mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: "Rafael J. Wysocki" <rjw@sisk.pl>
To: Yinghai Lu <yinghai@kernel.org>
Cc: ACPI Devel Maling List <linux-acpi@vger.kernel.org>,
	LKML <linux-kernel@vger.kernel.org>,
	Bjorn Helgaas <bhelgaas@google.com>, Jiang Liu <liuj97@gmail.com>,
	"Alexander E. Patrakov" <patrakov@gmail.com>
Subject: Re: [PATCH 1/3] ACPI / dock: Rework the handling of notifications
Date: Sun, 30 Jun 2013 16:18:42 +0200	[thread overview]
Message-ID: <1494881.y5W1zydDGl@vostro.rjw.lan> (raw)
In-Reply-To: <CAE9FiQXc9WKDZdABOB8p0soT+OTcRbJb7JTgrPz6ffd2dyPEHA@mail.gmail.com>

On Saturday, June 29, 2013 09:12:59 AM Yinghai Lu wrote:
> On Sat, Jun 29, 2013 at 4:16 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > On Friday, June 28, 2013 04:34:21 PM Yinghai Lu wrote:
> >> On Fri, Jun 28, 2013 at 3:53 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> >> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> >> >
> >> > The ACPI dock driver uses register_acpi_bus_notifier() which
> >> > installs a notifier triggered globally for all system notifications.
> >> > That first of all is inefficient, because the dock driver is only
> >> > interested in notifications associated with the devices it handles,
> >> > but it has to handle all system notifies for all devices.  Moreover,
> >> > it does that even if no docking stations are present in the system
> >> > (CONFIG_ACPI_DOCK set is sufficient for that to happen).  Besides,
> >> > that is inconvenient, because it requires the driver to do extra work
> >> > for each notification to find the target dock station object.
> >> >
> >> > For these reasons, rework the dock driver to install a notify
> >> > handler individually for each dock station in the system using
> >> > acpi_install_notify_handler().  This allows the dock station
> >> > object to be passed directly to the notify handler and makes it
> >> > possible to simplify the dock driver quite a bit.  It also
> >> > reduces the overhead related to the handling of all system
> >> > notifies when CONFIG_ACPI_DOCK is set.
> >>
> >> original change to use register_acpi_bus_notifier, have two assumption
> >> 1.  two dock_station will have same handle.
> >
> > Well, that would mean that dock_add() might be called twice for the same handle
> > and I don't see how that's possible.
> >
> > Moreover, even if that were possible, the loop in acpi_dock_notifier_call()
> > would break after finding the *first* matching handle anyway, so
> > acpi_dock_deferred_cb() wouldn't be called for the second dock station with
> > the same handle, if there were two.
> 
> related commit:
> commit 6bd00a61ab63d4ceb635ae0316353c11c900b8d8
> Author: Shaohua Li <shaohua.li@intel.com>
> Date:   Thu Aug 28 10:04:29 2008 +0800
> 
>     ACPI: introduce notifier change to avoid duplicates
> 
>     The battery driver already registers notification handler.
>     To avoid registering notification handler again,
>     introduce a notifier chain in global system notifier handler
>     and use it in dock driver.
> 
> so it is not two dock station have same handle. it is battery acpi driver.
> 
> but notitifer installing is changed
> 
> commit d94066910943837558d2a461c6766da981260bf0
> Author: Bjorn Helgaas <bjorn.helgaas@hp.com>
> Date:   Thu Apr 30 09:35:47 2009 -0600
> 
>     ACPI: battery: use .notify method instead of installing handler directly
> 
>     This patch adds a .notify() method.  The presence of .notify() causes
>     Linux/ACPI to manage event handlers and notify handlers on our behalf,
>     so we don't have to install and remove them ourselves.
> 
>     This driver apparently relies on seeing ALL notify events, not just
>     device-specific ones (because it used ACPI_ALL_NOTIFY).  We use the
>     ACPI_DRIVER_ALL_NOTIFY_EVENTS driver flag to request all events.
> 
>     Signed-off-by: Bjorn Helgaas <bjorn.helgaas@hp.com>
>     CC: Alexey Starikovskiy <alexey.y.starikovskiy@linux.intel.com>
>     Signed-off-by: Len Brown <len.brown@intel.com>
> 
> 
> >
> >> 2. acpi subsystem: non root acpi device only can have one system
> >> notifier installed.
> >
> > No, that limitation is long gone.  We removed it when we were working on ACPI
> > wakeup support for runtime PM.
> 
> looks like i misread that...
> 
>  * NOTES:       The Root namespace object may have only one handler for each
>  *              type of notify (System/Device). Device/Thermal/Processor objects
>  *              may have one device notify handler, and multiple system notify
>  *              handlers.
> 
> device could have multiple system notify handlers.
> 
> So
> 
> Acked-by: Yinghai Lu <yinghai@kernel.org>

Thanks (and for the other ACKs too)!


  reply	other threads:[~2013-06-30 14:09 UTC|newest]

Thread overview: 23+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-06-28 19:44 [PATCH 0/4] ACPI / dock: Cleanups and simplifications Rafael J. Wysocki
2013-06-28 19:45 ` [PATCH 1/4] ACPI / dock: Drop the hp_lock mutex from struct dock_station Rafael J. Wysocki
2013-06-28 19:46 ` [PATCH 2/4] ACPI / dock: Rework and simplify find_dock_devices() Rafael J. Wysocki
2013-06-28 20:08   ` Yinghai Lu
2013-06-28 19:47 ` [PATCH 3/4] ACPI / dock: Walk list in reverse order during removal of devices Rafael J. Wysocki
2013-06-28 20:14   ` Yinghai Lu
2013-06-28 19:48 ` [PATCH 4/4] ACPI / dock: Simplify dock_init_hotplug() and dock_release_hotplug() Rafael J. Wysocki
2013-06-28 20:16   ` Yinghai Lu
2013-06-28 22:47 ` [PATCH 0/3] ACPI / dock: Notification rework and memory leak fix Rafael J. Wysocki
2013-06-28 22:53   ` [PATCH 1/3] ACPI / dock: Rework the handling of notifications Rafael J. Wysocki
2013-06-28 23:34     ` Yinghai Lu
2013-06-29 11:16       ` Rafael J. Wysocki
2013-06-29 16:12         ` Yinghai Lu
2013-06-30 14:18           ` Rafael J. Wysocki [this message]
2013-07-01 20:21     ` Bjorn Helgaas
2013-07-02  1:28       ` Rafael J. Wysocki
2013-06-28 22:54   ` [PATCH 2/3] ACPI: Drop ACPI bus notifier call chain Rafael J. Wysocki
2013-06-28 22:54   ` [PATCH 3/3] ACPI / dock: Do not leak memory on falilures to add a dock station Rafael J. Wysocki
2013-07-03 23:23   ` [PATCH 0/4] ACPI / dock: One fix and more cleanups Rafael J. Wysocki
2013-07-03 23:25     ` [PATCH 1/4] ACPI / dock: Actually define acpi_dock_init() as void Rafael J. Wysocki
2013-07-03 23:26     ` [PATCH 2/4] ACPI / dock: Do not check CONFIG_ACPI_DOCK_MODULE Rafael J. Wysocki
2013-07-03 23:27     ` [PATCH 3/4] ACPI / dock / PCI: Drop ACPI dock notifier chain Rafael J. Wysocki
2013-07-03 23:28     ` [PATCH 4/4] ACPI / dock: Drop unnecessary local variable from dock_add() Rafael J. Wysocki

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=1494881.y5W1zydDGl@vostro.rjw.lan \
    --to=rjw@sisk.pl \
    --cc=bhelgaas@google.com \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=liuj97@gmail.com \
    --cc=patrakov@gmail.com \
    --cc=yinghai@kernel.org \
    /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

all inboxes | Powered by JetHome®