| Summary: | vfs_mountroot_shuffle sets wrong mnt_stat.mntonname | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Base System | Reporter: | Denis V. Meltsaykin <dvm.avgoor> | ||||
| Component: | kern | Assignee: | freebsd-bugs (Nobody) <bugs> | ||||
| Status: | New --- | ||||||
| Severity: | Affects Some People | ||||||
| Priority: | --- | ||||||
| Version: | 10.0-STABLE | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| Attachments: |
|
||||||
Created attachment 146853 [details] patch In case if you have configured /.mount.conf with several rootfs options like (for example): .md /mfs_root.lz readonly ufs:/dev/ufs/rootimg .timeout 3 ufs:/dev/ufs/bootfail .ask after fail at first rootfs kernel proceeds to next and loads it correct. But after loading system there are wrong options in mounted filesystems: /dev/ufs/failsafe on / (ufs, local, read-only) /dev/ufs/bootfail on /.mount (ufs, local, read-only) /dev/ufs/boot on /.mount (ufs, local, read-only) And a real ufs/boot is under /.mount/.mount/ instead of /.mount In this case you can't unmount it or remount to RW. Underlying filesystems become useless. This happens because of vfs_mountroot_shuffle() sets mporoot->mnt_stat.f_mntonname unconditional, despite the real FS hierarchy. I have attached something looks like a patch, please review it. I hope it will help. Thank you.