mb/google/rex: Enable Fast V-Mode for MTL-U 15W

This patch sets the Fast V-Mode (FVM) configuration parameter as
suggested in Intel doc 640982. As per the doc, Intel MTL-U 15W CPU
supports FVM on IA and SA.

Fast V-Mode (FVM): Intel Meteor Lake introduces the ability to manage
the peak power events it calls "reactive peak power management".

The Fast V-Mode is one such technique to perform the reactive peak power
management. It relies on the detector integrated inside the processor
which senses when the processor load current exceeds a present threshold
by monitoring the processor power domain IMVP (Intel Mobile Voltage
Positioning) VR sense point.

The baseline ITRIP for IA is 66A and 21A for SA.

BUG=b:286809233
TEST=Able to build and boot google/rex without seeing any performance
regression.

Change-Id: Ia7157bddf2e9586e4a91cc55e48693561072cd05
Signed-off-by: Subrata Banik <subratabanik@google.com>
Reviewed-on: https://review.coreboot.org/c/coreboot/+/75763
Reviewed-by: Kapil Porwal <kapilporwal@google.com>
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
diff --git a/src/mainboard/google/rex/variants/baseboard/rex/devicetree.cb b/src/mainboard/google/rex/variants/baseboard/rex/devicetree.cb
index b21c011..ec57ee7 100644
--- a/src/mainboard/google/rex/variants/baseboard/rex/devicetree.cb
+++ b/src/mainboard/google/rex/variants/baseboard/rex/devicetree.cb
@@ -73,6 +73,17 @@
 	register "pch_hda_idisp_link_frequency" = "HDA_LINKFREQ_96MHZ"
 	register "pch_hda_idisp_codec_enable" = "1"
 
+	# As per doc 640982, Intel MTL-U 15W CPU supports FVM on IA and SA
+	# The ICC Limit is represented in 1/4 A increments, i.e., a value of 400 = 100A
+	# For IA VR configuration
+	register "enable_fast_vmode[VR_DOMAIN_IA]" = "1"
+	register "cep_enable[VR_DOMAIN_IA]" = "1"
+	register "fast_vmode_i_trip[VR_DOMAIN_IA]" = "264" # 66A
+	# For SA VR configuration
+	register "enable_fast_vmode[VR_DOMAIN_SA]" = "1"
+	register "cep_enable[VR_DOMAIN_SA]" = "1"
+	register "fast_vmode_i_trip[VR_DOMAIN_SA]" = "84" # 21A
+
 	device domain 0 on
 		device ref igpu on end
 		device ref dtt on end