[Bug target/14813] New: The wrong init order

[Bug target/14813] New: The wrong init order

[hjl at lucon dot org]

The change in crtend.asm causes the wrong init oder.

-- 
           Summary: The wrong init order
           Product: gcc
           Version: 3.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hjl at lucon dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: ia64-unknown-linux-gnu
  GCC host triplet: ia64-unknown-linux-gnu
GCC target triplet: ia64-unknown-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] The wrong init order

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2004-04-01 20:00 -------
Created an attachment (id=6025)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=6025&action=view)
A testcase

# make
/usr/gcc-3.4/bin/gcc -O -g   -c -o bar.o bar.c
/usr/gcc-3.4/bin/gcc -static  bar.o   -o bar
./bar
preinit array 0
preinit array 1
preinit array 2
init array 0
init array 1
init array 2
init
fini array 2
fini array 1
fini array 0
fini

The right one should be

[hjl@gnu-4 init]$ make CC=gcc
gcc -O -g   -c -o bar.o bar.c
/tmp/ccyDlcv7.s: Assembler messages:
/tmp/ccyDlcv7.s:184: Warning: ignoring incorrect section type for
.preinit_array/tmp/ccyDlcv7.s:292: Warning: ignoring incorrect section type for
.init_array
/tmp/ccyDlcv7.s:400: Warning: ignoring incorrect section type for .fini_array
gcc -static  bar.o   -o bar
./bar
preinit array 0
preinit array 1
preinit array 2
init
init array 0
init array 1
init array 2
fini array 2
fini array 1
fini array 0
fini


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4.0 Regression] The wrong init order

[hjl at lucon dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|The wrong init order        |[3.4.0 Regression] The wrong
                   |                            |init order


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-04-01 20:05 -------
The last change to crtend.asm was over a year ago why was this not found way before the release?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
            Summary|[3.4.0 Regression] The wrong|[3.4/3.5 Regression] The
                   |init order                  |wrong init order
   Target Milestone|---                         |3.4.1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2004-04-01 21:05 -------
Created an attachment (id=6026)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=6026&action=view)
A patch

This patch should work for both 3.4 and 3.5.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[jakub at gcc dot gnu dot org]

------- Additional Comments From jakub at gcc dot gnu dot org  2004-04-06 09:42 -------
H.J., can you please post your patches to gcc-patches and put URL into
gcc-patches archive into the PRs instead of just attaching them to PRs and
not mentioning them on gcc-patches at all?



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2004-04-06 18:09 -------
The bug was introduced by

http://gcc.gnu.org/ml/gcc-cvs/2003-05/msg00274.html

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2004-04-12 23:06 -------
Created an attachment (id=6080)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=6080&action=view)
An updated patch

__do_jv_register_classes should be place before
__do_global_ctors_aux to preserve the init order
the same as gcc 3.3.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #6026 is|0                           |1
           obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[hjl at lucon dot org]

------- Additional Comments From hjl at lucon dot org  2004-04-13 16:48 -------
A testcase is posted at

http://gcc.gnu.org/ml/gcc-patches/2004-04/msg00788.html

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[wilson at specifixinc dot com]

------- Additional Comments From wilson at specifixinc dot com  2004-04-21 22:37 -------
Subject: Re:  [3.4/3.5 Regression] The wrong init order

pinskia at gcc dot gnu dot org wrote:
> ------- Additional Comments From pinskia at gcc dot gnu dot org  2004-04-01 20:05 -------
> The last change to crtend.asm was over a year ago why was this not found way before the release?

This only fails if an application or library uses init_array/fini_array.

GCC still uses ctors/dtors sections instead of the preferred and 
standardized init_array/fini_array sections.  So the only way that you 
can trigger this problem is if someone explicitly uses attribute section 
init_array, or if you link with code compiled by another compiler.

I suspect the latter method is how HJ found the problem, e.g. he may 
have noticed the problem after trying to link icc compiled code with 
gcc compiled code.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-04-21 23:19 -------
Subject: Bug 14813

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	gcc-3_4-branch
Changes by:	hjl@gcc.gnu.org	2004-04-21 23:19:09

Modified files:
	gcc            : ChangeLog 
	gcc/config/ia64: crtbegin.asm crtend.asm 

Log message:
	2004-04-21  H.J. Lu  <hongjiu.lu@intel.com>
	
	PR target/14813
	Backport from mainline:
	
	* config/ia64/crtend.asm: Move pointer to __do_global_ctors_aux
	in .init_array section to ...
	* config/ia64/crtbegin.asm: Here.
	
	* config/ia64/crtend.asm: Mark __do_global_ctors_aux global
	and hidden if HAVE_INITFINI_ARRAY is defined.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=2.2326.2.405&r2=2.2326.2.406
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/ia64/crtbegin.asm.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.16&r2=1.16.16.1
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/config/ia64/crtend.asm.diff?cvsroot=gcc&only_with_tag=gcc-3_4-branch&r1=1.12&r2=1.12.16.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813

[Bug target/14813] [3.4/3.5 Regression] The wrong init order

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-04-21 23:22 -------
Fixed for 3.4.1.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14813