linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Rik van Riel <riel@redhat.com>
To: Andrew Morton <akpm@linux-foundation.org>,
	Ebru Akagunduz <ebru.akagunduz@gmail.com>
Cc: linux-mm@kvack.org, kirill.shutemov@linux.intel.com,
	n-horiguchi@ah.jp.nec.com, aarcange@redhat.com,
	iamjoonsoo.kim@lge.com, xiexiuqi@huawei.com, gorcunov@openvz.org,
	linux-kernel@vger.kernel.org, mgorman@suse.de,
	rientjes@google.com, vbabka@suse.cz,
	aneesh.kumar@linux.vnet.ibm.com, hughd@google.com,
	hannes@cmpxchg.org, mhocko@suse.cz, boaz@plexistor.com,
	raindel@mellanox.com
Subject: Re: [RFC 3/3] mm: make swapin readahead to improve thp collapse rate
Date: Tue, 16 Jun 2015 23:20:20 -0400	[thread overview]
Message-ID: <5580E774.3070307@redhat.com> (raw)
In-Reply-To: <20150616141540.adc40130139151bf19f07ff9@linux-foundation.org>

On 06/16/2015 05:15 PM, Andrew Morton wrote:
> On Sun, 14 Jun 2015 18:04:43 +0300 Ebru Akagunduz <ebru.akagunduz@gmail.com> wrote:
> 
>> This patch makes swapin readahead to improve thp collapse rate.
>> When khugepaged scanned pages, there can be a few of the pages
>> in swap area.
>>
>> With the patch THP can collapse 4kB pages into a THP when
>> there are up to max_ptes_swap swap ptes in a 2MB range.
>>
>> The patch was tested with a test program that allocates
>> 800MB of memory, writes to it, and then sleeps. I force
>> the system to swap out all. Afterwards, the test program
>> touches the area by writing, it skips a page in each
>> 20 pages of the area.
>>
>> Without the patch, system did not swap in readahead.
>> THP rate was %47 of the program of the memory, it
>> did not change over time.
>>
>> With this patch, after 10 minutes of waiting khugepaged had
>> collapsed %99 of the program's memory.
>>
>> ...
>>
>> +/*
>> + * Bring missing pages in from swap, to complete THP collapse.
>> + * Only done if khugepaged_scan_pmd believes it is worthwhile.
>> + *
>> + * Called and returns without pte mapped or spinlocks held,
>> + * but with mmap_sem held to protect against vma changes.
>> + */
>> +
>> +static void __collapse_huge_page_swapin(struct mm_struct *mm,
>> +					struct vm_area_struct *vma,
>> +					unsigned long address, pmd_t *pmd,
>> +					pte_t *pte)
>> +{
>> +	unsigned long _address;
>> +	pte_t pteval = *pte;
>> +	int swap_pte = 0;
>> +
>> +	pte = pte_offset_map(pmd, address);
>> +	for (_address = address; _address < address + HPAGE_PMD_NR*PAGE_SIZE;
>> +	     pte++, _address += PAGE_SIZE) {
>> +		pteval = *pte;
>> +		if (is_swap_pte(pteval)) {
>> +			swap_pte++;
>> +			do_swap_page(mm, vma, _address, pte, pmd, 0x0, pteval);
>> +			/* pte is unmapped now, we need to map it */
>> +			pte = pte_offset_map(pmd, _address);
>> +		}
>> +	}
>> +	pte--;
>> +	pte_unmap(pte);
>> +	trace_mm_collapse_huge_page_swapin(mm, vma->vm_start, swap_pte);
>> +}
> 
> This is doing a series of synchronous reads.  That will be sloooow on
> spinning disks.
>
> This function should be significantly faster if it first gets all the
> necessary I/O underway.  I don't think we have a function which exactly
> does this.  Perhaps generalise swapin_readahead() or open-code
> something like

Looking at do_swap_page() and __lock_page_or_retry(), I guess
there already is a way to do the above.

Passing a "flags" of FAULT_FLAG_ALLOW_RETRY|FAULT_FLAG_RETRY_NOWAIT
to do_swap_page() should result in do_swap_page() returning with
the pte unmapped and the mmap_sem still held if the page was not
immediately available to map into the pte (trylock_page succeeds).

Ebru, can you try passing the above as the flags argument to
do_swap_page(), and see what happens?

-- 
All rights reversed

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2015-06-17  3:20 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-06-14 15:04 [RFC 0/3] mm: make swapin readahead to gain more thp performance Ebru Akagunduz
2015-06-14 15:04 ` [RFC 1/3] mm: add tracepoint for scanning pages Ebru Akagunduz
2015-06-15  1:04   ` Rik van Riel
2015-06-14 15:04 ` [RFC 2/3] mm: make optimistic check for swapin readahead Ebru Akagunduz
2015-06-15  5:40   ` Leon Romanovsky
2015-06-15  5:43     ` Rik van Riel
2015-06-15  6:08       ` Leon Romanovsky
2015-06-15  6:35         ` Rik van Riel
2015-06-15 14:05   ` Rik van Riel
2015-06-15 16:07     ` Leon Romanovsky
2015-06-14 15:04 ` [RFC 3/3] mm: make swapin readahead to improve thp collapse rate Ebru Akagunduz
2015-06-15 13:59   ` Rik van Riel
2015-06-16 21:15   ` Andrew Morton
2015-06-17  3:20     ` Rik van Riel [this message]
2015-06-17 17:38       ` Ebru Akagunduz

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=5580E774.3070307@redhat.com \
    --to=riel@redhat.com \
    --cc=aarcange@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=aneesh.kumar@linux.vnet.ibm.com \
    --cc=boaz@plexistor.com \
    --cc=ebru.akagunduz@gmail.com \
    --cc=gorcunov@openvz.org \
    --cc=hannes@cmpxchg.org \
    --cc=hughd@google.com \
    --cc=iamjoonsoo.kim@lge.com \
    --cc=kirill.shutemov@linux.intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mgorman@suse.de \
    --cc=mhocko@suse.cz \
    --cc=n-horiguchi@ah.jp.nec.com \
    --cc=raindel@mellanox.com \
    --cc=rientjes@google.com \
    --cc=vbabka@suse.cz \
    --cc=xiexiuqi@huawei.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).