linux-kernel - Re: Debian build polishing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 8 Mar 2019 12:57:03 -0500
From:   "Theodore Ts'o" <tytso@....edu>
To:     "Enrico Weigelt, metux IT consult" <info@...ux.net>
Cc:     linux-kernel@...r.kernel.org, yamada.masahiro@...ionext.com,
        michal.lkml@...kovi.net, apw@...onical.com, joe@...ches.com,
        linux-kbuild@...r.kernel.org, linux-riscv@...ts.infradead.org
Subject: Re: Debian build polishing

On Fri, Mar 08, 2019 at 01:44:19PM +0100, Enrico Weigelt, metux IT consult wrote:
> One point still puzzling me: once the debian/rules is applied and
> somebody calls `make deb-pkg`, he'll end up w/ unclean tree, as
> now a git-tracked file is changed.

This is not something I've noticed, but I build my Debian packages
like this:

make O=/build/linux-build bindeb-pkg

This works really well for me, since all of the build-artificats land
in /build, and I can use a HDD (or a PD-HDD when building using Google
Compute Engine) for /build, while I use a SSD for my source tree.  I
find that using a HDD for a target of a build doesn't really slow
things down, and this allows me to save $$$ (when using a Cloud VM)
and reduce flash wearout and capital cost (on my personal machines).

So I really hope your patches don't break this.  Also, are there any
changes the performance of building the Debian packages before and
after your changes?  And are there any differences in the packages in
terms of any pre-or-post install/removal scripts?

There are a lot of things I really dislike about the "official" Debian
kernel build processes (they're optimzied for distribution release
engineers, not kernel developers), so I'm really hoping that making
things more like the "official Debian way" doesn't break some of the
things I really like about the existing "make bindeb-pkg" build
system.

Cheers,

						- Ted