FreeBSD_HEAD_i386 - Build #893 - Fixed

FreeBSD_HEAD_i386 - Build #893 - Fixed:

Build information: https://jenkins.FreeBSD.org/job/FreeBSD_HEAD_i386/893/
Full change log: https://jenkins.FreeBSD.org/job/FreeBSD_HEAD_i386/893/changes
Full build log: https://jenkins.FreeBSD.org/job/FreeBSD_HEAD_i386/893/console

Change summaries:

286945 by melifaro:
Check value return from lle_create() for NULL.
This bug sneaked unnoticed in r286722.

Reported by:	adrian

286944 by ian:
Enable the watchdog driver on imx6, now that it works.

286943 by ian:
Make the imx watchdog actually work, by setting WDOG_CR_WDE (enable bit).
Also, follow the rules from watchdog(9) about what values to return in
various situations (especially, don't touch *error when asked to set a
non-zero timeout that isn't achievable on the hardware).

286942 by ian:
Add compatible strings for all the hardware this driver works with.

Also, move the READ/WRITE bus space access macros from the header into the
source file, and rename them to RD2/WR2 to make it clear they're 16-bit
accessors.  (READ/WRITE just don't seem like good names to be in a public
header file.)

286941 by jilles:
wordexp(): Improve some error codes.

Distinguish between WRDE_BADVAL and WRDE_SYNTAX based on when the error
occurred (parsing or execution), not based on whether WRDE_UNDEF was passed.

Also, return WRDE_NOSPACE for a few more unexpected results from sh.

286940 by jhb:
ino_t is unsigned, so use uintmax_t instead of intmax_t when printing it.

Submitted by:	bde (sort of)

286939 by jhb:
Always use %j with an intmax_t cast to print time_t values.  time_t is
longer than long on 32-bit platforms with a 64-bit time_t.

Inspired by:	mail from bde

286938 by jhb:
Various style and whitespace fixes.

286937 by jhb:
Use nitems().