From: Michal Hocko <mhocko@suse.com>
To: Oscar Salvador <osalvador@suse.de>
Cc: Andrew Morton <akpm@linux-foundation.org>,
Vlastimil Babka <vbabka@suse.cz>,
David Hildenbrand <david@redhat.com>,
Muchun Song <songmuchun@bytedance.com>,
Mike Kravetz <mike.kravetz@oracle.com>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v5 3/5] mm: Make alloc_contig_range handle free hugetlb pages
Date: Wed, 17 Mar 2021 15:22:58 +0100 [thread overview]
Message-ID: <YFIQwlt0ndU0R05z@dhcp22.suse.cz> (raw)
In-Reply-To: <20210317111251.17808-4-osalvador@suse.de>
On Wed 17-03-21 12:12:49, Oscar Salvador wrote:
> alloc_contig_range will fail if it ever sees a HugeTLB page within the
> range we are trying to allocate, even when that page is free and can be
> easily reallocated.
> This has proved to be problematic for some users of alloc_contic_range,
> e.g: CMA and virtio-mem, where those would fail the call even when those
> pages lay in ZONE_MOVABLE and are free.
>
> We can do better by trying to replace such page.
>
> Free hugepages are tricky to handle so as to no userspace application
> notices disruption, we need to replace the current free hugepage with
> a new one.
>
> In order to do that, a new function called alloc_and_dissolve_huge_page
> is introduced.
> This function will first try to get a new fresh hugepage, and if it
> succeeds, it will replace the old one in the free hugepage pool.
>
> All operations are being handled under hugetlb_lock, so no races are
Slightly confusing because allocation which is a part of the process is
certainly not done under the lock.
"The free page replacement is done under hugetlb_lock, so no external
user of hugetlb will notice the change. There is one tricky case when
page's refcount is 0 because it is in the process of being released.
A mising PageHugeFreed bit will tell us that freeing is in flight so we
retry after dropping the hugetlb_lock. The race window should be small
and the next retry should make a forward progress.
> possible. The only exception is when page's refcount is 0, but it still
> has not been flagged as PageHugeFreed.
> E.g, below scenario:
>
> CPU0 CPU1
> __free_huge_page() isolate_or_dissolve_huge_page
> PageHuge() == T
> alloc_and_dissolve_huge_page
> alloc_fresh_huge_page()
> spin_lock(hugetlb_lock)
> // PageHuge() && !PageHugeFreed &&
> // !PageCount()
> spin_unlock(hugetlb_lock)
> spin_lock(hugetlb_lock)
> 1) update_and_free_page
> PageHuge() == F
> __free_pages()
> 2) enqueue_huge_page
> SetPageHugeFreed()
> spin_unlock(&hugetlb_lock)
> spin_lock(hugetlb_lock)
> 1) PageHuge() == F (freed by case#1 from CPU0)
> 2) PageHuge() == T
> PageHugeFreed() == T
> - proceed with replacing the page
>
> In the case above we retry as the window race is quite small and we have high
> chances to succeed next time.
>
> With regard to the allocation, we restrict it to the node the page belongs
> to with __GFP_THISNODE, meaning we do not fallback on other node's zones.
>
> Note that gigantic hugetlb pages are fenced off since there is a cyclic
> dependency between them and alloc_contig_range.
>
> Signed-off-by: Oscar Salvador <osalvador@suse.de>
> Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
> Acked-by: Michal Hocko <mhocko@suse.com>
my ack still applies.
--
Michal Hocko
SUSE Labs
next prev parent reply other threads:[~2021-03-17 14:30 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-17 11:12 [PATCH v5 0/5] Make alloc_contig_range handle Hugetlb pages Oscar Salvador
2021-03-17 11:12 ` [PATCH v5 1/5] mm,page_alloc: Bail out earlier on -ENOMEM in alloc_contig_migrate_range Oscar Salvador
2021-03-17 14:05 ` Michal Hocko
2021-03-17 14:42 ` David Hildenbrand
2021-03-17 14:49 ` Michal Hocko
2021-03-18 11:04 ` Oscar Salvador
2021-03-18 11:37 ` Michal Hocko
2021-03-17 11:12 ` [PATCH v5 2/5] mm,compaction: Let isolate_migratepages_{range,block} return error codes Oscar Salvador
2021-03-17 14:12 ` Michal Hocko
2021-03-17 14:38 ` Oscar Salvador
2021-03-17 14:59 ` Michal Hocko
2021-03-18 9:50 ` Vlastimil Babka
2021-03-18 10:22 ` Michal Hocko
2021-03-18 11:10 ` Vlastimil Babka
2021-03-18 11:36 ` Michal Hocko
2021-03-19 9:57 ` Oscar Salvador
2021-03-19 10:14 ` Vlastimil Babka
2021-03-19 10:26 ` Oscar Salvador
2021-03-17 11:12 ` [PATCH v5 3/5] mm: Make alloc_contig_range handle free hugetlb pages Oscar Salvador
2021-03-17 14:22 ` Michal Hocko [this message]
2021-03-17 11:12 ` [PATCH v5 4/5] mm: Make alloc_contig_range handle in-use " Oscar Salvador
2021-03-17 14:26 ` Michal Hocko
2021-03-18 8:54 ` Oscar Salvador
2021-03-18 9:29 ` Michal Hocko
2021-03-18 9:59 ` Oscar Salvador
2021-03-18 10:12 ` Michal Hocko
2021-03-17 11:12 ` [PATCH v5 5/5] mm,page_alloc: Drop unnecessary checks from pfn_range_valid_contig Oscar Salvador
2021-03-17 11:15 ` David Hildenbrand
2021-03-17 14:31 ` Michal Hocko
2021-03-17 14:36 ` David Hildenbrand
2021-03-17 15:03 ` Michal Hocko
2021-03-18 8:44 ` Oscar Salvador
2021-03-18 8:55 ` Michal Hocko
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=YFIQwlt0ndU0R05z@dhcp22.suse.cz \
--to=mhocko@suse.com \
--cc=akpm@linux-foundation.org \
--cc=david@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mike.kravetz@oracle.com \
--cc=osalvador@suse.de \
--cc=songmuchun@bytedance.com \
--cc=vbabka@suse.cz \
/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).