[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAE9FiQWOj2QcW9M6c8WZi-EJkUqUeSC4VsDfWUVch+wDLxTjEw@mail.gmail.com>
Date: Fri, 19 Oct 2012 13:36:46 -0700
From: Yinghai Lu <yinghai@...nel.org>
To: Tang Chen <tangchen@...fujitsu.com>
Cc: jiang.liu@...wei.com, bhelgaas@...gle.com, lenb@...nel.org,
izumi.taku@...fujitsu.com, isimatu.yasuaki@...fujitsu.com,
linux-acpi@...r.kernel.org, linux-pci@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/1] Update acpi_root_bridge_list in container hotplug path.
On Fri, Oct 19, 2012 at 1:49 AM, Tang Chen <tangchen@...fujitsu.com> wrote:
> Is it just container and pci_root_bridge hot-remove need to call
> acpi_bus_trim() twice ? For normal device without sub-device, I think
> it is OK to call acpi_bus_trim(device, 1).
>
> The reason why I'm asking this question is:
>
> I saw in acpi_bus_hot_remove_device(), it almost does the same things
> you did in handle_root_bridge_removal(), except calling acpi_bus_trim()
> twice. And there are more than one path could do container hot-remove.
>
> If I add a container_device_remove() doing the similar things, it could
> be duplicated. So, shall we just remove handle_root_bridge_removal(),
> and only use acpi_bus_hot_remove_device() ?
>
> Of course, we need to call acpi_bus_trim() twice in
> acpi_bus_hot_remove_device().
yes. we could use acpi_bus_hot_remove_device to simply
acpi_bus_hot_remove_device()
but that eject_event should have device instead of passing handle, and
later check if that device is there or not.
like attached two patches.
Download attachment "root_bridge_remove_1.patch" of type "application/octet-stream" (2197 bytes)
Download attachment "root_bridge_remove_2.patch" of type "application/octet-stream" (2519 bytes)
Powered by blists - more mailing lists