Hi,
I hope you’re well. I want to clone the Linux GitHub kernel. According to Linux stable tree mirror at github — Greg Kroah-Hartman I must clone GitHub - gregkh/linux: Linux kernel stable tree mirror · GitHub . Does Kroah-Hartman GitHub mirror have any Git Large File Storage LFS file(s)?
Welcome to the forums @temika
Greg KH’s GitHub repository is a read-only mirror of the Linux stable tree. It does not use Git LFS, so you can clone it normally with standard Git and there is no need for the Git LFS client.
The repository is large because of the kernel’s long development history, the large number of branches and tags, and the size of the source tree itself.
If you do not need the full history, a shallow clone such as git clone --depth 1 https://github.com/gregkh/linux.git can save time and disk space.
Are you building a custom kernel or just exploring the source?
Hi Hydn,
Thank you very much for your prompt reply. Exploring for now, but I aim to build a custom kernel.
Hi @temika that’s a wonderful objective!
A kernel is going to take a while to build. I have a few thoughts and comments on that; I have done it before, though not very recently, and not with Greg’s specific work.
One thing to consider is whether you are going to build a kernel that you may use or distribute to various different systems or you want to build a kernel for a very specific system, because the modules that you build can and should vary depending on what you need.
If you are building only for a specific system, I recommend figuring out exactly what hardware specification and peripherals you want to utilize, and include only these in your build; this will greatly decrease the size of the kernel, and if you are careful to optimize it too, you can come up with a fast kernel for your system.
If, however you are targeting more than one system, then a general purpose configuration is better and it’s appropriate to build most, but maybe not ALL possible modules; again, build for what the use case will be. General purpose systems will typically include most things, but a server based kernel may add certain things, perhaps to optimize accessibility to a high speed disk farm, and omit some things that a laptop user may want.
Therefore an understanding of the module configuration is helpful. I’m already “rusty” on the particulars, but I’m at least aware of these types of things to consider when you decide to build your own kernel. Best wishes in your efforts, and do please write when you’ve built your first effort!