[PATCHv2,4/4] gdb: Allow struct fields named double

Message ID 41fdb503cf18c86986aaf70dd2387c42f52502ee.1545172667.git.andrew.burgess@embecosm.com
State New, archived
Headers

Commit Message

Andrew Burgess Dec. 18, 2018, 10:40 p.m. UTC
  The 64-bit RISC-V target currently models the floating point registers
as having the following type:

    union riscv_double
    {
        builtin_type_ieee_single float;
        builtin_type_ieee_double double;
    }

Notice the choice of names for the fields of this struct, possibly not
ideal choices, as these are not valid field names in C.  However, this
type is only ever defined within GDB (or in the target description),
and no restriction seems to exist on the field names in that case.

The problem though is that currently:

    (gdb) info registers $ft0
    ft0            {float = 0, double = 0}	(raw 0x0000000000000000)
    (gdb) p $ft0.float
    $1 = 0
    (gdb) p $ft0.double
    A syntax error in expression, near `double'.

We can access the 'float' field, but not the 'double' field.  This is
because the string 'double' is handled differently to the string
'float' in c-exp.y.

In both cases the string '$ft0' is parsed as a VARIABLE expression.

In the 'float' case, the string 'float' becomes a generic NAME token
in 'lex_one_token', which then allows the rule "exp '.' name" to match
and the field name lookup to occur.

The 'double' case is different.  In order to allow parsing of the type
string 'long double', the 'double' string becomes the token
DOUBLE_KEYWORD.  At this point there's no rule to match "exp '.'
DOUBLE_KEYWORD", so we can never lookup the field named 'double'.

We could rename the fields for RISC-V, and maybe that would be the
best solution.  However, its not hard to allow for fields named
'double', which is what this patch does.

A new case is added to the 'field_name' rule to match the
DOUBLE_KEYWORD, and create a suitable 'struct stoken'.  With this done
the "exp '.'  field_name" pattern can now match, and we can lookup the
double field.

With this patch in place I now see this behaviour:

    (gdb) info registers $ft0
    ft0            {float = 0, double = 0}	(raw 0x0000000000000000)
    (gdb) p $ft0.float
    $1 = 0
    (gdb) p $ft0.double
    $2 = 0

I've gone ahead and handled INT_KEYWORD, LONG, SHORT, SIGNED_KEYWORD,
and UNSIGNED as well within field_name.

I've added a new test for this functionality.

This change was tested on x86-64 GNU/Linux with no regressions.

gdb/ChangeLog:

	* c-exp.y (field_name): Allow DOUBLE_KEYWORD, INT_KEYWORD, LONG,
	SHORT, SIGNED_KEYWORD, and UNSIGNED tokens to act as a field
	names.
	(typename_stoken): New function.

gdb/testsuite/ChangeLog:

	* gdb.dwarf2/dw2-unusual-field-names.c: New file.
	* gdb.dwarf2/dw2-unusual-field-names.exp: New file.
---
 gdb/ChangeLog                                      |   7 ++
 gdb/c-exp.y                                        |  25 +++-
 gdb/testsuite/ChangeLog                            |   5 +
 gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.c |  32 +++++
 .../gdb.dwarf2/dw2-unusual-field-names.exp         | 132 +++++++++++++++++++++
 5 files changed, 200 insertions(+), 1 deletion(-)
 create mode 100644 gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.c
 create mode 100644 gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.exp
  

Comments

Tom Tromey Dec. 21, 2018, 5:19 p.m. UTC | #1
>>>>> "Andrew" == Andrew Burgess <andrew.burgess@embecosm.com> writes:

Andrew> gdb/ChangeLog:

Andrew> 	* c-exp.y (field_name): Allow DOUBLE_KEYWORD, INT_KEYWORD, LONG,
Andrew> 	SHORT, SIGNED_KEYWORD, and UNSIGNED tokens to act as a field
Andrew> 	names.
Andrew> 	(typename_stoken): New function.

Andrew> gdb/testsuite/ChangeLog:

Andrew> 	* gdb.dwarf2/dw2-unusual-field-names.c: New file.
Andrew> 	* gdb.dwarf2/dw2-unusual-field-names.exp: New file.

Thanks for doing this.  This is ok.  I wanted to mention that I think
it's a good idea to do this because the C language also serves as a
fallback either when language support isn't available or when one wants
to inspect things in a low-level way, and in this case it can be useful
to have some extensions to C.

Andrew> +foreach field_name { double float char byte long int short unsigned signed } {
Andrew> +    run_test $field_name
Andrew> +}

I'd like to verify that there aren't any duplicate test names resulting
from this approach.  If there aren't, this is fine; if there are, I
think run_test could be wrapped in a with_test_prefix.

thanks,
Tom
  

Patch

diff --git a/gdb/c-exp.y b/gdb/c-exp.y
index b6b57c8c4d4..44e8eeadd7a 100644
--- a/gdb/c-exp.y
+++ b/gdb/c-exp.y
@@ -113,6 +113,7 @@  static int type_aggregate_p (struct type *);
 static int parse_number (struct parser_state *par_state,
 			 const char *, int, int, YYSTYPE *);
 static struct stoken operator_stoken (const char *);
+static struct stoken typename_stoken (const char *);
 static void check_parameter_typelist (VEC (type_ptr) *);
 static void write_destructor_name (struct parser_state *par_state,
 				   struct stoken);
@@ -1643,9 +1644,21 @@  oper:	OPERATOR NEW
 			}
 	;
 
-
+/* This rule exists in order to allow some tokens that would not normally
+   match the 'name' rule to appear as fields within a struct.  The example
+   that initially motivated this was the RISC-V target which models the
+   floating point registers as a union with fields called 'float' and
+   'double'.  The 'float' string becomes a TYPENAME token and can appear
+   anywhere a 'name' can, however 'double' is its own token,
+   DOUBLE_KEYWORD, and doesn't match the 'name' rule.*/
 field_name
 	:	name
+	|	DOUBLE_KEYWORD { $$ = typename_stoken ("double"); }
+	|	INT_KEYWORD { $$ = typename_stoken ("int"); }
+	|	LONG { $$ = typename_stoken ("long"); }
+	|	SHORT { $$ = typename_stoken ("short"); }
+	|	SIGNED_KEYWORD { $$ = typename_stoken ("signed"); }
+	|	UNSIGNED { $$ = typename_stoken ("unsigned"); }
 	;
 
 name	:	NAME { $$ = $1.stoken; }
@@ -1718,6 +1731,16 @@  operator_stoken (const char *op)
   return st;
 };
 
+/* Returns a stoken of the type named TYPE.  */
+
+static struct stoken
+typename_stoken (const char *type)
+{
+  struct stoken st = { type, 0 };
+  st.length = strlen (type);
+  return st;
+};
+
 /* Return true if the type is aggregate-like.  */
 
 static int
diff --git a/gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.c b/gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.c
new file mode 100644
index 00000000000..091e475fb3f
--- /dev/null
+++ b/gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.c
@@ -0,0 +1,32 @@ 
+/* This testcase is part of GDB, the GNU debugger.
+
+   Copyright 2018 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/>.  */
+
+struct foo
+{
+  int field;
+};
+
+struct foo obj = { 0 };
+
+struct foo *ptr = &obj;
+
+int
+main (void)
+{
+  return obj.field;
+}
+
diff --git a/gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.exp b/gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.exp
new file mode 100644
index 00000000000..6443acee249
--- /dev/null
+++ b/gdb/testsuite/gdb.dwarf2/dw2-unusual-field-names.exp
@@ -0,0 +1,132 @@ 
+# Copyright 2018 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 that GDB can support accesing fields of a structure if the
+# fields have non-standard names.  Specifically, if the names are
+# reserved C type names like 'double', 'float', 'int', etc.
+#
+# We don't expect to that such structures should be seen in real C
+# code, but in some cases GDB will generate artificial structures, and
+# in some cases, these type names are the obvious choice for field
+# names.
+#
+# One specific example is RISC-V, the 64-bit floating point registers
+# are represented as a structure with the field names 'float' and
+# 'double'.
+
+load_lib dwarf.exp
+
+# This test can only be run on targets which support DWARF-2 and use
+# gas.
+if {![dwarf2_support]} {
+    return 0
+}
+
+standard_testfile dw2-unusual-field-names.c dw2-unusual-field-names.S
+set asm_file [standard_output_file $srcfile2]
+
+# We need to know the size of integer and address types in order to
+# write some of the debugging info we'd like to generate.
+#
+# For that, we ask GDB by debugging our test program.  Any program
+# would do, but since we already have one specifically for this
+# testcase, might as well use that.
+if [prepare_for_testing "failed to prepare" ${testfile} ${srcfile}] {
+    return -1
+}
+set int_size [get_sizeof "int" -1]
+
+# Rebuild the test binary with the single field within the structure
+# renamed to FIELD_NAME, then test that we can access the field
+# through '.' and through '->'.
+proc run_test { field_name } {
+    global asm_file testfile srcfile subdir srcdir
+    global int_size
+
+    Dwarf::assemble $asm_file {
+	global srcdir subdir srcfile
+	global field_name int_size
+
+	cu {} {
+	    DW_TAG_compile_unit {
+                {DW_AT_language @DW_LANG_C}
+                {DW_AT_name     dw2-unusual-field-names.c}
+                {DW_AT_comp_dir /tmp}
+	    } {
+		declare_labels itype ptype stype
+
+		itype: DW_TAG_base_type {
+		    {DW_AT_byte_size $int_size DW_FORM_sdata}
+		    {DW_AT_encoding  @DW_ATE_signed}
+		    {DW_AT_name int}
+		}
+
+		stype: DW_TAG_structure_type {
+		    {DW_AT_name "foo"}
+		    {DW_AT_byte_size $int_size DW_FORM_sdata}
+		} {
+		    member {
+			{name $field_name}
+			{type :$itype}
+			{data_member_location 0 data1}
+		    }
+		}
+
+		ptype: DW_TAG_pointer_type {
+		    {DW_AT_type :$stype}
+		}
+
+		DW_TAG_variable {
+		    {DW_AT_name obj}
+		    {DW_AT_type :$stype}
+		    {DW_AT_location {
+			DW_OP_addr [gdb_target_symbol obj]
+		    } SPECIAL_expr}
+		    {external 1 flag}
+		}
+
+		DW_TAG_variable {
+		    {DW_AT_name ptr}
+		    {DW_AT_type :$ptype}
+		    {DW_AT_location {
+			DW_OP_addr [gdb_target_symbol ptr]
+		    } SPECIAL_expr}
+		    {external 1 flag}
+		}
+	    }
+	}
+    }
+
+    if { [prepare_for_testing "failed to prepare" ${testfile} \
+	      [list $srcfile $asm_file] {nodebug}] } {
+	return -1
+    }
+
+    if ![runto_main] {
+	return -1
+    }
+
+    gdb_test "p obj.$field_name" " = 0" \
+	"access a field named '$field_name' directly"
+
+    gdb_test "p ptr->$field_name" " = 0" \
+	"access a field named '$field_name' through a pointer"
+
+    gdb_exit
+}
+
+foreach field_name { double float char byte long int short unsigned signed } {
+    run_test $field_name
+}