Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH] mm/filemap: Adding missing mem_cgroup_uncharge() to
 __add_to_page_cache_locked()
To:     Matthew Wilcox <willy@infradead.org>
Cc:     Andrew Morton <akpm@linux-foundation.org>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Alex Shi <alex.shi@linux.alibaba.com>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
References: <20210125042441.20030-1-longman@redhat.com>
 <20210125043631.GD308988@casper.infradead.org>
From:   Waiman Long <longman@redhat.com>
Organization: Red Hat
Message-ID: <1f234e4f-be0d-ecf9-7934-4be3de0d6fe5@redhat.com>
Date:   Mon, 25 Jan 2021 09:09:26 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.6.0
MIME-Version: 1.0
In-Reply-To: <20210125043631.GD308988@casper.infradead.org>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-US
Precedence: bulk

On 1/24/21 11:36 PM, Matthew Wilcox wrote:
> On Sun, Jan 24, 2021 at 11:24:41PM -0500, Waiman Long wrote:
>> diff --git a/mm/filemap.c b/mm/filemap.c
>> index 5c9d564317a5..aa0e0fb04670 100644
>> --- a/mm/filemap.c
>> +++ b/mm/filemap.c
>> @@ -835,6 +835,7 @@ noinline int __add_to_page_cache_locked(struct page *page,
>>   	XA_STATE(xas, &mapping->i_pages, offset);
>>   	int huge = PageHuge(page);
>>   	int error;
>> +	bool charged = false;
> I don't think we need this extra bool.
>
>> @@ -896,6 +898,8 @@ noinline int __add_to_page_cache_locked(struct page *page,
>>   
>>   	if (xas_error(&xas)) {
>>   		error = xas_error(&xas);
>> +		if (charged)
>> +			mem_cgroup_uncharge(page);
>>   		goto error;
>>   	}
> Better:
>
> -		goto error;
> +		goto uncharge;
> ...
> +uncharge:
> +	if (!huge)
> +		mem_cgroup_uncharge(page);
>   error:
> ...
>
That was my original plan. After finding out there was a potentially 
conflicting patch in linux-next:

commit 7a02fa97b897 ("secretmem: add memcg accounting")

@@ -839,7 +840,7 @@ noinline int __add_to_page_cache_locked(struct page 
*page,
         page->mapping = mapping;
         page->index = offset;

-       if (!huge) {
+       if (!huge && !page_is_secretmem(page)) {
                 error = mem_cgroup_charge(page, current->mm, gfp);
                 if (error)
                         goto error;

Adding a boolean is an easy way out without conflicting it.

Cheers,
Longman