linux-stable.git/kernel/signal.c, branch v6.9.2

Merge tag 'mm-nonmm-stable-2024-03-14-09-36' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm

2024-03-15T01:03:09+00:00

Pull non-MM updates from Andrew Morton:

 - Kuan-Wei Chiu has developed the well-named series "lib min_heap: Min
   heap optimizations".

 - Kuan-Wei Chiu has also sped up the library sorting code in the series
   "lib/sort: Optimize the number of swaps and comparisons".

 - Alexey Gladkov has added the ability for code running within an IPC
   namespace to alter its IPC and MQ limits. The series is "Allow to
   change ipc/mq sysctls inside ipc namespace".

 - Geert Uytterhoeven has contributed some dhrystone maintenance work in
   the series "lib: dhry: miscellaneous cleanups".

 - Ryusuke Konishi continues nilfs2 maintenance work in the series

	"nilfs2: eliminate kmap and kmap_atomic calls"
	"nilfs2: fix kernel bug at submit_bh_wbc()"

 - Nathan Chancellor has updated our build tools requirements in the
   series "Bump the minimum supported version of LLVM to 13.0.1".

 - Muhammad Usama Anjum continues with the selftests maintenance work in
   the series "selftests/mm: Improve run_vmtests.sh".

 - Oleg Nesterov has done some maintenance work against the signal code
   in the series "get_signal: minor cleanups and fix".

Plus the usual shower of singleton patches in various parts of the tree.
Please see the individual changelogs for details.

* tag 'mm-nonmm-stable-2024-03-14-09-36' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (77 commits)
  nilfs2: prevent kernel bug at submit_bh_wbc()
  nilfs2: fix failure to detect DAT corruption in btree and direct mappings
  ocfs2: enable ocfs2_listxattr for special files
  ocfs2: remove SLAB_MEM_SPREAD flag usage
  assoc_array: fix the return value in assoc_array_insert_mid_shortcut()
  buildid: use kmap_local_page()
  watchdog/core: remove sysctl handlers from public header
  nilfs2: use div64_ul() instead of do_div()
  mul_u64_u64_div_u64: increase precision by conditionally swapping a and b
  kexec: copy only happens before uchunk goes to zero
  get_signal: don't initialize ksig->info if SIGNAL_GROUP_EXIT/group_exec_task
  get_signal: hide_si_addr_tag_bits: fix the usage of uninitialized ksig
  get_signal: don't abuse ksig->info.si_signo and ksig->sig
  const_structs.checkpatch: add device_type
  Normalise "name (ad@dr)" MODULE_AUTHORs to "name "
  dyndbg: replace kstrdup() + strchr() with kstrdup_and_replace()
  list: leverage list_is_head() for list_entry_is_head()
  nilfs2: MAINTAINERS: drop unreachable project mirror site
  smp: make __smp_processor_id() 0-argument macro
  fat: fix uninitialized field in nostale filehandles
  ...

get_signal: don't initialize ksig->info if SIGNAL_GROUP_EXIT/group_exec_task

2024-03-06T21:07:39+00:00

This initialization is incomplete and unnecessary, neither do_group_exit()
nor PF_USER_WORKER need ksig->info.

Link: https://lkml.kernel.org/r/20240226165653.GA20834@redhat.com
Signed-off-by: Oleg Nesterov 
Cc: Christian Brauner 
Cc: Eric W. Biederman 
Cc: Peter Collingbourne 
Cc: Wen Yang 
Signed-off-by: Andrew Morton

get_signal: hide_si_addr_tag_bits: fix the usage of uninitialized ksig

2024-03-06T21:07:39+00:00

ksig->ka and ksig->info are not initialized if get_signal() returns 0 or
if the caller is PF_USER_WORKER.

Check signr != 0 before SA_EXPOSE_TAGBITS and move the "out" label down.

The latter means that ksig->sig won't be initialized if a PF_USER_WORKER
thread gets a fatal signal but this is fine, PF_USER_WORKER's don't use
ksig. And there is nothing new, in this case ksig->ka and ksig-info are
not initialized anyway. Add a comment.

Link: https://lkml.kernel.org/r/20240226165650.GA20829@redhat.com
Signed-off-by: Oleg Nesterov 
Cc: Christian Brauner 
Cc: Eric W. Biederman 
Cc: Peter Collingbourne 
Cc: Wen Yang 
Signed-off-by: Andrew Morton

get_signal: don't abuse ksig->info.si_signo and ksig->sig

2024-03-06T21:07:39+00:00

Patch series "get_signal: minor cleanups and fix".

Lets remove this clear_siginfo() right now.  It is incomplete (and thus
looks confusing) and unnecessary.  Also, PF_USER_WORKER's already don't
get a fully initialized ksig anyway.


This patch (of 3):

Cleanup and preparation for the next changes.

get_signal() uses signr or ksig->info.si_signo or ksig->sig in a chaotic
way, this looks confusing. Change it to always use signr.

Link: https://lkml.kernel.org/r/20240226165612.GA20787@redhat.com
Link: https://lkml.kernel.org/r/20240226165647.GA20826@redhat.com
Signed-off-by: Oleg Nesterov 
Cc: Christian Brauner 
Cc: Eric W. Biederman 
Cc: Peter Collingbourne 
Cc: Wen Yang 
Signed-off-by: Andrew Morton

pidfd: allow to override signal scope in pidfd_send_signal()

2024-02-21T08:46:08+00:00

Right now we determine the scope of the signal based on the type of
pidfd. There are use-cases where it's useful to override the scope of
the signal. For example in [1]. Add flags to determine the scope of the
signal:

(1) PIDFD_SIGNAL_THREAD: send signal to specific thread reference by @pidfd
(2) PIDFD_SIGNAL_THREAD_GROUP: send signal to thread-group of @pidfd
(2) PIDFD_SIGNAL_PROCESS_GROUP: send signal to process-group of @pidfd

Since we now allow specifying PIDFD_SEND_PROCESS_GROUP for
pidfd_send_signal() to send signals to process groups we need to adjust
the check restricting si_code emulation by userspace to account for
PIDTYPE_PGID.

Reviewed-by: Oleg Nesterov 
Link: https://github.com/systemd/systemd/issues/31093 [1]
Link: https://lore.kernel.org/r/20240210-chihuahua-hinzog-3945b6abd44a@brauner
Link: https://lore.kernel.org/r/20240214123655.GB16265@redhat.com
Signed-off-by: Christian Brauner

pidfd: change pidfd_send_signal() to respect PIDFD_THREAD

2024-02-10T21:37:23+00:00

Turn kill_pid_info() into kill_pid_info_type(), this allows to pass any
pid_type to group_send_sig_info(), despite its name it should work fine
even if type = PIDTYPE_PID.

Change pidfd_send_signal() to use PIDTYPE_PID or PIDTYPE_TGID depending
on PIDFD_THREAD.

While at it kill another TODO comment in pidfd_show_fdinfo(). As Christian
expains fdinfo reports f_flags, userspace can already detect PIDFD_THREAD.

Reviewed-by: Tycho Andersen 
Signed-off-by: Oleg Nesterov 
Link: https://lore.kernel.org/r/20240209130650.GA8048@redhat.com
Signed-off-by: Christian Brauner

signal: fill in si_code in prepare_kill_siginfo()

2024-02-10T21:37:18+00:00

So that do_tkill() can use this helper too. This also simplifies
the next patch.

TODO: perhaps we can kill prepare_kill_siginfo() and change the
callers to use SEND_SIG_NOINFO,  but this needs some changes in
__send_signal_locked() and TP_STORE_SIGINFO().

Reviewed-by: Tycho Andersen 
Signed-off-by: Oleg Nesterov 
Link: https://lore.kernel.org/r/20240209130620.GA8039@redhat.com
Signed-off-by: Christian Brauner

pidfd: change do_notify_pidfd() to use __wake_up(poll_to_key(EPOLLIN))

2024-02-06T12:52:51+00:00

rather than wake_up_all(). This way do_notify_pidfd() won't wakeup the
POLLHUP-only waiters which wait for pid_task() == NULL.

TODO:
    - as Christian pointed out, this asks for the new wake_up_all_poll()
      helper, it can already have other users.

    - we can probably discriminate the PIDFD_THREAD and non-PIDFD_THREAD
      waiters, but this needs more work. See
      https://lore.kernel.org/all/20240205140848.GA15853@redhat.com/

Signed-off-by: Oleg Nesterov 
Link: https://lore.kernel.org/r/20240205141348.GA16539@redhat.com
Reviewed-by: Tycho Andersen 
Signed-off-by: Christian Brauner

pidfd: implement PIDFD_THREAD flag for pidfd_open()

2024-02-02T12:12:28+00:00

With this flag:

	- pidfd_open() doesn't require that the target task must be
	  a thread-group leader

	- pidfd_poll() succeeds when the task exits and becomes a
	  zombie (iow, passes exit_notify()), even if it is a leader
	  and thread-group is not empty.

	  This means that the behaviour of pidfd_poll(PIDFD_THREAD,
	  pid-of-group-leader) is not well defined if it races with
	  exec() from its sub-thread; pidfd_poll() can succeed or not
	  depending on whether pidfd_task_exited() is called before
	  or after exchange_tids().

	  Perhaps we can improve this behaviour later, pidfd_poll()
	  can probably take sig->group_exec_task into account. But
	  this doesn't really differ from the case when the leader
	  exits before other threads (so pidfd_poll() succeeds) and
	  then another thread execs and pidfd_poll() will block again.

thread_group_exited() is no longer used, perhaps it can die.

Co-developed-by: Tycho Andersen 
Signed-off-by: Oleg Nesterov 
Link: https://lore.kernel.org/r/20240131132602.GA23641@redhat.com
Tested-by: Tycho Andersen 
Reviewed-by: Tycho Andersen 
Signed-off-by: Christian Brauner

pidfd: don't do_notify_pidfd() if !thread_group_empty()

2024-02-02T12:12:28+00:00

do_notify_pidfd() makes no sense until the whole thread group exits, change
do_notify_parent() to check thread_group_empty().

This avoids the unnecessary do_notify_pidfd() when tsk is not a leader, or
it exits before other threads, or it has a ptraced EXIT_ZOMBIE sub-thread.

Signed-off-by: Oleg Nesterov 
Link: https://lore.kernel.org/r/20240127132407.GA29136@redhat.com
Reviewed-by: Tycho Andersen 
Signed-off-by: Christian Brauner