8:22 a.m.
While doing some testing I ran into a surprising crash-build
bug on ppc64, where on a RHEL5 ppc64 host, a simple "make"
command would build a 32-bit PPC crash utility. But on RHEL6,
it built as 64-bit PPC64 crash utility as expected. (This was due
to a crash-6.0.3 change to support "make target=PPC" on
a ppc64 host).
Anyway, the reason is that in the RHEL5 environment, if an
application is compiled without specifying -m32 or -m64,
it builds a 32-bit binary, whereas on RHEL6 it builds a
64-bit binary by default:
RHEL5:
# cat doit.c
main() {}
# gcc doit.c
# file a.out
a.out: ELF 32-bit MSB executable, PowerPC or cisco 4500, version 1 (SYSV), for GNU/Linux
2.6.9, dynamically linked (uses shared libs), not stripped
# gcc --version
gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-52)
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.
# uname -r
2.6.18-308.el5
#
RHEL6:
# cat doit.c
main() {}
# gcc doit.c
# file a.out
a.out: ELF 64-bit MSB executable, 64-bit PowerPC or cisco 7500, version 1 (SYSV),
dynamically linked (uses shared libs), for GNU/Linux 2.6.18, not stripped
# gcc --version
gcc (GCC) 4.4.6 20110731 (Red Hat 4.4.6-3)
Copyright (C) 2010 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.
# uname -r
2.6.32-220.el6.ppc64
#
What controls that? Is it the compiler, some environment setting, or what?
Dave
7:25 p.m.
The default for RHEL5 is 32bit:
http://www.ibm.com/developerworks/wikis/display/LinuxP/Performance+FAQs#P...
The default for RHEL6 presumably changed to be 64bit like SLES11. To quote big blue (when
talking about RHEL5): "It is logical that the next version of RHEL would likely move
to a 64-bit default."
-----Original Message-----
From: crash-utility-bounces(a)redhat.com [mailto:crash-utility-bounces@redhat.com] On Behalf
Of Dave Anderson
Sent: Tuesday, June 26, 2012 11:23 PM
To: Discussion list for crash utility usage, maintenance and development
Subject: [Crash-utility] PPC64 compiler question
While doing some testing I ran into a surprising crash-build
bug on ppc64, where on a RHEL5 ppc64 host, a simple "make"
command would build a 32-bit PPC crash utility. But on RHEL6,
it built as 64-bit PPC64 crash utility as expected. (This was due
to a crash-6.0.3 change to support "make target=PPC" on
a ppc64 host).
Anyway, the reason is that in the RHEL5 environment, if an
application is compiled without specifying -m32 or -m64,
it builds a 32-bit binary, whereas on RHEL6 it builds a
64-bit binary by default:
RHEL5:
# cat doit.c
main() {}
# gcc doit.c
# file a.out
a.out: ELF 32-bit MSB executable, PowerPC or cisco 4500, version 1 (SYSV), for GNU/Linux
2.6.9, dynamically linked (uses shared libs), not stripped
# gcc --version
gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-52)
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.
# uname -r
2.6.18-308.el5
#
RHEL6:
# cat doit.c
main() {}
# gcc doit.c
# file a.out
a.out: ELF 64-bit MSB executable, 64-bit PowerPC or cisco 7500, version 1 (SYSV),
dynamically linked (uses shared libs), for GNU/Linux 2.6.18, not stripped
# gcc --version
gcc (GCC) 4.4.6 20110731 (Red Hat 4.4.6-3)
Copyright (C) 2010 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.
# uname -r
2.6.32-220.el6.ppc64
#
What controls that? Is it the compiler, some environment setting, or what?
Dave
--
Crash-utility mailing list
Crash-utility(a)redhat.com
https://www.redhat.com/mailman/listinfo/crash-utility
11:25 p.m.
Yes, The default user space for RHEL 5 and SLES10 is 32bit, but we always
build 64bit crash since PPC64 support is added. Top level makefile should
have something like:
ifeq (${TARGET}, PPC64)
CONF_FLAGS = -m64
endif
So, this change has to be modified to support cross-compile.
Thanks
Haren
"Seymour, Shane
M"
<shane.seymour@hp To
.com> "Discussion list for crash utility
Sent by: usage, maintenance and
crash-utility-bou development"
nces(a)redhat.com <crash-utility(a)redhat.com>
cc
06/26/2012 05:25 Subject
PM Re: [Crash-utility] PPC64 compiler
question
Please respond to
"Discussion list
for crash utility
usage,
maintenance and
development"
<crash-utility@re
dhat.com>
The default for RHEL5 is 32bit:
http://www.ibm.com/developerworks/wikis/display/LinuxP/Performance
+FAQs#PerformanceFAQs-WhyistheRHEL5defaultforapplications32bitandtheSLES11default64bit%3F
The default for RHEL6 presumably changed to be 64bit like SLES11. To quote
big blue (when talking about RHEL5): "It is logical that the next version
of RHEL would likely move to a 64-bit default."
-----Original Message-----
From: crash-utility-bounces(a)redhat.com [
mailto:crash-utility-bounces@redhat.com] On Behalf Of Dave Anderson
Sent: Tuesday, June 26, 2012 11:23 PM
To: Discussion list for crash utility usage, maintenance and development
Subject: [Crash-utility] PPC64 compiler question
While doing some testing I ran into a surprising crash-build
bug on ppc64, where on a RHEL5 ppc64 host, a simple "make"
command would build a 32-bit PPC crash utility. But on RHEL6,
it built as 64-bit PPC64 crash utility as expected. (This was due
to a crash-6.0.3 change to support "make target=PPC" on
a ppc64 host).
Anyway, the reason is that in the RHEL5 environment, if an
application is compiled without specifying -m32 or -m64,
it builds a 32-bit binary, whereas on RHEL6 it builds a
64-bit binary by default:
RHEL5:
# cat doit.c
main() {}
# gcc doit.c
# file a.out
a.out: ELF 32-bit MSB executable, PowerPC or cisco 4500, version 1 (SYSV),
for GNU/Linux 2.6.9, dynamically linked (uses shared libs), not stripped
# gcc --version
gcc (GCC) 4.1.2 20080704 (Red Hat 4.1.2-52)
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.
# uname -r
2.6.18-308.el5
#
RHEL6:
# cat doit.c
main() {}
# gcc doit.c
# file a.out
a.out: ELF 64-bit MSB executable, 64-bit PowerPC or cisco 7500, version 1
(SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.18, not
stripped
# gcc --version
gcc (GCC) 4.4.6 20110731 (Red Hat 4.4.6-3)
Copyright (C) 2010 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.
# uname -r
2.6.32-220.el6.ppc64
#
What controls that? Is it the compiler, some environment setting, or what?
Dave
--
Crash-utility mailing list
Crash-utility(a)redhat.com
https://www.redhat.com/mailman/listinfo/crash-utility
--
Crash-utility mailing list
Crash-utility(a)redhat.com
https://www.redhat.com/mailman/listinfo/crash-utility
11:10 a.m.
----- Original Message -----
Yes, The default user space for RHEL 5 and SLES10 is 32bit, but we
always build 64bit crash since PPC64 support is added. Top level
makefile should have something like:
ifeq (${TARGET}, PPC64)
CONF_FLAGS = -m64
endif
So, this change has to be modified to support cross-compile.
Thanks
Haren
Actually the crash-6.0.3 Makefile change that set "ifeq (${TARGET}, PPC64)"
is the problem at hand. The very first time that "make" is run on a virgin
crash source tree the "TARGET" is not known/set. So on a RHEL5 system,
configure.c is compiled as a 32-bit binary, and therefore thinks it's
32-bit host. The Makefile needs to be restored to the way it used to
work -- by checking "ifeq (${ARCH}, ppc64)" -- where ARCH gets set in the
Makefile just above it based upon a "uname -m" shell call. When I originally
tested the crash-6.0.3 patches that were posted, I used a RHEL6 ppc64 host,
where it worked by luck.
For cross-compile, if the user enters "make target=PPC", then the configure
program will override the host ARCH. And upon subsequent builds, it will
override the host ARCH by checking the "gdb-7.3.1/crash.target" file
that was created by the first build. That way, the target is "sticky"
so that you don't have to keep entering "make target=PPC" after it
has been done the first time.
Dave
10:57 a.m.
----- Original Message -----
The default for RHEL5 is 32bit:
http://www.ibm.com/developerworks/wikis/display/LinuxP/Performance+FAQs#P...
The default for RHEL6 presumably changed to be 64bit like SLES11. To
quote big blue (when talking about RHEL5): "It is logical that the
next version of RHEL would likely move to a 64-bit default."
Right -- but is the change built into the gcc binary? Or is there
some other file/environment-variable or whatever that gcc looks at?
Dave
4386
days inactive
4388
days old
devel@lists.crash-utility.osci.io
4 comments
3 participants
participants (3)
-
Dave Anderson -
Haren Myneni -
Seymour, Shane M