[Bug bootstrap/11318] New: [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[Bug bootstrap/11318] New: [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[ehrhardt at mathematik dot uni-ulm dot de]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 2367 bytes --]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11318

           Summary: [3.4 regresssion] Spurious bootstrap comparison failures
                    on sparc-sun-solaris
           Product: gcc
           Version: 3.4
            Status: UNCONFIRMED
          Keywords: build
          Severity: critical
          Priority: P2
         Component: bootstrap
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: ehrhardt at mathematik dot uni-ulm dot de
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: sparc-sun-solaris2.9
  GCC host triplet: sparc-sun-solaris2.9
GCC target triplet: sparc-sun-solaris2.9

I've been unable to do a clean bootstrap on sparc-sun-solaris with
3.4 for quite some time now. The problem is that I keep getting
spurious bootstrap comparison failures for some files. The files that
are affected seem to vary depending on the circumstances of the bootstrap.

However, the differences found are always due to an embedded temporary
file name in the object. The bootstrap comparison seems to expect this
difference but expects it at the beginning of the object file.
With my environment these differences appear somewhere in the middle
of the object file.

To illustrate this I ran fix-header.o and stage2/fix-header.o through
od -bc. The diff -u of these files looks like this:

--- /home/thales/ehrhardt/fix-header.od Wed Jun 25 13:14:16 2003
+++ /home/thales/ehrhardt/fix-header.od2        Wed Jun 25 13:14:14 2003
@@ -3716,8 +3716,8 @@
              \0 001   ²  \0  \0  \0  \0  \0  \0   `   Û      \0 001   ²
 0144100 000 000 000 000 000 000 000 000 144 000 000 000 000 000 000 000
          \0  \0  \0  \0  \0  \0  \0  \0   d  \0  \0  \0  \0  \0  \0  \0
-0144120 000 057 164 155 160 057 143 143 062 151 102 155 110 150 056 163
-         \0   /   t   m   p   /   c   c   2   i   B   m   H   h   .   s
+0144120 000 057 164 155 160 057 143 143 113 105 067 171 166 123 056 163
+         \0   /   t   m   p   /   c   c   K   E   7   y   v   S   .   s
 0144140 000 057 150 157 155 145 057 164 150 141 154 145 163 057 145 150
          \0   /   h   o   m   e   /   t   h   a   l   e   s   /   e   h
 0144160 162 150 141 162 144 164 057 147 143 143 055 063 056 064 055 143

I'm using GNU-as and GNU-ld.

   regards  Christian

[Bug bootstrap/11318] [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[ehrhardt at mathematik dot uni-ulm dot de]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11318


ehrhardt at mathematik dot uni-ulm dot de changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |zack at codesourcery dot com
                   |dot org                     |
             Status|UNCONFIRMED                 |NEW


------- Additional Comments From ehrhardt at mathematik dot uni-ulm dot de  2003-06-25 19:04 -------
I tracked the cause of this to a missing .file asm directive on
sparc. The patch below fixes bootstrap for me again. Zack: You were
the last to make changes in this area, what do you think about the
following patch?

Index: sparc/sparc.c
===================================================================
RCS file: /cvsroot/gcc/gcc/gcc/config/sparc/sparc.c,v
retrieving revision 1.250
diff -c -r1.250 sparc.c
*** sparc/sparc.c       19 Jun 2003 06:05:34 -0000      1.250
--- sparc/sparc.c       25 Jun 2003 19:00:19 -0000
***************
*** 255,260 ****
--- 255,263 ----
  #undef TARGET_ADDRESS_COST
  #define TARGET_ADDRESS_COST hook_int_rtx_0
  
+ #undef TARGET_ASM_FILE_START_FILE_DIRECTIVE
+ #define TARGET_ASM_FILE_START_FILE_DIRECTIVE true
+ 
  struct gcc_target targetm = TARGET_INITIALIZER;
  
  /* Validate and override various options, and do some machine dependent


   regards  Christian

[Bug bootstrap/11318] [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[zack at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11318


zack at gcc dot gnu dot org changed:

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


------- Additional Comments From zack at gcc dot gnu dot org  2003-06-25 19:17 -------
That should be unnecessary; sparc-sun-solaris should pick up the elfos.h
TARGET_ASM_FILE_START_FILE_DIRECTIVE setting.  Setting it unconditionally
in sparc/sparc.c will break older (a.out) targets that don't want a .file
directive.

Please investigate why the elfos.h definition is not getting used.

[Bug bootstrap/11318] [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[ehrhardt at mathematik dot uni-ulm dot de]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11318



------- Additional Comments From ehrhardt at mathematik dot uni-ulm dot de  2003-06-26 09:07 -------
Subject: Re:  [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

On Wed, Jun 25, 2003 at 07:17:31PM -0000, zack at gcc dot gnu dot org wrote:
> That should be unnecessary; sparc-sun-solaris should pick up the elfos.h
> TARGET_ASM_FILE_START_FILE_DIRECTIVE setting.  Setting it unconditionally
> in sparc/sparc.c will break older (a.out) targets that don't want a .file
> directive.

Unless I'm missing something gcc/config/elfos.h doesn't have a definition
for the TARGET_ASM_FILE_START_FILE_DIRECTIVE macro. The macro is then defined
to 0 in target-def.h. Does this patch look better?

Index: elfos.h
===================================================================
RCS file: /cvsroot/gcc/gcc/gcc/config/elfos.h,v
retrieving revision 1.61
diff -c -r1.61 elfos.h
*** elfos.h     19 Jun 2003 21:47:02 -0000      1.61
--- elfos.h     26 Jun 2003 09:04:08 -0000
***************
*** 230,235 ****
--- 230,239 ----
  #undef        TARGET_ASM_SELECT_SECTION
  #define TARGET_ASM_SELECT_SECTION default_elf_select_section
  
+ /* Start each asm-file with a .file directive. */
+ #undef TARGET_ASM_FILE_START_FILE_DIRECTIVE
+ #define TARGET_ASM_FILE_START_FILE_DIRECTIVE true
+ 
  /* Define the strings used for the special svr4 .type and .size directives.
     These strings generally do not vary from one system running svr4 to
     another, but if a given system (e.g. m88k running svr) needs to use


   regards  Christian

[Bug bootstrap/11318] [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[zack at codesourcery dot com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11318



------- Additional Comments From zack at codesourcery dot com  2003-06-26 16:04 -------
Subject: Re:  [3.4 regresssion] Spurious bootstrap
 comparison failures on sparc-sun-solaris

"ehrhardt at mathematik dot uni-ulm dot de" <gcc-bugzilla@gcc.gnu.org> writes:

> Unless I'm missing something gcc/config/elfos.h doesn't have a definition
> for the TARGET_ASM_FILE_START_FILE_DIRECTIVE macro. The macro is then defined
> to 0 in target-def.h. Does this patch look better?

It's right here:

| #undef  SET_ASM_OP
| #define SET_ASM_OP      "\t.set\t"
|
| /* Most svr4 assemblers want a .file directive at the beginning of
|   their input file.  */
| #define ASM_FILE_START_FILE_DIRECTIVE true

zw

[Bug bootstrap/11318] [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[ehrhardt at mathematik dot uni-ulm dot de]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11318



------- Additional Comments From ehrhardt at mathematik dot uni-ulm dot de  2003-06-27 08:24 -------
Subject: Re:  [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris


We're getting closer:

On Thu, Jun 26, 2003 at 04:04:52PM -0000, zack at codesourcery dot com wrote:
> > Unless I'm missing something gcc/config/elfos.h doesn't have a definition
> > for the TARGET_ASM_FILE_START_FILE_DIRECTIVE macro. [ ... ]
            ^^^^^^^
> It's right here:
> [ ... ]
> | #define ASM_FILE_START_FILE_DIRECTIVE true
           ^^

TARGET_ASM_FILE_START_FILE_DIRECTIVE is used all over gcc and documented
   in gcc.int.
ASM_FILE_START_FILE_DIRECTIVE (note the missing TARGET_ prefix) is defined
   in elfos.h and a few other files but never really used.

   regards  Christian

[Bug bootstrap/11318] [3.4 regresssion] Spurious bootstrap comparison failures on sparc-sun-solaris

[ehrhardt at mathematik dot uni-ulm dot de]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11318


ehrhardt at mathematik dot uni-ulm dot de changed:

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


------- Additional Comments From ehrhardt at mathematik dot uni-ulm dot de  2003-06-28 08:49 -------
Fixed by Zack's patch. Thanks.