v.value);
+ intermediate_buffer += source_offset_bits / 8;
}
break;
@@ -1903,22 +1900,21 @@ read_pieced_value (struct value *v)
/* These bits show up as zeros -- but do not cause the value
to be considered optimized-out. */
case DWARF_VALUE_IMPLICIT_POINTER:
- break;
+ goto skip_copy;
case DWARF_VALUE_OPTIMIZED_OUT:
mark_value_bits_optimized_out (v, offset, this_size_bits);
- break;
+ goto skip_copy;
default:
internal_error (__FILE__, __LINE__, _("invalid location type"));
}
- if (p->location != DWARF_VALUE_OPTIMIZED_OUT
- && p->location != DWARF_VALUE_IMPLICIT_POINTER)
- copy_bitwise (contents, dest_offset_bits,
- intermediate_buffer, source_offset_bits % 8,
- this_size_bits, bits_big_endian);
+ copy_bitwise (contents, dest_offset_bits,
+ intermediate_buffer, source_offset_bits % 8,
+ this_size_bits, bits_big_endian);
+ skip_copy:
offset += this_size_bits;
}
}
diff --git a/gdb/testsuite/gdb.dwarf2/nonvar-access.exp b/gdb/testsuite/gdb.dwarf2/nonvar-access.exp
index 3cb5c49..69e01ca 100644
--- a/gdb/testsuite/gdb.dwarf2/nonvar-access.exp
+++ b/gdb/testsuite/gdb.dwarf2/nonvar-access.exp
@@ -128,14 +128,26 @@ Dwarf::assemble $asm_file {
{name def_t}
{type :$struct_t_label}
{location {
- const1u 0
+ const2s -184
stack_value
bit_piece 9 0
- const1s -1
+ const4u 1752286
stack_value
bit_piece 23 0
} SPECIAL_expr}
}
+ # Composite location with some empty pieces.
+ DW_TAG_variable {
+ {name part_def_a}
+ {type :$array_a9_label}
+ {location {
+ piece 3
+ const4u 0xf1927314
+ stack_value
+ piece 4
+ piece 2
+ } SPECIAL_expr}
+ }
# Implicit location: immediate value.
DW_TAG_variable {
{name def_implicit_s}
@@ -197,9 +209,12 @@ gdb_test "print/x *(char (*)\[5\]) implicit_a_ptr" \
# Byte-aligned fields, pieced together from DWARF stack values.
gdb_test "print def_s" " = \\{a = 0, b = -1\\}"
+switch $endian { big {set val 0x92} little {set val 0x73} }
+gdb_test "print/x part_def_a\[4\]" " = $val"
+gdb_test "print/x part_def_a\[8\]" " = <optimized out>"
# Non-byte-aligned fields, pieced together from DWARF stack values.
-gdb_test "print def_t" " = \\{a = 0, b = -1\\}"
+gdb_test "print def_t" " = \\{a = -184, b = 1752286\\}"
# Simple variable without location.
gdb_test "print undef_int" " = <optimized out>"
--
2.3.0
next reply other threads:[~2017-03-08 18:26 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-08 18:26 Andreas Arnez [this message]
2017-03-10 17:43 ` Ulrich Weigand
2017-03-10 19:24 ` Andreas Arnez
2017-03-10 17:57 ` Ulrich Weigand
2017-03-10 19:27 ` Andreas Arnez
2017-03-10 20:01 ` Ulrich Weigand
2017-03-13 12:18 ` Andreas Arnez
2017-03-10 19:29 ` Pedro Alves
2017-03-13 12:24 ` Andreas Arnez
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=m3o9xboca5.fsf@oc1027705133.ibm.com \
--to=arnez@linux.vnet.ibm.com \
--cc=gdb-patches@sourceware.org \
/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