mm/memory_hotplug: fix online/offline_pages called w.o. mem_hotplug_lock
authorDavid Hildenbrand <david@redhat.com>
Tue, 30 Oct 2018 22:10:29 +0000 (15:10 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 31 Oct 2018 15:54:17 +0000 (08:54 -0700)
commit381eab4a6ee81266f8dddc62e57376c7e584e5b8
treedf4cf8ea76a7d8d85f13576bb95a75ec6b974612
parent8df1d0e4a265f25dc1e7e7624ccdbcb4a6630c89
mm/memory_hotplug: fix online/offline_pages called w.o. mem_hotplug_lock

There seem to be some problems as result of 30467e0b3be ("mm, hotplug:
fix concurrent memory hot-add deadlock"), which tried to fix a possible
lock inversion reported and discussed in [1] due to the two locks
a) device_lock()
b) mem_hotplug_lock

While add_memory() first takes b), followed by a) during
bus_probe_device(), onlining of memory from user space first took a),
followed by b), exposing a possible deadlock.

In [1], and it was decided to not make use of device_hotplug_lock, but
rather to enforce a locking order.

The problems I spotted related to this:

1. Memory block device attributes: While .state first calls
   mem_hotplug_begin() and the calls device_online() - which takes
   device_lock() - .online does no longer call mem_hotplug_begin(), so
   effectively calls online_pages() without mem_hotplug_lock.

2. device_online() should be called under device_hotplug_lock, however
   onlining memory during add_memory() does not take care of that.

In addition, I think there is also something wrong about the locking in

3. arch/powerpc/platforms/powernv/memtrace.c calls offline_pages()
   without locks. This was introduced after 30467e0b3be. And skimming over
   the code, I assume it could need some more care in regards to locking
   (e.g. device_online() called without device_hotplug_lock. This will
   be addressed in the following patches.

Now that we hold the device_hotplug_lock when
- adding memory (e.g. via add_memory()/add_memory_resource())
- removing memory (e.g. via remove_memory())
- device_online()/device_offline()

We can move mem_hotplug_lock usage back into
online_pages()/offline_pages().

Why is mem_hotplug_lock still needed? Essentially to make
get_online_mems()/put_online_mems() be very fast (relying on
device_hotplug_lock would be very slow), and to serialize against
addition of memory that does not create memory block devices (hmm).

[1] http://driverdev.linuxdriverproject.org/pipermail/ driverdev-devel/
    2015-February/065324.html

This patch is partly based on a patch by Vitaly Kuznetsov.

Link: http://lkml.kernel.org/r/20180925091457.28651-4-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Pavel Tatashin <pavel.tatashin@microsoft.com>
Reviewed-by: Rashmica Gupta <rashmica.g@gmail.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Len Brown <lenb@kernel.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "K. Y. Srinivasan" <kys@microsoft.com>
Cc: Haiyang Zhang <haiyangz@microsoft.com>
Cc: Stephen Hemminger <sthemmin@microsoft.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: Rashmica Gupta <rashmica.g@gmail.com>
Cc: Michael Neuling <mikey@neuling.org>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: Kate Stewart <kstewart@linuxfoundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Philippe Ombredanne <pombredanne@nexb.com>
Cc: Pavel Tatashin <pavel.tatashin@microsoft.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: YASUAKI ISHIMATSU <yasu.isimatu@gmail.com>
Cc: Mathieu Malaterre <malat@debian.org>
Cc: John Allen <jallen@linux.vnet.ibm.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
drivers/base/memory.c
mm/memory_hotplug.c