When passing an empty string to filepath, it would test if the first character
is '/' which is an invalid access through std::string_view.

Signed-off-by: Patrick Donnelly <[email protected]>
Fixes: https://tracker.ceph.com/issues/66373
(cherry picked from commit e62fce4)

This gets converted in messenger douts.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit da50c6e)

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit d6a5656)

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 8e88fb0)

Convert ms_dispatch to ms_dispatch2 to enable indicating that a map message is
acknowledged and instead of processed (or deliberately not processed).

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit f060ee8)

This avoids messages like:

    2025-02-18T05:31:17.738+0000 7f5206546640  0 ms_deliver_dispatch: unhandled message 0x5632d05f0700 fsmap(e 9) from mon.0 v2:172.21.3.230:40412/0

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 56cb057)

Instead of marking the message as handled, give another component (or Client) a
chance to process.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 616fdd1)

At the time this construction seemed safe since the caller should have a
reference but it could conveivably be the only ref. We don't want the ref count
to reach 0.

Additionally, catch errors so this callback is genuinely noexcept.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 395263c)

For some newer versions of cython, it appears it requires explicitly specifying
noexcept but old versions of Cython 0.29.Z do not understand that attribute.

See: https://cython.readthedocs.io/en/latest/src/userguide/migrating_to_cy30.html#exception-values-and-noexcept

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 90ac740)

invoking cmake with -DWITH_LIBCEPHFS=OFF fails to configure the client target:

CMake Error at src/client/CMakeLists.txt:13 (target_link_libraries):
  Target "client" links to:

    Boost::locale

  but the target was not found.  Possible reasons include:

    * There is a typo in the target name.
    * A find_package call is missing for an IMPORTED target.
    * An ALIAS target is missing.

because the client target is not conditional on WITH_LIBCEPHFS in src/CMakeLists.txt:

add_subdirectory(client)

if(WITH_LIBCEPHFS)

because client is also needed for ceph-fuse, make the client and its
dependencies depend on WITH_LIBCEPHFS OR WITH_FUSE

Signed-off-by: Casey Bodley <[email protected]>
(cherry picked from commit 70eb1d6)

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit a33faa7)

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit d62a379)

Unfortunately, it's not easy to refactor this test into a shared method without
setting up an explicit test class which has been avoided up to this point. So
I'm going to just copy the code. Sorry.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 28bdb8e)

In particular: there's no reason to do a getcwd after chdir.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 3bfd614)

This was missed in the path_walk refactor. readdir is not the only way to "get"
dentry names.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit cda2a6e)

libcephfs semantics require that the old path to the cwd be returned when
getcwd is encounters an unlinked directory in the current working directory.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit c7c5089)

When we wrap a dname after a path walk, we cannot do another path walk.
Originally this method was setup to permit that but with (upcoming) encryption
it cannot work. The dname from the path_walk is in the encrypted namespace so
we cannot do another lookup in a second path_walk.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit af8d267)

Fixes: https://tracker.ceph.com/issues/70409
Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit db27845)

Fixes: 9fbf37a
Fixes: https://tracker.ceph.com/issues/70409
Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 5fc94be)

Conflicts:
	src/mds/journal.cc: code movement caused by referent inode changeset

For inspection from `dump tree` MDS command.

Signed-off-by: Patrick Donnelly <[email protected]>
(cherry picked from commit 7ad1937)