[RFA] Allow "info address" of a template parameter

Message ID 20180720131942.6958-1-tom@tromey.com
State New, archived
Headers

Commit Message

Tom Tromey July 20, 2018, 1:19 p.m. UTC
  PR symtab/16842 shows that gdb will crash when the user tries to
invoke "info address" of a template parameter.

The bug here is that dwarf2read.c does not set the symtab on the
template parameter symbols.  This is pedantically correct, given that
the template symbols do not appear in a symtab.  However, gdb
primarily uses the symtab backlink to find the symbol's objfile.  So,
this patch simply sets the symtab on these symbols.

Tested by the buildbot.

gdb/ChangeLog
2018-07-20  Tom Tromey  <tom@tromey.com>

	PR symtab/16842.
	* dwarf2read.c (read_func_scope): Set symtab on template parameter
	symbols.
	(process_structure_scope): Likewise.

gdb/testsuite/ChangeLog
2018-07-20  Tom Tromey  <tom@tromey.com>

	PR symtab/16842.
	* gdb.cp/temargs.exp: Test "info address" of a template
	parameter.
---
 gdb/ChangeLog                    |  7 +++++++
 gdb/dwarf2read.c                 | 24 +++++++++++++++++++++++-
 gdb/testsuite/ChangeLog          |  6 ++++++
 gdb/testsuite/gdb.cp/temargs.exp |  8 ++++++++
 4 files changed, 44 insertions(+), 1 deletion(-)
  

Comments

Keith Seitz Aug. 2, 2018, 7:28 p.m. UTC | #1
On 07/20/2018 06:19 AM, Tom Tromey wrote:
> gdb/ChangeLog
> 2018-07-20  Tom Tromey  <tom@tromey.com>
> 
> 	PR symtab/16842.
> 	* dwarf2read.c (read_func_scope): Set symtab on template parameter
> 	symbols.
> 	(process_structure_scope): Likewise.
> 
> gdb/testsuite/ChangeLog
> 2018-07-20  Tom Tromey  <tom@tromey.com>
> 
> 	PR symtab/16842.
> 	* gdb.cp/temargs.exp: Test "info address" of a template
> 	parameter.

Since this doesn't appear to be checked-in yet...

This looks reasonable to me, FWIW. I urge you to approve your patch. ;-)

Keith
  
Tom Tromey Aug. 2, 2018, 10:06 p.m. UTC | #2
>>>>> "Keith" == Keith Seitz <keiths@redhat.com> writes:

Keith> On 07/20/2018 06:19 AM, Tom Tromey wrote:
>> gdb/ChangeLog
>> 2018-07-20  Tom Tromey  <tom@tromey.com>
>> 
>> PR symtab/16842.
>> * dwarf2read.c (read_func_scope): Set symtab on template parameter
>> symbols.
>> (process_structure_scope): Likewise.
>> 
>> gdb/testsuite/ChangeLog
>> 2018-07-20  Tom Tromey  <tom@tromey.com>
>> 
>> PR symtab/16842.
>> * gdb.cp/temargs.exp: Test "info address" of a template
>> parameter.

Keith> Since this doesn't appear to be checked-in yet...

Keith> This looks reasonable to me, FWIW. I urge you to approve your patch. ;-)

Thanks, I'm going to check it in.

Tom
  

Patch

diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index 302a0f493a4..c3656d2ca3c 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,10 @@ 
+2018-07-20  Tom Tromey  <tom@tromey.com>
+
+	PR symtab/16842.
+	* dwarf2read.c (read_func_scope): Set symtab on template parameter
+	symbols.
+	(process_structure_scope): Likewise.
+
 2018-07-19  Pedro Alves  <palves@redhat.com>
 
 	* guile/guile-internal.h (gdbscm_scm_to_c_string): Now returns a
diff --git a/gdb/dwarf2read.c b/gdb/dwarf2read.c
index b7933de49cf..8f5fb597161 100644
--- a/gdb/dwarf2read.c
+++ b/gdb/dwarf2read.c
@@ -13714,6 +13714,13 @@  read_func_scope (struct die_info *die, struct dwarf2_cu *cu)
       memcpy (templ_func->template_arguments,
 	      template_args.data (),
 	      (templ_func->n_template_arguments * sizeof (struct symbol *)));
+
+      /* Make sure that the symtab is set on the new symbols.  Even
+	 though they don't appear in this symtab directly, other parts
+	 of gdb assume that symbols do, and this is reasonably
+	 true.  */
+      for (struct symbol *sym : template_args)
+	symbol_set_symtab (sym, symbol_symtab (templ_func));
     }
 
   /* In C++, we can have functions nested inside functions (e.g., when
@@ -15832,6 +15839,7 @@  process_structure_scope (struct die_info *die, struct dwarf2_cu *cu)
      discriminant_info.  */
   bool is_variant_part = TYPE_FLAG_DISCRIMINATED_UNION (type);
   sect_offset discr_offset;
+  bool has_template_parameters = false;
 
   if (is_variant_part)
     {
@@ -15879,6 +15887,7 @@  process_structure_scope (struct die_info *die, struct dwarf2_cu *cu)
       /* Attach template arguments to type.  */
       if (!template_args.empty ())
 	{
+	  has_template_parameters = true;
 	  ALLOCATE_CPLUS_STRUCT_TYPE (type);
 	  TYPE_N_TEMPLATE_ARGUMENTS (type) = template_args.size ();
 	  TYPE_TEMPLATE_ARGUMENTS (type)
@@ -16028,7 +16037,20 @@  process_structure_scope (struct die_info *die, struct dwarf2_cu *cu)
      attribute, and a declaration attribute.  */
   if (dwarf2_attr (die, DW_AT_byte_size, cu) != NULL
       || !die_is_declaration (die, cu))
-    new_symbol (die, type, cu);
+    {
+      struct symbol *sym = new_symbol (die, type, cu);
+
+      if (has_template_parameters)
+	{
+	  /* Make sure that the symtab is set on the new symbols.
+	     Even though they don't appear in this symtab directly,
+	     other parts of gdb assume that symbols do, and this is
+	     reasonably true.  */
+	  for (int i = 0; i < TYPE_N_TEMPLATE_ARGUMENTS (type); ++i)
+	    symbol_set_symtab (TYPE_TEMPLATE_ARGUMENT (type, i),
+			       symbol_symtab (sym));
+	}
+    }
 }
 
 /* Assuming DIE is an enumeration type, and TYPE is its associated type,
diff --git a/gdb/testsuite/ChangeLog b/gdb/testsuite/ChangeLog
index 79ef76faf76..7039518ac3d 100644
--- a/gdb/testsuite/ChangeLog
+++ b/gdb/testsuite/ChangeLog
@@ -1,3 +1,9 @@ 
+2018-07-20  Tom Tromey  <tom@tromey.com>
+
+	PR symtab/16842.
+	* gdb.cp/temargs.exp: Test "info address" of a template
+	parameter.
+
 2018-07-19  Tom de Vries  <tdevries@suse.de>
 
 	* gdb.mi/list-thread-groups-available.exp (cores_re): Fix quoting in
diff --git a/gdb/testsuite/gdb.cp/temargs.exp b/gdb/testsuite/gdb.cp/temargs.exp
index 8c87872225e..5a109102d75 100644
--- a/gdb/testsuite/gdb.cp/temargs.exp
+++ b/gdb/testsuite/gdb.cp/temargs.exp
@@ -133,6 +133,10 @@  gdb_test "ptype T" "unsigned char" "test type of T in func"
 if $have_older_template_gcc { setup_xfail "*-*-*" }
 gdb_test "print I" " = 91" "test value of I in func"
 
+# PR symtab/16842 - gdb used to crash here.
+if $have_older_template_gcc { setup_xfail "*-*-*" }
+gdb_test "info addr I" "Symbol \"I\" is constant." "test address of I in templ_m"
+
 if $have_older_template_gcc { setup_xfail "*-*-*" }
 gdb_test "print P == &a_global" " = true" "test value of P in func"
 
@@ -151,6 +155,10 @@  gdb_test "ptype T" "double" "test type of T in templ_m"
 if $have_older_template_gcc { setup_xfail "*-*-*" }
 gdb_test "print I" " = 23" "test value of I in templ_m"
 
+# PR symtab/16842 - gdb used to crash here.
+if $have_older_template_gcc { setup_xfail "*-*-*" }
+gdb_test "info addr I" "Symbol \"I\" is constant." "test address of I in templ_m"
+
 if $have_older_template_gcc { setup_xfail "*-*-*" }
 gdb_test "print P == &a_global" " = true" "test value of P in templ_m"