Re: mm: mkfs.ext4 invoked oom-killer on i386 - pagecache_get_page

[Hugh Dickins]

On Thu, 21 May 2020, Johannes Weiner wrote:
> 
> Very much appreciate you guys tracking it down so quickly. Sorry about
> the breakage.
> 
> I think mem_cgroup_disabled() checks are pretty good markers of public
> entry points to the memcg API, so I'd prefer that even if a bit more
> verbose. What do you think?

An explicit mem_cgroup_disabled() check would be fine, but I must admit,
the patch below is rather too verbose for my own taste.  Your call.

> 
> ---
> From cd373ec232942a9bc43ee5e7d2171352019a58fb Mon Sep 17 00:00:00 2001
> From: Hugh Dickins <hughd@xxxxxxxxxx>
> Date: Thu, 21 May 2020 14:58:36 -0400
> Subject: [PATCH] mm: memcontrol: prepare swap controller setup for integration
>  fix
> 
> Fix crash with cgroup_disable=memory:
> 
> > > > > + mkfs -t ext4 /dev/disk/by-id/ata-TOSHIBA_MG04ACA100N_Y8NRK0BPF6XF
> > > > > mke2fs 1.43.8 (1-Jan-2018)
> > > > > Creating filesystem with 244190646 4k blocks and 61054976 inodes
> > > > > Filesystem UUID: 3bb1a285-2cb4-44b4-b6e8-62548f3ac620
> > > > > Superblock backups stored on blocks:
> > > > > 32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
> > > > > 4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968,
> > > > > 102400000, 214990848
> > > > > Allocating group tables:    0/7453                           done
> > > > > Writing inode tables:    0/7453                           done
> > > > > Creating journal (262144 blocks): [   35.502102] BUG: kernel NULL
> > > > > pointer dereference, address: 000000c8
> > > > > [   35.508372] #PF: supervisor read access in kernel mode
> > > > > [   35.513506] #PF: error_code(0x0000) - not-present page
> > > > > [   35.518638] *pde = 00000000
> > > > > [   35.521514] Oops: 0000 [#1] SMP
> > > > > [   35.524652] CPU: 0 PID: 145 Comm: kswapd0 Not tainted
> > > > > 5.7.0-rc6-next-20200519+ #1
> > > > > [   35.532121] Hardware name: Supermicro SYS-5019S-ML/X11SSH-F, BIOS
> > > > > 2.2 05/23/2018
> > > > > [   35.539507] EIP: mem_cgroup_get_nr_swap_pages+0x28/0x60
> 
> do_memsw_account() used to be automatically false when the cgroup
> controller was disabled. Now that it's replaced by
> cgroup_memory_noswap, for which this isn't true, make the
> mem_cgroup_disabled() checks explicit in the swap control API.
> 
> [hannes@xxxxxxxxxxx: use mem_cgroup_disabled() in all API functions]
> Reported-by: Naresh Kamboju <naresh.kamboju@xxxxxxxxxx>
> Debugged-by: Hugh Dickins <hughd@xxxxxxxxxx>
> Debugged-by: Michal Hocko <mhocko@xxxxxxxxxx>
> Signed-off-by: Johannes Weiner <hannes@xxxxxxxxxxx>
> ---
>  mm/memcontrol.c | 47 +++++++++++++++++++++++++++++++++++++++++------
>  1 file changed, 41 insertions(+), 6 deletions(-)

I'm certainly not against a mem_cgroup_disabled() check in the only
place that's been observed to need it, as a fixup to merge into your
original patch; but this seems rather an over-reaction - and I'm a
little surprised that setting mem_cgroup_disabled() doesn't just
force cgroup_memory_noswap, saving repetitious checks elsewhere
(perhaps there's a difficulty in that, I haven't looked).

Historically, I think we've added mem_cgroup_disabled() checks
(accessing a cacheline we'd rather avoid) where they're necessary,
rather than at every "interface".

And you seem to be in a very "goto out" mood today - we all have
our "goto out" days, alternating with our "return 0" days :)

Hugh

> 
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index 3e000a316b59..850bca380562 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -6811,6 +6811,9 @@ void mem_cgroup_swapout(struct page *page, swp_entry_t entry)
>  	VM_BUG_ON_PAGE(PageLRU(page), page);
>  	VM_BUG_ON_PAGE(page_count(page), page);
>  
> +	if (mem_cgroup_disabled())
> +		return;
> +
>  	if (cgroup_subsys_on_dfl(memory_cgrp_subsys))
>  		return;
>  
> @@ -6876,6 +6879,10 @@ int mem_cgroup_try_charge_swap(struct page *page, swp_entry_t entry)
>  	struct mem_cgroup *memcg;
>  	unsigned short oldid;
>  
> +	if (mem_cgroup_disabled())
> +		return 0;
> +
> +	/* Only cgroup2 has swap.max */
>  	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
>  		return 0;
>  
> @@ -6920,6 +6927,9 @@ void mem_cgroup_uncharge_swap(swp_entry_t entry, unsigned int nr_pages)
>  	struct mem_cgroup *memcg;
>  	unsigned short id;
>  
> +	if (mem_cgroup_disabled())
> +		return;
> +
>  	id = swap_cgroup_record(entry, 0, nr_pages);
>  	rcu_read_lock();
>  	memcg = mem_cgroup_from_id(id);
> @@ -6940,12 +6950,25 @@ long mem_cgroup_get_nr_swap_pages(struct mem_cgroup *memcg)
>  {
>  	long nr_swap_pages = get_nr_swap_pages();
>  
> -	if (cgroup_memory_noswap || !cgroup_subsys_on_dfl(memory_cgrp_subsys))
> -		return nr_swap_pages;
> +	if (mem_cgroup_disabled())
> +		goto out;
> +
> +	/* Swap control disabled */
> +	if (cgroup_memory_noswap)
> +		goto out;
> +
> +	/*
> +	 * Only cgroup2 has swap.max, cgroup1 does mem+sw accounting,
> +	 * which does not place restrictions specifically on swap.
> +	 */
> +	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
> +		goto out;
> +
>  	for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg))
>  		nr_swap_pages = min_t(long, nr_swap_pages,
>  				      READ_ONCE(memcg->swap.max) -
>  				      page_counter_read(&memcg->swap));
> +out:
>  	return nr_swap_pages;
>  }
>  
> @@ -6957,18 +6980,30 @@ bool mem_cgroup_swap_full(struct page *page)
>  
>  	if (vm_swap_full())
>  		return true;
> -	if (cgroup_memory_noswap || !cgroup_subsys_on_dfl(memory_cgrp_subsys))
> -		return false;
> +
> +	if (mem_cgroup_disabled())
> +		goto out;
> +
> +	/* Swap control disabled */
> +	if (cgroup_memory_noswap)
> +		goto out;
> +
> +	/*
> +	 * Only cgroup2 has swap.max, cgroup1 does mem+sw accounting,
> +	 * which does not place restrictions specifically on swap.
> +	 */
> +	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
> +		goto out;
>  
>  	memcg = page->mem_cgroup;
>  	if (!memcg)
> -		return false;
> +		goto out;
>  
>  	for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg))
>  		if (page_counter_read(&memcg->swap) * 2 >=
>  		    READ_ONCE(memcg->swap.max))
>  			return true;
> -
> +out:
>  	return false;
>  }
>  
> -- 
> 2.26.2