Control: tag -1 moreinfo Hi,
On Fri, 23 Sep 2022 at 16:47:19 +0900, Yuya Nishihara wrote:
> This is the output of 'cryptsetup luksResume', at version 2:2.5.0-3:
Thanks for the debug log! Did you obtain that by adding ‘--debug’ to
/lib/cryptsetup/functions:resume_mapping() and rebuilding the initramfs?
This is where our traces differ:
> # Detected dm-ioctl version 4.47.0.
> # Detected dm-crypt version 1.24.0.
> # Device-mapper backend running with UDEV support enabled.
While I got
# Detected dm-ioctl version 4.46.0.
# Detected dm-crypt version 1.24.0.
# Udev is not running. Not using udev synchronisation code.
# Device-mapper backend running with UDEV support disabled.
Which suggests that my resume environment is lacking /run/udev/control,
while yours doesn't:
https://sources.debian.org/src/lvm2/2.03.16-1/device_mapper/libdm-common.c/?hl=2336#L2336
https://sources.debian.org/src/systemd/251.4-3/src/libudev/libudev-queue.c/?hl=133#L133
Could you please apply /tmp/cryptsetup-suspend-wrapper.diff to
/lib/cryptsetup/scripts/suspend/cryptsetup-suspend-wrapper so we see
what the list of mounted filesystems and the content of /run/udev looks
like? The resume environment shouldn't contain /run/udev since we're no
longer bind-mounting /run.
If for some reason something automatically creates /run/udev/control
then I think a fix would be to pass ‘DM_DISABLE_UDEV=y’ to the
resume_device() environment.
Cheers,
--
Guilhem.
signature.asc
Description: PGP signature
