Patch "[PATCH v2 for-4.9 34/40] slub: move synchronize_sched out of slab_mutex on shrink" has been added to the 4.9-stable tree

[<gregkh@linuxfoundation.org>]

This is a note to let you know that I've just added the patch titled

    [PATCH v2 for-4.9 34/40] slub: move synchronize_sched out of slab_mutex on shrink

to the 4.9-stable tree which can be found at:
    http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     slub-move-synchronize_sched-out-of-slab_mutex-on-shrink.patch
and it can be found in the queue-4.9 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@vger.kernel.org> know about it.


>From foo@baz Mon Mar 20 11:41:01 CET 2017
From: alexander.levin@verizon.com
Date: Fri, 17 Mar 2017 00:48:31 +0000
Subject: [PATCH v2 for-4.9 34/40] slub: move synchronize_sched out of slab_mutex on shrink
To: "gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>
Cc: "stable@vger.kernel.org" <stable@vger.kernel.org>
Message-ID: <20170317004812.26960-34-alexander.levin@verizon.com>

From: Vladimir Davydov <vdavydov.dev@gmail.com>

[ Upstream commit 89e364db71fb5e7fc8d93228152abfa67daf35fa ]

synchronize_sched() is a heavy operation and calling it per each cache
owned by a memory cgroup being destroyed may take quite some time.  What
is worse, it's currently called under the slab_mutex, stalling all works
doing cache creation/destruction.

Actually, there isn't much point in calling synchronize_sched() for each
cache - it's enough to call it just once - after setting cpu_partial for
all caches and before shrinking them.  This way, we can also move it out
of the slab_mutex, which we have to hold for iterating over the slab
cache list.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=172991
Link: http://lkml.kernel.org/r/0a10d71ecae3db00fb4421bcd3f82bcc911f4be4.1475329751.git.vdavydov.dev@gmail.com
Signed-off-by: Vladimir Davydov <vdavydov.dev@gmail.com>
Reported-by: Doug Smythies <dsmythies@telus.net>
Acked-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Pekka Enberg <penberg@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 mm/slab.c        |    4 ++--
 mm/slab.h        |    2 +-
 mm/slab_common.c |   27 +++++++++++++++++++++++++--
 mm/slob.c        |    2 +-
 mm/slub.c        |   19 ++-----------------
 5 files changed, 31 insertions(+), 23 deletions(-)

--- a/mm/slab.c
+++ b/mm/slab.c
@@ -2332,7 +2332,7 @@ out:
 	return nr_freed;
 }
 
-int __kmem_cache_shrink(struct kmem_cache *cachep, bool deactivate)
+int __kmem_cache_shrink(struct kmem_cache *cachep)
 {
 	int ret = 0;
 	int node;
@@ -2352,7 +2352,7 @@ int __kmem_cache_shrink(struct kmem_cach
 
 int __kmem_cache_shutdown(struct kmem_cache *cachep)
 {
-	return __kmem_cache_shrink(cachep, false);
+	return __kmem_cache_shrink(cachep);
 }
 
 void __kmem_cache_release(struct kmem_cache *cachep)
--- a/mm/slab.h
+++ b/mm/slab.h
@@ -146,7 +146,7 @@ static inline unsigned long kmem_cache_f
 
 int __kmem_cache_shutdown(struct kmem_cache *);
 void __kmem_cache_release(struct kmem_cache *);
-int __kmem_cache_shrink(struct kmem_cache *, bool);
+int __kmem_cache_shrink(struct kmem_cache *);
 void slab_kmem_cache_release(struct kmem_cache *);
 
 struct seq_file;
--- a/mm/slab_common.c
+++ b/mm/slab_common.c
@@ -573,6 +573,29 @@ void memcg_deactivate_kmem_caches(struct
 	get_online_cpus();
 	get_online_mems();
 
+#ifdef CONFIG_SLUB
+	/*
+	 * In case of SLUB, we need to disable empty slab caching to
+	 * avoid pinning the offline memory cgroup by freeable kmem
+	 * pages charged to it. SLAB doesn't need this, as it
+	 * periodically purges unused slabs.
+	 */
+	mutex_lock(&slab_mutex);
+	list_for_each_entry(s, &slab_caches, list) {
+		c = is_root_cache(s) ? cache_from_memcg_idx(s, idx) : NULL;
+		if (c) {
+			c->cpu_partial = 0;
+			c->min_partial = 0;
+		}
+	}
+	mutex_unlock(&slab_mutex);
+	/*
+	 * kmem_cache->cpu_partial is checked locklessly (see
+	 * put_cpu_partial()). Make sure the change is visible.
+	 */
+	synchronize_sched();
+#endif
+
 	mutex_lock(&slab_mutex);
 	list_for_each_entry(s, &slab_caches, list) {
 		if (!is_root_cache(s))
@@ -584,7 +607,7 @@ void memcg_deactivate_kmem_caches(struct
 		if (!c)
 			continue;
 
-		__kmem_cache_shrink(c, true);
+		__kmem_cache_shrink(c);
 		arr->entries[idx] = NULL;
 	}
 	mutex_unlock(&slab_mutex);
@@ -755,7 +778,7 @@ int kmem_cache_shrink(struct kmem_cache
 	get_online_cpus();
 	get_online_mems();
 	kasan_cache_shrink(cachep);
-	ret = __kmem_cache_shrink(cachep, false);
+	ret = __kmem_cache_shrink(cachep);
 	put_online_mems();
 	put_online_cpus();
 	return ret;
--- a/mm/slob.c
+++ b/mm/slob.c
@@ -634,7 +634,7 @@ void __kmem_cache_release(struct kmem_ca
 {
 }
 
-int __kmem_cache_shrink(struct kmem_cache *d, bool deactivate)
+int __kmem_cache_shrink(struct kmem_cache *d)
 {
 	return 0;
 }
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -3887,7 +3887,7 @@ EXPORT_SYMBOL(kfree);
  * being allocated from last increasing the chance that the last objects
  * are freed in them.
  */
-int __kmem_cache_shrink(struct kmem_cache *s, bool deactivate)
+int __kmem_cache_shrink(struct kmem_cache *s)
 {
 	int node;
 	int i;
@@ -3899,21 +3899,6 @@ int __kmem_cache_shrink(struct kmem_cach
 	unsigned long flags;
 	int ret = 0;
 
-	if (deactivate) {
-		/*
-		 * Disable empty slabs caching. Used to avoid pinning offline
-		 * memory cgroups by kmem pages that can be freed.
-		 */
-		s->cpu_partial = 0;
-		s->min_partial = 0;
-
-		/*
-		 * s->cpu_partial is checked locklessly (see put_cpu_partial),
-		 * so we have to make sure the change is visible.
-		 */
-		synchronize_sched();
-	}
-
 	flush_all(s);
 	for_each_kmem_cache_node(s, node, n) {
 		INIT_LIST_HEAD(&discard);
@@ -3970,7 +3955,7 @@ static int slab_mem_going_offline_callba
 
 	mutex_lock(&slab_mutex);
 	list_for_each_entry(s, &slab_caches, list)
-		__kmem_cache_shrink(s, false);
+		__kmem_cache_shrink(s);
 	mutex_unlock(&slab_mutex);
 
 	return 0;


Patches currently in stable-queue which might be from gregkh@linuxfoundation.org are

queue-4.9/pci-add-comments-about-rom-bar-updating.patch
queue-4.9/acpi-blacklist-make-dell-latitude-3350-ethernet-work.patch
queue-4.9/s390-zcrypt-introduce-cex6-toleration.patch
queue-4.9/dccp-tcp-fix-routing-redirect-race.patch
queue-4.9/vrf-fix-use-after-free-in-vrf_xmit.patch
queue-4.9/tcp-fix-various-issues-for-sockets-morphing-to-listen-state.patch
queue-4.9/block-allow-write_same-commands-with-the-sg_io-ioctl.patch
queue-4.9/strparser-destroy-workqueue-on-module-exit.patch
queue-4.9/powerpc-mm-fix-build-break-when-cma-n-spapr_tce_iommu-y.patch
queue-4.9/vfio-spapr-postpone-default-window-creation.patch
queue-4.9/vfio-spapr-add-a-helper-to-create-default-dma-window.patch
queue-4.9/pci-do-any-vf-bar-updates-before-enabling-the-bars.patch
queue-4.9/usb-gadget-udc-atmel-remove-memory-leak.patch
queue-4.9/x86-hyperv-handle-unknown-nmis-on-one-cpu-when-unknown_nmi_panic.patch
queue-4.9/net-tunnel-set-inner-protocol-in-network-gro-hooks.patch
queue-4.9/serial-8250_pci-detach-low-level-driver-during-pci-error-recovery.patch
queue-4.9/powerpc-iommu-stop-using-current-in-mm_iommu_xxx.patch
queue-4.9/tun-fix-premature-pollout-notification-on-tun-devices.patch
queue-4.9/vxlan-correctly-validate-vxlan-id-against-vxlan_n_vid.patch
queue-4.9/bpf-fix-regression-on-verifier-pruning-wrt-map-lookups.patch
queue-4.9/tcp-dccp-block-bh-for-syn-processing.patch
queue-4.9/net-sched-act_skbmod-remove-unneeded-rcu_read_unlock-in-tcf_skbmod_dump.patch
queue-4.9/dccp-fix-memory-leak-during-tear-down-of-unsuccessful-connection-request.patch
queue-4.9/xen-do-not-re-use-pirq-number-cached-in-pci-device-msi-msg-data.patch
queue-4.9/vxlan-lock-rcu-on-tx-path.patch
queue-4.9/mlxsw-spectrum_router-avoid-potential-packets-loss.patch
queue-4.9/mpls-do-not-decrement-alive-counter-for-unregister-events.patch
queue-4.9/net-phy-avoid-deadlock-during-phy_error.patch
queue-4.9/uapi-fix-linux-packet_diag.h-userspace-compilation-error.patch
queue-4.9/pci-separate-vf-bar-updates-from-standard-bar-updates.patch
queue-4.9/pci-ignore-bar-updates-on-virtual-functions.patch
queue-4.9/geneve-lock-rcu-on-tx-path.patch
queue-4.9/dccp-fix-use-after-free-in-dccp_feat_activate_values.patch
queue-4.9/l2tp-avoid-use-after-free-caused-by-l2tp_ip_backlog_recv.patch
queue-4.9/powerpc-mm-iommu-vfio-spapr-put-pages-on-vfio-container-shutdown.patch
queue-4.9/bpf-fix-state-equivalence.patch
queue-4.9/scsi-ibmvscsis-clean-up-properly-if-target_submit_cmd-tmr-fails.patch
queue-4.9/drm-nouveau-disp-gp102-fix-cursor-overlay-immediate-channel-indices.patch
queue-4.9/pci-update-bars-using-property-bits-appropriate-for-type.patch
queue-4.9/scsi-ibmvscsis-synchronize-cmds-at-remove-time.patch
queue-4.9/vfio-spapr-postpone-allocation-of-userspace-version-of-tce-table.patch
queue-4.9/ibmveth-calculate-gso_segs-for-large-packets.patch
queue-4.9/net-mlx5e-do-not-reduce-lro-wqe-size-when-not-using-build_skb.patch
queue-4.9/net-sched-actions-decrement-module-reference-count-after-table-flush.patch
queue-4.9/pci-don-t-update-vf-bars-while-vf-memory-space-is-enabled.patch
queue-4.9/ipv4-mask-tos-for-input-route.patch
queue-4.9/net-fix-socket-refcounting-in-skb_complete_tx_timestamp.patch
queue-4.9/net-bridge-allow-ipv6-when-multicast-flood-is-disabled.patch
queue-4.9/net-mlx5e-fix-wrong-cqe-decompression.patch
queue-4.9/net-net_enable_timestamp-can-be-called-from-irq-contexts.patch
queue-4.9/igb-workaround-for-igb-i210-firmware-issue.patch
queue-4.9/drivers-hv-ring_buffer-count-on-wrap-around-mappings-in-get_next_pkt_raw-v2.patch
queue-4.9/drm-nouveau-disp-nv50-specify-ctrl-user-separately-when-constructing-classes.patch
queue-4.9/ipv6-make-ecmp-route-replacement-less-greedy.patch
queue-4.9/ipv6-avoid-write-to-a-possibly-cloned-skb.patch
queue-4.9/pci-remove-pci_resource_bar-and-pci_iov_resource_bar.patch
queue-4.9/mpls-send-route-delete-notifications-when-router-module-is-unloaded.patch
queue-4.9/dmaengine-iota-ioat_alloc_chan_resources-should-not-perform-sleeping-allocations.patch
queue-4.9/scsi-ibmvscsis-return-correct-partition-name-to-client.patch
queue-4.9/vti6-return-gre_key-for-vti6.patch
queue-4.9/vfio-spapr-reference-mm-in-tce_container.patch
queue-4.9/scsi-ibmvscsis-rearrange-functions-for-future-patches.patch
queue-4.9/dccp-unlock-sock-before-calling-sk_free.patch
queue-4.9/bpf-fix-mark_reg_unknown_value-for-spilled-regs-on-map-value-marking.patch
queue-4.9/powerpc-iommu-pass-mm_struct-to-init-cleanup-helpers.patch
queue-4.9/slub-move-synchronize_sched-out-of-slab_mutex-on-shrink.patch
queue-4.9/net-mlx5e-register-unregister-vport-representors-on-interface-attach-detach.patch
queue-4.9/pci-decouple-ioresource_rom_enable-and-pci_rom_address_enable.patch
queue-4.9/net-don-t-call-strlen-on-the-user-buffer-in-packet_bind_spkt.patch
queue-4.9/bpf-detect-identical-ptr_to_map_value_or_null-registers.patch
queue-4.9/scsi-ibmvscsis-issues-from-dan-carpenter-smatch.patch
queue-4.9/vxlan-don-t-allow-overwrite-of-config-src-addr.patch
queue-4.9/acpi-blacklist-add-_rev-quirks-for-dell-precision-5520-and-3520.patch
queue-4.9/bridge-drop-netfilter-fake-rtable-unconditionally.patch
queue-4.9/igb-add-i211-to-i210-phy-workaround.patch
queue-4.9/drm-nouveau-disp-nv50-split-chid-into-chid.ctrl-and-chid.user.patch
queue-4.9/net-fix-socket-refcounting-in-skb_complete_wifi_ack.patch
queue-4.9/scsi-ibmvscsis-synchronize-cmds-at-tpg_enable_store-time.patch
queue-4.9/ipv6-orphan-skbs-in-reassembly-unit.patch
queue-4.9/act_connmark-avoid-crashing-on-malformed-nlattrs-with-null-parms.patch
queue-4.9/uvcvideo-uvc_scan_fallback-for-webcams-with-broken-chain.patch