novox/mesh-host
8
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Name a filesystem by its kernel id, never its device name (hq ADR 0325 fourth review)

Browse Source 
An ext4 or xfs library mounted by UUID can come back from a reboot as another device
(/dev/sdb1 as /dev/sdc1); naming the filesystem by mountinfo's source read that as a
move, stripped the warranted library's entries and asked for a new Approve. A
filesystem is now named by statfs's f_fsid (from the UUID, or the ZFS dataset), its
type and its mount point. Entries left on the filesystem a folder is on now are
merged back or held, never walked away, and a TakenOn naming a mount point is not
taken.
This commit is contained in:
jschoubben
2026-10-12 19:10:18 +02:00
parent 21cd036232
commit f00f9c2a2f
6 changed files with 147 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files
internal/apply/keeper_linux.go
+12 -2
View File
@@ -87,14 +87,24 @@ func dropLeft(path string, f store.KeptFolder) []string {
if _, err := lstatWithin(path); err == nil {
holding = mountHolding(path, MountPoints())
}
holdingID := ""
if holding != "" {
holdingID = PlacedID(holding)
}
var said []string
var still []store.KeptLeft
var mine []store.KeptUser
for _, l := range f.Left {
same := l.Mount == f.Mount && (l.Placed == "" || l.Placed == f.Placed)
switch {
case f.Kept && l.Mount == f.Mount:
case f.Kept && same:
// Left on the filesystem the folder is kept on now: the folder's own entries.
mine = append(mine, l.Users...)
case l.Mount != holding:
case same:
// On the filesystem the folder is on now, not kept yet: never walked away; they are the folder's
// once it is kept there again.
still = append(still, l)
case l.Mount != holding || (l.Placed != "" && l.Placed != holdingID):
still = append(still, l)
default:
drop := map[uint32]bool{}