Re: [PATCH] mm, add_memory_resource: hold device_hotplug lock over mem_hotplug_{begin, done}

From: Dan Williams
Date: Thu Mar 09 2017 - 01:36:47 EST

On Mon, Mar 6, 2017 at 12:22 AM, Heiko Carstens
<heiko.carstens@xxxxxxxxxx> wrote:
> Hello Dan,
>> > If you look at commit 5e33bc4165f3 ("driver core / ACPI: Avoid device hot
>> > remove locking issues") then lock_device_hotplug_sysfs() was introduced to
>> > avoid a different subtle deadlock, but it also sleeps uninterruptible, but
>> > not for more than 5ms ;)
>> >
>> > However I'm not sure if the device hotplug lock should also be used to fix
>> > an unrelated bug that was introduced with the get_online_mems() /
>> > put_online_mems() interface. Should it?
>> No, I don't think it should.
>> I like your proposed direction of creating a new lock internal to
>> mem_hotplug_begin() to protect active_writer, and stop relying on
>> lock_device_hotplug to serve this purpose.
>> > If so, we need to sprinkle around a couple of lock_device_hotplug() calls
>> > near mem_hotplug_begin() calls, like Sebastian already started, and give it
>> > additional semantics (protecting mem_hotplug.active_writer), and hope it
>> > doesn't lead to deadlocks anywhere.
>> I'll put your proposed patch through some testing.
> On s390 it _seems_ to work. Did it pass your testing too?
> If so I would send a patch with proper patch description for inclusion.

Looks ok here. No lockdep warnings running it through it paces with
the persistent memory use case.