From: David Hildenbrand <david@redhat.com>
To: Balbir Singh <bsingharora@gmail.com>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
linux-doc@vger.kernel.org, linuxppc-dev@lists.ozlabs.org,
linux-acpi@vger.kernel.org, xen-devel@lists.xenproject.org,
devel@linuxdriverproject.org,
Andrew Morton <akpm@linux-foundation.org>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Boris Ostrovsky <boris.ostrovsky@oracle.com>,
Dan Williams <dan.j.williams@intel.com>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
Haiyang Zhang <haiyangz@microsoft.com>,
Heiko Carstens <heiko.carstens@de.ibm.com>,
John Allen <jallen@linux.vnet.ibm.com>,
Jonathan Corbet <corbet@lwn.net>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Juergen Gross <jgross@suse.com>,
Kate Stewart <kstewart@linuxfoundation.org>,
"K. Y. Srinivasan" <kys@microsoft.com>,
Len Brown <lenb@kernel.org>,
Martin Schwidefsky <schwidefsky@de.ibm.com>,
Mathieu Malaterre <malat@debian.org>,
Michael Ellerman <mpe@ellerman.id.au>,
Michael Neuling <mikey@neuling.org>,
Michal Hocko <mhocko@suse.com>,
Nathan Fontenot <nfont@linux.vnet.ibm.com>,
Oscar Salvador <osalvador@suse.de>,
Paul Mackerras <paulus@samba.org>,
Pavel Tatashin <pasha.tatashin@oracle.com>,
Pavel Tatashin <pavel.tatashin@microsoft.com>,
Philippe Ombredanne <pombredanne@nexb.com>,
"Rafael J. Wysocki" <rjw@rjwysocki.net>,
Rashmica Gupta <rashmica.g@gmail.com>,
Stephen Hemminger <sthemmin@microsoft.com>,
Thomas Gleixner <tglx@linutronix.de>,
Vlastimil Babka <vbabka@suse.cz>,
YASUAKI ISHIMATSU <yasu.isimatu@gmail.com>
Subject: Re: [PATCH v1 0/6] mm: online/offline_pages called w.o. mem_hotplug_lock
Date: Wed, 19 Sep 2018 09:35:07 +0200 [thread overview]
Message-ID: <f3a13f6a-b34c-8561-884a-23fd9aa60331@redhat.com> (raw)
In-Reply-To: <20180919012207.GD8537@350D>
Am 19.09.18 um 03:22 schrieb Balbir Singh:
> On Tue, Sep 18, 2018 at 01:48:16PM +0200, David Hildenbrand wrote:
>> Reading through the code and studying how mem_hotplug_lock is to be used,
>> I noticed that there are two places where we can end up calling
>> device_online()/device_offline() - online_pages()/offline_pages() without
>> the mem_hotplug_lock. And there are other places where we call
>> device_online()/device_offline() without the device_hotplug_lock.
>>
>> While e.g.
>> echo "online" > /sys/devices/system/memory/memory9/state
>> is fine, e.g.
>> echo 1 > /sys/devices/system/memory/memory9/online
>> Will not take the mem_hotplug_lock. However the device_lock() and
>> device_hotplug_lock.
>>
>> E.g. via memory_probe_store(), we can end up calling
>> add_memory()->online_pages() without the device_hotplug_lock. So we can
>> have concurrent callers in online_pages(). We e.g. touch in online_pages()
>> basically unprotected zone->present_pages then.
>>
>> Looks like there is a longer history to that (see Patch #2 for details),
>> and fixing it to work the way it was intended is not really possible. We
>> would e.g. have to take the mem_hotplug_lock in device/base/core.c, which
>> sounds wrong.
>>
>> Summary: We had a lock inversion on mem_hotplug_lock and device_lock().
>> More details can be found in patch 3 and patch 6.
>>
>> I propose the general rules (documentation added in patch 6):
>>
>> 1. add_memory/add_memory_resource() must only be called with
>> device_hotplug_lock.
>> 2. remove_memory() must only be called with device_hotplug_lock. This is
>> already documented and holds for all callers.
>> 3. device_online()/device_offline() must only be called with
>> device_hotplug_lock. This is already documented and true for now in core
>> code. Other callers (related to memory hotplug) have to be fixed up.
>> 4. mem_hotplug_lock is taken inside of add_memory/remove_memory/
>> online_pages/offline_pages.
>>
>> To me, this looks way cleaner than what we have right now (and easier to
>> verify). And looking at the documentation of remove_memory, using
>> lock_device_hotplug also for add_memory() feels natural.
>>
>
> That seems reasonable, but also implies that device_online() would hold
> back add/remove memory, could you please also document what mode
> read/write the locks need to be held? For example can the device_hotplug_lock
> be held in read mode while add/remove memory via (mem_hotplug_lock) is held
> in write mode?
device_hotplug_lock is an ordinary mutex. So no option there.
Only mem_hotplug_lock is a per CPU RW mutex. And as of now it only
exists to not require get_online_mems()/put_online_mems() to take the
device_hotplug_lock. Which is perfectly valid, because these users only
care about memory (not any other devices) not suddenly vanish. And that
RW lock makes things fast.
Any modifications (online/offline/add/remove) require the
mem_hotplug_lock in write.
I can add some more details to documentation in patch #6.
"... we should always hold the mem_hotplug_lock (via
mem_hotplug_begin/mem_hotplug_done) in write mode to serialize memory
hotplug" ..."
"In addition, mem_hotplug_lock (in contrast to device_hotplug_lock) in
read mode allows for a quite efficient get_online_mems/put_online_mems
implementation, so code accessing memory can protect from that memory
vanishing."
Would that work for you?
Thanks!
>
> Balbir Singh.
>
>
--
Thanks,
David / dhildenb
next prev parent reply other threads:[~2018-09-19 7:35 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-18 11:48 [PATCH v1 0/6] mm: online/offline_pages called w.o. mem_hotplug_lock David Hildenbrand
2018-09-18 11:48 ` [PATCH v1 1/6] mm/memory_hotplug: make remove_memory() take the device_hotplug_lock David Hildenbrand
2018-09-18 21:18 ` Rafael J. Wysocki
2018-09-18 11:48 ` [PATCH v1 2/6] mm/memory_hotplug: make add_memory() " David Hildenbrand
2018-09-18 21:19 ` Rafael J. Wysocki
2018-09-18 11:48 ` [PATCH v1 3/6] mm/memory_hotplug: fix online/offline_pages called w.o. mem_hotplug_lock David Hildenbrand
2018-09-18 11:48 ` [PATCH v1 4/6] powerpc/powernv: hold device_hotplug_lock when calling device_online() David Hildenbrand
2018-09-18 11:48 ` [PATCH v1 5/6] powerpc/powernv: hold device_hotplug_lock in memtrace_offline_pages() David Hildenbrand
2018-09-18 11:48 ` [PATCH v1 6/6] memory-hotplug.txt: Add some details about locking internals David Hildenbrand
2018-09-19 1:22 ` [PATCH v1 0/6] mm: online/offline_pages called w.o. mem_hotplug_lock Balbir Singh
2018-09-19 7:35 ` David Hildenbrand [this message]
2018-09-23 2:34 ` Balbir Singh
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=f3a13f6a-b34c-8561-884a-23fd9aa60331@redhat.com \
--to=david@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=benh@kernel.crashing.org \
--cc=boris.ostrovsky@oracle.com \
--cc=bsingharora@gmail.com \
--cc=corbet@lwn.net \
--cc=dan.j.williams@intel.com \
--cc=devel@linuxdriverproject.org \
--cc=gregkh@linuxfoundation.org \
--cc=haiyangz@microsoft.com \
--cc=heiko.carstens@de.ibm.com \
--cc=iamjoonsoo.kim@lge.com \
--cc=jallen@linux.vnet.ibm.com \
--cc=jgross@suse.com \
--cc=kstewart@linuxfoundation.org \
--cc=kys@microsoft.com \
--cc=lenb@kernel.org \
--cc=linux-acpi@vger.kernel.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=malat@debian.org \
--cc=mhocko@suse.com \
--cc=mikey@neuling.org \
--cc=mpe@ellerman.id.au \
--cc=nfont@linux.vnet.ibm.com \
--cc=osalvador@suse.de \
--cc=pasha.tatashin@oracle.com \
--cc=paulus@samba.org \
--cc=pavel.tatashin@microsoft.com \
--cc=pombredanne@nexb.com \
--cc=rashmica.g@gmail.com \
--cc=rjw@rjwysocki.net \
--cc=schwidefsky@de.ibm.com \
--cc=sthemmin@microsoft.com \
--cc=tglx@linutronix.de \
--cc=vbabka@suse.cz \
--cc=xen-devel@lists.xenproject.org \
--cc=yasu.isimatu@gmail.com \
/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).