linux-kernel - Re: [PATCH v4 3/4] mm/page_reporting: Allow driver to specify reporting order

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: <fef3e1a0-4a3a-70a2-233d-cd4bda3c8fab@redhat.com>
Date:   Fri, 25 Jun 2021 14:24:06 +1000
From:   Gavin Shan <gshan@...hat.com>
To:     Alexander Duyck <alexander.duyck@...il.com>
Cc:     linux-mm <linux-mm@...ck.org>, LKML <linux-kernel@...r.kernel.org>,
        David Hildenbrand <david@...hat.com>,
        "Michael S. Tsirkin" <mst@...hat.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Anshuman Khandual <anshuman.khandual@....com>,
        Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will@...nel.org>, shan.gavin@...il.com
Subject: Re: [PATCH v4 3/4] mm/page_reporting: Allow driver to specify
 reporting order

On 6/25/21 2:00 PM, Gavin Shan wrote:
> On 6/25/21 11:19 AM, Alexander Duyck wrote:
>> On Thu, Jun 24, 2021 at 4:46 PM Gavin Shan <gshan@...hat.com> wrote:
>>>
>>> The page reporting order (threshold) is sticky to @pageblock_order
>>> by default. The page reporting can never be triggered because the
>>> freeing page can't come up with a free area like that huge. The
>>> situation becomes worse when the system memory becomes heavily
>>> fragmented.
>>>
>>> For example, the following configurations are used on ARM64 when 64KB
>>> base page size is enabled. In this specific case, the page reporting
>>> won't be triggered until the freeing page comes up with a 512MB free
>>> area. That's hard to be met, especially when the system memory becomes
>>> heavily fragmented.
>>>
>>>     PAGE_SIZE:          64KB
>>>     HPAGE_SIZE:         512MB
>>>     pageblock_order:    13       (512MB)
>>>     MAX_ORDER:          14
>>>
>>> This allows the drivers to specify the page reporting order when the
>>> page reporting device is registered. It falls back to @pageblock_order
>>> if it's not specified by the driver. The existing users (hv_balloon
>>> and virtio_balloon) don't specify it and @pageblock_order is still
>>> taken as their page reporting order. So this shouldn't introduce any
>>> functional changes.
>>>
>>> Signed-off-by: Gavin Shan <gshan@...hat.com>
>>> Reviewed-by: Alexander Duyck <alexanderduyck@...com>
>>> ---
>>>   include/linux/page_reporting.h | 3 +++
>>>   mm/page_reporting.c            | 6 ++++++
>>>   2 files changed, 9 insertions(+)
>>>
>>> diff --git a/include/linux/page_reporting.h b/include/linux/page_reporting.h
>>> index 3b99e0ec24f2..fe648dfa3a7c 100644
>>> --- a/include/linux/page_reporting.h
>>> +++ b/include/linux/page_reporting.h
>>> @@ -18,6 +18,9 @@ struct page_reporting_dev_info {
>>>
>>>          /* Current state of page reporting */
>>>          atomic_t state;
>>> +
>>> +       /* Minimal order of page reporting */
>>> +       unsigned int order;
>>>   };
>>>
>>>   /* Tear-down and bring-up for page reporting devices */
>>> diff --git a/mm/page_reporting.c b/mm/page_reporting.c
>>> index 34bf4d26c2c4..382958eef8a9 100644
>>> --- a/mm/page_reporting.c
>>> +++ b/mm/page_reporting.c
>>> @@ -329,6 +329,12 @@ int page_reporting_register(struct page_reporting_dev_info *prdev)
>>>                  goto err_out;
>>>          }
>>>
>>> +       /*
>>> +        * Update the page reporting order if it's specified by driver.
>>> +        * Otherwise, it falls back to @pageblock_order.
>>> +        */
>>> +       page_reporting_order = prdev->order ? : pageblock_order;
>>> +
>>
>> An alternative to this would be to look at setting up some
>> comparisons. I might add another variable and do something like:
>> order = prdev->order ? : pageblock_order;
>> if (order < page_reporting_order)
>>      page_reporting_order = order;
>>
>> You could essentially do something similar in the previous patch but
>> just use pageblock_order directly rather than having to add a local
>> variable.
>>
>> That way if you need to still pull down the page reporting order you
>> can do so without prdev->order or pageblock_order overwriting the
>> value and pushing it back up.
>>
> 
> Thanks, Alex. Lets do both in v5, which will be posted shortly.
> 

Alex, I just posted v5 to have the checks you suggested. Could
you help to have a quick scan. It's pointless to let Andrew
drop the patches and apply the last one again :)

Thanks,
Gavin