Re: [PATCH] staging: rtl8723bs: Remove ACPI table declaration

[Greg Kroah-Hartman]

On Tue, Sep 25, 2018 at 11:44:37PM -0700, Nathan Chancellor wrote:
> On Wed, Sep 26, 2018 at 08:22:45AM +0200, Greg Kroah-Hartman wrote:
> > On Tue, Sep 25, 2018 at 10:39:10PM -0700, Nathan Chancellor wrote:
> > > Clang warns that the acpi_id declaration is not going to be emitted
> > > in the final assembly:
> > > 
> > > drivers/staging/rtl8723bs/os_dep/sdio_intf.c:25:36: warning: variable
> > > 'acpi_ids' is not needed and will not be emitted
> > > [-Wunneeded-internal-declaration]
> > > static const struct acpi_device_id acpi_ids[] = {
> > >                                    ^
> > > 1 warning generated.
> > > 
> > > This is because it's marked as static const and it is not used anywhere
> > > in this file. Doing a git grep on this driver for 'acpi' shows that this
> > > declaration has been unused since the driver's initial induction. Remove
> > > it since it's not doing anything.
> > > 
> > > Link: https://github.com/ClangBuiltLinux/linux/issues/169
> > > Signed-off-by: Nathan Chancellor <natechancellor@xxxxxxxxx>
> > > ---
> > >  drivers/staging/rtl8723bs/os_dep/sdio_intf.c | 6 ------
> > >  1 file changed, 6 deletions(-)
> > > 
> > > diff --git a/drivers/staging/rtl8723bs/os_dep/sdio_intf.c b/drivers/staging/rtl8723bs/os_dep/sdio_intf.c
> > > index 6d02904de63f..d473f9bd08c3 100644
> > > --- a/drivers/staging/rtl8723bs/os_dep/sdio_intf.c
> > > +++ b/drivers/staging/rtl8723bs/os_dep/sdio_intf.c
> > > @@ -22,13 +22,7 @@ static const struct sdio_device_id sdio_ids[] =
> > >  	{ SDIO_DEVICE(0x024c, 0xb723), },
> > >  	{ /* end: all zeroes */				},
> > >  };
> > > -static const struct acpi_device_id acpi_ids[] = {
> > > -	{"OBDA8723", 0x0000},
> > > -	{}
> > > -};
> > > -
> > >  MODULE_DEVICE_TABLE(sdio, sdio_ids);
> > > -MODULE_DEVICE_TABLE(acpi, acpi_ids);
> > 
> > You just removed the ability for the driver to be automatically loaded
> > if that acpi id is present.
> > 
> 
> I am not sure I understand. Every instance of acpi_device_id that I
> looked at in the kernel before sending this patch uses MODULE_DEVICE_TABLE
> but then that acpi_device_id declaration is always used in the driver
> definition either under the acpi_match_table member or ids member
> depending on what type of driver it is. Should this one do that as well?

I don't know, but it's not always necessary.

> Is that even possible with an sdio driver? I apologize if I am not
> making sense, I'm not super familiar with these interfaces.
> 
> I also read 'Documentation/acpi/enumeration.txt' which makes it seem
> like the declaration needs to be added to the device definition as well.

If it is a real acpi driver, yes, you need to actually register it with
the acpi core.

But, if it is just using the auto-load functionality of "if this acpi
string is found, load the module!" then the code is correct as-is.

Yeah, it's a hack, but does work and helps out for auto-loading drivers
for "odd" hardware that is not always self-describing.

thanks,

greg k-h