[Bug ada/17986] New: Compile error for make.adb breaks bootstrap

[Bug ada/17986] New: Compile error for make.adb breaks bootstrap

[aaronavay62 at aaronwl dot com]

During bootstrap with Ada and libada enabled:

../../xgcc -B../../ -c -O2 -g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes
-Wmissing-prototypes -fno-common      -gnatpg -gnata -I- -I../rts -I.
-I/aaronwl/cs/compilers/gcc/src/cvs/head/gcc/gcc/ada
/aaronwl/cs/compilers/gcc/src/cvs/head/gcc/gcc/ada/make.adb -o make.o

raised STORAGE_ERROR : EXCEPTION_STACK_OVERFLOW
make[4]: *** [make.o] Error 1

This error is generated after running for about 50 seconds.

It can be worked around by changing -O2 to -O1.

Configured with: ../../src/cvs/head/gcc/configure
--enable-languages=ada,c,c++,f95,java,objc --with-dwarf2 --enable-libmudflap
--enable-libgcj --with-win32-nlsapi=unicows --enable-libstdcxx-allocator=mt
--enable-libstdcxx-debug --prefix=/aaronwl/cs/env/mingw-head-head-head-20041012
--with-included-gettext
--with-libiconv-prefix=/aaronwl/cs/internat/iconv/install
--with-gmp=/aaronwl/cs/env/mingw-head-head-head-20041012
--with-mpfr=/aaronwl/cs/env/mingw-head-head-head-20041012
Thread model: win32
gcc version 4.0.0 20041013 (experimental)

i686-pc-mingw32
Windows XP SP2 Professional
Pentium 4 256MB
binutils mainline 2.15.92 20040920
mingwrt mainline 20041012
msvcrt 7.0.2600.2180
w32api mainline 20041012
libiconv 1.9.1
Environment and misc. tools from Cygwin Current Thu Oct  7 19:20:20 2004

-- 
           Summary: Compile error for make.adb breaks bootstrap
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: ada
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: aaronavay62 at aaronwl dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i686-pc-mingw32
  GCC host triplet: i686-pc-mingw32
GCC target triplet: i686-pc-mingw32


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|ada                         |tree-optimization
           Keywords|                            |build, ice-on-valid-code
            Summary|Compile error for make.adb  |[4.0 Regression] Compile
                   |breaks bootstrap            |error for make.adb breaks
                   |                            |bootstrap
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2004-10-14 19:22 -------
Present on sparc-sun-solaris2.8, with same workaround.  Preliminary
investigation suggests a PHI nodes sharing problem.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ebotcazou at gcc dot gnu dot
                   |                            |org
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-10-14 19:22:29
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[ebotcazou at gcc dot gnu dot org]

------- Additional Comments From ebotcazou at gcc dot gnu dot org  2004-10-19 15:03 -------
The file still takes a few minutes to compile but the situation has improved
since all the gnattools build for me as of today, both on sparc-sun-solaris2.8
and i586-mandrake-linux-gnu, whereas they previously didn't.  Could you try
again with an updated tree?


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 19:25 -------
Actually this is a tree optimization problem, SSA_NAMEs cause too many function levels in the GC code 
from 
PR18226:
Program received signal SIGSEGV, Segmentation fault.
gt_ggc_mx_lang_tree_node (x_p=0x3f9edc0) at ./gtype-ada.h:27
27        if (ggc_test_and_set_mark (x))
#0  gt_ggc_mx_lang_tree_node (x_p=0x3f9edc0) at ./gtype-ada.h:27
        x = (union lang_tree_node * const) 0x3f9edc0
        __FUNCTION__ = "gt_ggc_mx_lang_tree_node"
#1  0x0042b215 in gt_ggc_mx_lang_tree_node (x_p=0x41510a0) at ./gtype-ada.h:178
        x = (union lang_tree_node * const) 0x41510a0
        __FUNCTION__ = "gt_ggc_mx_lang_tree_node"
#2  0x0042b215 in gt_ggc_mx_lang_tree_node (x_p=0x3c84af0) at ./gtype-ada.h:178
        x = (union lang_tree_node * const) 0x3c84af0
        __FUNCTION__ = "gt_ggc_mx_lang_tree_node"
...

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW
   Last reconfirmed|2004-10-14 19:22:29         |2004-10-29 19:25:49
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 19:26 -------
*** Bug 18226 has been marked as a duplicate of this bug. ***

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 19:32 -------
Diego is there any way to reduce the number of SSA_NAMEs are TREE_CHAINed against each other:
              gt_ggc_m_9tree_node ((*x).t.ssa_name.common.chain);

Did we forgot to clear TREE_CHAIN of the SSA_NAME when grabbing from free_ssanames?

I will test a patch which does this because it looks like it is causing this problem.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dnovillo at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 19:34 -------
This patch should help but it might fix all the problems:
Index: tree-ssanames.c
===============================================================
====
RCS file: /cvs/gcc/gcc/gcc/tree-ssanames.c,v
retrieving revision 2.16
diff -u -p -r2.16 tree-ssanames.c
--- tree-ssanames.c	29 Oct 2004 08:41:09 -0000	2.16
+++ tree-ssanames.c	29 Oct 2004 19:32:49 -0000
@@ -215,6 +215,7 @@ make_ssa_name (tree var, tree stmt)
   SSA_NAME_DEF_STMT (t) = stmt;
   SSA_NAME_PTR_INFO (t) = NULL;
   SSA_NAME_IN_FREE_LIST (t) = 0;
+  TREE_CHAIN (t) = NULL_TREE;
 
   return t;
 }


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 19:36 -------
Forgot to say I am testing this patch right now (it might also solve a fortran problem with GC failure 
which shows up sometimes).

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 19:46 -------
Note the patch will not work as SSA_NAME_DEF_STMT is the same as TREE_CHAIN.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 20:34 -------
Ok, this patch to the Ada front-end should help:
Index: ada/ada-tree.h
===============================================================
====
RCS file: /cvs/gcc/gcc/gcc/ada/ada-tree.h,v
retrieving revision 1.26
diff -u -p -r1.26 ada-tree.h
--- ada/ada-tree.h	21 Sep 2004 15:31:00 -0000	1.26
+++ ada/ada-tree.h	29 Oct 2004 20:31:01 -0000
@@ -34,7 +34,11 @@ enum gnat_tree_code {
 #undef DEFTREECODE
 
 /* Ada uses the lang_decl and lang_type fields to hold a tree.  */
-union lang_tree_node GTY((desc ("0"))) {union tree_node GTY((tag ("0"))) t; };
+union lang_tree_node GTY((desc ("0")),
+                          chain_next ("(union lang_tree_node *)TREE_CHAIN (&%t.t)")))
+{
+  union tree_node GTY((tag ("0"))) t;
+};
 struct lang_decl GTY(()) {tree t; };
 struct lang_type GTY(()) {tree t; };
 
It should help.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 20:46 -------
The patch which I pasted is wrong, instead of:
TREE_CHAIN (&%t.t)
it should be:
TREE_CHAIN (&%h.t)

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-29 22:06 -------
Created an attachment (id=7433)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7433&action=view)
patch which helps a lot

The patch which works and improves the GC behavior for Ada.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug tree-optimization/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[steven at gcc dot gnu dot org]

------- Additional Comments From steven at gcc dot gnu dot org  2004-10-29 22:23 -------
This should blow away GC times from the Ada compile times. 
 
I'd be interested to see -ftime-report results from before 
and after the patch, if someone can produce them... 
 

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug ada/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-10-30 00:18 -------
I posted the patch here: <http://gcc.gnu.org/ml/gcc-patches/2004-10/msg02639.html>.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|tree-optimization           |ada
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug ada/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-11-04 19:12 -------
*** Bug 18301 has been marked as a duplicate of this bug. ***

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ro at techfak dot uni-
                   |                            |bielefeld dot de


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug ada/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[ro at gcc dot gnu dot org]

------- Additional Comments From ro at gcc dot gnu dot org  2004-11-19 23:17 -------
This problems bites me on several platforms (sparc-sun-solaris10,
i386-sun-solaris2.10, powerpc-apple-darwin7.6.0).  Perhaps you could ping your
patch so this gets fixed?

  Rainer


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug ada/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-11-22 13:50 -------
Subject: Bug 17986

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	pinskia@gcc.gnu.org	2004-11-22 13:50:41

Modified files:
	gcc/ada        : ChangeLog ada-tree.h 

Log message:
	2004-11-22  Andrew Pinski  <pinskia@physics.uc.edu>
	
	PR ada/17986
	* ada-tree.h (lang_tree_node): Set chain_next to be the chain of the
	tree.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/ChangeLog.diff?cvsroot=gcc&r1=1.608&r2=1.609
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/ada-tree.h.diff?cvsroot=gcc&r1=1.26&r2=1.27



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986

[Bug ada/17986] [4.0 Regression] Compile error for make.adb breaks bootstrap

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-11-22 13:51 -------
Fixed.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17986