Hello list,
While working on building a driver update module RPM for CentOS 6 x86_64 I noticed that that the Module.symvers file included in the non-debug kernel-devels for the latest CentOS-Plus x86_64 kernels seems to be much shorter than that for the debug version and also than both version for the i686 architecture
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686/Module.symvers | sort | uniq -c 6124 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686.debug/Module.symvers | sort | uniq -c 6177 EXPORT_SYMBOL 3734 EXPORT_SYMBOL_GPL
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64/Module.symvers | sort | uniq -c *2716 EXPORT_SYMBOL* *1957 EXPORT_SYMBOL_GPL*
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64.debug/Module.symvers | sort | uniq -c 6147 EXPORT_SYMBOL 3916 EXPORT_SYMBOL_GPL
This prevents the find find-requires.ksyms script (part of the %kernel_module_package utility set) from creating KABI dependencies properly on x86_64, for kernel module packages build according RHEL driver update program (i.e. kmod-* package), in the sense that they are build without any KABI-requires in the resulting kmod-RPM.
The standard CentOS kernel-devel (presumably RHEL upstream?) however, contains similar numbers of symbols in Module.symvers also in the x86_64 architecture:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.el6.i686/Module.symvers | sort | uniq -c 6119 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.el6.x86_64/Module.symvers | sort | uniq -c 6089 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL
Could anyone confirm whether or not this is by design? Or should I report this as a bug?
TIA & Regards, Thomas
On Sun, Feb 26, 2012 at 9:13 AM, Thomas Jacob jacob@internet24.de wrote:
Hello list,
While working on building a driver update module RPM for CentOS 6 x86_64 I noticed that that the Module.symvers file included in the non-debug kernel-devels for the latest CentOS-Plus x86_64 kernels seems to be much shorter than that for the debug version and also than both version for the i686 architecture
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686/Module.symvers | sort | uniq -c 6124 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686.debug/Module.symvers | sort | uniq -c 6177 EXPORT_SYMBOL 3734 EXPORT_SYMBOL_GPL
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64/Module.symvers | sort | uniq -c *2716 EXPORT_SYMBOL* *1957 EXPORT_SYMBOL_GPL*
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64.debug/Module.symvers | sort | uniq -c 6147 EXPORT_SYMBOL 3916 EXPORT_SYMBOL_GPL
This prevents the find find-requires.ksyms script (part of the %kernel_module_package utility set) from creating KABI dependencies properly on x86_64, for kernel module packages build according RHEL driver update program (i.e. kmod-* package), in the sense that they are build without any KABI-requires in the resulting kmod-RPM.
The standard CentOS kernel-devel (presumably RHEL upstream?) however, contains similar numbers of symbols in Module.symvers also in the x86_64 architecture:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.el6.i686/Module.symvers | sort | uniq -c 6119 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.el6.x86_64/Module.symvers | sort | uniq -c 6089 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL
Could anyone confirm whether or not this is by design? Or should I report this as a bug?
By all means, please file a bug report. This is certainly not by design. :-(
Akemi
On Sun, Feb 26, 2012 at 9:13 AM, Thomas Jacob jacob@internet24.de wrote:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64/Module.symvers | sort | uniq -c *2716 EXPORT_SYMBOL* *1957 EXPORT_SYMBOL_GPL*
However this is not a centosplus kernel. It should say:
/usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers
And I see:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers | sort | uniq -c 6193 EXPORT_SYMBOL 3897 EXPORT_SYMBOL_GPL
Akemi
On Feb 26, 2012, at 6:41 PM, Akemi Yagi wrote:
On Sun, Feb 26, 2012 at 9:13 AM, Thomas Jacob jacob@internet24.de wrote:
However this is not a centosplus kernel. It should say:
/usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers
And I see:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers | sort | uniq -c 6193 EXPORT_SYMBOL 3897 EXPORT_SYMBOL_GPL
Oh yes, sorry about that, I seem to have mixed this up with the "updates" packages of CentOS proper.
So is this an RHEL upstream problem then? Sorry I am rather new to the CentOS package building intrinsics ....
Akemi _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
On Sun, Feb 26, 2012 at 9:54 AM, Thomas Jacob jacob@internet24.de wrote:
On Feb 26, 2012, at 6:41 PM, Akemi Yagi wrote:
On Sun, Feb 26, 2012 at 9:13 AM, Thomas Jacob jacob@internet24.de wrote:
However this is not a centosplus kernel. It should say:
/usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers
And I see:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers | sort | uniq -c 6193 EXPORT_SYMBOL 3897 EXPORT_SYMBOL_GPL
Oh yes, sorry about that, I seem to have mixed this up with the "updates" packages of CentOS proper.
So is this an RHEL upstream problem then? Sorry I am rather new to the CentOS package building intrinsics ....
I don't think this problem exists in the RHEL kernel. Could someone check the distro CentOS 2.6.32-220.4.2.el6 kernel?
Akemi
On 02/26/2012 07:58 PM, Akemi Yagi wrote:
On Sun, Feb 26, 2012 at 9:54 AM, Thomas Jacobjacob@internet24.de wrote:
On Feb 26, 2012, at 6:41 PM, Akemi Yagi wrote:
On Sun, Feb 26, 2012 at 9:13 AM, Thomas Jacobjacob@internet24.de wrote:
However this is not a centosplus kernel. It should say:
/usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers
And I see:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers | sort | uniq -c 6193 EXPORT_SYMBOL 3897 EXPORT_SYMBOL_GPL
Oh yes, sorry about that, I seem to have mixed this up with the "updates" packages of CentOS proper.
So is this an RHEL upstream problem then? Sorry I am rather new to the CentOS package building intrinsics ....
I don't think this problem exists in the RHEL kernel. Could someone check the distro CentOS 2.6.32-220.4.2.el6 kernel?
[wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.2.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [root@DB ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686/Module.symvers | sort | uniq -c 6124 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL
On 02/27/2012 12:08 AM, Manuel Wolfshant wrote:
On 02/26/2012 07:58 PM, Akemi Yagi wrote:
On Sun, Feb 26, 2012 at 9:54 AM, Thomas Jacobjacob@internet24.de wrote:
On Feb 26, 2012, at 6:41 PM, Akemi Yagi wrote:
On Sun, Feb 26, 2012 at 9:13 AM, Thomas Jacobjacob@internet24.de wrote:
However this is not a centosplus kernel. It should say:
/usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers
And I see:
$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.centos.plus.x86_64/Module.symvers | sort | uniq -c 6193 EXPORT_SYMBOL 3897 EXPORT_SYMBOL_GPL
Oh yes, sorry about that, I seem to have mixed this up with the "updates" packages of CentOS proper.
So is this an RHEL upstream problem then? Sorry I am rather new to the CentOS package building intrinsics ....
I don't think this problem exists in the RHEL kernel. Could someone check the distro CentOS 2.6.32-220.4.2.el6 kernel?
[wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.2.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [root@DB ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686/Module.symvers | sort | uniq -c 6124 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL
Sorry for not mentioning: all 3 above are pure centos kernels. And the number of symbols does not coincide with the ones from the original post...
Guten Tag Manuel,
you threw in a 686 kernel at the 3rd prompt :-) - Still the numbers differ from the OP...
[root@pegasus ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL
[wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.2.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [root@DB ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686/Module.symvers | sort | uniq -c 6124 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL
CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
On 02/27/2012 01:37 AM, Thomas Göttgens wrote:
Guten Tag Manuel,
you threw in a 686 kernel at the 3rd prompt :-) - Still the numbers differ from the OP...
On purpose. Results taken from a different machine with different arch. The original poster used results for 32 bits and I wanted to rule out smtg specific to a particular arch.
[root@pegasus ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL
[wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.2.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [root@DB ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686/Module.symvers | sort | uniq -c 6124 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
Sorry for not mentioning that on the list also, but there is no bug, the whole thing was related to a corrupted file on my system, cf. http://bugs.centos.org/view.php?id=5550
On Feb 27, 2012, at 9:40 AM, Manuel Wolfshant wrote:
On 02/27/2012 01:37 AM, Thomas Göttgens wrote:
Guten Tag Manuel,
you threw in a 686 kernel at the 3rd prompt :-) - Still the numbers differ from the OP...
On purpose. Results taken from a different machine with different arch. The original poster used results for 32 bits and I wanted to rule out smtg specific to a particular arch.
[root@pegasus ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL
[wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.2.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [wolfy@wolfy2 V101]$ awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.1.el6.x86_64/Module.symvers | sort | uniq -c 6094 EXPORT_SYMBOL 3888 EXPORT_SYMBOL_GPL [root@DB ~]# awk '{print $4;}' /usr/src/kernels/2.6.32-220.4.2.el6.i686/Module.symvers | sort | uniq -c 6124 EXPORT_SYMBOL 3706 EXPORT_SYMBOL_GPL _______________________________________________ CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
-- Manuel Wolfshant linux registered user #131416 IT manager NoBug Consulting SRL A: Yes.
Q: Are you sure?
A: Because it reverses the logical flow of conversation.
Q: Why is top posting frowned upon?
CentOS-devel mailing list CentOS-devel@centos.org http://lists.centos.org/mailman/listinfo/centos-devel
-
Akemi Yagi -
Manuel Wolfshant -
Thomas Göttgens -
Thomas Jacob