Source: openmpi Followup-For: Bug #814451 This issue has started becoming a nuisance for many packages recently (this bug was filed in 2016, so I'm not certain why it's only recently started making more trouble. Possibly because we recently starting building h5py with MPI support).
Addressing Gilles' questions: - is the mlx4_core.ko or similar module loaded ? For me mlx4_core is not loaded, but ib_core is loaded - if yes, is there any reason for it to be loaded ? It's built into the default kernel image, linux-image-5.3.0-2-amd64 5.3.9-2 The Debian kernel build configuration can be found at https://salsa.debian.org/kernel-team/linux https://salsa.debian.org/kernel-team/linux/tree/master/debian I can't see any explicit reference there to ib_core, but CONFIG_MLX5_INFINIBAND=m is set in config/config - is the __linux__ macro defined on debian ? I can't tell from the config files if it is or is not. - does the /sys/class/infiniband directory exists ? It does exist. $ ls -ld /sys/class/infiniband drwxr-xr-x 2 root root 0 Nov 15 14:46 /sys/class/infiniband - is this directory empty ? It is empty. $ ls -l /sys/class/infiniband total 0 -- System Information: Debian Release: bullseye/sid APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 5.3.0-2-amd64 (SMP w/4 CPU cores) Kernel taint flags: TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE Locale: LANG=en_AU.UTF-8, LC_CTYPE=en_AU.UTF-8 (charmap=UTF-8), LANGUAGE=en_AU:en (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system) LSM: AppArmor: enabled
