From: "Kirill A. Shutemov" <kirill@shutemov.name>
To: Jan Kara <jack@suse.cz>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
"Theodore Ts'o" <tytso@mit.edu>,
Andreas Dilger <adilger.kernel@dilger.ca>,
Jan Kara <jack@suse.com>,
Andrew Morton <akpm@linux-foundation.org>,
Alexander Viro <viro@zeniv.linux.org.uk>,
Hugh Dickins <hughd@google.com>,
Andrea Arcangeli <aarcange@redhat.com>,
Dave Hansen <dave.hansen@intel.com>,
Vlastimil Babka <vbabka@suse.cz>,
Matthew Wilcox <willy@infradead.org>,
Ross Zwisler <ross.zwisler@linux.intel.com>,
linux-ext4@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
linux-block@vger.kernel.org
Subject: Re: [PATCHv3 15/41] filemap: handle huge pages in do_generic_file_read()
Date: Wed, 2 Nov 2016 11:32:04 +0300 [thread overview]
Message-ID: <20161102083204.GB13949@node.shutemov.name> (raw)
In-Reply-To: <20161101163940.GA5459@quack2.suse.cz>
On Tue, Nov 01, 2016 at 05:39:40PM +0100, Jan Kara wrote:
> On Mon 31-10-16 21:10:35, Kirill A. Shutemov wrote:
> > > If I understand the motivation right, it is mostly about being able to mmap
> > > PMD-sized chunks to userspace. So my naive idea would be that we could just
> > > implement it by allocating PMD sized chunks of pages when adding pages to
> > > page cache, we don't even have to read them all unless we come from PMD
> > > fault path.
> >
> > Well, no. We have one PG_{uptodate,dirty,writeback,mappedtodisk,etc}
> > per-hugepage, one common list of buffer heads...
> >
> > PG_dirty and PG_uptodate behaviour inhered from anon-THP (where handling
> > it otherwise doesn't make sense) and handling it differently for file-THP
> > is nightmare from maintenance POV.
>
> But the complexity of two different page sizes for page cache and *each*
> filesystem that wants to support it does not make the maintenance easy
> either.
I think with time we can make small pages just a subcase of huge pages.
And some generalization can be made once more than one filesystem with
backing storage will adopt huge pages.
> So I'm not convinced that using the same rules for anon-THP and
> file-THP is a clear win.
We already have file-THP with the same rules: tmpfs. Backing storage is
what changes the picture.
> And if we have these two options neither of which has negligible
> maintenance cost, I'd also like to see more justification for why it is
> a good idea to have file-THP for normal filesystems. Do you have any
> performance numbers that show it is a win under some realistic workload?
See below. As usual with huge pages, they make sense when you plenty of
memory.
> I'd also note that having PMD-sized pages has some obvious disadvantages as
> well:
>
> 1) I'm not sure buffer head handling code will quite scale to 512 or even
> 2048 buffer_heads on a linked list referenced from a page. It may work but
> I suspect the performance will suck.
Yes, buffer_head list doesn't scale. That's the main reason (along with 4)
why syscall-based IO sucks. We spend a lot of time looking for desired
block.
We need to switch to some other data structure for storing buffer_heads.
Is there a reason why we have list there in first place?
Why not just array?
I will look into it, but this sounds like a separate infrastructure change
project.
> 2) PMD-sized pages result in increased space & memory usage.
Space? Do you mean disk space? Not really: we still don't write beyond
i_size or into holes.
Behaviour wrt to holes may change with mmap()-IO as we have less
granularity, but the same can be seen just between different
architectures: 4k vs. 64k base page size.
> 3) In ext4 we have to estimate how much metadata we may need to modify when
> allocating blocks underlying a page in the worst case (you don't seem to
> update this estimate in your patch set). With 2048 blocks underlying a page,
> each possibly in a different block group, it is a lot of metadata forcing
> us to reserve a large transaction (not sure if you'll be able to even
> reserve such large transaction with the default journal size), which again
> makes things slower.
I didn't saw this on profiles. And xfstests looks fine. I probably need to
run them with 1k blocks once again.
> 4) As you have noted some places like write_begin() still depend on 4k
> pages which creates a strange mix of places that use subpages and that use
> head pages.
Yes, this need to be addressed to restore syscall-IO performance and take
advantage of huge pages.
But again, it's an infrastructure change that would likely affect
interface between VFS and filesystems. It deserves a separate patchset.
> All this would be a non-issue (well, except 2 I guess) if we just didn't
> expose filesystems to the fact that something like file-THP exists.
The numbers below generated with fio. The working set is relatively small,
so it fits into page cache and writing set doesn't hit dirty_ratio.
I think the mmap performance should be enough to justify initial inclusion
of an experimental feature: it useful for workloads that targets mmap()-IO.
It will take time to get feature mature anyway.
Configuration:
- 2x E5-2697v2, 64G RAM;
- INTEL SSDSC2CW24;
- IO request size is 4k;
- 8 processes, 512MB data set each;
Workload
read/write baseline stddev huge=always stddev change
--------------------------------------------------------------------------------
sync-read
read 21439.00 348.14 20297.33 259.62 -5.33%
sync-write
write 6833.20 147.08 3630.13 52.86 -46.88%
sync-readwrite
read 4377.17 17.53 2366.33 19.52 -45.94%
write 4378.50 17.83 2365.80 19.94 -45.97%
sync-randread
read 5491.20 66.66 14664.00 288.29 167.05%
sync-randwrite
write 6396.13 98.79 2035.80 8.17 -68.17%
sync-randrw
read 2927.30 115.81 1036.08 34.67 -64.61%
write 2926.47 116.45 1036.11 34.90 -64.60%
libaio-read
read 254.36 12.49 258.63 11.29 1.68%
libaio-write
write 4979.20 122.75 2904.77 17.93 -41.66%
libaio-readwrite
read 2738.57 142.72 2045.80 4.12 -25.30%
write 2729.93 141.80 2039.77 3.79 -25.28%
libaio-randread
read 113.63 2.98 210.63 5.07 85.37%
libaio-randwrite
write 4456.10 76.21 1649.63 7.00 -62.98%
libaio-randrw
read 97.85 8.03 877.49 28.27 796.80%
write 97.55 7.99 874.83 28.19 796.77%
mmap-read
read 20654.67 304.48 24696.33 1064.07 19.57%
mmap-write
write 8652.33 272.44 13187.33 499.10 52.41%
mmap-readwrite
read 6620.57 16.05 9221.60 399.56 39.29%
write 6623.63 16.34 9222.13 399.31 39.23%
mmap-randread
read 6717.23 1360.55 21939.33 326.38 226.61%
mmap-randwrite
write 3204.63 253.66 12371.00 61.49 286.03%
mmap-randrw
read 2150.50 78.00 7682.67 188.59 257.25%
write 2149.50 78.00 7685.40 188.35 257.54%
--
Kirill A. Shutemov
next prev parent reply other threads:[~2016-11-02 8:32 UTC|newest]
Thread overview: 75+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-15 11:54 [PATCHv3 00/41] ext4: support of huge pages Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 01/41] tools: Add WARN_ON_ONCE Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 02/41] radix tree test suite: Allow GFP_ATOMIC allocations to fail Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 03/41] radix-tree: Add radix_tree_join Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 04/41] radix-tree: Add radix_tree_split Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 05/41] radix-tree: Add radix_tree_split_preload() Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 06/41] radix-tree: Handle multiorder entries being deleted by replace_clear_tags Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 07/41] mm, shmem: swich huge tmpfs to multi-order radix-tree entries Kirill A. Shutemov
2016-09-16 12:07 ` Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 08/41] Revert "radix-tree: implement radix_tree_maybe_preload_order()" Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 09/41] page-flags: relax page flag policy for few flags Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 10/41] mm, rmap: account file thp pages Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 11/41] thp: try to free page's buffers before attempt split Kirill A. Shutemov
2016-10-11 15:40 ` Jan Kara
2016-10-11 21:43 ` Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 12/41] thp: handle write-protection faults for file THP Kirill A. Shutemov
2016-10-11 15:47 ` Jan Kara
2016-10-11 21:47 ` Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 13/41] truncate: make sure invalidate_mapping_pages() can discard huge pages Kirill A. Shutemov
2016-10-11 15:58 ` Jan Kara
2016-10-11 21:53 ` Kirill A. Shutemov
2016-10-12 6:43 ` Jan Kara
2016-10-24 10:41 ` Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 14/41] filemap: allocate huge page in page_cache_read(), if allowed Kirill A. Shutemov
2016-10-11 16:15 ` Jan Kara
2016-10-11 21:57 ` Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 15/41] filemap: handle huge pages in do_generic_file_read() Kirill A. Shutemov
2016-10-13 9:33 ` Jan Kara
2016-10-31 18:10 ` Kirill A. Shutemov
2016-11-01 16:39 ` Jan Kara
2016-11-02 8:32 ` Kirill A. Shutemov [this message]
2016-11-02 14:37 ` Christoph Hellwig
2016-11-03 20:40 ` Jan Kara
2016-11-07 11:07 ` Kirill A. Shutemov
2016-11-07 14:59 ` Christoph Hellwig
2016-11-02 14:36 ` Christoph Hellwig
2016-11-03 17:56 ` Jan Kara
2016-11-07 11:13 ` Kirill A. Shutemov
2016-11-07 15:01 ` Christoph Hellwig
2016-11-07 16:03 ` Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 16/41] filemap: allocate huge page in pagecache_get_page(), if allowed Kirill A. Shutemov
2016-09-15 11:54 ` [PATCHv3 17/41] filemap: handle huge pages in filemap_fdatawait_range() Kirill A. Shutemov
2016-10-13 9:44 ` Jan Kara
2016-10-13 12:08 ` Kirill A. Shutemov
2016-10-13 13:18 ` Jan Kara
2016-10-24 11:36 ` Kirill A. Shutemov
2016-10-30 17:31 ` Jan Kara
2016-09-15 11:55 ` [PATCHv3 18/41] HACK: readahead: alloc huge pages, if allowed Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 19/41] block: define BIO_MAX_PAGES to HPAGE_PMD_NR if huge page cache enabled Kirill A. Shutemov
2016-09-16 12:09 ` Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 20/41] mm: make write_cache_pages() work on huge pages Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 21/41] thp: introduce hpage_size() and hpage_mask() Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 22/41] thp: do not threat slab pages as huge in hpage_{nr_pages,size,mask} Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 23/41] fs: make block_read_full_page() be able to read huge page Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 24/41] fs: make block_write_{begin,end}() be able to handle huge pages Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 25/41] fs: make block_page_mkwrite() aware about " Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 26/41] truncate: make truncate_inode_pages_range() " Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 27/41] truncate: make invalidate_inode_pages2_range() " Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 28/41] mm, hugetlb: switch hugetlbfs to multi-order radix-tree entries Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 29/41] ext4: make ext4_mpage_readpages() hugepage-aware Kirill A. Shutemov
2016-09-15 12:27 ` Andreas Dilger
2016-09-16 12:17 ` Kirill A. Shutemov
2016-09-16 12:10 ` Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 30/41] ext4: make ext4_writepage() work on huge pages Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 31/41] ext4: handle huge pages in ext4_page_mkwrite() Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 32/41] ext4: handle huge pages in __ext4_block_zero_page_range() Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 33/41] ext4: make ext4_block_write_begin() aware about huge pages Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 34/41] ext4: handle huge pages in ext4_da_write_end() Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 35/41] ext4: make ext4_da_page_release_reservation() aware about huge pages Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 36/41] ext4: handle writeback with " Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 37/41] ext4: make EXT4_IOC_MOVE_EXT work " Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 38/41] ext4: fix SEEK_DATA/SEEK_HOLE for " Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 39/41] ext4: make fallocate() operations work with " Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 40/41] mm, fs, ext4: expand use of page_mapping() and page_to_pgoff() Kirill A. Shutemov
2016-09-15 11:55 ` [PATCHv3 41/41] ext4, vfs: add huge= mount option Kirill A. Shutemov
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=20161102083204.GB13949@node.shutemov.name \
--to=kirill@shutemov.name \
--cc=aarcange@redhat.com \
--cc=adilger.kernel@dilger.ca \
--cc=akpm@linux-foundation.org \
--cc=dave.hansen@intel.com \
--cc=hughd@google.com \
--cc=jack@suse.com \
--cc=jack@suse.cz \
--cc=kirill.shutemov@linux.intel.com \
--cc=linux-block@vger.kernel.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=ross.zwisler@linux.intel.com \
--cc=tytso@mit.edu \
--cc=vbabka@suse.cz \
--cc=viro@zeniv.linux.org.uk \
--cc=willy@infradead.org \
/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).