-
-
Notifications
You must be signed in to change notification settings - Fork 191
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Linux x230 cleanup remix #1184
Linux x230 cleanup remix #1184
Conversation
129f9f0
to
7eb3c24
Compare
Letting compile and then will :
If all above functions, do freespace comparison:
Then
|
x230-hotp-maximized comparison (linux-x230-maximized.config linux configuration)
New:
New - old: |
x230-hotp-verification comparison (linux-x230-legacy.config kernel configuration)
New:
new - old gain: |
Some Might be related to CONFIG_PREEMPT being deactivated.
|
Rechecking historical differences with https://www.kernelconfig.io/ |
952d2eb
to
620bf98
Compare
This is aimed to ease kernel version bump, having a whole picture of config changes between boards and kernel version defaults as opposed to current savedefconfig stored format
620bf98
to
a7cf73c
Compare
Passed checklist. Nothing broke and we gained around 300k from the kernel cuts in 4.14. |
Interesting enough, combined with #1121 we gained nearly 1mb! x230-hotp-maximized comparison (linux-x230-maximized.config linux configuration)
Local build
|
We can see for x230-flash.config that some things were flatten out to be similar with legacy and maximized (less changes). |
Test for linuxboot#590 (recommended as default for -9e per linuxboot#1184)
Test for linuxboot#590 (recommended as default for -9e per linuxboot#1184)
Continuing work under https://github.com/tlaurion/heads/tree/reassess_1184 will be forced push here when satisfactory. A lot of things are build as modules, which are taking useless compile time since those modules are not packed under initrd. For the moment, I will test x230-hotp-maximized to make sure there is no regression. No use of having either RETPOLINE, nor MEMBARRIER not anything of that sort under Heads, since everything is considered whiptail driven and when user does things from the unauthenticated recovery shell, we currently trust him to do the right thing under that volatile session which otherwise would change what is measured under Heads and/or would DOS the system (wiping disk is totally possible as of now). Goal of the present work is just to build a comparative for the 3 linux configs under old 4.14 kernel, aiming to move to the same kernel used by librem boards (5.x), to both reduce CI compilation time (reuse common built stuff to economize build time) and to have merged efforts which happened under #1232, where comparing different versions is really difficult. Goal now is to have reduced 4.14 versions to have baseline for #1121 final size gains comparisons prior of merge there, and finally merge this to go forward including NVMe support, BRTFS, LVM Thin etc, MMC card reader support which is frequently asked since avail on hardware, leading to generalize USB support to support Other Medium boot (boot from external MMC reader ISOs, external hard drive ISOs or secondary hard drive medias). |
Included under #1381 |
Continuing #1116 on x230 linux config tuning.
As opposed to currently stored linux config, those are not stored in defconfig but full oldconfig format.
Rationale:
As explained under #1116 (comment) the reasoning is to facilitate linux kernel version bumps and for users to understand what changed between those directly from the configuration files.
Addition:
Removal:
This is based on top of #1178 which will hopefully be merged soon, after which this will be rebased.
The only relevant commit is WiP here, which will change until satisfaction of size comparison and functionality is met, not causing regression. Then those kernel configurations will be used to have kernel version bump, with clear differences this time.