Create the GELIBOOT GEOM_ELI flag

[allanjude]

Summary

This flag indicates that the user wishes to use the GELIBOOT feature, to boot from a fully encrypted root file system.
Currently, GELIBOOT does not support key files, and when it does, they will be done a bit differently.

Due to the design of GELI, and the desire for secrecy, the GELI metadata does not know if key files are used or not, it just adds the key material (if it exists) to the HMAC before the optional user password, so there is no way to tell if a GELI partition has key files or not.
Since the GELIBOOT code in boot2 and the loader do not support keys, they will only attempt to attach if this flag is set

This will stop GELIBOOT from prompting for passwords to GELIs that it cannot decrypt, disrupting the boot process

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=208251

[oshogbo]

Small comment.

sbin/geom/class/eli/geom_eli.c

937

md.md_flags & G_ELI_FLAG_GELIBOOT is not a boolean equation you should compare it. Ideal would be:
(md.md_flags & G_ELI_FLAG_GELIBOOT) == G_ELI_FLAG_GELIBOOT

Sometimes good enough version would be:
(md.md_flags & G_ELI_FLAG_GELIBOOT) != 0

940

Same.

sys/geom/eli/g_eli_ctl.c

485

same.

489

same.

[oshogbo]

Ah I see now that all in the file is done the same comparison so not sure if you should change that.

[wblock]

sbin/geom/class/eli/geli.8

297

Couldn't this just be:

Enable booting from encrypted root filesystem.

298

This does not seem necessary, and "unencrypted" combined with the earlier "encrypted" makes for a confusing logical "not".

299

Simplify, passive -> active, and avoid "you" and "your":

The the boot loader prompts for the passphrase and loads
.Xr loader 8
from the encrypted partition.

494

Repeating this identically here does not seem too helpful.

501

Let's define what it does so the user does not need to know a new term:

Deactivate booting from an encrypted partition.

[allanjude]

In D5867#125174, @oshogbo wrote:

If you think it is worth it to fix these, I'll do it as a separate review after this one

[allanjude]

Address manpage feedback from wblock

[ed]

sbin/geom/class/eli/geom_eli.c

1001

It seems like we're proliferating tri-state options: yes, no and unspecified. Would it make sense to just come up with an enum somewhere at the top of this file and use those values instead of -1, 0 and 1?

[allanjude]

sbin/geom/class/eli/geom_eli.c

1001

Probably a good idea. I can do this as a separate commit after.