[SOLVED] Compiling Eoan gives debian/rules .compile_mark error

[Feni]

Hi there,

I'm trying to update my tutorial for removing the RMRR check to Eoan because Intel changed their intel-iommu driver, which means I need to make a new patch.
(https://forum.proxmox.com/threads/t...ntel-iommu-driver-to-remove-rmrr-check.36374/)

However, compiling the latest pve-kernel git causes an error, even if I just run the make clean without modifications:

Commands:

cd /usr/src/
git clone git://git.proxmox.com/git/pve-kernel.git
cd pve-kernel
make

Error:

<snip>
  LD [M]  ubuntu/xr-usb-serial/xr_usb_serial_common.ko
  LD [M]  virt/lib/irqbypass.ko
make[2]: Leaving directory '/usr/src/pve-kernel/build/ubuntu-eoan'
make[1]: *** [debian/rules:98: .compile_mark] Error 2
make[1]: Leaving directory '/usr/src/pve-kernel/build'
dpkg-buildpackage: error: debian/rules build subprocess returned exit status 2
make: *** [Makefile:58: pve-kernel-5.3.10-1-pve_5.3.10-1_amd64.deb] Error 2

This error persists no matter what I do, am I missing something or is the pve-kernel.git currenly broken @Thomas Lamprecht ?

 

[t.lamprecht]

However, compiling the latest pve-kernel git causes an error, even if I just run the make clean without modifications:

This is just a follow up error message, the real one may be a lot earlier in the log, please record it all with

make 2>&1 | tee build.log

and post the log here or in a paste service.

Note, you need (temporarily) several GB of free space to compile the kernel successfully, i.e. ~ 30 - 50 GB..

 

[Feni]

This is just a follow up error message, the real one may be a lot earlier in the log, please record it all with

make 2>&1 | tee build.log

and post the log here or in a paste service.

Note, you need (temporarily) several GB of free space to compile the kernel successfully, i.e. ~ 30 - 50 GB..

Thanks for the quick reply, I will follow-up with a reply in an hour or so.

 

[t.lamprecht]

Thanks for the quick reply, I will follow-up with a reply in an hour or so.

Ah yeah, the joy of waiting for kernels to compile ^^

 

[t.lamprecht]

Oh, and before I forget it: I just pushed out a new 5.3 kernel to the git repo about an hour or so ago, so you may want to base then off that one, just FYI.

 

[Feni]

Oh, and before I forget it: I just pushed out a new 5.3 kernel to the git repo about an hour or so ago, so you may want to base then off that one, just FYI.

Well, that went a bit faster than I feared luckily. Same result however, see the log for details: https://drive.google.com/file/d/1fZAQ4fWCxtTracaRGcc3l5KHebRsoL7b/view?usp=sharing

Only error I noticed other than the one I posted already:

/bin/sh: 1: lz4c: not found
make[4]: *** [arch/x86/boot/compressed/Makefile:146: arch/x86/boot/compressed/vmlinux.bin.lz4] Error 127
make[4]: *** Deleting file 'arch/x86/boot/compressed/vmlinux.bin.lz4'
make[3]: *** [arch/x86/boot/Makefile:112: arch/x86/boot/compressed/vmlinux] Error 2
make[2]: *** [arch/x86/Makefile:284: bzImage] Error 2
make[2]: *** Waiting for unfinished jobs....

 

[t.lamprecht]

/bin/sh: 1: lz4c: not found

Yes, that's the one. It seems the compression for vmlinuz changed from gzip to LZ4, you can solve it for now installing lz4

apt install lz4

Ill re-check a few things and either add lz4 to the build dependency list or switch back to gzip, thanks for the hint!

(I had lz4c installed in my build environment already, so did not noticed that..)

 

[Feni]

Yes, that's the one. It seems the compression for vmlinuz changed from gzip to LZ4, you can solve it for now installing lz4

apt install lz4

Ill re-check a few things and either add lz4 to the build dependency list or switch back to gzip, thanks for the hint!

(I had lz4c installed in my build environment already, so did not noticed that..)

Yeah, it's a new feature indeed in Eoan:

The default kernel compression algorithm was changed to lz4 on most architectures, while the default initramfs compression algorithm was changed to lz4 on all architectures, which should improve boot speed

https://www.linuxuprising.com/2019/10/whats-new-in-ubuntu-1910-eoan-ermine.html
http://smackerelofopinion.blogspot.com/2019/09/boot-speed-improvements-for-ubuntu-1910.html

Thanks for the assistance, compile seems to run smoothly now!

 

[t.lamprecht]

Yeah, it's a new feature indeed in Eoan:

Yes, just also found the commit and the linked report https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1840934
But it's a trade-off of having 25% larger image on /boot vs. only somewhat better boot speed over GZIP..

Not sure if a kernel load time itself of 0.5s instead of 0.75s is such a good improvement, when the trade-off is that one can fit one or two kernel less on the /boot partition (if it's an EFI partition, for example).. Somewhat mixed feelings here, but for now I'll document the lz4 build dependency