nxp: adding support of soc lx2160a

* NXP SoC is 16 A-72 core SoC.
* SoC specific defines are defined in:
  - soc.def
  - soc.h
* Called for BL2 and BL31 setup, SoC specific setup are implemented in:
  - soc.c
* platform specific helper functions implemented at:
  - aarch64/lx2160a_helpers.S
* platform specific functions used by 'plat/nxp/commpon/psci',
  etc. are implemented at:
  - aarch64/lx2160a.S
* platform specific implementation for handling PSCI_SYSTEM_RESET2:
  - aarch64/lx2160a_warm_rst.S

Signed-off-by: rocket <rod.dorris@nxp.com>
Signed-off-by: Pankaj Gupta <pankaj.gupta@nxp.com>
Change-Id: Ib40086f9d9079ed9b22967baff518c6df9f408b8
diff --git a/plat/nxp/soc-lx2160a/aarch64/lx2160a.S b/plat/nxp/soc-lx2160a/aarch64/lx2160a.S
new file mode 100644
index 0000000..4679fc2
--- /dev/null
+++ b/plat/nxp/soc-lx2160a/aarch64/lx2160a.S
@@ -0,0 +1,1824 @@
+/*
+ * Copyright 2018-2020 NXP
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ *
+ */
+
+.section .text, "ax"
+
+#include <asm_macros.S>
+
+#include <lib/psci/psci.h>
+#include <nxp_timer.h>
+#include <plat_gic.h>
+#include <pmu.h>
+
+#include <bl31_data.h>
+#include <plat_psci.h>
+#include <platform_def.h>
+
+.global soc_init_start
+.global soc_init_percpu
+.global soc_init_finish
+.global _set_platform_security
+.global _soc_set_start_addr
+
+.global _soc_core_release
+.global _soc_ck_disabled
+.global _soc_core_restart
+.global _soc_core_prep_off
+.global _soc_core_entr_off
+.global _soc_core_exit_off
+.global _soc_sys_reset
+.global _soc_sys_off
+.global _soc_core_prep_stdby
+.global _soc_core_entr_stdby
+.global _soc_core_exit_stdby
+.global _soc_core_prep_pwrdn
+.global _soc_core_entr_pwrdn
+.global _soc_core_exit_pwrdn
+.global _soc_clstr_prep_stdby
+.global _soc_clstr_exit_stdby
+.global _soc_clstr_prep_pwrdn
+.global _soc_clstr_exit_pwrdn
+.global _soc_sys_prep_stdby
+.global _soc_sys_exit_stdby
+.global _soc_sys_prep_pwrdn
+.global _soc_sys_pwrdn_wfi
+.global _soc_sys_exit_pwrdn
+
+.equ TZPC_BASE,			  0x02200000
+.equ TZPCDECPROT_0_SET_BASE, 0x02200804
+.equ TZPCDECPROT_1_SET_BASE, 0x02200810
+.equ TZPCDECPROT_2_SET_BASE, 0x0220081C
+
+#define CLUSTER_3_CORES_MASK 0xC0
+#define CLUSTER_3_IN_RESET  1
+#define CLUSTER_3_NORMAL	0
+
+/* cluster 3 handling no longer based on frequency, but rather on RCW[850],
+ * which is bit 18 of RCWSR27
+ */
+#define CLUSTER_3_RCW_BIT  0x40000
+
+/* retry count for clock-stop acks */
+.equ CLOCK_RETRY_CNT,  800
+
+/* disable prefetching in the A72 core */
+#define  CPUACTLR_DIS_LS_HW_PRE	0x100000000000000
+#define  CPUACTLR_DIS_L2_TLB_PRE   0x200000
+
+/* Function starts the initialization tasks of the soc,
+ * using secondary cores if they are available
+ *
+ * Called from C, saving the non-volatile regs
+ * save these as pairs of registers to maintain the
+ * required 16-byte alignment on the stack
+ *
+ * in:
+ * out:
+ * uses x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11
+ */
+func soc_init_start
+	stp  x4,  x5,  [sp, #-16]!
+	stp  x6,  x7,  [sp, #-16]!
+	stp  x8,  x9,  [sp, #-16]!
+	stp  x10, x11, [sp, #-16]!
+	stp  x12, x13, [sp, #-16]!
+	stp  x18, x30, [sp, #-16]!
+
+	/* make sure the personality has been
+	 * established by releasing cores that
+	 * are marked "to-be-disabled" from reset
+	 */
+	bl  release_disabled  		/* 0-9 */
+
+	/* init the task flags */
+	bl  _init_task_flags   		/* 0-1 */
+
+	/* set SCRATCHRW7 to 0x0 */
+	ldr  x0, =DCFG_SCRATCHRW7_OFFSET
+	mov  x1, xzr
+	bl   _write_reg_dcfg
+
+1:
+	/* restore the aarch32/64 non-volatile registers */
+	ldp  x18, x30, [sp], #16
+	ldp  x12, x13, [sp], #16
+	ldp  x10, x11, [sp], #16
+	ldp  x8,  x9,  [sp], #16
+	ldp  x6,  x7,  [sp], #16
+	ldp  x4,  x5,  [sp], #16
+	ret
+endfunc soc_init_start
+
+
+/* Function performs any soc-specific initialization that is needed on
+ * a per-core basis.
+ * in:  none
+ * out: none
+ * uses x0, x1, x2, x3
+ */
+func soc_init_percpu
+	stp  x4,  x30,  [sp, #-16]!
+
+	bl   plat_my_core_mask
+	mov  x2, x0				/* x2 = core mask */
+
+	/* Check if this core is marked for prefetch disable
+	 */
+	mov   x0, #PREFETCH_DIS_OFFSET
+	bl	_get_global_data		/* 0-1 */
+	tst   x0, x2
+	b.eq  1f
+	bl	_disable_ldstr_pfetch_A72	/* 0 */
+1:
+	mov  x0, #NXP_PMU_ADDR
+	bl enable_timer_base_to_cluster
+	ldp  x4,  x30,  [sp], #16
+	ret
+endfunc soc_init_percpu
+
+
+/* Function completes the initialization tasks of the soc
+ * in:
+ * out:
+ * uses x0, x1, x2, x3, x4
+ */
+func soc_init_finish
+	stp  x4,  x30,  [sp, #-16]!
+
+	ldp   x4,  x30,  [sp], #16
+	ret
+endfunc soc_init_finish
+
+
+/* Function sets the security mechanisms in the SoC to implement the
+ * Platform Security Policy
+ */
+func _set_platform_security
+	mov  x8, x30
+
+#if (!SUPPRESS_TZC)
+	/* initialize the tzpc */
+	bl   init_tzpc
+#endif
+
+#if (!SUPPRESS_SEC)
+	/* initialize secmon */
+#ifdef NXP_SNVS_ENABLED
+	mov x0, #NXP_SNVS_ADDR
+	bl  init_sec_mon
+#endif
+#endif
+
+	mov  x30, x8
+	ret
+endfunc _set_platform_security
+
+
+/* Function writes a 64-bit address to bootlocptrh/l
+ * in:  x0, 64-bit address to write to BOOTLOCPTRL/H
+ * uses x0, x1, x2
+ */
+func _soc_set_start_addr
+	/* Get the 64-bit base address of the dcfg block */
+	ldr  x2, =NXP_DCFG_ADDR
+
+	/* write the 32-bit BOOTLOCPTRL register */
+	mov  x1, x0
+	str  w1, [x2, #DCFG_BOOTLOCPTRL_OFFSET]
+
+	/* write the 32-bit BOOTLOCPTRH register */
+	lsr  x1, x0, #32
+	str  w1, [x2, #DCFG_BOOTLOCPTRH_OFFSET]
+	ret
+endfunc _soc_set_start_addr
+
+/* Function releases a secondary core from reset
+ * in:   x0 = core_mask_lsb
+ * out:  none
+ * uses: x0, x1, x2, x3
+ */
+func _soc_core_release
+	mov   x3, x30
+
+	ldr  x1, =NXP_SEC_REGFILE_ADDR
+	/* write to CORE_HOLD to tell
+	 * the bootrom that this core is
+	 * expected to run.
+	 */
+	str  w0, [x1, #CORE_HOLD_OFFSET]
+
+	/* read-modify-write BRRL to release core */
+	mov  x1, #NXP_RESET_ADDR
+	ldr  w2, [x1, #BRR_OFFSET]
+
+	/* x0 = core mask */
+	orr  w2, w2, w0
+	str  w2, [x1, #BRR_OFFSET]
+	dsb  sy
+	isb
+
+	/* send event */
+	sev
+	isb
+
+	mov   x30, x3
+	ret
+endfunc _soc_core_release
+
+
+/* Function determines if a core is disabled via COREDISABLEDSR
+ * in:  w0  = core_mask_lsb
+ * out: w0  = 0, core not disabled
+ *	  w0 != 0, core disabled
+ * uses x0, x1
+ */
+func _soc_ck_disabled
+
+	/* get base addr of dcfg block */
+	ldr  x1, =NXP_DCFG_ADDR
+
+	/* read COREDISABLEDSR */
+	ldr  w1, [x1, #DCFG_COREDISABLEDSR_OFFSET]
+
+	/* test core bit */
+	and  w0, w1, w0
+
+	ret
+endfunc _soc_ck_disabled
+
+
+/* Part of CPU_ON
+ * Function restarts a core shutdown via _soc_core_entr_off
+ * in:  x0 = core mask lsb (of the target cpu)
+ * out: x0 == 0, on success
+ *	  x0 != 0, on failure
+ * uses x0, x1, x2, x3, x4, x5, x6
+ */
+func _soc_core_restart
+	mov  x6, x30
+	mov  x4, x0
+
+	/* pgm GICD_CTLR - enable secure grp0  */
+	mov  x5, #NXP_GICD_ADDR
+	ldr  w2, [x5, #GICD_CTLR_OFFSET]
+	orr  w2, w2, #GICD_CTLR_EN_GRP_0
+	str  w2, [x5, #GICD_CTLR_OFFSET]
+	dsb sy
+	isb
+
+	/* poll on RWP til write completes */
+4:
+	ldr  w2, [x5, #GICD_CTLR_OFFSET]
+	tst  w2, #GICD_CTLR_RWP
+	b.ne 4b
+
+	/* x4 = core mask lsb
+	* x5 = gicd base addr
+	*/
+	mov  x0, x4
+	bl   get_mpidr_value
+
+	/* x0 = mpidr of target core
+	* x4 = core mask lsb of target core
+	* x5 = gicd base addr
+	*/
+
+	/* generate target list bit */
+	and  x1, x0, #MPIDR_AFFINITY0_MASK
+	mov  x2, #1
+	lsl  x2, x2, x1
+
+	/* get the affinity1 field */
+	and  x1, x0, #MPIDR_AFFINITY1_MASK
+	lsl  x1, x1, #8
+	orr  x2, x2, x1
+
+	/* insert the INTID for SGI15 */
+	orr  x2, x2, #ICC_SGI0R_EL1_INTID
+
+	/* fire the SGI */
+	msr  ICC_SGI0R_EL1, x2
+	dsb  sy
+	isb
+
+	/* load '0' on success */
+	mov  x0, xzr
+
+	mov  x30, x6
+	ret
+endfunc _soc_core_restart
+
+
+/* Part of CPU_OFF
+ * Function programs SoC & GIC registers in preparation for shutting down
+ * the core
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0, x1, x2, x3, x4, x5, x6, x7
+ */
+func _soc_core_prep_off
+	mov  x8, x30
+	mov  x7, x0		/* x7 = core mask lsb */
+
+	mrs  x1, CORTEX_A72_ECTLR_EL1
+
+	/* set smp and disable L2 snoops in cpuectlr */
+	orr  x1, x1, #CPUECTLR_SMPEN_EN
+	orr  x1, x1, #CPUECTLR_DISABLE_TWALK_PREFETCH
+	bic  x1, x1, #CPUECTLR_INS_PREFETCH_MASK
+	bic  x1, x1, #CPUECTLR_DAT_PREFETCH_MASK
+
+	/* set retention control in cpuectlr */
+	bic  x1, x1, #CPUECTLR_TIMER_MASK
+	orr  x1, x1, #CPUECTLR_TIMER_8TICKS
+	msr  CORTEX_A72_ECTLR_EL1, x1
+
+	/* get redistributor rd base addr for this core */
+	mov  x0, x7
+	bl   get_gic_rd_base
+	mov  x6, x0
+
+	/* get redistributor sgi base addr for this core */
+	mov  x0, x7
+	bl   get_gic_sgi_base
+	mov  x5, x0
+
+	/* x5 = gicr sgi base addr
+ 	 * x6 = gicr rd  base addr
+	 * x7 = core mask lsb
+	 */
+
+	/* disable SGI 15 at redistributor - GICR_ICENABLER0 */
+	mov  w3, #GICR_ICENABLER0_SGI15
+	str  w3, [x5, #GICR_ICENABLER0_OFFSET]
+2:
+	/* poll on rwp bit in GICR_CTLR */
+	ldr  w4, [x6, #GICR_CTLR_OFFSET]
+	tst  w4, #GICR_CTLR_RWP
+	b.ne 2b
+
+	/* disable GRP1 interrupts at cpu interface */
+	msr  ICC_IGRPEN1_EL3, xzr
+
+	/* disable GRP0 ints at cpu interface */
+	msr  ICC_IGRPEN0_EL1, xzr
+
+	/* program the redistributor - poll on GICR_CTLR.RWP as needed */
+
+	/* define SGI 15 as Grp0 - GICR_IGROUPR0 */
+	ldr  w4, [x5, #GICR_IGROUPR0_OFFSET]
+	bic  w4, w4, #GICR_IGROUPR0_SGI15
+	str  w4, [x5, #GICR_IGROUPR0_OFFSET]
+
+	/* define SGI 15 as Grp0 - GICR_IGRPMODR0 */
+	ldr  w3, [x5, #GICR_IGRPMODR0_OFFSET]
+	bic  w3, w3, #GICR_IGRPMODR0_SGI15
+	str  w3, [x5, #GICR_IGRPMODR0_OFFSET]
+
+	/* set priority of SGI 15 to highest (0x0) - GICR_IPRIORITYR3 */
+	ldr  w4, [x5, #GICR_IPRIORITYR3_OFFSET]
+	bic  w4, w4, #GICR_IPRIORITYR3_SGI15_MASK
+	str  w4, [x5, #GICR_IPRIORITYR3_OFFSET]
+
+	/* enable SGI 15 at redistributor - GICR_ISENABLER0 */
+	mov  w3, #GICR_ISENABLER0_SGI15
+	str  w3, [x5, #GICR_ISENABLER0_OFFSET]
+	dsb  sy
+	isb
+3:
+	/* poll on rwp bit in GICR_CTLR */
+	ldr  w4, [x6, #GICR_CTLR_OFFSET]
+	tst  w4, #GICR_CTLR_RWP
+	b.ne 3b
+
+	/* quiesce the debug interfaces */
+	mrs  x3, osdlr_el1
+	orr  x3, x3, #OSDLR_EL1_DLK_LOCK
+	msr  osdlr_el1, x3
+	isb
+
+	/* enable grp0 ints */
+	mov  x3, #ICC_IGRPEN0_EL1_EN
+	msr  ICC_IGRPEN0_EL1, x3
+
+	/* x5 = gicr sgi base addr
+	 * x6 = gicr rd  base addr
+	 * x7 = core mask lsb
+	 */
+
+	/* clear any pending interrupts */
+	mvn  w1, wzr
+	str  w1, [x5, #GICR_ICPENDR0_OFFSET]
+
+	/* make sure system counter is enabled */
+	ldr  x3, =NXP_TIMER_ADDR
+	ldr  w0, [x3, #SYS_COUNTER_CNTCR_OFFSET]
+	tst  w0, #SYS_COUNTER_CNTCR_EN
+	b.ne 4f
+	orr  w0, w0, #SYS_COUNTER_CNTCR_EN
+	str  w0, [x3, #SYS_COUNTER_CNTCR_OFFSET]
+4:
+	/* enable the core timer and mask timer interrupt */
+	mov  x1, #CNTP_CTL_EL0_EN
+	orr  x1, x1, #CNTP_CTL_EL0_IMASK
+	msr  cntp_ctl_el0, x1
+
+	isb
+	mov  x30, x8
+	ret
+endfunc _soc_core_prep_off
+
+
+/* Part of CPU_OFF:
+ * Function performs the final steps to shutdown the core
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0, x1, x2, x3, x4, x5
+ */
+func _soc_core_entr_off
+	mov  x5, x30
+	mov  x4, x0
+
+1:
+	/* enter low-power state by executing wfi */
+	wfi
+
+	/* see if SGI15 woke us up */
+	mrs  x2, ICC_IAR0_EL1
+	mov  x3, #ICC_IAR0_EL1_SGI15
+	cmp  x2, x3
+	b.ne 2f
+
+	/* deactivate the intrrupts. */
+	msr ICC_EOIR0_EL1, x2
+
+2:
+	/* check if core is turned ON */
+	mov  x0, x4
+	/* Fetched the core state in x0 */
+	bl   _getCoreState
+
+	cmp  x0, #CORE_WAKEUP
+	b.ne 1b
+
+	/* Reached here, exited the wfi */
+
+	mov  x30, x5
+	ret
+endfunc _soc_core_entr_off
+
+
+/* Part of CPU_OFF:
+ * Function starts the process of starting a core back up
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0, x1, x2, x3, x4, x5, x6
+ */
+func _soc_core_exit_off
+	mov  x6, x30
+	mov  x5, x0
+
+	/* disable forwarding of GRP0 ints at cpu interface */
+	msr  ICC_IGRPEN0_EL1, xzr
+
+	/* get redistributor sgi base addr for this core */
+	mov  x0, x5
+	bl   get_gic_sgi_base
+	mov  x4, x0
+
+	/* x4 = gicr sgi base addr
+	 * x5 = core mask
+	 */
+
+	/* disable SGI 15 at redistributor - GICR_ICENABLER0 */
+	mov  w1, #GICR_ICENABLER0_SGI15
+	str  w1, [x4, #GICR_ICENABLER0_OFFSET]
+
+	/* get redistributor rd base addr for this core */
+	mov  x0, x5
+	bl   get_gic_rd_base
+	mov  x4, x0
+
+2:
+	/* poll on rwp bit in GICR_CTLR */
+	ldr  w2, [x4, #GICR_CTLR_OFFSET]
+	tst  w2, #GICR_CTLR_RWP
+	b.ne 2b
+
+	/* unlock the debug interfaces */
+	mrs  x3, osdlr_el1
+	bic  x3, x3, #OSDLR_EL1_DLK_LOCK
+	msr  osdlr_el1, x3
+	isb
+
+	dsb sy
+	isb
+	mov  x30, x6
+	ret
+endfunc _soc_core_exit_off
+
+
+/* Function requests a reset of the entire SOC
+ * in:  none
+ * out: none
+ * uses: x0, x1, x2, x3, x4, x5, x6
+ */
+func _soc_sys_reset
+	mov  x6, x30
+
+	ldr  x2, =NXP_RST_ADDR
+	/* clear the RST_REQ_MSK and SW_RST_REQ */
+
+	mov  w0, #0x00000000
+	str  w0, [x2, #RSTCNTL_OFFSET]
+
+	/* initiate the sw reset request */
+	mov  w0, #SW_RST_REQ_INIT
+	str  w0, [x2, #RSTCNTL_OFFSET]
+
+	/* In case this address range is mapped as cacheable,
+	 * flush the write out of the dcaches.
+	 */
+	add  x2, x2, #RSTCNTL_OFFSET
+	dc   cvac, x2
+	dsb  st
+	isb
+
+	/* Function does not return */
+	b  .
+endfunc _soc_sys_reset
+
+
+/* Part of SYSTEM_OFF:
+ * Function turns off the SoC clocks
+ * Note: Function is not intended to return, and the only allowable
+ *	   recovery is POR
+ * in:  none
+ * out: none
+ * uses x0, x1, x2, x3
+ */
+func _soc_sys_off
+
+	/* A-009810: LPM20 entry sequence might cause
+	 * spurious timeout reset request
+	 * workaround: MASK RESET REQ RPTOE
+	 */
+	ldr  x0, =NXP_RESET_ADDR
+	ldr  w1, =RSTRQMR_RPTOE_MASK
+	str  w1, [x0, #RST_RSTRQMR1_OFFSET]
+
+	/* disable sec, QBman, spi and qspi */
+	ldr  x2, =NXP_DCFG_ADDR
+	ldr  x0, =DCFG_DEVDISR1_OFFSET
+	ldr  w1, =DCFG_DEVDISR1_SEC
+	str  w1, [x2, x0]
+	ldr  x0, =DCFG_DEVDISR3_OFFSET
+	ldr  w1, =DCFG_DEVDISR3_QBMAIN
+	str  w1, [x2, x0]
+	ldr  x0, =DCFG_DEVDISR4_OFFSET
+	ldr  w1, =DCFG_DEVDISR4_SPI_QSPI
+	str  w1, [x2, x0]
+
+	/* set TPMWAKEMR0 */
+	ldr  x0, =TPMWAKEMR0_ADDR
+	mov  w1, #0x1
+	str  w1, [x0]
+
+	/* disable icache, dcache, mmu @ EL1 */
+	mov  x1, #SCTLR_I_C_M_MASK
+	mrs  x0, sctlr_el1
+	bic  x0, x0, x1
+	msr  sctlr_el1, x0
+
+	/* disable L2 prefetches */
+	mrs  x0, CORTEX_A72_ECTLR_EL1
+	bic  x1, x1, #CPUECTLR_TIMER_MASK
+	orr  x0, x0, #CPUECTLR_SMPEN_EN
+	orr  x0, x0, #CPUECTLR_TIMER_8TICKS
+	msr  CORTEX_A72_ECTLR_EL1, x0
+	isb
+
+	/* disable CCN snoop domain */
+	mov  x1, #NXP_CCN_HN_F_0_ADDR
+	ldr  x0, =CCN_HN_F_SNP_DMN_CTL_MASK
+	str  x0, [x1, #CCN_HN_F_SNP_DMN_CTL_CLR_OFFSET]
+3:
+	ldr  w2, [x1, #CCN_HN_F_SNP_DMN_CTL_OFFSET]
+	cmp  w2, #0x2
+	b.ne 3b
+
+	mov  x3, #NXP_PMU_ADDR
+
+4:
+	ldr  w1, [x3, #PMU_PCPW20SR_OFFSET]
+	cmp  w1, #PMU_IDLE_CORE_MASK
+	b.ne 4b
+
+	mov  w1, #PMU_IDLE_CLUSTER_MASK
+	str  w1, [x3, #PMU_CLAINACTSETR_OFFSET]
+
+1:
+	ldr  w1, [x3, #PMU_PCPW20SR_OFFSET]
+	cmp  w1, #PMU_IDLE_CORE_MASK
+	b.ne 1b
+
+	mov  w1, #PMU_FLUSH_CLUSTER_MASK
+	str  w1, [x3, #PMU_CLL2FLUSHSETR_OFFSET]
+
+2:
+	ldr  w1, [x3, #PMU_CLL2FLUSHSR_OFFSET]
+	cmp  w1, #PMU_FLUSH_CLUSTER_MASK
+	b.ne 2b
+
+	mov  w1, #PMU_FLUSH_CLUSTER_MASK
+	str  w1, [x3, #PMU_CLSL2FLUSHCLRR_OFFSET]
+
+	mov  w1, #PMU_FLUSH_CLUSTER_MASK
+	str  w1, [x3, #PMU_CLSINACTSETR_OFFSET]
+
+	mov  x2, #DAIF_SET_MASK
+	mrs  x1, spsr_el1
+	orr  x1, x1, x2
+	msr  spsr_el1, x1
+
+	mrs  x1, spsr_el2
+	orr  x1, x1, x2
+	msr  spsr_el2, x1
+
+	/* force the debug interface to be quiescent */
+	mrs  x0, osdlr_el1
+	orr  x0, x0, #0x1
+	msr  osdlr_el1, x0
+
+	/* invalidate all TLB entries at all 3 exception levels */
+	tlbi alle1
+	tlbi alle2
+	tlbi alle3
+
+	/* x3 = pmu base addr */
+
+	/* request lpm20 */
+	ldr  x0, =PMU_POWMGTCSR_OFFSET
+	ldr  w1, =PMU_POWMGTCSR_VAL
+	str  w1, [x3, x0]
+
+5:
+	wfe
+	b.eq  5b
+endfunc _soc_sys_off
+
+
+/* Part of CPU_SUSPEND
+ * Function puts the calling core into standby state
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0
+ */
+func _soc_core_entr_stdby
+
+	dsb  sy
+	isb
+	wfi
+
+	ret
+endfunc _soc_core_entr_stdby
+
+
+/* Part of CPU_SUSPEND
+ * Function performs SoC-specific programming prior to standby
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0, x1
+ */
+func _soc_core_prep_stdby
+
+	/* clear CORTEX_A72_ECTLR_EL1[2:0] */
+	mrs  x1, CORTEX_A72_ECTLR_EL1
+	bic  x1, x1, #CPUECTLR_TIMER_MASK
+	msr  CORTEX_A72_ECTLR_EL1, x1
+
+	ret
+endfunc _soc_core_prep_stdby
+
+
+/* Part of CPU_SUSPEND
+ * Function performs any SoC-specific cleanup after standby state
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses none
+ */
+func _soc_core_exit_stdby
+
+	ret
+endfunc _soc_core_exit_stdby
+
+
+/* Part of CPU_SUSPEND
+ * Function performs SoC-specific programming prior to power-down
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses none
+ */
+func _soc_core_prep_pwrdn
+
+	/* make sure system counter is enabled */
+	ldr  x2, =NXP_TIMER_ADDR
+	ldr  w0, [x2, #SYS_COUNTER_CNTCR_OFFSET]
+	tst  w0, #SYS_COUNTER_CNTCR_EN
+	b.ne 1f
+	orr  w0, w0, #SYS_COUNTER_CNTCR_EN
+	str  w0, [x2, #SYS_COUNTER_CNTCR_OFFSET]
+1:
+
+	/* enable dynamic retention control (CPUECTLR[2:0])
+	 * set the SMPEN bit (CPUECTLR[6])
+	 */
+	mrs  x1, CORTEX_A72_ECTLR_EL1
+	bic  x1, x1, #CPUECTLR_RET_MASK
+	orr  x1, x1, #CPUECTLR_TIMER_8TICKS
+	orr  x1, x1, #CPUECTLR_SMPEN_EN
+	msr  CORTEX_A72_ECTLR_EL1, x1
+
+	isb
+	ret
+endfunc _soc_core_prep_pwrdn
+
+
+/* Part of CPU_SUSPEND
+ * Function puts the calling core into a power-down state
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0
+ */
+func _soc_core_entr_pwrdn
+
+	/* X0 = core mask lsb */
+
+	dsb  sy
+	isb
+	wfi
+
+	ret
+endfunc _soc_core_entr_pwrdn
+
+
+/* Part of CPU_SUSPEND
+ * Function performs any SoC-specific cleanup after power-down state
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses none
+ */
+func _soc_core_exit_pwrdn
+
+	ret
+endfunc _soc_core_exit_pwrdn
+
+
+/* Part of CPU_SUSPEND
+ * Function performs SoC-specific programming prior to standby
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0, x1
+ */
+func _soc_clstr_prep_stdby
+
+	/* clear CORTEX_A72_ECTLR_EL1[2:0] */
+	mrs  x1, CORTEX_A72_ECTLR_EL1
+	bic  x1, x1, #CPUECTLR_TIMER_MASK
+	msr  CORTEX_A72_ECTLR_EL1, x1
+
+	ret
+endfunc _soc_clstr_prep_stdby
+
+
+/* Part of CPU_SUSPEND
+ * Function performs any SoC-specific cleanup after standby state
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses none
+ */
+func _soc_clstr_exit_stdby
+
+	ret
+endfunc _soc_clstr_exit_stdby
+
+
+/* Part of CPU_SUSPEND
+ * Function performs SoC-specific programming prior to power-down
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses none
+ */
+func _soc_clstr_prep_pwrdn
+
+	/* make sure system counter is enabled */
+	ldr  x2, =NXP_TIMER_ADDR
+	ldr  w0, [x2, #SYS_COUNTER_CNTCR_OFFSET]
+	tst  w0, #SYS_COUNTER_CNTCR_EN
+	b.ne 1f
+	orr  w0, w0, #SYS_COUNTER_CNTCR_EN
+	str  w0, [x2, #SYS_COUNTER_CNTCR_OFFSET]
+1:
+
+	/* enable dynamic retention control (CPUECTLR[2:0])
+	 * set the SMPEN bit (CPUECTLR[6])
+	 */
+	mrs  x1, CORTEX_A72_ECTLR_EL1
+	bic  x1, x1, #CPUECTLR_RET_MASK
+	orr  x1, x1, #CPUECTLR_TIMER_8TICKS
+	orr  x1, x1, #CPUECTLR_SMPEN_EN
+	msr  CORTEX_A72_ECTLR_EL1, x1
+
+	isb
+	ret
+endfunc _soc_clstr_prep_pwrdn
+
+
+/* Part of CPU_SUSPEND
+ * Function performs any SoC-specific cleanup after power-down state
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses none
+ */
+func _soc_clstr_exit_pwrdn
+
+	ret
+endfunc _soc_clstr_exit_pwrdn
+
+
+/* Part of CPU_SUSPEND
+ * Function performs SoC-specific programming prior to standby
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0, x1
+ */
+func _soc_sys_prep_stdby
+
+	/* clear CORTEX_A72_ECTLR_EL1[2:0] */
+	mrs  x1, CORTEX_A72_ECTLR_EL1
+	bic  x1, x1, #CPUECTLR_TIMER_MASK
+	msr  CORTEX_A72_ECTLR_EL1, x1
+	ret
+endfunc _soc_sys_prep_stdby
+
+
+/* Part of CPU_SUSPEND
+ * Function performs any SoC-specific cleanup after standby state
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses none
+ */
+func _soc_sys_exit_stdby
+
+	ret
+endfunc _soc_sys_exit_stdby
+
+
+/* Part of CPU_SUSPEND
+ * Function performs SoC-specific programming prior to
+ * suspend-to-power-down
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0, x1
+ */
+func _soc_sys_prep_pwrdn
+
+	mrs   x1, CORTEX_A72_ECTLR_EL1
+	/* make sure the smp bit is set */
+	orr   x1, x1, #CPUECTLR_SMPEN_MASK
+	/* set the retention control */
+	orr   x1, x1, #CPUECTLR_RET_8CLK
+	/* disable tablewalk prefetch */
+	orr   x1, x1, #CPUECTLR_DISABLE_TWALK_PREFETCH
+	msr   CORTEX_A72_ECTLR_EL1, x1
+	isb
+
+	ret
+endfunc _soc_sys_prep_pwrdn
+
+
+/* Part of CPU_SUSPEND
+ * Function puts the calling core, and potentially the soc, into a
+ * low-power state
+ * in:  x0 = core mask lsb
+ * out: x0 = 0, success
+ *	  x0 < 0, failure
+ * uses x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x10, x11, x12, x13, x14,
+ *	  x15, x16, x17, x18, x19, x20, x21, x28
+ */
+func _soc_sys_pwrdn_wfi
+	mov  x28, x30
+
+	/* disable cluster snooping in the CCN-508 */
+	ldr  x1, =NXP_CCN_HN_F_0_ADDR
+	ldr  x7, [x1, #CCN_HN_F_SNP_DMN_CTL_OFFSET]
+	mov  x6, #CCN_HNF_NODE_COUNT
+1:
+	str  x7, [x1, #CCN_HN_F_SNP_DMN_CTL_CLR_OFFSET]
+	sub  x6, x6, #1
+	add  x1, x1, #CCN_HNF_OFFSET
+	cbnz x6, 1b
+
+	/* x0  = core mask
+	 * x7  = hnf sdcr
+	 */
+
+	ldr  x1, =NXP_PMU_CCSR_ADDR
+	ldr  x2, =NXP_PMU_DCSR_ADDR
+
+	/* enable the stop-request-override */
+	mov  x3, #PMU_POWMGTDCR0_OFFSET
+	mov  x4, #POWMGTDCR_STP_OV_EN
+	str  w4, [x2, x3]
+
+	/* x0  = core mask
+	 * x1  = NXP_PMU_CCSR_ADDR
+	 * x2  = NXP_PMU_DCSR_ADDR
+	 * x7  = hnf sdcr
+	 */
+
+	/* disable prefetching in the A72 core */
+	mrs  x8, CORTEX_A72_CPUACTLR_EL1
+	tst  x8, #CPUACTLR_DIS_LS_HW_PRE
+	b.ne 2f
+	dsb  sy
+	isb
+	/* disable data prefetch */
+	orr  x16, x8, #CPUACTLR_DIS_LS_HW_PRE
+	/* disable tlb prefetch */
+	orr  x16, x16, #CPUACTLR_DIS_L2_TLB_PRE
+	msr  CORTEX_A72_CPUACTLR_EL1, x16
+	isb
+
+	/* x0  = core mask
+	 * x1  = NXP_PMU_CCSR_ADDR
+	 * x2  = NXP_PMU_DCSR_ADDR
+	 * x7  = hnf sdcr
+	 * x8  = cpuactlr
+	 */
+
+2:
+	/* save hnf-sdcr and cpuactlr to stack */
+	stp  x7,  x8,  [sp, #-16]!
+
+	/* x0  = core mask
+	 * x1  = NXP_PMU_CCSR_ADDR
+	 * x2  = NXP_PMU_DCSR_ADDR
+	 */
+
+	/* save the IPSTPCRn registers to stack */
+	mov  x15, #PMU_IPSTPCR0_OFFSET
+	ldr  w9,  [x1, x15]
+	mov  x16, #PMU_IPSTPCR1_OFFSET
+	ldr  w10, [x1, x16]
+	mov  x17, #PMU_IPSTPCR2_OFFSET
+	ldr  w11, [x1, x17]
+	mov  x18, #PMU_IPSTPCR3_OFFSET
+	ldr  w12, [x1, x18]
+	mov  x19, #PMU_IPSTPCR4_OFFSET
+	ldr  w13, [x1, x19]
+	mov  x20, #PMU_IPSTPCR5_OFFSET
+	ldr  w14, [x1, x20]
+
+	stp  x9,  x10,  [sp, #-16]!
+	stp  x11, x12,  [sp, #-16]!
+	stp  x13, x14,  [sp, #-16]!
+
+	/* x0  = core mask
+	 * x1  = NXP_PMU_CCSR_ADDR
+	 * x2  = NXP_PMU_DCSR_ADDR
+	 * x15 = PMU_IPSTPCR0_OFFSET
+	 * x16 = PMU_IPSTPCR1_OFFSET
+	 * x17 = PMU_IPSTPCR2_OFFSET
+	 * x18 = PMU_IPSTPCR3_OFFSET
+	 * x19 = PMU_IPSTPCR4_OFFSET
+	 * x20 = PMU_IPSTPCR5_OFFSET
+	 */
+
+	/* load the full clock mask for IPSTPCR0 */
+	ldr  x3, =DEVDISR1_MASK
+	/* get the exclusions */
+	mov  x21, #PMU_IPPDEXPCR0_OFFSET
+	ldr  w4, [x1, x21]
+	/* apply the exclusions to the mask */
+	bic  w7, w3, w4
+	/* stop the clocks in IPSTPCR0 */
+	str  w7, [x1, x15]
+
+	/* use same procedure for IPSTPCR1-IPSTPCR5 */
+
+	/* stop the clocks in IPSTPCR1 */
+	ldr  x5, =DEVDISR2_MASK
+	mov  x21, #PMU_IPPDEXPCR1_OFFSET
+	ldr  w6, [x1, x21]
+	bic  w8, w5, w6
+	str  w8, [x1, x16]
+
+	/* stop the clocks in IPSTPCR2 */
+	ldr  x3, =DEVDISR3_MASK
+	mov  x21, #PMU_IPPDEXPCR2_OFFSET
+	ldr  w4, [x1, x21]
+	bic  w9, w3, w4
+	str  w9, [x1, x17]
+
+	/* stop the clocks in IPSTPCR3 */
+	ldr  x5,  =DEVDISR4_MASK
+	mov  x21, #PMU_IPPDEXPCR3_OFFSET
+	ldr  w6,  [x1, x21]
+	bic  w10, w5, w6
+	str  w10, [x1, x18]
+
+	/* stop the clocks in IPSTPCR4
+	 *   - exclude the ddr clocks as we are currently executing
+	 *	 out of *some* memory, might be ddr
+	 *   - exclude the OCRAM clk so that we retain any code/data in
+	 *	 OCRAM
+	 *   - may need to exclude the debug clock if we are testing
+	 */
+	ldr  x3, =DEVDISR5_MASK
+	mov  w6, #DEVDISR5_MASK_ALL_MEM
+	bic  w3, w3, w6
+
+	mov  w5, #POLICY_DEBUG_ENABLE
+	cbz  w5, 3f
+	mov  w6, #DEVDISR5_MASK_DBG
+	bic  w3, w3, w6
+3:
+	mov  x21, #PMU_IPPDEXPCR4_OFFSET
+	ldr  w4,  [x1, x21]
+	bic  w11, w3, w4
+	str  w11, [x1, x19]
+
+	/* stop the clocks in IPSTPCR5 */
+	ldr  x5,  =DEVDISR6_MASK
+	mov  x21, #PMU_IPPDEXPCR5_OFFSET
+	ldr  w6,  [x1, x21]
+	bic  w12, w5, w6
+	str  w12, [x1, x20]
+
+	/* x0  = core mask
+	 * x1  = NXP_PMU_CCSR_ADDR
+	 * x2  = NXP_PMU_DCSR_ADDR
+	 * x7  = IPSTPCR0
+	 * x8  = IPSTPCR1
+	 * x9  = IPSTPCR2
+	 * x10 = IPSTPCR3
+	 * x11 = IPSTPCR4
+	 * x12 = IPSTPCR5
+	 */
+
+	/* poll until the clocks are stopped in IPSTPACKSR0 */
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x21, #PMU_IPSTPACKSR0_OFFSET
+4:
+	ldr  w5, [x1, x21]
+	cmp  w5, w7
+	b.eq 5f
+	sub  w4, w4, #1
+	cbnz w4, 4b
+
+	/* poll until the clocks are stopped in IPSTPACKSR1 */
+5:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x21, #PMU_IPSTPACKSR1_OFFSET
+6:
+	ldr  w5, [x1, x21]
+	cmp  w5, w8
+	b.eq 7f
+	sub  w4, w4, #1
+	cbnz w4, 6b
+
+	/* poll until the clocks are stopped in IPSTPACKSR2 */
+7:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x21, #PMU_IPSTPACKSR2_OFFSET
+8:
+	ldr  w5, [x1, x21]
+	cmp  w5, w9
+	b.eq 9f
+	sub  w4, w4, #1
+	cbnz w4, 8b
+
+	/* poll until the clocks are stopped in IPSTPACKSR3 */
+9:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x21, #PMU_IPSTPACKSR3_OFFSET
+10:
+	ldr  w5, [x1, x21]
+	cmp  w5, w10
+	b.eq 11f
+	sub  w4, w4, #1
+	cbnz w4, 10b
+
+	/* poll until the clocks are stopped in IPSTPACKSR4 */
+11:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x21, #PMU_IPSTPACKSR4_OFFSET
+12:
+	ldr  w5, [x1, x21]
+	cmp  w5, w11
+	b.eq 13f
+	sub  w4, w4, #1
+	cbnz w4, 12b
+
+	/* poll until the clocks are stopped in IPSTPACKSR5 */
+13:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x21, #PMU_IPSTPACKSR5_OFFSET
+14:
+	ldr  w5, [x1, x21]
+	cmp  w5, w12
+	b.eq 15f
+	sub  w4, w4, #1
+	cbnz w4, 14b
+
+	/* x0  = core mask
+	 * x1  = NXP_PMU_CCSR_ADDR
+	 * x2  = NXP_PMU_DCSR_ADDR
+	 * x7  = IPSTPCR0
+	 * x8  = IPSTPCR1
+	 * x9  = IPSTPCR2
+	 * x10 = IPSTPCR3
+	 * x11 = IPSTPCR4
+	 * x12 = IPSTPCR5
+	 */
+
+15:
+	mov  x3, #NXP_DCFG_ADDR
+
+	/* save the devdisr registers to stack */
+	ldr  w13, [x3, #DCFG_DEVDISR1_OFFSET]
+	ldr  w14, [x3, #DCFG_DEVDISR2_OFFSET]
+	ldr  w15, [x3, #DCFG_DEVDISR3_OFFSET]
+	ldr  w16, [x3, #DCFG_DEVDISR4_OFFSET]
+	ldr  w17, [x3, #DCFG_DEVDISR5_OFFSET]
+	ldr  w18, [x3, #DCFG_DEVDISR6_OFFSET]
+
+	stp  x13, x14,  [sp, #-16]!
+	stp  x15, x16,  [sp, #-16]!
+	stp  x17, x18,  [sp, #-16]!
+
+	/* power down the IP in DEVDISR1 - corresponds to IPSTPCR0 */
+	str  w7,  [x3, #DCFG_DEVDISR1_OFFSET]
+
+	/* power down the IP in DEVDISR2 - corresponds to IPSTPCR1 */
+	str  w8, [x3, #DCFG_DEVDISR2_OFFSET]
+
+	/* power down the IP in DEVDISR3 - corresponds to IPSTPCR2 */
+	str  w9,  [x3, #DCFG_DEVDISR3_OFFSET]
+
+	/* power down the IP in DEVDISR4 - corresponds to IPSTPCR3 */
+	str  w10, [x3, #DCFG_DEVDISR4_OFFSET]
+
+	/* power down the IP in DEVDISR5 - corresponds to IPSTPCR4 */
+	str  w11, [x3, #DCFG_DEVDISR5_OFFSET]
+
+	/* power down the IP in DEVDISR6 - corresponds to IPSTPCR5 */
+	str  w12, [x3, #DCFG_DEVDISR6_OFFSET]
+
+	/* setup register values for the cache-only sequence */
+	mov  x4, #NXP_DDR_ADDR
+	mov  x5, #NXP_DDR2_ADDR
+	mov  x6, x11
+	mov  x7, x17
+	ldr  x12, =PMU_CLAINACTSETR_OFFSET
+	ldr  x13, =PMU_CLSINACTSETR_OFFSET
+	ldr  x14, =PMU_CLAINACTCLRR_OFFSET
+	ldr  x15, =PMU_CLSINACTCLRR_OFFSET
+
+	/* x0  = core mask
+	 * x1  = NXP_PMU_CCSR_ADDR
+	 * x2  = NXP_PMU_DCSR_ADDR
+	 * x3  = NXP_DCFG_ADDR
+	 * x4  = NXP_DDR_ADDR
+	 * x5  = NXP_DDR2_ADDR
+	 * w6  = IPSTPCR4
+	 * w7  = DEVDISR5
+	 * x12 = PMU_CLAINACTSETR_OFFSET
+	 * x13 = PMU_CLSINACTSETR_OFFSET
+	 * x14 = PMU_CLAINACTCLRR_OFFSET
+	 * x15 = PMU_CLSINACTCLRR_OFFSET
+	 */
+
+	mov  x8, #POLICY_DEBUG_ENABLE
+	cbnz x8, 29f
+	/* force the debug interface to be quiescent */
+	mrs  x9, OSDLR_EL1
+	orr  x9, x9, #0x1
+	msr  OSDLR_EL1, x9
+
+	/* enter the cache-only sequence */
+29:
+	bl   final_pwrdown
+
+	/* when we are here, the core has come out of wfi and the
+	 * ddr is back up
+	 */
+
+	mov  x8, #POLICY_DEBUG_ENABLE
+	cbnz x8, 30f
+	/* restart the debug interface */
+	mrs  x9, OSDLR_EL1
+	mov  x10, #1
+	bic  x9, x9, x10
+	msr  OSDLR_EL1, x9
+
+	/* get saved DEVDISR regs off stack */
+30:
+	ldp  x17, x18, [sp], #16
+	ldp  x15, x16, [sp], #16
+	ldp  x13, x14, [sp], #16
+	/* restore DEVDISR regs */
+	str  w18, [x3, #DCFG_DEVDISR6_OFFSET]
+	str  w17, [x3, #DCFG_DEVDISR5_OFFSET]
+	str  w16, [x3, #DCFG_DEVDISR4_OFFSET]
+	str  w15, [x3, #DCFG_DEVDISR3_OFFSET]
+	str  w14, [x3, #DCFG_DEVDISR2_OFFSET]
+	str  w13, [x3, #DCFG_DEVDISR1_OFFSET]
+	isb
+
+	/* get saved IPSTPCRn regs off stack */
+	ldp  x13, x14, [sp], #16
+	ldp  x11, x12, [sp], #16
+	ldp  x9,  x10, [sp], #16
+
+	/* restore IPSTPCRn regs */
+	mov  x15, #PMU_IPSTPCR5_OFFSET
+	str  w14, [x1, x15]
+	mov  x16, #PMU_IPSTPCR4_OFFSET
+	str  w13, [x1, x16]
+	mov  x17, #PMU_IPSTPCR3_OFFSET
+	str  w12, [x1, x17]
+	mov  x18, #PMU_IPSTPCR2_OFFSET
+	str  w11, [x1, x18]
+	mov  x19, #PMU_IPSTPCR1_OFFSET
+	str  w10, [x1, x19]
+	mov  x20, #PMU_IPSTPCR0_OFFSET
+	str  w9,  [x1, x20]
+	isb
+
+	/* poll on IPSTPACKCRn regs til IP clocks are restarted */
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x15, #PMU_IPSTPACKSR5_OFFSET
+16:
+	ldr  w5, [x1, x15]
+	and  w5, w5, w14
+	cbz  w5, 17f
+	sub  w4, w4, #1
+	cbnz w4, 16b
+
+17:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x15, #PMU_IPSTPACKSR4_OFFSET
+18:
+	ldr  w5, [x1, x15]
+	and  w5, w5, w13
+	cbz  w5, 19f
+	sub  w4, w4, #1
+	cbnz w4, 18b
+
+19:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x15, #PMU_IPSTPACKSR3_OFFSET
+20:
+	ldr  w5, [x1, x15]
+	and  w5, w5, w12
+	cbz  w5, 21f
+	sub  w4, w4, #1
+	cbnz w4, 20b
+
+21:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x15, #PMU_IPSTPACKSR2_OFFSET
+22:
+	ldr  w5, [x1, x15]
+	and  w5, w5, w11
+	cbz  w5, 23f
+	sub  w4, w4, #1
+	cbnz w4, 22b
+
+23:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x15, #PMU_IPSTPACKSR1_OFFSET
+24:
+	ldr  w5, [x1, x15]
+	and  w5, w5, w10
+	cbz  w5, 25f
+	sub  w4, w4, #1
+	cbnz w4, 24b
+
+25:
+	mov  w4,  #CLOCK_RETRY_CNT
+	mov  x15, #PMU_IPSTPACKSR0_OFFSET
+26:
+	ldr  w5, [x1, x15]
+	and  w5, w5, w9
+	cbz  w5, 27f
+	sub  w4, w4, #1
+	cbnz w4, 26b
+
+27:
+	/* disable the stop-request-override */
+	mov  x8, #PMU_POWMGTDCR0_OFFSET
+	mov  w9, #POWMGTDCR_STP_OV_EN
+	str  w9, [x2, x8]
+	isb
+
+	/* get hnf-sdcr and cpuactlr off stack */
+	ldp  x7, x8, [sp], #16
+
+	/* restore cpuactlr */
+	msr  CORTEX_A72_CPUACTLR_EL1, x8
+	isb
+
+	/* restore snooping in the hnf nodes */
+	ldr  x9, =NXP_CCN_HN_F_0_ADDR
+	mov  x6, #CCN_HNF_NODE_COUNT
+28:
+	str  x7, [x9, #CCN_HN_F_SNP_DMN_CTL_SET_OFFSET]
+	sub  x6, x6, #1
+	add  x9, x9, #CCN_HNF_OFFSET
+	cbnz x6, 28b
+	isb
+
+	mov  x30, x28
+	ret
+endfunc _soc_sys_pwrdn_wfi
+
+
+/* Part of CPU_SUSPEND
+ * Function performs any SoC-specific cleanup after power-down
+ * in:  x0 = core mask lsb
+ * out: none
+ * uses x0,
+ */
+func _soc_sys_exit_pwrdn
+
+	mrs   x1, CORTEX_A72_ECTLR_EL1
+	/* make sure the smp bit is set */
+	orr   x1, x1, #CPUECTLR_SMPEN_MASK
+	/* clr the retention control */
+	mov   x2, #CPUECTLR_RET_8CLK
+	bic   x1, x1, x2
+	/* enable tablewalk prefetch */
+	mov   x2, #CPUECTLR_DISABLE_TWALK_PREFETCH
+	bic   x1, x1, x2
+	msr   CORTEX_A72_ECTLR_EL1, x1
+	isb
+
+	ret
+endfunc _soc_sys_exit_pwrdn
+
+
+/* Function will pwrdown ddr and the final core - it will do this
+ * by loading itself into the icache and then executing from there
+ * in:
+ *   x0  = core mask
+ *   x1  = NXP_PMU_CCSR_ADDR
+ *   x2  = NXP_PMU_DCSR_ADDR
+ *   x3  = NXP_DCFG_ADDR
+ *   x4  = NXP_DDR_ADDR
+ *   x5  = NXP_DDR2_ADDR
+ *   w6  = IPSTPCR4
+ *   w7  = DEVDISR5
+ *   x12 = PMU_CLAINACTSETR_OFFSET
+ *   x13 = PMU_CLSINACTSETR_OFFSET
+ *   x14 = PMU_CLAINACTCLRR_OFFSET
+ *   x15 = PMU_CLSINACTCLRR_OFFSET
+ * out: none
+ * uses x0, x1, x2, x3, x4, x5, x6, x7, x8, x9, x13, x14, x15, x16,
+ *	  x17, x18
+ */
+
+/* 4Kb aligned */
+.align 12
+func final_pwrdown
+
+	mov  x0, xzr
+	b	touch_line_0
+start_line_0:
+	mov  x0, #1
+	/* put ddr controller 1 into self-refresh */
+	ldr  w8, [x4, #DDR_CFG_2_OFFSET]
+	orr  w8, w8, #CFG_2_FORCE_REFRESH
+	str  w8, [x4, #DDR_CFG_2_OFFSET]
+
+	/* put ddr controller 2 into self-refresh */
+	ldr  w8, [x5, #DDR_CFG_2_OFFSET]
+	orr  w8, w8, #CFG_2_FORCE_REFRESH
+	str  w8, [x5, #DDR_CFG_2_OFFSET]
+
+	/* stop the clocks in both ddr controllers */
+	mov  w10, #DEVDISR5_MASK_DDR
+	mov  x16, #PMU_IPSTPCR4_OFFSET
+	orr  w9,  w6, w10
+	str  w9,  [x1, x16]
+	isb
+
+	mov  x17, #PMU_IPSTPACKSR4_OFFSET
+touch_line_0:
+	cbz  x0, touch_line_1
+
+start_line_1:
+	/* poll IPSTPACKSR4 until
+	 * ddr controller clocks are stopped.
+	 */
+1:
+	ldr  w8, [x1, x17]
+	and  w8, w8, w10
+	cmp  w8, w10
+	b.ne 1b
+
+	/* shut down power to the ddr controllers */
+	orr w9, w7, #DEVDISR5_MASK_DDR
+	str w9, [x3, #DCFG_DEVDISR5_OFFSET]
+
+	/* disable cluster acp ports */
+	mov  w8, #CLAINACT_DISABLE_ACP
+	str  w8, [x1, x12]
+
+	/* disable skyros ports */
+	mov  w9, #CLSINACT_DISABLE_SKY
+	str  w9, [x1, x13]
+	isb
+
+touch_line_1:
+	cbz  x0, touch_line_2
+
+start_line_2:
+	isb
+3:
+	wfi
+
+	/* if we are here then we are awake
+	 * - bring this device back up
+	 */
+
+	/* enable skyros ports */
+	mov  w9, #CLSINACT_DISABLE_SKY
+	str  w9, [x1, x15]
+
+	/* enable acp ports */
+	mov  w8, #CLAINACT_DISABLE_ACP
+	str  w8, [x1, x14]
+	isb
+
+	/* bring up the ddr controllers */
+	str w7, [x3, #DCFG_DEVDISR5_OFFSET]
+	isb
+	str w6,  [x1, x16]
+	isb
+
+	nop
+touch_line_2:
+	cbz  x0, touch_line_3
+
+start_line_3:
+	/* poll IPSTPACKSR4 until
+	 * ddr controller clocks are running
+	 */
+	mov w10, #DEVDISR5_MASK_DDR
+2:
+	ldr  w8, [x1, x17]
+	and  w8, w8, w10
+	cbnz w8, 2b
+
+	/* take ddr controller 2 out of self-refresh */
+	mov w8, #CFG_2_FORCE_REFRESH
+	ldr w9, [x5, #DDR_CFG_2_OFFSET]
+	bic w9, w9, w8
+	str w9, [x5, #DDR_CFG_2_OFFSET]
+
+	/* take ddr controller 1 out of self-refresh */
+	ldr w9, [x4, #DDR_CFG_2_OFFSET]
+	bic w9, w9, w8
+	str w9, [x4, #DDR_CFG_2_OFFSET]
+	isb
+
+	nop
+	nop
+	nop
+touch_line_3:
+	cbz  x0, start_line_0
+
+	/* execute here after ddr is back up */
+
+	ret
+endfunc final_pwrdown
+
+/* Function returns CLUSTER_3_NORMAL if the cores of cluster 3 are
+ * to be handled normally, and it returns CLUSTER_3_IN_RESET if the cores
+ * are to be held in reset
+ * in:  none
+ * out: x0 = #CLUSTER_3_NORMAL,   cluster 3 treated normal
+ *	  x0 = #CLUSTER_3_IN_RESET, cluster 3 cores held in reset
+ * uses x0, x1, x2
+ */
+func cluster3InReset
+
+	/* default return is treat cores normal */
+	mov  x0, #CLUSTER_3_NORMAL
+
+	/* read RCW_SR27 register */
+	mov  x1, #NXP_DCFG_ADDR
+	ldr  w2, [x1, #RCW_SR27_OFFSET]
+
+	/* test the cluster 3 bit */
+	tst  w2, #CLUSTER_3_RCW_BIT
+	b.eq 1f
+
+	/* if we are here, then the bit was set */
+	mov  x0, #CLUSTER_3_IN_RESET
+1:
+	ret
+endfunc cluster3InReset
+
+
+/* Function checks to see if cores which are to be disabled have been
+ * released from reset - if not, it releases them
+ * Note: there may be special handling of cluster 3 cores depending upon the
+ *	   sys clk frequency
+ * in:  none
+ * out: none
+ * uses x0, x1, x2, x3, x4, x5, x6, x7, x8, x9
+ */
+func release_disabled
+	mov  x9, x30
+
+	/* check if we need to keep cluster 3 cores in reset */
+	bl   cluster3InReset		/*  0-2  */
+	mov  x8, x0
+
+	/* x8 = cluster 3 handling */
+
+	/* read COREDISABLESR */
+	mov  x0, #NXP_DCFG_ADDR
+	ldr  w4, [x0, #DCFG_COREDISABLEDSR_OFFSET]
+	cmp  x8, #CLUSTER_3_IN_RESET
+	b.ne 4f
+
+	/* the cluster 3 cores are to be held in reset, so remove
+	 * them from the disable mask
+	 */
+	bic  x4, x4, #CLUSTER_3_CORES_MASK
+4:
+	/* get the number of cpus on this device */
+	mov   x6, #PLATFORM_CORE_COUNT
+
+	mov  x0, #NXP_RESET_ADDR
+	ldr  w5, [x0, #BRR_OFFSET]
+
+	/* load the core mask for the first core */
+	mov  x7, #1
+
+	/* x4 = COREDISABLESR
+	 * x5 = BRR
+	 * x6 = loop count
+	 * x7 = core mask bit
+	 */
+2:
+	/* check if the core is to be disabled */
+	tst  x4, x7
+	b.eq 1f
+
+	/* see if disabled cores have already been released from reset */
+	tst  x5, x7
+	b.ne 5f
+
+	/* if core has not been released, then release it (0-3) */
+	mov  x0, x7
+	bl   _soc_core_release
+
+	/* record the core state in the data area (0-3) */
+	mov  x0, x7
+	mov  x1, #CORE_STATE_DATA
+	mov  x2, #CORE_DISABLED
+	bl   _setCoreData
+
+1:
+	/* see if this is a cluster 3 core */
+	mov   x3, #CLUSTER_3_CORES_MASK
+	tst   x3, x7
+	b.eq  5f
+
+	/* this is a cluster 3 core - see if it needs to be held in reset */
+	cmp  x8, #CLUSTER_3_IN_RESET
+	b.ne 5f
+
+	/* record the core state as disabled in the data area (0-3) */
+	mov  x0, x7
+	mov  x1, #CORE_STATE_DATA
+	mov  x2, #CORE_DISABLED
+	bl   _setCoreData
+
+5:
+	/* decrement the counter */
+	subs  x6, x6, #1
+	b.le  3f
+
+	/* shift the core mask to the next core */
+	lsl   x7, x7, #1
+	/* continue */
+	b	 2b
+3:
+	cmp  x8, #CLUSTER_3_IN_RESET
+	b.ne 6f
+
+	/* we need to hold the cluster 3 cores in reset,
+	 * so mark them in the COREDISR and COREDISABLEDSR registers as
+	 * "disabled", and the rest of the sw stack will leave them alone
+	 * thinking that they have been disabled
+	 */
+	mov  x0, #NXP_DCFG_ADDR
+	ldr  w1, [x0, #DCFG_COREDISR_OFFSET]
+	orr  w1, w1, #CLUSTER_3_CORES_MASK
+	str  w1, [x0, #DCFG_COREDISR_OFFSET]
+
+	ldr  w2, [x0, #DCFG_COREDISABLEDSR_OFFSET]
+	orr  w2, w2, #CLUSTER_3_CORES_MASK
+	str  w2, [x0, #DCFG_COREDISABLEDSR_OFFSET]
+	dsb  sy
+	isb
+
+#if (PSCI_TEST)
+	/* x0 = NXP_DCFG_ADDR : read COREDISABLESR */
+	ldr  w4, [x0, #DCFG_COREDISABLEDSR_OFFSET]
+	/* read COREDISR */
+	ldr  w3, [x0, #DCFG_COREDISR_OFFSET]
+#endif
+
+6:
+	mov  x30, x9
+	ret
+
+endfunc release_disabled
+
+
+/* Function setc up the TrustZone Address Space Controller (TZASC)
+ * in:  none
+ * out: none
+ * uses x0, x1
+ */
+func init_tzpc
+
+	/* set Non Secure access for all devices protected via TZPC */
+
+	/* decode Protection-0 Set Reg */
+	ldr	x1, =TZPCDECPROT_0_SET_BASE
+	/* set decode region to NS, Bits[7:0] */
+	mov	w0, #0xFF
+	str	w0, [x1]
+
+	/* decode Protection-1 Set Reg */
+	ldr	x1, =TZPCDECPROT_1_SET_BASE
+	/* set decode region to NS, Bits[7:0] */
+	mov	w0, #0xFF
+	str	w0, [x1]
+
+	/* decode Protection-2 Set Reg */
+	ldr	x1, =TZPCDECPROT_2_SET_BASE
+	/* set decode region to NS, Bits[7:0] */
+	mov	w0, #0xFF
+	str	w0, [x1]
+
+	/* entire SRAM as NS */
+	/* secure RAM region size Reg */
+	ldr	x1, =TZPC_BASE
+	/* 0x00000000 = no secure region */
+	mov	w0, #0x00000000
+	str	w0, [x1]
+
+	ret
+endfunc init_tzpc
+
+/* write a register in the DCFG block
+ * in:  x0 = offset
+ * in:  w1 = value to write
+ * uses x0, x1, x2
+ */
+func _write_reg_dcfg
+	ldr  x2, =NXP_DCFG_ADDR
+	str  w1, [x2, x0]
+	ret
+endfunc _write_reg_dcfg
+
+
+/* read a register in the DCFG block
+ * in:  x0 = offset
+ * out: w0 = value read
+ * uses x0, x1, x2
+ */
+func _read_reg_dcfg
+	ldr  x2, =NXP_DCFG_ADDR
+	ldr  w1, [x2, x0]
+	mov  w0, w1
+	ret
+endfunc _read_reg_dcfg
+
+
+/* Function returns an mpidr value for a core, given a core_mask_lsb
+ * in:  x0 = core mask lsb
+ * out: x0 = affinity2:affinity1:affinity0, where affinity is 8-bits
+ * uses x0, x1
+ */
+func get_mpidr_value
+
+	/* convert a core mask to an SoC core number */
+	clz  w0, w0
+	mov  w1, #31
+	sub  w0, w1, w0
+
+	/* get the mpidr core number from the SoC core number */
+	mov  w1, wzr
+	tst  x0, #1
+	b.eq 1f
+	orr  w1, w1, #1
+
+1:
+	/* extract the cluster number */
+	lsr  w0, w0, #1
+	orr  w0, w1, w0, lsl #8
+
+	ret
+endfunc get_mpidr_value
+
+
+/* Function returns the redistributor base address for the core specified
+ * in x1
+ * in:  x0 - core mask lsb of specified core
+ * out: x0 = redistributor rd base address for specified core
+ * uses x0, x1, x2
+ */
+func get_gic_rd_base
+	clz  w1, w0
+	mov  w2, #0x20
+	sub  w2, w2, w1
+	sub  w2, w2, #1
+
+	ldr  x0, =NXP_GICR_ADDR
+	mov  x1, #GIC_RD_OFFSET
+
+	/* x2 = core number
+	 * loop counter
+	 */
+2:
+	cbz  x2, 1f
+	add  x0, x0, x1
+	sub  x2, x2, #1
+	b	2b
+1:
+	ret
+endfunc get_gic_rd_base
+
+
+/* Function returns the redistributor base address for the core specified
+ * in x1
+ * in:  x0 - core mask lsb of specified core
+ * out: x0 = redistributor sgi base address for specified core
+ * uses x0, x1, x2
+ */
+func get_gic_sgi_base
+	clz  w1, w0
+	mov  w2, #0x20
+	sub  w2, w2, w1
+	sub  w2, w2, #1
+
+	ldr  x0, =NXP_GICR_SGI_ADDR
+	mov  x1, #GIC_SGI_OFFSET
+
+	/* loop counter */
+2:
+	cbz  x2, 1f		/* x2 = core number */
+	add  x0, x0, x1
+	sub  x2, x2, #1
+	b	2b
+1:
+	ret
+endfunc get_gic_sgi_base
+
+/* Function writes a register in the RESET block
+ * in:  x0 = offset
+ * in:  w1 = value to write
+ * uses x0, x1, x2
+ */
+func _write_reg_reset
+	ldr  x2, =NXP_RESET_ADDR
+	str  w1, [x2, x0]
+	ret
+endfunc _write_reg_reset
+
+
+/* Function reads a register in the RESET block
+ * in:  x0 = offset
+ * out: w0 = value read
+ * uses x0, x1
+ */
+func _read_reg_reset
+	ldr  x1, =NXP_RESET_ADDR
+	ldr  w0, [x1, x0]
+	ret
+endfunc _read_reg_reset