Bind mount
A Linux bind mount exposes an existing file or directory subtree at another location in the mount hierarchy.
A bind mount provides another route to existing filesystem objects, not a copy or point-in-time image. Changes through a writable route can remain visible through a read-only route to the same data. [1]
Ordinary MS_BIND does not reproduce submounts beneath the source directory. Adding MS_REC also reproduces eligible submounts; unbindable mounts and their subtrees are pruned. Recursive binding therefore does not promise a complete copy of every visible nested mount. [2] [3]
On the classic mount(2) interface, MS_BIND alone inherits the source mount options. A bind remount can change per-mount flags without changing the underlying superblock, allowing a read-only bind path beside a writable path. This access restriction is not device-wide write protection. [2] [1]
Relations
Citations
- util-linux v2.41.1: mount(8), bind operation and read-only setting notes
- Linux man-pages: mount(2), bind, recursive bind and per-mount remount flags
Sources & caveats
- The command manual is pinned to util-linux v2.41.1. The system-call manual is rolling documentation checked on the access date; no privileged mounting was performed.