[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAGrbwDStsL7=fvm9B3prYh4fQmOb9fTKQ5JQviNa-enjQ3xBcg@mail.gmail.com>
Date: Tue, 7 Oct 2025 01:07:54 +0100
From: Dmitry Safonov <dima@...sta.com>
To: dima@...sta.com
Cc: Nathan Chancellor <nathan@...nel.org>, Nicolas Schier <nicolas.schier@...ux.dev>,
linux-kbuild@...r.kernel.org, linux-kernel@...r.kernel.org,
David Disseldorp <ddiss@...e.de>, Nicolas Schier <nsc@...nel.org>
Subject: Re: [PATCH RFC] gen_init_cpio: Do fsync() only on regular files
On Tue, Oct 7, 2025 at 12:55 AM Dmitry Safonov via B4 Relay
<devnull+dima.arista.com@...nel.org> wrote:
>
> From: Dmitry Safonov <dima@...sta.com>
>
> Here at Arista gen_init_cpio is used in testing in order to create
> an initramfs for specific tests. Most notably, there is a test that does
> essentially a fork-bomb in kdump/panic kernel, replacing build-time
> generated init script: instead of doing makedumpfile, it does call
> shell tests.
>
> In comparison to usr/Makefile, which creates an intermediate .cpio file,
> the Makefile that generates initrd for tests is a one-liner:
> > file lib/kdump ${src_dir}/oom-crashkernel 0644 0 0 | usr/gen_init_cpio /dev/stdin | xz -z -c -e -Ccrc32 > ${target_dir}/oom-crashkernel.initrd
>
> As fsync() on a pipe fd returns -EINVAL, that stopped working.
To clarify: because any sane bash script with pipes should use
set -eo pipefail
which indeed is set in the script that generates initrd.
Thanks,
Dmitry
Powered by blists - more mailing lists