[PATCH] mm, memcg: remove hotplug locking from try_charge

From: Michal Hocko
Date: Wed Sep 13 2017 - 05:00:41 EST

From: Michal Hocko <mhocko@xxxxxxxx>

The following lockdep splat has been noticed during LTP testing

[21002.630252] ======================================================
[21002.637148] WARNING: possible circular locking dependency detected
[21002.644045] 4.13.0-rc3-next-20170807 #12 Not tainted
[21002.649583] ------------------------------------------------------
[21002.656492] a.out/4771 is trying to acquire lock:
[21002.661742] (cpu_hotplug_lock.rw_sem){++++++}, at: [<ffffffff812b4668>] drain_all_stock.part.35+0x18/0x140
[21002.672629]
[21002.672629] but task is already holding lock:
[21002.679137] (&mm->mmap_sem){++++++}, at: [<ffffffff8106eb35>] __do_page_fault+0x175/0x530
[21002.688371]
[21002.688371] which lock already depends on the new lock.
[21002.688371]
[21002.697505]
[21002.697505] the existing dependency chain (in reverse order) is:
[21002.705856]
[21002.705856] -> #3 (&mm->mmap_sem){++++++}:
[21002.712080] lock_acquire+0xc9/0x230
[21002.716661] __might_fault+0x70/0xa0
[21002.721241] _copy_to_user+0x23/0x70
[21002.725814] filldir+0xa7/0x110
[21002.729988] xfs_dir2_sf_getdents.isra.10+0x20c/0x2c0 [xfs]
[21002.736840] xfs_readdir+0x1fa/0x2c0 [xfs]
[21002.742042] xfs_file_readdir+0x30/0x40 [xfs]
[21002.747485] iterate_dir+0x17a/0x1a0
[21002.752057] SyS_getdents+0xb0/0x160
[21002.756638] entry_SYSCALL_64_fastpath+0x1f/0xbe
[21002.762371]
[21002.762371] -> #2 (&type->i_mutex_dir_key#3){++++++}:
[21002.769661] lock_acquire+0xc9/0x230
[21002.774239] down_read+0x51/0xb0
[21002.778429] lookup_slow+0xde/0x210
[21002.782903] walk_component+0x160/0x250
[21002.787765] link_path_walk+0x1a6/0x610
[21002.792625] path_openat+0xe4/0xd50
[21002.797100] do_filp_open+0x91/0x100
[21002.801673] file_open_name+0xf5/0x130
[21002.806429] filp_open+0x33/0x50
[21002.810620] kernel_read_file_from_path+0x39/0x80
[21002.816459] _request_firmware+0x39f/0x880
[21002.821610] request_firmware_direct+0x37/0x50
[21002.827151] request_microcode_fw+0x64/0xe0
[21002.832401] reload_store+0xf7/0x180
[21002.836974] dev_attr_store+0x18/0x30
[21002.841641] sysfs_kf_write+0x44/0x60
[21002.846318] kernfs_fop_write+0x113/0x1a0
[21002.851374] __vfs_write+0x37/0x170
[21002.855849] vfs_write+0xc7/0x1c0
[21002.860128] SyS_write+0x58/0xc0
[21002.864313] do_syscall_64+0x6c/0x1f0
[21002.868973] return_from_SYSCALL_64+0x0/0x7a
[21002.874317]
[21002.874317] -> #1 (microcode_mutex){+.+.+.}:
[21002.880748] lock_acquire+0xc9/0x230
[21002.885322] __mutex_lock+0x88/0x960
[21002.889894] mutex_lock_nested+0x1b/0x20
[21002.894854] microcode_init+0xbb/0x208
[21002.899617] do_one_initcall+0x51/0x1a9
[21002.904481] kernel_init_freeable+0x208/0x2a7
[21002.909922] kernel_init+0xe/0x104
[21002.914298] ret_from_fork+0x2a/0x40
[21002.918867]
[21002.918867] -> #0 (cpu_hotplug_lock.rw_sem){++++++}:
[21002.926058] __lock_acquire+0x153c/0x1550
[21002.931112] lock_acquire+0xc9/0x230
[21002.935688] cpus_read_lock+0x4b/0x90
[21002.940353] drain_all_stock.part.35+0x18/0x140
[21002.945987] try_charge+0x3ab/0x6e0
[21002.950460] mem_cgroup_try_charge+0x7f/0x2c0
[21002.955902] shmem_getpage_gfp+0x25f/0x1050
[21002.961149] shmem_fault+0x96/0x200
[21002.965621] __do_fault+0x1e/0xa0
[21002.969905] __handle_mm_fault+0x9c3/0xe00
[21002.975056] handle_mm_fault+0x16e/0x380
[21002.980013] __do_page_fault+0x24a/0x530
[21002.984968] do_page_fault+0x30/0x80
[21002.989537] page_fault+0x28/0x30
[21002.993812]
[21002.993812] other info that might help us debug this:
[21002.993812]
[21003.002744] Chain exists of:
[21003.002744] cpu_hotplug_lock.rw_sem --> &type->i_mutex_dir_key#3 --> &mm->mmap_sem
[21003.002744]
[21003.016238] Possible unsafe locking scenario:
[21003.016238]
[21003.022843] CPU0 CPU1
[21003.027896] ---- ----
[21003.032948] lock(&mm->mmap_sem);
[21003.036741] lock(&type->i_mutex_dir_key#3);
[21003.044419] lock(&mm->mmap_sem);
[21003.051025] lock(cpu_hotplug_lock.rw_sem);
[21003.055788]
[21003.055788] *** DEADLOCK ***
[21003.055788]
[21003.062393] 2 locks held by a.out/4771:
[21003.066675] #0: (&mm->mmap_sem){++++++}, at: [<ffffffff8106eb35>] __do_page_fault+0x175/0x530
[21003.076391] #1: (percpu_charge_mutex){+.+...}, at: [<ffffffff812b4c97>] try_charge+0x397/0x6e0

The problem is very similar to the one fixed by a459eeb7b852 ("mm,
page_alloc: do not depend on cpu hotplug locks inside the allocator").
We are taking hotplug locks while we can be sitting on top of basically
arbitrary locks. This just calls for problems.

We can get rid of {get,put}_online_cpus, fortunately. We do not have to
be worried about races with memory hotplug because drain_local_stock,
which is called from both the WQ draining and the memory hotplug
contexts, is always operating on the local cpu stock with IRQs disabled.

The only thing to be careful about is that the target memcg doesn't
vanish while we are still in drain_all_stock so take a reference on
it.

Reported-and-tested-by: Artem Savkov <asavkov@xxxxxxxxxx>
Signed-off-by: Michal Hocko <mhocko@xxxxxxxx>
---
Hi,
this has been reported [1] and the first fix posted [2]. I am reposting
it with the comment updated based on the review feedback from Thomas.
The issue is there for quite some time but I didn't get to check when
it has been introduced. Nor I have marked it for stable tree backport
because I am not sure it is serious enough. I will help with backporting
if there is a demand for it.

Can we have this merged?

[1] http://lkml.kernel.org/r/20170807140947.nhfz2gel6wytl6ia@xxxxxxxxxxxxxxxxxxxxxxxxx
[2] http://lkml.kernel.org/r/20170830154315.sa57wasw64rvnuhe@xxxxxxxxxxxxxx

mm/memcontrol.c | 20 +++++++++++++++-----
1 file changed, 15 insertions(+), 5 deletions(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 15af3da5af02..696c6529e900 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -1777,6 +1777,10 @@ static void drain_local_stock(struct work_struct *dummy)
struct memcg_stock_pcp *stock;
unsigned long flags;

+ /*
+ * The only protection from memory hotplug vs. drain_stock races is
+ * that we always operate on local CPU stock here with IRQ disabled
+ */
local_irq_save(flags);

stock = this_cpu_ptr(&memcg_stock);
@@ -1821,27 +1825,33 @@ static void drain_all_stock(struct mem_cgroup *root_memcg)
/* If someone's already draining, avoid adding running more workers. */
if (!mutex_trylock(&percpu_charge_mutex))
return;
- /* Notify other cpus that system-wide "drain" is running */
- get_online_cpus();
+ /*
+ * Notify other cpus that system-wide "drain" is running
+ * We do not care about races with the cpu hotplug because cpu down
+ * as well as workers from this path always operate on the local
+ * per-cpu data. CPU up doesn't touch memcg_stock at all.
+ */
curcpu = get_cpu();
for_each_online_cpu(cpu) {
struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu);
struct mem_cgroup *memcg;

memcg = stock->cached;
- if (!memcg || !stock->nr_pages)
+ if (!memcg || !stock->nr_pages || !css_tryget(&memcg->css))
continue;
- if (!mem_cgroup_is_descendant(memcg, root_memcg))
+ if (!mem_cgroup_is_descendant(memcg, root_memcg)) {
+ css_put(&memcg->css);
continue;
+ }
if (!test_and_set_bit(FLUSHING_CACHED_CHARGE, &stock->flags)) {
if (cpu == curcpu)
drain_local_stock(&stock->work);
else
schedule_work_on(cpu, &stock->work);
}
+ css_put(&memcg->css);
}
put_cpu();
- put_online_cpus();
mutex_unlock(&percpu_charge_mutex);
}

--
2.14.1