[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Lzip-bug] I have some practical question about lzip patch for linu

From: wrotycz
Subject: Re: [Lzip-bug] I have some practical question about lzip patch for linux kernel (about lzip-ed kernel modules)
Date: Fri, 20 Apr 2018 06:15:00 +0200
User-agent: GWP-Draft

If the changes required are easy, they may be ready in a few days. Else
it may take longer.

BTW, has anybody verified if this change is worth the effort? I.e., has
LZMA compression any advantages over gzip compression for kernel modules?

I can test it if it's not too complicated. I mean I know how to compile kernel and run new one but don't know what's kmod is all about so you'll need to give some source where I can read about it bit more.

When it comes to modules I made small test and result should be self explaining:
~/tmp/modules $ for DIR in `ls -1`; do du -sk $DIR; done

# unstripped
242644	236.95[MiB]	modules
69848	68.21[MiB]	modules-gz
55848	54.54[MiB]	modules-lz
55812	54.50[MiB]	modules-xz

# stripped
123776	120.87[MiB]	modules-s
45652	44.58[MiB]	modules-s-gz
41844	40.86[MiB]	modules-s-lz
41756	40.78[MiB]	modules-s-xz

So the difference for unstripped modules is 20% less (than gzip) for lzma but for unstripped modules it's 8%. 
Is it only 8% or <antonym to only> 8% is question of interpretation in this case but generally compressing unused modules is definitely good thing. In my case modules in memory take 14MiB, so saving 180/80[MiB] is definitely a good thing even today when computers have tons of RAM.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]