Arch Linux User Repository

Hi. Since the package became an orphan, I adopted it to continue what I actively maintain on https://aur.archlinux.org/pkgbase/linux-mainline-git, now on this AUR package.

The build script and config file will be up-to-date with Arch's official Linux package unless an update is required to be functional.

But I intentionally removed the documentation package because it consumes too much building time.

This package supports certain customization if compiled by the user. Check for the README file inside the package repo, or the .INSTALL message after package installation to learn how.

@sickcodes, further to @sandy8925, look for the following at the end of your build:

==> Finished making: linux-git 5.10rc6.r0.gb65054597872-1 (Mon 30 Nov 2020 06:45:15 GMT)

This tells you what your true build version is, as the version will automatically update to the latest commit on the master branch every time it is built.

@sickcodes I'd suggest adding loglevel to your kernel parameters and setting it to 4 or above. I think the default is 3 IIRC. Probably 6 or 7 would be best.

Hopefully, that'll give you more information during the boot process as to what if failing, or pausing, and maybe why. I do test it on my own laptop, and it is capable of booting, while the current config is based on Arch Linux 5.9 kernel, with a few extra options set for the 5.10 release.

@sickcodes - Sometimes failure to boot is due to upstream code having a bug. Try building a newer or older commit/RC and see if that helps. Usually they fix it up in a later RC and before stable release.

Super fast build:

MAKEFLAGS=-j$(( $(getconf _NPROCESSORS_ONLN)*2 )) make

Any tips on getting this to boot?

It rc2 seems to hang on boot.

Closest I got was to the root recovery shell.

Tried all of these from rEFInd boot loader as well as super grub disk EFI USB boot:

https://gist.github.com/sickcodes/3b614cbbb78db3dcec2f763608251b86

Super fast build:

yay linux-git -G
cd linux-git
N="$(($(nproc)*2))"
sed -i -e 's/make\ /make\ -j'$N'\ /g' PKGBUILD
makepkg -si

We're back in business. The package is back to normal size for me and I've been using the build for a few hours on my laptop. I'll take a look at supporting the -doc package over the weekend.

@iamazat Thank you for taking the time to look at this. I did try a reversion of the 5.10rc changes but as you've shown, it was the syncing of the 5.9 changes which caused it. I was wondering at the time about looking to see if there are changes to PKGBUILD at the time which would be worth migrating as well, but wanted to take it one step at a time! Looks like I should have just gone all out!

I've copied the INSTALL_MOD_STRIP setting to the build step, and a diff comparison also shows a step to run strip on vmlinuz, so I've ported that across too.

I have Jenkins building the updated package now, and if that works I'll sync the additional changes and add support for linux-git-docs package it seems.

Thank you greatly for your help and hopefully this will fix it.

Just for reference for anyone using this package - it looks like a configuration change between 5.9 and 5.10 is resulting in a significant increase in the size of the modules being created: The vmlinuz is still fine for me at about 9MB but the overall package is close to 1GB.

So to summarize (and update my previous comment):

• CONFIG_DEBUG_INFO had been enabled in 1
• no INSTALL_MOD_STRIP like in upstream 2

So after adding INSTALL_MOD_STRIP installed is 80.40 MiB, modules stripped (and initrd fits my 200MB boot parition) but vmlinux 700MB and I'm totally fine with this (since CONFIG_DEBUG_INFO enabled, and this will improve debugging capabilities)

Hi!

There are few issues with the config:

• .config:126:warning: unexpected data: CONFG_CGROUP_SCHED=y

• and CONFIG_DEBUG_INFO_BTF=y, in 1 you enabled it with the comment that w/o it the kernel size increased, looks like the problem is in that CONFIG_DEBUG_INFO had been enabled 2.