Buildroot Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Fiona Klute via buildroot <buildroot@buildroot.org>
To: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Kory Maincent <kory.maincent@bootlin.com>,
	buildroot@buildroot.org, Julien Olivain <ju.o@free.fr>
Subject: Re: [Buildroot] [PATCH v2] package/uboot-tools: Bump to version 2025.10
Date: Tue, 4 Nov 2025 22:43:39 +0100	[thread overview]
Message-ID: <3c366bcd-29ce-49da-a077-dda299470e3b@gmx.de> (raw)
In-Reply-To: <20251103235356.0b00bb4b@windsurf>

Hi Thomas, everyone!

Am 03.11.25 um 23:53 schrieb Thomas Petazzoni:
> Hello,
> 
> On Mon, 3 Nov 2025 23:46:36 +0100
> Fiona Klute <fiona.klute@gmx.de> wrote:
> 
>> You might rightly ask why I'm using both, but a) some examples in the
>> U-Boot documentation do [2], and b) that's how I noticed the issue in
>> the previous patch, so I kept it to notice similar issues. Point is,
>> some people might want to use other hash algorithms.
>>
>> Enabling CONFIG_TOOLS_CRC32 in autoconf.h the same way as
>> CONFIG_TOOLS_SHA256 works, but U-Boot supports a few more hash
>> algorithms that some people might want to use. Unless I missed
>> something, the full list should be:
>>
>> CONFIG_TOOLS_CRC16
>> CONFIG_TOOLS_CRC32
>> CONFIG_TOOLS_MD5
>> CONFIG_TOOLS_SHA1
>> CONFIG_TOOLS_SHA256
>> CONFIG_TOOLS_SHA384
>> CONFIG_TOOLS_SHA512
>>
>> Question is: What's the best way to fix this?
> 
> If it's host-uboot-tools, just enabling support for everything.

Host-uboot-tools is where it's an issue for me personally, but if anyone 
is using target uboot-tools to build FIT images it'll affect that, too 
(and possibly other things that need hashes). A patch to that just 
enables them all would be simple enough.

>> I can send a simple patch that adds all the (currently available) hash
>> algorithms to autoconf.h. But that feels kind of messy, plus there's the
>> question what else may be broken (just unnoticed so far).
>>
>> I see in the commit log that many years ago uboot-tools used KConfig and
>> that was then ripped out in 2015 due to licensing concerns I don't quite
>> understand (commit daf2c705a7252b662afe0100bfe1a3ebf75c1c2f, the
>> licensing concern is stated without explanation in
>> package/uboot-tools/0002-tools-only-in-no-dot-config-targets.patch).
> 
> The licensing concern *may* be out of date.
> 
> For some background, read
> https://en.wikipedia.org/wiki/OpenSSL#Licensing. Basically, the OpenSSL
> 1.x license is incompatible with the GPL license, so you can't link
> GPLv2 code (U-Boot code) with OpenSSL 1.x. However, the license of
> OpenSSL was changed, and as of OpenSSL 3.x, it's released under Apache
> 2.0, which is compatible with the GPL. Since we're now using OpenSSL
> 3.x in Buildroot, I believe this licensing concern is moot.

Thank you, it makes a lot more sense now. :-)

>> Interestingly the commit message for
>> 2ebf652589491ac2f8f3825afa5f75156c88b0a0 (2022) states it'd be better to
>> switch (back) to KConfig. Unless someone can explain the license
>> concern, I'm inclined to agree (but also with the statement it'd be a
>> major change).
>>
>> We could also add settings in Buildroot config, but that seems like it'd
>> combine the worst of the other two options.
> 
> To me it doesn't make sense to add an option for each and every hash.
> What should be made optional is the OpenSSL dependency. So for me it's
> OK to have uboot-tools built by default without FIT signature support
> (doesn't require OpenSSL), and an option to enable FIT signature
> support (which needs OpenSSL, and that should enable every possible
> hash supported by U-Boot).

This isn't actually about signatures, at least not primarily, it's about 
checksums for images (kernel, initramfs, DTB, …) in the FIT image. They 
are needed for unsigned images, too, but do not need OpenSSL. So the 
"FIT support" option would be the one to use to enable/disable the 
hashes, if we want to make them optional (right now SHA-1 and SHA-256 
are enabled unconditionally).

Best regards,
Fiona

_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

      reply	other threads:[~2025-11-04 21:43 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-03 13:53 [Buildroot] [PATCH v2] package/uboot-tools: Bump to version 2025.10 Kory Maincent via buildroot
2025-11-03 21:02 ` Julien Olivain via buildroot
2025-11-03 22:46 ` Fiona Klute via buildroot
2025-11-03 22:53   ` Thomas Petazzoni via buildroot
2025-11-04 21:43     ` Fiona Klute via buildroot [this message]

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=3c366bcd-29ce-49da-a077-dda299470e3b@gmx.de \
    --to=buildroot@buildroot.org \
    --cc=fiona.klute@gmx.de \
    --cc=ju.o@free.fr \
    --cc=kory.maincent@bootlin.com \
    --cc=thomas.petazzoni@bootlin.com \
    /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