[Bug bootstrap/43733] New: bootstrap fails building libgfortran on Solaris x86 with GNU as

[Bug bootstrap/43733] New: bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

../gcc-4.4.3/configure --prefix=/opt/gcc-4.4.3-64/gcc-4.4.3
--enable-languages=c,fortran --with-system-zlib --with-arch=core2
--with-as=/usr/sfw/bin/gas --with-gnu-as --with-ld=/usr/ccs/bin/ld
--without-gnu-ld

Bootstrap fails with

gmake[5]: Entering directory
`/var/tmp/gcc/tmp/i386-pc-solaris2.10/amd64/libgfortran'
gmake  all-am
gmake[6]: Entering directory
`/var/tmp/gcc/tmp/i386-pc-solaris2.10/amd64/libgfortran'
/bin/bash ./libtool --tag=CC --mode=link /var/tmp/gcc/tmp/./gcc/xgcc
-B/var/tmp/gcc/tmp/./gcc/
-B/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/bin/
-B/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/lib/ -isystem
/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/include -isystem
/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/sys-include -std=gnu99 -Wall
-Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -Wextra
-Wwrite-strings -fcx-fortran-rules -g -O2    -m64  -m64 -o libgfortranbegin.la
-rpath /opt/gcc-4.4.3-64/gcc-4.4.3/lib/gcc/i386-pc-solaris2.10/4.4.3/amd64
-static fmain.lo
libtool: link: ar cr .libs/libgfortranbegin.a  fmain.o
libtool: link: ranlib .libs/libgfortranbegin.a
libtool: link: ( cd ".libs" && rm -f "libgfortranbegin.la" && ln -s
"../libgfortranbegin.la" "libgfortranbegin.la" )
if /bin/bash ./libtool --tag=CC --mode=compile /var/tmp/gcc/tmp/./gcc/xgcc
-B/var/tmp/gcc/tmp/./gcc/
-B/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/bin/
-B/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/lib/ -isystem
/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/include -isystem
/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/sys-include -DHAVE_CONFIG_H -I.
-I../../../../gcc-4.4.3/libgfortran -I. 
-iquote../../../../gcc-4.4.3/libgfortran/io
-I../../../../gcc-4.4.3/libgfortran/../gcc
-I../../../../gcc-4.4.3/libgfortran/../gcc/config -I../../.././gcc
-D_GNU_SOURCE  -std=gnu99 -Wall -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition -Wextra -Wwrite-strings -fcx-fortran-rules -g -O2   
-m64 -MT date_and_time.lo -MD -MP -MF ".deps/date_and_time.Tpo" -c -o
date_and_time.lo `test -f 'intrinsics/date_and_time.c' || echo
'../../../../gcc-4.4.3/libgfortran/'`intrinsics/date_and_time.c; \
        then mv -f ".deps/date_and_time.Tpo" ".deps/date_and_time.Plo"; else rm
-f ".deps/date_and_time.Tpo"; exit 1; fi
libtool: compile:  /var/tmp/gcc/tmp/./gcc/xgcc -B/var/tmp/gcc/tmp/./gcc/
-B/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/bin/
-B/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/lib/ -isystem
/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/include -isystem
/opt/gcc-4.4.3-64/gcc-4.4.3/i386-pc-solaris2.10/sys-include -DHAVE_CONFIG_H -I.
-I../../../../gcc-4.4.3/libgfortran -I.
-iquote../../../../gcc-4.4.3/libgfortran/io
-I../../../../gcc-4.4.3/libgfortran/../gcc
-I../../../../gcc-4.4.3/libgfortran/../gcc/config -I../../.././gcc
-D_GNU_SOURCE -std=gnu99 -Wall -Wstrict-prototypes -Wmissing-prototypes
-Wold-style-definition -Wextra -Wwrite-strings -fcx-fortran-rules -g -O2 -m64
-MT date_and_time.lo -MD -MP -MF .deps/date_and_time.Tpo -c
../../../../gcc-4.4.3/libgfortran/intrinsics/date_and_time.c  -fPIC -DPIC -o
.libs/date_and_time.o
/var/tmp//cc6vGz5w.s: Assembler messages:
/var/tmp//cc6vGz5w.s:1451: Error: suffix or operands invalid for `sahf'
gmake[6]: *** [date_and_time.lo] Error 1
gmake[6]: Leaving directory
`/var/tmp/gcc/tmp/i386-pc-solaris2.10/amd64/libgfortran'
gmake[5]: *** [all] Error 2
gmake[5]: Leaving directory
`/var/tmp/gcc/tmp/i386-pc-solaris2.10/amd64/libgfortran'
gmake[4]: *** [multi-do] Error 1
gmake[4]: Leaving directory `/var/tmp/gcc/tmp/i386-pc-solaris2.10/libgfortran'
gmake[3]: *** [all-multi] Error 2
gmake[3]: Leaving directory `/var/tmp/gcc/tmp/i386-pc-solaris2.10/libgfortran'
gmake[2]: *** [all] Error 2
gmake[2]: Leaving directory `/var/tmp/gcc/tmp/i386-pc-solaris2.10/libgfortran'
gmake[1]: *** [all-target-libgfortran] Error 2
gmake[1]: Leaving directory `/var/tmp/gcc/tmp'
gmake: *** [all] Error 2


The same configuration works fine with "--enable-languages=c" or without
"--with-as=/usr/sfw/bin/gas --with-gnu-as"
It only fails if building fortran and using GNU as


CONFIG_SHELL=/usr/bin/bash

GNU bash, version 3.00.16(1)-release (i386-pc-solaris2.10)
Copyright (C) 2004 Free Software Foundation, Inc.

$ /usr/ccs/bin/ld -V
ld: Software Generation Utilities - Solaris Link Editors: 5.10-1.490

$ /usr/ccs/bin/as -V
as: SunOS 5.10 119961-05 Patch 03/16/2009

$ /usr/sfw/bin/gas --version
GNU assembler 2.15
Copyright 2002 Free Software Foundation, Inc.
This program is free software; you may redistribute it under the terms of
the GNU General Public License.  This program has absolutely no warranty.
This assembler was configured for a target of `i386-pc-solaris2.10'.

$ make --version
GNU Make 3.81
Copyright (C) 2006  Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

This program built for i386-pc-solaris2.10

N.B. I have /usr/xpg4/bin first in my PATH to work around Bug 38923


-- 
           Summary: bootstrap fails building libgfortran on Solaris x86 with
                    GNU as
           Product: gcc
           Version: 4.4.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: redi at gcc dot gnu dot org
 GCC build triplet: i386-pc-solaris2.10
  GCC host triplet: i386-pc-solaris2.10
GCC target triplet: i386-pc-solaris2.10


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #1 from redi at gcc dot gnu dot org  2010-04-12 16:51 -------
P.S. I'm using in-tree gmp 4.4.3 and mpfr 2.4.2

I'm currently trying another bootstrap without --with-arch=core2


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #2 from ubizjak at gmail dot com  2010-04-12 17:11 -------
configure should detect if assembler supports sahf mnemonic.

In your build directory, check gcc/config.log for:

configure:23019: checking assembler for sahf mnemonic
configure:23028: /usr/local/x86_64-unknown-linux-gnu/bin/as    -o conftest.o
conftest.s >&5
configure:23031: $? = 0
configure:23042: result: yes

Based on this result, gcc outputs:

(define_insn "x86_sahf_1"
  [(set (reg:CC FLAGS_REG)
        (unspec:CC [(match_operand:HI 0 "register_operand" "a")]
                   UNSPEC_SAHF))]
  "TARGET_SAHF"
{
#ifdef HAVE_AS_IX86_SAHF
  return "sahf";
#else
  return ASM_BYTE "0x9e";
#endif
}


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #3 from redi at gcc dot gnu dot org  2010-04-12 17:15 -------
(In reply to comment #1)
> I'm currently trying another bootstrap without --with-arch=core2

That worked OK, so it's only the combination of /usr/sfw/bin/gas, fortran and
arch=core2 that files 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #4 from redi at gcc dot gnu dot org  2010-04-12 17:15 -------
configure:22928: checking assembler for sahf mnemonic
configure:22937: /usr/sfw/bin/gas  -o conftest.o conftest.s >&5
configure:22940: $? = 0
configure:22951: result: yes


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #5 from ubizjak at gmail dot com  2010-04-12 17:24 -------
Looks to me that this is libgfortran messing with choosen assembler.

The bootstrap would die way before libgfortran is touched otherwise.

I'm changing component to libgfortran in the hope that fortraners will confirm
(and eventually fix ;) this PR.


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|bootstrap                   |libfortran


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #6 from redi at gcc dot gnu dot org  2010-04-12 17:25 -------
running the failing command with -save-temps shows that sahf isused here:

.LM312:
    fprem
    fnstsw  %ax
    sahf
    jp  .L103
    fstp    %st(1)
    fstpl   -56(%rbp)
    movsd   -56(%rbp), %xmm1
    ucomisd %xmm1, %xmm1
    jp  .L109
    jne .L109


adding -v confirms that /usr/sfw/bin/gas is being used as the assembler, not
some other as


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #7 from redi at gcc dot gnu dot org  2010-04-12 17:30 -------
the failure is while building the 64bit libgfortran, is sahf valid in 64-bit
mode?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #8 from ubizjak at gmail dot com  2010-04-12 17:47 -------
(In reply to comment #3)
> (In reply to comment #1)
> > I'm currently trying another bootstrap without --with-arch=core2
> 
> That worked OK, so it's only the combination of /usr/sfw/bin/gas, fortran and
> arch=core2 that files 

-march=core2 is required to generate sahf. There are targets (generic k8) that
die with SIGILL when they execute 0x9e opcode. Look for lahf_lm in
/proc/cpuinfo.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #9 from ubizjak at gmail dot com  2010-04-12 17:56 -------
(In reply to comment #6)
> running the failing command with -save-temps shows that sahf isused here:
> 
> .LM312:
>     fprem
>     fnstsw  %ax
>     sahf
>     jp  .L103
>     fstp    %st(1)
>     fstpl   -56(%rbp)
>     movsd   -56(%rbp), %xmm1
>     ucomisd %xmm1, %xmm1
>     jp  .L109
>     jne .L109
> 
> 
> adding -v confirms that /usr/sfw/bin/gas is being used as the assembler, not
> some other as

Let me guess, /usr/sfw/bin/gas compiles this asm file OK, while /usr/ccs/bin/ld
doesn't?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #10 from ubizjak at gmail dot com  2010-04-12 17:57 -------
(In reply to comment #9)

> Let me guess, /usr/sfw/bin/gas compiles this asm file OK, while /usr/ccs/bin/ld
> doesn't?

Well, /usr/ccs/bin/as, of course.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #11 from redi at gcc dot gnu dot org  2010-04-12 18:02 -------
(In reply to comment #10)
> (In reply to comment #9)
> 
> > Let me guess, /usr/sfw/bin/gas compiles this asm file OK, while /usr/ccs/bin/ld
> > doesn't?
> 
> Well, /usr/ccs/bin/as, of course.
> 

gas barfs on it, this is the from the output of manually running the failing
xgcc command, with -v and -save-temps added:

 /usr/sfw/bin/gas -v -I. -I../../../../gcc-4.4.3/libgfortran -I.
-I../../../../gcc-4.4.3/libgfortran/../gcc
-I../../../../gcc-4.4.3/libgfortran/../gcc/config -I../../.././gcc
--traditional-format -V -Qy --64 -s -o .libs/date_and_time.o date_and_time.s
GNU assembler version 2.15 (i386-pc-solaris2.10) using BFD version 2.15
date_and_time.s: Assembler messages:
date_and_time.s:1451: Error: suffix or operands invalid for `sahf'


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #12 from redi at gcc dot gnu dot org  2010-04-12 18:05 -------
(In reply to comment #3)
> (In reply to comment #1)
> > I'm currently trying another bootstrap without --with-arch=core2
> 
> That worked OK, so it's only the combination of /usr/sfw/bin/gas, fortran and
> arch=core2 that files 

oops - stupid fingers - I meant fails, not files!


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #13 from redi at gcc dot gnu dot org  2010-04-12 18:08 -------
I can't read or write assembler, but searching the interweb tells me that sahf
is not valid in 64-bit mode, e.g.
http://www.x86-64.org/pipermail/discuss/2004-April/004615.html and also
"Introduction to 80x86 Assembly Language and Computer Architecture" by Richard
C. Detmer

Why is that instruction being issued when compiling amd64/libgfortran?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #14 from ubizjak at gmail dot com  2010-04-12 18:16 -------
(In reply to comment #13)
> I can't read or write assembler, but searching the interweb tells me that sahf
> is not valid in 64-bit mode, e.g.
> http://www.x86-64.org/pipermail/discuss/2004-April/004615.html and also
> "Introduction to 80x86 Assembly Language and Computer Architecture" by Richard
> C. Detmer
> 
> Why is that instruction being issued when compiling amd64/libgfortran?

Don't worry about this insn, it works for core2.

The problem is that your ./configure claims that assembler supports this
mnemonic (Comment #4) while in comment #11, it doesn't. Can you debug this a
bit, I don't have binutils-2.15 here anymore.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #15 from redi at gcc dot gnu dot org  2010-04-12 18:21 -------
(In reply to comment #14)
> > Why is that instruction being issued when compiling amd64/libgfortran?
> 
> Don't worry about this insn, it works for core2.

GNU as 2.15 doesn't believe you :-)

$ echo  sahf > test.s
$ /usr/sfw/bin/gas test.s
$ /usr/sfw/bin/gas --64 test.s
test.s: Assembler messages:
test.s:1: Error: suffix or operands invalid for `sahf'

configure only tests it the first way, which is 32-bit

there should be a configure test using --64 if it's going to be issued in
64-bit mode


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #16 from ubizjak at gmail dot com  2010-04-12 18:42 -------
Can you try this (untested) patch?

Index: acinclude.m4
===================================================================
--- acinclude.m4        (revision 158225)
+++ acinclude.m4        (working copy)
@@ -452,6 +452,9 @@
     dnl Always pass --32 to ia32 Linux assembler.
     gcc_cv_as_flags="--32"
     ;;
+  x86_64-*-*)
+    gcc_cv_as_flags="--64"
+    ;;
   *)
     gcc_cv_as_flags=" "
     ;;


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|libfortran                  |bootstrap


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[ubizjak at gmail dot com]

------- Comment #17 from ubizjak at gmail dot com  2010-04-12 20:02 -------
(In reply to comment #15)

> GNU as 2.15 doesn't believe you :-)
> 
> $ echo  sahf > test.s
> $ /usr/sfw/bin/gas test.s
> $ /usr/sfw/bin/gas --64 test.s
> test.s: Assembler messages:
> test.s:1: Error: suffix or operands invalid for `sahf'

Well - dear GNU as 2.15 - sahf doesn't have any operands. ;)

This looks like a bug in binutils 2.15, because otherwise:

$ as --64 tt.s
tt.s: Assembler messages:
tt.s:1: Error: no such instruction: `evil_nonexistent_insn'

OTOH, the patch from comment #16 is a bad idea, it will cripple non-GAS
assemblers (and as discussed elsewhere, it isn't effective for solaris anyway).

I'd suggest you just upgrade your binutils... -march=core2 switches -mssse3 and
2.15 will be immediately out of luck for any vectorized code due to the usage
of SSSE3 permute insns. 


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #18 from redi at gcc dot gnu dot org  2010-04-13 09:21 -------
Instead I've decided to use --with-arch=nocona --with-tune=core2, since that
avoids having to deploy a new binutils to every server where I want to deploy
gcc


-- 

redi at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|bootstrap                   |libfortran


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug libfortran/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[burnus at gcc dot gnu dot org]

------- Comment #19 from burnus at gcc dot gnu dot org  2010-04-13 10:12 -------
> This looks like a bug in binutils 2.15

Can we thus close the bug? Or remains there something to do in libgfortran?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails building libgfortran on Solaris x86 with GNU as

[redi at gcc dot gnu dot org]

------- Comment #20 from redi at gcc dot gnu dot org  2010-04-13 11:21 -------
oops - I didn't mean to set the component back to libfortran, that must have
happened when I refreshed the page and my browser "helpfully" kept that
selected. I've reverted it to bootstrap

I don't think it should be closed: the installation docs for Solaris x86
recommend to use the default binutils 2.15 and say it works fine, but that's
not the case if you use --with-arch=core2, that should be documented (or even
better, fixed)


-- 

redi at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|libfortran                  |bootstrap


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[redi at gcc dot gnu dot org]

------- Comment #21 from redi at gcc dot gnu dot org  2010-04-13 12:26 -------
more accurate summary


-- 

redi at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|bootstrap fails building    |bootstrap fails on Solaris
                   |libgfortran on Solaris x86  |10 x86 with GNU as 2.15 and
                   |with GNU as                 |--with-arch=core2


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[ro at gcc dot gnu dot org]

------- Comment #22 from ro at gcc dot gnu dot org  2010-04-28 18:13 -------
(In reply to comment #20)

> I don't think it should be closed: the installation docs for Solaris x86
> recommend to use the default binutils 2.15 and say it works fine, but that's
> not the case if you use --with-arch=core2, that should be documented (or even
> better, fixed)

I think there are different issues here: gas 2.15 works fine *in a default
configuration*.  Once you start messing around with --with-arch, you can
probably
easily fine an arch value that this ancient gas (8 years by now) doesn't
support.
I don't think it's useful to document every possible variation here, especially
since install.texi clearly states that current gas 2.20.1 works, too.

The other (and primary) issue probably is that --with-arch=core2 might simply
be wrong for a bi-arch compiler (as is --with-arch=pentium4).  Could you try
this with --with-arch-32=core2 instead?


-- 

ro at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[redi at gcc dot gnu dot org]

------- Comment #23 from redi at gcc dot gnu dot org  2010-04-28 18:38 -------
core2 works fine for bi-arch linux builds, but I'll try your suggestion on
Solaris asap

I assume it will work, but 64bit code will not use the core2 instructions,
which might be suboptimal (I want to use the cpu to the full potential!)

I imagine it's also possible for the same problem to occur using -march=native,
if the processor flags indicate core2 instructions will work.

That said, 2.15 is an ancient gas (even though it's very common on Solaris 10)
so I'm not too worried about this bug being fixed and I agree it's not worth
changing the config docs.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[ubizjak at gmail dot com]

------- Comment #24 from ubizjak at gmail dot com  2010-05-19 06:47 -------
Can you try this patch:

Index: configure.ac
===================================================================
--- configure.ac        (revision 159481)
+++ configure.ac        (working copy)
@@ -3254,9 +3254,10 @@

     gcc_GAS_CHECK_FEATURE([sahf mnemonic],
       gcc_cv_as_ix86_sahf,,,
-      [sahf],,
+      [.code64
+       sahf],,
       [AC_DEFINE(HAVE_AS_IX86_SAHF, 1,
-        [Define if your assembler supports the sahf mnemonic.])])
+        [Define if your assembler supports the sahf mnemonic in 64bit
mode.])])

     gcc_GAS_CHECK_FEATURE([swap suffix],
       gcc_cv_as_ix86_swap,,,


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[ubizjak at gmail dot com]

------- Comment #25 from ubizjak at gmail dot com  2010-05-19 12:21 -------
Created an attachment (id=20698)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=20698&action=view)
complete patch

Can you please test the complete patch if it solves your bootstrap failure?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[redi at gcc dot gnu dot org]

------- Comment #26 from redi at gcc dot gnu dot org  2010-05-20 10:18 -------
I tweaked the patch to apply it to 4.4.3 (replacing ASM_BYTE with .byte) and it
fixes the bootstrap failure for this config:

--enable-languages=c,c++,fortran --with-arch=core2 --with-as=/usr/sfw/bin/gas
--with-gnu-as --with-ld=/usr/ccs/bin/ld --without-gnu-ld


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug bootstrap/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[uros at gcc dot gnu dot org]

------- Comment #27 from uros at gcc dot gnu dot org  2010-05-20 18:52 -------
Subject: Bug 43733

Author: uros
Date: Thu May 20 18:52:03 2010
New Revision: 159636

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=159636
Log:
        PR target/43733
        * configure.ac (gcc_cv_as_ix86_sahf): Switch to 64bit mode.
        * configure: Regenerate.
        * config.in: Regenerate.
        * config/i386/i386.md (x86_sahf_1): Conditionally output 0x9e
        instead of sahf only for 64bit targets.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config.in
    trunk/gcc/config/i386/i386.md
    trunk/gcc/configure
    trunk/gcc/configure.ac


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug target/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[ubizjak at gmail dot com]

-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |ubizjak at gmail dot com
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
          Component|bootstrap                   |target
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2010-05-20 18:52:52
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug target/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[uros at gcc dot gnu dot org]

------- Comment #28 from uros at gcc dot gnu dot org  2010-05-24 16:23 -------
Subject: Bug 43733

Author: uros
Date: Mon May 24 16:22:30 2010
New Revision: 159785

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=159785
Log:
        Backport from mainline:
        2010-05-20  Uros Bizjak  <ubizjak@gmail.com>

        PR target/43733
        * configure.ac (gcc_cv_as_ix86_sahf): Switch to 64bit mode.
        * configure: Regenerate.
        * config.in: Regenerate.
        * config/i386/i386.md (x86_sahf_1): Conditionally output 0x9e
        instead of sahf only for 64bit targets.


Modified:
    branches/gcc-4_5-branch/gcc/ChangeLog
    branches/gcc-4_5-branch/gcc/config.in
    branches/gcc-4_5-branch/gcc/config/i386/i386.md
    branches/gcc-4_5-branch/gcc/configure
    branches/gcc-4_5-branch/gcc/configure.ac


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug target/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[uros at gcc dot gnu dot org]

------- Comment #29 from uros at gcc dot gnu dot org  2010-05-24 16:26 -------
Subject: Bug 43733

Author: uros
Date: Mon May 24 16:26:17 2010
New Revision: 159786

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=159786
Log:
        Backport from mainline:
        2010-05-20  Uros Bizjak  <ubizjak@gmail.com>

        PR target/43733
        * configure.ac (gcc_cv_as_ix86_sahf): Switch to 64bit mode.
        * configure: Regenerate.
        * config.in: Regenerate.
        * config/i386/i386.md (x86_sahf_1): Conditionally output 0x9e
        instead of sahf only for 64bit targets.


Modified:
    branches/gcc-4_4-branch/gcc/ChangeLog
    branches/gcc-4_4-branch/gcc/config.in
    branches/gcc-4_4-branch/gcc/config/i386/i386.md
    branches/gcc-4_4-branch/gcc/configure
    branches/gcc-4_4-branch/gcc/configure.ac


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug target/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[uros at gcc dot gnu dot org]

------- Comment #30 from uros at gcc dot gnu dot org  2010-05-24 16:29 -------
Subject: Bug 43733

Author: uros
Date: Mon May 24 16:28:53 2010
New Revision: 159787

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=159787
Log:
        Backport from mainline:
        2010-05-20  Uros Bizjak  <ubizjak@gmail.com>

        PR target/43733
        * configure.ac (gcc_cv_as_ix86_sahf): Switch to 64bit mode.
        * configure: Regenerate.
        * config.in: Regenerate.
        * config/i386/i386.md (x86_sahf_1): Conditionally output 0x9e
        instead of sahf only for 64bit targets.


Modified:
    branches/gcc-4_3-branch/gcc/ChangeLog
    branches/gcc-4_3-branch/gcc/config.in
    branches/gcc-4_3-branch/gcc/config/i386/i386.md
    branches/gcc-4_3-branch/gcc/configure
    branches/gcc-4_3-branch/gcc/configure.ac


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733

[Bug target/43733] bootstrap fails on Solaris 10 x86 with GNU as 2.15 and --with-arch=core2

[ubizjak at gmail dot com]

------- Comment #31 from ubizjak at gmail dot com  2010-05-24 16:31 -------
Fixed.


-- 

ubizjak at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
           Keywords|                            |build
         Resolution|                            |FIXED
   Target Milestone|---                         |4.3.6


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43733