Re: [Buildroot] [PATCH 1/1] package/e2fsprogs: fix fsverity build error slightly differently

From: Thomas Petazzoni via buildroot <buildroot@buildroot.org>
To: Markus Mayer <mmayer@broadcom.com>
Cc: Julien Olivain <ju.o@free.fr>, Bernd Kuhls <bernd@kuhls.net>,
	 Buildroot List <buildroot@buildroot.org>
Subject: Re: [Buildroot] [PATCH 1/1] package/e2fsprogs: fix fsverity build error slightly differently
Date: Mon, 2 Feb 2026 14:18:45 +0100	[thread overview]
Message-ID: <aYCjvj6p819w-_E7@windsurf> (raw)
In-Reply-To: <20251205003008.153577-2-mmayer@broadcom.com>

Hello Markus,

On Thu, Dec 04, 2025 at 04:30:04PM -0800, Markus Mayer via buildroot wrote:
> Turns out Thomas and I were working concurrently on two different ways
> to fix the e2fsprogs 1.47.3 build error with regards to fsverity not
> being present in older kernels.
> 
> Thomas' patch disables the fsverity functionality in e2fsprogs based on
> the kernel headers used for compilation while my patch aims to provide
> the missing definitions if the kernel doesn't.
> 
> By providing the missing definitions whenever the kernel headers don't,
> fsverity related features can be compiled in even if the kernel headers
> are too old. This allows binaries to run on newer kernels with full
> functionality, but it relies on e2fsprogs to properly handle -ENOTSUPP
> errors at runtime if the kernel doesn't support the ioctl.[1]
> 
> We both submitted our respective patches to Ted T'so but haven't heard
> back as of yet. After a brief discussion on the Buildroot mailing list,
> we are trying my approach for now.
> 
> [1]: Compiling binaries against older kernel headers, but exclusively
> running them on newer kernels, can happen if one's toolchain comes with
> older kernel headers in the sysroot for increased compatibility across
> multiple kernel versions (e.g. the same toolchain & sysroot are used to
> build root file systems with Linux 5.4 and 6.12).
> 
> Signed-off-by: Markus Mayer <mmayer@broadcom.com>

Since then, my patch has been accepted and merged by Ted Tso:

  https://github.com/tytso/e2fsprogs/commit/4aba12dc19a95626eb7fa1df2d19c5235c2bc848

And it was integrated into Buildroot as well. Since we no longer have
any build failure in Buildroot, I think to be convinced that another
approach is needed, you should first convince upstream first that your
approach is better. Therefore, I will mark your Buildroot patch as
Rejected, but if you get your patches merged upstream, we will
obviously be happy to backport them in Buildroot (until a new
e2fsprogs release contains them, of course).

Best regards,

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot