svn commit: r292788 - in head/sys: geom/part sys

[Konstantin Belousov]

On Mon, Dec 28, 2015 at 11:32:29AM +0300, Andrey V. Elsukov wrote:
> On 27.12.15 21:12, Allan Jude wrote:
> >  enum g_part_alias {
> >  	G_PART_ALIAS_APPLE_BOOT,	/* An Apple boot partition entry. */
> > +	G_PART_ALIAS_APPLE_CORE_STORAGE,/* An Apple Core Storage partition. */
> >  	G_PART_ALIAS_APPLE_HFS,		/* An HFS+ file system entry. */
> >  	G_PART_ALIAS_APPLE_LABEL,	/* An Apple label partition entry. */
> >  	G_PART_ALIAS_APPLE_RAID,	/* An Apple RAID partition entry. */
> >  	G_PART_ALIAS_APPLE_RAID_OFFLINE,/* An Apple RAID (offline) part entry.*/
> >  	G_PART_ALIAS_APPLE_TV_RECOVERY,	/* An Apple TV recovery part entry. */
> >  	G_PART_ALIAS_APPLE_UFS,		/* An Apple UFS partition entry. */
> > +	G_PART_ALIAS_BIOS_BOOT,		/* A GRUB 2 boot partition entry. */
> > +	G_PART_ALIAS_CHROMEOS_FIRMWARE,	/* A ChromeOS firmware part. entry. */
> > +	G_PART_ALIAS_CHROMEOS_KERNEL,	/* A ChromeOS Kernel part. entry. */
> > +	G_PART_ALIAS_CHROMEOS_RESERVED,	/* ChromeOS. Reserved for future use. */
> > +	G_PART_ALIAS_CHROMEOS_ROOT,	/* A ChromeOS root part. entry. */
> 
> This enum could be used by third-party kernel module and in general it
> is not good to change the order of elements here. This can break `gpart
> show` output for such modules. Just for the future. :)

If the enum values are considered part of KBI, then explicit assignment
of the enum values is much preferred.