[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <651c836a-327c-f91d-fc5d-ab6f59c1b8db@huawei.com>
Date: Wed, 29 Jan 2020 10:27:30 +0000
From: John Garry <john.garry@...wei.com>
To: "Rafael J. Wysocki" <rafael@...nel.org>
CC: Olof Johansson <olof@...om.net>,
"Rafael J. Wysocki" <rjw@...ysocki.net>,
Len Brown <lenb@...nel.org>,
Jeremy Linton <jeremy.linton@....com>,
"Arnd Bergmann" <arnd@...db.de>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
ACPI Devel Maling List <linux-acpi@...r.kernel.org>,
Hanjun Guo <guohanjun@...wei.com>,
"Greg Kroah-Hartman" <gregkh@...uxfoundation.org>
Subject: Re: [PATCH RFC 2/2] soc: Add a basic ACPI generic driver
>>
>>> Any reasons for not putting it into drivers/acpi/pptt.c specifically?
>>> .
>>
>> I don't think so.
>>
>> One thing is that the code does a one-time scan of the PPTT to find all
>> processor package nodes with ID structures to register the soc devices -
>> so we would need some new call from from acpi_init() for that.
>
Hi Rafael,
> Or an extra initcall or similar. [Calls from acpi_init() are basically
> for things that need to be strictly ordered in a specific way for some
> reason.]>
> Why would that be a problem?
I don't see a problem if we want to use a soc driver, but that is
starting to look unlikely.
Alternatively, if we want to create some folder under
/sys/firmware/acpi, any restriction comes from the folder location.
For a folder like /sys/firmware/acpi/pptt, we need to ensure acpi_kobj
is initialized; acpi_kobj is set from subsys_init(acpi_init), so
module_init() for pptt module would suffice.
However if we wanted to make pptt folder a sub-folder from those created
in acpi_sysfs_init() - then we would need to make that parent folder
kobj non-static. Again, module_init() would suffice.
Thanks,
John
Powered by blists - more mailing lists