make-kpkg doesn't work properly when compiling i386 kernel on amd64 host
Hello, I am trying to build a kernel for my i386 PC on a amd64 server. I follow a old instrusion at http://lists.debian.org/debian-user/2008/07/msg00926.html . It can build the kernel itself correctly, but the problem is script tools such as genksyms are build as amd64 format. I notice that Linux Makefile mentions an environment variable named HOSTCC. I tried to set the variable, but it still failed because of linker issue. Can make-kpkg do this job correctly? What extra parameter I should add?
Re: make-kpkg doesn't work properly when compiling i386 kernel on amd64 host
On 04/17/2013 12:44 PM, 曹思亮 wrote: Hello, I am trying to build a kernel for my i386 PC on a amd64 server. You might also try to build it inside an i386 chroot-environment on your amd64 server. I follow a old instrusion at http://lists.debian.org/debian-user/2008/07/msg00926.html . It can build the kernel itself correctly, but the problem is script tools such as genksyms are build as amd64 format. I notice that Linux Makefile mentions an environment variable named HOSTCC. I tried to set the variable, but it still failed because of linker issue. Can make-kpkg do this job correctly? What extra parameter I should add? Unfortunately, I do not know. -- http://masysma.ohost.de/ -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/516ef20f.8080...@web.de
Re: make-kpkg doesn't work properly when compiling i386 kernel on amd64 host
In article xs4all.CAHurxuitAusyBJ9aqo0-Ncegvp4wHXb_dxdSjJQue-FF=_k...@mail.gmail.com you write: Hello, I am trying to build a kernel for my i386 PC on a amd64 server. I follow a old instrusion at http://lists.debian.org/debian-user/2008/07/msg00926.html . It can build the kernel itself correctly, but the problem is script tools such as genksyms are build as amd64 format. I notice that Linux Makefile mentions an environment variable named HOSTCC. I tried to set the variable, but it still failed because of linker issue. Can make-kpkg do this job correctly? What extra parameter I should add? Yes, you can do this. 1. add ARCH=i386 to make config and similar commands, for example: cd linux-version cp .../where/ever/your/dotconfig/file/is/dotconfig .config make oldconfig ARCH=i386 2. Run make-kpkg with an i386 specific environment and flags, like: CONCURRENCY_LEVEL=3 DEB_HOST_ARCH=i386 make-kpkg \ --cross-compile - --arch i386 --rootcmd fakeroot \ --revision REVISION_YOU_USE --append-to-version APPEND_YOU_USE \ --initrd kernel_image The important thing is to set DEB_HOST_ARCH=i386 in the environment of make-kpkg and add the flags --cross-compile - --arch i386 Mike. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201304172256.r3hmuwnn019...@xs8.xs4all.nl
Re: Debian way of compiling a kernel.
In 20110209223754.5fa03...@ws82.int.tlc, Dan Serban wrote: I ask. What is the real ... accepted ... and suggested method that I follow, I don't understand why kernel-package looks deprecated, or what have you, but any information would be appreciated. If you want to use Debian's configuration and Debian's patches I'd look into downloading the source package and modifying it, but that can be a daunting task; there are a number of Debian-isms to learn along the way. Once you know what you are doing, you can update the debian/changelog, use a target in debian/rules to prepare a .orig.tar, and use dpkg-buildpkg to get a set of binary packages that are significantly similar to the ones from the kernel packaging team. If you just want a .deb to install, I've heard there's a makefile target in the kernel tarball that works fine. I believe but can't confirm that the .debs generated by the makefile in the kernel tarball will properly invoke the postint scripts that are used to update grub.cfg, menu.lst, or the lilo boot sector. The wiki has some pretty good information, too: http://wiki.debian.org/DebianKernel http://wiki.debian.org/HowToRebuildAnOfficialDebianKernelPackage http://wiki.debian.org/DebianKernelCustomCompilation All of those pages seem to reference: http://kernel-handbook.alioth.debian.org/ Specifically: http://kernel-handbook.alioth.debian.org/ch-common-tasks.html which I'm pretty sure is the official documentation produced by the Debian Linux Kernel packaging team. -- Boyd Stephen Smith Jr. ,= ,-_-. =. b...@iguanasuicide.net ((_/)o o(\_)) ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-' http://iguanasuicide.net/\_/ signature.asc Description: This is a digitally signed message part.
Re: Debian way of compiling a kernel.
On 2011-02-10 09:18 +0100, Boyd Stephen Smith Jr. wrote: If you just want a .deb to install, I've heard there's a makefile target in the kernel tarball that works fine. That target is called deb-pkg, i.e. you type make deb-pkg and get a nice Debian package. I believe but can't confirm that the .debs generated by the makefile in the kernel tarball will properly invoke the postint scripts that are used to update grub.cfg, menu.lst, or the lilo boot sector. The maintainer scripts in the generated package just run the hooks under /etc/kernel.d, in Squeeze and later those should take care of generating an initramfs, updating the bootloader's configuration etc. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87oc6kb89k@turtle.gmx.de
Re: Debian way of compiling a kernel.
Boyd Stephen Smith Jr.: If you just want a .deb to install, I've heard there's a makefile target in the kernel tarball that works fine. I believe but can't confirm that the .debs generated by the makefile in the kernel tarball will properly invoke the postint scripts that are used to update grub.cfg, menu.lst, or the lilo boot sector. I routinely compile vanilla kernels from git just using make, make deb-pkg and when installing the resulting package, grub2 and the initramfs get updated just fine. J. -- I wear a lot of leather but would never wear fur. [Agree] [Disagree] http://www.slowlydownward.com/NODATA/data_enter2.html signature.asc Description: Digital signature
Re: Debian way of compiling a kernel.
Ughn.. think google just discarded my post instead of sending. Don't want to retype; but here's the link: http://users.wowway.com/~zlinuxman/Kernel.htm -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/AANLkTikwHint7Z5tS82Pkk=+f3p-ec6dbeg0nkgpq...@mail.gmail.com
Debian way of compiling a kernel.
Ages ago, when amd64 wasn't part of the debian collection, I used to compile kernels myself using make-kpkg. This worked wonderfully, when I had to debug driver patches etc. (all is now of course stable). Since then I've forgotten this process, but this is not my problem. I wanted to test a fixed DSDT acpi table against a motherboard I am having flaky io issues with, I know the kernel is intended to handle these problems but I did want to see if it helped. When I hit the wiki page, I was surprised to see the procedure of building a debian kernel, nothing I have done before to handle building a slightly modified deb kernel. I ask. What is the real ... accepted ... and suggested method that I follow, I don't understand why kernel-package looks deprecated, or what have you, but any information would be appreciated. Thank you. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20110209223754.5fa03...@ws82.int.tlc
Re: Trouble compiling generic kernel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Monday 26 July 2010, Sven Joachim svenj...@gmx.de was heard to say: You need to install the fakeroot package to build Debian packages as an ordinary user. In Squeeze, dpkg-dev recommends fakeroot for that reason. Well, that solves that question. I thought the fakeroot package was for that extra bit of security, because I had it confused with chroot and that whole jail thing. Got it! I'll try again on my main system (the Lenny one) again today with all the advice given. I think it will work this time! And this may be the first time I didn't get advice from Ron Johnson. :^) Peace, may your aim never waver, Curt- - -- Those who torment us for our own good will torment us without end, for they do so with the approval of their consciences. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEVAwUBTE14zy9Y35yItIgBAQIMUQf8CYY9pwPWAclrgaICSGmRna/YrNXntV8z 19bcK4Ac/mtK2TXuvwgzRieRS8VpBN7aEmlm5qbn8PvPl6sB/e86gQfCRVISnTym 6yKJmtfp2QbugPMSyhPIJpsoZp4ot7lZd9GPj664gMtwFe5/+2TAfshRjNlahzxb Tz4C5ErAjN7DPtxXLp0hYDDdWi7PFl8a+juL9BgKoaToiiAUF7R15JVL1OnbzTd4 6k2sCbSs4Rndy2f6j6+/boDOvrYflg9iLn8bbWS4oY5CR3F+aunzfmvQRsKPXxlL lXsVAH8M9vSSj83L7n8GxNODOtq/P/4sIjpfl5hvFET38sDC8VTnXw== =KDgy -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201007260800.15737.howl...@priss.com
Re: Trouble compiling generic kernel
On 2010-07-26 14:00 +0200, Curt Howland wrote: On Monday 26 July 2010, Sven Joachim svenj...@gmx.de was heard to say: You need to install the fakeroot package to build Debian packages as an ordinary user. In Squeeze, dpkg-dev recommends fakeroot for that reason. Well, that solves that question. I thought the fakeroot package was for that extra bit of security, because I had it confused with chroot and that whole jail thing. Got it! I'll try again on my main system (the Lenny one) again today with all the advice given. I think it will work this time! And this may be the first time I didn't get advice from Ron Johnson. :^) There's another advice I need to give: while dpkg-dev will automatically use fakeroot when necessary, make-kpkg currently does not unless you set ROOT_CMD=fakeroot in the environment. You can put this setting into ~/.kernel-pkg.conf (I have it there for ages, and forgot about it). See kernel-pkg.conf(5) for other settings you may want to put into ~/.kernel-pkg.conf. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87mxtel1xm@turtle.gmx.de
Re: Trouble compiling generic kernel
On 07/26/2010 11:04 AM, Sven Joachim wrote: There's another advice I need to give: while dpkg-dev will automatically use fakeroot when necessary, make-kpkg currently does not unless you set ROOT_CMD=fakeroot in the environment. You can put this setting into ~/.kernel-pkg.conf (I have it there for ages, and forgot about it). One can also add a command line option, --rootcmd=fakeroot or one can simply run the whole 'make-kpkg' command as the argument to fakeroot: fakeroot make-kpkg [...] Dave W. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c4da6ee.8010...@sbcglobal.net
Re: Trouble compiling generic kernel
On 2010-07-25 05:22 +0200, Stan Hoeppner wrote: Curt Howland put forth on 7/24/2010 8:49 PM: Not compiling often (as you can tell), I note that the compile is using only one cpu of 4 at a time. I'm sure there is a parallel compilation tweak somewhere. Setting CONCURRENCY_LEVEL=4 in your user environment should fix this, at least, it used to. This only works if you use kernel-package to build the kernel. The generic way is to use make -j4. If you regularly do your kernel work while logged on as root, you'd need to append it to /root/.bashrc or /root/.profile There is no need to ever build the kernel as root. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87k4okni2i@turtle.gmx.de
Re: Trouble compiling generic kernel
On 2010-07-25 03:49 +0200, Curt Howland wrote: Not compiling often (as you can tell), I note that the compile is using only one cpu of 4 at a time. I'm sure there is a parallel compilation tweak somewhere. Yes, the -j switch for make. If you want to build a Debian package that you can install and test, use make -j4 deb-pkg. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87eiesnhvt@turtle.gmx.de
Re: Trouble compiling generic kernel
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Sunday 25 July 2010, Sven Joachim svenj...@gmx.de was heard to
say:
Setting CONCURRENCY_LEVEL=4 in your user environment should fix
this, at least, it used to.
This only works if you use kernel-package to build the kernel. The
generic way is to use make -j4.
Excellent advice to the novice.
There is no need to ever build the kernel as root.
Strangely enough, I wasn't. Not quite _that_ novice I guess. :^)
Now on to something new: Build failure! Who'd'a thought? I changed
back to single-concurrency just to make sure that it wasn't
related, make clean, then tried again with the same result.
Any suggestions?
==
$ make-kpkg --revision=curt0.1 kernel_image
{scrolled so many as I'm sure you know... Then,}
make[1]: Leaving directory `/usr/src/linux-2.6.35-rc6'
/usr/bin/make -j1 ARCH=i386 \
-C Documentation/lguest
make[1]: Entering directory
`/usr/src/linux-2.6.35-rc6/Documentation/lguest'
cc -m32 -Wall -Wmissing-declarations -Wmissing-prototypes -O3 -I../../include
-I../../arch/x86/include -U_FORTIFY_SOURCE
lguest.c -o lguest
lguest.c:21:25: error: sys/eventfd.h: No such file or directory
lguest.c: In function ‘create_thread’:
lguest.c:1026: warning: implicit declaration of function ‘eventfd’
make[1]: *** [lguest] Error 1
make[1]: Leaving directory
`/usr/src/linux-2.6.35-rc6/Documentation/lguest'
make: *** [debian/stamp/build/kernel] Error 2
==
- --
Those who torment us for our own good will torment us without end,
for they do so with the approval of their consciences.
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.9 (GNU/Linux)
iQEVAwUBTExlkC9Y35yItIgBAQIkZQf/XJqNfWBom12LWS6mBIM1aBjjoQaDBFlv
cTCZoMbPlnkVZiq8QokP3PL+ChPTmf63NV3grbUNKSfDMPADrNhhs0OsjBDbY5zd
DDoga+1wIUwTMwCiNH1McqFM682R36Z9SwqdJOCU1Cii8aBwQ7IchcJP8dGU9+3+
SUVnlprhsXTYsTNHxa8+0AiZPOdkSWCotii26Mo4ljeaIgCiTmCahD9nLiLt9OOw
VnkiE0OWHDybZ4JK6JDHYK67Gk6xIUSvndUYIhjYAgxuh9TCicaxn2LrWagwScCH
sMxYVtmGP0iHz41kP1daSpYiHbvMlpZHxdt00TGQnR3yWAoIeRBapA==
=4oN7
-END PGP SIGNATURE-
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/201007251225.52779.howl...@priss.com
Re: Trouble compiling generic kernel
On 07/25/2010 07:25 PM, Curt Howland wrote:
==
$ make-kpkg --revision=curt0.1 kernel_image
{scrolled so many as I'm sure you know... Then,}
make[1]: Leaving directory `/usr/src/linux-2.6.35-rc6'
/usr/bin/make -j1 ARCH=i386 \
-C Documentation/lguest
make[1]: Entering directory
`/usr/src/linux-2.6.35-rc6/Documentation/lguest'
cc -m32 -Wall -Wmissing-declarations -Wmissing-prototypes -O3 -I../../include
-I../../arch/x86/include -U_FORTIFY_SOURCE
lguest.c -o lguest
lguest.c:21:25: error: sys/eventfd.h: No such file or directory
lguest.c: In function ‘create_thread’:
lguest.c:1026: warning: implicit declaration of function ‘eventfd’
make[1]: *** [lguest] Error 1
make[1]: Leaving directory
`/usr/src/linux-2.6.35-rc6/Documentation/lguest'
make: *** [debian/stamp/build/kernel] Error 2
==
If you are certain that the kernel tree isn't missing some files, then this
is a situation for bugzilla.kernel.org . :)
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4c4c6c30.8030...@gmail.com
Re: Trouble compiling generic kernel
On 2010-07-25 18:54 +0200, Rares Aioanei wrote:
On 07/25/2010 07:25 PM, Curt Howland wrote:
==
$ make-kpkg --revision=curt0.1 kernel_image
{scrolled so many as I'm sure you know... Then,}
make[1]: Leaving directory `/usr/src/linux-2.6.35-rc6'
/usr/bin/make -j1 ARCH=i386 \
-C Documentation/lguest
make[1]: Entering directory
`/usr/src/linux-2.6.35-rc6/Documentation/lguest'
cc -m32 -Wall -Wmissing-declarations -Wmissing-prototypes -O3
-I../../include -I../../arch/x86/include -U_FORTIFY_SOURCE
lguest.c -o lguest
lguest.c:21:25: error: sys/eventfd.h: No such file or directory
lguest.c: In function ‘create_thread’:
lguest.c:1026: warning: implicit declaration of function ‘eventfd’
make[1]: *** [lguest] Error 1
make[1]: Leaving directory
`/usr/src/linux-2.6.35-rc6/Documentation/lguest'
make: *** [debian/stamp/build/kernel] Error 2
==
If you are certain that the kernel tree isn't missing some files, then this
is a situation for bugzilla.kernel.org . :)
Not really. The libc6-dev package is too old, lacking sys/eventfd.
Sven
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/8739v7mq11@turtle.gmx.de
Re: Trouble compiling generic kernel
On 2010-07-25 18:25 +0200, Curt Howland wrote:
Now on to something new: Build failure! Who'd'a thought? I changed
back to single-concurrency just to make sure that it wasn't
related, make clean, then tried again with the same result.
Any suggestions?
==
$ make-kpkg --revision=curt0.1 kernel_image
When using kernel-package, make sure that you install the latest version
(12.036). Earlier versions will likely not work, in particular the
Lenny version is known to be broken with kernels = 2.6.33.
{scrolled so many as I'm sure you know... Then,}
make[1]: Leaving directory `/usr/src/linux-2.6.35-rc6'
/usr/bin/make -j1 ARCH=i386 \
-C Documentation/lguest
make[1]: Entering directory
`/usr/src/linux-2.6.35-rc6/Documentation/lguest'
cc -m32 -Wall -Wmissing-declarations -Wmissing-prototypes -O3 -I../../include
-I../../arch/x86/include -U_FORTIFY_SOURCE
lguest.c -o lguest
lguest.c:21:25: error: sys/eventfd.h: No such file or directory
Oh yeah, sys/eventfd.h is in the libc6-dev package. But not in Lenny,
only in Squeeze. Since you probably don't want to upgrade libc6, you
should build the kernel without CONFIG_LGUEST: under Virtualization,
uncheck Linux hypervisor example code (NEW). This should hopefully do
the trick.
Sven
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/87tynnlb4u@turtle.gmx.de
Re: Trouble compiling generic kernel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Sunday 25 July 2010, Rares Aioanei was heard to say: If you are certain that the kernel tree isn't missing some files, then this is a situation for bugzilla.kernel.org . :) As certain as I can be, having been careful not to delete anything. Just make menuconfig then make-kpkg... Golly, I feel so ... awed! Li'l ol' me? Post to kernel.org? Gee wilikers. - -- Those who torment us for our own good will torment us without end, for they do so with the approval of their consciences. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEVAwUBTEx0Di9Y35yItIgBAQJ4Bgf/U1yR0EgGdj304Z1VTKspR36Mz2c4H7TQ dNVOFydw6RY+vf0i8HxHZBNX3N5r49BRp06lt+Huk9Ed9wwuoN5bBE3lSeLdFbdf 9s2wjRfix8zt0Z/FUCIq61pNRznGnnxZX5K0nfzQdcRrDspBGYLnAQi1H3bdxSLS xupJsUNdjIuj9ryh4kb5DmHaCrGN7+mwWVd9dC7Rc6CfT+NxgfJCUWgcREdRBhgf BPhc4wnIUlEdmCFKMrnmV8F8wuLSUbEG5Efd4HKC5pUWw4VCvQwF/SlIlZbqjaJm dizD6k9Lw2qT2YLHsu6ztyAnF3Km0MZQmgl7cCRGy++0q48UJLkyrg== =+K/1 -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201007251327.42289.howl...@priss.com
Re: Trouble compiling generic kernel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Sunday 25 July 2010, Rares Aioanei was heard to say: If you are certain that the kernel tree isn't missing some files, then this is a situation for bugzilla.kernel.org . :) Well, no, it must be just me. I pulled down and tried the 2.6.34.1 mainline kernel and received an almost identical error, but the same sys/eventfd.h line: === make[1]: Leaving directory `/usr/src/linux-2.6.34.1' /usr/bin/make -j4 ARCH=i386 \ -C Documentation/lguest make[1]: Entering directory `/usr/src/linux-2.6.34.1/Documentation/lguest' cc -m32 -Wall -Wmissing-declarations -Wmissing-prototypes -O3 -I../../include -I../../arch/x86/include -U_FORTIFY_SOURCE lguest.c -o lguest lguest.c:21:25: error: sys/eventfd.h: No such file or directory lguest.c: In function ‘create_thread’: lguest.c:1026: warning: implicit declaration of function ‘eventfd’ make[1]: *** [lguest] Error 1 make[1]: Leaving directory `/usr/src/linux-2.6.34.1/Documentation/lguest' make: *** [debian/stamp/build/kernel] Error 2 === I wonder if the problem is from using make-kpkg somehow - -- Those who torment us for our own good will torment us without end, for they do so with the approval of their consciences. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEVAwUBTEx6CS9Y35yItIgBAQJifgf/Xuz7T5zLuERyrhbH+qvi0jBpcU7AyOli HjvIKXnZqFIb3BEhFxVIWR3A1wXwmaUXa35sHS2QwSkxH4LJ5JbHU9HIEMSs0hoS jKpGXWK84YnYyE6HZk99gHn9tmnCUirKneS8gzx9g4+hHD32Xtiq4bFzN5yOj+4+ CGfySDGhhWhB+9s8L7Sk0U7QrQ/USlVBH3XvZBBTYBbrP2FhE4FZURHj11OWsfGi B9qpNq8s8L7+5hTjyPBKU5oMc61XMMKwvirkyPtiqSMXRvPaKr2nuzwTx+2f8Une e6WMV5HRJRqMLHgu6xe361U31BuNh8aRCy5LvRLe4dD9lCpKH5R28w== =VqCE -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201007251353.13713.howl...@priss.com
Re: Trouble compiling generic kernel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Sunday 25 July 2010, Sven Joachim svenj...@gmx.de was heard to say: Not really. The libc6-dev package is too old, lacking sys/eventfd. Got it. I will give up with this machine. I don't want to upgrade to Squeeze due to preferring KDE3. Curt- - -- Those who torment us for our own good will torment us without end, for they do so with the approval of their consciences. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEVAwUBTEyCwC9Y35yItIgBAQJMxgf+Js5gSXNl/05x09wt7nW4oC4790plhiBq Z8UlJoeonJPHeyDxHwqRTAYwdjNwBRSF/oixZ1wYLwFh73ytB2bgnxRchtDIHeDr i+JL78DVoj6buQPHGAVfDC74t/WjezasL1R2EiHW0xyQzYBYkmGcsQ72zvEhdULj mzFR0kX6TZhrsVvaQ1Pk/cjFi68ekVbCD7813yHHS4dBplR43gtYKco8QhuFXkvs uCbyhO0gIDxPS3z4Orkoi0cx0QrnJrCeZG2IkpVhEUIe8Raa8UEJUTVODfk1byw9 mKoZCL+G9Ao7ZMs1jIzNyUiEr7v9LBVbMPcPkhuACrDXypaRdMCODw== =wnUa -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201007251430.29846.howl...@priss.com
Re: Trouble compiling generic kernel
Well, I've run into another problem with the compile that folks might find interesting. All while running as the only user account on this machine (other than root), I ran into a very interesting permissions error at the end of the make-kpkg kernel_image process. I did the compile as root successfully after seeing this error the first time, just to see if it was user related, and the compile completed successfully. In fact, I'm running 2.6.35-rc6 right now, no problems. But I erased the directory and un-tarred the generic kernel source again just to be sure, and received the same permission error again. I'm just not sure if this is a debian make-kpkg error or not... Here's the error and ls -a of the directories I think its referring to: = echo done debian/stamp/build/kernel /usr/bin/make -f ./debian/rules debian/stamp/binary/pre-linux-image-2.6.35-rc6-curt.02 make[1]: Entering directory `/usr/src/linux-2.6.35-rc6' == making target debian/stamp/install/linux-image-2.6.35-rc6-curt.02 [new prereqs: ]== This is kernel package version 12.036. rm -f -r .//usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02 .//usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02.deb install -p -d -o root -g root -m 755 /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postinst.d /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/preinst.d \ /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postrm.d /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/prerm.d install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postinst.d': Operation not permitted install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/preinst.d': Operation not permitted install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postrm.d': Operation not permitted install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/prerm.d': Operation not permitted make[1]: *** [debian/stamp/install/linux-image-2.6.35-rc6-curt.02] Error 1 make[1]: Leaving directory `/usr/src/linux-2.6.35-rc6' make: *** [kernel_image] Error 2 $ dir /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/ total 16 drwx-- 2 curt curt 4096 Jul 25 21:12 postinst.d drwx-- 2 curt curt 4096 Jul 25 21:12 postrm.d drwx-- 2 curt curt 4096 Jul 25 21:12 preinst.d drwx-- 2 curt curt 4096 Jul 25 21:12 prerm.d $ dir /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/ total 4 drwxr-xr-x 6 curt curt 4096 Jul 25 21:12 kernel = -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktikwv7mrwgotrllvhbvhywhgfoqqr3bwel7ka...@mail.gmail.com
Re: Trouble compiling generic kernel
On Mon, Jul 26, 2010 at 2:32 AM, Curt Howland howl...@priss.com wrote: Well, I've run into another problem with the compile that folks might find interesting. All while running as the only user account on this machine (other than root), I ran into a very interesting permissions error at the end of the make-kpkg kernel_image process. I did the compile as root successfully after seeing this error the first time, just to see if it was user related, and the compile completed successfully. In fact, I'm running 2.6.35-rc6 right now, no problems. But I erased the directory and un-tarred the generic kernel source again just to be sure, and received the same permission error again. I'm just not sure if this is a debian make-kpkg error or not... Here's the error and ls -a of the directories I think its referring to: = echo done debian/stamp/build/kernel /usr/bin/make -f ./debian/rules debian/stamp/binary/pre-linux-image-2.6.35-rc6-curt.02 make[1]: Entering directory `/usr/src/linux-2.6.35-rc6' == making target debian/stamp/install/linux-image-2.6.35-rc6-curt.02 [new prereqs: ]== This is kernel package version 12.036. rm -f -r .//usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02 .//usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02.deb install -p -d -o root -g root -m 755 /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postinst.d /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/preinst.d \ /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postrm.d /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/prerm.d install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postinst.d': Operation not permitted install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/preinst.d': Operation not permitted install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postrm.d': Operation not permitted install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/prerm.d': Operation not permitted make[1]: *** [debian/stamp/install/linux-image-2.6.35-rc6-curt.02] Error 1 make[1]: Leaving directory `/usr/src/linux-2.6.35-rc6' make: *** [kernel_image] Error 2 $ dir /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/ total 16 drwx-- 2 curt curt 4096 Jul 25 21:12 postinst.d drwx-- 2 curt curt 4096 Jul 25 21:12 postrm.d drwx-- 2 curt curt 4096 Jul 25 21:12 preinst.d drwx-- 2 curt curt 4096 Jul 25 21:12 prerm.d $ dir /usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/ total 4 drwxr-xr-x 6 curt curt 4096 Jul 25 21:12 kernel = I had the same error when i was compiling the 2.6.34.1, both inside and outside the /usr/src/ dir, and again worked fine when i did it as root. I don’t think a normal user can changer owner and permissions can they? Regards, Angus ps thinking about it, you would have had to untar or cp the source as root in order to get it into the /usr/src folder right, if so wouldn’t it end up belonging to root? (im still a bit new to how *nix permissions work, so pardon my ignorance) -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/aanlktiknzvhyqiixkhzs-2gxgwytiuudrhsxqxdc6...@mail.gmail.com
Re: Trouble compiling generic kernel
On 2010-07-26 03:32 +0200, Curt Howland wrote: Well, I've run into another problem with the compile that folks might find interesting. All while running as the only user account on this machine (other than root), I ran into a very interesting permissions error at the end of the make-kpkg kernel_image process. I did the compile as root successfully after seeing this error the first time, just to see if it was user related, and the compile completed successfully. In fact, I'm running 2.6.35-rc6 right now, no problems. But I erased the directory and un-tarred the generic kernel source again just to be sure, and received the same permission error again. I'm just not sure if this is a debian make-kpkg error or not... You need to install the fakeroot package to build Debian packages as an ordinary user. In Squeeze, dpkg-dev recommends fakeroot for that reason. install: cannot change owner and permissions of `/usr/src/linux-2.6.35-rc6/debian/linux-image-2.6.35-rc6-curt.02/etc/kernel/postinst.d': Operation not permitted What fakeroot does is to install a wrapper around chown and other system calls to pretend they succeeded. It also starts a daemon that remembers the new owner, so that the files are actually owned by that user (usually root) when the package is created. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/871vaqn8sl@turtle.gmx.de
Trouble compiling generic kernel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi. Lenny, up to date with security and backports. I pulled down 2.6.35-rc6.tar.bz2 from kernel.org, and would like to try compiling it to see what's Coming Soon To A Disto Near Me. Its been about 9 years since the last time I compiled a kernel, but I'm having trouble I didn't have then. For some reason, none of the make things that I've used in the past are working now: === $ make xconfig CHECK qt * * Unable to find the QT3 installation. Please make sure that * the QT3 development package is correctly installed and * either install pkg-config or set the QTDIR environment * variable to the correct location. * make[1]: *** No rule to make target `scripts/kconfig/.tmp_qtcheck', needed by `scripts/kconfig/qconf.o'. Stop. make: *** [xconfig] Error 2 === === $ make menuconfig *** Unable to find the ncurses libraries or the *** required header files. *** 'make menuconfig' requires the ncurses libraries. *** *** Install ncurses (ncurses-devel) and try again. *** make[1]: *** [scripts/kconfig/dochecklxdialog] Error 1 make: *** [menuconfig] Error 2 === === $ make gconfig * * Unable to find the GTK+ installation. Please make sure that * the GTK+ 2.0 development package is correctly installed... * You need gtk+-2.0, glib-2.0 and libglade-2.0. * make[1]: *** No rule to make target `scripts/kconfig/.tmp_gtkcheck', needed by `scripts/kconfig/gconf.o'. Stop. make: *** [gconfig] Error 2 === make config did work, but working through everything in text is enough to make a sane person INSANE. I was surprised with the failure of menuconfig, since I have all of ncurses installed and the ncurses-devel suggestion isn't a Debian package name anyway. Suggestions? - -- Those who torment us for our own good will torment us without end, for they do so with the approval of their consciences. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEVAwUBTEtMeS9Y35yItIgBAQKOfAf/UDR2ITg8Idcu/2Z3yxRalWOkJ51DeQtl XJ+7SxYj11LvSQAWdltbRAABK8BeFq9ylzdFxT+ncxUvwCnZT+iqBxMDquzj3smA vyzAe7xcAm9th4Sqw+krr5nWzovj6bPDmkhNhIknvGVEcFAmaPgihL6GEgAjcAbA Tvxu8OqPnS5tnGJcBNpMtV2htUmSju8QuGI4e8IQkygRX/46dgQbI0tnU3O8+1YK 3X59amdukZxgl6lsVh3C/kGq5yL+YhKfNRRtf52VtN0UTe8x01jMoJrrIkI8IjVf gGPqFPjD3IDwdvgiD273MVvIW56dbIMyGN58W6FAZsg9IdZK4eXIvA== =y66Y -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201007241626.33562.howl...@priss.com
Re: Trouble compiling generic kernel
On 2010-07-24 22:26 +0200, Curt Howland wrote: Hi. Lenny, up to date with security and backports. I pulled down 2.6.35-rc6.tar.bz2 from kernel.org, and would like to try compiling it to see what's Coming Soon To A Disto Near Me. Its been about 9 years since the last time I compiled a kernel, but I'm having trouble I didn't have then. For some reason, none of the make things that I've used in the past are working now: This is due to missing development packages. === $ make xconfig CHECK qt * * Unable to find the QT3 installation. Please make sure that * the QT3 development package is correctly installed and * either install pkg-config or set the QTDIR environment * variable to the correct location. * make[1]: *** No rule to make target `scripts/kconfig/.tmp_qtcheck', needed by `scripts/kconfig/qconf.o'. Stop. make: *** [xconfig] Error 2 === This target needs the libqt3-mt-dev package (and possibly others). === $ make menuconfig *** Unable to find the ncurses libraries or the *** required header files. *** 'make menuconfig' requires the ncurses libraries. *** *** Install ncurses (ncurses-devel) and try again. *** make[1]: *** [scripts/kconfig/dochecklxdialog] Error 1 make: *** [menuconfig] Error 2 === This target needs libncurses5-dev. === $ make gconfig * * Unable to find the GTK+ installation. Please make sure that * the GTK+ 2.0 development package is correctly installed... * You need gtk+-2.0, glib-2.0 and libglade-2.0. * make[1]: *** No rule to make target `scripts/kconfig/.tmp_gtkcheck', needed by `scripts/kconfig/gconf.o'. Stop. make: *** [gconfig] Error 2 === This target needs libglade2-dev (and possibly other packages). make config did work, but working through everything in text is enough to make a sane person INSANE. Yes, this is only for people who are or want to become crazy. I was surprised with the failure of menuconfig, since I have all of ncurses installed and the ncurses-devel suggestion isn't a Debian package name anyway. The -devel namespace comes from Red Hat based distributions, in Debian development packages always end in -dev and very often begin with lib. Suggestions? Install libncurses5-dev and use make menuconfig or make nconfig (the latter is new in 2.6.35). Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87bp9wegr3@turtle.gmx.de
Re: Trouble compiling generic kernel
Curt writes: I was surprised with the failure of menuconfig, since I have all of ncurses installed and the ncurses-devel suggestion isn't a Debian package name anyway. Suggestions? libncurses5-dev -- John Hasler -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87pqyctx33@thumper.dhh.gt.org
Re: Trouble compiling generic kernel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On Saturday 24 July 2010, Sven Joachim svenj...@gmx.de was heard to say: Suggestions? Install libncurses5-dev and use make menuconfig or make nconfig (the latter is new in 2.6.35). Got it, and it works fine. Thank you. Not compiling often (as you can tell), I note that the compile is using only one cpu of 4 at a time. I'm sure there is a parallel compilation tweak somewhere. If it is easy, fine. If not, then forget about it, I'm not in any rush. Curt- - -- Those who torment us for our own good will torment us without end, for they do so with the approval of their consciences. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) iQEVAwUBTEuYNC9Y35yItIgBAQII7wf+P5vId/fhyN1YSvVjQF6x0S4fHBCfvHCf qnsgceZuuTjnRZAhy+pFmJ20avvf9bLcD88RPgBYrcSCHB2pYOXN03JbIS26VXyj Y4YMz6XXZIFuE+GQaINEKd+Yfs5gXxTNq9HlL34TQx0foY0Pvu9MCS/DJPsZOjM4 BtIIifDA0Rpb0dCBiwGT7X3OSxTmPbRG9R7VHOWH2L3TMA+WfAsgV1wXjMf9dMDA peZvCimNDMEDVRDMEDf5fXU+8UqytuZYEhbZRCEXtmbwTEv3Fbw8sSIad0zuV6wV 6FU3i1KWRRfJYMOo2s3SD9W1T4LYkWe1KYVmxyL/dxMFnN+lXUe9gg== =Qc2b -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/201007242149.47058.howl...@priss.com
Re: Trouble compiling generic kernel
Curt Howland put forth on 7/24/2010 8:49 PM: Not compiling often (as you can tell), I note that the compile is using only one cpu of 4 at a time. I'm sure there is a parallel compilation tweak somewhere. Setting CONCURRENCY_LEVEL=4 in your user environment should fix this, at least, it used to. From the bash command line, for the current session only: export CONCURRENCY_LEVEL=4 To make it permanent you need to edit /home/[user]/.bash_profile and append: export CONCURRENCY_LEVEL=4 If you regularly do your kernel work while logged on as root, you'd need to append it to /root/.bashrc or /root/.profile -- Stan -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4c4badf5.6030...@hardwarefreak.com
RE: Trouble compiling generic kernel
On Sat, 24 Jul 2010 22:22:29 -0500 s...@hardwarefreak.com wrote: Setting CONCURRENCY_LEVEL=4 in your user environment should fix this, at least, it used to. From the bash command line, for the current session only: export CONCURRENCY_LEVEL=4 To make it permanent you need to edit /home/[user]/.bash_profile and append: export CONCURRENCY_LEVEL=4 If you regularly do your kernel work while logged on as root, you'd need to append it to /root/.bashrc or /root/.profile You might also want to read http://www.wowway.com/~zlinuxman/Kernel.htm or from google's cache at http://webcache.googleusercontent.com/search?q=cache:7D7TfZqA1wMJ:www.wowway.com/~zlinuxman/Kernel.htm+http://www.wowway.com/~zlinuxman/Kernel.htmcd=4hl=enct=clnkgl=ca -M _ Game on: Challenge friends to great games on Messenger http://go.microsoft.com/?linkid=9734387
Re: Trouble compiling generic kernel
On Sat, 24 Jul 2010 22:22:29 -0500 Stan Hoeppner s...@hardwarefreak.com wrote: Curt Howland put forth on 7/24/2010 8:49 PM: Not compiling often (as you can tell), I note that the compile is using only one cpu of 4 at a time. I'm sure there is a parallel compilation tweak somewhere. Setting CONCURRENCY_LEVEL=4 in your user environment should fix this, at least, it used to. From the bash command line, for the current session only: export CONCURRENCY_LEVEL=4 To make it permanent you need to edit /home/[user]/.bash_profile and append: export CONCURRENCY_LEVEL=4 If you regularly do your kernel work while logged on as root, you'd need to append it to /root/.bashrc or /root/.profile And assuming you're using kernel-package, from its manpage: --jobs number -j number Set the environment variable CONCURRENCY_LEVEL to number. Celejar -- foffl.sourceforge.net - Feeds OFFLine, an offline RSS/Atom aggregator mailmin.sourceforge.net - remote access via secure (OpenPGP) email ssuds.sourceforge.net - A Simple Sudoku Solver and Generator -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20100725001136.f19fdb7f.cele...@gmail.com
Re: Compiling Linux kernel
On 2009-12-25 07:55 +0100, Daniel Bareiro wrote: Sorry to send this message twice, but I thought that for some reason it had not arrived at the list. Although it seems that both messages arrived with a delay of six hours. This can be due to some moderation of the list? This list is not moderated, and your messages have a very low spamassassin score, so they should have arrived immediately. You can look in the Received: headers to see where they spent the time. And why in this case yes we must use these options if running an amd64 kernel in userland 32 is not necessary to use ARCH=x86_64 in the invocations of make when compiling of the traditional way? Well, thinking a little more about this subject, the cause by which this happens perhaps is that when make-kpkg consults the general architecture of the system, it obtains in (b.1) as in (b.2) that is i386. For that reason in both cases it is necessary to use --cross-compile and --arch. Nevertheless when being used the compilation of the traditional way, this becomes by outside any own control of Debian and the architecture that will be used by default is the one of running kernel. Is correct this asseveration? Yes, kernel-package relies on the information dpkg-architecture provides whereas the kernel Makefile trusts uname -m to determine the architecture. On systems with a 64-bit kernel and 32-bit userland this gives different information: , | % dpkg-architecture -qDEB_BUILD_ARCH | i386 | % uname -m | x86_64 ` Merry Christmas, Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Compiling Linux kernel
Hi, Sven. On Friday, 25 December 2009 09:53:53 +0100, Sven Joachim wrote: Sorry to send this message twice, but I thought that for some reason it had not arrived at the list. Although it seems that both messages arrived with a delay of six hours. This can be due to some moderation of the list? This list is not moderated, and your messages have a very low spamassassin score, so they should have arrived immediately. You can look in the Received: headers to see where they spent the time. It seems that the problem was to the return: Received: from liszt.debian.org (EHLO liszt.debian.org) [82.195.75.100] by mx0.gmx.net (mx052) with SMTP; 24 Dec 2009 20:10:35 +0100 -- Received: from localhost (localhost [127.0.0.1]) by liszt.debian.org (Postfix) with QMQP id 5547A2D0D8E; Thu, 24 Dec 2009 12:47:49 + (UTC) -- Received: from localhost (localhost [127.0.0.1]) by liszt.debian.org (Postfix) with ESMTP id CF5AD2D0D5C for lists-debian-u...@liszt.debian.org; Thu, 24 Dec 2009 12:47:42 + (UTC) Received: from liszt.debian.org ([127.0.0.1]) by localhost (lists.debian.org [127.0.0.1]) (amavisd-new, port 2525) with ESMTP id 7OfJDoFZMriA for lists-debian-u...@liszt.debian.org; Thu, 24 Dec 2009 12:47:35 + (UTC) Now I'm seeing that the email crosses liszt.debian.org several times, which draws attention to me. And why in this case yes we must use these options if running an amd64 kernel in userland 32 is not necessary to use ARCH=x86_64 in the invocations of make when compiling of the traditional way? Well, thinking a little more about this subject, the cause by which this happens perhaps is that when make-kpkg consults the general architecture of the system, it obtains in (b.1) as in (b.2) that is i386. For that reason in both cases it is necessary to use --cross-compile and --arch. Nevertheless when being used the compilation of the traditional way, this becomes by outside any own control of Debian and the architecture that will be used by default is the one of running kernel. Is correct this asseveration? Yes, kernel-package relies on the information dpkg-architecture provides whereas the kernel Makefile trusts uname -m to determine the architecture. On systems with a 64-bit kernel and 32-bit userland this gives different information: , | % dpkg-architecture -qDEB_BUILD_ARCH | i386 | % uname -m | x86_64 ` Thanks to all by the contributions. These were very useful. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On Wednesday, 23 December 2009 11:19:26 +0100, Sven Joachim wrote: b) Debian way compilation: b.1) Having booted an i386 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # fakeroot make-kpkg clean --cross-compile - -arch amd64 # fakeroot make-kpkg --initrd --cross-compile - -arch amd64 \ --append-to-version=-dgb kernel_image kernel_headers It does not present questions during make and it boots ok enabling the options mentioned above. Another thing that I have noticed is that the packages are generated with a suffix i386, although after boot the kernel is x86_64: linux-headers-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb linux-image-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb Yes, because you are building on i386 (as in dpkg --print-architecture). This is what you want, BTW. Ah, ok. The suffix doesn't depend on the specific architecture of kernel but of all the distribution that one is using. For that reason with Ubuntu amd64 yes the packages are generated with suffix amd64 :-) b.2) Having booted an amd64 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make menuconfig # fakeroot make-kpkg clean # fakeroot make-kpkg --initrd --append-to-version=-dgb kernel_image \ kernel_headers You still need to specify --cross-compile and --arch here. And why in this case yes we must use these options if running an amd64 kernel in userland 32 is not necessary to use ARCH=x86_64 in the invocations of make when compiling of the traditional way? Thanks for your reply and merry Christmas. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On 2009-12-17 at 21:32:59 -300, Daniel Bareiro wrote: Hi all! I am trying to compile Linux 2.6.32 with the source code of kernel.org. Kernel that I'm using at the moment is 2.6.26-2-686 of the Debian GNU/Linux repositories. In order to generate the configuration, I've copied the file corresponding to this kernel to the directory of sources, /usr/src/linux, and I've executed make menuconfig: As some others have already pointed out, I believe your basic problem is trying to use a config file from an old release with a new release. That is not recommended. But on the general subject of kernel building in Debian, you might want to check out the following web site: http://www.wowway.com/~zlinuxman/Kernel.htm -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Compiling Linux kernel
Hi, Sven. On Wednesday, 23 December 2009 11:19:26 +0100, Sven Joachim wrote: b) Debian way compilation: b.1) Having booted an i386 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # fakeroot make-kpkg clean --cross-compile - -arch amd64 # fakeroot make-kpkg --initrd --cross-compile - -arch amd64 \ --append-to-version=-dgb kernel_image kernel_headers It does not present questions during make and it boots ok enabling the options mentioned above. Another thing that I have noticed is that the packages are generated with a suffix i386, although after boot the kernel is x86_64: linux-headers-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb linux-image-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb Yes, because you are building on i386 (as in dpkg --print-architecture). This is what you want, BTW. Ah, ok. The suffix doesn't depend on the specific architecture of kernel but of all the distribution that one is using. For that reason with Ubuntu amd64 yes the packages are generated with suffix amd64 :-) b.2) Having booted an amd64 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make menuconfig # fakeroot make-kpkg clean # fakeroot make-kpkg --initrd --append-to-version=-dgb kernel_image \ kernel_headers You still need to specify --cross-compile and --arch here. And why in this case yes we must use these options if running an amd64 kernel in userland 32 is not necessary to use ARCH=x86_64 in the invocations of make when compiling of the traditional way? Thanks for your reply and merry Christmas. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
Sorry to send this message twice, but I thought that for some reason it had not arrived at the list. Although it seems that both messages arrived with a delay of six hours. This can be due to some moderation of the list? On Thursday, 24 December 2009 09:45:54 -0300, Daniel Bareiro wrote: b.2) Having booted an amd64 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make menuconfig # fakeroot make-kpkg clean # fakeroot make-kpkg --initrd --append-to-version=-dgb kernel_image \ kernel_headers You still need to specify --cross-compile and --arch here. And why in this case yes we must use these options if running an amd64 kernel in userland 32 is not necessary to use ARCH=x86_64 in the invocations of make when compiling of the traditional way? Well, thinking a little more about this subject, the cause by which this happens perhaps is that when make-kpkg consults the general architecture of the system, it obtains in (b.1) as in (b.2) that is i386. For that reason in both cases it is necessary to use --cross-compile and --arch. Nevertheless when being used the compilation of the traditional way, this becomes by outside any own control of Debian and the architecture that will be used by default is the one of running kernel. Is correct this asseveration? Merry Christmas. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On Saturday, 19 December 2009 16:19:31 -0300, Daniel Bareiro wrote: I assume that it must have differences between both kernels versions; for that reason, as I've mentioned in another mail of this thread, after to have copied the file, I followed a similar procedure to which mentioned above, but with the ARCH=x86_64 variable: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make But in spite of to have used 'make menuconfig', 'make' did 'restart config' beginning to do questions to me. If you are not running a 64-bit kernel, you have to specify ARCH=x86_64 for _all_ make invocations. Well. I didn't know that there was to do it with all. Using the following steps I was not interrogated: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make ARCH=x86_64 # make ARCH=x86_64 modules_install # make ARCH=x86_64 install But when boot, the process is interrupted with the following message: request_module runaway loop modprobe binfmt-464c It draws attention to me that this has not happened after compiling after to have booted with amd64 kernel using the same environment (KVM VM with host amd64). The unique difference in the process was the aggregate of the ARCH=x86_64 variable with each invocation of make. Reading [1] and [2], I already found the cause of this problem. The configuration in Executable file formats / Emulations must be the following one in order to use a kernel x86_64 in userland 32. [*] Kernel support for ELF binaries [ ] Write ELF core dumps with partial segments M Kernel support for MISC binaries [*] IA32 Emulation * IA32 a.out support I didn't have this problem after to boot with kernel amd64 because to configure the new kernel I was based on the configuration of running amd64 kernel, in which these options already were setted of this way. Then, making an analysis of the tests that I was doing, I reached these results: a) Traditional compilation: a.1) Having booted an i386 kernel and userland 32: # make ARCH=x86_64 menuconfig # make ARCH=x86_64 # make ARCH=x86_64 modules_install # make ARCH=x86_64 install It does not present questions during make and it boots ok enabling the options mentioned above. a.2) Having booted an amd64 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make # make modules_install # make install It does not present questions during make and it boots ok. b) Debian way compilation: b.1) Having booted an i386 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # fakeroot make-kpkg clean --cross-compile - -arch amd64 # fakeroot make-kpkg --initrd --cross-compile - -arch amd64 \ --append-to-version=-dgb kernel_image kernel_headers It does not present questions during make and it boots ok enabling the options mentioned above. Another thing that I have noticed is that the packages are generated with a suffix i386, although after boot the kernel is x86_64: linux-headers-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb linux-image-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb b.2) Having booted an amd64 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make menuconfig # fakeroot make-kpkg clean # fakeroot make-kpkg --initrd --append-to-version=-dgb kernel_image \ kernel_headers After to execute this last line, the process starts but I begin to be interrogated again by configuration options. If in this environment (having booted an amd64 kernel an userland 32) I use the syntax mentioned in (b.1), I don't have this problem. It would seem to be that for some reason fakeroot make-kpkg finds a complete environment of 32 bits. Still using the syntax of (b.1) in this environment, I obtain the packages with a suffix i386 again. linux-headers-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb linux-image-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb Regards, Daniel [1] http://www.informaticien.be/index.ks?page=forum_topicid=5388index=0 [2] http://saalwaechter-notes.blogspot.com/2008/10/requestmodule-runaway-loop-modprobe.html -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On 2009-12-23 11:07 +0100, Daniel Bareiro wrote: Reading [1] and [2], I already found the cause of this problem. The configuration in Executable file formats / Emulations must be the following one in order to use a kernel x86_64 in userland 32. [*] Kernel support for ELF binaries [ ] Write ELF core dumps with partial segments M Kernel support for MISC binaries [*] IA32 Emulation Now that you mention it, I recall that this isn't set up automatically when you come from a 32-bit configuration. b) Debian way compilation: b.1) Having booted an i386 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # fakeroot make-kpkg clean --cross-compile - -arch amd64 # fakeroot make-kpkg --initrd --cross-compile - -arch amd64 \ --append-to-version=-dgb kernel_image kernel_headers It does not present questions during make and it boots ok enabling the options mentioned above. Another thing that I have noticed is that the packages are generated with a suffix i386, although after boot the kernel is x86_64: linux-headers-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb linux-image-2.6.32-dgb_2.6.32-dgb-10.00.Custom_i386.deb Yes, because you are building on i386 (as in dpkg --print-architecture). This is what you want, BTW. b.2) Having booted an amd64 kernel and userland 32: # cp /boot/config-`uname -r` ./.config # make menuconfig # fakeroot make-kpkg clean # fakeroot make-kpkg --initrd --append-to-version=-dgb kernel_image \ kernel_headers You still need to specify --cross-compile and --arch here. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Compiling Linux kernel
Hi, Sven. On Friday, 18 December 2009 17:34:22 +0100, Sven Joachim wrote: I was trying installing and booting 2.6.26-2-amd64 kernel and then compiling 2.6.32 kernel of the traditional way: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make menuconfig # make In this case I didn't use the ARCH=x86_64 variable and after to execute 'make menuselect' to select the configuration parameters, when invoking 'make', this worked without doing any type of question. Could this be due to I booted previously with amd64 kernel? Not really, make menuconfig selects the defaults for options which did not exist in previous kernels. Note that it will use the architecture of the running kernel if you do not specify ARCH. Perfect. I assume that it must have differences between both kernels versions; for that reason, as I've mentioned in another mail of this thread, after to have copied the file, I followed a similar procedure to which mentioned above, but with the ARCH=x86_64 variable: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make But in spite of to have used 'make menuconfig', 'make' did 'restart config' beginning to do questions to me. If you are not running a 64-bit kernel, you have to specify ARCH=x86_64 for _all_ make invocations. Well. I didn't know that there was to do it with all. Using the following steps I was not interrogated: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make ARCH=x86_64 # make ARCH=x86_64 modules_install # make ARCH=x86_64 install But when boot, the process is interrupted with the following message: request_module runaway loop modprobe binfmt-464c It draws attention to me that this has not happened after compiling after to have booted with amd64 kernel using the same environment (KVM VM with host amd64). The unique difference in the process was the aggregate of the ARCH=x86_64 variable with each invocation of make. Thanks for your reply. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On 2009-12-18 01:32 +0100, Daniel Bareiro wrote: I am trying to compile Linux 2.6.32 with the source code of kernel.org. Kernel that I'm using at the moment is 2.6.26-2-686 of the Debian GNU/Linux repositories. In order to generate the configuration, I've copied the file corresponding to this kernel to the directory of sources, /usr/src/linux, and I've executed make menuconfig: - # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig - It would be better to start with the configuration of Debian's -amd64 images if you want to use a 64-bit kernel, or even install and boot the 2.6.26-2-amd64 kernel. But when I execute 'make', restart config becomes and I'm interrogated on some details of the configuration: - antares:/usr/src/linux# make HOSTLD scripts/kconfig/conf scripts/kconfig/conf -s arch/x86/Kconfig * * Restart config... * * * RCU Subsystem * RCU Implementation 1. Tree-based hierarchical RCU (TREE_RCU) choice[1]: 1 Enable tracing for RCU (RCU_TRACE) [N/y/?] n Tree-based hierarchical RCU fanout value (RCU_FANOUT) [32] (NEW) - Something similar happens when compiling using the Debian way. That's because there are numerous changes between 2.6.26 and 2.6.32, so your old configuration cannot be used as-is. You may want to run make ARCH=x86_64 menuconfig first, so that the defaults are used. Otherwise you will have to answer literally thousands of questions. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Compiling Linux kernel
Hi, Sven. On Friday, 18 December 2009 09:15:43 +0100, Sven Joachim wrote: I am trying to compile Linux 2.6.32 with the source code of kernel.org. Kernel that I'm using at the moment is 2.6.26-2-686 of the Debian GNU/Linux repositories. In order to generate the configuration, I've copied the file corresponding to this kernel to the directory of sources, /usr/src/linux, and I've executed make menuconfig: - # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig - It would be better to start with the configuration of Debian's -amd64 images if you want to use a 64-bit kernel, or even install and boot the 2.6.26-2-amd64 kernel. I was trying installing and booting 2.6.26-2-amd64 kernel and then compiling 2.6.32 kernel of the traditional way: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make menuconfig # make In this case I didn't use the ARCH=x86_64 variable and after to execute 'make menuselect' to select the configuration parameters, when invoking 'make', this worked without doing any type of question. Could this be due to I booted previously with amd64 kernel? But when I execute 'make', restart config becomes and I'm interrogated on some details of the configuration: - antares:/usr/src/linux# make HOSTLD scripts/kconfig/conf scripts/kconfig/conf -s arch/x86/Kconfig * * Restart config... * * * RCU Subsystem * RCU Implementation 1. Tree-based hierarchical RCU (TREE_RCU) choice[1]: 1 Enable tracing for RCU (RCU_TRACE) [N/y/?] n Tree-based hierarchical RCU fanout value (RCU_FANOUT) [32] (NEW) - Something similar happens when compiling using the Debian way. That's because there are numerous changes between 2.6.26 and 2.6.32, so your old configuration cannot be used as-is. You may want to run make ARCH=x86_64 menuconfig first, so that the defaults are used. Otherwise you will have to answer literally thousands of questions. I assume that it must have differences between both kernels versions; for that reason, as I've mentioned in another mail of this thread, after to have copied the file, I followed a similar procedure to which mentioned above, but with the ARCH=x86_64 variable: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make But in spite of to have used 'make menuconfig', 'make' did 'restart config' beginning to do questions to me. Also I tried what Kumer suggested to me of cancel with Ctrl+C when initiating the questions and to execute 'make menuconfig' again, but this only works in the case of not using ARCH=x86_64 with 'make menuconfig' in the second time that it is invoked. But when not using this variable, the processor family returns to be like Pentium Pro and in the .config file I get: # CONFIG_64BIT is not set CONFIG_X86_32=y # CONFIG_X86_64 is not set CONFIG_X86=y I tried selecting in the ncurses menu the option Opteron/Athlon64/Hammer/K8 for the processor family, but even so I don't see changes in the values of the variables that I've indicated above. Thanks for your reply. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On 2009-12-18 16:56 +0100, Daniel Bareiro wrote: I was trying installing and booting 2.6.26-2-amd64 kernel and then compiling 2.6.32 kernel of the traditional way: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make menuconfig # make In this case I didn't use the ARCH=x86_64 variable and after to execute 'make menuselect' to select the configuration parameters, when invoking 'make', this worked without doing any type of question. Could this be due to I booted previously with amd64 kernel? Not really, make menuconfig selects the defaults for options which did not exist in previous kernels. Note that it will use the architecture of the running kernel if you do not specify ARCH. I assume that it must have differences between both kernels versions; for that reason, as I've mentioned in another mail of this thread, after to have copied the file, I followed a similar procedure to which mentioned above, but with the ARCH=x86_64 variable: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make But in spite of to have used 'make menuconfig', 'make' did 'restart config' beginning to do questions to me. If you are not running a 64-bit kernel, you have to specify ARCH=x86_64 for _all_ make invocations. Sven -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Compiling Linux kernel
Hi all! I am trying to compile Linux 2.6.32 with the source code of kernel.org. Kernel that I'm using at the moment is 2.6.26-2-686 of the Debian GNU/Linux repositories. In order to generate the configuration, I've copied the file corresponding to this kernel to the directory of sources, /usr/src/linux, and I've executed make menuconfig: - # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig - But when I execute 'make', restart config becomes and I'm interrogated on some details of the configuration: - antares:/usr/src/linux# make HOSTLD scripts/kconfig/conf scripts/kconfig/conf -s arch/x86/Kconfig * * Restart config... * * * RCU Subsystem * RCU Implementation 1. Tree-based hierarchical RCU (TREE_RCU) choice[1]: 1 Enable tracing for RCU (RCU_TRACE) [N/y/?] n Tree-based hierarchical RCU fanout value (RCU_FANOUT) [32] (NEW) - Something similar happens when compiling using the Debian way. I have the impression that 'make' is taking like reference the configuration in /boot/ and the one that is in /usr/src/linux and when finding differences, it interrogates to me on some things. Is it possible? Under other circumstances, which can be the problem? Thanks in advance for your replies. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On Thu, Dec 17, 2009 at 09:32:59PM -0300, Daniel Bareiro wrote: Hi all! I am trying to compile Linux 2.6.32 with the source code of kernel.org. Kernel that I'm using at the moment is 2.6.26-2-686 of the Debian GNU/Linux repositories. In order to generate the configuration, I've copied the file corresponding to this kernel to the directory of sources, /usr/src/linux, and I've executed make menuconfig: - # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig - But when I execute 'make', restart config becomes and I'm interrogated on some details of the configuration: - antares:/usr/src/linux# make HOSTLD scripts/kconfig/conf scripts/kconfig/conf -s arch/x86/Kconfig * * Restart config... * * * RCU Subsystem * RCU Implementation 1. Tree-based hierarchical RCU (TREE_RCU) choice[1]: 1 Enable tracing for RCU (RCU_TRACE) [N/y/?] n Tree-based hierarchical RCU fanout value (RCU_FANOUT) [32] (NEW) - Something similar happens when compiling using the Debian way. I have the impression that 'make' is taking like reference the configuration in /boot/ and the one that is in /usr/src/linux and when finding differences, it interrogates to me on some things. Is it possible? Under other circumstances, which can be the problem? Thanks in advance for your replies. Well, the kernel build does check what your current config is, and based on that, asks you some new questions. One way I get around this is: cd kernel-build-directory cp /boot/config-$(uname -r) .config make menuconfig # choose the options you want. Run make-kpkg to build my kernel deb. The make menuconfig step refreshes your config, so that you can skip the interrogation step. HTH. Kumar -- mar...@bdsi.com (no longer valid - where are you now, Martin?) -- from /usr/src/linux/drivers/cdrom/mcd.c signature.asc Description: Digital signature
Re: Compiling Linux kernel
Hi, Kumar. On Thursday, 17 December 2009 18:40:07 -0600, Kumar Appaiah wrote: I am trying to compile Linux 2.6.32 with the source code of kernel.org. Kernel that I'm using at the moment is 2.6.26-2-686 of the Debian GNU/Linux repositories. In order to generate the configuration, I've copied the file corresponding to this kernel to the directory of sources, /usr/src/linux, and I've executed make menuconfig: - # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig - But when I execute 'make', restart config becomes and I'm interrogated on some details of the configuration: - antares:/usr/src/linux# make HOSTLD scripts/kconfig/conf scripts/kconfig/conf -s arch/x86/Kconfig * * Restart config... * * * RCU Subsystem * RCU Implementation 1. Tree-based hierarchical RCU (TREE_RCU) choice[1]: 1 Enable tracing for RCU (RCU_TRACE) [N/y/?] n Tree-based hierarchical RCU fanout value (RCU_FANOUT) [32] (NEW) - Something similar happens when compiling using the Debian way. I have the impression that 'make' is taking like reference the configuration in /boot/ and the one that is in /usr/src/linux and when finding differences, it interrogates to me on some things. Is it possible? Under other circumstances, which can be the problem? Well, the kernel build does check what your current config is, and based on that, asks you some new questions. One way I get around this is: cd kernel-build-directory cp /boot/config-$(uname -r) .config make menuconfig # choose the options you want. Run make-kpkg to build my kernel deb. The make menuconfig step refreshes your config, so that you can skip the interrogation step. These steps are similar to which I'm doing, although not to the Debian way: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make The interrogation step comes when doing make. But it seems that it is obtaining some things of the configuration in /boot/, because in the questions I see (without form to be able to modify) that the family is Pentium-Pro (M686) and in /usr/src/linux/.config I have: CONFIG_64BIT=y # CONFIG_X86_32 is not set CONFIG_X86_64=y CONFIG_X86=y Thanks for your reply. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
On Thu, Dec 17, 2009 at 10:26:39PM -0300, Daniel Bareiro wrote: Well, the kernel build does check what your current config is, and based on that, asks you some new questions. One way I get around this is: cd kernel-build-directory cp /boot/config-$(uname -r) .config make menuconfig # choose the options you want. Run make-kpkg to build my kernel deb. The make menuconfig step refreshes your config, so that you can skip the interrogation step. These steps are similar to which I'm doing, although not to the Debian way: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make The interrogation step comes when doing make. But it seems that it is obtaining some things of the configuration in /boot/, because in the questions I see (without form to be able to modify) that the family is Pentium-Pro (M686) and in /usr/src/linux/.config I have: CONFIG_64BIT=y # CONFIG_X86_32 is not set CONFIG_X86_64=y CONFIG_X86=y Well, then did you try using make menuconfig? I would guess that that would refresh your .config, thereby skipping the interrogation. In my case: [ku...@bluemoon ~/Software/Kernel/linux-git master] make [snip] scripts/kconfig/conf -s arch/x86/Kconfig * * Restart config... * * * RCU Subsystem * RCU Implementation 1. Tree-based hierarchical RCU (TREE_RCU) 2. Preemptable tree-based hierarchical RCU (TREE_PREEMPT_RCU) (NEW) choice[1-2]: I now hit Ctrl+C. Next, I do make menuconfig I just select Exit. When asked if I can save the configuration file, I say yes. Now, observe: [ku...@bluemoon ~/Software/Kernel/linux-git master] make scripts/kconfig/conf -s arch/x86/Kconfig CHK include/linux/version.h UPD include/linux/version.h ... The interrogation step was skipped. Hope this helps. Kumar -- Where in the US is Linus? He was in the Promise Land. -- David S. Miller da...@caip.rutgers.edu signature.asc Description: Digital signature
Re: Compiling Linux kernel
Hi, Kumar. On Thursday, 17 December 2009 19:34:09 -0600, Kumar Appaiah wrote: On Thu, Dec 17, 2009 at 10:26:39PM -0300, Daniel Bareiro wrote: Well, the kernel build does check what your current config is, and based on that, asks you some new questions. One way I get around this is: cd kernel-build-directory cp /boot/config-$(uname -r) .config make menuconfig # choose the options you want. Run make-kpkg to build my kernel deb. The make menuconfig step refreshes your config, so that you can skip the interrogation step. These steps are similar to which I'm doing, although not to the Debian way: # cd /usr/src/linux-2.6.32 # cp /boot/config-`uname -r` ./.config # make ARCH=x86_64 menuconfig # make The interrogation step comes when doing make. But it seems that it is obtaining some things of the configuration in /boot/, because in the questions I see (without form to be able to modify) that the family is Pentium-Pro (M686) and in /usr/src/linux/.config I have: CONFIG_64BIT=y # CONFIG_X86_32 is not set CONFIG_X86_64=y CONFIG_X86=y Well, then did you try using make menuconfig? I would guess that that would refresh your .config, thereby skipping the interrogation. In my case: [ku...@bluemoon ~/Software/Kernel/linux-git master] make [snip] scripts/kconfig/conf -s arch/x86/Kconfig * * Restart config... * * * RCU Subsystem * RCU Implementation 1. Tree-based hierarchical RCU (TREE_RCU) 2. Preemptable tree-based hierarchical RCU (TREE_PREEMPT_RCU) (NEW) choice[1-2]: I now hit Ctrl+C. Next, I do make menuconfig I just select Exit. When asked if I can save the configuration file, I say yes. Now, observe: [ku...@bluemoon ~/Software/Kernel/linux-git master] make scripts/kconfig/conf -s arch/x86/Kconfig CHK include/linux/version.h UPD include/linux/version.h ... The interrogation step was skipped. This only works in the case of not using ARCH=x86_64 with make menuconfig in the second time that is invoked. But when not using this variable, the processor family returns to be like Pentium Pro and in the .config file I get: # CONFIG_64BIT is not set CONFIG_X86_32=y # CONFIG_X86_64 is not set CONFIG_X86=y I tried selecting in the ncurses menu the option Opteron/Athlon64/Hammer/K8 for the processor family, but even so I do not see changes in the values of the variables that I've indicated above. Thanks for your reply. Regards, Daniel -- Fingerprint: BFB3 08D6 B4D1 31B2 72B9 29CE 6696 BF1B 14E6 1D37 Powered by Debian GNU/Linux Lenny - Linux user #188.598 signature.asc Description: Digital signature
Re: Compiling Linux kernel
Dear Daniel, On Thu, Dec 17, 2009 at 11:41:11PM -0300, Daniel Bareiro wrote: [snip menuconfig discussion] This only works in the case of not using ARCH=x86_64 with make menuconfig in the second time that is invoked. But when not using this variable, the processor family returns to be like Pentium Pro and in the .config file I get: # CONFIG_64BIT is not set CONFIG_X86_32=y # CONFIG_X86_64 is not set CONFIG_X86=y I tried selecting in the ncurses menu the option Opteron/Athlon64/Hammer/K8 for the processor family, but even so I do not see changes in the values of the variables that I've indicated above. Well, I guess I've run out of options. The easy (but dirty) way out is to run make menuconfig to use a known, good .config file, and then run make silentoldconfig, and answer the default to everything you don't think you need to care about (read hold the Enter key till it ends). This works for me, but I bet someone can tell us a better solution… Sorry for not being much help. Kumar -- After watching my newly-retired dad spend two weeks learning how to make a new folder, it became obvious that intuitive mostly means what the writer or speaker of intuitive likes. -- Bruce Ediger, bedi...@teal.csn.org, on X the intuitiveness of a Mac interface signature.asc Description: Digital signature
Re: compiling a kernel from kernel.org [SOLVED]
Hey, Thx everybody for your quick answers and friendly help. You were right. I removed all Xen options from the kernel config and linux-image-2.6.32-rc5_20091016-2_amd64.deb has been built. Installing it with dpkg -i ... worked fine. The only problem to be solved was the initrd. it has NOT been generated by dpkg during the installation. so I generated it using mkinitramfs -c -k 2.6.32-rc5. worked fine. update-grub - worked fine as well. Finally I can use KMS with my radeon 4670 :) Greetings Gregor Galwas On Tue, 20 Oct 2009 21:34:53 +0200, Manoj Srivastava sriva...@ieee.org wrote: Hi, Well, firstly, if you are going to be using the buildpackage target, instead of the far faster kernel_image target, you should either configure /etc/kernel-pkg.conf, adding your name and email, and have that in a keyring your gpg knows about, or pass the --us and --uc arguments on the command line. I think, unless you know what you are doing, try first with kernel_image target. Next, you seem to have Xen options turned on in your config, and thus make-kpkg is trying to create a xenu-linux image. This is undergoing some development, so if you want Xen images, please get the 12.024 version of kernel-package from Sid. If you are just trying to build a normal kernel, redo your .config not to have Xen stuff in it. manoj -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: compiling a kernel from kernel.org [SOLVED]
Gregor Galwas wrote: Hey, Thx everybody for your quick answers and friendly help. You were right. I removed all Xen options from the kernel config and linux-image-2.6.32-rc5_20091016-2_amd64.deb has been built. Installing it with dpkg -i ... worked fine. The only problem to be solved was the initrd. it has NOT been generated by dpkg during the installation. so I generated it using mkinitramfs -c -k 2.6.32-rc5. worked fine. update-grub - worked fine as well. Finally I can use KMS with my radeon 4670 :) Greetings Gregor Galwas Hi, regarding the initrd, you need to copy over to /etc/kernel/* the relevant scripts from /usr/share/kernel-package/examples/etc/kernel/* (look for initramfs or yaird if you use it). It will save you the mkinitramfs step. Also have a look at /usr/share/kernel-package/examples/etc/sample.kernel-img.conf to trigger links creation and such. Tom -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: compiling a kernel from kernel.org [SOLVED]
On Wed, Oct 21 2009, Gregor Galwas wrote:
The only problem to be solved was the initrd. it has NOT been
generated by dpkg during the installation.
so I generated it using mkinitramfs -c -k 2.6.32-rc5. worked
fine. update-grub - worked fine as well.
,[ Manual page make-kpkg(1) ]
| --initrd
| If make-kpkg is generating a kernel-image package, arrange to
| convey to the hook scripts run from the post installation
| maintainer scripts that this image requires an initrd, and that
| the initrd generation hook scripts should not short circuit
| early. Without this option, the example initramfs hook scripts
| bundled in with kernel-package will take no action on
| installation. The same effect can be achieved by setting the
| environment variable INITRD to any non empty value. Please note
| that unless there are hook scripts in /etc/kernel or added into
| the hook script parameter of /etc/kernel-img.conf. no initrd
| will be created (the bundled in example scripts are just
| examples -- user action is required before anything happens).
`
,[ /usr/share/doc/kernel-package/README.gz ]
| gotchas. Note that you will have to arrange for the actual
| initrd creation to take place by installing a script like
| /usr/share/kernel-package/examples/etc/kernel/post{inst,rm}.d/yaird
| or, alternately,
| /usr/share/kernel-package/examples/etc/kernel/post{inst,rm}.d/initramfs
| into the corresponding directories /etc/kernel/post{inst,rm}.d,
| since the kernel-postinst does not arrange for the initramfs
| creator to be called. You can thus select your own;
| initramfs-tools or yaird.
|
| Let me repeat:
| Since nothing is created automatically. you need to provide a hook
| script for things to happen when you install the kernel image
| package. The user provides such scripts. For example, to invoke
| mkinitramfs, I did:
| --8---cut here---start-8---
| cp /usr/share/kernel-package/examples/etc/kernel/postinst.d/yaird \
| /etc/kernel/postinst.d/
| cp /usr/share/kernel-package/examples/etc/kernel/postrm.d/yaird \
| /etc/kernel/postrm.d/
| --8---cut here---end---8---
|
| Or, alternately, you could do:
| --8---cut here---start-8---
| cp /usr/share/kernel-package/examples/etc/kernel/postinst.d/initramfs \
| /etc/kernel/postinst.d/
| cp /usr/share/kernel-package/examples/etc/kernel/postrm.d/initramfs \
| /etc/kernel/postrm.d/
| --8---cut here---end---8---
|
| These scripts above to nothing unless the corresponding
| packages are installed (initramfs-tools or yaird), so you could
| potentially cp both over -- as long as you never install both yaird
| and initramfs-tools at the same time.
|
| To run grub, I have in /etc/kernel-img.conf:
| --8---cut here---start-8---
| postinst_hook = update-grub
| postrm_hook = update-grub
| --8---cut here---end---8---
|
| You can look at other example in the examples directory:
| /usr/share/kernel-package/examples/ to see if there are other example
| script you want to cp into /etc/kernel -- and you can create your own
| scripts.
|
| For example, if you use linux-headers-* packages to compile third
| party modules so that you do not have to keep the sources directory
| around, you might be interested in:
| --8---cut here---start-8---
| /etc/kernel/header_postinst.d/link
| /etc/kernel/header_postrm.d/link
| /etc/kernel/header_prerm.d/link
| /etc/kernel/postinst.d/force-build-link
| /etc/kernel/postrm.d/force-build-link
| --8---cut here---end---8---
|
| These scripts will try to make sure that the symlink
| /lib/modules/$VERSION/build
| is sane -- that is points to the header packages whether you install
| the image packages first, or the header packages first -- and takes
| care of cleanup when either of the packages are installed.
`
Perhaps people who maintain FAQ's on this list add the above?
manoj
--
Some people have a great ambition: to build something that will last, at
least until they've finished building it.
Manoj Srivastava sriva...@acm.org http://www.golden-gryphon.com/
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
compiling a kernel from kernel.org
hello, in the new last days I've been trying to compile a recent kernel from kernel.org I fetched a kernel from http://www.kernel.org/ mainline: 2.6.32-rc5 2009-10-16 [Full Source] http://www.kernel.org/pub/linux/kernel/v2.6/testing/linux-2.6.32-rc5.tar.bz2 unpacked it using tar xjvf linux-2.6.32-rc5.tar.bz2 -C /usr/src linked : ln -s /usr/src/linux-2.6.32-rc5.tar.bz2 /usr/src/linux cp /boot/config-`uname -r` /usr/src/linux/.config make oldconfig then I started building the packages with make-kpkg buildpackage --revision 20091016 --initrd it started compiling it and it ended after a loinger time with : make[1]: Leaving directory `/usr/src/linux-2.6.32-rc5' dpkg-genchanges -b -mUnknown Kernel Package Maintainer unkn...@unconfigured.in.etc.kernel-pkg.conf ../linux-source-2.6.32-rc5_20091016_amd64.changes dpkg-genchanges: warning: package linux-image-2.6.32-rc5 in control file but not in files list dpkg-genchanges: warning: package linux-image-2.6.32-rc5-dbg in control file but not in files list dpkg-genchanges: warning: package linux-uml-2.6.32-rc5 in control file but not in files list dpkg-genchanges: warning: package linux-xen0-2.6.32-rc5 in control file but not in files list dpkg-genchanges: binary-only upload - not including any source code signfile linux-source-2.6.32-rc5_20091016_amd64.changes gpg: skipped Unknown Kernel Package Maintainer: secret key not available gpg: [stdin]: clearsign failed: secret key not available dpkg-buildpackage: binary only upload (no source included) dpkg-buildpackage: warning: Failed to sign .changes file make: *** [debian/stamp/build/buildpackage] Error 1 following packages have been created in /usr/src: linux-doc-2.6.32-rc5_20091016_all.deb linux-headers-2.6.32-rc5_20091016_amd64.deb linux-headers-2.6.32-rc5_2.6.32-rc5-10.00.Custom_amd64.deb linux-manual-2.6.32-rc5_20091016_all.deb linux-source-2.6.32-rc5_20091016_all.deb linux-source-2.6.32-rc5_20091016_amd64.changes linux-xenu-2.6.32-rc5_20091016_amd64.deb linux-xenu-2.6.32-rc5_2.6.32-rc5-10.00.Custom_amd64.deb ok, so far - it looked good even thouigh that xenu part annoyed me somehow. when I treid to install linux-xenu-2.6.32-rc5_20091016_amd64.deb I got: pegasos:/usr/src# dpkg -i linux-xenu-2.6.32-rc5_20091016_amd64.deb Selecting previously deselected package linux-xenu-2.6.32-rc5. (Reading database ... 215220 files and directories currently installed.) Unpacking linux-xenu-2.6.32-rc5 (from linux-xenu-2.6.32-rc5_20091016_amd64.deb) ... Done. Setting up linux-xenu-2.6.32-rc5 (20091016) ... Internal Error: Could not find image (/boot/vmlinuz-2.6.32-rc5) dpkg: error processing linux-xenu-2.6.32-rc5 (--install): subprocess installed post-installation script returned error exit status 2 Errors were encountered while processing: linux-xenu-2.6.32-rc5 Obviously something is badly wrong then. I am using Debian Unstable with kernel 2.6.31 (Linux pegasos 2.6.31-trunk-amd64 #1 SMP Mon Oct 12 23:36:11 UTC 2009 x86_64 GNU/Linux) and Grub2 anyone has an idea whats wrong here? Or am I doing it completely wrong? Thx for reading at least... Gregor Galwas -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: compiling a kernel from kernel.org
Hi, Well, firstly, if you are going to be using the buildpackage target, instead of the far faster kernel_image target, you should either configure /etc/kernel-pkg.conf, adding your name and email, and have that in a keyring your gpg knows about, or pass the --us and --uc arguments on the command line. I think, unless you know what you are doing, try first with kernel_image target. Next, you seem to have Xen options turned on in your config, and thus make-kpkg is trying to create a xenu-linux image. This is undergoing some development, so if you want Xen images, please get the 12.024 version of kernel-package from Sid. If you are just trying to build a normal kernel, redo your .config not to have Xen stuff in it. manoj -- This is an unauthorized cybernetic announcement. Manoj Srivastava sriva...@acm.org http://www.golden-gryphon.com/ 1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
problem compiling new kernel
Hi to Everyone, With my old SARGE system, as stated here last week, I was unable to boot newly compiled kernels, most likely because of the faulty initrd.img that I got using 'mkinitrd'. For this reason, I decided to install Lenny. Now, with Lenny, I haven't so far been able to go any further in recompiling usable kernels. 2.6.30.4 did compile without errors, but then I found that I had no 'mkinitrd' on my new system. How am I to generate an initrd.img ? I was unable to find any package that I could install using apt-get with the name 'mkinitrd', there is no such thing either in the cache base. However, when I installed Lenny (72 hours ago), the install program did generate an initrd.img-2.6.26-2-686 that is now installed in my /boot directory, and my /boot/grub/menu.lst file does specify this file in the boot menu. Thanks in advance for your help P.S. I did find a doc about some related subjects in the /Documentation directory of the downloaded kernel... but it was so old and obsolete that it was useless, dated 1998 and mentioning only lilo, not grub... this in a recent kernel doc (2008-2009) ! -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: problem compiling new kernel
bdebreil 写道: Hi to Everyone, With my old SARGE system, as stated here last week, I was unable to boot newly compiled kernels, most likely because of the faulty initrd.img that I got using 'mkinitrd'. For this reason, I decided to install Lenny. Now, with Lenny, I haven't so far been able to go any further in recompiling usable kernels. 2.6.30.4 did compile without errors, but then I found that I had no 'mkinitrd' on my new system. How am I to generate an initrd.img ? I was unable to find any package that I could install using apt-get with the name 'mkinitrd', there is no such thing either in the cache base. However, when I installed Lenny (72 hours ago), the install program did generate an initrd.img-2.6.26-2-686 that is now installed in my /boot directory, and my /boot/grub/menu.lst file does specify this file in the boot menu. Thanks in advance for your help P.S. I did find a doc about some related subjects in the /Documentation directory of the downloaded kernel... but it was so old and obsolete that it was useless, dated 1998 and mentioning only lilo, not grub... this in a recent kernel doc (2008-2009) ! update-initramfs maybe the tool you are after. Google it or man it. It should be installed on your system. Or you can install it with initramfs-tools. Good luck. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: problem compiling new kernel
On Fri, 28 Aug 2009 19:07:45 +0800 Niu Kun haoniu...@gmail.com wrote: bdebreil 写道: ... Now, with Lenny, I haven't so far been able to go any further in recompiling usable kernels. 2.6.30.4 did compile without errors, but then I found that I had no 'mkinitrd' on my new system. How am I to generate an initrd.img ? I was unable to find any package that I could install using apt-get with the name 'mkinitrd', there is no such thing either in the cache base. However, when I installed Lenny (72 hours ago), the install program did generate an initrd.img-2.6.26-2-686 that is now installed in my /boot directory, and my /boot/grub/menu.lst file does specify this file in the boot menu. ... update-initramfs maybe the tool you are after. Google it or man it. It should be installed on your system. Or you can install it with initramfs-tools. Assuming that you're building kernel packages with kernel-package, you may be being hit by this: The image postinst no longer runs the initramfs creation commands. Instead, there are example scripts provided that will perform the task. These scripts will work for official kernel images as well. /usr/share/doc/NEWS.Debian.gz Basically, the short version of what you need to do is: cp /usr/share/doc/kernel-package/examples/etc/kernel/postinst.d/initramfs /etc/kernel/postinst.d/ [This really needs to be in a faq somewhere, if it isn't already. It comes up frequently on the list.] Celejar -- mailmin.sourceforge.net - remote access via secure (OpenPGP) email ssuds.sourceforge.net - A Simple Sudoku Solver and Generator -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: problem compiling new kernel
On Fri, Aug 28 2009, Celejar wrote: Assuming that you're building kernel packages with kernel-package, you may be being hit by this: The image postinst no longer runs the initramfs creation commands. Instead, there are example scripts provided that will perform the task. These scripts will work for official kernel images as well. /usr/share/doc/NEWS.Debian.gz Basically, the short version of what you need to do is: cp /usr/share/doc/kernel-package/examples/etc/kernel/postinst.d/initramfs /etc/kernel/postinst.d/ You might also want to clean things up: cp /usr/share/doc/kernel-package/examples/etc/kernel/postrm.d/initramfs \ /etc/kernel/postrm.d/ This clean up files created in the postinst. manoj -- Now here's something you're really going to like! Rocket J. Squirrel Manoj Srivastava sriva...@acm.org http://www.golden-gryphon.com/ 1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Problem compiling custom kernel on debian
Hi, I am trying to compile Linux kernel 2.6.29-1 as described here http://kernel-handbook.alioth.debian.org/ch-common-tasks.html, section 4.4, i.e. I copy the /boot/config-2.6.26-1-686 to untarred linux kernel directory as .config and after doing make menuconfig, when I try to do make-kpkg clean I get error that file system is read only. Somehow just before running this command, the file system wasn´t readonly but after running this, my file system has become read only. Looking at the /etc/fstab, it shows the following line: /dev/hdc1 / ext3errors=remount-ro 0 1 which probably means that in case of errors mount the file system as read only...but not sure what is causing this. The full error log from make-kpkg clean is shown below. exec debian/rules DEBIAN_REVISION=5:10.Custom clean == making target CLN-common [new prereqs: ]== == making target unpatch_now [new prereqs: ]== == making target real_stamp_clean [new prereqs: unpatch_now]== running clean test ! -e scripts/package/builddeb || mv -f scripts/package/builddeb scripts/package/builddeb.kpkg-dist test ! -e scripts/package/Makefile || test -f scripts/package/Makefile.kpkg-dist || (mv -f scripts/package/Makefile scripts/package/Makefile.kpkg-dist (echo # Dummy file ; echo help:) scripts/package/Makefile) test ! -f .config || cp -pf .config ,,precious /usr/bin/makeARCH=i386 -C Documentation/lguest clean make[1]: Entering directory `/home/divkis01/downloads/linux-2.6.29.1/Documentation/lguest' rm -f lguest make[1]: Leaving directory `/home/divkis01/downloads/linux-2.6.29.1/Documentation/lguest' test ! -f Makefile || \ /usr/bin/makeARCH=i386 distclean make[1]: Entering directory `/home/divkis01/downloads/linux-2.6.29.1' CLEAN scripts/basic CLEAN scripts/kconfig CLEAN include/config find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/orphan.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/locking.h': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/inode-item.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/inode.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/inode-map.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/ref-cache.h': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/root-tree.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/ioctl.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/print-tree.h': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/ioctl.h': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/ordered-data.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/print-tree.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/ordered-data.h': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/locking.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/ref-cache.c': Input/output error find: `/home/divkis01/downloads/linux-2.6.29.1/fs/btrfs/struct-funcs.c': Input/output error make[1]: Leaving directory `/home/divkis01/downloads/linux-2.6.29.1' test ! -f scripts/package/builddeb.kpkg-dist || mv -f scripts/package/builddeb.kpkg-dist scripts/package/builddeb mv: cannot move `scripts/package/builddeb.kpkg-dist' to `scripts/package/builddeb': Read-only file system Any help is welcome, Thanks in advance, Divick -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Problem compiling custom kernel on debian
Divick Kishore: make-kpkg clean I get error that file system is read only. Somehow just before running this command, the file system wasn´t readonly but after running this, my file system has become read only. Your problem is most probably unrelated to your building a kernel. Take a look into /var/log/syslog and/or dmesg output. You will probably find messages about hard disk errors. Post them here if you have trouble interpreting them. Posting the output of 'smartctl -a /dev/hda' (from smartmontools, insert your hard drive device name) might help as well. J. -- Americans have a better life. [Agree] [Disagree] http://www.slowlydownward.com/NODATA/data_enter2.html signature.asc Description: Digital signature
Re: Problem compiling custom kernel on debian
Hi,
On Tue, Apr 21, 2009 at 1:22 PM, Jochen Schulz m...@well-adjusted.de
wrote:
Your problem is most probably unrelated to your building a kernel. Take
a look into /var/log/syslog and/or dmesg output. You will probably find
messages about hard disk errors. Post them here if you have trouble
interpreting them.
Strangely this starts off when I try to run make-kpkg. I don´t see any
error messages in the dmesg log, till I run this command. After I run
this, I see the following error messages in the dmesg log. I am not
sure what does that point to...i.e. how it can be fixed.
[ 629.127668] hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
[ 629.127668] hdc: dma_intr: error=0x40 { UncorrectableError },
LBAsect=1841587, sector=1841583
[ 629.127668] ide: failed opcode was: unknown
[ 629.127668] end_request: I/O error, dev hdc, sector 1841583
[ 629.127668] EXT3-fs error (device hdc1): ext3_get_inode_loc: unable
to read inode block - inode=58202, block=230190
[ 629.127668] Aborting journal on device hdc1.
[ 629.128573] ext3_abort called.
[ 629.128587] EXT3-fs error (device hdc1): ext3_journal_start_sb:
Detected aborted journal
[ 629.128599] Remounting filesystem read-only
[ 629.129010] Remounting filesystem read-only
[ 629.971851] hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
[ 629.971851] hdc: dma_intr: error=0x40 { UncorrectableError },
LBAsect=1841587, sector=1841583
[ 629.971851] ide: failed opcode was: unknown
[ 629.971851] end_request: I/O error, dev hdc, sector 1841583
[ 629.971851] EXT3-fs error (device hdc1): ext3_get_inode_loc: unable
to read inode block - inode=58199, block=230190
[ 630.980013] hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
[ 630.980013] hdc: dma_intr: error=0x40 { UncorrectableError },
LBAsect=1841587, sector=1841583
[ 630.980013] ide: failed opcode was: unknown
[ 630.980013] end_request: I/O error, dev hdc, sector 1841583
[ 630.980013] EXT3-fs error (device hdc1): ext3_get_inode_loc: unable
to read inode block - inode=58193, block=230190
[ 632.136021] hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
[ 632.136021] hdc: dma_intr: error=0x40 { UncorrectableError },
LBAsect=1841587, sector=1841583
[ 632.136021] ide: failed opcode was: unknown
[ 632.136021] end_request: I/O error, dev hdc, sector 1841583
[ 632.136021] EXT3-fs error (device hdc1): ext3_get_inode_loc: unable
to read inode block - inode=58195, block=230190
[ 633.295745] hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
[ 633.295771] hdc: dma_intr: error=0x40 { UncorrectableError },
LBAsect=1841587, sector=1841583
[ 633.295795] ide: failed opcode was: unknown
[ 633.295809] end_request: I/O error, dev hdc, sector 1841583
[ 633.295921] EXT3-fs error (device hdc1): ext3_get_inode_loc: unable
to read inode block - inode=58194, block=230190
[ 634.475282] psmouse.c: Wheel Mouse at isa0060/serio1/input0 lost
synchronization, throwing 3 bytes away.
[ 634.515097] hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
[ 634.515097] hdc: dma_intr: error=0x40 { UncorrectableError },
LBAsect=1841587, sector=1841583
[ 634.515097] ide: failed opcode was: unknown
[ 634.515097] end_request: I/O error, dev hdc, sector 1841583
[ 634.515097] EXT3-fs error (device hdc1): ext3_get_inode_loc: unable
to read inode block - inode=58206, block=230190
Posting the output of 'smartctl -a /dev/hda' (from smartmontools, insert
your hard drive device name) might help as well.
OK. I will try doing that as wellsince the current file system is
readonly, I can´t install anything till I reboot :(. BTW is there a
way to remount the root filesystem as rw after it has encountered this
error and becomes ro ?
Thanks a lot for your help,
Regards,
Divick
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: Problem compiling custom kernel on debian
Hi, Posting the output of 'smartctl -a /dev/hda' (from smartmontools, insert your hard drive device name) might help as well. The output from smatctl is fairly cryptic for me, posting it below: sudo smartctl -a /dev/hdc1 smartctl version 5.38 [i686-pc-linux-gnu] Copyright (C) 2002-8 Bruce Allen Home page is http://smartmontools.sourceforge.net/ === START OF INFORMATION SECTION === Device Model: SAMSUNG SV4084H Serial Number:0274J1FR139181 Firmware Version: PM100-12 User Capacity:40,822,161,408 bytes Device is:In smartctl database [for details use: -P show] ATA Version is: 4 ATA Standard is: ATA/ATAPI-4 T13 1153D revision 17 Local Time is:Tue Apr 21 21:37:59 2009 IST == WARNING: May need -F samsung or -F samsung2 enabled; see manual for details. SMART support is: Available - device has SMART capability. SMART support is: Enabled === START OF READ SMART DATA SECTION === SMART overall-health self-assessment test result: PASSED General SMART Values: Offline data collection status: (0x00) Offline data collection activity was never started. Auto Offline Data Collection: Disabled. Self-test execution status: ( 0) The previous self-test routine completed without error or no self-test has ever been run. Total time to complete Offline data collection: ( 960) seconds. Offline data collection capabilities:(0x1b) SMART execute Offline immediate. Auto Offline data collection on/off support. Suspend Offline collection upon new command. Offline surface scan supported. Self-test supported. No Conveyance Self-test supported. No Selective Self-test supported. SMART capabilities:(0x0003) Saves SMART data before entering power-saving mode. Supports SMART auto save timer. Error logging capability:(0x01) Error logging supported. No General Purpose Logging support. Short self-test routine recommended polling time:( 1) minutes. Extended self-test routine recommended polling time:( 16) minutes. SMART Attributes Data Structure revision number: 9 Vendor Specific SMART Attributes with Thresholds: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000a 100 100 051Old_age Always - 210 4 Start_Stop_Count0x0032 094 094 000Old_age Always - 6287 5 Reallocated_Sector_Ct 0x0033 253 253 010Pre-fail Always - 0 7 Seek_Error_Rate 0x000a 253 253 051Old_age Always - 0 8 Seek_Time_Performance 0x0024 253 253 000Old_age Offline - 0 9 Power_On_Hours 0x0032 097 097 000Old_age Always - 2136356 12 Power_Cycle_Count 0x0032 096 096 000Old_age Always - 4064 197 Current_Pending_Sector 0x0032 253 253 010Old_age Always - 0 198 Offline_Uncorrectable 0x0030 253 253 010Old_age Offline - 0 199 UDMA_CRC_Error_Count0x000a 200 200 000Old_age Always - 0 200 Multi_Zone_Error_Rate 0x000a 100 100 051Old_age Always - 0 201 Soft_Read_Error_Rate0x000a 100 100 051Old_age Always - 0 SMART Error Log Version: 1 Warning: ATA error count 238 inconsistent with error log pointer 5 ATA Error Count: 238 (device log contains only the most recent five errors) CR = Command Register [HEX] FR = Features Register [HEX] SC = Sector Count Register [HEX] SN = Sector Number Register [HEX] CL = Cylinder Low Register [HEX] CH = Cylinder High Register [HEX] DH = Device/Head Register [HEX] DC = Device Command Register [HEX] ER = Error register [HEX] ST = Status register [HEX] Powered_Up_Time is measured from power on, and printed as DDd+hh:mm:SS.sss where DD=days, hh=hours, mm=minutes, SS=sec, and sss=millisec. It wraps after 49.710 days. Error 238 occurred at disk power-on lifetime: 17802 hours (741 days + 18 hours) When the command that caused the error occurred, the device was active or idle. After command completion occurred, registers were: ER ST SC SN CL CH DH -- -- -- -- -- -- -- 40 88 08 af 19 1c e0 8 sectors at LBA = 0x001c19af = 1841583 Commands leading to the command that caused
Re: Problem compiling custom kernel on debian
Divick Kishore:
On Tue, Apr 21, 2009 at 1:22 PM, Jochen Schulz m...@well-adjusted.de
wrote:
Your problem is most probably unrelated to your building a kernel. Take
a look into /var/log/syslog and/or dmesg output. You will probably find
messages about hard disk errors. Post them here if you have trouble
interpreting them.
Strangely this starts off when I try to run make-kpkg. I don´t see any
error messages in the dmesg log, till I run this command. After I run
this, I see the following error messages in the dmesg log. I am not
sure what does that point to...i.e. how it can be fixed.
Probably the kernel sourcs are located on a damaged location on your
hard drive.
[ 629.127668] hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
[ 629.127668] hdc: dma_intr: error=0x40 { UncorrectableError },
LBAsect=1841587, sector=1841583
[ 629.127668] ide: failed opcode was: unknown
[ 629.127668] end_request: I/O error, dev hdc, sector 1841583
[ 629.127668] EXT3-fs error (device hdc1): ext3_get_inode_loc: unable
to read inode block - inode=58202, block=230190
I am not completely sure, but it looks like your hard drive is dying.
You can try to run a bad blocks test (see man fsck.ext3, I don't have
the syntax at hand).
OK. I will try doing that as wellsince the current file system is
readonly, I can´t install anything till I reboot :(. BTW is there a
way to remount the root filesystem as rw after it has encountered this
error and becomes ro ?
mount -o remount,rw /
But you are risking (further) data loss if you keep writing to the disk.
J.
--
If I am asked 'How are you' more than a million times in my life I
promise to explode.
[Agree] [Disagree]
http://www.slowlydownward.com/NODATA/data_enter2.html
signature.asc
Description: Digital signature
Re: Problem compiling custom kernel on debian
Divick Kishore: General SMART Values: Offline data collection status: (0x00) Offline data collection activity was never started. Auto Offline Data Collection: Disabled. Self-test execution status: ( 0) The previous self-test routine completed without error or no self-test has ever been run. You can make your disk run SMART self-tests to find out whether your disk thinks it is dying. The command is something like 'smartctl -t long /dev/hdX'. Vendor Specific SMART Attributes with Thresholds: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x000a 100 100 051Old_age Always - 210 -- snip I am not an expert, but the values look ok to me. Gemeral rule of thumb: compare WORST with THRESH. If WORST is lower, you might have a problem. If the attribute is of type Pre-fail, it is probably severe. 197 Current_Pending_Sector 0x0032 253 253 010Old_age Always - 0 No unreadable sectors are waiting to be relocated by the disk. This is good. Error 238 occurred at disk power-on lifetime: 17802 hours (741 days + 18 hours) When the command that caused the error occurred, the device was active or idle. After command completion occurred, registers were: ER ST SC SN CL CH DH -- -- -- -- -- -- -- 40 88 08 af 19 1c e0 8 sectors at LBA = 0x001c19af = 1841583 Commands leading to the command that caused the error were: CR FR SC SN CL CH DH DC Powered_Up_Time Command/Feature_Name -- -- -- -- -- -- -- -- c8 00 08 af 19 1c e0 00 00:07:19.000 READ DMA ca 00 40 87 05 44 e0 00 00:07:19.000 WRITE DMA ca 00 10 47 05 44 e0 00 00:07:19.000 WRITE DMA ca 00 08 2f 05 44 e0 00 00:07:19.000 WRITE DMA ca 00 20 07 05 44 e0 00 00:07:19.000 WRITE DMA I *think* these errors aren't necessarily bad, but I cannot really tell. What I would do: use the disk as little as possible, get a replacement and do a backup. J. -- I like my Toyota RAV4 because of the commanding view of the traffic jams. [Agree] [Disagree] http://www.slowlydownward.com/NODATA/data_enter2.html signature.asc Description: Digital signature
Re: Problem compiling 2.6.26 kernel
On 09/28/2008 02:49 PM, Arthur Barlow wrote: I did check the settings in .config. Both BLK_DEV_IDE and BLK_DEV_IDEDISK are set to y. As for the ATA/SATA settings, they were not set, but my disk is an older one and should not care, but I tried setting them and trying again, and I got the same result. The drive is a Western Digital 40G drive, model WDC WD400JB-00FMA0. As I mentioned it's using Reiserfs 3.6. Also I got the source code from Debian's repository. It's just a Debianized tar ball. That is quite confusing. As the boot messages fly by, you can stop them by using Control-S. Another Control-S allows the boot-up to continue. Use that method to pause the display long enough for you to see how your IDE drives are recognized by the kernel. I have messages like these: ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx ... hda: HDS728080PLAT20, ATA DISK drive ... ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 I would hope that you have something similar. Also try roots of /dev/hdb1 and /dev/hdc1 and /dev/sda1. Even though those are not likely to work, it's probably worth a try. Can we see the entire boot stanza used by LILO? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Problem compiling 2.6.26 kernel
On 09/27/2008 10:42 PM, Arthur Barlow wrote: I have used Debian for many years and I have always used the kernel-package program that is included with Debian to compile new kernels. I have an older Athlon PC that has the Reiserfs 3.6 on it. After I do the make-kpkg --revision= ## kernel-image and then use dpkg -i to install the new kernel, I keep getting a kernel panic. The message reads like this: VFS: Cannot open rootdevice 301 or unknown-block(3,1) Please append a correct root= boot option; here are available partitions: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(3,1) It appears to have trouble mounting the file systems. I've never run into this before. The boot manager is still lilo and there is a root=/dev/hda1 statement in the lilo.conf file. Any suggestions would be appreciated. Thanks. You seem to be the second person in two days who has complained about this. Has kernel.org changed how IDE devices are accessed? For the time being, you can probably boot using a root of /dev/sda1. The problem is probably that the kernel IDE driver (CONFIG_BLK_DEV_IDE) is not being loaded before libata (CONFIG_ATA). I'm running 2.6.26.5 with a config file that I've used for about a year; I generally get my kernels from kernel.org. It would be interesting to know where your kernel source comes from. Also, please give us the output of these commands: grep BLK_DEV_IDE .config grep CONFIG_S\?ATA .config BTW, I have both BLK_DEV_IDE and BLK_DEV_IDEDISK set to y. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Problem compiling 2.6.26 kernel
On Sun, Sep 28, 2008 at 02:28:43AM -0500, Mumia W.. wrote: On 09/27/2008 10:42 PM, Arthur Barlow wrote: I have used Debian for many years and I have always used the kernel-package program that is included with Debian to compile new kernels. I have an older Athlon PC that has the Reiserfs 3.6 on it. After I do the make-kpkg --revision= ## kernel-image and then use dpkg -i to install the new kernel, I keep getting a kernel panic. The message reads like this: VFS: Cannot open rootdevice 301 or unknown-block(3,1) Please append a correct root= boot option; here are available partitions: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(3,1) It appears to have trouble mounting the file systems. I've never run into this before. The boot manager is still lilo and there is a root=/dev/hda1 statement in the lilo.conf file. Any suggestions would be appreciated. Thanks. You seem to be the second person in two days who has complained about this. Has kernel.org changed how IDE devices are accessed? For the time being, you can probably boot using a root of /dev/sda1. The problem is probably that the kernel IDE driver (CONFIG_BLK_DEV_IDE) is not being loaded before libata (CONFIG_ATA). I'm running 2.6.26.5 with a config file that I've used for about a year; I generally get my kernels from kernel.org. Interestingly, I installed the latest 2.6.26.6 and I was burning a dvd, but I had to reboot the machine because the process got stuck, I could not kill 09 the process and it locked the dvd drive It would be interesting to know where your kernel source comes from. Also, please give us the output of these commands: grep BLK_DEV_IDE .config grep CONFIG_S\?ATA .config BTW, I have both BLK_DEV_IDE and BLK_DEV_IDEDISK set to y. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- A young girl once committed suicide because her mother refused her a new bonnet. Coroner's verdict: Death from excessive spunk. -- Sacramento Daily Union, September 13, 1860 signature.asc Description: Digital signature
Re: Problem compiling 2.6.26 kernel
Alex Samad wrote: On Sun, Sep 28, 2008 at 02:28:43AM -0500, Mumia W.. wrote: On 09/27/2008 10:42 PM, Arthur Barlow wrote: I have used Debian for many years and I have always used the kernel-package program that is included with Debian to compile new kernels. I have an older Athlon PC that has the Reiserfs 3.6 on it. After I do the make-kpkg --revision= ## kernel-image and then use dpkg -i to install the new kernel, I keep getting a kernel panic. The message reads like this: VFS: Cannot open rootdevice 301 or unknown-block(3,1) Please append a correct root= boot option; here are available partitions: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(3,1) It appears to have trouble mounting the file systems. I've never run into this before. The boot manager is still lilo and there is a root=/dev/hda1 statement in the lilo.conf file. Any suggestions would be appreciated. Thanks. You seem to be the second person in two days who has complained about this. Has kernel.org changed how IDE devices are accessed? For the time being, you can probably boot using a root of /dev/sda1. The problem is probably that the kernel IDE driver (CONFIG_BLK_DEV_IDE) is not being loaded before libata (CONFIG_ATA). I'm running 2.6.26.5 with a config file that I've used for about a year; I generally get my kernels from kernel.org. Interestingly, I installed the latest 2.6.26.6 and I was burning a dvd, but I had to reboot the machine because the process got stuck, I could not kill 09 the process and it locked the dvd drive It would be interesting to know where your kernel source comes from. Also, please give us the output of these commands: grep BLK_DEV_IDE .config grep CONFIG_S\?ATA .config BTW, I have both BLK_DEV_IDE and BLK_DEV_IDEDISK set to y. Trying 2.6.26.6. BTW 'uname -a' does not give that information: Linux debian 2.6.26-1-686 #1 SMP Wed Sep 10 16:46:13 UTC 2008 i686 GNU/Linux but if you subscribe to: http://klive.cpushare.com/ you get that info, but apparently few Debian users do because I generally only see myself. Other distros have a lot more subscribers. Hugo -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Re: Problem compiling 2.6.26 kernel
On Sep 28, 2008, at 2:23 AM, debian-user-digest- [EMAIL PROTECTED] wrote From: Mumia W.. [EMAIL PROTECTED] Date: September 28, 2008 12:28:43 AM PDT To: Debian User List debian-user@lists.debian.org Subject: Re: Problem compiling 2.6.26 kernel On 09/27/2008 10:42 PM, Arthur Barlow wrote: I have used Debian for many years and I have always used the kernel- package program that is included with Debian to compile new kernels. I have an older Athlon PC that has the Reiserfs 3.6 on it. After I do the make-kpkg --revision= ## kernel-image and then use dpkg -i to install the new kernel, I keep getting a kernel panic. The message reads like this: VFS: Cannot open rootdevice 301 or unknown-block(3,1) Please append a correct root= boot option; here are available partitions: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(3,1) It appears to have trouble mounting the file systems. I've never run into this before. The boot manager is still lilo and there is a root=/dev/hda1 statement in the lilo.conf file. Any suggestions would be appreciated. Thanks. You seem to be the second person in two days who has complained about this. Has kernel.org changed how IDE devices are accessed? For the time being, you can probably boot using a root of /dev/sda1. The problem is probably that the kernel IDE driver (CONFIG_BLK_DEV_IDE) is not being loaded before libata (CONFIG_ATA). I'm running 2.6.26.5 with a config file that I've used for about a year; I generally get my kernels from kernel.org. It would be interesting to know where your kernel source comes from. Also, please give us the output of these commands: grep BLK_DEV_IDE .config grep CONFIG_S\?ATA .config BTW, I have both BLK_DEV_IDE and BLK_DEV_IDEDISK set to y. I did check the settings in .config. Both BLK_DEV_IDE and BLK_DEV_IDEDISK are set to y. As for the ATA/SATA settings, they were not set, but my disk is an older one and should not care, but I tried setting them and trying again, and I got the same result. The drive is a Western Digital 40G drive, model WDC WD400JB-00FMA0. As I mentioned it's using Reiserfs 3.6. Also I got the source code from Debian's repository. It's just a Debianized tar ball.
Problem compiling 2.6.26 kernel
I have used Debian for many years and I have always used the kernel- package program that is included with Debian to compile new kernels. I have an older Athlon PC that has the Reiserfs 3.6 on it. After I do the make-kpkg --revision= ## kernel-image and then use dpkg -i to install the new kernel, I keep getting a kernel panic. The message reads like this: VFS: Cannot open rootdevice 301 or unknown-block(3,1) Please append a correct root= boot option; here are available partitions: Kernel panic - not syncing: VFS: Unable to mount root fs on unknown- block(3,1) It appears to have trouble mounting the file systems. I've never run into this before. The boot manager is still lilo and there is a root=/dev/hda1 statement in the lilo.conf file. Any suggestions would be appreciated. Thanks. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
after compiling the kernel 2.6.18 bash error message no X Window
Hello! My system is Debian Etch. After I downloaded the new linux-source-2.6.18, compiled it and installed the new custom kernel, I get error messages on console prompt when logged in: -bash: id: command not found -bash: [: : integer expression expected and after gdm login I get error message that there are errors on installation or my disk is full. My disk isn't full, I use df -H to see whether my disk is full. In .xsession-errors are these lines: /etc/gdm/PreSession/Default: \ Registering your session with wtmp and utmp /etc/gdm/PreSession/Default: running: /usr/bin/sessreg -a -w /var/log/wtmp -u /var/run/utmp -x \ /var/lib/gdm/:0.Xservers -h -l :0 csanyipal /etc/gdm/Xsession: Beginning session setup... /etc/gdm/Xsession: line 165: grep: command not found /etc/gdm/Xsession: line 73: ls: command not found /etc/gdm/Xsession: Executing /usr/bin/wmaker failed, \ will try to run x-terminal-emulator /etc/gdm/Xsession: line 197: exec: x-terminal-emulator: not found What is the solution for this problem? I appreciate any advices! -- Regards, Paul Csányi http://www.freewebs.com/csanyi-pal/index.htm -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Compiling a Nvidia module [was Re: Compiling a kernel]
On Mon, Oct 23, 2006 at 11:31:46PM -0500, cothrige wrote: With this install of Debian I decided to stick to what I know, and grabbed the binary installer direct from NVidia. I ran that, and in less than two minutes I was up and running. No complaints from Debian and no complaints from NVidia. Oh, except for one because I had tried to compile it before I updated the system fully and so I had a gcc which didn't match the kernel, but that was fixed in no time. Apparently the binary installer from NVidia messes with the libraries on the system and is not the recommended method for installing. Read http://home.comcast.net/~andrex/Debian-nVidia/ The Debian way is certainly a lot easier. Now where has that nvidia-glx package gone? -- Chris. == ... the official version cannot be abandoned because the implication of rejecting it is far too disturbing: that we are subject to a government conspiracy of `X-Files' proportions and insidiousness. Letter to the LA Times Magazine, September 18, 2005. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a Nvidia module [was Re: Compiling a kernel]
* Chris Bannister ([EMAIL PROTECTED]) wrote: Apparently the binary installer from NVidia messes with the libraries on the system and is not the recommended method for installing. Read http://home.comcast.net/~andrex/Debian-nVidia/ The Debian way is certainly a lot easier. Now where has that nvidia-glx package gone? Well, that is exactly the problem. I tried several times initially to install it via apt but it could never do it. IIRC, there was no nvidia-glx package and so it would not install. This was a test of an installation BTW which ended up being only temporary. When I finally committed and gave over my main partition to Debian I decided to simply try it the old fashioned way and see how it went. So far so good, though I expect that may change at any time. Many thanks for the howto though. The next time I compile a kernel I will try it the Debian way, just to see how I perhaps messed up the first time. Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a Nvidia module [was Re: Compiling a kernel]
On Wednesday 25 October 2006 13:29, Chris Bannister wrote: On Mon, Oct 23, 2006 at 11:31:46PM -0500, cothrige wrote: With this install of Debian I decided to stick to what I know, and grabbed the binary installer direct from NVidia. I ran that, and in less than two minutes I was up and running. No complaints from Debian and no complaints from NVidia. Oh, except for one because I had tried to compile it before I updated the system fully and so I had a gcc which didn't match the kernel, but that was fixed in no time. Apparently the binary installer from NVidia messes with the libraries on the system and is not the recommended method for installing. Read http://home.comcast.net/~andrex/Debian-nVidia/ The Debian way is certainly a lot easier. Now where has that nvidia-glx package gone? Nvidia's own gives me twice the frame rate of Debian's nvidia-glx (Sid). Nvidia is looking for some Xorg SDK to tell it where to place its module. Not finding that, it uses the default /usr/lib which means interference with/by xorg glx packages. Each time you put in new Xorg stuff, simply rerun their installer. A bit of a bother but as you said, a few minutes. Examin their advanced options. There may be a way to tell it where to install the glx stuff. Then simply have xorg.conf check modules there first :-) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a Nvidia module [was Re: Compiling a kernel]
On Wed October 25 2006 06:39, David Baron wrote: On Wednesday 25 October 2006 13:29, Chris Bannister wrote: On Mon, Oct 23, 2006 at 11:31:46PM -0500, cothrige wrote: With this install of Debian I decided to stick to what I know, and grabbed the binary installer direct from NVidia. I ran that, and in less than two minutes I was up and running. No complaints from Debian and no complaints from NVidia. Oh, except for one because I had tried to compile it before I updated the system fully and so I had a gcc which didn't match the kernel, but that was fixed in no time. Apparently the binary installer from NVidia messes with the libraries on the system and is not the recommended method for installing. Read http://home.comcast.net/~andrex/Debian-nVidia/ The Debian way is certainly a lot easier. Now where has that nvidia-glx package gone? Nvidia's own gives me twice the frame rate of Debian's nvidia-glx (Sid). Nvidia is looking for some Xorg SDK to tell it where to place its module. Not finding that, it uses the default /usr/lib which means interference with/by xorg glx packages. Each time you put in new Xorg stuff, simply rerun their installer. A bit of a bother but as you said, a few minutes. Examin their advanced options. There may be a way to tell it where to install the glx stuff. Then simply have xorg.conf check modules there first :-) When I (cough) installed nvidia's own installer I installed xorg-dev first so it could figure out where to put stuff. I'm running with an m-a installed module now and am having good results. Wish I had checked the frame rate but it seems to be working well in any case. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
[EMAIL PROTECTED] wrote: You have gotten a couple DIFFERENT approaches to installing a kernel on Debian. At least one comment should send up a warning: Yes, a level-minded user. On compiling with --initrd, I finally drank the coolade last year. Before I tried to have no modules, compiling needed modules into the kernel itself. Since everyone now compiles oodles of modules, mostly uneeded, but, unknown uneeded, instructions now usually talk about modules. As others in this thread have mentioned. I just finally succumbed to the Jim Jones thing in the last couple of days after years of building non-initrd kernels on Debian due to a strong dislike of devfs. I do not know if it is intentional on the part of the Debian development team to force initrd on the population, but, I have found it quite difficult, and very time consuming, trying to build and boot a non- initrd kernel from the 2.6.1x series. No need for the rtfm/google is your friend' stuff I have most of the current Debian docs from packages, self edited OpenOffice.org libraries, and web picked pdf's. All I wanted was to locally build a trusted 2.6.18 series kernel with CITI_NFS4 patches. After several build iterations on a -rc6 source tree. The one that finally booted without a 'kernel panic: vfs: error root fileysystem not found' error, was a 'make-kpkg --initrd' build. Per usual, the alsa system still won't greet me when KDE starts, though all the modules are loaded, but this thing runs as a kerberos slave with our ldap database replicated to it. So sound sucks, and I guess I don't need it (more like, I don't have time to screw with it). I guess while I'm ranting; Linux was promoted years ago as being able to run on outdated hardware. My budget does not allow me to purchase the latest hardware pushed down to us by Intel, AMD, and MS, every millisecond. But it sure seems like Linux (all distributions) and the 'BSD's for that matter, have developed a preference for recent (within the last three years) hardware. Maybe I'm confusing 'able' and 'useable'. Oh well, I think there might be something for me in 'man kernel-img.conf'. Michael -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
John O'Hagan wrote: On Sunday 22 October 2006 18:02, cothrige wrote: [...] In the past, as a Slackware user, I never installed an OS where I didn't immediately compile a new kernel. Slack uses a 2.4 kernel, and I use some peripheral items which seem to require, or at least greatly prefer a 2.6 kernel. The process I used was very simple, and I got quite used to it. I downloaded the sources from www.kernel.org and opened them up in /usr/src/. I then would run 'make menuconfig', 'make' and 'make modules_install.' I copied the bzImage into /boot, as well as the System.map and config file. I edited lilo.conf, ran /sbin/lilo and rebooted into the new kernel. All usually went well and I rarely had to look back. [...] Or, is there maybe a Debian tool to compile a kernel which is intended to be used rather than this classic method? While things seem fine with the kernel installed from apt, better than fine actually, I figure the day is going to come when I will need to compile a new kernel, and I would like to know if possible what to expect. Not to mention just plain how to do it. [...] Hi Patrick, I always compile my own kernels the Debian (testing) way like this: -Install the latest Debian linux-source package (currently linux-source-2.6.17); or you can use vanilla source as you describe Can't find linux-source using apt, may be kernel-source? I've found kernel-source-2.6.11. Is it the newest? -Make a symlink /usr/src/linux to the resulting folder /usr/src/linux-source-2.6.17 (is this step still necessary?) -Configure the kernel -In /usr/src/linux, run make-kpkg buildpackage kernel-image (there are other options, in the man page) -Install the resulting .deb packages in /usr/src with dpkg -i - Reboot into your new kernel This will build a kernel without an initrd, so you must compile in all drivers for the the boot disk(s). Or use the --initrd option. If you are recompiling a kernel with the same version name, you must move /lib/modules/[$KERNEL_VERSION] out of the way (you are warned if you forget!), or you can use the --revision or --append-to-version options to avoid this. HTH, John -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
(which has more thought even than these email-lists); eg, Compiling a New Kernel at http://www.debian.org/releases/stable/i386/ch08s05.html.en which corresponds to www.debian.org -- installation manual -- Jameson C. Burt, NJ9L Fairfax, Virginia, USA [EMAIL PROTECTED] http://www.coost.com (202) 690-0380 (work) LTSP.org: magic mysterious and awe-inspiring even though we know they are real and not supernatural -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
I look here when I compile my own kernel: http://kernel-handbook.alioth.debian.org/ /David. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
, I noticed that it must have overwitten a few Debian installed files. So when I upgrade Debian, I am prepared (but haven't seen) for possible graphics oddities, in which case I would rerun nvidia.com's executable to re-ovewrite those files. Exactly what I have planned to do in the case there is a problem. But, so far, there hasn't been one. And perhaps that will remain the case. You have heard from many of us, some complementing others information, some giving non-Debian approaches, and you don't really know who to listen to. YOU NEED TO GO TO A DEBIAN DOCUMENT (which has more thought even than these email-lists); eg, Compiling a New Kernel at http://www.debian.org/releases/stable/i386/ch08s05.html.en which corresponds to www.debian.org -- installation manual Many thanks for the info and the links. I will go and see what they say, for future reference. I am sure that there are immense loads of info out there, and I have tried to glean what I can before troubling the list. Though, I do have to compliment this list in general for being one of the most generous I have ever had the good fortune to interact with. The reputation which preceded Debian users was that of an arrogant lot with keyboard shortcuts to type RTFM with one stroke. But, this has been the opposite of my experience reading and posting here. If this list is any indication of the Debian user community in general, it is certainly one of the finest. Many thanks again for all the help, and I will be sure to read the info at the links you included. Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Compiling a kernel
I am sure this is a really stupid question, but having read through the reference and searched online (some searches involve such common terms they never return anything useful) I have really been unable to find a clear answer. I hope someone here can help. In the past, as a Slackware user, I never installed an OS where I didn't immediately compile a new kernel. Slack uses a 2.4 kernel, and I use some peripheral items which seem to require, or at least greatly prefer a 2.6 kernel. The process I used was very simple, and I got quite used to it. I downloaded the sources from www.kernel.org and opened them up in /usr/src/. I then would run 'make menuconfig', 'make' and 'make modules_install.' I copied the bzImage into /boot, as well as the System.map and config file. I edited lilo.conf, ran /sbin/lilo and rebooted into the new kernel. All usually went well and I rarely had to look back. However, I cannot find out if this will work in Debian. (I am using grub so obviously the lilo thing would not) The entire system behind everything seems so much more detailed and complex than Slack that I have my doubts this will work at all. Do I have to use kernel sources from Debian? And will this completely throw off dependency situations in apt? Or, is there maybe a Debian tool to compile a kernel which is intended to be used rather than this classic method? While things seem fine with the kernel installed from apt, better than fine actually, I figure the day is going to come when I will need to compile a new kernel, and I would like to know if possible what to expect. Not to mention just plain how to do it. Many thanks in advance, Patrick
Re: Compiling a kernel
cothrige [EMAIL PROTECTED] wrote: I am sure this is a really stupid question, but having read through the reference and searched online (some searches involve such common terms they never return anything useful) I have really been unable to find a clear answer. I hope someone here can help. I never compiled on Debian, but I watched many threads so I hope I can give you a few pointers. The Debian Way (tm) to compile a kernel is using make-kpkg. This will create a .deb which you can install using 'dpkg -i'. I haven't done this myself, but the archives of this list contain many examples. AFAICT this can be done with sources from kernel.org as well. The default Debian kernel needs an initrd because it has many things compiled as modules. You might be able to avoid the initrd if you compile most (or everything) in. YMMV You can compile a kernel from scratch using the classic method. After you copy the image in /boot just run update-grub and it will add it to your boot menu. You might want to read the notes in /boot/grub/menu.lst if you want/need to add boot options. HTH, Andrei -- If you can't explain it simply, you don't understand it well enough. (Albert Einstein) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
For more than a year I compile my kernels the way you described (universal vay) and I have no problems. Of course there is a debian way but it's not a must. Regards, Seweryn -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
Patrick, Its relatively easy .. and you can make it a bit easier on yourself. Untar from kernel.org in /usr/src be sure ncurses-dev and ncurses are present make menuconfig and configure your kernel now make (or make -j xx, where xx = # of cpu's if 1) [ fancy gcc hacks go here if your brave ] if modular - make modules_install make install (will copy vmlinuz and friends to /boot) I typically use grub, however I think it will update lilo for you if present and in use. Then make your initrd if needed and tweak as needed, verify /etc/modules is what you want it to be and you should be good to go. Cross your fingers and reboot. Really very little difference at all. make help is also very helpful. HTH Best, -Tim On Sun, 2006-10-22 at 03:02 -0500, cothrige wrote: I am sure this is a really stupid question, but having read through the reference and searched online (some searches involve such common terms they never return anything useful) I have really been unable to find a clear answer. I hope someone here can help. In the past, as a Slackware user, I never installed an OS where I didn't immediately compile a new kernel. Slack uses a 2.4 kernel, and I use some peripheral items which seem to require, or at least greatly prefer a 2.6 kernel. The process I used was very simple, and I got quite used to it. I downloaded the sources from www.kernel.org and opened them up in /usr/src/. I then would run 'make menuconfig', 'make' and 'make modules_install.' I copied the bzImage into /boot, as well as the System.map and config file. I edited lilo.conf, ran /sbin/lilo and rebooted into the new kernel. All usually went well and I rarely had to look back. However, I cannot find out if this will work in Debian. (I am using grub so obviously the lilo thing would not) The entire system behind everything seems so much more detailed and complex than Slack that I have my doubts this will work at all. Do I have to use kernel sources from Debian? And will this completely throw off dependency situations in apt? Or, is there maybe a Debian tool to compile a kernel which is intended to be used rather than this classic method? While things seem fine with the kernel installed from apt, better than fine actually, I figure the day is going to come when I will need to compile a new kernel, and I would like to know if possible what to expect. Not to mention just plain how to do it. Many thanks in advance, Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
On Sunday 22 October 2006 18:02, cothrige wrote: [...] In the past, as a Slackware user, I never installed an OS where I didn't immediately compile a new kernel. Slack uses a 2.4 kernel, and I use some peripheral items which seem to require, or at least greatly prefer a 2.6 kernel. The process I used was very simple, and I got quite used to it. I downloaded the sources from www.kernel.org and opened them up in /usr/src/. I then would run 'make menuconfig', 'make' and 'make modules_install.' I copied the bzImage into /boot, as well as the System.map and config file. I edited lilo.conf, ran /sbin/lilo and rebooted into the new kernel. All usually went well and I rarely had to look back. [...] Or, is there maybe a Debian tool to compile a kernel which is intended to be used rather than this classic method? While things seem fine with the kernel installed from apt, better than fine actually, I figure the day is going to come when I will need to compile a new kernel, and I would like to know if possible what to expect. Not to mention just plain how to do it. [...] Hi Patrick, I always compile my own kernels the Debian (testing) way like this: -Install the latest Debian linux-source package (currently linux-source-2.6.17); or you can use vanilla source as you describe -Make a symlink /usr/src/linux to the resulting folder /usr/src/linux-source-2.6.17 (is this step still necessary?) -Configure the kernel -In /usr/src/linux, run make-kpkg buildpackage kernel-image (there are other options, in the man page) -Install the resulting .deb packages in /usr/src with dpkg -i - Reboot into your new kernel This will build a kernel without an initrd, so you must compile in all drivers for the the boot disk(s). Or use the --initrd option. If you are recompiling a kernel with the same version name, you must move /lib/modules/[$KERNEL_VERSION] out of the way (you are warned if you forget!), or you can use the --revision or --append-to-version options to avoid this. HTH, John -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
* John O'Hagan ([EMAIL PROTECTED]) wrote: Hi Patrick, Hello John, I always compile my own kernels the Debian (testing) way like this: -Install the latest Debian linux-source package (currently linux-source-2.6.17); or you can use vanilla source as you describe -Make a symlink /usr/src/linux to the resulting folder /usr/src/linux-source-2.6.17 (is this step still necessary?) I have wondered this too. I always do it, but it seems like I read something somwhere suggesting that it is not needed. -Configure the kernel -In /usr/src/linux, run make-kpkg buildpackage kernel-image (there are other options, in the man page) -Install the resulting .deb packages in /usr/src with dpkg -i Will I still have to configure grub? And will update-grub work or will I have to manually edit menu.lst? I am still getting the hang of grub. It was never a part of Slackware and so I had become very comfortable with lilo. - Reboot into your new kernel This will build a kernel without an initrd, so you must compile in all drivers for the the boot disk(s). Or use the --initrd option. I prefer to go without initrd myself. Just seems a tad faster, though probably my imagination. If you are recompiling a kernel with the same version name, you must move /lib/modules/[$KERNEL_VERSION] out of the way (you are warned if you forget!), or you can use the --revision or --append-to-version options to avoid this. This is new to me. Will these flags stop clobbering of my modules? If so, that is a very cool feature. I forget to move my modules at least half of the time and it really stinks. I think it is because in Slack since I was going from a 2.4 to a 2.6 it never seemed to matter. I am going to do some reading on this. Many thanks for the help. Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
* Tim Post ([EMAIL PROTECTED]) wrote: Hello Tim, [snip] Then make your initrd if needed and tweak as needed, verify /etc/modules is what you want it to be and you should be good to go. Cross your fingers and reboot. This makes me think. Recently I have gotten in the habit, after installing the kernel and all the associated parts, of running generate-modprobe.conf to get all the modules covered. But, it seems that there is no modprobe.conf in Debian, and so I wonder if there is a similar tool or method to generate the information covering the modules? Before I ran this I would have to really hunt around to cover the alsa stuff and so on, and so it was a nice shortcut. Thanks a lot, Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 10/22/06 09:36, cothrige wrote: * John O'Hagan ([EMAIL PROTECTED]) wrote: Hi Patrick, Hello John, [snip] If you are recompiling a kernel with the same version name, you must move /lib/modules/[$KERNEL_VERSION] out of the way (you are warned if you forget!), or you can use the --revision or --append-to-version options to avoid this. Or edit the CONFIG_LOCALVERSION line of /usr/src/linux/.config. This is new to me. Will these flags stop clobbering of my modules? Yes. If so, that is a very cool feature. I forget to move my modules at least half of the time and it really stinks. I think it is because in Slack since I was going from a 2.4 to a 2.6 it never seemed to matter. I am going to do some reading on this. - -- Ron Johnson, Jr. Jefferson LA USA Is common sense really valid? For example, it is common sense to white-power racists that whites are superior to blacks, and that those with brown skins are mud people. However, that common sense is obviously wrong. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.5 (GNU/Linux) iD8DBQFFO4QjS9HxQb37XmcRApgKAJ9JxmTNiMF3qWjIyWqYVVhysdnHqQCgzFzB Gg12/+mviMDqkuNrTpKfTOw= =nrBp -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
Le dimanche 22 octobre 2006 16:43, cothrige a écrit : * Tim Post ([EMAIL PROTECTED]) wrote: Hello Tim, [snip] Then make your initrd if needed and tweak as needed, verify /etc/modules is what you want it to be and you should be good to go. Cross your fingers and reboot. This makes me think. Recently I have gotten in the habit, after installing the kernel and all the associated parts, of running generate-modprobe.conf to get all the modules covered. But, it seems that there is no modprobe.conf in Debian, and so I wonder if there is a similar tool or method to generate the information covering the modules? Before I ran this I would have to really hunt around to cover the alsa stuff and so on, and so it was a nice shortcut. This modprobe.conf is modularized in several files (you can add one) in /etc/modprobe.d/. pgpsn3q6uXsut.pgp Description: PGP signature
Re: Compiling a kernel
On Monday 23 October 2006 00:36, cothrige wrote: * John O'Hagan ([EMAIL PROTECTED]) wrote: [snip summary of Debian kernel compilation] Will I still have to configure grub? And will update-grub work or will I have to manually edit menu.lst? [...] Installing the kernel-package generated by make-kpkg will automatically detect and update grub, and add itself to menu.list. How easy is that? Regards, John -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
depmod should be called by the makefile upon make modules_install after a successful build. Its really as easy as make, make modules_install , make install and a mkinitrd (if you need one). If using GRUB, remember by default the selection menu is hidden. You'll need to comment out the hiddenmenu line, and of course turn on pretty colors. /etc/modules is a newline delimited list of modules that are loaded at boot time. You can further tweak how modules are loaded by modifying your initrd directly. You'll need to do this if you use things like AoE/iSCSI to access the root file system you pass to your kernel (i.e. root=/dev/etherd/e3.0) or any service that loads and unloads its own modules to start / stop but must be on in order to boot. You shouldn't need to unless you have a custom / unique setup, but here's how anyway - Just mount it : mkdir /tmp/initrd-mnt modprobe loop mount -o loop -t cramfs /boot/my.initrd-img.x.x.x /tmp/initrd-mnt Then simply pico the file named linuxrc, or edit the file containing the list of modules. You'll see the modprobe lines loading modules, feel free to add or detract :) This is really handy to get things like AoE/iSCSI or cluster file systems going prior to mounting the root fs or any other odd boot needs. Just remember, linuxrc isn't smart, its mission is only to load modules and pivot over to the root fs you specify... so make sure you add some logic to deal with mod load failures if they happen (if you tinker, that is). Kernel hacking is quite a bit easier on GNU distributions (like Debian) once you get used to the differences from whitebox / gentoo. Not saying its not fun on those flavors, I just happen to really prefer GNU. Many more comprehensive step by step kernel building tutorials are easily found via Google .. this should be enough to get you going with a degree of confidence however :) HTH Best, -Tim On Sun, 2006-10-22 at 09:43 -0500, cothrige wrote: * Tim Post ([EMAIL PROTECTED]) wrote: Hello Tim, [snip] Then make your initrd if needed and tweak as needed, verify /etc/modules is what you want it to be and you should be good to go. Cross your fingers and reboot. This makes me think. Recently I have gotten in the habit, after installing the kernel and all the associated parts, of running generate-modprobe.conf to get all the modules covered. But, it seems that there is no modprobe.conf in Debian, and so I wonder if there is a similar tool or method to generate the information covering the modules? Before I ran this I would have to really hunt around to cover the alsa stuff and so on, and so it was a nice shortcut. Thanks a lot, Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
* Tim Post ([EMAIL PROTECTED]) wrote: depmod should be called by the makefile upon make modules_install after a successful build. Its really as easy as make, make modules_install , make install and a mkinitrd (if you need one). If using GRUB, remember by default the selection menu is hidden. You'll need to comment out the hiddenmenu line, and of course turn on pretty colors. Okay, a touch of confusion on this. I am unfamiliar with grub, only having started using it with this install of Debian and that having been setup by the Debian installer. Right now the menu is not hidden, but are you saying that a 'make install' of the kernel will add hiddenmenu to my configuration? BTW, I have never used make install, kinda scary I suppose, and so don't know much about it. Probably won't use it now either, unless it is highly recommended. /etc/modules is a newline delimited list of modules that are loaded at boot time. What is the difference between this and the modprobe.d files? I notice that my system has both (alsa, nvidia and such in modprobe.d and loop is all that is in modules). The man pages didn't seem to clarify why both are needed. You can further tweak how modules are loaded by modifying your initrd directly. [snip] I have never used initrd, at least not when I have compiled a kernel. To be entirely honest I have never fully understood just what it does. I was under the impression it was for things like booting from reiser fs and having to load modules to do it. However, that always made me wonder why a person would just not compile in the reiser support, and so I have naturally assumed that I was dead wrong on why it existed and what it accomplished. But, since I have never seemed to need it, i.e. my system has always booted fairly predictably without it, I never tried to learn more. Kernel hacking is quite a bit easier on GNU distributions (like Debian) once you get used to the differences from whitebox / gentoo. Not saying its not fun on those flavors, I just happen to really prefer GNU. I am going to open my mouth and prove my total ignorance, as I have never even approached any kernel hacking. But, what is a 'whitebox' and how is it different from GNU? I would have thought that Gentoo was a GNU distro. Many thanks, Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
* Gilles Mocellin ([EMAIL PROTECTED]) wrote: This modprobe.conf is modularized in several files (you can add one) in /etc/modprobe.d/. Ahh yes, I see that. I would think I could run 'generate-modprobe.conf ~/modprobe.conf' and then split the info up as I need it. Shouldn't be too impossible I would think. Thanks, Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
* John O'Hagan ([EMAIL PROTECTED]) wrote: Installing the kernel-package generated by make-kpkg will automatically detect and update grub, and add itself to menu.list. How easy is that? Now that it is a nifty feature. I suppose there is certainly something to be said for the Debian approach to this after all. Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
cothrige [EMAIL PROTECTED] wrote: I have never used initrd, at least not when I have compiled a kernel. To be entirely honest I have never fully understood just what it does. I was under the impression it was for things like booting from reiser fs and having to load modules to do it. However, that always made me wonder why a person would just not compile in the reiser support, and so I have naturally assumed that I was dead wrong on why it existed and what it accomplished. But, since I have never seemed to need it, i.e. my system has always booted fairly predictably without it, I never tried to learn more. initrd's are especially useful for distros, because a kernel with all stuff compiled in is not an option (too big), but you still need some of the modules very early in the boot process, when the root filesystem is not accessible yet. For your particular system you can afford to compile everything in, because you know exactly what you need and what you don't. Regards, Andrei -- If you can't explain it simply, you don't understand it well enough. (Albert Einstein) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
cothrige wrote: * Tim Post ([EMAIL PROTECTED]) wrote: depmod should be called by the makefile upon make modules_install after a successful build. Its really as easy as make, make modules_install , make install and a mkinitrd (if you need one). If using GRUB, remember by default the selection menu is hidden. You'll need to comment out the hiddenmenu line, and of course turn on pretty colors. Okay, a touch of confusion on this. I am unfamiliar with grub, only having started using it with this install of Debian and that having been setup by the Debian installer. Right now the menu is not hidden, but are you saying that a 'make install' of the kernel will add hiddenmenu to my configuration? BTW, I have never used make install, kinda scary I suppose, and so don't know much about it. Probably won't use it now either, unless it is highly recommended. /etc/modules is a newline delimited list of modules that are loaded at boot time. What is the difference between this and the modprobe.d files? I notice that my system has both (alsa, nvidia and such in modprobe.d and loop is all that is in modules). The man pages didn't seem to clarify why both are needed. You can further tweak how modules are loaded by modifying your initrd directly. [snip] I have never used initrd, at least not when I have compiled a kernel. To be entirely honest I have never fully understood just what it does. I was under the impression it was for things like booting from reiser fs and having to load modules to do it. However, that always made me wonder why a person would just not compile in the reiser support, and so I have naturally assumed that I was dead wrong on why it existed and what it accomplished. But, since I have never seemed to need it, i.e. my system has always booted fairly predictably without it, I never tried to learn more. Kernel hacking is quite a bit easier on GNU distributions (like Debian) once you get used to the differences from whitebox / gentoo. Not saying its not fun on those flavors, I just happen to really prefer GNU. I am going to open my mouth and prove my total ignorance, as I have never even approached any kernel hacking. But, what is a 'whitebox' and how is it different from GNU? I would have thought that Gentoo was a GNU distro. Many thanks, Patrick Patrick, Not just reiserfs but say if your hard drive type was a modual (ide, sata, etc). Yes it is cleaner perhaps to hard code it in, but on the other hand, if you miss one item in the chain, you get kernel panic. I would suggest you install kernel-package and read the README in /usr/share/doc/kernel-package . It has a rather complete walk through on how to do it the debian way, including the initrd. Make-kpkg will in fact make all the moduals, add them, and the kernel and make a .deb you then use dpkg -i kernelname.deb to install, and that will modify grub for you (or lilo I suspose). Comes in real handy if you have more then one box with the same hardware, just transfer the .deb and install it. Plus now your system knows it is installed and if you use the make-kpgk --append-to=customename_or_number_here then apt will not replace your kernel (unless you happen to name it something that matches a apt-getable kernel. Real nice. I hope that helps. -- Damon L. Chesser [EMAIL PROTECTED] [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Compiling a kernel
* Andrei Popescu ([EMAIL PROTECTED]) wrote: initrd's are especially useful for distros, because a kernel with all stuff compiled in is not an option (too big), but you still need some of the modules very early in the boot process, when the root filesystem is not accessible yet. For your particular system you can afford to compile everything in, because you know exactly what you need and what you don't. That does make sense. I can see how it would make sense in that situation. It should have occurred to me, but I suppose I was just not taking the time to look at it from all sides. Thanks for the clarification. Patrick -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Selecting processor type when compiling the kernel
Hello, I have a laptop with Celeron M processor within. My question is: what processor type should I select when compiling the kernel? Pentium M even though it's just Celeron M?
Re: compiling sid kernel source 2.6.15 in sarge
Hiya LV, I have been using 2.6.15 for about a week now. It's fantastic. Seems to be faster than 2.6.8. I have scanning, sound, USB, cd burning, rsync, nfs. All works well. I used the kernel from kernel.org. I didn't bother with the ramdisk. Seems to be a waste of time. I followed the instructions from here: http://newbiedoc.sourceforge.net/tutorials/kernel-pkg/intro-kernel-pkg.html.en Not sure why you cannot compile. Try what I did and let us know how you get on. HTH -- Glenn Meehan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: compiling sid kernel source 2.6.15 in sarge
On 3/1/06, Glenn Meehan [EMAIL PROTECTED] wrote: I have been using 2.6.15 for about a week now. It's fantastic. Seems to be faster than 2.6.8. I have scanning, sound, USB, cd burning, rsync, nfs. All works well. I used the kernel from kernel.org. I didn't bother with the ramdisk. Seems to be a waste of time. I followed the instructions from here: http://newbiedoc.sourceforge.net/tutorials/kernel-pkg/intro-kernel-pkg.html.en Not sure why you cannot compile. Try what I did and let us know how you get on. I have compiled and could run vanilla kernel 2.6.15.4. But Problem ocuurred when I tried to compile ndiswrapper module. I installed ndiswrapper source from sarge debs. /usr/src/modules/ndiswrapper/wrapper.c:286:53: macro halt passed 1 arguments, but takes just 0 /usr/src/modules/ndiswrapper/wrapper.c: In function `miniport_halt': /usr/src/modules/ndiswrapper/wrapper.c:286: warning: statement with no effect make[5]: *** [/usr/src/modules/ndiswrapper/wrapper.o] Error 1 make[4]: *** [_module_/usr/src/modules/ndiswrapper] Error 2 make[4]: Leaving directory `/usr/src/linux-2.6.15.4' make[3]: *** [default] Error 2 make[3]: Leaving directory `/usr/src/modules/ndiswrapper' make[2]: *** [binary-modules] Error 2 make[2]: Leaving directory `/usr/src/modules/ndiswrapper' make[1]: *** [kdist_build] Error 2 make[1]: Leaving directory `/usr/src/modules/ndiswrapper' Module /usr/src/modules/ndiswrapper failed. Hit return to Continue kernel is kernel-image-2.6.15.4.20060301 and kernel-source is in /usr/src/linux-2.6.15.4 and kernel headers in /usr/src/kernel-headers-2.6.15.4.20060301 and ndiswrapper source in /usr/src/modules/ndiswrapper. How should I proceed to compile ndiswrapper module? Further if anybody can throw somelight why kernel compiled from u nstable source couldn't boot even? -- L.V.Gandhi http://lvgandhi.tripod.com/ linux user No.205042
compiling sid kernel source 2.6.15 in sarge
I was using sid kernel source from my sid partition. When I was running sarge, I untarred sid kernel source 2.6.15 in /usr/src. Added bootsplash patch did oldconfig using sarge's config file from /boot/.and answered mostly all to defaults. I thought 2.6.15 will take only linux-image, I tried make-kpkg linux-image. It didn't accepted it gave error. Then I did make-kpkg --initrd --append-to-version .$(date +'%Y%m%d') --revision 1 kernel_image kernel_headers modules_image. I had ndiswrapper modules folder in /usr/src/modules. After it made kernel_image kernel_headers, it failed to compile ndiswrapper. Then I removed ndiswrapper source and untarred sid ndiswrapper source. Now it compiled gave all three debs. both kAs sso as it is started, machine reboots. I would like to know why old ndiswrapper is not getting compiled? Why rebooting takes place? What mistake I have done? Are any modification is needed in source before it is compiled in sarge. I didn't use kernel.org source as it may not have debian specific patches. Am I right? -- L.V.Gandhi http://lvgandhi.tripod.com/ linux user No.205042
