@ai-ecoverse/wasm-mount

mount and umount for slicc's wasm realm. Processes mount through the kernel's own drivers (tmpfs, fsa, hostfs, installed types).

pnpm add -g @ai-ecoverse/wasm-mount
mkdir -p /mnt/scratch /mnt/f
mount -t tmpfs -o maxfile=1g none /mnt/scratch
mount -t fsa none /mnt/f        # empty ("nomedium") until "Insert folder" in the page
mount                           # list /proc/mounts
umount /mnt/scratch
umount -l /mnt/f                # detach even if busy

• mount [-l] [-t type] lists mounted filesystems.
• mount [-frwv] [-t type] [-o options] <source> <directory> mounts (-f fakes it). ro/rw set read-only; other options go to the driver (maxfile=1g, …).
• umount [-lfv] <directory>|<source>... unmounts; -l/-f detach at once.

Remount, bind and move mounts, mount -a and fstab lookups are not supported. The flags are a subset of util-linux's, with the same spellings and the same listing format. Exit codes follow util-linux: 1 for usage, 32 for mount failure.

Requires @ai-ecoverse/slicc-kernel ≥ 1.17.0 (process mounts, slicc-kernel#92); older kernels answer "mount(2) is not available".

1.0.0-5 (homescoop#666; 2026-10-20)

EPERM from mount or umount now reads Operation not permitted (not your mount point, or refused by the kernel's mount policy). Under users enforce a user may mount only over their own directories, and the old (process mounts are disabled here) was wrong for that case.