This allows users to specify exactly which modules to include in the initramfs for their system. They must remember to include dependancy modules themselves! (Eg sata_* depend on libata).
Created attachment 126048 [details, diff] 0002-Add-customized-modules_load-support.patch
*** Bug 154047 has been marked as a duplicate of this bug. ***
I think I'd prefer do this in a configuration file. In fact, I was thinking more of making a genkernel.conf option for allowing someone to do just this, but for both the modules_load file or the individual MODULES_* variables. Something like... MODULES_LOAD_PATH (unset uses default) OVERRIDE_MODULE_VARS=[y/N] Setting OVERRIDE_MODULES_VARS should allow a user to set one (or more) of the MODULES_* variables in genkernel.conf, itself. I'd love it if the contents of the variables were "stackable" using make.conf-like rules, but I'd be perfectly happy with simply forcing the user to put all the modules they want for each variable and pushing the stacking support off until some time later. The override should only override variables that are explicitly listed. The rest are pulled from the normal modules_load file(s). How does that sound?
I personally think this patch is fine as-is. It's a very simple approach that should work for pretty much everyone. We could even extend it a bit to allow for loading multiple files (for i in ${MODULES_CONFIG}; do source ${i}; done). It won't do stacking within the individual vars, but it will allow "inheriting" from previous files, so you can just redefine the var you want.
Have something like this being implemented? I use nouveau-drm, and wanted it for KMS-goodies loaded already during initramdisk-phase, but I only want the nouveau-module, the other ones I have compiled are not needed that early in the boot process...
PS: Adding keyword "Inclusion" to better show this bugs nature in searches...
... open since 2007-07-26, last discussed on 2008-11-15 ... why, why, WHY does Gentoo take soooo long to implement existing solutions like this _always_??
(In reply to comment #7) > ... open since 2007-07-26, last discussed on 2008-11-15 ... why, why, WHY > does Gentoo take soooo long to implement existing solutions like this > _always_?? Show me the existing solution. We have a 5 year old patch that probably does not even apply cleanly against genkernel anymore. And it seems like there was a light argue about if that patch had the right approach among the former maintainers of genkernel, which probably lead to them never applying it in the first case. So maybe try the patch out, and help port it so it works against the current genkernel codebase? Because honestly. Your message killed all motivation I could possibly find to work on this feature request. So either hope someone else find the motivation, or come back later and try a little more friendly bump about this feature request.
(In reply to comment #7) > ... open since 2007-07-26, last discussed on 2008-11-15 ... why, why, WHY > does Gentoo take soooo long to implement existing solutions like this > _always_?? genkernel is quite complex, we are understaffed and work without payment in our spare time. I can assure you that Debian's initramfs generator has open long standing, bugging issues, too. Not too pick at Debian, just to put this in relation.
Created attachment 347578 [details, diff] Add support MODULES_* override from /etc/genkernel.conf for genkernel 3.4.45 version
Yes, 9.5 years later, I have come back to my patch, and re-implemented it! [master 5745d52] Allow complete override of MODULES_* 2 files changed, 24 insertions(+), 3 deletions(-)