Hi, On Thu, Mar 23, 2023 at 10:36:21PM +0100, Salvatore Bonaccorso wrote: > Hi Eriberto, > > On Sat, Nov 19, 2022 at 11:53:59AM -0300, Eriberto wrote: > > Em sáb., 19 de nov. de 2022 às 11:36, Salvatore Bonaccorso > > <car...@debian.org> escreveu: > > > > > > Hi > > > > > > On Fri, Nov 11, 2022 at 06:33:23PM -0300, Joao Eriberto Mota Filho wrote: > > > > Package: src:linux > > > > Version: 5.19.11-1~bpo11+1 > > > > Severity: important > > > > > > > > Dear maintainer, > > > > > > > > I have a desktop with 3 polls over RAID1 (2 SSD, 2 HDD, plus 2 SSD). The > > > > current kernel on BPO creates an infinite loop when shutting down the > > > > system. I > > > > can see several messages like: > > > > > > > > systemd-shutdown[1]: Not all MD devices stopped. 1 left. > > > > systemd-shutdown[1]: Stopping MD devices. > > > > systemd-shutdown[1]: Stopping MD /dev/md2 (9:2) > > > > md: md2 stopped. > > > > systemd-shutdown[1]: Not all MD devices stopped. 1 left. > > > > systemd-shutdown[1]: Stopping MD devices. > > > > systemd-shutdown[1]: Stopping MD /dev/md2 (9:2) > > > > md: md2 stopped. > > > > systemd-shutdown[1]: Not all MD devices stopped. 1 left. > > > > systemd-shutdown[1]: Stopping MD devices. > > > > systemd-shutdown[1]: Stopping MD /dev/md2 (9:2) > > > > md: md2 stopped. > > > > systemd-shutdown[1]: Not all MD devices stopped. 1 left. > > > > systemd-shutdown[1]: Stopping MD devices. > > > > systemd-shutdown[1]: Stopping MD /dev/md2 (9:2) > > > > [...] > > > > block device autoloading is deprecated and will be removed > > > > block device autoloading is deprecated and will be removed > > > > block device autoloading is deprecated and will be removed > > > > block device autoloading is deprecated and will be removed > > > > block device autoloading is deprecated and will be removed > > > > block device autoloading is deprecated and will be removed > > > > block device autoloading is deprecated and will be removed > > > > block device autoloading is deprecated and will be removed > > > > [...] > > > > md: md2 stopped. > > > > md: md2 stopped. > > > > md: md2 stopped. > > > > md: md2 stopped. > > > > md: md2 stopped. > > > > md: md2 stopped. > > > > md: md2 stopped. > > > > [...] > > > > systemd-shutdown[1]: Not all MD devices stopped. 1 left. > > > > systemd-shutdown[1]: Stopping MD devices. > > > > systemd-shutdown[1]: Stopping MD /dev/md2 (9:2) > > > > md: md2 stopped. > > > > systemd-shutdown[1]: Not all MD devices stopped. 1 left. > > > > systemd-shutdown[1]: Stopping MD devices. > > > > systemd-shutdown[1]: Stopping MD /dev/md2 (9:2) > > > > md: md2 stopped. > > > > [...] > > > > > > > > There is a solution from Arch Linux[1]: > > > > > > > > "Arch disabled BLOCK_LEGACY_AUTOLOAD for 5.18 which broke mdraid". > > > > > > > > [1] https://bbs.archlinux.org/viewtopic.php?id=279383 > > > > > > > > Please, consider disabling the deprecated BLOCK_LEGACY_AUTOLOAD, > > > > enabled by > > > > default in current kernel on Debian: > > > > > > > > $ cat /boot/config-5.19.0-0.deb11.2-amd64 | grep BLOCK_LEGACY_AUTOLOAD > > > > CONFIG_BLOCK_LEGACY_AUTOLOAD=y > > > > > > > > Thanks in advance. > > > > > > I'm not sure, can we can do that (yet)? Some context about this is in > > > https://lore.kernel.org/all/yhe%2fc0k0fn9j8...@bombadil.infradead.org/ > > > . In fact back for 5.18-rc1 upstream has weakened the removal schedule > > > for block device autoloading and with 451f0b6f4c44 ("block: default > > > BLOCK_LEGACY_AUTOLOAD to y")[1]. > > > > > > Initially it was planned to make it for 5.19, see fbdee71bb5d8 > > > ("block: deprecate autoloading based on dev_t")[2]. > > > > > > [1]: > > > https://git.kernel.org/linus/451f0b6f4c44d7b649ae609157b114b71f6d7875 > > > [2]: > > > https://git.kernel.org/linus/fbdee71bb5d8d054e1bdb5af4c540f2cb86fe296 > > > > > > Regards, > > > Salvatore > > > > Thanks for the clarification Salvatore. Currently, this is a nightmare > > for me because I am being compelled to power off my computer via the > > energy switch. > > I noticed in mainline landed 6c0f5898836c ("md: select > BLOCK_LEGACY_AUTOLOAD") in 6.3-rc3, which was as well backported to > 6.1.21. > > https://git.kernel.org/linus/6c0f5898836c05c6d850a750ed7940ba29e4e6c5 > > So I guess we can/need to close this bug or mark it wontfix at least > for the time beeing. > > Thoughts?
Well scratch that about closing the bug. The bug still exists, it's just that we won't disable BLOCK_LEGACY_AUTOLOAD. I assume you are still able to reproduce the problem with the most current kernel available? Regards, Salvatore
