Re: [PATCH] mm, memcg: fix reclaim deadlock with writeback

["Kirill A. Shutemov" <kirill@shutemov.name>]

On Tue, Dec 11, 2018 at 02:26:45PM +0100, Michal Hocko wrote:
> From: Michal Hocko <mhocko@suse.com>
> 
> Liu Bo has experienced a deadlock between memcg (legacy) reclaim and the
> ext4 writeback
> task1:
> [<ffffffff811aaa52>] wait_on_page_bit+0x82/0xa0
> [<ffffffff811c5777>] shrink_page_list+0x907/0x960
> [<ffffffff811c6027>] shrink_inactive_list+0x2c7/0x680
> [<ffffffff811c6ba4>] shrink_node_memcg+0x404/0x830
> [<ffffffff811c70a8>] shrink_node+0xd8/0x300
> [<ffffffff811c73dd>] do_try_to_free_pages+0x10d/0x330
> [<ffffffff811c7865>] try_to_free_mem_cgroup_pages+0xd5/0x1b0
> [<ffffffff8122df2d>] try_charge+0x14d/0x720
> [<ffffffff812320cc>] memcg_kmem_charge_memcg+0x3c/0xa0
> [<ffffffff812321ae>] memcg_kmem_charge+0x7e/0xd0
> [<ffffffff811b68a8>] __alloc_pages_nodemask+0x178/0x260
> [<ffffffff8120bff5>] alloc_pages_current+0x95/0x140
> [<ffffffff81074247>] pte_alloc_one+0x17/0x40
> [<ffffffff811e34de>] __pte_alloc+0x1e/0x110
> [<ffffffffa06739de>] alloc_set_pte+0x5fe/0xc20
> [<ffffffff811e5d93>] do_fault+0x103/0x970
> [<ffffffff811e6e5e>] handle_mm_fault+0x61e/0xd10
> [<ffffffff8106ea02>] __do_page_fault+0x252/0x4d0
> [<ffffffff8106ecb0>] do_page_fault+0x30/0x80
> [<ffffffff8171bce8>] page_fault+0x28/0x30
> [<ffffffffffffffff>] 0xffffffffffffffff
> 
> task2:
> [<ffffffff811aadc6>] __lock_page+0x86/0xa0
> [<ffffffffa02f1e47>] mpage_prepare_extent_to_map+0x2e7/0x310 [ext4]
> [<ffffffffa08a2689>] ext4_writepages+0x479/0xd60
> [<ffffffff811bbede>] do_writepages+0x1e/0x30
> [<ffffffff812725e5>] __writeback_single_inode+0x45/0x320
> [<ffffffff81272de2>] writeback_sb_inodes+0x272/0x600
> [<ffffffff81273202>] __writeback_inodes_wb+0x92/0xc0
> [<ffffffff81273568>] wb_writeback+0x268/0x300
> [<ffffffff81273d24>] wb_workfn+0xb4/0x390
> [<ffffffff810a2f19>] process_one_work+0x189/0x420
> [<ffffffff810a31fe>] worker_thread+0x4e/0x4b0
> [<ffffffff810a9786>] kthread+0xe6/0x100
> [<ffffffff8171a9a1>] ret_from_fork+0x41/0x50
> [<ffffffffffffffff>] 0xffffffffffffffff
> 
> He adds
> : task1 is waiting for the PageWriteback bit of the page that task2 has
> : collected in mpd->io_submit->io_bio, and tasks2 is waiting for the LOCKED
> : bit the page which tasks1 has locked.
> 
> More precisely task1 is handling a page fault and it has a page locked
> while it charges a new page table to a memcg. That in turn hits a memory
> limit reclaim and the memcg reclaim for legacy controller is waiting on
> the writeback but that is never going to finish because the writeback
> itself is waiting for the page locked in the #PF path. So this is
> essentially ABBA deadlock.
> 
> Waiting for the writeback in legacy memcg controller is a workaround
> for pre-mature OOM killer invocations because there is no dirty IO
> throttling available for the controller. There is no easy way around
> that unfortunately. Therefore fix this specific issue by pre-allocating
> the page table outside of the page lock. We have that handy
> infrastructure for that already so simply reuse the fault-around pattern
> which already does this.
> 
> Reported-and-Debugged-by: Liu Bo <bo.liu@linux.alibaba.com>
> Signed-off-by: Michal Hocko <mhocko@suse.com>
> ---
> Hi,
> this has been originally reported here [1]. While it could get worked
> around in the fs, catching the allocation early sounds like a preferable
> approach. Liu Bo has noted that he is not able to reproduce this anymore
> because kmem accounting has been disabled in their workload but this
> should be quite straightforward to review.
> 
> There are probably other hidden __GFP_ACCOUNT | GFP_KERNEL allocations
> from under a fs page locked but they should be really rare. I am not
> aware of a better solution unfortunately.
> 
> I would appreciate if Kirril could have a look and double check I am not
> doing something stupid here.
> 
> Debugging lock_page deadlocks is an absolute PITA considering a lack of
> lockdep support so I would mark it for stable.
> 
> [1] http://lkml.kernel.org/r/1540858969-75803-1-git-send-email-bo.liu@linux.alibaba.com
>  mm/memory.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/mm/memory.c b/mm/memory.c
> index 4ad2d293ddc2..1a73d2d4659e 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -2993,6 +2993,17 @@ static vm_fault_t __do_fault(struct vm_fault *vmf)
>  	struct vm_area_struct *vma = vmf->vma;
>  	vm_fault_t ret;
>  
> +	/*
> +	 * Preallocate pte before we take page_lock because this might lead to
> +	 * deadlocks for memcg reclaim which waits for pages under writeback.
> +	 */
> +	if (pmd_none(*vmf->pmd) && !vmf->prealloc_pte) {
> +		vmf->prealloc_pte = pte_alloc_one(vmf->vma->vm>mm, vmf->address);
> +		if (!vmf->prealloc_pte)
> +			return VM_FAULT_OOM;
> +		smp_wmb(); /* See comment in __pte_alloc() */
> +	}
> +
>  	ret = vma->vm_ops->fault(vmf);
>  	if (unlikely(ret & (VM_FAULT_ERROR | VM_FAULT_NOPAGE | VM_FAULT_RETRY |
>  			    VM_FAULT_DONE_COW)))

Sorry, but I don't think it fixes anything. Just hides it a level deeper.

The trick with ->prealloc_pte works for faultaround because we can rely on
->map_pages() to not sleep and we know how it will setup page table entry.
Basically, core controls most of the path.

It's not the case with ->fault(). It is free to sleep and allocate
whatever it wants.

For instance, DAX page fault will setup page table entry on its own and
return VM_FAULT_NOPAGE. It uses vmf_insert_mixed() to setup the page table
and ignores your pre-allocated page table.

But it's just an example. The problem is that ->fault() is not bounded on
what it can do, unlike ->map_pages().

-- 
 Kirill A. Shutemov