[BUG] Read-only permission evidence becomes INTERNAL #32
Labels
No labels
alpha-1
bug
ci
docs
duplicate
enhancement
help wanted
invalid
performance
phase-6
question
refactor
security
wontfix
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
erix/vfsd#32
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Summary
Vfsd reports an internal error when a read-only mount refuses write permission. This hides the expected policy refusal from the caller.
Reproduction
From Vfsd, run:
The fixtures prepare a writable acquisition scope over a read-only mount and request write permission for the root, current directory, or an acquired file. The second fixture crosses into a read-only child mount.
Expected and actual behavior
The permission response must contain
VFS_ACQUISITION_EVIDENCE_READ_ONLY_V1(0x100) and no snapshot. The affected dispatcher passes this valid result through the generic result filter and returnsIPC_RESULT_INTERNAL(5). The mount check itself correctly refuses the request before provider I/O.Environment and identity
Affected: Vfsd 8c42847b. The host fixtures use explicitly configured mount policy and acquired handles.
Evidence and prior checks
The fix preserves
0x100only in the permission-evidence response family. Generic transport result validation remains unchanged. Both regression controls pass in CI run 356.Tracking
Related: mount-policy enforcement, Vfsd PR 3, Phase 6.
The permission response now preserves the read-only evidence code in its own response family. Fixed in this commit.