From: Pavel Tatashin <pasha.tatashin@oracle.com>
To: linux-kernel@vger.kernel.org, sparclinux@vger.kernel.org,
linux-mm@kvack.org, linuxppc-dev@lists.ozlabs.org,
linux-s390@vger.kernel.org, borntraeger@de.ibm.com,
heiko.carstens@de.ibm.com, davem@davemloft.net
Subject: [v3 0/9] parallelized "struct page" zeroing
Date: Fri, 5 May 2017 13:03:07 -0400 [thread overview]
Message-ID: <1494003796-748672-1-git-send-email-pasha.tatashin@oracle.com> (raw)
Changelog:
v2 - v3
- Addressed David's comments about one change per patch:
* Splited changes to platforms into 4 patches
* Made "do not zero vmemmap_buf" as a separate patch
v1 - v2
- Per request, added s390 to deferred "struct page" zeroing
- Collected performance data on x86 which proofs the importance to
keep memset() as prefetch (see below).
When deferred struct page initialization feature is enabled, we get a
performance gain of initializing vmemmap in parallel after other CPUs are
started. However, we still zero the memory for vmemmap using one boot CPU.
This patch-set fixes the memset-zeroing limitation by deferring it as well.
Performance gain on SPARC with 32T:
base: https://hastebin.com/ozanelatat.go
fix: https://hastebin.com/utonawukof.go
As you can see without the fix it takes: 97.89s to boot
With the fix it takes: 46.91 to boot.
Performance gain on x86 with 1T:
base: https://hastebin.com/uvifasohon.pas
fix: https://hastebin.com/anodiqaguj.pas
On Intel we save 10.66s/T while on SPARC we save 1.59s/T. Intel has
twice as many pages, and also fewer nodes than SPARC (sparc 32 nodes, vs.
intel 8 nodes).
It takes one thread 11.25s to zero vmemmap on Intel for 1T, so it should
take additional 11.25 / 8 = 1.4s (this machine has 8 nodes) per node to
initialize the memory, but it takes only additional 0.456s per node, which
means on Intel we also benefit from having memset() and initializing all
other fields in one place.
Pavel Tatashin (9):
sparc64: simplify vmemmap_populate
mm: defining memblock_virt_alloc_try_nid_raw
mm: add "zero" argument to vmemmap allocators
mm: do not zero vmemmap_buf
mm: zero struct pages during initialization
sparc64: teach sparc not to zero struct pages memory
x86: teach x86 not to zero struct pages memory
powerpc: teach platforms not to zero struct pages memory
s390: teach platforms not to zero struct pages memory
arch/powerpc/mm/init_64.c | 4 +-
arch/s390/mm/vmem.c | 5 ++-
arch/sparc/mm/init_64.c | 26 +++++++----------------
arch/x86/mm/init_64.c | 3 +-
include/linux/bootmem.h | 3 ++
include/linux/mm.h | 15 +++++++++++--
mm/memblock.c | 46 ++++++++++++++++++++++++++++++++++++------
mm/page_alloc.c | 3 ++
mm/sparse-vmemmap.c | 48 +++++++++++++++++++++++++++++---------------
9 files changed, 103 insertions(+), 50 deletions(-)
next reply other threads:[~2017-05-05 17:04 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-05-05 17:03 Pavel Tatashin [this message]
2017-05-05 17:03 ` [v3 1/9] sparc64: simplify vmemmap_populate Pavel Tatashin
2017-05-05 17:03 ` [v3 2/9] mm: defining memblock_virt_alloc_try_nid_raw Pavel Tatashin
2017-05-05 17:03 ` [v3 3/9] mm: add "zero" argument to vmemmap allocators Pavel Tatashin
2017-05-13 19:17 ` kbuild test robot
2017-05-05 17:03 ` [v3 4/9] mm: do not zero vmemmap_buf Pavel Tatashin
2017-05-05 17:03 ` [v3 5/9] mm: zero struct pages during initialization Pavel Tatashin
2017-05-05 17:03 ` [v3 6/9] sparc64: teach sparc not to zero struct pages memory Pavel Tatashin
2017-05-05 17:03 ` [v3 7/9] x86: teach x86 " Pavel Tatashin
2017-05-05 17:03 ` [v3 8/9] powerpc: teach platforms " Pavel Tatashin
2017-05-05 17:03 ` [v3 9/9] s390: " Pavel Tatashin
2017-05-08 11:36 ` Heiko Carstens
2017-05-15 18:24 ` Pasha Tatashin
2017-05-15 23:17 ` Heiko Carstens
2017-05-16 0:33 ` Pasha Tatashin
2017-05-09 18:12 ` [v3 0/9] parallelized "struct page" zeroing Michal Hocko
2017-05-09 18:54 ` Pasha Tatashin
2017-05-10 7:24 ` Michal Hocko
2017-05-10 13:42 ` Pasha Tatashin
2017-05-10 14:57 ` Michal Hocko
2017-05-10 15:01 ` Pasha Tatashin
2017-05-10 15:20 ` David Miller
2017-05-11 20:47 ` Pasha Tatashin
2017-05-11 20:59 ` Pasha Tatashin
2017-05-12 16:57 ` David Miller
2017-05-12 17:24 ` Pasha Tatashin
2017-05-12 17:37 ` David Miller
2017-05-16 23:50 ` Benjamin Herrenschmidt
2017-05-12 16:56 ` David Miller
2017-05-10 15:19 ` David Miller
2017-05-10 17:17 ` Matthew Wilcox
2017-05-10 18:00 ` David Miller
2017-05-10 21:11 ` Matthew Wilcox
2017-05-11 8:05 ` Michal Hocko
2017-05-11 14:35 ` David Miller
2017-05-15 18:12 ` Pasha Tatashin
2017-05-15 19:38 ` Michal Hocko
2017-05-15 20:44 ` Pasha Tatashin
2017-05-16 8:36 ` Michal Hocko
2017-05-26 16:45 ` Pasha Tatashin
2017-05-29 11:53 ` Michal Hocko
2017-05-30 17:16 ` Pasha Tatashin
2017-05-31 16:31 ` Michal Hocko
2017-05-31 16:51 ` David Miller
2017-06-01 3:35 ` Pasha Tatashin
2017-06-01 8:46 ` 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=1494003796-748672-1-git-send-email-pasha.tatashin@oracle.com \
--to=pasha.tatashin@oracle.com \
--cc=borntraeger@de.ibm.com \
--cc=davem@davemloft.net \
--cc=heiko.carstens@de.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-s390@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=sparclinux@vger.kernel.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).