From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugs-return-319836-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 8343 invoked by alias); 27 May 2010 20:51:04 -0000
Received: (qmail 8279 invoked by uid 48); 27 May 2010 20:50:47 -0000
Date: Thu, 27 May 2010 20:51:00 -0000
Message-ID: <20100527205047.8278.qmail@sourceware.org>
X-Bugzilla-Reason: CC
References: <bug-44188-10053@http.gcc.gnu.org/bugzilla/>
Subject: [Bug c++/44188] Fails to produce DW_AT_typedef for typedef of anonymous struct
In-Reply-To: <bug-44188-10053@http.gcc.gnu.org/bugzilla/>
Reply-To: gcc-bugzilla@gcc.gnu.org
To: gcc-bugs@gcc.gnu.org
From: "jakub at gcc dot gnu dot org" <gcc-bugzilla@gcc.gnu.org>
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
X-SW-Source: 2010-05/txt/msg03068.txt.bz2



------- Comment #4 from jakub at gcc dot gnu dot org  2010-05-27 20:50 -------
Not 100% sure yet, but I believe this commit broke bootstrap on both
x86_64-linux and i686-linux, e.g. in libobjc or in ada bits.
One of the ICEs I'm seeing is:
#0  fancy_abort (file=0x8d1d1e3 "../../gcc/dwarf2out.c", line=16990,
function=0x8d21b15 "add_byte_size_attribute") at ../../gcc/diagnostic.c:782
#1  0x0823486f in add_byte_size_attribute (die=0xf7d76a20, tree_node=<value
optimized out>) at ../../gcc/dwarf2out.c:16990
#2  0x082568d3 in gen_struct_or_union_type_die (type=0xf7d52cc0,
context_die=0xf7d633f0, usage=<value optimized out>)
    at ../../gcc/dwarf2out.c:19329
#3  gen_tagged_type_die (type=0xf7d52cc0, context_die=0xf7d633f0, usage=<value
optimized out>) at ../../gcc/dwarf2out.c:19502
#4  0x08257e9a in gen_typedef_die (decl=0xf7d600d8, context_die=0xf7d633f0) at
../../gcc/dwarf2out.c:19430
#5  0x08255a59 in gen_decl_die (decl=0xf7d600d8, origin=<value optimized out>,
context_die=0xf7d633f0) at ../../gcc/dwarf2out.c:20186

gen_typedef_die is called on:

 <type_decl 0xf7d600d8 id
    type <pointer_type 0xf7d52cc0 id
        type <record_type 0xf7d52c00 objc_object asm_written type_0 SI
            size <integer_cst 0xf7ce8258 constant 32>
            unit size <integer_cst 0xf7ce8090 constant 4>
            align 32 symtab -136878016 alias set -1 canonical type 0xf7d52c00
fields <field_decl 0xf7d62a6c class_pointer>
            pointer_to_this <pointer_type 0xf7d52cc0 id> chain <type_decl
0xf7d60000 D.1196>>
        asm_written unsigned SI size <integer_cst 0xf7ce8258 32> unit size
<integer_cst 0xf7ce8090 4>
        align 32 symtab -136877536 alias set -1 canonical type 0xf7d52cc0>
    asm_written VOID file /usr/src/gcc/libobjc/objc/objc.h line 72 col 4
    align 1>

Given the comments, I wonder if:
--- dwarf2out.c.jj3     2010-05-27 21:48:57.000000000 +0200
+++ dwarf2out.c 2010-05-27 22:50:22.000000000 +0200
@@ -19427,7 +19427,9 @@ gen_typedef_die (tree decl, dw_die_ref c
             generate that DIE right away. add_type_attribute
             called below will then pick (via lookup_type_die) that
             anonymous struct DIE.  */
-         gen_tagged_type_die (type, context_die, DINFO_USAGE_DIR_USE);
+         if (RECORD_OR_UNION_TYPE_P (type)
+             || TREE_CODE (type) == ENUMERAL_TYPE)
+           gen_tagged_type_die (type, context_die, DINFO_USAGE_DIR_USE);
        }

       add_type_attribute (type_die, type, TREE_READONLY (decl),

isn't intended, calling gen_tagged_type_die on a POINTER_TYPE is weird.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44188