Drop the -fundamental suffix from src/fundamental/ headers in favor of names
that match their src/basic/ or src/shared/ counterparts (e.g.
macro-fundamental.h -> macro.h, assert-fundamental.h -> assert-util.h,
cleanup-fundamental.h -> cleanup-util.h). Rename src/basic/{btrfs,label}.{c,h}
to use the -util suffix to match the existing shared/btrfs-util and
shared/label-util siblings. Rename src/shared/mkdir-label.{c,h} to mkdir.{c,h}
and src/shared/tmpfile-util-label.{c,h} to tmpfile-util.{c,h} to match the
corresponding src/basic names.
This saves us from having to come up with separate names for files that do
the same thing across tiers, and it makes it easier to move stuff between
src/fundamental, src/basic and src/shared: consumers just #include "foo.h"
and pick up whichever tier their -I path resolves to first, so call sites
don't need to be updated when an API moves between layers.
Where a higher-tier wrapper exists (e.g. src/basic/macro.h wrapping
src/fundamental/macro.h), the wrapper uses an explicit "../fundamental/foo.h"
or "../basic/foo.h" relative include for the lower-tier header. We can't use
GCC's #include_next directive for this — when the wrapper is reachable both
via same-dir-as-source lookup and via -I (e.g. -Isrc/shared) for the
directory it lives in, #include_next advances by exactly one slot in libcpp's
internal directory chain and lands on the same physical directory it was
already in, never reaching the lower-tier sibling (see make_cpp_dir() in
gcc/libcpp/files.cc:1986).
To make sure the right headers are always picked up, the include directories
are reordered so that e.g. src/shared always takes priority over src/basic and
similar for the other directories.
Co-developed-by: Claude Opus 4.7 <noreply@anthropic.com>
Renames src/shared/bpf-dlopen.{c,h} to src/shared/bpf-util.{c,h} and
folds the former src/shared/bpf-compat.h (struct forward decl and
compat_bpf_map_create() helper) into the new header.
Aligns dlopen_bpf() with the standard wrapper pattern: drops the
manual dlopen_safe()/dlsym_many_or_warn()/TAKE_PTR(dl) plumbing and
the bespoke 'cached' int in favor of dlopen_many_sym_or_warn() inside
a FOREACH_STRING() soname-fallback loop.
Unifies declaration of the version-specific symbols (bpf_create_map,
bpf_map_create, bpf_object__next_map, bpf_token_create) into a single
DISABLE_WARNING_REDUNDANT_DECLS block in the header, and alphabetically
merges the DLSYM_PROTOTYPE list. DLSYM_OPTIONAL is used to load each
one — call sites already handle NULL (compat_bpf_map_create() and the
sym_bpf_object__next_map guard in userns-restrict.c). bpf_token_create
additionally defaults to a missing_bpf_token_create() stub returning
-ENOSYS, so callers can branch on the errno instead of NULL-checking
the pointer.
Updates test-bpf-token to match: drops the compile-time
LIBBPF_MAJOR_VERSION ≥ 1.5 gate and the direct <bpf/bpf.h> include in
favor of dlopen_bpf() + sym_bpf_token_create(), and treats -ENOSYS as
the test-skip path (covering both 'libbpf too old' and 'kernel lacks
BPF_TOKEN_CREATE support').
Until very recently all types defined by execute.h started with "Exec"
in the name. I think that was useful, since it made clear that the types
are associated with the ExecContext infrastructure. Let's hence restore
this.
(If we every move these types out of execute.h we should drop the "Exec"
prefix again. But today is not that day.)
No real code changes, just dumb renaming.
Every dlopen_xxx() helper now takes an int log_level argument. It is
passed through to dlopen_many_sym_or_warn() (which in turn propagates it
to dlopen_verbose() for the library-not-installed case), and is used by
the fallback stub when support for the library is not compiled in to
emit a "<lib> support is not compiled in." message at the caller's level.
Callers pass LOG_DEBUG when gracefully degrading, or a higher level when
the failure should surface, and no longer need to log redundantly at the
call site.
As part of this, dlopen_bpf_full() (which already took a log_level) is
merged into dlopen_bpf() rather than keeping both.
The static inline fallbacks used to live in the headers, which required
pulling log.h in from every header that declared a dlopen_xxx(). Move
them into the .c files instead: the declaration is always outside the
#if HAVE_XXX block, the impl sits outside the outer #if HAVE_XXX wrap
with its own internal #if HAVE_XXX/#else/#endif, and apparmor-util.c,
idn-util.c, libmount-util.c and pam-util.c are now always compiled so
they can host their stubs.
Split the single directory fd that chaseat() used to take into two separate
fds: a root_fd that sets the chroot boundary (symlinks may not escape it,
absolute symlinks resolve relative to it), and a dir_fd that path resolution
starts from. This makes the chroot semantics of chaseat() explicit at every
call site instead of encoding them in the CHASE_AT_RESOLVE_IN_ROOT flag,
which is removed. It also decouples the starting directory from the root
boundary, so callers can descend from any inode inside the tree without
having to reopen the root separately.
XAT_FDROOT passed as root_fd means "no containment" (host root); as dir_fd
it means "start at root_fd". For a smoother transition, AT_FDCWD is also
accepted as root_fd and treated as XAT_FDROOT. When root_fd points to a
directory that is actually the host root, it is normalized to XAT_FDROOT
up front so the existing shortcut path can kick in.
Absolute paths returned by chaseat() are now relative to root_fd, and
relative paths are relative to dir_fd. The result is absolute only when
there is no chroot boundary (root_fd is XAT_FDROOT), or when an absolute
symlink made resolution jump out of the dir_fd subtree; otherwise callers
get a relative path they can feed straight back into an openat()-style
call against dir_fd. Specifically, when dir_fd == root_fd and we're not
operating on the host's root directory, we return a relative path even if
we received an absolute path or resolved an absolute symlink to allow
passing the path directly to openat() style functions. We do this to not
have to go modify every caller of chaseat() to make sure they deal properly
with any absolute paths they might receive. Only when root_fd != dir_fd do
we have to return an absolute path to indicate that the path is relative to
root_fd and not dir_fd.
The shortcut that skips the per-component walk is reworked around a new
chase_xopenat() helper that funnels CHASE_NOFOLLOW, CHASE_MUST_BE_* and
CHASE_TRIGGER_AUTOFS through xopenat_full()'s O_NOFOLLOW, O_DIRECTORY,
XO_REGULAR, XO_SOCKET and XO_TRIGGER_AUTOMOUNT flags. As a result these
flags no longer force us off the shortcut and can be dropped from
CHASE_NO_SHORTCUT_MASK, and the old openat_opath_with_automount() helper
goes away. A CHASE_MUST_BE_ANY alias is introduced for shortcut callers
(stat/access paths) that don't go through xopenat_full() and still need
to bail on those flags locally.
All *_and_* helpers built on top of chaseat() (chase_and_openat,
chase_and_opendirat, chase_and_statat, chase_and_accessat,
chase_and_fopenat_unlocked, chase_and_unlinkat, chase_and_open_parent_at)
gain the same root_fd parameter, and every call site in the tree is
ported to the new signature.
chase_and_open() is also fixed to correctly handle CHASE_EXTRACT_FILENAME
without CHASE_PARENT.
History has clearly shown that we're terrible at keeping
the size of the dont_close array up-to-date. Hence let's
step away from a hardcoded max size for that, instead
always collect all fds in the array initializer and
let compiler figure it out, taking advantage of the fact
that close_all_fds() gracefully handles invalid fds in
the 'except' array.
Follow-up for 12001b1bf0
It is part of the keep_fds array already, which is intended
for fds that need to survive until the very end of executor
runtime, i.e. across the second close_all_fds() call.
The owned fds should be closed before close_all_fds() along the lines
of exec_params/runtime_close(), otherwise they would be left dangling
afterwards.
In https://github.com/systemd/systemd/issues/28515, multiple people report that
services that have PrivateTmp=yes and DefaultDependencies=no fail to create the
temporary directories under /tmp, when /tmp is e.g. a bind mount or some other
kind of mount that takes more time.
Before PrivateTmp=disconnected was added, we didn't have a nice solution:
DefaultDependencies=no is used to start services very early, so we wouldn't
want to add a dependency on /tmp automatically. With PrivateTmp=disconnected we
have a fairly nice solution. Let's "upgrade" to this mode automatically.
Strictly speaking, it is a small compat break, but in practice it's unlikely to
matter for early-boot services whether their /tmp is private or disconnected.
The dependency on /tmp that is checked is After. I think this is enough,
since any tmp.mount would be pulled in by local-fs.target and the rest of
the transaction anyway, so we don't need to check more than After.
The asserts are relaxed, because now the two settings can now diverge
in either way.
Resolves https://github.com/systemd/systemd/issues/28515.
[yhndnzj: fix unit_add_exec_dependencies() to handle the new
combination; add a comment in exec_needs_sys_admin()]
The only reason why ExecSharedRuntime.(var_)tmp_dir might
be suppressed is when (/var)/tmp/ is listed in InaccessiblePaths=,
in which case mount namespace is needed anyways. Moreover,
the current logic imposes a chicken-and-egg problem:
at various places in pid1 we'd need to check whether
the unit runs in a mountns, but the ExecSharedRuntime
may not be allocated for it just yet.
This introduces PinnedResources as a structure combining pinned
references to a root directory, root image, or root mstack. This is not
only easier to work with, but essential to make certain unpriv things
work, as we need some mechanism to pin resources before we drop into a
userns which might possibly not provide access anymore to those
resources.
Hence this does two things: introduce the new structure, and immediately
hook it up so that we pin things properly before dropping into userns,
and then makes use of this after dropping the right way, and enables
unpriv userns operation.
The concept is generic enough to eventually implement extension images +
mount images with the same structure, but in order to keep the changes
managable this is left for another time.
(This also makes one further clean-up: client-side verity-reuse checks
are moved server side if we are unpriv. Previously we'd do them client
side, but they were doomed to fail because of lack of privs. Hence let's
drop the client side if we are unpriv and purely do them server-side in
that case.)
So far we opened a new Varlink connection for every mountfsd/nsresourced
method call. Given each tool only does a very small number of calls
(usually 1…5) on them and the connections are cheap this is not too
wasteful. Nonetheless, let's do something about it, and allow reusing
the connection for multiple calls.
This not only makes things a bit more efficient, but has one more
important benefit: Varlink connections pin the security context of the
client when connecting. This means that varlink method calls done with a
connection established while some code was privileged will still operate
as privieged once privs are dropped, until the connection is closed.
This pinning effect is really nice, as it gives us behaviour in a
"capability system" like scheme. Later code is going to use that to
continue doing certain priv userns ops even after unsharing userns and
becoming fully unpriv.
We have two very similar checks in place: in some contexts we want to
know if *any* RootDirectory= is configured, in the other we want to
suppress if it is configured to our regular root. Let's add a helper for
both (even if we only need it once), to make the mirrored behaviour
clear.
Currently, pick_filter_image_any matches any image
with any suffix, which is way more than it should
be doing. It should only be matching images with
the .raw suffix.
Let's address this shortcoming by allowing to pass
multiple filters to path_pick(), and define
pick_filter_image_any as the combination of the
raw and directory image filters.
Fixes#40083
Transparent Hugepages (THP) is a Linux kernel feature that manages
memory using larger pages (2MB on x86, compared to the default 4KB).
The main goal is to improve memory management efficiency and system
performance, especially for memory-intensive applications.
However, it can cause drawbacks in some scenarios, such as memory
regression and latency spikes. THP policy is governed for the entire
system via /sys/kernel/mm/transparent_hugepage/enabled.
However, it can be overridden for individual workloads via prctl(2)
call.
MemoryTHP= is used to disable THPs at exec-invoke to stop
providing THPs for workloads where the drawbacks outweigh the advantages.
When set to "disable", MemoryTHP= disables THPs completely for the
process, irrespecitive of global THP controls.
When set to "madvise", MemoryTHP= disables THPs for the process except
when specifically madvised by the process with MADV_HUGEPAGE or MADV_COLLAPSE.
libpam uses pamh in its function declarations for
the plugin API so let's use the same name in our
tree as well.
Making sure the plugin function definitions match
the plugin function declarations is required to
enable clang-tidy's
readability-inconsistent-declaration-parameter-name
check, but to keep things consistent everywhere we
opt to use pamh tree-wide.
Major distributions already have libseccomp 2.5.x or newer.
Let's bump to the required minimum version to 2.4.0, which provides
SCMP_ACT_KILL_PROCESS, SCMP_ACT_LOG, SCMP_ARCH_PARISC, and
SCMP_ARCH_PARISC64.
Note, libseccomp 2.4.0 was released on 2019-03-15.
See also #38608.
Allow CPUSchedulingPolicy to be set to "ext". SCHED_EXT is a new
scheduling policy in Linux v6.12 that allows processes to be scheduled
using custom BPF schedulers instead of the default in-kernel ones.
Selectively setting the SCHED_EXT policy is useful for systems running
in "partial mode" where not all processes are run using a custom
scheduler.
Fallback to SCHED_OTHER and print an error message for systems where
SCHED_EXT isn't available.
Follow-up for 171ceb4a00
In setup_output() we assume stdout has been set up properly
before stderr, hence the stdout we're inheriting from must
be writable (or more precisely, would have been adjusted to be).
Hence no need to duplicate it.
Note that exec_input_is_inheritable() rightfully refuses EXEC_INPUT_FILE,
in which case std_output would have been reset in service_fix_stdio()
already.
While at it, use the generic fallback logic of first trying user manager
stdout when stdin is not writable.
Now that we dlopen() so many deps, it might happen by accident that we
end up dlopen()ening stuff when we entered a container, which we should
really avoid, to not mix host and container libraries.
Let's add a global variable we can set when we want to block dlopen() to
ever succeed. This is then checked primarily in
dlopen_many_sym_or_warn(), where we'll generate EPERM plus a log
message.
There are a couple of other places we invoke dlopen(), without going
through dlopen_many_sym_or_warn(). This adds the same check there.
Daan De Meyer
Lennart Poettering
Christian Brauner
Zbigniew Jędrzejewski-Szmek
Mike Yuan
Zbigniew Jędrzejewski-Szmek
Lennart Poettering
Daan De Meyer
Usama Arif
Yu Watanabe
Matt Fleming