[PATCH] Fix Ada value printing on PPC64

[PATCH] Fix Ada value printing on PPC64

[Tom Tromey]

The val_print removal patches introduced an Ada regression on PPC64
(probably any big-endian system).

The issue comes because value_field does not understand that Ada
wrapper fields can be bitfields that wrap a non-scalar type.  In this
case the value is already left-justified, so the justification done
there does the wrong thing.

Perhaps it would be good, eventually, to change value_field to
understand this case.  In the meantime this implements an Ada-specific
solution.

gdb/ChangeLog
2020-04-30  Tom Tromey  <tromey@adacore.com>

	* ada-lang.c (ada_value_primitive_field): Now public.
	* ada-lang.h (ada_value_primitive_field): Declare.
	* ada-valprint.c (print_field_values): Use
	ada_value_primitive_field for wrapper fields.
---
 gdb/ChangeLog      | 7 +++++++
 gdb/ada-lang.c     | 5 +----
 gdb/ada-lang.h     | 5 +++++
 gdb/ada-valprint.c | 3 ++-
 4 files changed, 15 insertions(+), 5 deletions(-)

diff --git a/gdb/ada-lang.c b/gdb/ada-lang.c
index bfbc69084ec..e25930fca21 100644
--- a/gdb/ada-lang.c
+++ b/gdb/ada-lang.c
@@ -204,9 +204,6 @@ static struct symbol *standard_lookup (const char *, const struct block *,
 static struct value *ada_search_struct_field (const char *, struct value *, int,
                                               struct type *);
 
-static struct value *ada_value_primitive_field (struct value *, int, int,
-                                                struct type *);
-
 static int find_struct_field (const char *, struct type *, int,
                               struct type **, int *, int *, int *, int *);
 
@@ -7125,7 +7122,7 @@ ada_in_variant (LONGEST val, struct type *type, int field_num)
    fields.  FIELDNO says which field.   Differs from value_primitive_field
    only in that it can handle packed values of arbitrary type.  */
 
-static struct value *
+struct value *
 ada_value_primitive_field (struct value *arg1, int offset, int fieldno,
                            struct type *arg_type)
 {
diff --git a/gdb/ada-lang.h b/gdb/ada-lang.h
index bb9e3c3027b..c0a71091e35 100644
--- a/gdb/ada-lang.h
+++ b/gdb/ada-lang.h
@@ -234,6 +234,11 @@ extern struct bound_minimal_symbol ada_lookup_simple_minsym (const char *);
 
 extern int ada_scan_number (const char *, int, LONGEST *, int *);
 
+extern struct value *ada_value_primitive_field (struct value *arg1,
+						int offset,
+						int fieldno,
+						struct type *arg_type);
+
 extern struct type *ada_parent_type (struct type *);
 
 extern int ada_is_ignored_field (struct type *, int);
diff --git a/gdb/ada-valprint.c b/gdb/ada-valprint.c
index 31f3a50b340..010446d70f8 100644
--- a/gdb/ada-valprint.c
+++ b/gdb/ada-valprint.c
@@ -609,7 +609,8 @@ print_field_values (struct value *value, struct value *outer_value,
 
       if (ada_is_wrapper_field (type, i))
 	{
-	  struct value *field_val = value_field (value, i);
+	  struct value *field_val = ada_value_primitive_field (value, 0,
+							       i, type);
 	  comma_needed =
 	    print_field_values (field_val, field_val,
 				stream, recurse, options,
-- 
2.21.1

Re: [PATCH] Fix Ada value printing on PPC64

[Tom Tromey]

>>>>> "Tom" == Tom Tromey <tromey@adacore.com> writes:

Tom> The val_print removal patches introduced an Ada regression on PPC64
Tom> (probably any big-endian system).

Tom> The issue comes because value_field does not understand that Ada
Tom> wrapper fields can be bitfields that wrap a non-scalar type.  In this
Tom> case the value is already left-justified, so the justification done
Tom> there does the wrong thing.

Tom> Perhaps it would be good, eventually, to change value_field to
Tom> understand this case.  In the meantime this implements an Ada-specific
Tom> solution.

I'm checking this in.

Tom