mm: soft-offline: close the race against page allocation
authorNaoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Fri, 24 Aug 2018 00:00:42 +0000 (17:00 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 24 Aug 2018 01:48:43 +0000 (18:48 -0700)
commitd4ae9916ea2947341180d2b538f48875ff393a86
treee478e5e6e84de25fade8f2995c9ee49cacd5a488
parent6bc9b56433b76e40d11099338d27fbc5cd2935ca
mm: soft-offline: close the race against page allocation

A process can be killed with SIGBUS(BUS_MCEERR_AR) when it tries to
allocate a page that was just freed on the way of soft-offline.  This is
undesirable because soft-offline (which is about corrected error) is
less aggressive than hard-offline (which is about uncorrected error),
and we can make soft-offline fail and keep using the page for good
reason like "system is busy."

Two main changes of this patch are:

- setting migrate type of the target page to MIGRATE_ISOLATE. As done
  in free_unref_page_commit(), this makes kernel bypass pcplist when
  freeing the page. So we can assume that the page is in freelist just
  after put_page() returns,

- setting PG_hwpoison on free page under zone->lock which protects
  freelists, so this allows us to avoid setting PG_hwpoison on a page
  that is decided to be allocated soon.

[akpm@linux-foundation.org: tweak set_hwpoison_free_buddy_page() comment]
Link: http://lkml.kernel.org/r/1531452366-11661-3-git-send-email-n-horiguchi@ah.jp.nec.com
Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Reported-by: Xishi Qiu <xishi.qiuxishi@alibaba-inc.com>
Tested-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: <zy.zhengyi@alibaba-inc.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
include/linux/page-flags.h
include/linux/swapops.h
mm/memory-failure.c
mm/migrate.c
mm/page_alloc.c