Kernel Compile Issue

Kernel Compile Issue

[Aruna Hewapathirane]

Hello everyone,

I am trying to compile a old kernel and it compiles fine. I used:
time make -j$(nproc) and this is a-ok and works.
make modules_install works fine. Now when I try to install the newly
compiled kernel I get this error below.

Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64)

it says Consult /var/lib/dkms/open-vm-tools/9.4.6/build/make.log for
more information.
and this file shows me what is below:

==============================================================
DKMS make.log for open-vm-tools-9.4.6 for kernel 3.18.2-testing (x86_64)
Tue Jun 22 22:30:14 EDT 2021
/bin/sh: 1: test: -I./arch/x86/include: unexpected operator
Using standalone build system.
Makefile:167: Makefile.normal: No such file or directory
make[2]: *** No rule to make target 'Makefile.normal'.  Stop.
/bin/sh: 1: test: -I./arch/x86/include: unexpected operator
Using standalone build system.
Makefile:167: Makefile.normal: No such file or directory
make[2]: *** No rule to make target 'Makefile.normal'.  Stop.
===============================================================

I can use all the help I can get right now. Anyone ever experience
this ? If so is
is there a fix or workaround ? What causes this error ?

I have multiple kernels on this ancient box and would really prefer
not to have to
wipe everything and reinstall from scratch.

The only thing I did recently was install Virtualbox and Wireshark.
Both work no issues.

If anyone has any pointers would be most welcome and much appreciated.

Thanks - Aruna

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Thomas Bracht Laumann Jespersen]

> I am trying to compile a old kernel and it compiles fine. I used:
> time make -j$(nproc) and this is a-ok and works.
> make modules_install works fine. Now when I try to install the newly
> compiled kernel I get this error below.
> 
> Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64)

Just to be clear: This is what you get when running "sudo make install" ?

> it says Consult /var/lib/dkms/open-vm-tools/9.4.6/build/make.log for
> more information.
> and this file shows me what is below:
> 
> ==============================================================
> DKMS make.log for open-vm-tools-9.4.6 for kernel 3.18.2-testing (x86_64)
> Tue Jun 22 22:30:14 EDT 2021
> /bin/sh: 1: test: -I./arch/x86/include: unexpected operator
> Using standalone build system.
> Makefile:167: Makefile.normal: No such file or directory
> make[2]: *** No rule to make target 'Makefile.normal'.  Stop.
> /bin/sh: 1: test: -I./arch/x86/include: unexpected operator
> Using standalone build system.
> Makefile:167: Makefile.normal: No such file or directory
> make[2]: *** No rule to make target 'Makefile.normal'.  Stop.
> ===============================================================

What are the contents of this Makefile (line 167)? Looks like it's expecting a
"Makefile.normal" that it can't find - I don't know what this file is or where
it would come from.

All the best,
-- Thomas

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Aruna Hewapathirane]

On Thu, Jun 24, 2021 at 6:21 AM Thomas Bracht Laumann Jespersen
<t@laumann.xyz> wrote:
>
> > I am trying to compile a old kernel and it compiles fine. I used:
> > time make -j$(nproc) and this is a-ok and works.
> > make modules_install works fine. Now when I try to install the newly
> > compiled kernel I get this error below.
> >
> > Error! Bad return status for module build on kernel: 3.18.2-testing (x86_64)
>
> Just to be clear: This is what you get when running "sudo make install" ?

Yes,  I was getting that error when running "sudo make install". I
have manually removed
all instances of VirtualBox and any guest additions and that warning
went away. But now I have
another problem. After compilation grub menu shows the new kernel.
When I select the new
kernel it goes through all the usual steps of booting then hangs at
the graphical login screen where
I no longer have a mouse and Ctrl+Alt+F1 does not work either .

I have given below the output of: sudo time make install 2>&1 | tee
make_install.txt

==============================================================
sh ./arch/x86/boot/install.sh 3.18.2-testing arch/x86/boot/bzImage \
    System.map "/boot"
run-parts: executing /etc/kernel/postinst.d/apt-auto-removal
3.18.2-testing /boot/vmlinuz-3.18.2-testing
run-parts: executing /etc/kernel/postinst.d/initramfs-tools
3.18.2-testing /boot/vmlinuz-3.18.2-testing
update-initramfs: Generating /boot/initrd.img-3.18.2-testing
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-2.fw for
module r8169
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-1.fw for
module r8169
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-2.fw for
module r8169
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-1.fw for
module r8169
run-parts: executing /etc/kernel/postinst.d/zz-update-grub
3.18.2-testing /boot/vmlinuz-3.18.2-testing
Generating grub configuration file ...
Found background image: /usr/share/images/desktop-base/desktop-grub.png
Found linux image: /boot/vmlinuz-3.18.2-testing
Found initrd image: /boot/initrd.img-3.18.2-testing
Found linux image: /boot/vmlinuz-3.18.2-testing.old
Found initrd image: /boot/initrd.img-3.18.2-testing
Found linux image: /boot/vmlinuz-3.16.0-4-amd64
Found initrd image: /boot/initrd.img-3.16.0-4-amd64
Found memtest86+ image: /boot/memtest86+.bin
Found memtest86+ multiboot image: /boot/memtest86+_multiboot.bin
Found Windows 7 (loader) on /dev/sda1
Found Ubuntu 12.04.5 LTS (12.04) on /dev/sdb1
Found Debian GNU/Linux (6.0.5) on /dev/sdb3
Found Debian GNU/Linux (8.0) on /dev/sdb5
Found Ubuntu 14.04.3 LTS (14.04) on /dev/sdb6
done
1.33user 0.71system 0:29.75elapsed 6%CPU (0avgtext+0avgdata 16520maxresident)k
251456inputs+74824outputs (52major+1632571minor)pagefaults 0swaps

==============================================================

Like I said I have a lot of old kernels lying around this box :-)

> > it says Consult /var/lib/dkms/open-vm-tools/9.4.6/build/make.log for
> > more information.
> > and this file shows me what is below:
> >
> > ==============================================================
> > DKMS make.log for open-vm-tools-9.4.6 for kernel 3.18.2-testing (x86_64)
> > Tue Jun 22 22:30:14 EDT 2021
> > /bin/sh: 1: test: -I./arch/x86/include: unexpected operator
> > Using standalone build system.
> > Makefile:167: Makefile.normal: No such file or directory
> > make[2]: *** No rule to make target 'Makefile.normal'.  Stop.
> > /bin/sh: 1: test: -I./arch/x86/include: unexpected operator
> > Using standalone build system.
> > Makefile:167: Makefile.normal: No such file or directory
> > make[2]: *** No rule to make target 'Makefile.normal'.  Stop.
> > ===============================================================
>
> What are the contents of this Makefile (line 167)? Looks like it's expecting a
> "Makefile.normal" that it can't find - I don't know what this file is or where
> it would come from.

The contents of the Makefile in ./arch/x86/include at line 167:

166 - # prevent gcc from generating any FP code by mistake
167 - KBUILD_CFLAGS += -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
168 - KBUILD_CFLAGS += $(call cc-option,-mno-avx,)

170 - KBUILD_CFLAGS += $(mflags-y)
171 - KBUILD_AFLAGS += $(mflags-y)

> All the best,
> -- Thomas

You have pointed me in the right direction so thank you. Ten plus years
I have been dabbling with Linux and I just learnt something new. I had no idea
about this:

aruna@debian:/media/aruna/kernel-stuff/linux-3.18.2$ ls -alh
/etc/kernel/postinst.d
total 20K
drwxr-xr-x 2 root root 4.0K Jun 23 06:18 .
drwxr-xr-x 7 root root 4.0K Mar 31 00:45 ..
-rwxr-xr-x 1 root root 2.8K Apr 13  2015 apt-auto-removal
-rwxr-xr-x 1 root root  858 Mar  1  2015 initramfs-tools
-rwxr-xr-x 1 root root  641 Dec 14  2015 zz-update-grub

Thank you Thomas I will keep trying to fix the no mouse issue now.

Aruna

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Cindy Sue Causey]

On 6/24/21, Aruna Hewapathirane <aruna.hewapathirane@gmail.com> wrote:
> On Thu, Jun 24, 2021 at 6:21 AM Thomas Bracht Laumann Jespersen
> <t@laumann.xyz> wrote:
>>
>> > I am trying to compile a old kernel and it compiles fine. I used:
>> > time make -j$(nproc) and this is a-ok and works.
>> > make modules_install works fine. Now when I try to install the newly
>> > compiled kernel I get this error below.
>> >
>> > Error! Bad return status for module build on kernel: 3.18.2-testing
>> > (x86_64)
>>
>> Just to be clear: This is what you get when running "sudo make install" ?
>
> Yes,  I was getting that error when running "sudo make install". I
> have manually removed
> all instances of VirtualBox and any guest additions and that warning
> went away. But now I have
> another problem. After compilation grub menu shows the new kernel.
> When I select the new
> kernel it goes through all the usual steps of booting then hangs at
> the graphical login screen where
> I no longer have a mouse and Ctrl+Alt+F1 does not work either .


Hi from a perennial lurker here! Disclaimer up front: My apologies if
you're already doing this step.

In my case, my experience is with Debian doing similar versions of
this over the last year or so. I tripped over that I was mismatching
e.g. my initrd.img and vmlinuz instances in an attempt at separating
out all things "/boot" while running LILO instead of GRUB2 for 14 or
so different bootable partitions.

My stumbled upon fix was to run "update-initramfs -u" then double
check to make sure that that latest update matched the location LILO
was using to call up those partition-specific files. As an example, I
have a "lilo" subdirectory under /boot then each operating system has
its own deeper subdirectory. Those subdirectories hold copies of each
system's initrd.img and vmlinuz.

Ever since I started very carefully double checking everything after
each update that triggered LILO to run, I haven't been locked out
since, not once. Prior to this (really just a few weeks ago), I faced
every possible form of lockout that you're describing plus
occasionally also hitting that blank screen where the login prompt
doesn't even bother to appear. The experience has been so empowering
that I'm preparing to test drive syslinux, too, something I've
actually "feared" until a couple weeks ago. :)

Hope this helps someone at some point... somewhere.. somehow.

Cindy :)
-- 
* runs with birdseed *

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Aruna Hewapathirane]

On Thu, Jun 24, 2021 at 12:13 PM Cindy Sue Causey
<butterflybytes@gmail.com> wrote:
>
> On 6/24/21, Aruna Hewapathirane <aruna.hewapathirane@gmail.com> wrote:
> > On Thu, Jun 24, 2021 at 6:21 AM Thomas Bracht Laumann Jespersen
> > <t@laumann.xyz> wrote:
> >>
> >> > I am trying to compile a old kernel and it compiles fine. I used:
> >> > time make -j$(nproc) and this is a-ok and works.
> >> > make modules_install works fine. Now when I try to install the newly
> >> > compiled kernel I get this error below.
> >> >
> >> > Error! Bad return status for module build on kernel: 3.18.2-testing
> >> > (x86_64)
> >>
> >> Just to be clear: This is what you get when running "sudo make install" ?
> >
> > Yes,  I was getting that error when running "sudo make install". I
> > have manually removed
> > all instances of VirtualBox and any guest additions and that warning
> > went away. But now I have
> > another problem. After compilation grub menu shows the new kernel.
> > When I select the new
> > kernel it goes through all the usual steps of booting then hangs at
> > the graphical login screen where
> > I no longer have a mouse and Ctrl+Alt+F1 does not work either .
>
>
> Hi from a perennial lurker here! Disclaimer up front: My apologies if
> you're already doing this step.

> out all things "/boot" while running LILO instead of GRUB2 for 14 or
> so different bootable partitions.

Hi, I thought 'I' was the only person insane enough to have multiple partitions
that run Linux ? 14 or so partitions ? Yikes ! Let's see how many I have managed
to create over the years :

aruna@debian:~$ lsblk
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda      8:0    0 232.9G  0 disk
├─sda1   8:1    0   100M  0 part
├─sda2   8:2    0 212.2G  0 part /media/aruna/OS
├─sda3   8:3    0     1K  0 part
├─sda5   8:5    0  19.7G  0 part /
└─sda6   8:6    0   903M  0 part [SWAP]
sdb      8:16   0 931.5G  0 disk
├─sdb1   8:17   0   250G  0 part /media/aruna/Ubuntu 12.04.4 L
├─sdb2   8:18   0   250G  0 part /media/aruna/kernel-stuff
├─sdb3   8:19   0  16.8G  0 part /media/aruna/back-up
├─sdb4   8:20   0     1K  0 part
├─sdb5   8:21   0   204G  0 part
/media/aruna/3c8fce17-5948-43e7-9662-921e65c7faa7
├─sdb6   8:22   0 196.9G  0 part
/media/aruna/200f9328-ae6b-4fe9-9d94-24ffd391bbe9
├─sdb7   8:23   0   6.2G  0 part
└─sdb8   8:24   0   7.7G  0 part
sdh      8:112  1   3.7G  0 disk
├─sdh1   8:113  1   2.4G  0 part /media/aruna/d-live 10.10.0 ma amd64
└─sdh2   8:114  1   2.5M  0 part
sr0     11:0    1  1024M  0 rom

Darn! You beat me by one partition but it is a close second eh ? Can't
count the sdh
that has the Debian 10 live CD I just plugged in :-)

> In my case, my experience is with Debian doing similar versions of
> this over the last year or so. I tripped over that I was mismatching
> e.g. my initrd.img and vmlinuz instances in an attempt at separating

Believe me I am feeling the pain you must have felt at that time right now.
And this is really interesting what you have shared, makes me wonder now.

> My stumbled upon fix was to run "update-initramfs -u" then double
> check to make sure that that latest update matched the location LILO
> was using to call up those partition-specific files. As an example, I
> have a "lilo" subdirectory under /boot then each operating system has
> its own deeper subdirectory. Those subdirectories hold copies of each
> system's initrd.img and vmlinuz.

I have always used GRUB2 so now I have to read about "lilo" I guess. Thanks.

> Ever since I started very carefully double checking everything after
> each update that triggered LILO to run, I haven't been locked out
> since, not once. Prior to this (really just a few weeks ago), I faced
> every possible form of lockout that you're describing plus
> occasionally also hitting that blank screen where the login prompt
> doesn't even bother to appear. The experience has been so empowering
> that I'm preparing to test drive syslinux, too, something I've
> actually "feared" until a couple weeks ago. :)

You 'feared' syslinux ? There was a time long years ago I had accidentally
opened up a shell/terminal and I just sat there for a good 30 minutes unsure
what the hell do I do now ? I was too afraid to x-out and close that
window because
I had no idea what it would or would not do. And I had no clue what to
type in that
bash shell either. Those were the good old days :-)

> Hope this helps someone at some point... somewhere.. somehow.

It has most definitely helped me. I will be back with more questions as I
proceed to try and troubleshoot this mess I have created.

Thank you very much for sharing - Aruna






>
> Cindy :)
> --
> * runs with birdseed *
>
> _______________________________________________
> Kernelnewbies mailing list
> Kernelnewbies@kernelnewbies.org
> https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Thomas Bracht Laumann Jespersen]

> Yes,  I was getting that error when running "sudo make install". I
> have manually removed
> all instances of VirtualBox and any guest additions and that warning
> went away.

Interesting...

> When I select the new
> kernel it goes through all the usual steps of booting then hangs at
> the graphical login screen where
> I no longer have a mouse and Ctrl+Alt+F1 does not work either .

Is there a way you could increase the kernel log level to debug, and capture the
logs during boot? That might be informative.

> I have given below the output of: sudo time make install 2>&1 | tee
> make_install.txt
> 
> ==============================================================
> sh ./arch/x86/boot/install.sh 3.18.2-testing arch/x86/boot/bzImage \
>     System.map "/boot"
> run-parts: executing /etc/kernel/postinst.d/apt-auto-removal
> 3.18.2-testing /boot/vmlinuz-3.18.2-testing
> run-parts: executing /etc/kernel/postinst.d/initramfs-tools
> 3.18.2-testing /boot/vmlinuz-3.18.2-testing
> update-initramfs: Generating /boot/initrd.img-3.18.2-testing
> W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-2.fw for
> module r8169
> W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-1.fw for
> module r8169
> W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-2.fw for
> module r8169
> W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-1.fw for
> module r8169
> run-parts: executing /etc/kernel/postinst.d/zz-update-grub
> 3.18.2-testing /boot/vmlinuz-3.18.2-testing
> Generating grub configuration file ...
> Found background image: /usr/share/images/desktop-base/desktop-grub.png
> Found linux image: /boot/vmlinuz-3.18.2-testing
> Found initrd image: /boot/initrd.img-3.18.2-testing
> Found linux image: /boot/vmlinuz-3.18.2-testing.old
> Found initrd image: /boot/initrd.img-3.18.2-testing
> Found linux image: /boot/vmlinuz-3.16.0-4-amd64
> Found initrd image: /boot/initrd.img-3.16.0-4-amd64
> Found memtest86+ image: /boot/memtest86+.bin
> Found memtest86+ multiboot image: /boot/memtest86+_multiboot.bin
> Found Windows 7 (loader) on /dev/sda1
> Found Ubuntu 12.04.5 LTS (12.04) on /dev/sdb1
> Found Debian GNU/Linux (6.0.5) on /dev/sdb3
> Found Debian GNU/Linux (8.0) on /dev/sdb5
> Found Ubuntu 14.04.3 LTS (14.04) on /dev/sdb6
> done
> 1.33user 0.71system 0:29.75elapsed 6%CPU (0avgtext+0avgdata 16520maxresident)k
> 251456inputs+74824outputs (52major+1632571minor)pagefaults 0swaps
> 
> ==============================================================
> 
> Like I said I have a lot of old kernels lying around this box :-)

And Windowses :-) But nothing looks particularly erroneous about the output.

> The contents of the Makefile in ./arch/x86/include at line 167:
> 
> 166 - # prevent gcc from generating any FP code by mistake
> 167 - KBUILD_CFLAGS += -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
> 168 - KBUILD_CFLAGS += $(call cc-option,-mno-avx,)

Interesting, I see no mention of "Makefile.normal" here. But maybe that's
because you removed whatever required it?

> You have pointed me in the right direction so thank you. Ten plus years
> I have been dabbling with Linux and I just learnt something new. I had no idea
> about this:
> 
> aruna@debian:/media/aruna/kernel-stuff/linux-3.18.2$ ls -alh
> /etc/kernel/postinst.d
> total 20K
> drwxr-xr-x 2 root root 4.0K Jun 23 06:18 .
> drwxr-xr-x 7 root root 4.0K Mar 31 00:45 ..
> -rwxr-xr-x 1 root root 2.8K Apr 13  2015 apt-auto-removal
> -rwxr-xr-x 1 root root  858 Mar  1  2015 initramfs-tools
> -rwxr-xr-x 1 root root  641 Dec 14  2015 zz-update-grub
> 
> Thank you Thomas I will keep trying to fix the no mouse issue now.

No problem, I wasn't sure if I was helping :-)

-- Thomas

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Aruna Hewapathirane]

On Thu, Jun 24, 2021 at 1:29 PM Thomas Bracht Laumann Jespersen
<t@laumann.xyz> wrote:
>
> > Yes,  I was getting that error when running "sudo make install". I
> > have manually removed
> > all instances of VirtualBox and any guest additions and that warning
> > went away.
>
> Interesting...
>
Interesting ? Don't forget I compiled and installed via make install, so I  had
to manually remove the following entries:

/boot/vmlinuz*KERNEL-VERSION*
/boot/initrd*KERNEL-VERSION*
/boot/System-map*KERNEL-VERSION*
/boot/config-*KERNEL-VERSION*
/lib/modules/*KERNEL-VERSION*/
/var/lib/initramfs/*KERNEL-VERSION*/

Then update the grub configuration:
sudo update-grub2

Then look for any loose ends using 'locate':
locate -b -e *dkms 3.18.2-*

You get the idea. Was definitely not fun :-)

> > When I select the new
> > kernel it goes through all the usual steps of booting then hangs at
> > the graphical login screen where
> > I no longer have a mouse and Ctrl+Alt+F1 does not work either .
>
> Is there a way you could increase the kernel log level to debug, and capture the
> logs during boot? That might be informative.

Hmm... should be possible. Never needed to do this before so another
learning experience for me here. I will get back to you on this.

> > I have given below the output of: sudo time make install 2>&1 | tee
> > make_install.txt
> >
> > ==============================================================
> > sh ./arch/x86/boot/install.sh 3.18.2-testing arch/x86/boot/bzImage \
> >     System.map "/boot"
> > run-parts: executing /etc/kernel/postinst.d/apt-auto-removal
> > 3.18.2-testing /boot/vmlinuz-3.18.2-testing
> > run-parts: executing /etc/kernel/postinst.d/initramfs-tools
> > 3.18.2-testing /boot/vmlinuz-3.18.2-testing
> > update-initramfs: Generating /boot/initrd.img-3.18.2-testing
> > W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-2.fw for
> > module r8169
> > W: Possible missing firmware /lib/firmware/rtl_nic/rtl8107e-1.fw for
> > module r8169
> > W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-2.fw for
> > module r8169
> > W: Possible missing firmware /lib/firmware/rtl_nic/rtl8168h-1.fw for
> > module r8169
> > run-parts: executing /etc/kernel/postinst.d/zz-update-grub
> > 3.18.2-testing /boot/vmlinuz-3.18.2-testing
> > Generating grub configuration file ...
> > Found background image: /usr/share/images/desktop-base/desktop-grub.png
> > Found linux image: /boot/vmlinuz-3.18.2-testing
> > Found initrd image: /boot/initrd.img-3.18.2-testing
> > Found linux image: /boot/vmlinuz-3.18.2-testing.old
> > Found initrd image: /boot/initrd.img-3.18.2-testing
> > Found linux image: /boot/vmlinuz-3.16.0-4-amd64
> > Found initrd image: /boot/initrd.img-3.16.0-4-amd64
> > Found memtest86+ image: /boot/memtest86+.bin
> > Found memtest86+ multiboot image: /boot/memtest86+_multiboot.bin
> > Found Windows 7 (loader) on /dev/sda1
> > Found Ubuntu 12.04.5 LTS (12.04) on /dev/sdb1
> > Found Debian GNU/Linux (6.0.5) on /dev/sdb3
> > Found Debian GNU/Linux (8.0) on /dev/sdb5
> > Found Ubuntu 14.04.3 LTS (14.04) on /dev/sdb6
> > done
> > 1.33user 0.71system 0:29.75elapsed 6%CPU (0avgtext+0avgdata 16520maxresident)k
> > 251456inputs+74824outputs (52major+1632571minor)pagefaults 0swaps
> >
> > ==============================================================
> >
> > Like I said I have a lot of old kernels lying around this box :-)
>
> And Windowses :-) But nothing looks particularly erroneous about the output.

I used to work from home as a remote agent for a call center and they
had a policy
we had to use Windows no exceptions.  I tried to convert them to Linux
and you all
know how that went :-)

> > The contents of the Makefile in ./arch/x86/include at line 167:
> >
> > 166 - # prevent gcc from generating any FP code by mistake
> > 167 - KBUILD_CFLAGS += -mno-sse -mno-mmx -mno-sse2 -mno-3dnow
> > 168 - KBUILD_CFLAGS += $(call cc-option,-mno-avx,)
>
> Interesting, I see no mention of "Makefile.normal" here. But maybe that's
> because you removed whatever required it?

Could well be ... but 'locate' shows me this:

aruna@debian:~$ locate Makefile.normal
/usr/src/open-vm-tools-9.4.6/vmci/Makefile.normal
/var/lib/dkms/open-vm-tools/9.4.6/build/vmci/Makefile.normal

Does that help or make any sense ?

> > You have pointed me in the right direction so thank you. Ten plus years
> > I have been dabbling with Linux and I just learnt something new. I had no idea
> > about this:
> >
> > aruna@debian:/media/aruna/kernel-stuff/linux-3.18.2$ ls -alh
> > /etc/kernel/postinst.d
> > total 20K
> > drwxr-xr-x 2 root root 4.0K Jun 23 06:18 .
> > drwxr-xr-x 7 root root 4.0K Mar 31 00:45 ..
> > -rwxr-xr-x 1 root root 2.8K Apr 13  2015 apt-auto-removal
> > -rwxr-xr-x 1 root root  858 Mar  1  2015 initramfs-tools
> > -rwxr-xr-x 1 root root  641 Dec 14  2015 zz-update-grub
> >
> > Thank you Thomas I will keep trying to fix the no mouse issue now.
>
> No problem, I wasn't sure if I was helping :-)

Yes sir you most certainly did help.

>
> -- Thomas

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Valdis Klētnieks]

[-- Attachment #1.1: Type: text/plain, Size: 546 bytes --]

On Thu, 24 Jun 2021 14:19:04 -0400, Aruna Hewapathirane said:

> Could well be ... but 'locate' shows me this:
>
> aruna@debian:~$ locate Makefile.normal
> /usr/src/open-vm-tools-9.4.6/vmci/Makefile.normal
> /var/lib/dkms/open-vm-tools/9.4.6/build/vmci/Makefile.normal

Note that 'locate' doesn't actually walk through the file system, it greps
through a file created by 'updatedb' which usually gets run from a crontab
entry.  As a result, 'locate' can report the existence of files that have actually
been removed after the last updatedb run.


[-- Attachment #1.2: Type: application/pgp-signature, Size: 832 bytes --]

[-- Attachment #2: Type: text/plain, Size: 170 bytes --]

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Aruna Hewapathirane]

On Thu, Jun 24, 2021 at 7:09 PM Valdis Klētnieks
<valdis.kletnieks@vt.edu> wrote:
>
> On Thu, 24 Jun 2021 14:19:04 -0400, Aruna Hewapathirane said:
>
> > Could well be ... but 'locate' shows me this:
> >
> > aruna@debian:~$ locate Makefile.normal
> > /usr/src/open-vm-tools-9.4.6/vmci/Makefile.normal
> > /var/lib/dkms/open-vm-tools/9.4.6/build/vmci/Makefile.normal
>
> Note that 'locate' doesn't actually walk through the file system, it greps
> through a file created by 'updatedb' which usually gets run from a crontab
> entry.  As a result, 'locate' can report the existence of files that have actually
> been removed after the last updatedb run.

Oh dear... no wonder locate was lightning fast and I was thinking even
for linux this
is crazy fast speed. Now I know...  Thank you Valdis and any thoughts
on how I can
get my mouse to work at the graphical login screen again ?

make j$(nproc) is ok
make modules_install is ok
make install is ok

but after selecting the new kernel at the GRUB2 menu it hangs at the graphical
login and I am stumped.

Soo... that is why 'locate' is so damn fast :-)

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Valdis Klētnieks]

[-- Attachment #1.1: Type: text/plain, Size: 689 bytes --]

On Thu, 24 Jun 2021 19:36:08 -0400, Aruna Hewapathirane said:

> but after selecting the new kernel at the GRUB2 menu it hangs at the graphical
> login and I am stumped.

Does control-alt-F2, -F3, -F4 get you line mode consoles? If so, you
should be able to login, and get dmesg, the X server log, and the logs
for whatever greeter program runs for graphical login.

If you can't get a line mode console, check if it's answering ping, and
if it's answering any tcp or udp based services (if you have any enabled).
Another thing to look at is netconsole to send kernel messages to a syslog
server running on another machine (or netcat at the destination if it doesn't
have syslog running),

[-- Attachment #1.2: Type: application/pgp-signature, Size: 832 bytes --]

[-- Attachment #2: Type: text/plain, Size: 170 bytes --]

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Aruna Hewapathirane]

On Thu, Jun 24, 2021 at 8:51 PM Valdis Klētnieks
<valdis.kletnieks@vt.edu> wrote:
>
> On Thu, 24 Jun 2021 19:36:08 -0400, Aruna Hewapathirane said:
>
> > but after selecting the new kernel at the GRUB2 menu it hangs at the graphical
> > login and I am stumped.
>
> Does control-alt-F2, -F3, -F4 get you line mode consoles? If so, you
> should be able to login, and get dmesg, the X server log, and the logs
> for whatever greeter program runs for graphical login.

No my keyboard is locked not responsive just the numLock led is on, so
unable to control-alt-F2 :-(

> If you can't get a line mode console, check if it's answering ping, and
> if it's answering any tcp or udp based services (if you have any enabled).
> Another thing to look at is netconsole to send kernel messages to a syslog
> server running on another machine (or netcat at the destination if it doesn't
> have syslog running),

I found a 500Gb ssd inside one of the old machines and I started from scratch.
Pulled out the two cables from the harddisks and plugged in the ssd, installed
debian-10.10.0-amd64-netinst.iso then tried to compile kernel 5.12.13 from
kernel.org and voila et success.

And under 13 minutes to full compile.
Also compiled linux-next-next-20210625 no issues, everything is good again.

Someday I want to try and figure out what is going on with the original system.

Oh, and Thomas Syslog tells me :

Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs'
Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules.
Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service
entered failed state.

Does that make sense to you or anyone ? Anyways things are all good. I
just have to pull out cables
and hook up the ssd when I feel like compiling a kernel, which I can
live with :-)

Cindy Sue Causey I was wondering is there a 'reason' you are sticking
with 'lilo' and not
moving to grub2 ?

Thank you everyone for all the help and support - Aruna

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Valdis Klētnieks]

[-- Attachment #1.1: Type: text/plain, Size: 1304 bytes --]

On Fri, 25 Jun 2021 16:05:45 -0400, Aruna Hewapathirane said:

> Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs'
> Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules.
> Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service entered failed state.
>
> Does that make sense to you or anyone ?

For reasons understandable only to systemd, it tried to modprobe vmhgfs and
didn't find it in the initramfs.   Does this configuration actually *need* a
vmware shared folder in order to boot?  If so, you'll need to tell your dracut/
mkinitramfs/whatever command to include it. If not, you'll need to tell it to
not look for it.

(Particularly annoying -  dracut will include that systemd unit even if you
have a kernel that doesn't need any modules to boot because all the requisites
are built-in, you said --no-kernel to not include kernel modules in the
initramfs, and you gave it -o "kernel-modules kernel-modules-extra kernel-network-modules"
to omit those dracut modules from the run. I submitted a patch that was
rejected, because apparently *some* places build an initramfs with 'dracut
--no-kernel' and then append another cpio image that contains kernel modules,
and the patch would break that.  This is why we can't have nice things.)


[-- Attachment #1.2: Type: application/pgp-signature, Size: 832 bytes --]

[-- Attachment #2: Type: text/plain, Size: 170 bytes --]

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Aruna Hewapathirane]

On Sat, Jun 26, 2021 at 2:39 AM Valdis Klētnieks
<valdis.kletnieks@vt.edu> wrote:
>
> On Fri, 25 Jun 2021 16:05:45 -0400, Aruna Hewapathirane said:
>
> > Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs'
> > Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules.
> > Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service entered failed state.
> >
> > Does that make sense to you or anyone ?
>
> For reasons understandable only to systemd, it tried to modprobe vmhgfs and
> didn't find it in the initramfs.

I think I am beginning to understand what went wrong, I installed
VirtualBox and all
my troubles started after this. Everything was smooth until I tried to
compile a kernel.
This is when things went sideways. So if your going to be compiling
kernels stay away
from VirtualBox on Debian.

So as much as I do not care for systemd because it deviates from the
original philosophy of do one
thing but do that one thing really really well and better than anyone
else policy to do everything under
the sun ( and so far seems to do everything under the sun really
well). It was 'me' who did this 'not'
systemd. I am human, humans err ?

Systemd is impressive and scares me at times but then I guess it comes
down to personal choice ?
I keep asking myself 'if' something bad happens to systemd, what are
our options ? Think about it ?
It is not like tweaking init using text files and echo eh ?

> Does this configuration actually *need* a  vmware shared folder in order to boot?
Uh-uh no it does not.

I guess  it is looking for that 'vmhgfs' because I installed VirtualBox.
And something really interesting, I deleted VirtualBox or so I
thought. I recompiled the kernel
and checked, same 'bleeping' error. So reinstall VirtualBox and like
some Voo Doo magic all
my operating systems showed up which tells me Virtual Box keeps its
folder and settings in
your home folder. One lives and learns :-)

> If so, you'll need to tell your dracut/  mkinitramfs/whatever command to include it.
> If not, you'll need to tell it to not look for it.

oh-kay so here are the latest developments. I told initramfs to load
crc32c, told make menuconfig
to disable Virtualization then recompiled and re-test and my kernel
drops to a initramfs shell ( some progress yay )
then I ask initramfs to load 'vmhgfs' and it complains about 'can't
find module not in blah-blah' and reverts to the
original error, so I am back to square one.

sudo vim /etc/initramfs-tools/modules ( in Debian ) for anyone curious :-)

> (Particularly annoying -  dracut will include that systemd unit even if you
> have a kernel that doesn't need any modules to boot because all the requisites
> are built-in, you said --no-kernel to not include kernel modules in the
> initramfs, and you gave it -o "kernel-modules kernel-modules-extra kernel-network-modules"
> to omit those dracut modules from the run. I submitted a patch that was
> rejected, because apparently *some* places build an initramfs with 'dracut
> --no-kernel' and then append another cpio image that contains kernel modules,
> and the patch would break that.  This is why we can't have nice things.)
>
Valdis, if I apply that patch to 'my' kernel and it works then problem
solved right  ? What do you think ?
And you do have nice things my friend, remember a young lady some
years back wrote in her facebook...
"Drove a Jag, did not crash"  ? :-)

I am going to keep working on this as and when I have the time I want
to know how to fix this.
We know it is because I installed VirtualBox, now I gotta find a fix :-(

Thanks everybody - Aruna

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Re: Kernel Compile Issue

[Thomas Bracht Laumann Jespersen]

> > > Jun 25 10:46:42 debian systemd-modules-load[201]: Failed to find module 'vmhgfs'
> > > Jun 25 10:46:42 debian systemd[1]: Failed to start Load Kernel Modules.
> > > Jun 25 10:46:42 debian systemd[1]: Unit systemd-modules-load.service entered failed state.
> > >
> > > Does that make sense to you or anyone ?
> >
> > For reasons understandable only to systemd, it tried to modprobe vmhgfs and
> > didn't find it in the initramfs.

What Valdis says: I would expect that you can find a 'vmhgfs.ko' file somewhere
in /lib/modules that systemd is trying to load, but can't because it wasn't
included in the initramfs.

*Goes looking for 'vmhgfs.ko' in package manager*

Oh, so the vmhgfs.ko module is installed with open-vm-tools, not
virtualbox-modules. TIL.

Also it appears that vmhgfs is for VMware, not VirtualBox.

> > Does this configuration actually *need* a  vmware shared folder in order to boot?
> Uh-uh no it does not.
> 
> I guess  it is looking for that 'vmhgfs' because I installed VirtualBox.
> And something really interesting, I deleted VirtualBox or so I
> thought. I recompiled the kernel
> and checked, same 'bleeping' error. So reinstall VirtualBox and like
> some Voo Doo magic all
> my operating systems showed up which tells me Virtual Box keeps its
> folder and settings in
> your home folder. One lives and learns :-)

Just so we're on the same page: This module is for VMware, not VirtualBox. So I
wouldn't expect (un)installing VirtualBox would affect this module.

> > If so, you'll need to tell your dracut/  mkinitramfs/whatever command to include it.
> > If not, you'll need to tell it to not look for it.

Exactly. I don't know if/how you do this with systemd, but if you don't need the
module at boot, systemd shouldn't be trying to load it.

> I am going to keep working on this as and when I have the time I want
> to know how to fix this.
> We know it is because I installed VirtualBox, now I gotta find a fix :-(

Silly idea maybe (because I'm no systemd expert): In OpenRC you can blacklist
modules in /etc/modprobe.d. Maybe something like this [0]:

	echo "blacklist vmhgfs" >> /etc/modprobe.d/vmware

Looks like you can even add it as a boot parameter. And I would assume that it
works with systemd as well, because Arch Linux uses systemd by default.

[0] https://wiki.archlinux.org/title/Kernel_modules#Blacklisting

Best of luck!
-- Thomas

_______________________________________________
Kernelnewbies mailing list
Kernelnewbies@kernelnewbies.org
https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies