mbox series

[v3,00/10] Set casefold/fscrypt dentry operations through sb->s_d_op

Message ID 20240111225816.18117-1-krisman@suse.de
Headers show
Series Set casefold/fscrypt dentry operations through sb->s_d_op | expand

Message

Gabriel Krisman Bertazi Jan. 11, 2024, 10:58 p.m. UTC 
Hi,

Thank you, Eric and Al, for your feedback on the previous version.

This version has some big changes: instead of only configuring on the
case-insensitive case, we do it for case-sensitive fscrypt as well, and
disable d_revalidate as needed.  This pretty much reverses the way
fscrypt operated (only enable d_revalidate for dentries that require
it), but has the advantage we can be consistent among variations of
case-insensitive/sensitive, encrypted/unencrypted configurations.

You'll find the code is simpler than v1 and v2.  I dropped the dcache
patch because now we always try to disable DCACHE_OP_REVALIDATE while
holding the d_lock already, so I do it inline; I also changed the way we
drop d_revalidate when the key is made available, because we couldn't
really do it the way I originally proposed on the RCU case, which would
require falling back to non-RCU lookup just to disable d_revalidate; I
also included a patch fixing the overlayfs issue that I mentioned on the
previous thread.  While unrelated to the rest of the patchset, it is a
quick fix that I might merge earlier if you are happy with it.

More details can be found in the per-patch changelog.

This survived fstests on ext4 and f2fs.  I also verified that fscrypt
continues to work when combined to overlayfs as Eric requested.

..
original cover letter:

When case-insensitive and fscrypt were adapted to work together, we moved the
code that sets the dentry operations for case-insensitive dentries(d_hash and
d_compare) to happen from a helper inside ->lookup.  This is because fscrypt
wants to set d_revalidate only on some dentries, so it does it only for them in
d_revalidate.

But, case-insensitive hooks are actually set on all dentries in the filesystem,
so the natural place to do it is through s_d_op and let d_alloc handle it [1].
In addition, doing it inside the ->lookup is a problem for case-insensitive
dentries that are not created through ->lookup, like those coming
open-by-fhandle[2], which will not see the required d_ops.

This patchset therefore reverts to using sb->s_d_op to set the dentry operations
for case-insensitive filesystems.  In order to set case-insensitive hooks early
and not require every dentry to have d_revalidate in case-insensitive
filesystems, it introduces a patch suggested by Al Viro to disable d_revalidate
on some dentries on the fly.

It survives fstests encrypt and quick groups without regressions.  Based on
v6.7-rc1.

[1] https://lore.kernel.org/linux-fsdevel/20231123195327.GP38156@ZenIV/
[2] https://lore.kernel.org/linux-fsdevel/20231123171255.GN38156@ZenIV/

Gabriel Krisman Bertazi (10):
  ovl: Reject mounting case-insensitive filesystems
  fscrypt: Share code between functions that prepare lookup
  fscrypt: Drop d_revalidate for valid dentries during lookup
  fscrypt: Drop d_revalidate once the key is added
  libfs: Merge encrypted_ci_dentry_ops and ci_dentry_ops
  libfs: Add helper to choose dentry operations at mount
  ext4: Configure dentry operations at dentry-creation time
  f2fs: Configure dentry operations at dentry-creation time
  ubifs: Configure dentry operations at dentry-creation time
  libfs: Drop generic_set_encrypted_ci_d_ops

 fs/ceph/dir.c           |  2 +-
 fs/ceph/file.c          |  2 +-
 fs/crypto/hooks.c       | 62 +++++++++++++++++++++--------------------
 fs/ext4/namei.c         |  1 -
 fs/ext4/super.c         |  1 +
 fs/f2fs/namei.c         |  1 -
 fs/f2fs/super.c         |  1 +
 fs/libfs.c              | 61 +++++++++++-----------------------------
 fs/overlayfs/params.c   | 13 +++++++--
 fs/ubifs/dir.c          |  1 -
 fs/ubifs/super.c        |  1 +
 include/linux/fs.h      | 11 +++++++-
 include/linux/fscrypt.h | 51 ++++++++++++++++++++-------------
 13 files changed, 106 insertions(+), 102 deletions(-)