All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Thomas Schmitt" <scdbackup@gmx.net>
To: grub-devel@gnu.org
Subject: Re: [PATCH v3 2/3] mkrescue: add argument --fixed-time to get reproducible uuids
Date: Tue, 15 Dec 2015 17:48:33 +0100	[thread overview]
Message-ID: <28882584413021611678@scdbackup.webframe.org> (raw)
In-Reply-To: <567035BB.7050204@gmail.com>

Hi,

Andrei Borzenkov wrote:
> I am not sure we should stretch reproducible builds that far. ISO
> image created by grub-mkrescue is not binary.

I was approached by Debian's reproducible-builds project because
they wanted to be able to create reproducible test ISOs.
  http://lists.alioth.debian.org/pipermail/reproducible-builds/Week-of-Mon-20150601/001693.html


Vladimir Serbinenko wrote:
> > We need to find a way to reliably find boot
> > disk without depending on current time.

Andrei Borzenkov wrote:
> Well, UUID of isofs used by GRUB is not unique in any sense,

I understand that we rely on the improbability that two
competing ISOs got created in the same second.
So an explicitely chosen "UUID" must be sufficiently random
on the first production to distinguish non-identical images.
Re-productions should then use the same "UUID".

The "UUID" is stored in the ISO as timestamp string of form
YYYYMMDDhhmmsscc with decimal digits. E.g. 2015121517395800
"cc" means centi-seconds, which would be usable to expand the
"UUID" space by a factor of 100.
It cannot be forwarded as time_t, though.
One would need finer time granularity or a second integer variable
which would bring the "cc" part down to the composition of the
xorriso command. (xorrisofs option --modification-date, i assume)


> so it is not really much worse than it was before.
> Having reliable way to identify boot device imply some unique property
> of boot device which automatically conflict with idea of identical images.

Yep. Having identical images would mean that they are the same
in any aspect. So here we should have no problem, i think.
(Adventurous testers could now try what happens if they present
 their machine two copies of the same ISO on two devices.)


Have a nice day :)

Thomas



  parent reply	other threads:[~2015-12-15 16:47 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-04 16:10 [PATCH 0/3] reproducible builds Alexander Couzens
2015-12-04 16:10 ` [PATCH 1/3] mkstandalone: add argument --fixed-time to override mtime of files Alexander Couzens
2015-12-04 18:00   ` Vladimir 'φ-coder/phcoder' Serbinenko
2015-12-05  6:35     ` Andrei Borzenkov
2015-12-04 16:10 ` [PATCH 2/3] mkrescue: add argument --fixed-time to get reproducible uuids Alexander Couzens
2015-12-04 16:10 ` [PATCH 3/3] Makefile: use FIXED_TIMESTAMP for mkstandalone if set Alexander Couzens
2015-12-04 16:48   ` Alexander Couzens
2015-12-04 17:09     ` [PATCH] Makefile/coreboot use SOURCE_DATE_EPOCH as time source " Alexander Couzens
2015-12-04 18:01   ` [PATCH 3/3] Makefile: use FIXED_TIMESTAMP for mkstandalone " Vladimir 'φ-coder/phcoder' Serbinenko
2015-12-04 18:32 ` [PATCH v3 0/3] reproducible builds Alexander Couzens
2015-12-05  6:28   ` Andrei Borzenkov
2015-12-05 11:43     ` Alexander Couzens
2015-12-04 18:32 ` [PATCH v3 1/3] mkstandalone: add argument --fixed-time to override mtime of files Alexander Couzens
2015-12-14 14:47   ` Vladimir 'φ-coder/phcoder' Serbinenko
2015-12-04 18:32 ` [PATCH v3 2/3] mkrescue: add argument --fixed-time to get reproducible uuids Alexander Couzens
2015-12-14 15:22   ` Vladimir 'φ-coder/phcoder' Serbinenko
2015-12-15 15:46     ` Andrei Borzenkov
2015-12-15 16:02       ` Vladimir 'phcoder' Serbinenko
2015-12-15 16:48       ` Thomas Schmitt [this message]
2015-12-04 18:32 ` [PATCH v3 3/3] Makefile/coreboot use SOURCE_DATE_EPOCH as time source if set Alexander Couzens
2015-12-14 15:23   ` Vladimir 'φ-coder/phcoder' Serbinenko

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=28882584413021611678@scdbackup.webframe.org \
    --to=scdbackup@gmx.net \
    --cc=grub-devel@gnu.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.