public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers
@ 2004-02-27  1:04 rearnsha at gcc dot gnu dot org
  2004-02-27  1:05 ` [Bug optimization/14312] " rearnsha at gcc dot gnu dot org
                   ` (8 more replies)
  0 siblings, 9 replies; 10+ messages in thread
From: rearnsha at gcc dot gnu dot org @ 2004-02-27  1:04 UTC (permalink / raw)
  To: gcc-bugs

The attached function no-longer generates tail calls in tree-ssa.  This is a
regression from 3.5.0 (the problem seems to be some intermediate variables that
are created by earlier passes and which are marked non-static, addressable).

Compile the attached code with

  arm-elf-gcc -O2 -S test.c


void bar(int i);
void baz(int *);
void foo(int *x) { if (*x < 0) { baz (x); return; } bar(*x);
}

observe that the resulting code is 

foo:
        @ args = 0, pretend = 0, frame = 0
        @ frame_needed = 1, uses_anonymous_args = 0
        mov     ip, sp
        stmfd   sp!, {fp, ip, lr, pc}
        ldr     r3, [r0, #0]
        cmp     r3, #0
        sub     fp, ip, #4
        blt     .L5
        mov     r0, r3
        bl      bar
        ldmfd   sp, {fp, sp, pc}
.L5:
        bl      baz
        ldmfd   sp, {fp, sp, pc}

when it should be
        ldr     r3, [r0, #0]
        cmp     r3, #0
        @ lr needed for prologue
        blt     .L4
        mov     r0, r3
        b       bar
.L4:
        b       baz

-- 
           Summary: [tree-ssa regression] tailcalls not being generated when
                    functions contain pointers
           Product: gcc
           Version: tree-ssa
            Status: UNCONFIRMED
          Keywords: pessimizes-code
          Severity: normal
          Priority: P2
         Component: optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: rearnsha at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org
GCC target triplet: arm-elf


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
@ 2004-02-27  1:05 ` rearnsha at gcc dot gnu dot org
  2004-02-27  2:01 ` [Bug optimization/14312] [tree-ssa] [regression] " pinskia at gcc dot gnu dot org
                   ` (7 subsequent siblings)
  8 siblings, 0 replies; 10+ messages in thread
From: rearnsha at gcc dot gnu dot org @ 2004-02-27  1:05 UTC (permalink / raw)
  To: gcc-bugs



-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
      Known to fail|                            |tree-ssa
      Known to work|                            |3.4.0 3.5.0
   Last reconfirmed|0000-00-00 00:00:00         |2004-02-27 01:05:54
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
  2004-02-27  1:05 ` [Bug optimization/14312] " rearnsha at gcc dot gnu dot org
@ 2004-02-27  2:01 ` pinskia at gcc dot gnu dot org
  2004-02-27 12:38 ` rth at gcc dot gnu dot org
                   ` (6 subsequent siblings)
  8 siblings, 0 replies; 10+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-02-27  2:01 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From pinskia at gcc dot gnu dot org  2004-02-27 02:01 -------
Confirmed.  Next time please do not confirm your own bug.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
 GCC target triplet|arm-elf                     |
      Known to work|3.4.0 3.5.0                 |3.4.0
            Summary|[tree-ssa regression]       |[tree-ssa] [regression]
                   |tailcalls not being         |tailcalls not being
                   |generated when functions    |generated when functions
                   |contain pointers            |contain pointers
   Target Milestone|---                         |tree-ssa


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
  2004-02-27  1:05 ` [Bug optimization/14312] " rearnsha at gcc dot gnu dot org
  2004-02-27  2:01 ` [Bug optimization/14312] [tree-ssa] [regression] " pinskia at gcc dot gnu dot org
@ 2004-02-27 12:38 ` rth at gcc dot gnu dot org
  2004-02-27 13:16 ` dnovillo at redhat dot com
                   ` (5 subsequent siblings)
  8 siblings, 0 replies; 10+ messages in thread
From: rth at gcc dot gnu dot org @ 2004-02-27 12:38 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From rth at gcc dot gnu dot org  2004-02-27 12:37 -------
Diego, the problem here is that we reject FOO for tail-call optimization because
we find "TMT.1" is a "local variable" with TREE_ADDRESSABLE set.  Clearly we need
to be ignoring memory tags.

Will you please invent a way to recoginize them and ignore them in
suitable_for_tail_opt_p?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |dnovillo at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
                   ` (2 preceding siblings ...)
  2004-02-27 12:38 ` rth at gcc dot gnu dot org
@ 2004-02-27 13:16 ` dnovillo at redhat dot com
  2004-02-27 18:39 ` cvs-commit at gcc dot gnu dot org
                   ` (4 subsequent siblings)
  8 siblings, 0 replies; 10+ messages in thread
From: dnovillo at redhat dot com @ 2004-02-27 13:16 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From dnovillo at redhat dot com  2004-02-27 13:16 -------
Subject: Re:  [tree-ssa] [regression] tailcalls not
	being generated when functions contain pointers

On Fri, 2004-02-27 at 07:38, rth at gcc dot gnu dot org wrote:

> Will you please invent a way to recoginize them and ignore them in
> suitable_for_tail_opt_p?
>
There's nothing to invent, really.  suitable_for_tail_opt_p was doing a
very weak test.  It should really test whether the local variable is
call-clobbered, which is computed using escape analysis.

We ought to check other places in the optimizers where we use
addressability instead of is_call_clobbered.

Fixed with this patch.  Will commit if testing doesn't show any
regressions.


Diego.

        PR optimization/14312
        * tree-tailcall.c (suitable_for_tail_opt_p): Call
        is_call_clobbered instead of checking addressability.

Index: tree-tailcall.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/Attic/tree-tailcall.c,v
retrieving revision 1.1.2.19
diff -d -c -p -r1.1.2.19 tree-tailcall.c
*** tree-tailcall.c     25 Feb 2004 03:22:47 -0000      1.1.2.19
--- tree-tailcall.c     27 Feb 2004 13:14:14 -0000
*************** suitable_for_tail_opt_p (void)
*** 72,87 ****
    if (current_function_stdarg)
      return false;

!   /* No local variable should have its address taken, as otherwise it might
!      be passed to the recursive call.  This of course is overly
!      conservative and should be replaced by a dataflow analysis later.  */
    for (i = 0; i < (int) VARRAY_ACTIVE_SIZE (referenced_vars); i++)
      {
        tree var = VARRAY_TREE (referenced_vars, i);

        if (decl_function_context (var) == current_function_decl
!         && !TREE_STATIC (var)
!         && TREE_ADDRESSABLE (var))
        return false;
      }

--- 72,86 ----
    if (current_function_stdarg)
      return false;

!   /* No local variable should be call-clobbered.  We ignore any kind
!      of memory tag, as these are not real variables.  */
    for (i = 0; i < (int) VARRAY_ACTIVE_SIZE (referenced_vars); i++)
      {
        tree var = VARRAY_TREE (referenced_vars, i);

        if (decl_function_context (var) == current_function_decl
!         && var_ann (var)->mem_tag_kind == NOT_A_TAG
!         && is_call_clobbered (var))
        return false;
      }



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
                   ` (3 preceding siblings ...)
  2004-02-27 13:16 ` dnovillo at redhat dot com
@ 2004-02-27 18:39 ` cvs-commit at gcc dot gnu dot org
  2004-02-27 18:45 ` dnovillo at gcc dot gnu dot org
                   ` (3 subsequent siblings)
  8 siblings, 0 replies; 10+ messages in thread
From: cvs-commit at gcc dot gnu dot org @ 2004-02-27 18:39 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-02-27 18:39 -------
Subject: Bug 14312

CVSROOT:	/cvs/gcc
Module name:	gcc
Branch: 	tree-ssa-20020619-branch
Changes by:	dnovillo@gcc.gnu.org	2004-02-27 18:39:01

Modified files:
	gcc            : ChangeLog.tree-ssa tree-pretty-print.c 
	                 tree-ssa-alias.c tree-tailcall.c 
	gcc/testsuite  : ChangeLog.tree-ssa 
Added files:
	gcc/testsuite/gcc.dg/tree-ssa: tailcall-2.c 

Log message:
	PR optimization/14312
	* tree-pretty-print.c (dump_generic_node): Mark tail calls.
	* tree-ssa-alias.c (compute_may_aliases): Restore call to
	dump_referenced_vars.
	(compute_points_to_and_addr_escape): If the address of a
	variable V is stored into a non-pointer variable, mark V as
	call-clobbered.
	* tree-tailcall.c (suitable_for_tail_opt_p): Check for
	call-clobbered instead of TREE_ADDRESSABLE.
	Ignore memory tags.
	(optimize_tail_call): Add newline to dump output.
	
	testsuite/ChangeLog.tree-ssa
	
	* gcc.dg/tree-ssa/tailcall-2.c: New test.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ChangeLog.tree-ssa.diff?cvsroot=gcc&only_with_tag=tree-ssa-20020619-branch&r1=1.1.2.1231&r2=1.1.2.1232
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-pretty-print.c.diff?cvsroot=gcc&only_with_tag=tree-ssa-20020619-branch&r1=1.1.2.76&r2=1.1.2.77
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-ssa-alias.c.diff?cvsroot=gcc&only_with_tag=tree-ssa-20020619-branch&r1=1.1.2.6&r2=1.1.2.7
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/tree-tailcall.c.diff?cvsroot=gcc&only_with_tag=tree-ssa-20020619-branch&r1=1.1.2.19&r2=1.1.2.20
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.tree-ssa.diff?cvsroot=gcc&only_with_tag=tree-ssa-20020619-branch&r1=1.1.2.177&r2=1.1.2.178
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/gcc.dg/tree-ssa/tailcall-2.c.diff?cvsroot=gcc&only_with_tag=tree-ssa-20020619-branch&r1=NONE&r2=1.1.2.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
                   ` (4 preceding siblings ...)
  2004-02-27 18:39 ` cvs-commit at gcc dot gnu dot org
@ 2004-02-27 18:45 ` dnovillo at gcc dot gnu dot org
  2004-03-04  3:00 ` dnovillo at gcc dot gnu dot org
                   ` (2 subsequent siblings)
  8 siblings, 0 replies; 10+ messages in thread
From: dnovillo at gcc dot gnu dot org @ 2004-02-27 18:45 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From dnovillo at gcc dot gnu dot org  2004-02-27 18:45 -------

Fixed. http://gcc.gnu.org/ml/gcc-patches/2004-02/msg02622.html

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
                   ` (5 preceding siblings ...)
  2004-02-27 18:45 ` dnovillo at gcc dot gnu dot org
@ 2004-03-04  3:00 ` dnovillo at gcc dot gnu dot org
  2004-03-07  2:03 ` pinskia at gcc dot gnu dot org
  2004-03-09 15:05 ` dnovillo at gcc dot gnu dot org
  8 siblings, 0 replies; 10+ messages in thread
From: dnovillo at gcc dot gnu dot org @ 2004-03-04  3:00 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From dnovillo at gcc dot gnu dot org  2004-03-04 03:00 -------

The fix seems to have broken bootstraps on PPC.  Re-opening.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
                   ` (6 preceding siblings ...)
  2004-03-04  3:00 ` dnovillo at gcc dot gnu dot org
@ 2004-03-07  2:03 ` pinskia at gcc dot gnu dot org
  2004-03-09 15:05 ` dnovillo at gcc dot gnu dot org
  8 siblings, 0 replies; 10+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-03-07  2:03 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-07 02:03 -------
Patch that most likely will fix the bootstrap problem is here: <http://gcc.gnu.org/ml/gcc/
2004-03/msg00370.html>.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Bug optimization/14312] [tree-ssa] [regression] tailcalls not being generated when functions contain pointers
  2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
                   ` (7 preceding siblings ...)
  2004-03-07  2:03 ` pinskia at gcc dot gnu dot org
@ 2004-03-09 15:05 ` dnovillo at gcc dot gnu dot org
  8 siblings, 0 replies; 10+ messages in thread
From: dnovillo at gcc dot gnu dot org @ 2004-03-09 15:05 UTC (permalink / raw)
  To: gcc-bugs


------- Additional Comments From dnovillo at gcc dot gnu dot org  2004-03-09 15:05 -------

Fixed with 

        * calls.c (initialize_argument_information): Add
        parameter may_tail_call.  Set to false for invisible
        pass-by-reference arguments that require stack
        allocation.
        Update all users.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14312


^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2004-03-09 15:05 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-02-27  1:04 [Bug optimization/14312] New: [tree-ssa regression] tailcalls not being generated when functions contain pointers rearnsha at gcc dot gnu dot org
2004-02-27  1:05 ` [Bug optimization/14312] " rearnsha at gcc dot gnu dot org
2004-02-27  2:01 ` [Bug optimization/14312] [tree-ssa] [regression] " pinskia at gcc dot gnu dot org
2004-02-27 12:38 ` rth at gcc dot gnu dot org
2004-02-27 13:16 ` dnovillo at redhat dot com
2004-02-27 18:39 ` cvs-commit at gcc dot gnu dot org
2004-02-27 18:45 ` dnovillo at gcc dot gnu dot org
2004-03-04  3:00 ` dnovillo at gcc dot gnu dot org
2004-03-07  2:03 ` pinskia at gcc dot gnu dot org
2004-03-09 15:05 ` dnovillo at gcc dot gnu dot org

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).