Re: linux-next: Tree for Aug 19 (scsi/libsas/)

["Gustavo A. R. Silva" <gustavoars@kernel.org>]

Hi,

Please, see my comments below...

On Wed, Aug 19, 2020 at 06:09:34PM +1000, Stephen Rothwell wrote:
> Hi Randy,
> 
> On Tue, 18 Aug 2020 23:30:36 -0700 Randy Dunlap <rdunlap@infradead.org> wrote:
> >
> > Is this some kind of mis-merge?
> > 
> > In sas_discover.c:
> > 
> > 	case SAS_SATA_DEV:
> > 	case SAS_SATA_PM:
> > #ifdef CONFIG_SCSI_SAS_ATA
> > 		error = sas_discover_sata(dev);
> > 		break;
> > #else
> > 		pr_notice("ATA device seen but CONFIG_SCSI_SAS_ATA=N so cannot attach\n");
> > 		fallthrough;
> > #endif
> > 		fallthrough;	/* only for the #else condition above */
> 
> No, that comes from commit
> 
>   58e813cceabd ("treewide: Use fallthrough pseudo-keyword")
> 
> from the kspp-gustavo tree.
> 
> >   CC [M]  drivers/scsi/libsas/sas_discover.o
> > In file included from ./../include/linux/compiler_types.h:65:0,
> >                  from <command-line>:0:
> > ../drivers/scsi/libsas/sas_discover.c: In function 'sas_discover_domain':
> > ../include/linux/compiler_attributes.h:214:41: warning: attribute 'fallthrough' not preceding a case label or default label
> >  # define fallthrough                    __attribute__((__fallthrough__))
> >                                          ^
> > ../drivers/scsi/libsas/sas_discover.c:469:3: note: in expansion of macro 'fallthrough'
> >    fallthrough;
> >    ^~~~~~~~~~~
> >   CC      drivers/ide/ide-eh.o
> > ../include/linux/compiler_attributes.h:214:41: error: invalid use of attribute 'fallthrough'
> >  # define fallthrough                    __attribute__((__fallthrough__))
> >                                          ^
> > ../drivers/scsi/libsas/sas_discover.c:471:3: note: in expansion of macro 'fallthrough'
> >    fallthrough; /* only for the #else condition above */
> >    ^~~~~~~~~~~
> 

I didn't catch this build error locally and, unfortunately, kernel test
robot didn't either, see:

https://lore.kernel.org/lkml/5f3cc99a.HgvOW3rH0mD0RmkM%25lkp@intel.com/

This is fixed in my -next tree now.

Sorry for the inconvenience.

Thanks!
--
Gustavo