cortex_a.c

Go to the documentation of this file.

// SPDX-License-Identifier: GPL-2.0-or-later
 
/***************************************************************************
 *   Copyright (C) 2005 by Dominic Rath                                    *
 *   Dominic.Rath@gmx.de                                                   *
 *                                                                         *
 *   Copyright (C) 2006 by Magnus Lundin                                   *
 *   lundin@mlu.mine.nu                                                    *
 *                                                                         *
 *   Copyright (C) 2008 by Spencer Oliver                                  *
 *   spen@spen-soft.co.uk                                                  *
 *                                                                         *
 *   Copyright (C) 2009 by Dirk Behme                                      *
 *   dirk.behme@gmail.com - copy from cortex_m3                            *
 *                                                                         *
 *   Copyright (C) 2010 Øyvind Harboe                                      *
 *   oyvind.harboe@zylin.com                                               *
 *                                                                         *
 *   Copyright (C) ST-Ericsson SA 2011                                     *
 *   michel.jaouen@stericsson.com : smp minimum support                    *
 *                                                                         *
 *   Copyright (C) Broadcom 2012                                           *
 *   ehunter@broadcom.com : Cortex-R4 support                              *
 *                                                                         *
 *   Copyright (C) 2013 Kamal Dasu                                         *
 *   kdasu.kdev@gmail.com                                                  *
 *                                                                         *
 *   Copyright (C) 2016 Chengyu Zheng                                      *
 *   chengyu.zheng@polimi.it : watchpoint support                          *
 *                                                                         *
 *   Cortex-A8(tm) TRM, ARM DDI 0344H                                      *
 *   Cortex-A9(tm) TRM, ARM DDI 0407F                                      *
 *   Cortex-A4(tm) TRM, ARM DDI 0363E                                      *
 *   Cortex-A15(tm)TRM, ARM DDI 0438C                                      *
 *   Architecture Reference Manual, ARMv7-A and ARMv7-R, ARM DDI 0406C.d   *
 *                                                                         *
 ***************************************************************************/
 
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
 
#include "breakpoints.h"
#include "cortex_a.h"
#include "register.h"
#include "armv7a_mmu.h"
#include "target_request.h"
#include "target_type.h"
#include "arm_coresight.h"
#include "arm_opcodes.h"
#include "arm_semihosting.h"
#include "jtag/interface.h"
#include "transport/transport.h"
#include "smp.h"
#include <helper/bits.h>
#include <helper/nvp.h>
#include <helper/time_support.h>
#include <helper/align.h>
 
static int cortex_a_poll(struct target *target);
static int cortex_a_debug_entry(struct target *target);
static int cortex_a_restore_context(struct target *target, bool bpwp);
static int cortex_a_set_breakpoint(struct target *target,
    struct breakpoint *breakpoint, uint8_t matchmode);
static int cortex_a_set_context_breakpoint(struct target *target,
    struct breakpoint *breakpoint, uint8_t matchmode);
static int cortex_a_set_hybrid_breakpoint(struct target *target,
    struct breakpoint *breakpoint);
static int cortex_a_unset_breakpoint(struct target *target,
    struct breakpoint *breakpoint);
static int cortex_a_wait_dscr_bits(struct target *target, uint32_t mask,
    uint32_t value, uint32_t *dscr);
static int cortex_a_mmu(struct target *target, bool *enabled);
static int cortex_a_mmu_modify(struct target *target, bool enable);
static int cortex_a_virt2phys(struct target *target,
    target_addr_t virt, target_addr_t *phys);
static int cortex_a_read_cpu_memory(struct target *target,
    uint32_t address, uint32_t size, uint32_t count, uint8_t *buffer);
 
static unsigned int ilog2(unsigned int x)
{
    unsigned int y = 0;
    x /= 2;
    while (x) {
        ++y;
        x /= 2;
        }
    return y;
}
 
/*  restore cp15_control_reg at resume */
static int cortex_a_restore_cp15_control_reg(struct target *target)
{
    int retval = ERROR_OK;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    if (cortex_a->cp15_control_reg != cortex_a->cp15_control_reg_curr) {
        cortex_a->cp15_control_reg_curr = cortex_a->cp15_control_reg;
        /* LOG_INFO("cp15_control_reg: %8.8" PRIx32, cortex_a->cp15_control_reg); */
        retval = armv7a->arm.mcr(target, 15,
                0, 0,   /* op1, op2 */
                1, 0,   /* CRn, CRm */
                cortex_a->cp15_control_reg);
    }
    return retval;
}
 
/*
 * Set up ARM core for memory access.
 * If !phys_access, switch to SVC mode and make sure MMU is on
 * If phys_access, switch off mmu
 */
static int cortex_a_prep_memaccess(struct target *target, bool phys_access)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    bool mmu_enabled = false;
 
    if (!phys_access) {
        arm_dpm_modeswitch(&armv7a->dpm, ARM_MODE_SVC);
        cortex_a_mmu(target, &mmu_enabled);
        if (mmu_enabled)
            cortex_a_mmu_modify(target, true);
        if (cortex_a->dacrfixup_mode == CORTEX_A_DACRFIXUP_ON) {
            /* overwrite DACR to all-manager */
            armv7a->arm.mcr(target, 15,
                    0, 0, 3, 0,
                    0xFFFFFFFF);
        }
    } else {
        cortex_a_mmu(target, &mmu_enabled);
        if (mmu_enabled)
            cortex_a_mmu_modify(target, false);
    }
    return ERROR_OK;
}
 
/*
 * Restore ARM core after memory access.
 * If !phys_access, switch to previous mode
 * If phys_access, restore MMU setting
 */
static int cortex_a_post_memaccess(struct target *target, bool phys_access)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    if (!phys_access) {
        if (cortex_a->dacrfixup_mode == CORTEX_A_DACRFIXUP_ON) {
            /* restore */
            armv7a->arm.mcr(target, 15,
                    0, 0, 3, 0,
                    cortex_a->cp15_dacr_reg);
        }
        arm_dpm_modeswitch(&armv7a->dpm, ARM_MODE_ANY);
    } else {
        bool mmu_enabled = false;
        cortex_a_mmu(target, &mmu_enabled);
        if (mmu_enabled)
            cortex_a_mmu_modify(target, true);
    }
    return ERROR_OK;
}
 
 
/*  modify cp15_control_reg in order to enable or disable mmu for :
 *  - virt2phys address conversion
 *  - read or write memory in phys or virt address */
static int cortex_a_mmu_modify(struct target *target, bool enable)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = target_to_armv7a(target);
    int retval = ERROR_OK;
    bool need_write = false;
 
    if (enable) {
        /*  if mmu enabled at target stop and mmu not enable */
        if (!(cortex_a->cp15_control_reg & 0x1U)) {
            LOG_ERROR("trying to enable mmu on target stopped with mmu disable");
            return ERROR_FAIL;
        }
        if ((cortex_a->cp15_control_reg_curr & 0x1U) == 0) {
            cortex_a->cp15_control_reg_curr |= 0x1U;
            need_write = true;
        }
    } else {
        if ((cortex_a->cp15_control_reg_curr & 0x1U) == 0x1U) {
            cortex_a->cp15_control_reg_curr &= ~0x1U;
            need_write = true;
        }
    }
 
    if (need_write) {
        LOG_DEBUG("%s, writing cp15 ctrl: %" PRIx32,
            enable ? "enable mmu" : "disable mmu",
            cortex_a->cp15_control_reg_curr);
 
        retval = armv7a->arm.mcr(target, 15,
                0, 0,   /* op1, op2 */
                1, 0,   /* CRn, CRm */
                cortex_a->cp15_control_reg_curr);
    }
    return retval;
}
 
/*
 * Cortex-A Basic debug access, very low level assumes state is saved
 */
static int cortex_a_init_debug_access(struct target *target)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
    uint32_t dscr;
    int retval;
 
    /* lock memory-mapped access to debug registers to prevent
     * software interference */
    retval = mem_ap_write_u32(armv7a->debug_ap,
            armv7a->debug_base + ARM_CS_LAR, 0);
    if (retval != ERROR_OK)
        return retval;
 
    /* Disable cacheline fills and force cache write-through in debug state */
    retval = mem_ap_write_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCCR, 0);
    if (retval != ERROR_OK)
        return retval;
 
    /* Disable TLB lookup and refill/eviction in debug state */
    retval = mem_ap_write_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSMCR, 0);
    if (retval != ERROR_OK)
        return retval;
 
    retval = dap_run(armv7a->debug_ap->dap);
    if (retval != ERROR_OK)
        return retval;
 
    /* Enabling of instruction execution in debug mode is done in debug_entry code */
 
    /* Resync breakpoint registers */
 
    /* Enable halt for breakpoint, watchpoint and vector catch */
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, &dscr);
    if (retval != ERROR_OK)
        return retval;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, dscr | DSCR_HALT_DBG_MODE);
    if (retval != ERROR_OK)
        return retval;
 
    /* Since this is likely called from init or reset, update target state information*/
    return cortex_a_poll(target);
}
 
static int cortex_a_wait_instrcmpl(struct target *target, uint32_t *dscr, bool force)
{
    /* Waits until InstrCmpl_l becomes 1, indicating instruction is done.
     * Writes final value of DSCR into *dscr. Pass force to force always
     * reading DSCR at least once. */
    struct armv7a_common *armv7a = target_to_armv7a(target);
    int retval;
 
    if (force) {
        retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DSCR, dscr);
        if (retval != ERROR_OK) {
            LOG_ERROR("Could not read DSCR register");
            return retval;
        }
    }
 
    retval = cortex_a_wait_dscr_bits(target, DSCR_INSTR_COMP, DSCR_INSTR_COMP, dscr);
    if (retval != ERROR_OK)
        LOG_ERROR("Error waiting for InstrCompl=1");
    return retval;
}
 
/* To reduce needless round-trips, pass in a pointer to the current
 * DSCR value.  Initialize it to zero if you just need to know the
 * value on return from this function; or DSCR_INSTR_COMP if you
 * happen to know that no instruction is pending.
 */
static int cortex_a_exec_opcode(struct target *target,
    uint32_t opcode, uint32_t *dscr_p)
{
    uint32_t dscr;
    int retval;
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    dscr = dscr_p ? *dscr_p : 0;
 
    LOG_DEBUG("exec opcode 0x%08" PRIx32, opcode);
 
    /* Wait for InstrCompl bit to be set */
    retval = cortex_a_wait_instrcmpl(target, dscr_p, false);
    if (retval != ERROR_OK)
        return retval;
 
    retval = mem_ap_write_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_ITR, opcode);
    if (retval != ERROR_OK)
        return retval;
 
    /* Wait for InstrCompl bit to be set */
    retval = cortex_a_wait_instrcmpl(target, &dscr, true);
    if (retval != ERROR_OK) {
        LOG_ERROR("Error waiting for cortex_a_exec_opcode");
        return retval;
    }
 
    if (dscr_p)
        *dscr_p = dscr;
 
    return retval;
}
 
/*
 * Cortex-A implementation of Debug Programmer's Model
 *
 * NOTE the invariant:  these routines return with DSCR_INSTR_COMP set,
 * so there's no need to poll for it before executing an instruction.
 *
 * NOTE that in several of these cases the "stall" mode might be useful.
 * It'd let us queue a few operations together... prepare/finish might
 * be the places to enable/disable that mode.
 */
 
static inline struct cortex_a_common *dpm_to_a(struct arm_dpm *dpm)
{
    return container_of(dpm, struct cortex_a_common, armv7a_common.dpm);
}
 
static int cortex_a_write_dcc(struct cortex_a_common *a, uint32_t data)
{
    LOG_DEBUG("write DCC 0x%08" PRIx32, data);
    return mem_ap_write_u32(a->armv7a_common.debug_ap,
            a->armv7a_common.debug_base + CPUDBG_DTRRX, data);
}
 
static int cortex_a_read_dcc(struct cortex_a_common *a, uint32_t *data,
    uint32_t *dscr_p)
{
    uint32_t dscr = DSCR_INSTR_COMP;
    int retval;
 
    if (dscr_p)
        dscr = *dscr_p;
 
    /* Wait for DTRRXfull */
    retval = cortex_a_wait_dscr_bits(a->armv7a_common.arm.target,
            DSCR_DTR_TX_FULL, DSCR_DTR_TX_FULL, &dscr);
    if (retval != ERROR_OK) {
        LOG_ERROR("Error waiting for read dcc");
        return retval;
    }
 
    retval = mem_ap_read_atomic_u32(a->armv7a_common.debug_ap,
            a->armv7a_common.debug_base + CPUDBG_DTRTX, data);
    if (retval != ERROR_OK)
        return retval;
    /* LOG_DEBUG("read DCC 0x%08" PRIx32, *data); */
 
    if (dscr_p)
        *dscr_p = dscr;
 
    return retval;
}
 
static int cortex_a_dpm_prepare(struct arm_dpm *dpm)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t dscr;
    int retval;
 
    /* set up invariant:  INSTR_COMP is set after ever DPM operation */
    retval = cortex_a_wait_instrcmpl(dpm->arm->target, &dscr, true);
    if (retval != ERROR_OK) {
        LOG_ERROR("Error waiting for dpm prepare");
        return retval;
    }
 
    /* this "should never happen" ... */
    if (dscr & DSCR_DTR_RX_FULL) {
        LOG_ERROR("DSCR_DTR_RX_FULL, dscr 0x%08" PRIx32, dscr);
        /* Clear DCCRX */
        retval = cortex_a_exec_opcode(
                a->armv7a_common.arm.target,
                ARMV4_5_MRC(14, 0, 0, 0, 5, 0),
                &dscr);
        if (retval != ERROR_OK)
            return retval;
    }
 
    return retval;
}
 
static int cortex_a_dpm_finish(struct arm_dpm *dpm)
{
    /* REVISIT what could be done here? */
    return ERROR_OK;
}
 
static int cortex_a_instr_write_data_dcc(struct arm_dpm *dpm,
    uint32_t opcode, uint32_t data)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    int retval;
    uint32_t dscr = DSCR_INSTR_COMP;
 
    retval = cortex_a_write_dcc(a, data);
    if (retval != ERROR_OK)
        return retval;
 
    return cortex_a_exec_opcode(
            a->armv7a_common.arm.target,
            opcode,
            &dscr);
}
 
static int cortex_a_instr_write_data_rt_dcc(struct arm_dpm *dpm,
    uint8_t rt, uint32_t data)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t dscr = DSCR_INSTR_COMP;
    int retval;
 
    if (rt > 15)
        return ERROR_TARGET_INVALID;
 
    retval = cortex_a_write_dcc(a, data);
    if (retval != ERROR_OK)
        return retval;
 
    /* DCCRX to Rt, "MCR p14, 0, R0, c0, c5, 0", 0xEE000E15 */
    return cortex_a_exec_opcode(
            a->armv7a_common.arm.target,
            ARMV4_5_MRC(14, 0, rt, 0, 5, 0),
            &dscr);
}
 
static int cortex_a_instr_write_data_r0(struct arm_dpm *dpm,
    uint32_t opcode, uint32_t data)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t dscr = DSCR_INSTR_COMP;
    int retval;
 
    retval = cortex_a_instr_write_data_rt_dcc(dpm, 0, data);
    if (retval != ERROR_OK)
        return retval;
 
    /* then the opcode, taking data from R0 */
    return cortex_a_exec_opcode(a->armv7a_common.arm.target, opcode,
                &dscr);
}
 
static int cortex_a_instr_write_data_r0_r1(struct arm_dpm *dpm,
    uint32_t opcode, uint64_t data)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t dscr = DSCR_INSTR_COMP;
    int retval;
 
    retval = cortex_a_instr_write_data_rt_dcc(dpm, 0, data & 0xffffffffULL);
    if (retval != ERROR_OK)
        return retval;
 
    retval = cortex_a_instr_write_data_rt_dcc(dpm, 1, data >> 32);
    if (retval != ERROR_OK)
        return retval;
 
    /* then the opcode, taking data from R0, R1 */
    return cortex_a_exec_opcode(a->armv7a_common.arm.target, opcode,
                &dscr);
}
 
static int cortex_a_instr_cpsr_sync(struct arm_dpm *dpm)
{
    struct target *target = dpm->arm->target;
    uint32_t dscr = DSCR_INSTR_COMP;
 
    /* "Prefetch flush" after modifying execution status in CPSR */
    return cortex_a_exec_opcode(target,
            ARMV4_5_MCR(15, 0, 0, 7, 5, 4),
            &dscr);
}
 
static int cortex_a_instr_read_data_dcc(struct arm_dpm *dpm,
    uint32_t opcode, uint32_t *data)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    int retval;
    uint32_t dscr = DSCR_INSTR_COMP;
 
    /* the opcode, writing data to DCC */
    retval = cortex_a_exec_opcode(
            a->armv7a_common.arm.target,
            opcode,
            &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    return cortex_a_read_dcc(a, data, &dscr);
}
 
static int cortex_a_instr_read_data_rt_dcc(struct arm_dpm *dpm,
    uint8_t rt, uint32_t *data)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t dscr = DSCR_INSTR_COMP;
    int retval;
 
    if (rt > 15)
        return ERROR_TARGET_INVALID;
 
    retval = cortex_a_exec_opcode(
            a->armv7a_common.arm.target,
            ARMV4_5_MCR(14, 0, rt, 0, 5, 0),
            &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    return cortex_a_read_dcc(a, data, &dscr);
}
 
static int cortex_a_instr_read_data_r0(struct arm_dpm *dpm,
    uint32_t opcode, uint32_t *data)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t dscr = DSCR_INSTR_COMP;
    int retval;
 
    /* the opcode, writing data to R0 */
    retval = cortex_a_exec_opcode(
            a->armv7a_common.arm.target,
            opcode,
            &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* write R0 to DCC */
    return cortex_a_instr_read_data_rt_dcc(dpm, 0, data);
}
 
static int cortex_a_instr_read_data_r0_r1(struct arm_dpm *dpm,
    uint32_t opcode, uint64_t *data)
{
    uint32_t lo, hi;
    int retval;
 
    /* the opcode, writing data to RO, R1 */
    retval = cortex_a_instr_read_data_r0(dpm, opcode, &lo);
    if (retval != ERROR_OK)
        return retval;
 
    *data = lo;
 
    /* write R1 to DCC */
    retval = cortex_a_instr_read_data_rt_dcc(dpm, 1, &hi);
    if (retval != ERROR_OK)
        return retval;
 
    *data |= (uint64_t)hi << 32;
 
    return retval;
}
 
static int cortex_a_bpwp_enable(struct arm_dpm *dpm, unsigned int index_t,
    uint32_t addr, uint32_t control)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t vr = a->armv7a_common.debug_base;
    uint32_t cr = a->armv7a_common.debug_base;
    int retval;
 
    switch (index_t) {
    case 0 ... 15:  /* breakpoints */
        vr += CPUDBG_BVR_BASE;
        cr += CPUDBG_BCR_BASE;
        break;
    case 16 ... 31: /* watchpoints */
        vr += CPUDBG_WVR_BASE;
        cr += CPUDBG_WCR_BASE;
        index_t -= 16;
        break;
    default:
        return ERROR_FAIL;
    }
    vr += 4 * index_t;
    cr += 4 * index_t;
 
    LOG_DEBUG("A: bpwp enable, vr %08" PRIx32 " cr %08" PRIx32, vr, cr);
 
    retval = mem_ap_write_atomic_u32(a->armv7a_common.debug_ap,
            vr, addr);
    if (retval != ERROR_OK)
        return retval;
    return mem_ap_write_atomic_u32(a->armv7a_common.debug_ap, cr, control);
}
 
static int cortex_a_bpwp_disable(struct arm_dpm *dpm, unsigned int index_t)
{
    struct cortex_a_common *a = dpm_to_a(dpm);
    uint32_t cr;
 
    switch (index_t) {
    case 0 ... 15:
        cr = a->armv7a_common.debug_base + CPUDBG_BCR_BASE;
        break;
    case 16 ... 31:
        cr = a->armv7a_common.debug_base + CPUDBG_WCR_BASE;
        index_t -= 16;
        break;
    default:
        return ERROR_FAIL;
    }
    cr += 4 * index_t;
 
    LOG_DEBUG("A: bpwp disable, cr %08" PRIx32, cr);
 
    /* clear control register */
    return mem_ap_write_atomic_u32(a->armv7a_common.debug_ap, cr, 0);
}
 
static int cortex_a_dpm_setup(struct cortex_a_common *a, uint32_t didr)
{
    struct arm_dpm *dpm = &a->armv7a_common.dpm;
    int retval;
 
    dpm->arm = &a->armv7a_common.arm;
    dpm->didr = didr;
 
    dpm->prepare = cortex_a_dpm_prepare;
    dpm->finish = cortex_a_dpm_finish;
 
    dpm->instr_write_data_dcc = cortex_a_instr_write_data_dcc;
    dpm->instr_write_data_r0 = cortex_a_instr_write_data_r0;
    dpm->instr_write_data_r0_r1 = cortex_a_instr_write_data_r0_r1;
    dpm->instr_cpsr_sync = cortex_a_instr_cpsr_sync;
 
    dpm->instr_read_data_dcc = cortex_a_instr_read_data_dcc;
    dpm->instr_read_data_r0 = cortex_a_instr_read_data_r0;
    dpm->instr_read_data_r0_r1 = cortex_a_instr_read_data_r0_r1;
 
    dpm->bpwp_enable = cortex_a_bpwp_enable;
    dpm->bpwp_disable = cortex_a_bpwp_disable;
 
    retval = arm_dpm_setup(dpm);
    if (retval == ERROR_OK)
        retval = arm_dpm_initialize(dpm);
 
    return retval;
}
static struct target *get_cortex_a(struct target *target, int32_t coreid)
{
    struct target_list *head;
 
    foreach_smp_target(head, target->smp_targets) {
        struct target *curr = head->target;
        if ((curr->coreid == coreid) && (curr->state == TARGET_HALTED))
            return curr;
    }
    return target;
}
static int cortex_a_halt(struct target *target);
 
static int cortex_a_halt_smp(struct target *target)
{
    int retval = 0;
    struct target_list *head;
 
    foreach_smp_target(head, target->smp_targets) {
        struct target *curr = head->target;
        if ((curr != target) && (curr->state != TARGET_HALTED)
            && target_was_examined(curr))
            retval += cortex_a_halt(curr);
    }
    return retval;
}
 
static int update_halt_gdb(struct target *target)
{
    struct target *gdb_target = NULL;
    struct target_list *head;
    struct target *curr;
    int retval = 0;
 
    if (target->gdb_service && target->gdb_service->core[0] == -1) {
        target->gdb_service->target = target;
        target->gdb_service->core[0] = target->coreid;
        retval += cortex_a_halt_smp(target);
    }
 
    if (target->gdb_service)
        gdb_target = target->gdb_service->target;
 
    foreach_smp_target(head, target->smp_targets) {
        curr = head->target;
        /* skip calling context */
        if (curr == target)
            continue;
        if (!target_was_examined(curr))
            continue;
        /* skip targets that were already halted */
        if (curr->state == TARGET_HALTED)
            continue;
        /* Skip gdb_target; it alerts GDB so has to be polled as last one */
        if (curr == gdb_target)
            continue;
 
        /* avoid recursion in cortex_a_poll() */
        curr->smp = 0;
        cortex_a_poll(curr);
        curr->smp = 1;
    }
 
    /* after all targets were updated, poll the gdb serving target */
    if (gdb_target && gdb_target != target)
        cortex_a_poll(gdb_target);
    return retval;
}
 
/*
 * Cortex-A Run control
 */
 
static int cortex_a_poll(struct target *target)
{
    int retval = ERROR_OK;
    uint32_t dscr;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    enum target_state prev_target_state = target->state;
    /*  toggle to another core is done by gdb as follow */
    /*  maint packet J core_id */
    /*  continue */
    /*  the next polling trigger an halt event sent to gdb */
    if ((target->state == TARGET_HALTED) && (target->smp) &&
        (target->gdb_service) &&
        (!target->gdb_service->target)) {
        target->gdb_service->target =
            get_cortex_a(target, target->gdb_service->core[1]);
        target_call_event_callbacks(target, TARGET_EVENT_HALTED);
        return retval;
    }
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, &dscr);
    if (retval != ERROR_OK)
        return retval;
    cortex_a->cpudbg_dscr = dscr;
 
    if (DSCR_RUN_MODE(dscr) == (DSCR_CORE_HALTED | DSCR_CORE_RESTARTED)) {
        if (prev_target_state != TARGET_HALTED) {
            /* We have a halting debug event */
            LOG_TARGET_DEBUG(target, "Target halted");
            target->state = TARGET_HALTED;
 
            retval = cortex_a_debug_entry(target);
            if (retval != ERROR_OK)
                return retval;
 
            if (target->smp) {
                retval = update_halt_gdb(target);
                if (retval != ERROR_OK)
                    return retval;
            }
 
            if (prev_target_state == TARGET_DEBUG_RUNNING) {
                target_call_event_callbacks(target, TARGET_EVENT_DEBUG_HALTED);
            } else { /* prev_target_state is RUNNING, UNKNOWN or RESET */
                if (arm_semihosting(target, &retval) != 0)
                    return retval;
 
                target_call_event_callbacks(target,
                    TARGET_EVENT_HALTED);
            }
        }
    } else
        target->state = TARGET_RUNNING;
 
    return retval;
}
 
static int cortex_a_halt(struct target *target)
{
    int retval;
    uint32_t dscr;
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    /*
     * Tell the core to be halted by writing DRCR with 0x1
     * and then wait for the core to be halted.
     */
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DRCR, DRCR_HALT);
    if (retval != ERROR_OK)
        return retval;
 
    dscr = 0; /* force read of dscr */
    retval = cortex_a_wait_dscr_bits(target, DSCR_CORE_HALTED,
            DSCR_CORE_HALTED, &dscr);
    if (retval != ERROR_OK) {
        LOG_TARGET_ERROR(target, "Error waiting for halt");
        return retval;
    }
 
    target->debug_reason = DBG_REASON_DBGRQ;
 
    return ERROR_OK;
}
 
static int cortex_a_internal_restore(struct target *target, bool current,
    target_addr_t *address, bool handle_breakpoints, bool debug_execution)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
    int retval;
    uint32_t resume_pc;
 
    if (!debug_execution)
        target_free_all_working_areas(target);
 
#if 0
    if (debug_execution) {
        /* Disable interrupts */
        /* We disable interrupts in the PRIMASK register instead of
         * masking with C_MASKINTS,
         * This is probably the same issue as Cortex-M3 Errata 377493:
         * C_MASKINTS in parallel with disabled interrupts can cause
         * local faults to not be taken. */
        buf_set_u32(armv7m->core_cache->reg_list[ARMV7M_PRIMASK].value, 0, 32, 1);
        armv7m->core_cache->reg_list[ARMV7M_PRIMASK].dirty = true;
        armv7m->core_cache->reg_list[ARMV7M_PRIMASK].valid = true;
 
        /* Make sure we are in Thumb mode */
        buf_set_u32(armv7m->core_cache->reg_list[ARMV7M_XPSR].value, 0, 32,
            buf_get_u32(armv7m->core_cache->reg_list[ARMV7M_XPSR].value, 0,
            32) | (1 << 24));
        armv7m->core_cache->reg_list[ARMV7M_XPSR].dirty = true;
        armv7m->core_cache->reg_list[ARMV7M_XPSR].valid = true;
    }
#endif
 
    /* current = true: continue on current pc, otherwise continue at <address> */
    resume_pc = buf_get_u32(arm->pc->value, 0, 32);
    if (!current)
        resume_pc = *address;
    else
        *address = resume_pc;
 
    /* Make sure that the Armv7 gdb thumb fixups does not
     * kill the return address
     */
    switch (arm->core_state) {
    case ARM_STATE_ARM:
        resume_pc &= 0xFFFFFFFC;
        break;
    case ARM_STATE_THUMB:
    case ARM_STATE_THUMB_EE:
        /* When the return address is loaded into PC
         * bit 0 must be 1 to stay in Thumb state
         */
        resume_pc |= 0x1;
        break;
    case ARM_STATE_JAZELLE:
        LOG_TARGET_ERROR(target, "How do I resume into Jazelle state??");
        return ERROR_FAIL;
    case ARM_STATE_AARCH64:
        LOG_TARGET_ERROR(target, "Shouldn't be in AARCH64 state");
        return ERROR_FAIL;
    }
    LOG_TARGET_DEBUG(target, "resume pc = 0x%08" PRIx32, resume_pc);
    buf_set_u32(arm->pc->value, 0, 32, resume_pc);
    arm->pc->dirty = true;
    arm->pc->valid = true;
 
    /* restore dpm_mode at system halt */
    arm_dpm_modeswitch(&armv7a->dpm, ARM_MODE_ANY);
    /* called it now before restoring context because it uses cpu
     * register r0 for restoring cp15 control register */
    retval = cortex_a_restore_cp15_control_reg(target);
    if (retval != ERROR_OK)
        return retval;
    retval = cortex_a_restore_context(target, handle_breakpoints);
    if (retval != ERROR_OK)
        return retval;
    target->debug_reason = DBG_REASON_NOTHALTED;
    target->state = TARGET_RUNNING;
 
    /* registers are now invalid */
    register_cache_invalidate(arm->core_cache);
 
#if 0
    /* the front-end may request us not to handle breakpoints */
    if (handle_breakpoints) {
        /* Single step past breakpoint at current address */
        breakpoint = breakpoint_find(target, resume_pc);
        if (breakpoint) {
            LOG_DEBUG("unset breakpoint at 0x%8.8x", breakpoint->address);
            cortex_m3_unset_breakpoint(target, breakpoint);
            cortex_m3_single_step_core(target);
            cortex_m3_set_breakpoint(target, breakpoint);
        }
    }
 
#endif
    return retval;
}
 
static int cortex_a_internal_restart(struct target *target)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
    int retval;
    uint32_t dscr;
    /*
     * * Restart core and wait for it to be started.  Clear ITRen and sticky
     * * exception flags: see ARMv7 ARM, C5.9.
     *
     * REVISIT: for single stepping, we probably want to
     * disable IRQs by default, with optional override...
     */
 
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    if ((dscr & DSCR_INSTR_COMP) == 0)
        LOG_TARGET_ERROR(target, "DSCR InstrCompl must be set before leaving debug!");
 
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, dscr & ~DSCR_ITR_EN);
    if (retval != ERROR_OK)
        return retval;
 
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DRCR, DRCR_RESTART |
            DRCR_CLEAR_EXCEPTIONS);
    if (retval != ERROR_OK)
        return retval;
 
    dscr = 0; /* force read of dscr */
    retval = cortex_a_wait_dscr_bits(target, DSCR_CORE_RESTARTED,
            DSCR_CORE_RESTARTED, &dscr);
    if (retval != ERROR_OK) {
        LOG_TARGET_ERROR(target, "Error waiting for resume");
        return retval;
    }
 
    target->debug_reason = DBG_REASON_NOTHALTED;
    target->state = TARGET_RUNNING;
 
    /* registers are now invalid */
    register_cache_invalidate(arm->core_cache);
 
    return ERROR_OK;
}
 
static int cortex_a_restore_smp(struct target *target, bool handle_breakpoints)
{
    int retval = ERROR_OK;
    struct target_list *head;
    target_addr_t address;
 
    foreach_smp_target(head, target->smp_targets) {
        struct target *curr = head->target;
        if ((curr != target) && (curr->state != TARGET_RUNNING)
            && target_was_examined(curr)) {
            /*  resume current address , not in step mode */
            int retval2 = cortex_a_internal_restore(curr, true, &address,
                    handle_breakpoints, false);
 
            if (retval2 == ERROR_OK)
                retval2 = cortex_a_internal_restart(curr);
 
            if (retval2 == ERROR_OK)
                target_call_event_callbacks(curr, TARGET_EVENT_RESUMED);
 
            if (retval == ERROR_OK)
                retval = retval2;   // save the first error
        }
    }
    return retval;
}
 
static int cortex_a_resume(struct target *target, bool current,
    target_addr_t address, bool handle_breakpoints, bool debug_execution)
{
    int retval = 0;
    /* dummy resume for smp toggle in order to reduce gdb impact  */
    if ((target->smp) && (target->gdb_service->core[1] != -1)) {
        /*   simulate a start and halt of target */
        target->gdb_service->target = NULL;
        target->gdb_service->core[0] = target->gdb_service->core[1];
        /*  fake resume at next poll we play the  target core[1], see poll*/
        target_call_event_callbacks(target, TARGET_EVENT_RESUMED);
        return 0;
    }
    cortex_a_internal_restore(target, current, &address, handle_breakpoints,
        debug_execution);
    if (target->smp) {
        target->gdb_service->core[0] = -1;
        retval = cortex_a_restore_smp(target, handle_breakpoints);
        if (retval != ERROR_OK)
            return retval;
    }
    cortex_a_internal_restart(target);
 
    if (!debug_execution) {
        target->state = TARGET_RUNNING;
        target_call_event_callbacks(target, TARGET_EVENT_RESUMED);
        LOG_TARGET_DEBUG(target, "target resumed at " TARGET_ADDR_FMT, address);
    } else {
        target->state = TARGET_DEBUG_RUNNING;
        target_call_event_callbacks(target, TARGET_EVENT_DEBUG_RESUMED);
        LOG_TARGET_DEBUG(target, "target debug resumed at " TARGET_ADDR_FMT, address);
    }
 
    return ERROR_OK;
}
 
static int cortex_a_debug_entry(struct target *target)
{
    uint32_t dscr;
    int retval = ERROR_OK;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
 
    LOG_TARGET_DEBUG(target, "dscr = 0x%08" PRIx32, cortex_a->cpudbg_dscr);
 
    /* REVISIT surely we should not re-read DSCR !! */
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* REVISIT see A TRM 12.11.4 steps 2..3 -- make sure that any
     * imprecise data aborts get discarded by issuing a Data
     * Synchronization Barrier:  ARMV4_5_MCR(15, 0, 0, 7, 10, 4).
     */
 
    /* Enable the ITR execution once we are in debug mode */
    dscr |= DSCR_ITR_EN;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Examine debug reason */
    arm_dpm_report_dscr(&armv7a->dpm, cortex_a->cpudbg_dscr);
 
    /* First load register accessible through core debug port */
    retval = arm_dpm_read_current_registers(&armv7a->dpm);
    if (retval != ERROR_OK)
        return retval;
 
    if (arm->spsr) {
        /* read SPSR */
        retval = arm_dpm_read_reg(&armv7a->dpm, arm->spsr, 17);
        if (retval != ERROR_OK)
            return retval;
    }
 
    /* save address of instruction that triggered the watchpoint? */
    if (target->debug_reason == DBG_REASON_WATCHPOINT) {
        /*
         * On v7.1 Debug architecture or on synchronous (precise) watchpoints,
         * WFAR is not used. Take the instruction address from halted PC.
         * See ARM DDI 0406C.d chapter C5.2.2 "Effect of entering Debug state
         * on CP15 registers and the DBGWFAR".
         */
        if (((armv7a->dpm.didr >> 16) & 0xf) > 4 ||
                DSCR_ENTRY(cortex_a->cpudbg_dscr) == DSCR_ENTRY_PRECISE_WATCHPT) {
            armv7a->dpm.wp_addr = buf_get_u32(arm->pc->value, 0, 32);
        } else {
            uint32_t wfar;
 
            retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_WFAR,
                    &wfar);
            if (retval != ERROR_OK)
                return retval;
            arm_dpm_report_wfar(&armv7a->dpm, wfar);
        }
    }
 
#if 0
/* TODO, Move this */
    uint32_t cp15_control_register, cp15_cacr, cp15_nacr;
    cortex_a_read_cp(target, &cp15_control_register, 15, 0, 1, 0, 0);
    LOG_DEBUG("cp15_control_register = 0x%08x", cp15_control_register);
 
    cortex_a_read_cp(target, &cp15_cacr, 15, 0, 1, 0, 2);
    LOG_DEBUG("cp15 Coprocessor Access Control Register = 0x%08x", cp15_cacr);
 
    cortex_a_read_cp(target, &cp15_nacr, 15, 0, 1, 1, 2);
    LOG_DEBUG("cp15 Nonsecure Access Control Register = 0x%08x", cp15_nacr);
#endif
 
    /* Are we in an exception handler */
/*  armv4_5->exception_number = 0; */
    if (armv7a->post_debug_entry) {
        retval = armv7a->post_debug_entry(target);
        if (retval != ERROR_OK)
            return retval;
    }
 
    return retval;
}
 
static int cortex_a_post_debug_entry(struct target *target)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    int retval;
 
    /* MRC p15,0,<Rt>,c1,c0,0 ; Read CP15 System Control Register */
    retval = armv7a->arm.mrc(target, 15,
            0, 0,   /* op1, op2 */
            1, 0,   /* CRn, CRm */
            &cortex_a->cp15_control_reg);
    if (retval != ERROR_OK)
        return retval;
    LOG_TARGET_DEBUG(target, "cp15_control_reg: %8.8" PRIx32,
                     cortex_a->cp15_control_reg);
    cortex_a->cp15_control_reg_curr = cortex_a->cp15_control_reg;
 
    if (!armv7a->is_armv7r)
        armv7a_read_ttbcr(target);
 
    if (!armv7a->armv7a_mmu.armv7a_cache.info_valid)
        armv7a_identify_cache(target);
 
    if (armv7a->is_armv7r) {
        armv7a->armv7a_mmu.mmu_enabled = false;
    } else {
        armv7a->armv7a_mmu.mmu_enabled = cortex_a->cp15_control_reg & 0x1U;
    }
    armv7a->armv7a_mmu.armv7a_cache.d_u_cache_enabled =
        cortex_a->cp15_control_reg & 0x4U;
    armv7a->armv7a_mmu.armv7a_cache.i_cache_enabled =
        cortex_a->cp15_control_reg & 0x1000U;
    cortex_a->curr_mode = armv7a->arm.core_mode;
 
    /* switch to SVC mode to read DACR */
    arm_dpm_modeswitch(&armv7a->dpm, ARM_MODE_SVC);
    armv7a->arm.mrc(target, 15,
            0, 0, 3, 0,
            &cortex_a->cp15_dacr_reg);
 
    LOG_DEBUG("cp15_dacr_reg: %8.8" PRIx32,
            cortex_a->cp15_dacr_reg);
 
    arm_dpm_modeswitch(&armv7a->dpm, ARM_MODE_ANY);
    return ERROR_OK;
}
 
static int cortex_a_set_dscr_bits(struct target *target,
        unsigned long bit_mask, unsigned long value)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
    uint32_t dscr;
 
    /* Read DSCR */
    int retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* clear bitfield */
    dscr &= ~bit_mask;
    /* put new value */
    dscr |= value & bit_mask;
 
    /* write new DSCR */
    return mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DSCR, dscr);
}
 
/*
 * Single-step on ARMv7a/r is implemented through a HW breakpoint that hits
 * every instruction at any address except the address of the current
 * instruction.
 * Such HW breakpoint is never hit in case of a single instruction that jumps
 * on itself (infinite loop), or a WFI or a WFE. In this case, halt the CPU
 * after a timeout.
 * The jump on itself would be executed several times before the timeout forces
 * the halt, but this is not an issue. In ARMv7a/r there are few "pathological"
 * instructions, listed below, that jumps on itself and that can have side
 * effects if executed more than once; but they are not considered as real use
 * cases generated by a compiler.
 * Some example:
 * - 'pop {pc}' or multi register 'pop' including PC, when the new PC value is
 *   the same value of current PC. The single step will not stop at the first
 *   'pop' and will continue taking values from the stack, modifying SP at each
 *   iteration.
 * - 'rfeda', 'rfedb', 'rfeia', 'rfeib', when the new PC value is the same
 *   value of current PC. The register provided to the instruction (usually SP)
 *   will be incremented or decremented at each iteration.
 *
 * TODO: fix exit in case of error, cleaning HW breakpoints.
 */
static int cortex_a_step(struct target *target, bool current, target_addr_t address,
    bool handle_breakpoints)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
    struct breakpoint *breakpoint = NULL;
    struct breakpoint stepbreakpoint;
    struct reg *r;
    int retval;
 
    if (target->state != TARGET_HALTED) {
        LOG_TARGET_ERROR(target, "not halted");
        return ERROR_TARGET_NOT_HALTED;
    }
 
    /* current = true: continue on current pc, otherwise continue at <address> */
    r = arm->pc;
    if (!current)
        buf_set_u32(r->value, 0, 32, address);
    else
        address = buf_get_u32(r->value, 0, 32);
 
    /* The front-end may request us not to handle breakpoints.
     * But since Cortex-A uses breakpoint for single step,
     * we MUST handle breakpoints.
     */
    handle_breakpoints = true;
    if (handle_breakpoints) {
        breakpoint = breakpoint_find(target, address);
        if (breakpoint)
            cortex_a_unset_breakpoint(target, breakpoint);
    }
 
    /* Setup single step breakpoint */
    stepbreakpoint.address = address;
    stepbreakpoint.asid = 0;
    stepbreakpoint.length = (arm->core_state == ARM_STATE_THUMB)
        ? 2 : 4;
    stepbreakpoint.type = BKPT_HARD;
    stepbreakpoint.is_set = false;
 
    /* Disable interrupts during single step if requested */
    if (cortex_a->isrmasking_mode == CORTEX_A_ISRMASK_ON) {
        retval = cortex_a_set_dscr_bits(target, DSCR_INT_DIS, DSCR_INT_DIS);
        if (retval != ERROR_OK)
            return retval;
    }
 
    /* Break on IVA mismatch */
    cortex_a_set_breakpoint(target, &stepbreakpoint, 0x04);
 
    target->debug_reason = DBG_REASON_SINGLESTEP;
 
    retval = cortex_a_resume(target, true, address, false, false);
    if (retval != ERROR_OK)
        return retval;
 
    // poll at least once before starting the timeout
    retval = cortex_a_poll(target);
    if (retval != ERROR_OK)
        return retval;
 
    int64_t then = timeval_ms() + 100;
    while (target->state != TARGET_HALTED) {
        if (timeval_ms() > then)
            break;
 
        retval = cortex_a_poll(target);
        if (retval != ERROR_OK)
            return retval;
    }
 
    if (target->state != TARGET_HALTED) {
        LOG_TARGET_DEBUG(target, "timeout waiting for target halt, try halt");
 
        retval = cortex_a_halt(target);
        if (retval != ERROR_OK)
            return retval;
 
        retval = cortex_a_poll(target);
        if (retval != ERROR_OK)
            return retval;
 
        if (target->state != TARGET_HALTED) {
            LOG_TARGET_ERROR(target, "timeout waiting for target halt");
            return ERROR_FAIL;
        }
    }
 
    cortex_a_unset_breakpoint(target, &stepbreakpoint);
 
    /* Re-enable interrupts if they were disabled */
    if (cortex_a->isrmasking_mode == CORTEX_A_ISRMASK_ON) {
        retval = cortex_a_set_dscr_bits(target, DSCR_INT_DIS, 0);
        if (retval != ERROR_OK)
            return retval;
    }
 
 
    target->debug_reason = DBG_REASON_BREAKPOINT;
 
    if (breakpoint)
        cortex_a_set_breakpoint(target, breakpoint, 0);
 
    return ERROR_OK;
}
 
static int cortex_a_restore_context(struct target *target, bool bpwp)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    LOG_TARGET_DEBUG(target, " ");
 
    if (armv7a->pre_restore_context)
        armv7a->pre_restore_context(target);
 
    return arm_dpm_write_dirty_registers(&armv7a->dpm, bpwp);
}
 
/*
 * Cortex-A Breakpoint and watchpoint functions
 */
 
/* Setup hardware Breakpoint Register Pair */
static int cortex_a_set_breakpoint(struct target *target,
    struct breakpoint *breakpoint, uint8_t matchmode)
{
    int retval;
    int brp_i = 0;
    uint32_t control;
    uint8_t byte_addr_select = 0x0F;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct cortex_a_brp *brp_list = cortex_a->brp_list;
 
    if (breakpoint->is_set) {
        LOG_WARNING("breakpoint already set");
        return ERROR_OK;
    }
 
    if (breakpoint->type == BKPT_HARD) {
        while (brp_list[brp_i].used && (brp_i < cortex_a->brp_num))
            brp_i++;
        if (brp_i >= cortex_a->brp_num) {
            LOG_ERROR("ERROR Can not find free Breakpoint Register Pair");
            return ERROR_TARGET_RESOURCE_NOT_AVAILABLE;
        }
        breakpoint_hw_set(breakpoint, brp_i);
        if (breakpoint->length == 3) {
            /* Thumb-2 breakpoint: fixup to length 4 if word aligned,
             * set byte mask for length 2 if unaligned */
            if (IS_ALIGNED(breakpoint->address, 4))
                breakpoint->length = 4;
            else
                breakpoint->length = 2;
        }
        if (breakpoint->length == 2)
            byte_addr_select = (3 << (breakpoint->address & 0x02));
        control = ((matchmode & 0x7) << 20)
            | (byte_addr_select << 5)
            | (3 << 1) | 1;
        brp_list[brp_i].used = true;
        brp_list[brp_i].value = (breakpoint->address & 0xFFFFFFFC);
        brp_list[brp_i].control = control;
        retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_BVR_BASE + 4 * brp_list[brp_i].brpn,
                brp_list[brp_i].value);
        if (retval != ERROR_OK)
            return retval;
        retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_BCR_BASE + 4 * brp_list[brp_i].brpn,
                brp_list[brp_i].control);
        if (retval != ERROR_OK)
            return retval;
        LOG_DEBUG("brp %i control 0x%0" PRIx32 " value 0x%0" PRIx32, brp_i,
            brp_list[brp_i].control,
            brp_list[brp_i].value);
    } else if (breakpoint->type == BKPT_SOFT) {
        uint8_t code[4];
        if (breakpoint->length == 2) {
            /* length == 2: Thumb breakpoint */
            buf_set_u32(code, 0, 32, ARMV5_T_BKPT(0x11));
        } else if (breakpoint->length == 3) {
            /* length == 3: Thumb-2 breakpoint, actual encoding is
             * a regular Thumb BKPT instruction but we replace a
             * 32bit Thumb-2 instruction, so fix-up the breakpoint
             * length
             */
            buf_set_u32(code, 0, 32, ARMV5_T_BKPT(0x11));
            breakpoint->length = 4;
        } else {
            /* length == 4, normal ARM breakpoint */
            buf_set_u32(code, 0, 32, ARMV5_BKPT(0x11));
        }
 
        /*
         * ARMv7-A/R fetches instructions in little-endian on both LE and BE CPUs.
         * But Cortex-R4 and Cortex-R5 big-endian require BE instructions.
         * https://developer.arm.com/documentation/den0042/a/Coding-for-Cortex-R-Processors/Endianness
         * https://developer.arm.com/documentation/den0013/d/Porting/Endianness
         */
        if ((((cortex_a->cpuid & CPUDBG_CPUID_MASK) == CPUDBG_CPUID_CORTEX_R4) ||
            ((cortex_a->cpuid & CPUDBG_CPUID_MASK) == CPUDBG_CPUID_CORTEX_R5)) &&
            target->endianness == TARGET_BIG_ENDIAN) {
            // In place swapping is allowed
            buf_bswap32(code, code, 4);
        }
 
        retval = target_read_memory(target,
                breakpoint->address & 0xFFFFFFFE,
                breakpoint->length, 1,
                breakpoint->orig_instr);
        if (retval != ERROR_OK)
            return retval;
 
        /* make sure data cache is cleaned & invalidated down to PoC */
        armv7a_cache_flush_virt(target, breakpoint->address, breakpoint->length);
 
        retval = target_write_memory(target,
                breakpoint->address & 0xFFFFFFFE,
                breakpoint->length, 1, code);
        if (retval != ERROR_OK)
            return retval;
 
        /* update i-cache at breakpoint location */
        armv7a_l1_d_cache_inval_virt(target, breakpoint->address, breakpoint->length);
        armv7a_l1_i_cache_inval_virt(target, breakpoint->address, breakpoint->length);
 
        breakpoint->is_set = true;
    }
 
    return ERROR_OK;
}
 
static int cortex_a_set_context_breakpoint(struct target *target,
    struct breakpoint *breakpoint, uint8_t matchmode)
{
    int retval = ERROR_FAIL;
    int brp_i = 0;
    uint32_t control;
    uint8_t byte_addr_select = 0x0F;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct cortex_a_brp *brp_list = cortex_a->brp_list;
 
    if (breakpoint->is_set) {
        LOG_WARNING("breakpoint already set");
        return retval;
    }
    /*check available context BRPs*/
    while ((brp_list[brp_i].used ||
        (brp_list[brp_i].type != BRP_CONTEXT)) && (brp_i < cortex_a->brp_num))
        brp_i++;
 
    if (brp_i >= cortex_a->brp_num) {
        LOG_ERROR("ERROR Can not find free Breakpoint Register Pair");
        return ERROR_FAIL;
    }
 
    breakpoint_hw_set(breakpoint, brp_i);
    control = ((matchmode & 0x7) << 20)
        | (byte_addr_select << 5)
        | (3 << 1) | 1;
    brp_list[brp_i].used = true;
    brp_list[brp_i].value = (breakpoint->asid);
    brp_list[brp_i].control = control;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_BVR_BASE + 4 * brp_list[brp_i].brpn,
            brp_list[brp_i].value);
    if (retval != ERROR_OK)
        return retval;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_BCR_BASE + 4 * brp_list[brp_i].brpn,
            brp_list[brp_i].control);
    if (retval != ERROR_OK)
        return retval;
    LOG_DEBUG("brp %i control 0x%0" PRIx32 " value 0x%0" PRIx32, brp_i,
        brp_list[brp_i].control,
        brp_list[brp_i].value);
    return ERROR_OK;
 
}
 
static int cortex_a_set_hybrid_breakpoint(struct target *target, struct breakpoint *breakpoint)
{
    int retval = ERROR_FAIL;
    int brp_1 = 0;  /* holds the contextID pair */
    int brp_2 = 0;  /* holds the IVA pair */
    uint32_t control_ctx, control_iva;
    uint8_t ctx_byte_addr_select = 0x0F;
    uint8_t iva_byte_addr_select = 0x0F;
    uint8_t ctx_machmode = 0x03;
    uint8_t iva_machmode = 0x01;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct cortex_a_brp *brp_list = cortex_a->brp_list;
 
    if (breakpoint->is_set) {
        LOG_WARNING("breakpoint already set");
        return retval;
    }
    /*check available context BRPs*/
    while ((brp_list[brp_1].used ||
        (brp_list[brp_1].type != BRP_CONTEXT)) && (brp_1 < cortex_a->brp_num))
        brp_1++;
 
    LOG_DEBUG("brp(CTX) found num: %d", brp_1);
    if (brp_1 >= cortex_a->brp_num) {
        LOG_ERROR("ERROR Can not find free Breakpoint Register Pair");
        return ERROR_FAIL;
    }
 
    while ((brp_list[brp_2].used ||
        (brp_list[brp_2].type != BRP_NORMAL)) && (brp_2 < cortex_a->brp_num))
        brp_2++;
 
    LOG_DEBUG("brp(IVA) found num: %d", brp_2);
    if (brp_2 >= cortex_a->brp_num) {
        LOG_ERROR("ERROR Can not find free Breakpoint Register Pair");
        return ERROR_FAIL;
    }
 
    breakpoint_hw_set(breakpoint, brp_1);
    breakpoint->linked_brp = brp_2;
    control_ctx = ((ctx_machmode & 0x7) << 20)
        | (brp_2 << 16)
        | (0 << 14)
        | (ctx_byte_addr_select << 5)
        | (3 << 1) | 1;
    brp_list[brp_1].used = true;
    brp_list[brp_1].value = (breakpoint->asid);
    brp_list[brp_1].control = control_ctx;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_BVR_BASE + 4 * brp_list[brp_1].brpn,
            brp_list[brp_1].value);
    if (retval != ERROR_OK)
        return retval;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_BCR_BASE + 4 * brp_list[brp_1].brpn,
            brp_list[brp_1].control);
    if (retval != ERROR_OK)
        return retval;
 
    control_iva = ((iva_machmode & 0x7) << 20)
        | (brp_1 << 16)
        | (iva_byte_addr_select << 5)
        | (3 << 1) | 1;
    brp_list[brp_2].used = true;
    brp_list[brp_2].value = (breakpoint->address & 0xFFFFFFFC);
    brp_list[brp_2].control = control_iva;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_BVR_BASE + 4 * brp_list[brp_2].brpn,
            brp_list[brp_2].value);
    if (retval != ERROR_OK)
        return retval;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_BCR_BASE + 4 * brp_list[brp_2].brpn,
            brp_list[brp_2].control);
    if (retval != ERROR_OK)
        return retval;
 
    return ERROR_OK;
}
 
static int cortex_a_unset_breakpoint(struct target *target, struct breakpoint *breakpoint)
{
    int retval;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct cortex_a_brp *brp_list = cortex_a->brp_list;
 
    if (!breakpoint->is_set) {
        LOG_WARNING("breakpoint not set");
        return ERROR_OK;
    }
 
    if (breakpoint->type == BKPT_HARD) {
        if ((breakpoint->address != 0) && (breakpoint->asid != 0)) {
            int brp_i = breakpoint->number;
            int brp_j = breakpoint->linked_brp;
            if (brp_i >= cortex_a->brp_num) {
                LOG_DEBUG("Invalid BRP number in breakpoint");
                return ERROR_OK;
            }
            LOG_DEBUG("rbp %i control 0x%0" PRIx32 " value 0x%0" PRIx32, brp_i,
                brp_list[brp_i].control, brp_list[brp_i].value);
            brp_list[brp_i].used = false;
            brp_list[brp_i].value = 0;
            brp_list[brp_i].control = 0;
            retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_BCR_BASE + 4 * brp_list[brp_i].brpn,
                    brp_list[brp_i].control);
            if (retval != ERROR_OK)
                return retval;
            retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_BVR_BASE + 4 * brp_list[brp_i].brpn,
                    brp_list[brp_i].value);
            if (retval != ERROR_OK)
                return retval;
            if ((brp_j < 0) || (brp_j >= cortex_a->brp_num)) {
                LOG_DEBUG("Invalid BRP number in breakpoint");
                return ERROR_OK;
            }
            LOG_DEBUG("rbp %i control 0x%0" PRIx32 " value 0x%0" PRIx32, brp_j,
                brp_list[brp_j].control, brp_list[brp_j].value);
            brp_list[brp_j].used = false;
            brp_list[brp_j].value = 0;
            brp_list[brp_j].control = 0;
            retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_BCR_BASE + 4 * brp_list[brp_j].brpn,
                    brp_list[brp_j].control);
            if (retval != ERROR_OK)
                return retval;
            retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_BVR_BASE + 4 * brp_list[brp_j].brpn,
                    brp_list[brp_j].value);
            if (retval != ERROR_OK)
                return retval;
            breakpoint->linked_brp = 0;
            breakpoint->is_set = false;
            return ERROR_OK;
 
        } else {
            int brp_i = breakpoint->number;
            if (brp_i >= cortex_a->brp_num) {
                LOG_DEBUG("Invalid BRP number in breakpoint");
                return ERROR_OK;
            }
            LOG_DEBUG("rbp %i control 0x%0" PRIx32 " value 0x%0" PRIx32, brp_i,
                brp_list[brp_i].control, brp_list[brp_i].value);
            brp_list[brp_i].used = false;
            brp_list[brp_i].value = 0;
            brp_list[brp_i].control = 0;
            retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_BCR_BASE + 4 * brp_list[brp_i].brpn,
                    brp_list[brp_i].control);
            if (retval != ERROR_OK)
                return retval;
            retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_BVR_BASE + 4 * brp_list[brp_i].brpn,
                    brp_list[brp_i].value);
            if (retval != ERROR_OK)
                return retval;
            breakpoint->is_set = false;
            return ERROR_OK;
        }
    } else {
 
        /* make sure data cache is cleaned & invalidated down to PoC */
        armv7a_cache_flush_virt(target, breakpoint->address,
                        breakpoint->length);
 
        /* restore original instruction (kept in target endianness) */
        if (breakpoint->length == 4) {
            retval = target_write_memory(target,
                    breakpoint->address & 0xFFFFFFFE,
                    4, 1, breakpoint->orig_instr);
            if (retval != ERROR_OK)
                return retval;
        } else {
            retval = target_write_memory(target,
                    breakpoint->address & 0xFFFFFFFE,
                    2, 1, breakpoint->orig_instr);
            if (retval != ERROR_OK)
                return retval;
        }
 
        /* update i-cache at breakpoint location */
        armv7a_l1_d_cache_inval_virt(target, breakpoint->address,
                         breakpoint->length);
        armv7a_l1_i_cache_inval_virt(target, breakpoint->address,
                         breakpoint->length);
    }
    breakpoint->is_set = false;
 
    return ERROR_OK;
}
 
static int cortex_a_add_breakpoint(struct target *target,
    struct breakpoint *breakpoint)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    if ((breakpoint->type == BKPT_HARD) && (cortex_a->brp_num_available < 1)) {
        LOG_INFO("no hardware breakpoint available");
        return ERROR_TARGET_RESOURCE_NOT_AVAILABLE;
    }
 
    if (breakpoint->type == BKPT_HARD)
        cortex_a->brp_num_available--;
 
    return cortex_a_set_breakpoint(target, breakpoint, 0x00);   /* Exact match */
}
 
static int cortex_a_add_context_breakpoint(struct target *target,
    struct breakpoint *breakpoint)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    if ((breakpoint->type == BKPT_HARD) && (cortex_a->brp_num_available < 1)) {
        LOG_INFO("no hardware breakpoint available");
        return ERROR_TARGET_RESOURCE_NOT_AVAILABLE;
    }
 
    if (breakpoint->type == BKPT_HARD)
        cortex_a->brp_num_available--;
 
    return cortex_a_set_context_breakpoint(target, breakpoint, 0x02);   /* asid match */
}
 
static int cortex_a_add_hybrid_breakpoint(struct target *target,
    struct breakpoint *breakpoint)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    if ((breakpoint->type == BKPT_HARD) && (cortex_a->brp_num_available < 1)) {
        LOG_INFO("no hardware breakpoint available");
        return ERROR_TARGET_RESOURCE_NOT_AVAILABLE;
    }
 
    if (breakpoint->type == BKPT_HARD)
        cortex_a->brp_num_available--;
 
    return cortex_a_set_hybrid_breakpoint(target, breakpoint);  /* ??? */
}
 
 
static int cortex_a_remove_breakpoint(struct target *target, struct breakpoint *breakpoint)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
#if 0
/* It is perfectly possible to remove breakpoints while the target is running */
    if (target->state != TARGET_HALTED) {
        LOG_WARNING("target not halted");
        return ERROR_TARGET_NOT_HALTED;
    }
#endif
 
    if (breakpoint->is_set) {
        cortex_a_unset_breakpoint(target, breakpoint);
        if (breakpoint->type == BKPT_HARD)
            cortex_a->brp_num_available++;
    }
 
 
    return ERROR_OK;
}
 
static int cortex_a_set_watchpoint(struct target *target, struct watchpoint *watchpoint)
{
    int retval = ERROR_OK;
    int wrp_i = 0;
    uint32_t control;
    uint32_t address;
    uint8_t address_mask;
    uint8_t byte_address_select;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct cortex_a_wrp *wrp_list = cortex_a->wrp_list;
 
    if (watchpoint->is_set) {
        LOG_WARNING("watchpoint already set");
        return retval;
    }
 
    /* check available context WRPs */
    while (wrp_list[wrp_i].used && (wrp_i < cortex_a->wrp_num))
        wrp_i++;
 
    if (wrp_i >= cortex_a->wrp_num) {
        LOG_ERROR("ERROR Can not find free Watchpoint Register Pair");
        return ERROR_FAIL;
    }
 
    if (watchpoint->length == 0 || watchpoint->length > 0x80000000U ||
            (watchpoint->length & (watchpoint->length - 1))) {
        LOG_WARNING("watchpoint length must be a power of 2");
        return ERROR_FAIL;
    }
 
    if (watchpoint->address & (watchpoint->length - 1)) {
        LOG_WARNING("watchpoint address must be aligned at length");
        return ERROR_FAIL;
    }
 
    /* FIXME: ARM DDI 0406C: address_mask is optional. What to do if it's missing?  */
    /* handle wp length 1 and 2 through byte select */
    switch (watchpoint->length) {
    case 1:
        byte_address_select = BIT(watchpoint->address & 0x3);
        address = watchpoint->address & ~0x3;
        address_mask = 0;
        break;
 
    case 2:
        byte_address_select = 0x03 << (watchpoint->address & 0x2);
        address = watchpoint->address & ~0x3;
        address_mask = 0;
        break;
 
    case 4:
        byte_address_select = 0x0f;
        address = watchpoint->address;
        address_mask = 0;
        break;
 
    default:
        byte_address_select = 0xff;
        address = watchpoint->address;
        address_mask = ilog2(watchpoint->length);
        break;
    }
 
    uint8_t load_store_access_control;
    switch (watchpoint->rw) {
    case WPT_READ:
        load_store_access_control = 1;
        break;
    case WPT_WRITE:
        load_store_access_control = 2;
        break;
    case WPT_ACCESS:
        load_store_access_control = 3;
        break;
    default:
        LOG_ERROR("BUG: watchpoint->rw neither read, write nor access");
        return ERROR_FAIL;
    };
 
    watchpoint_set(watchpoint, wrp_i);
    control = (address_mask << 24) |
        (byte_address_select << 5) |
        (load_store_access_control << 3) |
        (0x3 << 1) | 1;
    wrp_list[wrp_i].used = true;
    wrp_list[wrp_i].value = address;
    wrp_list[wrp_i].control = control;
 
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_WVR_BASE + 4 * wrp_list[wrp_i].wrpn,
            wrp_list[wrp_i].value);
    if (retval != ERROR_OK)
        return retval;
 
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_WCR_BASE + 4 * wrp_list[wrp_i].wrpn,
            wrp_list[wrp_i].control);
    if (retval != ERROR_OK)
        return retval;
 
    LOG_DEBUG("wp %i control 0x%0" PRIx32 " value 0x%0" PRIx32, wrp_i,
            wrp_list[wrp_i].control,
            wrp_list[wrp_i].value);
 
    return ERROR_OK;
}
 
static int cortex_a_unset_watchpoint(struct target *target, struct watchpoint *watchpoint)
{
    int retval;
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct cortex_a_wrp *wrp_list = cortex_a->wrp_list;
 
    if (!watchpoint->is_set) {
        LOG_WARNING("watchpoint not set");
        return ERROR_OK;
    }
 
    int wrp_i = watchpoint->number;
    if (wrp_i >= cortex_a->wrp_num) {
        LOG_DEBUG("Invalid WRP number in watchpoint");
        return ERROR_OK;
    }
    LOG_DEBUG("wrp %i control 0x%0" PRIx32 " value 0x%0" PRIx32, wrp_i,
            wrp_list[wrp_i].control, wrp_list[wrp_i].value);
    wrp_list[wrp_i].used = false;
    wrp_list[wrp_i].value = 0;
    wrp_list[wrp_i].control = 0;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_WCR_BASE + 4 * wrp_list[wrp_i].wrpn,
            wrp_list[wrp_i].control);
    if (retval != ERROR_OK)
        return retval;
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_WVR_BASE + 4 * wrp_list[wrp_i].wrpn,
            wrp_list[wrp_i].value);
    if (retval != ERROR_OK)
        return retval;
    watchpoint->is_set = false;
 
    return ERROR_OK;
}
 
static int cortex_a_add_watchpoint(struct target *target, struct watchpoint *watchpoint)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    if (cortex_a->wrp_num_available < 1) {
        LOG_INFO("no hardware watchpoint available");
        return ERROR_TARGET_RESOURCE_NOT_AVAILABLE;
    }
 
    int retval = cortex_a_set_watchpoint(target, watchpoint);
    if (retval != ERROR_OK)
        return retval;
 
    cortex_a->wrp_num_available--;
    return ERROR_OK;
}
 
static int cortex_a_remove_watchpoint(struct target *target, struct watchpoint *watchpoint)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    if (watchpoint->is_set) {
        cortex_a->wrp_num_available++;
        cortex_a_unset_watchpoint(target, watchpoint);
    }
    return ERROR_OK;
}
 
 
/*
 * Cortex-A Reset functions
 */
 
static int cortex_a_assert_reset(struct target *target)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    LOG_DEBUG(" ");
 
    /* FIXME when halt is requested, make it work somehow... */
 
    /* This function can be called in "target not examined" state */
 
    /* Issue some kind of warm reset. */
    if (target_has_event_action(target, TARGET_EVENT_RESET_ASSERT))
        target_handle_event(target, TARGET_EVENT_RESET_ASSERT);
    else if (jtag_get_reset_config() & RESET_HAS_SRST) {
        /* REVISIT handle "pulls" cases, if there's
         * hardware that needs them to work.
         */
 
        /*
         * FIXME: fix reset when transport is not JTAG. This is a temporary
         * work-around for release v0.10 that is not intended to stay!
         */
        if (!transport_is_jtag() ||
                (target->reset_halt && (jtag_get_reset_config() & RESET_SRST_NO_GATING)))
            adapter_assert_reset();
 
    } else {
        LOG_ERROR("%s: how to reset?", target_name(target));
        return ERROR_FAIL;
    }
 
    /* registers are now invalid */
    if (armv7a->arm.core_cache)
        register_cache_invalidate(armv7a->arm.core_cache);
 
    target->state = TARGET_RESET;
 
    return ERROR_OK;
}
 
static int cortex_a_deassert_reset(struct target *target)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
    int retval;
 
    LOG_DEBUG(" ");
 
    /* be certain SRST is off */
    adapter_deassert_reset();
 
    if (target_was_examined(target)) {
        retval = cortex_a_poll(target);
        if (retval != ERROR_OK)
            return retval;
    }
 
    if (target->reset_halt) {
        if (target->state != TARGET_HALTED) {
            LOG_WARNING("%s: ran after reset and before halt ...",
                target_name(target));
            if (target_was_examined(target)) {
                retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                        armv7a->debug_base + CPUDBG_DRCR, DRCR_HALT);
                if (retval != ERROR_OK)
                    return retval;
            } else
                target->state = TARGET_UNKNOWN;
        }
    }
 
    return ERROR_OK;
}
 
static int cortex_a_set_dcc_mode(struct target *target, uint32_t mode, uint32_t *dscr)
{
    /* Changes the mode of the DCC between non-blocking, stall, and fast mode.
     * New desired mode must be in mode. Current value of DSCR must be in
     * *dscr, which is updated with new value.
     *
     * This function elides actually sending the mode-change over the debug
     * interface if the mode is already set as desired.
     */
    uint32_t new_dscr = (*dscr & ~DSCR_EXT_DCC_MASK) | mode;
    if (new_dscr != *dscr) {
        struct armv7a_common *armv7a = target_to_armv7a(target);
        int retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DSCR, new_dscr);
        if (retval == ERROR_OK)
            *dscr = new_dscr;
        return retval;
    } else {
        return ERROR_OK;
    }
}
 
static int cortex_a_wait_dscr_bits(struct target *target, uint32_t mask,
    uint32_t value, uint32_t *dscr)
{
    /* Waits until the specified bit(s) of DSCR take on a specified value. */
    struct armv7a_common *armv7a = target_to_armv7a(target);
    int64_t then;
    int retval;
 
    if ((*dscr & mask) == value)
        return ERROR_OK;
 
    then = timeval_ms();
    while (1) {
        retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DSCR, dscr);
        if (retval != ERROR_OK) {
            LOG_ERROR("Could not read DSCR register");
            return retval;
        }
        if ((*dscr & mask) == value)
            break;
        if (timeval_ms() > then + 1000) {
            LOG_ERROR("timeout waiting for DSCR bit change");
            return ERROR_FAIL;
        }
    }
    return ERROR_OK;
}
 
static int cortex_a_read_copro(struct target *target, uint32_t opcode,
    uint32_t *data, uint32_t *dscr)
{
    int retval;
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    /* Move from coprocessor to R0. */
    retval = cortex_a_exec_opcode(target, opcode, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Move from R0 to DTRTX. */
    retval = cortex_a_exec_opcode(target, ARMV4_5_MCR(14, 0, 0, 0, 5, 0), dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Wait until DTRTX is full (according to ARMv7-A/-R architecture
     * manual section C8.4.3, checking InstrCmpl_l is not sufficient; one
     * must also check TXfull_l). Most of the time this will be free
     * because TXfull_l will be set immediately and cached in dscr. */
    retval = cortex_a_wait_dscr_bits(target, DSCR_DTRTX_FULL_LATCHED,
            DSCR_DTRTX_FULL_LATCHED, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Read the value transferred to DTRTX. */
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DTRTX, data);
    if (retval != ERROR_OK)
        return retval;
 
    return ERROR_OK;
}
 
static int cortex_a_read_dfar_dfsr(struct target *target, uint32_t *dfar,
    uint32_t *dfsr, uint32_t *dscr)
{
    int retval;
 
    if (dfar) {
        retval = cortex_a_read_copro(target, ARMV4_5_MRC(15, 0, 0, 6, 0, 0), dfar, dscr);
        if (retval != ERROR_OK)
            return retval;
    }
 
    if (dfsr) {
        retval = cortex_a_read_copro(target, ARMV4_5_MRC(15, 0, 0, 5, 0, 0), dfsr, dscr);
        if (retval != ERROR_OK)
            return retval;
    }
 
    return ERROR_OK;
}
 
static int cortex_a_write_copro(struct target *target, uint32_t opcode,
    uint32_t data, uint32_t *dscr)
{
    int retval;
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    /* Write the value into DTRRX. */
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DTRRX, data);
    if (retval != ERROR_OK)
        return retval;
 
    /* Move from DTRRX to R0. */
    retval = cortex_a_exec_opcode(target, ARMV4_5_MRC(14, 0, 0, 0, 5, 0), dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Move from R0 to coprocessor. */
    retval = cortex_a_exec_opcode(target, opcode, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Wait until DTRRX is empty (according to ARMv7-A/-R architecture manual
     * section C8.4.3, checking InstrCmpl_l is not sufficient; one must also
     * check RXfull_l). Most of the time this will be free because RXfull_l
     * will be cleared immediately and cached in dscr. */
    retval = cortex_a_wait_dscr_bits(target, DSCR_DTRRX_FULL_LATCHED, 0, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    return ERROR_OK;
}
 
static int cortex_a_write_dfar_dfsr(struct target *target, uint32_t dfar,
    uint32_t dfsr, uint32_t *dscr)
{
    int retval;
 
    retval = cortex_a_write_copro(target, ARMV4_5_MCR(15, 0, 0, 6, 0, 0), dfar, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    retval = cortex_a_write_copro(target, ARMV4_5_MCR(15, 0, 0, 5, 0, 0), dfsr, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    return ERROR_OK;
}
 
static int cortex_a_dfsr_to_error_code(uint32_t dfsr)
{
    uint32_t status, upper4;
 
    if (dfsr & (1 << 9)) {
        /* LPAE format. */
        status = dfsr & 0x3f;
        upper4 = status >> 2;
        if (upper4 == 1 || upper4 == 2 || upper4 == 3 || upper4 == 15)
            return ERROR_TARGET_TRANSLATION_FAULT;
        else if (status == 33)
            return ERROR_TARGET_UNALIGNED_ACCESS;
        else
            return ERROR_TARGET_DATA_ABORT;
    } else {
        /* Normal format. */
        status = ((dfsr >> 6) & 0x10) | (dfsr & 0xf);
        if (status == 1)
            return ERROR_TARGET_UNALIGNED_ACCESS;
        else if (status == 5 || status == 7 || status == 3 || status == 6 ||
                status == 9 || status == 11 || status == 13 || status == 15)
            return ERROR_TARGET_TRANSLATION_FAULT;
        else
            return ERROR_TARGET_DATA_ABORT;
    }
}
 
static int cortex_a_write_cpu_memory_slow(struct target *target,
    uint32_t size, uint32_t count, const uint8_t *buffer, uint32_t *dscr)
{
    /* Writes count objects of size size from *buffer. Old value of DSCR must
     * be in *dscr; updated to new value. This is slow because it works for
     * non-word-sized objects. Avoid unaligned accesses as they do not work
     * on memory address space without "Normal" attribute. If size == 4 and
     * the address is aligned, cortex_a_write_cpu_memory_fast should be
     * preferred.
     * Preconditions:
     * - Address is in R0.
     * - R0 is marked dirty.
     */
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
    int retval;
 
    /* Mark register R1 as dirty, to use for transferring data. */
    arm_reg_current(arm, 1)->dirty = true;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Go through the objects. */
    while (count) {
        /* Write the value to store into DTRRX. */
        uint32_t data, opcode;
        if (size == 1)
            data = *buffer;
        else if (size == 2)
            data = target_buffer_get_u16(target, buffer);
        else
            data = target_buffer_get_u32(target, buffer);
        retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DTRRX, data);
        if (retval != ERROR_OK)
            return retval;
 
        /* Transfer the value from DTRRX to R1. */
        retval = cortex_a_exec_opcode(target, ARMV4_5_MRC(14, 0, 1, 0, 5, 0), dscr);
        if (retval != ERROR_OK)
            return retval;
 
        /* Write the value transferred to R1 into memory. */
        if (size == 1)
            opcode = ARMV4_5_STRB_IP(1, 0);
        else if (size == 2)
            opcode = ARMV4_5_STRH_IP(1, 0);
        else
            opcode = ARMV4_5_STRW_IP(1, 0);
        retval = cortex_a_exec_opcode(target, opcode, dscr);
        if (retval != ERROR_OK)
            return retval;
 
        /* Check for faults and return early. */
        if (*dscr & (DSCR_STICKY_ABORT_PRECISE | DSCR_STICKY_ABORT_IMPRECISE))
            return ERROR_OK; /* A data fault is not considered a system failure. */
 
        /* Wait until DTRRX is empty (according to ARMv7-A/-R architecture
         * manual section C8.4.3, checking InstrCmpl_l is not sufficient; one
         * must also check RXfull_l). Most of the time this will be free
         * because RXfull_l will be cleared immediately and cached in dscr. */
        retval = cortex_a_wait_dscr_bits(target, DSCR_DTRRX_FULL_LATCHED, 0, dscr);
        if (retval != ERROR_OK)
            return retval;
 
        /* Advance. */
        buffer += size;
        --count;
    }
 
    return ERROR_OK;
}
 
static int cortex_a_write_cpu_memory_fast(struct target *target,
    uint32_t count, const uint8_t *buffer, uint32_t *dscr)
{
    /* Writes count objects of size 4 from *buffer. Old value of DSCR must be
     * in *dscr; updated to new value. This is fast but only works for
     * word-sized objects at aligned addresses.
     * Preconditions:
     * - Address is in R0 and must be a multiple of 4.
     * - R0 is marked dirty.
     */
    struct armv7a_common *armv7a = target_to_armv7a(target);
    int retval;
 
    /* Switch to fast mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_FAST_MODE, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Latch STC instruction. */
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_ITR, ARMV4_5_STC(0, 1, 0, 1, 14, 5, 0, 4));
    if (retval != ERROR_OK)
        return retval;
 
    /* Transfer all the data and issue all the instructions. */
    return mem_ap_write_buf_noincr(armv7a->debug_ap, buffer,
            4, count, armv7a->debug_base + CPUDBG_DTRRX);
}
 
static int cortex_a_write_cpu_memory(struct target *target,
    uint32_t address, uint32_t size,
    uint32_t count, const uint8_t *buffer)
{
    /* Write memory through the CPU. */
    int retval, final_retval;
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
    uint32_t dscr, orig_dfar, orig_dfsr, fault_dscr, fault_dfar, fault_dfsr;
 
    LOG_DEBUG("Writing CPU memory address 0x%" PRIx32 " size %"  PRIu32 " count %"  PRIu32,
              address, size, count);
    if (target->state != TARGET_HALTED) {
        LOG_TARGET_ERROR(target, "not halted");
        return ERROR_TARGET_NOT_HALTED;
    }
 
    if (!count)
        return ERROR_OK;
 
    /* Clear any abort. */
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DRCR, DRCR_CLEAR_EXCEPTIONS);
    if (retval != ERROR_OK)
        return retval;
 
    /* Read DSCR. */
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Mark R0 as dirty. */
    arm_reg_current(arm, 0)->dirty = true;
 
    /* Read DFAR and DFSR, as they will be modified in the event of a fault. */
    retval = cortex_a_read_dfar_dfsr(target, &orig_dfar, &orig_dfsr, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Get the memory address into R0. */
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DTRRX, address);
    if (retval != ERROR_OK)
        return retval;
    retval = cortex_a_exec_opcode(target, ARMV4_5_MRC(14, 0, 0, 0, 5, 0), &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    if (size == 4 && (address % 4) == 0) {
        /* We are doing a word-aligned transfer, so use fast mode. */
        retval = cortex_a_write_cpu_memory_fast(target, count, buffer, &dscr);
    } else {
        /* Use slow path. Adjust size for aligned accesses */
        switch (address % 4) {
        case 1:
        case 3:
            count *= size;
            size = 1;
            break;
        case 2:
            if (size == 4) {
                count *= 2;
                size = 2;
            }
            break;
        case 0:
        default:
            break;
        }
        retval = cortex_a_write_cpu_memory_slow(target, size, count, buffer, &dscr);
    }
 
    final_retval = retval;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, &dscr);
    if (final_retval == ERROR_OK)
        final_retval = retval;
 
    /* Wait for last issued instruction to complete. */
    retval = cortex_a_wait_instrcmpl(target, &dscr, true);
    if (final_retval == ERROR_OK)
        final_retval = retval;
 
    /* Wait until DTRRX is empty (according to ARMv7-A/-R architecture manual
     * section C8.4.3, checking InstrCmpl_l is not sufficient; one must also
     * check RXfull_l). Most of the time this will be free because RXfull_l
     * will be cleared immediately and cached in dscr. However, don't do this
     * if there is fault, because then the instruction might not have completed
     * successfully. */
    if (!(dscr & DSCR_STICKY_ABORT_PRECISE)) {
        retval = cortex_a_wait_dscr_bits(target, DSCR_DTRRX_FULL_LATCHED, 0, &dscr);
        if (retval != ERROR_OK)
            return retval;
    }
 
    /* If there were any sticky abort flags, clear them. */
    if (dscr & (DSCR_STICKY_ABORT_PRECISE | DSCR_STICKY_ABORT_IMPRECISE)) {
        fault_dscr = dscr;
        mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DRCR, DRCR_CLEAR_EXCEPTIONS);
        dscr &= ~(DSCR_STICKY_ABORT_PRECISE | DSCR_STICKY_ABORT_IMPRECISE);
    } else {
        fault_dscr = 0;
    }
 
    /* Handle synchronous data faults. */
    if (fault_dscr & DSCR_STICKY_ABORT_PRECISE) {
        if (final_retval == ERROR_OK) {
            /* Final return value will reflect cause of fault. */
            retval = cortex_a_read_dfar_dfsr(target, &fault_dfar, &fault_dfsr, &dscr);
            if (retval == ERROR_OK) {
                LOG_ERROR("data abort at 0x%08" PRIx32 ", dfsr = 0x%08" PRIx32, fault_dfar, fault_dfsr);
                final_retval = cortex_a_dfsr_to_error_code(fault_dfsr);
            } else
                final_retval = retval;
        }
        /* Fault destroyed DFAR/DFSR; restore them. */
        retval = cortex_a_write_dfar_dfsr(target, orig_dfar, orig_dfsr, &dscr);
        if (retval != ERROR_OK)
            LOG_ERROR("error restoring dfar/dfsr - dscr = 0x%08" PRIx32, dscr);
    }
 
    /* Handle asynchronous data faults. */
    if (fault_dscr & DSCR_STICKY_ABORT_IMPRECISE) {
        if (final_retval == ERROR_OK)
            /* No other error has been recorded so far, so keep this one. */
            final_retval = ERROR_TARGET_DATA_ABORT;
    }
 
    /* If the DCC is nonempty, clear it. */
    if (dscr & DSCR_DTRTX_FULL_LATCHED) {
        uint32_t dummy;
        retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DTRTX, &dummy);
        if (final_retval == ERROR_OK)
            final_retval = retval;
    }
    if (dscr & DSCR_DTRRX_FULL_LATCHED) {
        retval = cortex_a_exec_opcode(target, ARMV4_5_MRC(14, 0, 1, 0, 5, 0), &dscr);
        if (final_retval == ERROR_OK)
            final_retval = retval;
    }
 
    /* Done. */
    return final_retval;
}
 
static int cortex_a_read_cpu_memory_slow(struct target *target,
    uint32_t size, uint32_t count, uint8_t *buffer, uint32_t *dscr)
{
    /* Reads count objects of size size into *buffer. Old value of DSCR must be
     * in *dscr; updated to new value. This is slow because it works for
     * non-word-sized objects. Avoid unaligned accesses as they do not work
     * on memory address space without "Normal" attribute. If size == 4 and
     * the address is aligned, cortex_a_read_cpu_memory_fast should be
     * preferred.
     * Preconditions:
     * - Address is in R0.
     * - R0 is marked dirty.
     */
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
    int retval;
 
    /* Mark register R1 as dirty, to use for transferring data. */
    arm_reg_current(arm, 1)->dirty = true;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Go through the objects. */
    while (count) {
        /* Issue a load of the appropriate size to R1. */
        uint32_t opcode, data;
        if (size == 1)
            opcode = ARMV4_5_LDRB_IP(1, 0);
        else if (size == 2)
            opcode = ARMV4_5_LDRH_IP(1, 0);
        else
            opcode = ARMV4_5_LDRW_IP(1, 0);
        retval = cortex_a_exec_opcode(target, opcode, dscr);
        if (retval != ERROR_OK)
            return retval;
 
        /* Issue a write of R1 to DTRTX. */
        retval = cortex_a_exec_opcode(target, ARMV4_5_MCR(14, 0, 1, 0, 5, 0), dscr);
        if (retval != ERROR_OK)
            return retval;
 
        /* Check for faults and return early. */
        if (*dscr & (DSCR_STICKY_ABORT_PRECISE | DSCR_STICKY_ABORT_IMPRECISE))
            return ERROR_OK; /* A data fault is not considered a system failure. */
 
        /* Wait until DTRTX is full (according to ARMv7-A/-R architecture
         * manual section C8.4.3, checking InstrCmpl_l is not sufficient; one
         * must also check TXfull_l). Most of the time this will be free
         * because TXfull_l will be set immediately and cached in dscr. */
        retval = cortex_a_wait_dscr_bits(target, DSCR_DTRTX_FULL_LATCHED,
                DSCR_DTRTX_FULL_LATCHED, dscr);
        if (retval != ERROR_OK)
            return retval;
 
        /* Read the value transferred to DTRTX into the buffer. */
        retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DTRTX, &data);
        if (retval != ERROR_OK)
            return retval;
        if (size == 1)
            *buffer = (uint8_t) data;
        else if (size == 2)
            target_buffer_set_u16(target, buffer, (uint16_t) data);
        else
            target_buffer_set_u32(target, buffer, data);
 
        /* Advance. */
        buffer += size;
        --count;
    }
 
    return ERROR_OK;
}
 
static int cortex_a_read_cpu_memory_fast(struct target *target,
    uint32_t count, uint8_t *buffer, uint32_t *dscr)
{
    /* Reads count objects of size 4 into *buffer. Old value of DSCR must be in
     * *dscr; updated to new value. This is fast but only works for word-sized
     * objects at aligned addresses.
     * Preconditions:
     * - Address is in R0 and must be a multiple of 4.
     * - R0 is marked dirty.
     */
    struct armv7a_common *armv7a = target_to_armv7a(target);
    uint32_t u32;
    int retval;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Issue the LDC instruction via a write to ITR. */
    retval = cortex_a_exec_opcode(target, ARMV4_5_LDC(0, 1, 0, 1, 14, 5, 0, 4), dscr);
    if (retval != ERROR_OK)
        return retval;
 
    count--;
 
    if (count > 0) {
        /* Switch to fast mode if not already in that mode. */
        retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_FAST_MODE, dscr);
        if (retval != ERROR_OK)
            return retval;
 
        /* Latch LDC instruction. */
        retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_ITR, ARMV4_5_LDC(0, 1, 0, 1, 14, 5, 0, 4));
        if (retval != ERROR_OK)
            return retval;
 
        /* Read the value transferred to DTRTX into the buffer. Due to fast
         * mode rules, this blocks until the instruction finishes executing and
         * then reissues the read instruction to read the next word from
         * memory. The last read of DTRTX in this call reads the second-to-last
         * word from memory and issues the read instruction for the last word.
         */
        retval = mem_ap_read_buf_noincr(armv7a->debug_ap, buffer,
                4, count, armv7a->debug_base + CPUDBG_DTRTX);
        if (retval != ERROR_OK)
            return retval;
 
        /* Advance. */
        buffer += count * 4;
    }
 
    /* Wait for last issued instruction to complete. */
    retval = cortex_a_wait_instrcmpl(target, dscr, false);
    if (retval != ERROR_OK)
        return retval;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Check for faults and return early. */
    if (*dscr & (DSCR_STICKY_ABORT_PRECISE | DSCR_STICKY_ABORT_IMPRECISE))
        return ERROR_OK; /* A data fault is not considered a system failure. */
 
    /* Wait until DTRTX is full (according to ARMv7-A/-R architecture manual
     * section C8.4.3, checking InstrCmpl_l is not sufficient; one must also
     * check TXfull_l). Most of the time this will be free because TXfull_l
     * will be set immediately and cached in dscr. */
    retval = cortex_a_wait_dscr_bits(target, DSCR_DTRTX_FULL_LATCHED,
            DSCR_DTRTX_FULL_LATCHED, dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Read the value transferred to DTRTX into the buffer. This is the last
     * word. */
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DTRTX, &u32);
    if (retval != ERROR_OK)
        return retval;
    target_buffer_set_u32(target, buffer, u32);
 
    return ERROR_OK;
}
 
static int cortex_a_read_cpu_memory(struct target *target,
    uint32_t address, uint32_t size,
    uint32_t count, uint8_t *buffer)
{
    /* Read memory through the CPU. */
    int retval, final_retval;
    struct armv7a_common *armv7a = target_to_armv7a(target);
    struct arm *arm = &armv7a->arm;
    uint32_t dscr, orig_dfar, orig_dfsr, fault_dscr, fault_dfar, fault_dfsr;
 
    LOG_DEBUG("Reading CPU memory address 0x%" PRIx32 " size %"  PRIu32 " count %"  PRIu32,
              address, size, count);
    if (target->state != TARGET_HALTED) {
        LOG_TARGET_ERROR(target, "not halted");
        return ERROR_TARGET_NOT_HALTED;
    }
 
    if (!count)
        return ERROR_OK;
 
    /* Clear any abort. */
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DRCR, DRCR_CLEAR_EXCEPTIONS);
    if (retval != ERROR_OK)
        return retval;
 
    /* Read DSCR */
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DSCR, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Mark R0 as dirty. */
    arm_reg_current(arm, 0)->dirty = true;
 
    /* Read DFAR and DFSR, as they will be modified in the event of a fault. */
    retval = cortex_a_read_dfar_dfsr(target, &orig_dfar, &orig_dfsr, &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    /* Get the memory address into R0. */
    retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DTRRX, address);
    if (retval != ERROR_OK)
        return retval;
    retval = cortex_a_exec_opcode(target, ARMV4_5_MRC(14, 0, 0, 0, 5, 0), &dscr);
    if (retval != ERROR_OK)
        return retval;
 
    if (size == 4 && (address % 4) == 0) {
        /* We are doing a word-aligned transfer, so use fast mode. */
        retval = cortex_a_read_cpu_memory_fast(target, count, buffer, &dscr);
    } else {
        /* Use slow path. Adjust size for aligned accesses */
        switch (address % 4) {
        case 1:
        case 3:
            count *= size;
            size = 1;
            break;
        case 2:
            if (size == 4) {
                count *= 2;
                size = 2;
            }
            break;
        case 0:
        default:
            break;
        }
        retval = cortex_a_read_cpu_memory_slow(target, size, count, buffer, &dscr);
    }
 
    final_retval = retval;
 
    /* Switch to non-blocking mode if not already in that mode. */
    retval = cortex_a_set_dcc_mode(target, DSCR_EXT_DCC_NON_BLOCKING, &dscr);
    if (final_retval == ERROR_OK)
        final_retval = retval;
 
    /* Wait for last issued instruction to complete. */
    retval = cortex_a_wait_instrcmpl(target, &dscr, true);
    if (final_retval == ERROR_OK)
        final_retval = retval;
 
    /* If there were any sticky abort flags, clear them. */
    if (dscr & (DSCR_STICKY_ABORT_PRECISE | DSCR_STICKY_ABORT_IMPRECISE)) {
        fault_dscr = dscr;
        mem_ap_write_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DRCR, DRCR_CLEAR_EXCEPTIONS);
        dscr &= ~(DSCR_STICKY_ABORT_PRECISE | DSCR_STICKY_ABORT_IMPRECISE);
    } else {
        fault_dscr = 0;
    }
 
    /* Handle synchronous data faults. */
    if (fault_dscr & DSCR_STICKY_ABORT_PRECISE) {
        if (final_retval == ERROR_OK) {
            /* Final return value will reflect cause of fault. */
            retval = cortex_a_read_dfar_dfsr(target, &fault_dfar, &fault_dfsr, &dscr);
            if (retval == ERROR_OK) {
                LOG_ERROR("data abort at 0x%08" PRIx32 ", dfsr = 0x%08" PRIx32, fault_dfar, fault_dfsr);
                final_retval = cortex_a_dfsr_to_error_code(fault_dfsr);
            } else
                final_retval = retval;
        }
        /* Fault destroyed DFAR/DFSR; restore them. */
        retval = cortex_a_write_dfar_dfsr(target, orig_dfar, orig_dfsr, &dscr);
        if (retval != ERROR_OK)
            LOG_ERROR("error restoring dfar/dfsr - dscr = 0x%08" PRIx32, dscr);
    }
 
    /* Handle asynchronous data faults. */
    if (fault_dscr & DSCR_STICKY_ABORT_IMPRECISE) {
        if (final_retval == ERROR_OK)
            /* No other error has been recorded so far, so keep this one. */
            final_retval = ERROR_TARGET_DATA_ABORT;
    }
 
    /* If the DCC is nonempty, clear it. */
    if (dscr & DSCR_DTRTX_FULL_LATCHED) {
        uint32_t dummy;
        retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DTRTX, &dummy);
        if (final_retval == ERROR_OK)
            final_retval = retval;
    }
    if (dscr & DSCR_DTRRX_FULL_LATCHED) {
        retval = cortex_a_exec_opcode(target, ARMV4_5_MRC(14, 0, 1, 0, 5, 0), &dscr);
        if (final_retval == ERROR_OK)
            final_retval = retval;
    }
 
    /* Done. */
    return final_retval;
}
 
 
/*
 * Cortex-A Memory access
 *
 * This is same Cortex-M3 but we must also use the correct
 * ap number for every access.
 */
 
static int cortex_a_read_phys_memory(struct target *target,
    target_addr_t address, uint32_t size,
    uint32_t count, uint8_t *buffer)
{
    int retval;
 
    if (!count || !buffer)
        return ERROR_COMMAND_SYNTAX_ERROR;
 
    LOG_DEBUG("Reading memory at real address " TARGET_ADDR_FMT "; size %" PRIu32 "; count %" PRIu32,
        address, size, count);
 
    /* read memory through the CPU */
    cortex_a_prep_memaccess(target, true);
    retval = cortex_a_read_cpu_memory(target, address, size, count, buffer);
    cortex_a_post_memaccess(target, true);
 
    return retval;
}
 
static int cortex_a_read_memory(struct target *target, target_addr_t address,
    uint32_t size, uint32_t count, uint8_t *buffer)
{
    int retval;
 
    /* cortex_a handles unaligned memory access */
    LOG_DEBUG("Reading memory at address " TARGET_ADDR_FMT "; size %" PRIu32 "; count %" PRIu32,
        address, size, count);
 
    cortex_a_prep_memaccess(target, false);
    retval = cortex_a_read_cpu_memory(target, address, size, count, buffer);
    cortex_a_post_memaccess(target, false);
 
    return retval;
}
 
static int cortex_a_write_phys_memory(struct target *target,
    target_addr_t address, uint32_t size,
    uint32_t count, const uint8_t *buffer)
{
    int retval;
 
    if (!count || !buffer)
        return ERROR_COMMAND_SYNTAX_ERROR;
 
    LOG_DEBUG("Writing memory to real address " TARGET_ADDR_FMT "; size %" PRIu32 "; count %" PRIu32,
        address, size, count);
 
    /* write memory through the CPU */
    cortex_a_prep_memaccess(target, true);
    retval = cortex_a_write_cpu_memory(target, address, size, count, buffer);
    cortex_a_post_memaccess(target, true);
 
    return retval;
}
 
static int cortex_a_write_memory(struct target *target, target_addr_t address,
    uint32_t size, uint32_t count, const uint8_t *buffer)
{
    int retval;
 
    /* cortex_a handles unaligned memory access */
    LOG_DEBUG("Writing memory at address " TARGET_ADDR_FMT "; size %" PRIu32 "; count %" PRIu32,
        address, size, count);
 
    cortex_a_prep_memaccess(target, false);
    retval = cortex_a_write_cpu_memory(target, address, size, count, buffer);
    cortex_a_post_memaccess(target, false);
    return retval;
}
 
static int cortex_a_read_buffer(struct target *target, target_addr_t address,
                uint32_t count, uint8_t *buffer)
{
    uint32_t size;
 
    /* Align up to maximum 4 bytes. The loop condition makes sure the next pass
     * will have something to do with the size we leave to it. */
    for (size = 1; size < 4 && count >= size * 2 + (address & size); size *= 2) {
        if (address & size) {
            int retval = target_read_memory(target, address, size, 1, buffer);
            if (retval != ERROR_OK)
                return retval;
            address += size;
            count -= size;
            buffer += size;
        }
    }
 
    /* Read the data with as large access size as possible. */
    for (; size > 0; size /= 2) {
        uint32_t aligned = count - count % size;
        if (aligned > 0) {
            int retval = target_read_memory(target, address, size, aligned / size, buffer);
            if (retval != ERROR_OK)
                return retval;
            address += aligned;
            count -= aligned;
            buffer += aligned;
        }
    }
 
    return ERROR_OK;
}
 
static int cortex_a_write_buffer(struct target *target, target_addr_t address,
                 uint32_t count, const uint8_t *buffer)
{
    uint32_t size;
 
    /* Align up to maximum 4 bytes. The loop condition makes sure the next pass
     * will have something to do with the size we leave to it. */
    for (size = 1; size < 4 && count >= size * 2 + (address & size); size *= 2) {
        if (address & size) {
            int retval = target_write_memory(target, address, size, 1, buffer);
            if (retval != ERROR_OK)
                return retval;
            address += size;
            count -= size;
            buffer += size;
        }
    }
 
    /* Write the data with as large access size as possible. */
    for (; size > 0; size /= 2) {
        uint32_t aligned = count - count % size;
        if (aligned > 0) {
            int retval = target_write_memory(target, address, size, aligned / size, buffer);
            if (retval != ERROR_OK)
                return retval;
            address += aligned;
            count -= aligned;
            buffer += aligned;
        }
    }
 
    return ERROR_OK;
}
 
static int cortex_a_handle_target_request(void *priv)
{
    struct target *target = priv;
    struct armv7a_common *armv7a = target_to_armv7a(target);
    int retval;
 
    if (!target_was_examined(target))
        return ERROR_OK;
    if (!target->dbg_msg_enabled)
        return ERROR_OK;
 
    if (target->state == TARGET_RUNNING) {
        uint32_t request;
        uint32_t dscr;
        retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DSCR, &dscr);
 
        /* check if we have data */
        int64_t then = timeval_ms();
        while ((dscr & DSCR_DTR_TX_FULL) && (retval == ERROR_OK)) {
            retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_DTRTX, &request);
            if (retval == ERROR_OK) {
                target_request(target, request);
                retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                        armv7a->debug_base + CPUDBG_DSCR, &dscr);
            }
            if (timeval_ms() > then + 1000) {
                LOG_ERROR("Timeout waiting for dtr tx full");
                return ERROR_FAIL;
            }
        }
    }
 
    return ERROR_OK;
}
 
/*
 * Cortex-A target information and configuration
 */
 
static int cortex_a_examine_first(struct target *target)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct adiv5_dap *swjdp = armv7a->arm.dap;
    struct adiv5_private_config *pc = target->private_config;
 
    int i;
    int retval = ERROR_OK;
    uint32_t didr, cpuid, dbg_osreg, dbg_idpfr1;
 
    if (!armv7a->debug_ap) {
        if (pc->ap_num == DP_APSEL_INVALID) {
            /* Search for the APB-AP - it is needed for access to debug registers */
            retval = dap_find_get_ap(swjdp, AP_TYPE_APB_AP, &armv7a->debug_ap);
            if (retval != ERROR_OK) {
                LOG_ERROR("Could not find APB-AP for debug access");
                return retval;
            }
        } else {
            armv7a->debug_ap = dap_get_ap(swjdp, pc->ap_num);
            if (!armv7a->debug_ap) {
                LOG_ERROR("Cannot get AP");
                return ERROR_FAIL;
            }
        }
    }
 
    retval = mem_ap_init(armv7a->debug_ap);
    if (retval != ERROR_OK) {
        LOG_ERROR("Could not initialize the APB-AP");
        return retval;
    }
 
    armv7a->debug_ap->memaccess_tck = 80;
 
    if (!target->dbgbase_set) {
        LOG_TARGET_DEBUG(target, "dbgbase is not set, trying to detect using the ROM table");
        /* Lookup Processor DAP */
        retval = dap_lookup_cs_component(armv7a->debug_ap, ARM_CS_C9_DEVTYPE_CORE_DEBUG,
                &armv7a->debug_base, target->coreid);
        if (retval != ERROR_OK) {
            LOG_TARGET_ERROR(target, "Can't detect dbgbase from the ROM table; you need to specify it explicitly");
            return retval;
        }
        LOG_DEBUG("Detected core %" PRId32 " dbgbase: " TARGET_ADDR_FMT,
              target->coreid, armv7a->debug_base);
    } else
        armv7a->debug_base = target->dbgbase;
 
    if ((armv7a->debug_base & (1UL<<31)) == 0)
        LOG_TARGET_WARNING(target,
            "Debug base address has bit 31 set to 0. Access to debug registers will likely fail!\n"
            "Please fix the target configuration");
 
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_DIDR, &didr);
    if (retval != ERROR_OK) {
        LOG_DEBUG("Examine %s failed", "DIDR");
        return retval;
    }
 
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
            armv7a->debug_base + CPUDBG_CPUID, &cpuid);
    if (retval != ERROR_OK) {
        LOG_DEBUG("Examine %s failed", "CPUID");
        return retval;
    }
 
    LOG_DEBUG("didr = 0x%08" PRIx32, didr);
    LOG_DEBUG("cpuid = 0x%08" PRIx32, cpuid);
 
    cortex_a->didr = didr;
    cortex_a->cpuid = cpuid;
 
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_PRSR, &dbg_osreg);
    if (retval != ERROR_OK)
        return retval;
    LOG_TARGET_DEBUG(target, "DBGPRSR  0x%" PRIx32, dbg_osreg);
 
    if ((dbg_osreg & PRSR_POWERUP_STATUS) == 0) {
        LOG_TARGET_ERROR(target, "powered down!");
        target->state = TARGET_UNKNOWN; /* TARGET_NO_POWER? */
        return ERROR_TARGET_INIT_FAILED;
    }
 
    if (dbg_osreg & PRSR_STICKY_RESET_STATUS)
        LOG_TARGET_DEBUG(target, "was reset!");
 
    /* Read DBGOSLSR and check if OSLK is implemented */
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_OSLSR, &dbg_osreg);
    if (retval != ERROR_OK)
        return retval;
    LOG_TARGET_DEBUG(target, "DBGOSLSR 0x%" PRIx32, dbg_osreg);
 
    /* check if OS Lock is implemented */
    if ((dbg_osreg & OSLSR_OSLM) == OSLSR_OSLM0 || (dbg_osreg & OSLSR_OSLM) == OSLSR_OSLM1) {
        /* check if OS Lock is set */
        if (dbg_osreg & OSLSR_OSLK) {
            LOG_TARGET_DEBUG(target, "OSLock set! Trying to unlock");
 
            retval = mem_ap_write_atomic_u32(armv7a->debug_ap,
                            armv7a->debug_base + CPUDBG_OSLAR,
                            0);
            if (retval == ERROR_OK)
                retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                            armv7a->debug_base + CPUDBG_OSLSR, &dbg_osreg);
 
            /* if we fail to access the register or cannot reset the OSLK bit, bail out */
            if (retval != ERROR_OK || (dbg_osreg & OSLSR_OSLK) != 0) {
                LOG_TARGET_ERROR(target, "OSLock sticky, core not powered?");
                target->state = TARGET_UNKNOWN; /* TARGET_NO_POWER? */
                return ERROR_TARGET_INIT_FAILED;
            }
        }
    }
 
    retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                 armv7a->debug_base + CPUDBG_ID_PFR1, &dbg_idpfr1);
    if (retval != ERROR_OK)
        return retval;
 
    if (dbg_idpfr1 & 0x000000f0) {
        LOG_TARGET_DEBUG(target, "has security extensions");
        armv7a->arm.core_type = ARM_CORE_TYPE_SEC_EXT;
    }
    if (dbg_idpfr1 & 0x0000f000) {
        LOG_TARGET_DEBUG(target, "has virtualization extensions");
        /*
         * overwrite and simplify the checks.
         * virtualization extensions require implementation of security extension
         */
        armv7a->arm.core_type = ARM_CORE_TYPE_VIRT_EXT;
    }
 
    /* Avoid recreating the registers cache */
    if (!target_was_examined(target)) {
        retval = cortex_a_dpm_setup(cortex_a, didr);
        if (retval != ERROR_OK)
            return retval;
    }
 
    /* Setup Breakpoint Register Pairs */
    cortex_a->brp_num = ((didr >> 24) & 0x0F) + 1;
    cortex_a->brp_num_context = ((didr >> 20) & 0x0F) + 1;
    cortex_a->brp_num_available = cortex_a->brp_num;
    free(cortex_a->brp_list);
    cortex_a->brp_list = calloc(cortex_a->brp_num, sizeof(struct cortex_a_brp));
/*  cortex_a->brb_enabled = ????; */
    for (i = 0; i < cortex_a->brp_num; i++) {
        cortex_a->brp_list[i].used = false;
        if (i < (cortex_a->brp_num-cortex_a->brp_num_context))
            cortex_a->brp_list[i].type = BRP_NORMAL;
        else
            cortex_a->brp_list[i].type = BRP_CONTEXT;
        cortex_a->brp_list[i].value = 0;
        cortex_a->brp_list[i].control = 0;
        cortex_a->brp_list[i].brpn = i;
    }
 
    LOG_DEBUG("Configured %i hw breakpoints", cortex_a->brp_num);
 
    /* Setup Watchpoint Register Pairs */
    cortex_a->wrp_num = ((didr >> 28) & 0x0F) + 1;
    cortex_a->wrp_num_available = cortex_a->wrp_num;
    free(cortex_a->wrp_list);
    cortex_a->wrp_list = calloc(cortex_a->wrp_num, sizeof(struct cortex_a_wrp));
    for (i = 0; i < cortex_a->wrp_num; i++) {
        cortex_a->wrp_list[i].used = false;
        cortex_a->wrp_list[i].value = 0;
        cortex_a->wrp_list[i].control = 0;
        cortex_a->wrp_list[i].wrpn = i;
    }
 
    LOG_DEBUG("Configured %i hw watchpoints", cortex_a->wrp_num);
 
    /* select debug_ap as default */
    swjdp->apsel = armv7a->debug_ap->ap_num;
 
    target_set_examined(target);
    return ERROR_OK;
}
 
static int cortex_a_examine(struct target *target)
{
    int retval = ERROR_OK;
 
    /* Reestablish communication after target reset */
    retval = cortex_a_examine_first(target);
 
    /* Configure core debug access */
    if (retval == ERROR_OK)
        retval = cortex_a_init_debug_access(target);
 
    return retval;
}
 
/*
 *  Cortex-A target creation and initialization
 */
 
static int cortex_a_init_target(struct command_context *cmd_ctx,
    struct target *target)
{
    /* examine_first() does a bunch of this */
    arm_semihosting_init(target);
    return ERROR_OK;
}
 
static int cortex_a_init_arch_info(struct target *target,
    struct cortex_a_common *cortex_a, struct adiv5_dap *dap)
{
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
 
    /* Setup struct cortex_a_common */
    cortex_a->common_magic = CORTEX_A_COMMON_MAGIC;
    armv7a->arm.dap = dap;
 
    /* register arch-specific functions */
    armv7a->examine_debug_reason = NULL;
 
    armv7a->post_debug_entry = cortex_a_post_debug_entry;
 
    armv7a->pre_restore_context = NULL;
 
    armv7a->armv7a_mmu.read_physical_memory = cortex_a_read_phys_memory;
 
 
/*  arm7_9->handle_target_request = cortex_a_handle_target_request; */
 
    /* REVISIT v7a setup should be in a v7a-specific routine */
    armv7a_init_arch_info(target, armv7a);
    target_register_timer_callback(cortex_a_handle_target_request, 1,
        TARGET_TIMER_TYPE_PERIODIC, target);
 
    return ERROR_OK;
}
 
static int cortex_a_target_create(struct target *target)
{
    struct cortex_a_common *cortex_a;
    struct adiv5_private_config *pc;
 
    if (!target->private_config)
        return ERROR_FAIL;
 
    pc = (struct adiv5_private_config *)target->private_config;
 
    cortex_a = calloc(1, sizeof(struct cortex_a_common));
    if (!cortex_a) {
        LOG_ERROR("Out of memory");
        return ERROR_FAIL;
    }
    cortex_a->common_magic = CORTEX_A_COMMON_MAGIC;
    cortex_a->armv7a_common.is_armv7r = false;
    cortex_a->armv7a_common.arm.arm_vfp_version = ARM_VFP_V3;
 
    return cortex_a_init_arch_info(target, cortex_a, pc->dap);
}
 
static int cortex_r4_target_create(struct target *target)
{
    struct cortex_a_common *cortex_a;
    struct adiv5_private_config *pc;
 
    pc = (struct adiv5_private_config *)target->private_config;
    if (adiv5_verify_config(pc) != ERROR_OK)
        return ERROR_FAIL;
 
    cortex_a = calloc(1, sizeof(struct cortex_a_common));
    if (!cortex_a) {
        LOG_ERROR("Out of memory");
        return ERROR_FAIL;
    }
    cortex_a->common_magic = CORTEX_A_COMMON_MAGIC;
    cortex_a->armv7a_common.is_armv7r = true;
 
    return cortex_a_init_arch_info(target, cortex_a, pc->dap);
}
 
static void cortex_a_deinit_target(struct target *target)
{
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
    struct armv7a_common *armv7a = &cortex_a->armv7a_common;
    struct arm_dpm *dpm = &armv7a->dpm;
    uint32_t dscr;
    int retval;
 
    if (target_was_examined(target)) {
        /* Disable halt for breakpoint, watchpoint and vector catch */
        retval = mem_ap_read_atomic_u32(armv7a->debug_ap,
                armv7a->debug_base + CPUDBG_DSCR, &dscr);
        if (retval == ERROR_OK)
            mem_ap_write_atomic_u32(armv7a->debug_ap,
                    armv7a->debug_base + CPUDBG_DSCR,
                    dscr & ~DSCR_HALT_DBG_MODE);
    }
 
    if (armv7a->debug_ap)
        dap_put_ap(armv7a->debug_ap);
 
    free(cortex_a->wrp_list);
    free(cortex_a->brp_list);
    arm_free_reg_cache(dpm->arm);
    free(dpm->dbp);
    free(dpm->dwp);
    free(target->private_config);
    free(cortex_a);
}
 
static int cortex_a_mmu(struct target *target, bool *enabled)
{
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    if (target->state != TARGET_HALTED) {
        LOG_TARGET_ERROR(target, "not halted");
        return ERROR_TARGET_NOT_HALTED;
    }
 
    if (armv7a->is_armv7r)
        *enabled = false;
    else
        *enabled = target_to_cortex_a(target)->armv7a_common.armv7a_mmu.mmu_enabled;
 
    return ERROR_OK;
}
 
static int cortex_a_virt2phys(struct target *target,
    target_addr_t virt, target_addr_t *phys)
{
    int retval;
    bool mmu_enabled = false;
 
    /*
     * If the MMU was not enabled at debug entry, there is no
     * way of knowing if there was ever a valid configuration
     * for it and thus it's not safe to enable it. In this case,
     * just return the virtual address as physical.
     */
    cortex_a_mmu(target, &mmu_enabled);
    if (!mmu_enabled) {
        *phys = virt;
        return ERROR_OK;
    }
 
    /* mmu must be enable in order to get a correct translation */
    retval = cortex_a_mmu_modify(target, true);
    if (retval != ERROR_OK)
        return retval;
    return armv7a_mmu_translate_va_pa(target, (uint32_t)virt,
                            phys, 1);
}
 
COMMAND_HANDLER(cortex_a_handle_cache_info_command)
{
    struct target *target = get_current_target(CMD_CTX);
    struct armv7a_common *armv7a = target_to_armv7a(target);
 
    return armv7a_handle_cache_info_command(CMD,
            &armv7a->armv7a_mmu.armv7a_cache);
}
 
 
COMMAND_HANDLER(cortex_a_handle_dbginit_command)
{
    struct target *target = get_current_target(CMD_CTX);
    if (!target_was_examined(target)) {
        LOG_ERROR("target not examined yet");
        return ERROR_FAIL;
    }
 
    return cortex_a_init_debug_access(target);
}
 
COMMAND_HANDLER(handle_cortex_a_mask_interrupts_command)
{
    struct target *target = get_current_target(CMD_CTX);
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    static const struct nvp nvp_maskisr_modes[] = {
        { .name = "off", .value = CORTEX_A_ISRMASK_OFF },
        { .name = "on", .value = CORTEX_A_ISRMASK_ON },
        { .name = NULL, .value = -1 },
    };
    const struct nvp *n;
 
    if (CMD_ARGC > 0) {
        n = nvp_name2value(nvp_maskisr_modes, CMD_ARGV[0]);
        if (!n->name) {
            LOG_ERROR("Unknown parameter: %s - should be off or on", CMD_ARGV[0]);
            return ERROR_COMMAND_SYNTAX_ERROR;
        }
 
        cortex_a->isrmasking_mode = n->value;
    }
 
    n = nvp_value2name(nvp_maskisr_modes, cortex_a->isrmasking_mode);
    command_print(CMD, "cortex_a interrupt mask %s", n->name);
 
    return ERROR_OK;
}
 
COMMAND_HANDLER(handle_cortex_a_dacrfixup_command)
{
    struct target *target = get_current_target(CMD_CTX);
    struct cortex_a_common *cortex_a = target_to_cortex_a(target);
 
    static const struct nvp nvp_dacrfixup_modes[] = {
        { .name = "off", .value = CORTEX_A_DACRFIXUP_OFF },
        { .name = "on", .value = CORTEX_A_DACRFIXUP_ON },
        { .name = NULL, .value = -1 },
    };
    const struct nvp *n;
 
    if (CMD_ARGC > 0) {
        n = nvp_name2value(nvp_dacrfixup_modes, CMD_ARGV[0]);
        if (!n->name)
            return ERROR_COMMAND_SYNTAX_ERROR;
        cortex_a->dacrfixup_mode = n->value;
 
    }
 
    n = nvp_value2name(nvp_dacrfixup_modes, cortex_a->dacrfixup_mode);
    command_print(CMD, "cortex_a domain access control fixup %s", n->name);
 
    return ERROR_OK;
}
 
static const struct command_registration cortex_a_exec_command_handlers[] = {
    {
        .name = "cache_info",
        .handler = cortex_a_handle_cache_info_command,
        .mode = COMMAND_EXEC,
        .help = "display information about target caches",
        .usage = "",
    },
    {
        .name = "dbginit",
        .handler = cortex_a_handle_dbginit_command,
        .mode = COMMAND_EXEC,
        .help = "Initialize core debug",
        .usage = "",
    },
    {
        .name = "maskisr",
        .handler = handle_cortex_a_mask_interrupts_command,
        .mode = COMMAND_ANY,
        .help = "mask cortex_a interrupts",
        .usage = "['on'|'off']",
    },
    {
        .name = "dacrfixup",
        .handler = handle_cortex_a_dacrfixup_command,
        .mode = COMMAND_ANY,
        .help = "set domain access control (DACR) to all-manager "
            "on memory access",
        .usage = "['on'|'off']",
    },
    {
        .chain = armv7a_mmu_command_handlers,
    },
    {
        .chain = smp_command_handlers,
    },
 
    COMMAND_REGISTRATION_DONE
};
static const struct command_registration cortex_a_command_handlers[] = {
    {
        .chain = arm_command_handlers,
    },
    {
        .chain = armv7a_command_handlers,
    },
    {
        .name = "cortex_a",
        .mode = COMMAND_ANY,
        .help = "Cortex-A command group",
        .usage = "",
        .chain = cortex_a_exec_command_handlers,
    },
    COMMAND_REGISTRATION_DONE
};
 
struct target_type cortexa_target = {
    .name = "cortex_a",
 
    .poll = cortex_a_poll,
    .arch_state = armv7a_arch_state,
 
    .halt = cortex_a_halt,
    .resume = cortex_a_resume,
    .step = cortex_a_step,
 
    .assert_reset = cortex_a_assert_reset,
    .deassert_reset = cortex_a_deassert_reset,
 
    /* REVISIT allow exporting VFP3 registers ... */
    .get_gdb_arch = arm_get_gdb_arch,
    .get_gdb_reg_list = arm_get_gdb_reg_list,
 
    .read_memory = cortex_a_read_memory,
    .write_memory = cortex_a_write_memory,
 
    .read_buffer = cortex_a_read_buffer,
    .write_buffer = cortex_a_write_buffer,
 
    .checksum_memory = arm_checksum_memory,
    .blank_check_memory = arm_blank_check_memory,
 
    .run_algorithm = armv4_5_run_algorithm,
 
    .add_breakpoint = cortex_a_add_breakpoint,
    .add_context_breakpoint = cortex_a_add_context_breakpoint,
    .add_hybrid_breakpoint = cortex_a_add_hybrid_breakpoint,
    .remove_breakpoint = cortex_a_remove_breakpoint,
    .add_watchpoint = cortex_a_add_watchpoint,
    .remove_watchpoint = cortex_a_remove_watchpoint,
 
    .commands = cortex_a_command_handlers,
    .target_create = cortex_a_target_create,
    .target_jim_configure = adiv5_jim_configure,
    .init_target = cortex_a_init_target,
    .examine = cortex_a_examine,
    .deinit_target = cortex_a_deinit_target,
 
    .read_phys_memory = cortex_a_read_phys_memory,
    .write_phys_memory = cortex_a_write_phys_memory,
    .mmu = cortex_a_mmu,
    .virt2phys = cortex_a_virt2phys,
};
 
static const struct command_registration cortex_r4_exec_command_handlers[] = {
    {
        .name = "dbginit",
        .handler = cortex_a_handle_dbginit_command,
        .mode = COMMAND_EXEC,
        .help = "Initialize core debug",
        .usage = "",
    },
    {
        .name = "maskisr",
        .handler = handle_cortex_a_mask_interrupts_command,
        .mode = COMMAND_EXEC,
        .help = "mask cortex_r4 interrupts",
        .usage = "['on'|'off']",
    },
 
    COMMAND_REGISTRATION_DONE
};
static const struct command_registration cortex_r4_command_handlers[] = {
    {
        .chain = arm_command_handlers,
    },
    {
        .name = "cortex_r4",
        .mode = COMMAND_ANY,
        .help = "Cortex-R4 command group",
        .usage = "",
        .chain = cortex_r4_exec_command_handlers,
    },
    COMMAND_REGISTRATION_DONE
};
 
struct target_type cortexr4_target = {
    .name = "cortex_r4",
 
    .poll = cortex_a_poll,
    .arch_state = armv7a_arch_state,
 
    .halt = cortex_a_halt,
    .resume = cortex_a_resume,
    .step = cortex_a_step,
 
    .assert_reset = cortex_a_assert_reset,
    .deassert_reset = cortex_a_deassert_reset,
 
    /* REVISIT allow exporting VFP3 registers ... */
    .get_gdb_arch = arm_get_gdb_arch,
    .get_gdb_reg_list = arm_get_gdb_reg_list,
 
    .read_memory = cortex_a_read_phys_memory,
    .write_memory = cortex_a_write_phys_memory,
 
    .checksum_memory = arm_checksum_memory,
    .blank_check_memory = arm_blank_check_memory,
 
    .run_algorithm = armv4_5_run_algorithm,
 
    .add_breakpoint = cortex_a_add_breakpoint,
    .add_context_breakpoint = cortex_a_add_context_breakpoint,
    .add_hybrid_breakpoint = cortex_a_add_hybrid_breakpoint,
    .remove_breakpoint = cortex_a_remove_breakpoint,
    .add_watchpoint = cortex_a_add_watchpoint,
    .remove_watchpoint = cortex_a_remove_watchpoint,
 
    .commands = cortex_r4_command_handlers,
    .target_create = cortex_r4_target_create,
    .target_jim_configure = adiv5_jim_configure,
    .init_target = cortex_a_init_target,
    .examine = cortex_a_examine,
    .deinit_target = cortex_a_deinit_target,
};