Cfgcleanup fix

Cfgcleanup fix

[Zdenek Dvorak]

Hello.

The following patch fixes updating of liveness information in 
try_crossjump_to_edge; this fixes Ada bootstrap failure on
sparc-sun-solaris2.8.

Bootstrapped and regtested on i686. Also bootstrapped on
sparc-sun-solaris2.8 (c and ada only).

Zdenek Dvorak

Changelog:
	* cfgcleanup.c (try_crossjump_to_edge): Fix updating of liveness
	information.

Index: cfgcleanup.c
===================================================================
RCS file: /cvsroot/gcc/gcc/gcc/cfgcleanup.c,v
retrieving revision 1.62
diff -c -3 -p -r1.62 cfgcleanup.c
*** cfgcleanup.c	10 Jun 2002 21:36:12 -0000	1.62
--- cfgcleanup.c	22 Jun 2002 03:23:31 -0000
*************** try_crossjump_to_edge (mode, e1, e2)
*** 1310,1320 ****
  {
    int nmatch;
    basic_block src1 = e1->src, src2 = e2->src;
!   basic_block redirect_to;
    rtx newpos1, newpos2;
    edge s;
-   rtx last;
-   rtx label;
  
    /* Search backward through forwarder blocks.  We don't need to worry
       about multiple entry or chained forwarders, as they will be optimized
--- 1310,1318 ----
  {
    int nmatch;
    basic_block src1 = e1->src, src2 = e2->src;
!   basic_block redirect_to, redirect_from, to_remove;
    rtx newpos1, newpos2;
    edge s;
  
    /* Search backward through forwarder blocks.  We don't need to worry
       about multiple entry or chained forwarders, as they will be optimized
*************** try_crossjump_to_edge (mode, e1, e2)
*** 1442,1469 ****
  
    if (GET_CODE (newpos1) == NOTE)
      newpos1 = NEXT_INSN (newpos1);
-   last = src1->end;
  
!   /* Emit the jump insn.  */
!   label = block_label (redirect_to);
!   emit_jump_insn_after (gen_jump (label), src1->end);
!   JUMP_LABEL (src1->end) = label;
!   LABEL_NUSES (label)++;
! 
!   /* Delete the now unreachable instructions.  */
!   delete_insn_chain (newpos1, last);
! 
!   /* Make sure there is a barrier after the new jump.  */
!   last = next_nonnote_insn (src1->end);
!   if (!last || GET_CODE (last) != BARRIER)
!     emit_barrier_after (src1->end);
! 
!   /* Update CFG.  */
!   while (src1->succ)
!     remove_edge (src1->succ);
!   make_single_succ_edge (src1, redirect_to, 0);
  
!   update_forwarder_flag (src1);
  
    return true;
  }
--- 1440,1453 ----
  
    if (GET_CODE (newpos1) == NOTE)
      newpos1 = NEXT_INSN (newpos1);
  
!   redirect_from = split_block (src1, PREV_INSN (newpos1))->src;
!   to_remove = redirect_from->succ->dest;
  
!   redirect_edge_and_branch_force (redirect_from->succ, redirect_to);
!   flow_delete_block (to_remove);
! 
!   update_forwarder_flag (redirect_from);
  
    return true;
  }

Re: Cfgcleanup fix

[Jan Hubicka]

> Hello.
> 
> The following patch fixes updating of liveness information in 
> try_crossjump_to_edge; this fixes Ada bootstrap failure on
> sparc-sun-solaris2.8.
> 
> Bootstrapped and regtested on i686. Also bootstrapped on
> sparc-sun-solaris2.8 (c and ada only).
> 
> Zdenek Dvorak
> 
> Changelog:
> 	* cfgcleanup.c (try_crossjump_to_edge): Fix updating of liveness
> 	information.
> 
..
> *************** try_crossjump_to_edge (mode, e1, e2)
> *** 1442,1469 ****
>   
>     if (GET_CODE (newpos1) == NOTE)
>       newpos1 = NEXT_INSN (newpos1);
> -   last = src1->end;
>   
> !   /* Emit the jump insn.  */
> !   label = block_label (redirect_to);
> !   emit_jump_insn_after (gen_jump (label), src1->end);
> !   JUMP_LABEL (src1->end) = label;
> !   LABEL_NUSES (label)++;
> ! 
> !   /* Delete the now unreachable instructions.  */
> !   delete_insn_chain (newpos1, last);
> ! 
> !   /* Make sure there is a barrier after the new jump.  */
> !   last = next_nonnote_insn (src1->end);
> !   if (!last || GET_CODE (last) != BARRIER)
> !     emit_barrier_after (src1->end);
> ! 
> !   /* Update CFG.  */
> !   while (src1->succ)
> !     remove_edge (src1->succ);
> !   make_single_succ_edge (src1, redirect_to, 0);
>   
> !   update_forwarder_flag (src1);
>   
>     return true;
>   }
> --- 1440,1453 ----
>   
>     if (GET_CODE (newpos1) == NOTE)
>       newpos1 = NEXT_INSN (newpos1);
>   
> !   redirect_from = split_block (src1, PREV_INSN (newpos1))->src;
> !   to_remove = redirect_from->succ->dest;
>   
> !   redirect_edge_and_branch_force (redirect_from->succ, redirect_to);
> !   flow_delete_block (to_remove);
> ! 
> !   update_forwarder_flag (redirect_from);

This looks good. The main purpose for the old code was to avoid creating
and destroying of basic block at time it had linear time complexity.  I
have to check rest of code for similar hacks.

Honza
>   
>     return true;
>   }

Re: Cfgcleanup fix

[Rainer Orth]

Zdenek Dvorak writes:

> The following patch fixes updating of liveness information in 
> try_crossjump_to_edge; this fixes Ada bootstrap failure on
> sparc-sun-solaris2.8.
> 
> Bootstrapped and regtested on i686. Also bootstrapped on
> sparc-sun-solaris2.8 (c and ada only).

Indeed: I've successfully bootstrapped with this patch on
sparc-sun-solaris2.8 and i386-pc-solaris2.8 with Ada included.  There are a
few new regressions compared to 20020617, but they almost certainly
unrelated to this patch:

On i386-pc-solaris2.8:

+FAIL: g77.f-torture/execute/f90-intrinsic-bit.f compilation,  -Os 
+FAIL: gcc.c-torture/compile/20001226-1.c,  -O1  
+FAIL: gcc.c-torture/compile/20001226-1.c,  -O2  
+FAIL: gcc.c-torture/compile/20001226-1.c,  -O3 -fomit-frame-pointer  
+FAIL: gcc.c-torture/compile/20001226-1.c,  -Os  

On sparc-sun-solaris2.8:

+FAIL: g77.f-torture/execute/f90-intrinsic-bit.f compilation,  -Os 
+FAIL: g77.f-torture/execute/f90-intrinsic-bit.f compilation,  -Os 
+FAIL: gcc.c-torture/compile/20001226-1.c,  -O2  
+FAIL: gcc.c-torture/compile/20001226-1.c,  -O3 -fomit-frame-pointer  
+FAIL: gcc.c-torture/compile/20001226-1.c,  -Os  
+FAIL: gcc.c-torture/execute/20020108-1.c compilation,  -Os 
+FAIL: gcc.c-torture/execute/20020108-1.c compilation,  -Os 

These have been reported by various people already.

+FAIL: gcc.c-torture/execute/simd-1.c compilation,  -O0 
+FAIL: gcc.c-torture/execute/simd-1.c compilation,  -O1 
+FAIL: gcc.c-torture/execute/simd-1.c compilation,  -O2 
+FAIL: gcc.c-torture/execute/simd-1.c compilation,  -O3 -fomit-frame-pointer 
+FAIL: gcc.c-torture/execute/simd-1.c compilation,  -O3 -g 
+FAIL: gcc.c-torture/execute/simd-1.c compilation,  -Os 

/vol/gnu/src/gcc/gcc-dist/gcc/testsuite/gcc.c-torture/execute/simd-1.c: In function `main':
/vol/gnu/src/gcc/gcc-dist/gcc/testsuite/gcc.c-torture/execute/simd-1.c:33: internal compiler error: Internal compiler error in simplify_gen_subreg, at simplify-rtx.c:2551

+FAIL: gcc.dg/simd-1.c (test for excess errors)

/vol/gnu/src/gcc/gcc-dist/gcc/testsuite/gcc.dg/simd-1.c:44: internal compiler error: Internal compiler error in emit_move_insn_1, at expr.c:3150

These only happen with -m64.

Thanks for your help.

	Rainer

Unreviewed Ada bootstrap patch for 3.2 [Re: Cfgcleanup fix]

[Rainer Orth]

I just noticed that the this patch

	http://gcc.gnu.org/ml/gcc-patches/2002-06/msg01899.html

has remained unreviewed for more than three weeks.  It is necessary to
allow Ada to bootstrap on at least sparc-sun-solaris2.8 and
i386-pc-solaris2.8.

	Rainer

Re: Unreviewed Ada bootstrap patch for 3.2 [Re: Cfgcleanup fix]

[Geert Bosch]

On Friday, July 5, 2002, at 11:14 , Rainer Orth wrote:

> I just noticed that the this patch
>
> 	http://gcc.gnu.org/ml/gcc-patches/2002-06/msg01899.html
>
> has remained unreviewed for more than three weeks.  It is necessary to
> allow Ada to bootstrap on at least sparc-sun-solaris2.8 and
> i386-pc-solaris2.8.

Just to be clear, even though the patch fixes the Ada bootstrap,
this is not an ada/ patch that I can review. It would be great if
this patch could get reviewed.

   -Geert

Re: Unreviewed Ada bootstrap patch for 3.2 [Re: Cfgcleanup fix]

[Zdenek Dvorak]

Hello.

> >I just noticed that the this patch
> >
> >	http://gcc.gnu.org/ml/gcc-patches/2002-06/msg01899.html
> >
> >has remained unreviewed for more than three weeks.  It is necessary to
> >allow Ada to bootstrap on at least sparc-sun-solaris2.8 and
> >i386-pc-solaris2.8.
> 
> Just to be clear, even though the patch fixes the Ada bootstrap,
> this is not an ada/ patch that I can review. It would be great if
> this patch could get reviewed.

It would be really fine if someone could finally take a look on this.
I'm not sure whether it is still crucial for Ada bootstrap (the bug
could be easily masked by some other commit in the meantime), but anyway
it would be nice if it got fixed properly.

Zdenek

Re: Unreviewed Ada bootstrap patch for 3.2 [Re: Cfgcleanup fix]

[Andreas Schwab]

Zdenek Dvorak <rakdver@atrey.karlin.mff.cuni.cz> writes:

|> Hello.
|> 
|> > >I just noticed that the this patch
|> > >
|> > >	http://gcc.gnu.org/ml/gcc-patches/2002-06/msg01899.html
|> > >
|> > >has remained unreviewed for more than three weeks.  It is necessary to
|> > >allow Ada to bootstrap on at least sparc-sun-solaris2.8 and
|> > >i386-pc-solaris2.8.
|> > 
|> > Just to be clear, even though the patch fixes the Ada bootstrap,
|> > this is not an ada/ patch that I can review. It would be great if
|> > this patch could get reviewed.
|> 
|> It would be really fine if someone could finally take a look on this.
|> I'm not sure whether it is still crucial for Ada bootstrap (the bug
|> could be easily masked by some other commit in the meantime), but anyway
|> it would be nice if it got fixed properly.

It is definitely still needed.  FWIW, it also successfully fixes
bootstrapping Ada on ia64.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux AG, Deutschherrnstr. 15-19, D-90429 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."