Merge 4.19.318 into android-4.19-stable

Changes in 4.19.318
	asm-generic: Move common compat types to asm-generic/compat.h
	media: dvb: as102-fe: Fix as10x_register_addr packing
	media: dvb-usb: dib0700_devices: Add missing release_firmware()
	IB/core: Implement a limit on UMAD receive List
	drm/amd/display: Skip finding free audio for unknown engine_id
	media: dw2102: Don't translate i2c read into write
	sctp: prefer struct_size over open coded arithmetic
	firmware: dmi: Stop decoding on broken entry
	Input: ff-core - prefer struct_size over open coded arithmetic
	net: dsa: mv88e6xxx: Correct check for empty list
	media: dvb-frontends: tda18271c2dd: Remove casting during div
	media: s2255: Use refcount_t instead of atomic_t for num_channels
	media: dvb-frontends: tda10048: Fix integer overflow
	i2c: i801: Annotate apanel_addr as __ro_after_init
	powerpc/64: Set _IO_BASE to POISON_POINTER_DELTA not 0 for CONFIG_PCI=n
	orangefs: fix out-of-bounds fsid access
	powerpc/xmon: Check cpu id in commands "c#", "dp#" and "dx#"
	jffs2: Fix potential illegal address access in jffs2_free_inode
	s390/pkey: Wipe sensitive data on failure
	tcp: take care of compressed acks in tcp_add_reno_sack()
	tcp: tcp_mark_head_lost is only valid for sack-tcp
	tcp: add ece_ack flag to reno sack functions
	net: tcp better handling of reordering then loss cases
	UPSTREAM: tcp: fix DSACK undo in fast recovery to call tcp_try_to_open()
	tcp_metrics: validate source addr length
	bonding: Fix out-of-bounds read in bond_option_arp_ip_targets_set()
	selftests: fix OOM in msg_zerocopy selftest
	selftests: make order checking verbose in msg_zerocopy selftest
	inet_diag: Initialize pad field in struct inet_diag_req_v2
	nilfs2: fix inode number range checks
	nilfs2: add missing check for inode numbers on directory entries
	mm: optimize the redundant loop of mm_update_owner_next()
	Bluetooth: Fix incorrect pointer arithmatic in ext_adv_report_evt
	can: kvaser_usb: Explicitly initialize family in leafimx driver_info struct
	fsnotify: Do not generate events for O_PATH file descriptors
	Revert "mm/writeback: fix possible divide-by-zero in wb_dirty_limits(), again"
	drm/nouveau: fix null pointer dereference in nouveau_connector_get_modes
	drm/amdgpu/atomfirmware: silence UBSAN warning
	bnx2x: Fix multiple UBSAN array-index-out-of-bounds
	media: dw2102: fix a potential buffer overflow
	i2c: pnx: Fix potential deadlock warning from del_timer_sync() call in isr
	nilfs2: fix incorrect inode allocation from reserved inodes
	drm/i915: make find_fw_domain work on intel_uncore
	tcp: fix incorrect undo caused by DSACK of TLP retransmit
	net: lantiq_etop: add blank line after declaration
	net: ethernet: lantiq_etop: fix double free in detach
	ppp: reject claimed-as-LCP but actually malformed packets
	ARM: davinci: Convert comma to semicolon
	USB: serial: option: add Telit generic core-dump composition
	USB: serial: option: add Telit FN912 rmnet compositions
	USB: serial: option: add Fibocom FM350-GL
	USB: serial: option: add support for Foxconn T99W651
	USB: serial: option: add Netprisma LCUK54 series modules
	USB: serial: option: add Rolling RW350-GL variants
	USB: Add USB_QUIRK_NO_SET_INTF quirk for START BP-850k
	usb: gadget: configfs: Prevent OOB read/write in usb_string_copy()
	USB: core: Fix duplicate endpoint bug by clearing reserved bits in the descriptor
	hpet: Support 32-bit userspace
	libceph: fix race between delayed_work() and ceph_monc_stop()
	tcp: refactor tcp_retransmit_timer()
	net: tcp: fix unexcepted socket die when snd_wnd is 0
	tcp: use signed arithmetic in tcp_rtx_probe0_timed_out()
	tcp: avoid too many retransmit packets
	SUNRPC: Fix RPC client cleaned up the freed pipefs dentries
	nilfs2: fix kernel bug on rename operation of broken directory
	i2c: rcar: bring hardware to known state when probing
	Linux 4.19.318

Change-Id: I6d2646a308c3f44976d00ee372e87568c3e40c23
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>

kernel/exit.c

@@ -494,6 +494,8 @@ void mm_update_next_owner(struct mm_struct *mm)
 	 * Search through everything else, we should not get here often.
 	 */
 	for_each_process(g) {
+		if (atomic_read(&mm->mm_users) <= 1)
+			break;
 		if (g->flags & PF_KTHREAD)
 			continue;
 		for_each_thread(g, c) {