[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20250918152830.438554-1-nschichan@freebox.fr>
Date: Thu, 18 Sep 2025 17:28:30 +0200
From: Nicolas Schichan <nschichan@...ebox.fr>
To: safinaskar@...il.com
Cc: akpm@...ux-foundation.org,
andy.shevchenko@...il.com,
axboe@...nel.dk,
brauner@...nel.org,
cyphar@...har.com,
devicetree@...r.kernel.org,
ecurtin@...hat.com,
email2tema@...il.com,
graf@...zon.com,
gregkh@...uxfoundation.org,
hca@...ux.ibm.com,
hch@....de,
hsiangkao@...ux.alibaba.com,
initramfs@...r.kernel.org,
jack@...e.cz,
julian.stecklina@...erus-technology.de,
kees@...nel.org,
linux-acpi@...r.kernel.org,
linux-alpha@...r.kernel.org,
linux-api@...r.kernel.org,
linux-arch@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org,
linux-block@...r.kernel.org,
linux-csky@...r.kernel.org,
linux-doc@...r.kernel.org,
linux-efi@...r.kernel.org,
linux-ext4@...r.kernel.org,
linux-fsdevel@...r.kernel.org,
linux-hexagon@...r.kernel.org,
linux-kernel@...r.kernel.org,
linux-m68k@...ts.linux-m68k.org,
linux-mips@...r.kernel.org,
linux-openrisc@...r.kernel.org,
linux-parisc@...r.kernel.org,
linux-riscv@...ts.infradead.org,
linux-s390@...r.kernel.org,
linux-sh@...r.kernel.org,
linux-snps-arc@...ts.infradead.org,
linux-um@...ts.infradead.org,
linuxppc-dev@...ts.ozlabs.org,
loongarch@...ts.linux.dev,
mcgrof@...nel.org,
mingo@...hat.com,
monstr@...str.eu,
mzxreary@...inter.de,
patches@...ts.linux.dev,
rob@...dley.net,
sparclinux@...r.kernel.org,
thomas.weissschuh@...utronix.de,
thorsten.blum@...ux.dev,
torvalds@...ux-foundation.org,
tytso@....edu,
viro@...iv.linux.org.uk,
x86@...nel.org,
nschichan@...ebox.fr
Subject: Re: [PATCH RESEND 00/62] initrd: remove classic initrd support
Hello,
> Intro
> ====
> This patchset removes classic initrd (initial RAM disk) support,
> which was deprecated in 2020.
This serie came a bit as a surprise, because even though the message
notifying of the initrd deprecation was added in July 2020, the message
was never displayed on our kernels.
When booting with root=/dev/ram0 in the kernel commandline,
handle_initrd() where the deprecation message resides is never called,
which is rather unfortunate (init/do_mounts_initrd.c):
if (rd_load_image("/initrd.image") && ROOT_DEV != Root_RAM0) {
init_unlink("/initrd.image");
handle_initrd(root_device_name); // shows the deprecation msg
return true;
}
It is likely we are not the alone booting with that particular
configuration, so other people are probably going to be surprised when
initrd support is removed, because they never saw the deprecation
message.
We do depend on initrd support a lot on our embedded platforms (more
than a million devices with a yearlyish upgrade to the latest
kernel). If it eventually becomes removed this is going to impact us.
We use an initrd squashfs4 image, because coming from a time where
embedded flash devices were fragile, we avoid having the root
filesystem directly mounted (even when read only) on the flash
block/mtd device, and have the bootloader load the root filesystem as
an initrd.
We use a squashfs4 because we can mount it and keep it compressed. The
kernel would decompress data on demand in the page cache, and evict it
as needed.
Regards,
--
Nicolas Schichan
Powered by blists - more mailing lists