Re: [PATCH 5/8] fix hotplug race

[Tejun Heo <teheo@suse.de>]

Tejun Heo wrote:
> Shaohua Li wrote:
>> hotplug notification handler and drivers' notification handler are all
>> running in one workqueue. Before hotplug removes an acpi device, the
>> device driver's notification handler is already be recorded to run just
>> after global notification handler. After hotplug notification handler
>> runs, acpica will notice a NULL notification handler and crash. This
>> patch runs hotplug in other workqueue and wait for all acpi notication
>> handlers finish. This is found in battery hotplug, but actually all
>> hotplug can be affected.
>>
>> Signed-off-by: Zhang Rui <rui.zhang@intel.com>
>> Signed-off-by: Shaohua Li <shaohua.li@intel.com> 
>> ---
>>  drivers/acpi/dock.c     |   24 +++++++++++++++++++++++-
>>  drivers/acpi/osl.c      |   44 +++++++++++++++++++++++++++++++++++++++-----
>>  include/acpi/acpiosxf.h |    3 +++
>>  3 files changed, 65 insertions(+), 6 deletions(-)
> 
> I don't understand the ACPI/dock specific changes but libata-acpi
> definitely looks much prettier this way.  Holger, does everything look okay?

Aiee.. This should have been a reply to PATCH 6/8 not PATCH 5/8.  Sorry.

-- 
tejun