Update clock names

[oh]

Summary

TI AM335x update clock names

Due to upstream has changed to use clock-output-names we need to update the names in our code aswell.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/arm/boot/dts/am33xx-clocks.dtsi?id=ec7aa25fa4839e05f73b17a838e46dc591df75f0

Split of https://reviews.freebsd.org/D27889

Will be described in https://wiki.freebsd.org/arm/BeagleBoneBlackMain

[imp]

Don't know about the names

[oh]

Sorry, missed dmtimer that was not part of D27889 but still needs to be updated

[oh]

sync the patch to 2452bcd8913bb45ec269d0a3219ca8bfc0c7a183. No changes in the actual patch, but ti_pruss.c was updated in ef9ffb8594eee294334ced627755bf5b46b48f9f

[oh]

Previously the review also did some cleanup (ti_pruss.c) and silences if its not in the bootverbose mode.

Now, only do the update of the DTS clock names

[rick_sloservers.com]

Tested on hardware as part of this stack: D46703, D46712, D46713, D46714, D41888, D46722, D46723, D46742, D46743, D46744 and D46746 rebased from stable/14 onto stable/15 (6b25004d1d18), KERNCONF=AM335X, on a BeagleBone Black Rev B3. The board boots 15.1-STABLE to multi-user with zero device attach failures. Full results and evidence: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=297800, discussion on D46723.

This revision is accepted, so a note before it lands: the version of this change on the malavon stablebbb/14 branch (commit 6d93b69bcf99) also fixes am335x_musb.c to request usbotg_fck rather than usbotg_fck@47c. That hunk is not in this diff, which touches four files against the branch commit's six.

Without it USB does not attach on a BeagleBone Black: clk_get_by_name() returns ENXIO and musbotg_attach() fails. I rebased from the Phabricator diffs, hit the failure, and diagnosed it before discovering it had already been fixed on the branch. The fix is yours, not mine. Worth folding in before this is committed, or landing alongside it.

The hunk, in musbotg_attach(), so it is a paste rather than a lookup:

-	err = clk_get_by_name(dev, "usbotg_fck@47c", &clk_usbotg_fck);
+	err = clk_get_by_name(dev, "usbotg_fck", &clk_usbotg_fck);
 	if (err) {
-		device_printf(dev, "Can not find usbotg_fck@47c\n");
+		device_printf(dev, "Can not find usbotg_fck\n");
 		return (ENXIO);
 	}
 	err = clk_enable(clk_usbotg_fck);
 	if (err) {
-		device_printf(dev, "Can not enable usbotg_fck@47c\n");
+		device_printf(dev, "Can not enable usbotg_fck\n");
 		return (ENXIO);
 	}

The gate is registered under the name from its clock-output-names property, so the unit-address form never matches. The FIXME comment above it could stay, reworded: the node still has no clocks property, and the gate still lives at offset 0x47c per TRM 8.1.12.2.

Also note commit 46bed482 on that branch is a one-line fixup to this revision; applying this diff without it leaves the build broken.