| .\" Copyright (c) 2025 Chengyu Zhu <hudsonzhu@tencent.com> |
| .\" |
| .TH MOUNT.EROFS 8 |
| .SH NAME |
| mount.erofs \- manage EROFS filesystem |
| .SH SYNOPSIS |
| \fBmount.erofs\fR [\fIOPTIONS\fR] \fISOURCE\fR \fIMOUNTPOINT\fR |
| .br |
| \fBmount.erofs\fR \fB\-u\fR \fITARGET\fR |
| .br |
| \fBmount.erofs\fR \fB\-\-reattach\fR \fITARGET\fR |
| .SH DESCRIPTION |
| EROFS is an enhanced lightweight read-only filesystem with modern designs |
| for scenarios which need high-performance read-only requirements. |
| .PP |
| \fBmount.erofs\fR is used to mount an EROFS filesystem from \fISOURCE\fR |
| (which can be an image file or block device) to a \fIMOUNTPOINT\fR. It supports multiple backends including |
| direct kernel mount, FUSE-based mount, and NBD (Network Block Device) for |
| remote sources like OCI images. |
| .SH OPTIONS |
| .TP |
| .B \-h, \-\-help |
| Display help message and exit. |
| .TP |
| .B \-V, \-\-version |
| Display version information and exit. |
| .TP |
| .BI "\-o " options |
| Comma-separated list of mount options. See \fBMOUNT OPTIONS\fR below. |
| .TP |
| .BI "\-t " type[.subtype] |
| Specify the filesystem type and optional subtype. The type should be |
| \fBerofs\fR. Available subtypes are: |
| .RS |
| .TP |
| .B fuse |
| Use FUSE-based mount via \fBerofsfuse\fR. |
| .TP |
| .B local |
| Force direct kernel mount (default if available). |
| .TP |
| .B nbd |
| Use NBD backend for remote sources (e.g., OCI images). |
| .RE |
| .TP |
| .B \-u |
| Unmount the filesystem at the specified target. |
| .TP |
| .B \-\-reattach |
| Reattach to an existing NBD device and restart the NBD service. |
| .SH MOUNT OPTIONS |
| Standard mount options: |
| .TP |
| .B ro |
| Mount the filesystem read-only (default). |
| .TP |
| .B rw |
| Mount the filesystem read-write (not supported for EROFS). |
| .TP |
| .B nosuid |
| Do not honor set-user-ID and set-group-ID bits. |
| .TP |
| .B suid |
| Honor set-user-ID and set-group-ID bits (default). |
| .TP |
| .B nodev |
| Do not interpret character or block special devices. |
| .TP |
| .B dev |
| Interpret character or block special devices (default). |
| .TP |
| .B noexec |
| Do not allow direct execution of any binaries. |
| .TP |
| .B exec |
| Allow execution of binaries (default). |
| .TP |
| .B noatime |
| Do not update inode access times. |
| .TP |
| .B atime |
| Update inode access times (default). |
| .TP |
| .B nodiratime |
| Do not update directory inode access times. |
| .TP |
| .B diratime |
| Update directory inode access times (default). |
| .TP |
| .B relatime |
| Update inode access times relative to modify or change time. |
| .TP |
| .B norelatime |
| Do not use relative atime updates. |
| .SH OCI-SPECIFIC OPTIONS |
| The following OCI-specific options are available: |
| .TP |
| .BI "oci.blob=" digest |
| Specify the OCI blob digest to mount. The digest should be in the format |
| \fBsha256:...\fR. Cannot be used together with \fBoci.layer\fR. |
| .TP |
| .BI "oci.layer=" index |
| Specify the OCI layer index to mount (0-based). Cannot be used together |
| with \fBoci.blob\fR. |
| .TP |
| .BI "oci.platform=" platform |
| Specify the target platform (default: \fBlinux/amd64\fR). |
| .TP |
| .BI "oci.username=" username |
| Username for OCI registry authentication. |
| .TP |
| .BI "oci.password=" password |
| Password for OCI registry authentication. |
| .TP |
| .BI "oci.tarindex=" path |
| Path to a tarball index file for hybrid tar+OCI mode. |
| .TP |
| .BI "oci.zinfo=" path |
| Path to a gzip zinfo file for random access to gzip-compressed tar layers. |
| .TP |
| .BI "oci.insecure" |
| Use HTTP instead of HTTPS to access the image registry. |
| .SH NOTES |
| .IP \(bu 2 |
| EROFS filesystems are read-only by nature. The \fBrw\fR option will be ignored. |
| .IP \(bu 2 |
| When mounting OCI images via NBD, the mount process creates a background |
| daemon to serve the NBD device. The daemon will automatically clean up when |
| the filesystem is unmounted. |
| .IP \(bu 2 |
| The \fB\-\-reattach\fR option is useful for recovering NBD mounts after a |
| system crash or when the NBD daemon was terminated unexpectedly. |
| .IP \(bu 2 |
| Kernel direct mount is used when mounting a regular file without |
| specifying a backend type. If file-based mounts is unsupported, |
| loop devices will be set up automatically. |
| .SH SEE ALSO |
| .BR mkfs.erofs (1), |
| .BR erofsfuse (1), |
| .BR dump.erofs (1), |
| .BR fsck.erofs (1), |
| .BR mount (8), |
| .BR umount (8) |
| .SH AVAILABILITY |
| \fBmount.erofs\fR is part of erofs-utils. |