Re: fs/acl: Add ACL_USER_OBJ permissions test
Sachin Sant <[email protected]>
| Newsgroups | gmane.linux.ltp |
|---|---|
| Message-ID | <[email protected]> |
On 03/06/26 3:13 pm, [email protected] wrote: > [PATCH 3/8] fs/acl: Add symlink ACL operations test This should be PATCH 6/8 in the series. The review comments are correct. The current test only proves that acl_set_file() and acl_get_file() succeed through the symlink path, but it does not verify that the ACL was actually applied to the target file. I’ll update it to read the ACL from both TESTSYMLINK and TESTFILE, and verify the expected entries/permissions match.” > acl_link01.c: Test does not verify what it claims > > The commit message says: > "acl_set_file() on a symlink sets ACL on the target file" > > But the test only checks the return value of acl_set_file() and > acl_get_file() through the symlink — it never verifies that the ACL > was actually set on TESTFILE (the symlink target). > > The retrieved ACL is read through the symlink and then immediately > freed without inspecting its contents or comparing it with what was > set. So the claim that "ACL operations follow symlinks rather than > operating on the link itself" is not actually proven. > > To properly verify this, also call acl_get_file(TESTFILE, ...) and > confirm the expected entries are present, or at minimum compare the > ACL retrieved via the symlink against the ACL retrieved via the > direct file path. > > --- -- Thanks - Sachin -- Mailing list info: https://lists.linux.it/listinfo/ltp