mailarchive of the ptxdist mailing list
 help / color / mirror / Atom feed
From: Michael Olbrich <m.olbrich@pengutronix.de>
To: "Baeuerle, Florian" <Florian.Baeuerle@allegion.com>
Cc: "ptxdist@pengutronix.de" <ptxdist@pengutronix.de>
Subject: Re: [ptxdist] [PATCH 2/3] barebox: add option for building barebox reproducibly
Date: Wed, 24 Oct 2018 10:19:30 +0200	[thread overview]
Message-ID: <20181024081930.ypwupyqrcese4rmf@pengutronix.de> (raw)
In-Reply-To: <7058ade9eef6f97b42f5e821ef7aa1ffd47ffeac.camel@allegion.com>

On Fri, Oct 12, 2018 at 12:21:14PM +0000, Baeuerle, Florian wrote:
> Am Freitag, den 12.10.2018, 14:01 +0200 schrieb Michael Olbrich:
> > On Thu, Oct 11, 2018 at 02:19:17PM +0000, Baeuerle, Florian wrote:
> > > Inject fake timestamps into the build via faketime to produce the exact
> > > same binaries.
> > > 
> > > Signed-off-by: Florian Bäuerle <florian.baeuerle@allegion.com>
> > > ---
> > >  platforms/barebox.in | 10 ++++++++++
> > >  rules/barebox.make   | 16 +++++++++++++++-
> > >  2 files changed, 25 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/platforms/barebox.in b/platforms/barebox.in
> > > index 645366e93..5bc2f08b6 100644
> > > --- a/platforms/barebox.in
> > > +++ b/platforms/barebox.in
> > > @@ -13,6 +13,7 @@ menuconfig BAREBOX
> > >  	select HOST_OPENSSL if BAREBOX_NEEDS_HOST_OPENSSL
> > >  	select HOST_IMX_CST if BAREBOX_NEEDS_HOST_IMX_CST
> > >  	select HOST_LZOP if BAREBOX_NEEDS_HOST_LZOP
> > > +	select HOST_FAKETIME if BAREBOX_REPRODUCIBLE
> > >  	prompt "barebox                       "
> > >  	bool
> > >  	help
> > > @@ -50,6 +51,15 @@ config BAREBOX_EXTRA_ENV_PATH
> > >  	  Directories and files to add to the builtin barebox environment.
> > >  	  Relative paths must be relative to the platformconfigdir.
> > >  
> > > +config BAREBOX_REPRODUCIBLE
> > > +	prompt "build reproducibly"
> > > +	bool
> > > +	default n
> > > +	help
> > > +	  With this, fake timestamps are injected to the barebox build, so the
> > > +	  resulting barebox images are identical when the sources haven't been
> > > +	  changed.
> > > +
> > >  config BAREBOX_BAREBOXENV
> > >  	prompt "install 'bareboxenv'"
> > >  	bool
> > > diff --git a/rules/barebox.make b/rules/barebox.make
> > > index a6ec3c394..ad160b3ba 100644
> > > --- a/rules/barebox.make
> > > +++ b/rules/barebox.make
> > > @@ -45,6 +45,20 @@ BAREBOX_ENV := \
> > >  	KCONFIG_NOTIMESTAMP=1 \
> > >  	pkg_wrapper_blacklist="$(BAREBOX_WRAPPER_BLACKLIST)"
> > >  
> > > +ifdef PTXCONF_BAREBOX_REPRODUCIBLE
> > > +BAREBOX_FAKE_TIMESTAMP := \
> > > +	$(shell echo $(BAREBOX_VERSION) | \
> > > +		sed -E "s/([0-9]{4})\.([0-9]{1,2}).*/\1-\2-01 00:00:00/")
> > > +BAREBOX_FAKETIME := faketime -f "$(BAREBOX_FAKE_TIMESTAMP)"
> > > +BAREBOX_ENV := $(BAREBOX_ENV) \
> > > +	KBUILD_BUILD_TIMESTAMP="$(BAREBOX_FAKE_TIMESTAMP)" \
> > > +	KBUILD_BUILD_VERSION="0" \
> > > +	KBUILD_BUILD_USER="ptxdist" \
> > > +	KBUILD_BUILD_HOST="ptxdist"
> > 
> > What is not covered by the KBUILD_BUILD_* variables? faketime is a big
> > hammer that can cause other problems. I'd like to understand what happens
> > here first.
> 
> According to my observations, only the compression. A few compressors are
> capable of not including timestamps (e.g. gzip: --no-name), but others are not
> (lzop).
> 
> To fix that, there are a few options I do not like:
> 
> 1. fix the in the barebox build system and implement ugly hacks so this works
> for lzop (overwriting an lzo header field) and other compressors that do not
> support this out of the box.
> 
> 2. fix this in ptxdist, probably by doing the same ugly stuff before copying
> images to IMAGEDIR
> 
> 3. fix the problems in lzop and other compressors that do not support faking or
> skipping those header fields.
> 
> The thirds option is still the best, but it will take ages to reach
> distributions.

I've looked at this a bit. For gzip, barebox already uses '-n', so that
should work correctly. Most other compressors don't add a timestamp, at
least in my quick tests. What remains is lzop.

We already have a BAREBOX_NEEDS_HOST_LZOP option, so the only thing needed
is a patch. Using SOURCE_DATE_EPOCH for mtime_low/mtime_high if available
(instead of the file time) should be trivial.

Michael

-- 
Pengutronix e.K.                           |                             |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
ptxdist mailing list
ptxdist@pengutronix.de

  reply	other threads:[~2018-10-24  8:19 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-10-11 14:19 [ptxdist] [PATCH 1/3] host-faketime: new package Baeuerle, Florian
2018-10-11 14:19 ` [ptxdist] [PATCH 2/3] barebox: add option for building barebox reproducibly Baeuerle, Florian
2018-10-12  7:59   ` Baeuerle, Florian
2018-10-12 12:01   ` Michael Olbrich
2018-10-12 12:21     ` Baeuerle, Florian
2018-10-24  8:19       ` Michael Olbrich [this message]
2018-10-26 13:46         ` [ptxdist] [PATCH 0/2] Barebox reproducible build + ptxdist setup configuration Baeuerle, Florian
2018-10-26 13:46           ` [ptxdist] [PATCH 1/4] config/setup: make reproducible builds configurable Baeuerle, Florian
2018-10-26 13:46           ` [ptxdist] [PATCH 2/4] barebox: depend on HOST_LZOP Baeuerle, Florian
2018-10-26 13:46           ` [ptxdist] [PATCH 3/4] lzop: add patch that allows overriding mtime Baeuerle, Florian
2018-10-26 13:46           ` [ptxdist] [PATCH 4/4] barebox: add support for reproducible build Baeuerle, Florian
2018-10-12 12:28     ` [ptxdist] [PATCH 2/3] barebox: add option for building barebox reproducibly Baeuerle, Florian
2018-10-16  8:59       ` Michael Olbrich
2018-10-16 11:06         ` Baeuerle, Florian
2018-10-16 12:51           ` Michael Olbrich
2018-10-16 14:11             ` Baeuerle, Florian
2018-10-16 19:06               ` Michael Olbrich
2018-10-16 10:09   ` Michael Olbrich
2018-10-16 10:57     ` Baeuerle, Florian
2018-10-16 12:38       ` Michael Olbrich
2018-10-16 14:27         ` Baeuerle, Florian
2018-10-16 19:20           ` Michael Olbrich
2018-10-16 17:15       ` Alexander Dahl
2018-10-16 19:13         ` Michael Olbrich
2018-10-11 14:19 ` [ptxdist] [PATCH 3/3] barebox_mlo: add option for building barebox_mlo reproducibly Baeuerle, Florian
2018-10-11 14:43 ` [ptxdist] [PATCH 1/3] host-faketime: new package Roland Hieber
2018-10-11 21:30   ` Jon Ringle
2018-10-16  7:33     ` Baeuerle, Florian
2018-10-16  8:47       ` Michael Olbrich
2018-10-12  7:47   ` Baeuerle, Florian
2018-10-16  8:51     ` Michael Olbrich
2018-10-16 11:34       ` Alexander Dahl
2018-10-16 12:30         ` Michael Olbrich
2018-10-16  8:53 ` Michael Olbrich

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20181024081930.ypwupyqrcese4rmf@pengutronix.de \
    --to=m.olbrich@pengutronix.de \
    --cc=Florian.Baeuerle@allegion.com \
    --cc=ptxdist@pengutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox