Re: [PATCH v2 1/2] acpi: support IBM SMBus CMI devices

["Darrick J. Wong" <djwong@us.ibm.com>]

On Tue, Oct 27, 2009 at 06:03:32PM +0100, Jean Delvare wrote:

> I'm only half please with this. You change the function named, but it
> doesn't follow the calling convention of acpi_dock_match(), which is a
> little confusing.
> 
> Anyway, I will need an ack from the ACPI people before I can pick this
> patch. Or maybe they should even push it upstream themselves.

I am confused.  Looking at that bunch of ifs, acpi_is_video_device returns 1
for a match and 0 for no match.  acpi_bay_match returns 0 for a match and
-ENODEV for no match, which just happens to work with the ACPI_SUCCESS macro.
acpi_dock_match returns ACPI error codes.  Each of the three existing tests has
different return value semantics, so it is not clear to me which one I should
use.

I didn't think it was correct for my probe function to use the ACPI_STATUS
macro unless it returned ACPI error codes... which it does not.  -ENODEV seemed
appropriate for "no device found".

Is it desirable to clean them all up to follow the same convention?

--D