[10:55] <oere> Dear all, I followed the steps described at https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel but I was not able to modify it so the generated kernel debs have an append local version extension ... I'm using pdebuild --use-pdebuild-internal and used 'dch --local +test' before building it ... is there any other usable documentation ... the background is to patch the kernel to support Beignet on Haswell (https://01.org/zh/beignet/download
[17:51] <arges> infinity: so for bug 1454687, the nx 842 module needs to be loaded in the initramfs to enable usage with zswap. If modules=most doesn't set that in initramfs-tools, is there an 'ubuntu' way to append to that anywhere?
[17:53] <infinity> arges: Except that we enable zswap after we pivot, not in the initramfs.
[17:53] <infinity> arges: This is another case of "bug submitter doesn't use Ubuntu, and assumed we do it the same way as RedHat".
[17:53] <arges> infinity: yea. i got that part when I saw dracut
[17:53] <arges> : )
[17:54] <infinity> arges: We do have an initramfs zswap hook but, confusingly, we only use it in the installer.  And I've been meaning to consolidate it with zram-config.
[17:55] <infinity> arges: But if these need to be loaded pre-zswap (but otherwise, not necessarily loaded at all), they belong conditionally in zram-config.
[17:56] <arges> infinity: ok
[17:57] <infinity> arges: Err, assuming zram and zswap are the same thing.  Maybe they're not.
[17:58] <arges> infinity: the other thing is the bug commenter said he's working on patch to make it change it requireing to be done at boot time
[17:58] <arges> so maybe wait for those patches
[18:00] <infinity> arges: Ahh, zswap != zram.  How confusing.
[18:02] <infinity> arges: Okay, so, maybe the compressors do need to be in the initrd for this, much like lz4 is.
[18:02] <infinity> arges: So, I'd grep initramfs-tools for lz4, and add these in the same spot.
[18:02] <arges> infinity: ok i'll target that package too
[18:03] <infinity> Which, of course, is being pulled in as a dep, so isn't mentioned at all in the source. :P
[18:04] <infinity> arges: Anyhow, it'd be the auto_add_modules function.
[18:05] <arges> infinity: ok i see it
[18:06] <infinity> arges: Don't worry about arch-restricting, it ignores modules that don't exist, IIRC.
[18:06] <infinity> arges: Though, easy to test my assertion on an x86 machine.
[18:07] <arges> infinity: yup. was going to ask. 
[19:01] <arges> infinity: does this seem reasonable? http://paste.ubuntu.com/11775115/
[19:15] <infinity> arges: Probably.  You'll need to do some testing to make sure that (a) it's needed, and (b) it actually does something.
[19:16] <infinity> arges: I, however, might sleep.  I've been up since yesterday morning.
[20:08] <arges> infinity: ok