[Bug target/24374] New: [4.0/4.1 Regression] Sibcalling optimization is happening in main

[Bug target/24374] New: [4.0/4.1 Regression] Sibcalling optimization is happening in main

[pinskia at gcc dot gnu dot org]

Take the following example:
int main1(void);

int main(void)
{
  return main1();
}
----
With 3.4.0, we got at -O2:
main:
        pushl   %ebp
        movl    %esp, %ebp
        subl    $8, %esp
        andl    $-16, %esp
        subl    $16, %esp
        call    main1
        leave
        ret

While with 4.0.0 and above we get:
main:
        pushl   %ebp
        movl    %esp, %ebp
        subl    $8, %esp
        andl    $-16, %esp
        subl    $16, %esp
        leave
        jmp     main1
        .size   main, .-main

This is wrong, we should not be sibcalling optimizing in main at all.


-- 
           Summary: [4.0/4.1 Regression] Sibcalling optimization is
                    happening in main
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P2
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: pinskia at gcc dot gnu dot org
GCC target triplet: i?86-*-*


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.0.3


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2005-10-15 17:40 -------
I should note that the reasons why you don't want to sibcalling optimize to
calls in main is so that the alignment fix up is not lost in the secondary
function.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[jkj at sco dot com]

------- Comment #2 from jkj at sco dot com  2005-10-15 20:02 -------
Created an attachment (id=9990)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9990&action=view)
This change to calls.c fixes the problem for me

I'm almost certain that there is a better way to do this, but this patch fixed
the problem for me.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[jkj at sco dot com]

------- Comment #3 from jkj at sco dot com  2005-10-15 20:14 -------
Created an attachment (id=9991)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9991&action=view)
Sorry I uploaded the wrong diff. this is against mainline.


-- 

jkj at sco dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #9990 is|0                           |1
           obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[jkj at sco dot com]

------- Comment #4 from jkj at sco dot com  2005-10-15 21:18 -------
Created an attachment (id=9993)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=9993&action=view)
New patch based on comments from Pinski


-- 

jkj at sco dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #9991 is|0                           |1
           obsolete|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[rth at gcc dot gnu dot org]

-- 

rth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |rth at gcc dot gnu dot org
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-10-16 04:24:58
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[mmitchel at gcc dot gnu dot org]

------- Comment #5 from mmitchel at gcc dot gnu dot org  2005-10-31 06:25 -------
Kean, I think you need to check TREE_PUBLIC (decl) as well; a "static" main
function isn't the main you're looking for.  Note that a "static" main is
permitted by gcc, with a warning, by default.

(It looks like the same bug appears in cfgexpand.c, by the way.  I'll open a
new PR for that.)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[jkj at sco dot com]

------- Comment #6 from jkj at sco dot com  2005-10-31 07:12 -------
rth has a completely different fix for this that is much more comprehensive.
http://gcc.gnu.org/ml/gcc/2005-10/msg00412.html has the details. I'm still
working on bringing my 4.1 tree up to speed so I can help him test this.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[rth at gcc dot gnu dot org]

------- Comment #7 from rth at gcc dot gnu dot org  2005-11-03 01:40 -------
Subject: Bug 24374

Author: rth
Date: Thu Nov  3 01:40:33 2005
New Revision: 106420

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=106420
Log:
        PR target/9350
        PR target/24374
        * dwarf2out.c (dwarf2out_reg_save_reg): New.
        (dwarf2out_frame_debug_expr): Return after dwarf_handle_frame_unspec.
        * function.c (assign_parms): Use calls.internal_arg_pointer.
        (expand_main_function): Remove FORCE_PREFERRED_STACK_BOUNDARY_IN_MAIN
        code.
        * target-def.h (TARGET_INTERNAL_ARG_POINTER): New.
        (TARGET_CALLS): Add it.
        * target.h (struct gcc_target): Add calls.internal_arg_pointer.
        * targhooks.c (default_internal_arg_pointer): New.
        * targhooks.h (default_internal_arg_pointer): Declare.
        * tree.h (dwarf2out_reg_save_reg): Declare.
        * doc/tm.texi (FORCE_PREFERRED_STACK_BOUNDARY_IN_MAIN): Remove.
        * config/i386/i386.c (dbx_register_map): Add return column.
        (dbx64_register_map, svr4_dbx_register_map): Likewise.
        (TARGET_INTERNAL_ARG_POINTER, ix86_internal_arg_pointer): New.
        (TARGET_DWARF_HANDLE_FRAME_UNSPEC, ix86_dwarf_handle_frame_unspec):
New.        (ix86_function_ok_for_sibcall): Disable if force_align_arg_pointer.
        (ix86_save_reg): Save force_align_arg_pointer.
        (ix86_emit_save_regs): Make regno unsigned.
        (ix86_emit_save_regs_using_mov): Likewise.
        (ix86_expand_prologue): Handle force_align_arg_pointer.
        (ix86_expand_epilogue): Likewise.
        * config/i386/i386.h: (dbx_register_map): Update.
        (dbx64_register_map, svr4_dbx_register_map): Update.
        (struct machine_function): Add force_align_arg_pointer.
        * config/i386/i386.md (UNSPEC_REG_SAVE, UNSPEC_DEF_CFA): New.
        (UNSPEC_TP, UNSPEC_TLS_GD, UNSPEC_TLS_LD_BASE): Renumber.
        (TARGET_PUSH_MEMORY peepholes): Disable if RTX_FRAME_RELATED_P.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/i386/i386.c
    trunk/gcc/config/i386/i386.h
    trunk/gcc/config/i386/i386.md
    trunk/gcc/doc/tm.texi
    trunk/gcc/dwarf2out.c
    trunk/gcc/function.c
    trunk/gcc/target-def.h
    trunk/gcc/target.h
    trunk/gcc/targhooks.c
    trunk/gcc/targhooks.h
    trunk/gcc/tree.h


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374

[Bug target/24374] [4.0/4.1 Regression] Sibcalling optimization is happening in main

[rth at gcc dot gnu dot org]

------- Comment #8 from rth at gcc dot gnu dot org  2005-11-03 01:42 -------
Fixed for mainline.  I have no particular need to see this backported to 4.0.
Any change there should be limited to ix86_function_ok_for_sibcall.


-- 

rth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|4.0.3                       |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24374