On Fri, Jul 25, 2025 at 01:45:51PM +0200, Hector Cao wrote: > > > > > > > > > What do you think of this design: > > > > > 1) Create a file msr.conf inside src/util/modules-load.d/ > > > > > 2) Modify src/util/meson.build to install this file to > > /etc/modules-load.d/ > > > > > > > > > > Optional: > > > > > - Do you think it is necessary to add the configuration option to > > enable > > > > > this modules-load file installation ? > > > > > > > > We should probably have a meson_options.txt entry as some distros have > > it > > > > as a built-in, so it would be wierd to see a file installed to load it. > > > > > > The other thing to keep in mind is that the file needs to be shipped > > > conditionally, on x86 only. > > > > > > Remind me again, why don't we just use virKModLoad() for this? It's > > > functionality that already exists and can easily be reused, and it > > > would make for an overall cleaner solution IMO. > > > > The modules-load.d file is processed at system startup in a well defined > > context. This is more appropriate than changing global system state in > > the middle of a libvirt API process to start a guest. > > > > > One thing I'm worried about with modules-load.d approach is that at libvirt > installation, the module > is not actually loaded, a reboot is necessary to make it happen. You need that to make systemd service activation happen too, so the need for a reboot isn't new. With regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
