chore: fvp_r:  Initial No-EL3 and MPU Implementation

For v8-R64, especially R82, creating code to run BL1 at EL2, using MPU.

Signed-off-by: Gary Morrison <gary.morrison@arm.com>
Change-Id: I439ac3915b982ad1e61d24365bdd1584b3070425
diff --git a/include/arch/aarch64/el2_common_macros.S b/include/arch/aarch64/el2_common_macros.S
new file mode 100644
index 0000000..c57a1ec
--- /dev/null
+++ b/include/arch/aarch64/el2_common_macros.S
@@ -0,0 +1,448 @@
+/*
+ * Copyright (c) 2015-2021, ARM Limited and Contributors. All rights reserved.
+ *
+ * SPDX-License-Identifier: BSD-3-Clause
+ */
+
+#ifndef EL2_COMMON_MACROS_S
+#define EL2_COMMON_MACROS_S
+
+#include <arch.h>
+#include <asm_macros.S>
+#include <context.h>
+#include <lib/xlat_tables/xlat_tables_defs.h>
+
+#include <platform_def.h>
+
+	/*
+	 * Helper macro to initialise system registers at EL2.
+	 */
+	.macro el2_arch_init_common
+
+	/* ---------------------------------------------------------------------
+	 * SCTLR_EL2 has already been initialised - read current value before
+	 * modifying.
+	 *
+	 * SCTLR_EL2.I: Enable the instruction cache.
+	 *
+	 * SCTLR_EL2.SA: Enable Stack Alignment check. A SP alignment fault
+	 *  exception is generated if a load or store instruction executed at
+	 *  EL2 uses the SP as the base address and the SP is not aligned to a
+	 *  16-byte boundary.
+	 *
+	 * SCTLR_EL2.A: Enable Alignment fault checking. All instructions that
+	 *  load or store one or more registers have an alignment check that the
+	 *  address being accessed is aligned to the size of the data element(s)
+	 *  being accessed.
+	 * ---------------------------------------------------------------------
+	 */
+	mov	x1, #(SCTLR_I_BIT | SCTLR_A_BIT | SCTLR_SA_BIT)
+	mrs	x0, sctlr_el2
+	orr	x0, x0, x1
+	msr	sctlr_el2, x0
+	isb
+
+	/* ---------------------------------------------------------------------
+	 * Initialise HCR_EL2, setting all fields rather than relying on HW.
+	 * All fields are architecturally UNKNOWN on reset. The following fields
+	 * do not change during the TF lifetime. The remaining fields are set to
+	 * zero here but are updated ahead of transitioning to a lower EL in the
+	 * function cm_init_context_common().
+	 *
+	 * HCR_EL2.TWE: Set to zero so that execution of WFE instructions at
+	 *  EL2, EL1 and EL0 are not trapped to EL2.
+	 *
+	 * HCR_EL2.TWI: Set to zero so that execution of WFI instructions at
+	 *  EL2, EL1 and EL0 are not trapped to EL2.
+	 *
+	 * HCR_EL2.HCD: Set to zero to enable HVC calls at EL1 and above,
+	 *  from both Security states and both Execution states.
+	 *
+	 * HCR_EL2.TEA: Set to one to route External Aborts and SError
+	 * Interrupts to EL2 when executing at any EL.
+	 *
+	 * HCR_EL2.{API,APK}: For Armv8.3 pointer authentication feature,
+	 * disable traps to EL2 when accessing key registers or using
+	 * pointer authentication instructions from lower ELs.
+	 * ---------------------------------------------------------------------
+	 */
+	mov_imm	x0, ((HCR_RESET_VAL | HCR_TEA_BIT) \
+			& ~(HCR_TWE_BIT | HCR_TWI_BIT | HCR_HCD_BIT))
+#if CTX_INCLUDE_PAUTH_REGS
+	/*
+	 * If the pointer authentication registers are saved during world
+	 * switches, enable pointer authentication everywhere, as it is safe to
+	 * do so.
+	 */
+	orr	x0, x0, #(HCR_API_BIT | HCR_APK_BIT)
+#endif  /* CTX_INCLUDE_PAUTH_REGS */
+	msr	hcr_el2, x0
+
+	/* ---------------------------------------------------------------------
+	 * Initialise MDCR_EL2, setting all fields rather than relying on
+	 * hw. Some fields are architecturally UNKNOWN on reset.
+	 *
+	 * MDCR_EL2.SDD: Set to one to disable AArch64 Secure self-hosted
+	 *  debug. Debug exceptions, other than Breakpoint Instruction
+	 *  exceptions, are disabled from all ELs in Secure state.
+	 *
+	 * MDCR_EL2.TDOSA: Set to zero so that EL2 and EL2 System register
+	 *  access to the powerdown debug registers do not trap to EL2.
+	 *
+	 * MDCR_EL2.TDA: Set to zero to allow EL0, EL1 and EL2 access to the
+	 *  debug registers, other than those registers that are controlled by
+	 *  MDCR_EL2.TDOSA.
+	 *
+	 * MDCR_EL2.TPM: Set to zero so that EL0, EL1, and EL2 System
+	 *  register accesses to all Performance Monitors registers do not trap
+	 *  to EL2.
+	 *
+	 * MDCR_EL2.SCCD: Set to one so that cycle counting by PMCCNTR_EL0
+	 *  is prohibited in Secure state. This bit is RES0 in versions of the
+	 *  architecture with FEAT_PMUv3p5 not implemented, setting it to 1
+	 *  doesn't have any effect on them.
+	 *
+	 * MDCR_EL2.MCCD: Set to one so that cycle counting by PMCCNTR_EL0
+	 *  is prohibited in EL2. This bit is RES0 in versions of the
+	 *  architecture with FEAT_PMUv3p7 not implemented, setting it to 1
+	 *  doesn't have any effect on them.
+	 *
+	 * MDCR_EL2.SPME: Set to zero so that event counting by the program-
+	 *  mable counters PMEVCNTR<n>_EL0 is prohibited in Secure state. If
+	 *  ARMv8.2 Debug is not implemented this bit does not have any effect
+	 *  on the counters unless there is support for the implementation
+	 *  defined authentication interface
+	 *  ExternalSecureNoninvasiveDebugEnabled().
+	 * ---------------------------------------------------------------------
+	 */
+	mov_imm	x0, ((MDCR_EL2_RESET_VAL | MDCR_SDD_BIT | \
+		      MDCR_SPD32(MDCR_SPD32_DISABLE) | MDCR_SCCD_BIT | \
+		      MDCR_MCCD_BIT) & ~(MDCR_SPME_BIT | MDCR_TDOSA_BIT | \
+		      MDCR_TDA_BIT | MDCR_TPM_BIT))
+
+	msr	mdcr_el2, x0
+
+	/* ---------------------------------------------------------------------
+	 * Initialise PMCR_EL0 setting all fields rather than relying
+	 * on hw. Some fields are architecturally UNKNOWN on reset.
+	 *
+	 * PMCR_EL0.LP: Set to one so that event counter overflow, that
+	 *  is recorded in PMOVSCLR_EL0[0-30], occurs on the increment
+	 *  that changes PMEVCNTR<n>_EL0[63] from 1 to 0, when ARMv8.5-PMU
+	 *  is implemented. This bit is RES0 in versions of the architecture
+	 *  earlier than ARMv8.5, setting it to 1 doesn't have any effect
+	 *  on them.
+	 *
+	 * PMCR_EL0.LC: Set to one so that cycle counter overflow, that
+	 *  is recorded in PMOVSCLR_EL0[31], occurs on the increment
+	 *  that changes PMCCNTR_EL0[63] from 1 to 0.
+	 *
+	 * PMCR_EL0.DP: Set to one so that the cycle counter,
+	 *  PMCCNTR_EL0 does not count when event counting is prohibited.
+	 *
+	 * PMCR_EL0.X: Set to zero to disable export of events.
+	 *
+	 * PMCR_EL0.D: Set to zero so that, when enabled, PMCCNTR_EL0
+	 *  counts on every clock cycle.
+	 * ---------------------------------------------------------------------
+	 */
+	mov_imm	x0, ((PMCR_EL0_RESET_VAL | PMCR_EL0_LP_BIT | \
+		      PMCR_EL0_LC_BIT | PMCR_EL0_DP_BIT) & \
+		    ~(PMCR_EL0_X_BIT | PMCR_EL0_D_BIT))
+
+	msr	pmcr_el0, x0
+
+	/* ---------------------------------------------------------------------
+	 * Enable External Aborts and SError Interrupts now that the exception
+	 * vectors have been setup.
+	 * ---------------------------------------------------------------------
+	 */
+	msr	daifclr, #DAIF_ABT_BIT
+
+	/* ---------------------------------------------------------------------
+	 * Initialise CPTR_EL2, setting all fields rather than relying on hw.
+	 * All fields are architecturally UNKNOWN on reset.
+	 *
+	 * CPTR_EL2.TCPAC: Set to zero so that any accesses to CPACR_EL1 do
+	 * not trap to EL2.
+	 *
+	 * CPTR_EL2.TTA: Set to zero so that System register accesses to the
+	 *  trace registers do not trap to EL2.
+	 *
+	 * CPTR_EL2.TFP: Set to zero so that accesses to the V- or Z- registers
+	 *  by Advanced SIMD, floating-point or SVE instructions (if implemented)
+	 *  do not trap to EL2.
+	 */
+
+	mov_imm x0, (CPTR_EL2_RESET_VAL & ~(TCPAC_BIT | TTA_BIT | TFP_BIT))
+	msr	cptr_el2, x0
+
+	/*
+	 * If Data Independent Timing (DIT) functionality is implemented,
+	 * always enable DIT in EL2
+	 */
+	mrs	x0, id_aa64pfr0_el1
+	ubfx	x0, x0, #ID_AA64PFR0_DIT_SHIFT, #ID_AA64PFR0_DIT_LENGTH
+	cmp	x0, #ID_AA64PFR0_DIT_SUPPORTED
+	bne	1f
+	mov	x0, #DIT_BIT
+	msr	DIT, x0
+1:
+	.endm
+
+/* -----------------------------------------------------------------------------
+ * This is the super set of actions that need to be performed during a cold boot
+ * or a warm boot in EL2. This code is shared by BL1 and BL31.
+ *
+ * This macro will always perform reset handling, architectural initialisations
+ * and stack setup. The rest of the actions are optional because they might not
+ * be needed, depending on the context in which this macro is called. This is
+ * why this macro is parameterised ; each parameter allows to enable/disable
+ * some actions.
+ *
+ *  _init_sctlr:
+ *	Whether the macro needs to initialise SCTLR_EL2, including configuring
+ *      the endianness of data accesses.
+ *
+ *  _warm_boot_mailbox:
+ *	Whether the macro needs to detect the type of boot (cold/warm). The
+ *	detection is based on the platform entrypoint address : if it is zero
+ *	then it is a cold boot, otherwise it is a warm boot. In the latter case,
+ *	this macro jumps on the platform entrypoint address.
+ *
+ *  _secondary_cold_boot:
+ *	Whether the macro needs to identify the CPU that is calling it: primary
+ *	CPU or secondary CPU. The primary CPU will be allowed to carry on with
+ *	the platform initialisations, while the secondaries will be put in a
+ *	platform-specific state in the meantime.
+ *
+ *	If the caller knows this macro will only be called by the primary CPU
+ *	then this parameter can be defined to 0 to skip this step.
+ *
+ * _init_memory:
+ *	Whether the macro needs to initialise the memory.
+ *
+ * _init_c_runtime:
+ *	Whether the macro needs to initialise the C runtime environment.
+ *
+ * _exception_vectors:
+ *	Address of the exception vectors to program in the VBAR_EL2 register.
+ *
+ * _pie_fixup_size:
+ *	Size of memory region to fixup Global Descriptor Table (GDT).
+ *
+ *	A non-zero value is expected when firmware needs GDT to be fixed-up.
+ *
+ * -----------------------------------------------------------------------------
+ */
+	.macro el2_entrypoint_common					\
+		_init_sctlr, _warm_boot_mailbox, _secondary_cold_boot,	\
+		_init_memory, _init_c_runtime, _exception_vectors,	\
+		_pie_fixup_size
+
+	.if \_init_sctlr
+		/* -------------------------------------------------------------
+		 * This is the initialisation of SCTLR_EL2 and so must ensure
+		 * that all fields are explicitly set rather than relying on hw.
+		 * Some fields reset to an IMPLEMENTATION DEFINED value and
+		 * others are architecturally UNKNOWN on reset.
+		 *
+		 * SCTLR.EE: Set the CPU endianness before doing anything that
+		 *  might involve memory reads or writes. Set to zero to select
+		 *  Little Endian.
+		 *
+		 * SCTLR_EL2.WXN: For the EL2 translation regime, this field can
+		 *  force all memory regions that are writeable to be treated as
+		 *  XN (Execute-never). Set to zero so that this control has no
+		 *  effect on memory access permissions.
+		 *
+		 * SCTLR_EL2.SA: Set to zero to disable Stack Alignment check.
+		 *
+		 * SCTLR_EL2.A: Set to zero to disable Alignment fault checking.
+		 *
+		 * SCTLR.DSSBS: Set to zero to disable speculation store bypass
+		 *  safe behaviour upon exception entry to EL2.
+		 * -------------------------------------------------------------
+		 */
+		mov_imm	x0, (SCTLR_RESET_VAL & ~(SCTLR_EE_BIT | SCTLR_WXN_BIT \
+				| SCTLR_SA_BIT | SCTLR_A_BIT | SCTLR_DSSBS_BIT))
+		msr	sctlr_el2, x0
+		isb
+	.endif /* _init_sctlr */
+
+#if DISABLE_MTPMU
+		bl	mtpmu_disable
+#endif
+
+	.if \_warm_boot_mailbox
+		/* -------------------------------------------------------------
+		 * This code will be executed for both warm and cold resets.
+		 * Now is the time to distinguish between the two.
+		 * Query the platform entrypoint address and if it is not zero
+		 * then it means it is a warm boot so jump to this address.
+		 * -------------------------------------------------------------
+		 */
+		bl	plat_get_my_entrypoint
+		cbz	x0, do_cold_boot
+		br	x0
+
+	do_cold_boot:
+	.endif /* _warm_boot_mailbox */
+
+	.if \_pie_fixup_size
+#if ENABLE_PIE
+		/*
+		 * ------------------------------------------------------------
+		 * If PIE is enabled fixup the Global descriptor Table only
+		 * once during primary core cold boot path.
+		 *
+		 * Compile time base address, required for fixup, is calculated
+		 * using "pie_fixup" label present within first page.
+		 * ------------------------------------------------------------
+		 */
+	pie_fixup:
+		ldr	x0, =pie_fixup
+		and	x0, x0, #~(PAGE_SIZE_MASK)
+		mov_imm	x1, \_pie_fixup_size
+		add	x1, x1, x0
+		bl	fixup_gdt_reloc
+#endif /* ENABLE_PIE */
+	.endif /* _pie_fixup_size */
+
+	/* ---------------------------------------------------------------------
+	 * Set the exception vectors.
+	 * ---------------------------------------------------------------------
+	 */
+	adr	x0, \_exception_vectors
+	msr	vbar_el2, x0
+	isb
+
+	/* ---------------------------------------------------------------------
+	 * It is a cold boot.
+	 * Perform any processor specific actions upon reset e.g. cache, TLB
+	 * invalidations etc.
+	 * ---------------------------------------------------------------------
+	 */
+	bl	reset_handler
+
+	el2_arch_init_common
+
+	.if \_secondary_cold_boot
+		/* -------------------------------------------------------------
+		 * Check if this is a primary or secondary CPU cold boot.
+		 * The primary CPU will set up the platform while the
+		 * secondaries are placed in a platform-specific state until the
+		 * primary CPU performs the necessary actions to bring them out
+		 * of that state and allows entry into the OS.
+		 * -------------------------------------------------------------
+		 */
+		bl	plat_is_my_cpu_primary
+		cbnz	w0, do_primary_cold_boot
+
+		/* This is a cold boot on a secondary CPU */
+		bl	plat_secondary_cold_boot_setup
+		/* plat_secondary_cold_boot_setup() is not supposed to return */
+		bl	el2_panic
+	do_primary_cold_boot:
+	.endif /* _secondary_cold_boot */
+
+	/* ---------------------------------------------------------------------
+	 * Initialize memory now. Secondary CPU initialization won't get to this
+	 * point.
+	 * ---------------------------------------------------------------------
+	 */
+
+	.if \_init_memory
+		bl	platform_mem_init
+	.endif /* _init_memory */
+
+	/* ---------------------------------------------------------------------
+	 * Init C runtime environment:
+	 *   - Zero-initialise the NOBITS sections. There are 2 of them:
+	 *       - the .bss section;
+	 *       - the coherent memory section (if any).
+	 *   - Relocate the data section from ROM to RAM, if required.
+	 * ---------------------------------------------------------------------
+	 */
+	.if \_init_c_runtime
+		adrp	x0, __BSS_START__
+		add	x0, x0, :lo12:__BSS_START__
+
+		adrp	x1, __BSS_END__
+		add	x1, x1, :lo12:__BSS_END__
+		sub	x1, x1, x0
+		bl	zeromem
+
+#if defined(IMAGE_BL1) || (defined(IMAGE_BL2) && BL2_AT_EL3 && BL2_IN_XIP_MEM)
+		adrp	x0, __DATA_RAM_START__
+		add	x0, x0, :lo12:__DATA_RAM_START__
+		adrp	x1, __DATA_ROM_START__
+		add	x1, x1, :lo12:__DATA_ROM_START__
+		adrp	x2, __DATA_RAM_END__
+		add	x2, x2, :lo12:__DATA_RAM_END__
+		sub	x2, x2, x0
+		bl	memcpy16
+#endif
+	.endif /* _init_c_runtime */
+
+	/* ---------------------------------------------------------------------
+	 * Use SP_EL0 for the C runtime stack.
+	 * ---------------------------------------------------------------------
+	 */
+	msr	spsel, #0
+
+	/* ---------------------------------------------------------------------
+	 * Allocate a stack whose memory will be marked as Normal-IS-WBWA when
+	 * the MMU is enabled. There is no risk of reading stale stack memory
+	 * after enabling the MMU as only the primary CPU is running at the
+	 * moment.
+	 * ---------------------------------------------------------------------
+	 */
+	bl	plat_set_my_stack
+
+#if STACK_PROTECTOR_ENABLED
+	.if \_init_c_runtime
+	bl	update_stack_protector_canary
+	.endif /* _init_c_runtime */
+#endif
+	.endm
+
+	.macro	apply_at_speculative_wa
+#if ERRATA_SPECULATIVE_AT
+	/*
+	 * Explicitly save x30 so as to free up a register and to enable
+	 * branching and also, save x29 which will be used in the called
+	 * function
+	 */
+	stp	x29, x30, [sp, #CTX_GPREGS_OFFSET + CTX_GPREG_X29]
+	bl	save_and_update_ptw_el1_sys_regs
+	ldp	x29, x30, [sp, #CTX_GPREGS_OFFSET + CTX_GPREG_X29]
+#endif
+	.endm
+
+	.macro	restore_ptw_el1_sys_regs
+#if ERRATA_SPECULATIVE_AT
+	/* -----------------------------------------------------------
+	 * In case of ERRATA_SPECULATIVE_AT, must follow below order
+	 * to ensure that page table walk is not enabled until
+	 * restoration of all EL1 system registers. TCR_EL1 register
+	 * should be updated at the end which restores previous page
+	 * table walk setting of stage1 i.e.(TCR_EL1.EPDx) bits. ISB
+	 * ensures that CPU does below steps in order.
+	 *
+	 * 1. Ensure all other system registers are written before
+	 *    updating SCTLR_EL1 using ISB.
+	 * 2. Restore SCTLR_EL1 register.
+	 * 3. Ensure SCTLR_EL1 written successfully using ISB.
+	 * 4. Restore TCR_EL1 register.
+	 * -----------------------------------------------------------
+	 */
+	isb
+	ldp	x28, x29, [sp, #CTX_EL1_SYSREGS_OFFSET + CTX_SCTLR_EL1]
+	msr	sctlr_el1, x28
+	isb
+	msr	tcr_el1, x29
+#endif
+	.endm
+
+#endif /* EL2_COMMON_MACROS_S */