linux-kernel - Re: [RFC PATCH] mm: fix the inaccurate memory statistics issue for users

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <eb7e85f3-90d5-428c-a93a-7e54ade1479c@linux.alibaba.com>
Date: Fri, 23 May 2025 13:47:28 +0800
From: Baolin Wang <baolin.wang@...ux.alibaba.com>
To: Donet Tom <donettom@...ux.ibm.com>, akpm@...ux-foundation.org,
 david@...hat.com, shakeelb@...gle.com
Cc: lorenzo.stoakes@...cle.com, Liam.Howlett@...cle.com, vbabka@...e.cz,
 rppt@...nel.org, surenb@...gle.com, mhocko@...e.com, linux-mm@...ck.org,
 linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH] mm: fix the inaccurate memory statistics issue for
 users



On 2025/5/23 13:25, Donet Tom wrote:
> 
> On 5/23/25 8:46 AM, Baolin Wang wrote:
>> On some large machines with a high number of CPUs running a 64K kernel,
>> we found that the 'RES' field is always 0 displayed by the top command
>> for some processes, which will cause a lot of confusion for users.
>>
>>      PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     
>> TIME+ COMMAND
>>   875525 root      20   0   12480      0      0 R   0.3   0.0   
>> 0:00.08 top
>>        1 root      20   0  172800      0      0 S   0.0   0.0   
>> 0:04.52 systemd
>>
>> The main reason is that the batch size of the percpu counter is quite 
>> large
>> on these machines, caching a significant percpu value, since 
>> converting mm's
>> rss stats into percpu_counter by commit f1a7941243c1 ("mm: convert 
>> mm's rss
>> stats into percpu_counter"). Intuitively, the batch number should be 
>> optimized,
>> but on some paths, performance may take precedence over statistical 
>> accuracy.
>> Therefore, introducing a new interface to add the percpu statistical 
>> count
>> and display it to users, which can remove the confusion. In addition, 
>> this
>> change is not expected to be on a performance-critical path, so the 
>> modification
>> should be acceptable.
>>
>> Signed-off-by: Baolin Wang <baolin.wang@...ux.alibaba.com>
>> ---
>>   fs/proc/task_mmu.c | 14 +++++++-------
>>   include/linux/mm.h |  5 +++++
>>   2 files changed, 12 insertions(+), 7 deletions(-)
>>
>> diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
>> index b9e4fbbdf6e6..f629e6526935 100644
>> --- a/fs/proc/task_mmu.c
>> +++ b/fs/proc/task_mmu.c
>> @@ -36,9 +36,9 @@ void task_mem(struct seq_file *m, struct mm_struct *mm)
>>       unsigned long text, lib, swap, anon, file, shmem;
>>       unsigned long hiwater_vm, total_vm, hiwater_rss, total_rss;
>> -    anon = get_mm_counter(mm, MM_ANONPAGES);
>> -    file = get_mm_counter(mm, MM_FILEPAGES);
>> -    shmem = get_mm_counter(mm, MM_SHMEMPAGES);
>> +    anon = get_mm_counter_sum(mm, MM_ANONPAGES);
> 
> 
> Hi Baolin Wang,
> 
> We also observed the same issue where the RSS value in /proc/PID/status
> was 0 on machines with a high number of CPUs. With this patch, the issue
> got fixedl

Yes, we also observed this issue.

> Rss value without this patch
> ----------------------------
>   # cat /proc/87406/status
> .....
> VmRSS:           0 kB
> RssAnon:           0 kB
> RssFile:           0 k
> 
> 
> Rss values with this patch
> --------------------------
>   # cat /proc/3055/status
> VmRSS:        2176 kB
> RssAnon:         512 kB
> RssFile:        1664 kB
> RssShmem:           0 kB
> 
> Tested-by Donet Tom <donettom@...ux.ibm.com>

Thanks for testing.