From: Toshi Kani <toshi.kani@hp.com>
To: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: linux-acpi@vger.kernel.org,
Vasilis Liaskovitis <vasilis.liaskovitis@profitbricks.com>,
Wen Congyang <wency@cn.fujitsu.com>,
Wen Congyang <wencongyang@gmail.com>,
isimatu.yasuaki@jp.fujitsu.com, lenb@kernel.org,
gregkh@linuxfoundation.org, linux-kernel@vger.kernel.org,
linux-mm@kvack.org
Subject: Re: [RFC PATCH v3 3/3] acpi_memhotplug: Allow eject to proceed on rebind scenario
Date: Wed, 28 Nov 2012 14:02:48 -0700 [thread overview]
Message-ID: <1354136568.26955.312.camel@misato.fc.hp.com> (raw)
In-Reply-To: <9212118.3s2xH6uJDI@vostro.rjw.lan>
> > > > > > > Consider the following case:
> > > > > > >
> > > > > > > We hotremove the memory device by SCI and unbind it from the driver at the same time:
> > > > > > >
> > > > > > > CPUa CPUb
> > > > > > > acpi_memory_device_notify()
> > > > > > > unbind it from the driver
> > > > > > > acpi_bus_hot_remove_device()
> > > > > >
> > > > > > Can we make acpi_bus_remove() to fail if a given acpi_device is not
> > > > > > bound with a driver? If so, can we make the unbind operation to perform
> > > > > > unbind only?
> > > > >
> > > > > acpi_bus_remove_device could check if the driver is present, and return -ENODEV
> > > > > if it's not present (dev->driver == NULL).
> > > > >
> > > > > But there can still be a race between an eject and an unbind operation happening
> > > > > simultaneously. This seems like a general problem to me i.e. not specific to an
> > > > > acpi memory device. How do we ensure an eject does not race with a driver unbind
> > > > > for other acpi devices?
> > > > >
> > > > > Is there a per-device lock in acpi-core or device-core that can prevent this from
> > > > > happening? Driver core does a device_lock(dev) on all operations, but this is
> > > > > probably not grabbed on SCI-initiated acpi ejects.
> > > >
> > > > Since driver_unbind() calls device_lock(dev->parent) before calling
> > > > device_release_driver(), I am wondering if we can call
> > > > device_lock(dev->dev->parent) at the beginning of acpi_bus_remove()
> > > > (i.e. before calling pre_remove) and fails if dev->driver is NULL. The
> > > > parent lock is otherwise released after device_release_driver() is done.
> > >
> > > I would be careful. You may introduce some subtle locking-related issues
> > > this way.
> >
> > Right. This requires careful inspection and testing. As far as the
> > locking is concerned, I am not keen on using fine grained locking for
> > hot-plug. It is much simpler and solid if we serialize such operations.
> >
> > > Besides, there may be an alternative approach to all this. For example,
> > > what if we don't remove struct device objects on eject? The ACPI handles
> > > associated with them don't go away in that case after all, do they?
> >
> > Umm... Sorry, I am not getting your point. The issue is that we need
> > to be able to fail a request when memory range cannot be off-lined.
> > Otherwise, we end up ejecting online memory range.
>
> Yes, this is the major one. The minor issue, however, is a race condition
> between unbinding a driver from a device and removing the device if I
> understand it correctly. Which will go away automatically if the device is
> not removed in the first place. Or so I would think. :-)
I see. I do not think whether or not the device is removed on eject
makes any difference here. The issue is that after driver_unbind() is
done, acpi_bus_hot_remove_device() no longer calls the ACPI memory
driver (hence, it cannot fail in prepare_remove), and goes ahead to call
_EJ0. If driver_unbind() did off-line the memory, this is OK. However,
it cannot off-line kernel memory ranges. So, we basically need to
either 1) serialize acpi_bus_hot_remove_device() and driver_unbind(), or
2) make acpi_bus_hot_remove_device() to fail if driver_unbind() is run
during the operation.
Thanks,
-Toshi
next prev parent reply other threads:[~2012-11-28 21:11 UTC|newest]
Thread overview: 92+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-11-23 17:50 [RFC PATCH v3 0/3] acpi: Introduce prepare_remove device operation Vasilis Liaskovitis
2012-11-23 17:50 ` [RFC PATCH v3 1/3] acpi: Introduce prepare_remove operation in acpi_device_ops Vasilis Liaskovitis
2012-11-27 0:10 ` Toshi Kani
2012-11-27 18:36 ` Vasilis Liaskovitis
2012-11-27 23:18 ` Rafael J. Wysocki
2012-11-23 17:50 ` [RFC PATCH v3 2/3] acpi_memhotplug: Add prepare_remove operation Vasilis Liaskovitis
2012-11-24 16:23 ` Wen Congyang
2012-11-23 17:50 ` [RFC PATCH v3 3/3] acpi_memhotplug: Allow eject to proceed on rebind scenario Vasilis Liaskovitis
2012-11-24 16:20 ` Wen Congyang
2012-11-26 8:36 ` Vasilis Liaskovitis
2012-11-26 9:11 ` Wen Congyang
2012-11-27 0:19 ` Toshi Kani
2012-11-27 18:32 ` Vasilis Liaskovitis
2012-11-27 22:03 ` Toshi Kani
2012-11-27 23:41 ` Rafael J. Wysocki
2012-11-28 16:01 ` Toshi Kani
2012-11-28 18:40 ` Rafael J. Wysocki
2012-11-28 21:02 ` Toshi Kani [this message]
2012-11-28 21:40 ` Rafael J. Wysocki
2012-11-28 21:40 ` Toshi Kani
2012-11-28 22:01 ` Rafael J. Wysocki
2012-11-28 22:04 ` Toshi Kani
2012-11-28 22:21 ` Rafael J. Wysocki
2012-11-28 22:16 ` Toshi Kani
2012-11-28 22:39 ` Rafael J. Wysocki
2012-11-28 22:46 ` Greg KH
2012-11-28 23:05 ` Rafael J. Wysocki
2012-11-28 23:10 ` Greg KH
2012-11-28 23:31 ` Rafael J. Wysocki
2012-11-28 23:49 ` Rafael J. Wysocki
2012-11-29 1:02 ` Toshi Kani
2012-11-29 1:15 ` Toshi Kani
2012-11-29 10:03 ` Rafael J. Wysocki
2012-11-29 11:30 ` Vasilis Liaskovitis
2012-11-29 16:57 ` Rafael J. Wysocki
2012-11-29 17:56 ` Toshi Kani
2012-11-29 20:25 ` Rafael J. Wysocki
2012-11-29 20:38 ` Toshi Kani
2012-11-29 21:23 ` Rafael J. Wysocki
2012-11-29 21:46 ` Toshi Kani
2012-11-29 22:11 ` Rafael J. Wysocki
2012-11-29 23:17 ` Toshi Kani
2012-11-30 0:13 ` Rafael J. Wysocki
2012-11-30 1:09 ` Toshi Kani
2012-11-29 16:43 ` Toshi Kani
2012-11-29 11:04 ` Vasilis Liaskovitis
2012-11-29 17:44 ` Toshi Kani
2012-12-06 9:30 ` Vasilis Liaskovitis
2012-12-06 12:50 ` Rafael J. Wysocki
2012-12-06 15:41 ` Toshi Kani
2012-12-06 20:32 ` Rafael J. Wysocki
2012-11-28 11:05 ` [RFC PATCH v3 0/3] acpi: Introduce prepare_remove device operation Hanjun Guo
2012-11-28 18:41 ` Toshi Kani
2012-11-29 4:48 ` Hanjun Guo
2012-11-29 22:27 ` Toshi Kani
2012-12-03 4:25 ` Hanjun Guo
2012-12-04 0:10 ` Toshi Kani
2012-12-04 9:16 ` Hanjun Guo
2012-12-04 23:23 ` Toshi Kani
2012-12-05 12:10 ` Hanjun Guo
2012-12-05 22:31 ` Toshi Kani
2012-12-06 16:47 ` Jiang Liu
2012-12-07 2:25 ` Toshi Kani
2012-12-06 16:40 ` Jiang Liu
2012-12-06 20:30 ` Rafael J. Wysocki
2012-12-07 2:57 ` Toshi Kani
2012-12-07 5:57 ` Jiang Liu
2012-12-08 1:08 ` Toshi Kani
2012-12-11 14:34 ` Jiang Liu
2012-12-13 14:42 ` Toshi Kani
2012-12-13 15:15 ` Jiang Liu
2012-12-15 1:19 ` Toshi Kani
2012-11-29 10:15 ` Rafael J. Wysocki
2012-11-29 11:36 ` Vasilis Liaskovitis
2012-12-06 16:59 ` Jiang Liu
2012-11-29 17:03 ` Toshi Kani
2012-11-29 20:30 ` Rafael J. Wysocki
2012-11-29 20:39 ` Toshi Kani
2012-11-29 20:56 ` Toshi Kani
2012-11-29 21:25 ` Rafael J. Wysocki
2012-12-06 17:10 ` Jiang Liu
2012-12-06 17:07 ` Jiang Liu
2012-12-06 17:01 ` Jiang Liu
2012-12-06 16:56 ` Jiang Liu
2012-12-06 16:00 ` Jiang Liu
2012-12-06 16:03 ` Toshi Kani
2012-12-06 16:25 ` Jiang Liu
2012-12-06 16:31 ` Toshi Kani
2012-12-06 16:52 ` Jiang Liu
2012-12-06 17:09 ` Toshi Kani
2012-12-06 17:30 ` Jiang Liu
2012-12-06 17:28 ` Toshi Kani
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=1354136568.26955.312.camel@misato.fc.hp.com \
--to=toshi.kani@hp.com \
--cc=gregkh@linuxfoundation.org \
--cc=isimatu.yasuaki@jp.fujitsu.com \
--cc=lenb@kernel.org \
--cc=linux-acpi@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=rjw@sisk.pl \
--cc=vasilis.liaskovitis@profitbricks.com \
--cc=wencongyang@gmail.com \
--cc=wency@cn.fujitsu.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).