Linux kernel "historical" git repository with full history

linux git linux-kernel kernel

18,292

Solution 1

Here is my setup.

I have a repository with a clone of the following remotes:

And the following grafts (info/grafts):

1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 e7e173af42dbf37b1d946f9ee00219cb3b2bea6a
7a2deb32924142696b8174cdf9b38cd72a11fc96 379a6be1eedb84ae0d476afbc4b4070383681178

With these grafts, I have an unbroken view of the kernel history since 0.01. The first graft glues together the very first release in Linus' repository with the corresponding release of tglx/history.git. The second graft glues together tglx/history.git and davej/history.git.

There are a few older versions missing, and the older versions have release granularity instead of patch granularity, but this is the best setup I know of.

Edit: Dave Jones pointed me to http://www.archive.org/details/git-history-of-linux, which seems to be exactly what you want.

Solution 2

Here is a review of available 2018 options with a focus on tag availability and date correctness.

https://archive.org/download/git-history-of-linux/full-history-linux.git.tar

Developed by Dave Jones, and made available on archive.org.

Covers early versions to 2010.
244,464 commits
Just 184 tags, covering versions in 2.6. The tags that should have been created for all versions seem to be missing.
Early commits have realistic dates, but incorrect times (11:00:00 199X -0600).
Some dates seem to be incorrect. For example, both 2.1.110 and 2.1.111 are dated Wed May 20 11:00:00 1998 -0600, although the latest file in the 2.1.111 snapshot is dated 1998-07-25 09:17.
The creation process is documented on GitHub and seems very thorough.

https://git.kernel.org/pub/scm/linux/kernel/git/tglx/history.git/

Created by Thomas Gleixner.

Covers 2.4.0 to 2.6.12-rc2.
Contains 170 tags covering 2.5.X and 2.6.X.
63,428 commits
Dates are correct.
Contains patches converted into commits.

https://github.com/mpe/linux-fullhistory

Created by Michael Ellerman, derived from work by Yoann Padioleau, based on historical trees reconstructed by Dave Jones and Thomas Gleixner, and Linus' mainline tree.

Covers full history
Provides only 558 tags, mostly starting at 2.0.0.
790,471 commits
Same issues with dates as in Dave Jones's repo.
Uses replace objects instead of grafts.

https://git.kernel.org/pub/scm/linux/kernel/git/history/history.git/

Owned by the Linux history team.

Covers early versions to 2.6.33-rc5.
1710 tags, starting with 0.10, covering most early versions.
244,774 commits
Most historic versions are incorrectly dated Fri Nov 23 15:09:04 2007 -0500.

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/

Modern Linux development.

Covers 2.6.12-rc2 (2005) until today
569 tags
777,419 commits (August 2018)
Proper commits

Solution 3

the referenced repos no longer exist. the new one is here: https://git.kernel.org/cgit/linux/kernel/git/history/history.git/

if you're like me and want to keep some repos sep, you can leverage alternates with the graft to do so:

# Same dir as main linux
$ git clone --bare git://git.kernel.org/pub/scm/linux/kernel/git/history/history.git
$ cd linux/.git/
$ echo ../../../history.git/objects >> objects/info/alternates
$ echo 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 e7e173af42dbf37b1d946f9ee00219cb3b2bea6a >> info/grafts
$ echo 7a2deb32924142696b8174cdf9b38cd72a11fc96 379a6be1eedb84ae0d476afbc4b4070383681178 >> info/grafts

Solution 4

The best what I've found is git://git.kernel.org/pub/scm/linux/kernel/git/davej/history.git. History tracking there starts from Linux-0.01 but many comments are poor something like "Import 2.1.38pre1".

Anyway there is a lot of knowledge.

Thanks for help!

View more solutions

18,292

Author by

ashrugger

Updated on June 16, 2022

Comments

ashrugger almost 2 years

I think many developers like to investigate sources with the help of git gui blame. As explained in the commit for Linux-2.6.12-rc2 (also mirrored at Github), it needs to have special historical Linux repository for this purpose.

Linux-2.6.12-rc2

Initial git repository build. I’m not bothering with the full history, even though we have it. We can create a separate “historical” git archive of that later if we want to, and in the meantime it’s about 3.2GB when imported into git — space that would just make the early git days unnecessarily complicated, when we don’t have a lot of good infrastructure for it.

Let it rip!

I have looked at a lot of the prepared historical repositories but I didn’t find one containing changes going back to version zero, so I gave up and am asking this question here.