21 files changed, 1224 insertions, 848 deletions

diff --git a/arch/x86/cpu/Makefile b/arch/x86/cpu/Makefile
index 54668aab24..85fd5e616e 100644
--- a/arch/x86/cpu/Makefile
+++ b/arch/x86/cpu/Makefile
@@ -9,9 +9,22 @@
 ifeq ($(CONFIG_$(SPL_)X86_64),y)
 extra-y	= start64.o
 else
+ifeq ($(CONFIG_$(SPL_TPL_)X86_16BIT_INIT),y)
 extra-y	= start.o
+else
+ifndef CONFIG_SPL
+extra-y	= start.o
+else
+ifdef CONFIG_SPL_BUILD
+extra-y	= start_from_tpl.o
+else
+extra-y	= start_from_spl.o
 endif
-extra-$(CONFIG_$(SPL_)X86_16BIT_INIT) += resetvec.o start16.o
+endif
+endif
+endif
+
+extra-$(CONFIG_$(SPL_TPL_)X86_16BIT_INIT) += resetvec.o start16.o
 
 obj-y	+= cpu.o cpu_x86.o
 
diff --git a/arch/x86/cpu/broadwell/Makefile b/arch/x86/cpu/broadwell/Makefile
index d3785aabdf..52d56c65be 100644
--- a/arch/x86/cpu/broadwell/Makefile
+++ b/arch/x86/cpu/broadwell/Makefile
@@ -3,7 +3,24 @@
 # Copyright (c) 2016 Google, Inc
 
 obj-y += adsp.o
-obj-y += cpu.o
+obj-$(CONFIG_$(SPL_TPL_)X86_16BIT_INIT) += cpu.o
+obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += cpu_full.o
+
+ifdef CONFIG_SPL
+ifndef CONFIG_SPL_BUILD
+obj-y += cpu_from_spl.o
+obj-y += cpu_full.o
+obj-y += refcode.o
+endif
+ifndef CONFIG_SPL_BUILD
+# obj-y += cpu_from_spl.o
+endif
+endif
+
+ifeq ($(CONFIG_$(SPL_TPL_)X86_32BIT_INIT),)
+#obj-y += cpu_from_spl.o
+endif
+
 obj-y += iobp.o
 obj-y += lpc.o
 obj-y += me.o
@@ -11,6 +28,6 @@ obj-y += northbridge.o
 obj-y += pch.o
 obj-y += pinctrl_broadwell.o
 obj-y += power_state.o
-obj-y += refcode.o
+obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += refcode.o
 obj-y += sata.o
-obj-y += sdram.o
+obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += sdram.o
diff --git a/arch/x86/cpu/broadwell/cpu.c b/arch/x86/cpu/broadwell/cpu.c
index 232fa40eb5..bb7c361408 100644
--- a/arch/x86/cpu/broadwell/cpu.c
+++ b/arch/x86/cpu/broadwell/cpu.c
@@ -12,75 +12,15 @@
 #include <asm/cpu_x86.h>
 #include <asm/cpu_common.h>
 #include <asm/intel_regs.h>
+#include <asm/lpc_common.h>
 #include <asm/msr.h>
+#include <asm/pci.h>
 #include <asm/post.h>
 #include <asm/turbo.h>
 #include <asm/arch/cpu.h>
 #include <asm/arch/pch.h>
 #include <asm/arch/rcb.h>
 
-struct cpu_broadwell_priv {
-	bool ht_disabled;
-};
-
-/* Convert time in seconds to POWER_LIMIT_1_TIME MSR value */
-static const u8 power_limit_time_sec_to_msr[] = {
-	[0]   = 0x00,
-	[1]   = 0x0a,
-	[2]   = 0x0b,
-	[3]   = 0x4b,
-	[4]   = 0x0c,
-	[5]   = 0x2c,
-	[6]   = 0x4c,
-	[7]   = 0x6c,
-	[8]   = 0x0d,
-	[10]  = 0x2d,
-	[12]  = 0x4d,
-	[14]  = 0x6d,
-	[16]  = 0x0e,
-	[20]  = 0x2e,
-	[24]  = 0x4e,
-	[28]  = 0x6e,
-	[32]  = 0x0f,
-	[40]  = 0x2f,
-	[48]  = 0x4f,
-	[56]  = 0x6f,
-	[64]  = 0x10,
-	[80]  = 0x30,
-	[96]  = 0x50,
-	[112] = 0x70,
-	[128] = 0x11,
-};
-
-/* Convert POWER_LIMIT_1_TIME MSR value to seconds */
-static const u8 power_limit_time_msr_to_sec[] = {
-	[0x00] = 0,
-	[0x0a] = 1,
-	[0x0b] = 2,
-	[0x4b] = 3,
-	[0x0c] = 4,
-	[0x2c] = 5,
-	[0x4c] = 6,
-	[0x6c] = 7,
-	[0x0d] = 8,
-	[0x2d] = 10,
-	[0x4d] = 12,
-	[0x6d] = 14,
-	[0x0e] = 16,
-	[0x2e] = 20,
-	[0x4e] = 24,
-	[0x6e] = 28,
-	[0x0f] = 32,
-	[0x2f] = 40,
-	[0x4f] = 48,
-	[0x6f] = 56,
-	[0x10] = 64,
-	[0x30] = 80,
-	[0x50] = 96,
-	[0x70] = 112,
-	[0x11] = 128,
-};
-
 int arch_cpu_init_dm(void)
 {
 	struct udevice *dev;
@@ -156,613 +96,13 @@ int print_cpuinfo(void)
 	return 0;
 }
 
-/*
- * The core 100MHz BLCK is disabled in deeper c-states. One needs to calibrate
- * the 100MHz BCLCK against the 24MHz BLCK to restore the clocks properly
- * when a core is woken up
- */
-static int pcode_ready(void)
-{
-	int wait_count;
-	const int delay_step = 10;
-
-	wait_count = 0;
-	do {
-		if (!(readl(MCHBAR_REG(BIOS_MAILBOX_INTERFACE)) &
-				MAILBOX_RUN_BUSY))
-			return 0;
-		wait_count += delay_step;
-		udelay(delay_step);
-	} while (wait_count < 1000);
-
-	return -ETIMEDOUT;
-}
-
-static u32 pcode_mailbox_read(u32 command)
-{
-	int ret;
-
-	ret = pcode_ready();
-	if (ret) {
-		debug("PCODE: mailbox timeout on wait ready\n");
-		return ret;
-	}
-
-	/* Send command and start transaction */
-	writel(command | MAILBOX_RUN_BUSY, MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
-
-	ret = pcode_ready();
-	if (ret) {
-		debug("PCODE: mailbox timeout on completion\n");
-		return ret;
-	}
-
-	/* Read mailbox */
-	return readl(MCHBAR_REG(BIOS_MAILBOX_DATA));
-}
-
-static int pcode_mailbox_write(u32 command, u32 data)
-{
-	int ret;
-
-	ret = pcode_ready();
-	if (ret) {
-		debug("PCODE: mailbox timeout on wait ready\n");
-		return ret;
-	}
-
-	writel(data, MCHBAR_REG(BIOS_MAILBOX_DATA));
-
-	/* Send command and start transaction */
-	writel(command | MAILBOX_RUN_BUSY, MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
-
-	ret = pcode_ready();
-	if (ret) {
-		debug("PCODE: mailbox timeout on completion\n");
-		return ret;
-	}
-
-	return 0;
-}
-
-/* @dev is the CPU device */
-static void initialize_vr_config(struct udevice *dev)
-{
-	int ramp, min_vid;
-	msr_t msr;
-
-	debug("Initializing VR config\n");
-
-	/* Configure VR_CURRENT_CONFIG */
-	msr = msr_read(MSR_VR_CURRENT_CONFIG);
-	/*
-	 * Preserve bits 63 and 62. Bit 62 is PSI4 enable, but it is only valid
-	 * on ULT systems
-	 */
-	msr.hi &= 0xc0000000;
-	msr.hi |= (0x01 << (52 - 32)); /* PSI3 threshold -  1A */
-	msr.hi |= (0x05 << (42 - 32)); /* PSI2 threshold -  5A */
-	msr.hi |= (0x14 << (32 - 32)); /* PSI1 threshold - 20A */
-	msr.hi |= (1 <<  (62 - 32)); /* Enable PSI4 */
-	/* Leave the max instantaneous current limit (12:0) to default */
-	msr_write(MSR_VR_CURRENT_CONFIG, msr);
-
-	/* Configure VR_MISC_CONFIG MSR */
-	msr = msr_read(MSR_VR_MISC_CONFIG);
-	/* Set the IOUT_SLOPE scalar applied to dIout in U10.1.9 format */
-	msr.hi &= ~(0x3ff << (40 - 32));
-	msr.hi |= (0x200 << (40 - 32)); /* 1.0 */
-	/* Set IOUT_OFFSET to 0 */
-	msr.hi &= ~0xff;
-	/* Set entry ramp rate to slow */
-	msr.hi &= ~(1 << (51 - 32));
-	/* Enable decay mode on C-state entry */
-	msr.hi |= (1 << (52 - 32));
-	/* Set the slow ramp rate */
-	msr.hi &= ~(0x3 << (53 - 32));
-	/* Configure the C-state exit ramp rate */
-	ramp = fdtdec_get_int(gd->fdt_blob, dev_of_offset(dev),
-			      "intel,slow-ramp", -1);
-	if (ramp != -1) {
-		/* Configured slow ramp rate */
-		msr.hi |= ((ramp & 0x3) << (53 - 32));
-		/* Set exit ramp rate to slow */
-		msr.hi &= ~(1 << (50 - 32));
-	} else {
-		/* Fast ramp rate / 4 */
-		msr.hi |= (0x01 << (53 - 32));
-		/* Set exit ramp rate to fast */
-		msr.hi |= (1 << (50 - 32));
-	}
-	/* Set MIN_VID (31:24) to allow CPU to have full control */
-	msr.lo &= ~0xff000000;
-	min_vid = fdtdec_get_int(gd->fdt_blob, dev_of_offset(dev),
-				 "intel,min-vid", 0);
-	msr.lo |= (min_vid & 0xff) << 24;
-	msr_write(MSR_VR_MISC_CONFIG, msr);
-
-	/*  Configure VR_MISC_CONFIG2 MSR */
-	msr = msr_read(MSR_VR_MISC_CONFIG2);
-	msr.lo &= ~0xffff;
-	/*
-	 * Allow CPU to control minimum voltage completely (15:8) and
-	 * set the fast ramp voltage in 10mV steps
-	 */
-	if (cpu_get_family_model() == BROADWELL_FAMILY_ULT)
-		msr.lo |= 0x006a; /* 1.56V */
-	else
-		msr.lo |= 0x006f; /* 1.60V */
-	msr_write(MSR_VR_MISC_CONFIG2, msr);
-
-	/* Set C9/C10 VCC Min */
-	pcode_mailbox_write(MAILBOX_BIOS_CMD_WRITE_C9C10_VOLTAGE, 0x1f1f);
-}
-
-static int calibrate_24mhz_bclk(void)
-{
-	int err_code;
-	int ret;
-
-	ret = pcode_ready();
-	if (ret)
-		return ret;
-
-	/* A non-zero value initiates the PCODE calibration */
-	writel(~0, MCHBAR_REG(BIOS_MAILBOX_DATA));
-	writel(MAILBOX_RUN_BUSY | MAILBOX_BIOS_CMD_FSM_MEASURE_INTVL,
-	       MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
-
-	ret = pcode_ready();
-	if (ret)
-		return ret;
-
-	err_code = readl(MCHBAR_REG(BIOS_MAILBOX_INTERFACE)) & 0xff;
-
-	debug("PCODE: 24MHz BLCK calibration response: %d\n", err_code);
-
-	/* Read the calibrated value */
-	writel(MAILBOX_RUN_BUSY | MAILBOX_BIOS_CMD_READ_CALIBRATION,
-	       MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
-
-	ret = pcode_ready();
-	if (ret)
-		return ret;
-
-	debug("PCODE: 24MHz BLCK calibration value: 0x%08x\n",
-	      readl(MCHBAR_REG(BIOS_MAILBOX_DATA)));
-
-	return 0;
-}
-
-static void configure_pch_power_sharing(void)
-{
-	u32 pch_power, pch_power_ext, pmsync, pmsync2;
-	int i;
-
-	/* Read PCH Power levels from PCODE */
-	pch_power = pcode_mailbox_read(MAILBOX_BIOS_CMD_READ_PCH_POWER);
-	pch_power_ext = pcode_mailbox_read(MAILBOX_BIOS_CMD_READ_PCH_POWER_EXT);
-
-	debug("PCH Power: PCODE Levels 0x%08x 0x%08x\n", pch_power,
-	      pch_power_ext);
-
-	pmsync = readl(RCB_REG(PMSYNC_CONFIG));
-	pmsync2 = readl(RCB_REG(PMSYNC_CONFIG2));
-
-	/*
-	 * Program PMSYNC_TPR_CONFIG PCH power limit values
-	 *  pmsync[0:4]   = mailbox[0:5]
-	 *  pmsync[8:12]  = mailbox[6:11]
-	 *  pmsync[16:20] = mailbox[12:17]
-	 */
-	for (i = 0; i < 3; i++) {
-		u32 level = pch_power & 0x3f;
-		pch_power >>= 6;
-		pmsync &= ~(0x1f << (i * 8));
-		pmsync |= (level & 0x1f) << (i * 8);
-	}
-	writel(pmsync, RCB_REG(PMSYNC_CONFIG));
-
-	/*
-	 * Program PMSYNC_TPR_CONFIG2 Extended PCH power limit values
-	 *  pmsync2[0:4]   = mailbox[23:18]
-	 *  pmsync2[8:12]  = mailbox_ext[6:11]
-	 *  pmsync2[16:20] = mailbox_ext[12:17]
-	 *  pmsync2[24:28] = mailbox_ext[18:22]
-	 */
-	pmsync2 &= ~0x1f;
-	pmsync2 |= pch_power & 0x1f;
-
-	for (i = 1; i < 4; i++) {
-		u32 level = pch_power_ext & 0x3f;
-		pch_power_ext >>= 6;
-		pmsync2 &= ~(0x1f << (i * 8));
-		pmsync2 |= (level & 0x1f) << (i * 8);
-	}
-	writel(pmsync2, RCB_REG(PMSYNC_CONFIG2));
-}
-
-static int bsp_init_before_ap_bringup(struct udevice *dev)
-{
-	int ret;
-
-	initialize_vr_config(dev);
-	ret = calibrate_24mhz_bclk();
-	if (ret)
-		return ret;
-	configure_pch_power_sharing();
-
-	return 0;
-}
-
-int cpu_config_tdp_levels(void)
-{
-	msr_t platform_info;
-
-	/* Bits 34:33 indicate how many levels supported */
-	platform_info = msr_read(MSR_PLATFORM_INFO);
-	return (platform_info.hi >> 1) & 3;
-}
-
-static void set_max_ratio(void)
-{
-	msr_t msr, perf_ctl;
-
-	perf_ctl.hi = 0;
-
-	/* Check for configurable TDP option */
-	if (turbo_get_state() == TURBO_ENABLED) {
-		msr = msr_read(MSR_NHM_TURBO_RATIO_LIMIT);
-		perf_ctl.lo = (msr.lo & 0xff) << 8;
-	} else if (cpu_config_tdp_levels()) {
-		/* Set to nominal TDP ratio */
-		msr = msr_read(MSR_CONFIG_TDP_NOMINAL);
-		perf_ctl.lo = (msr.lo & 0xff) << 8;
-	} else {
-		/* Platform Info bits 15:8 give max ratio */
-		msr = msr_read(MSR_PLATFORM_INFO);
-		perf_ctl.lo = msr.lo & 0xff00;
-	}
-	msr_write(IA32_PERF_CTL, perf_ctl);
-
-	debug("cpu: frequency set to %d\n",
-	      ((perf_ctl.lo >> 8) & 0xff) * CPU_BCLK);
-}
-
-int broadwell_init(struct udevice *dev)
-{
-	struct cpu_broadwell_priv *priv = dev_get_priv(dev);
-	int num_threads;
-	int num_cores;
-	msr_t msr;
-	int ret;
-
-	msr = msr_read(CORE_THREAD_COUNT_MSR);
-	num_threads = (msr.lo >> 0) & 0xffff;
-	num_cores = (msr.lo >> 16) & 0xffff;
-	debug("CPU has %u cores, %u threads enabled\n", num_cores,
-	      num_threads);
-
-	priv->ht_disabled = num_threads == num_cores;
-
-	ret = bsp_init_before_ap_bringup(dev);
-	if (ret)
-		return ret;
-
-	set_max_ratio();
-
-	return ret;
-}
-
-static void configure_mca(void)
-{
-	msr_t msr;
-	const unsigned int mcg_cap_msr = 0x179;
-	int i;
-	int num_banks;
-
-	msr = msr_read(mcg_cap_msr);
-	num_banks = msr.lo & 0xff;
-	msr.lo = 0;
-	msr.hi = 0;
-	/*
-	 * TODO(adurbin): This should only be done on a cold boot. Also, some
-	 * of these banks are core vs package scope. For now every CPU clears
-	 * every bank
-	 */
-	for (i = 0; i < num_banks; i++)
-		msr_write(MSR_IA32_MC0_STATUS + (i * 4), msr);
-}
-
-static void enable_lapic_tpr(void)
-{
-	msr_t msr;
-
-	msr = msr_read(MSR_PIC_MSG_CONTROL);
-	msr.lo &= ~(1 << 10);	/* Enable APIC TPR updates */
-	msr_write(MSR_PIC_MSG_CONTROL, msr);
-}
-
-
-static void configure_c_states(void)
-{
-	msr_t msr;
-
-	msr = msr_read(MSR_PMG_CST_CONFIG_CONTROL);
-	msr.lo |= (1 << 31);	/* Timed MWAIT Enable */
-	msr.lo |= (1 << 30);	/* Package c-state Undemotion Enable */
-	msr.lo |= (1 << 29);	/* Package c-state Demotion Enable */
-	msr.lo |= (1 << 28);	/* C1 Auto Undemotion Enable */
-	msr.lo |= (1 << 27);	/* C3 Auto Undemotion Enable */
-	msr.lo |= (1 << 26);	/* C1 Auto Demotion Enable */
-	msr.lo |= (1 << 25);	/* C3 Auto Demotion Enable */
-	msr.lo &= ~(1 << 10);	/* Disable IO MWAIT redirection */
-	/* The deepest package c-state defaults to factory-configured value */
-	msr_write(MSR_PMG_CST_CONFIG_CONTROL, msr);
-
-	msr = msr_read(MSR_MISC_PWR_MGMT);
-	msr.lo &= ~(1 << 0);	/* Enable P-state HW_ALL coordination */
-	msr_write(MSR_MISC_PWR_MGMT, msr);
-
-	msr = msr_read(MSR_POWER_CTL);
-	msr.lo |= (1 << 18);	/* Enable Energy Perf Bias MSR 0x1b0 */
-	msr.lo |= (1 << 1);	/* C1E Enable */
-	msr.lo |= (1 << 0);	/* Bi-directional PROCHOT# */
-	msr_write(MSR_POWER_CTL, msr);
-
-	/* C-state Interrupt Response Latency Control 0 - package C3 latency */
-	msr.hi = 0;
-	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_0_LIMIT;
-	msr_write(MSR_C_STATE_LATENCY_CONTROL_0, msr);
-
-	/* C-state Interrupt Response Latency Control 1 */
-	msr.hi = 0;
-	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_1_LIMIT;
-	msr_write(MSR_C_STATE_LATENCY_CONTROL_1, msr);
-
-	/* C-state Interrupt Response Latency Control 2 - package C6/C7 short */
-	msr.hi = 0;
-	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_2_LIMIT;
-	msr_write(MSR_C_STATE_LATENCY_CONTROL_2, msr);
-
-	/* C-state Interrupt Response Latency Control 3 - package C8 */
-	msr.hi = 0;
-	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_3_LIMIT;
-	msr_write(MSR_C_STATE_LATENCY_CONTROL_3, msr);
-
-	/* C-state Interrupt Response Latency Control 4 - package C9 */
-	msr.hi = 0;
-	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_4_LIMIT;
-	msr_write(MSR_C_STATE_LATENCY_CONTROL_4, msr);
-
-	/* C-state Interrupt Response Latency Control 5 - package C10 */
-	msr.hi = 0;
-	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_5_LIMIT;
-	msr_write(MSR_C_STATE_LATENCY_CONTROL_5, msr);
-}
-
-static void configure_misc(void)
-{
-	msr_t msr;
-
-	msr = msr_read(MSR_IA32_MISC_ENABLE);
-	msr.lo |= (1 << 0);	  /* Fast String enable */
-	msr.lo |= (1 << 3);	  /* TM1/TM2/EMTTM enable */
-	msr.lo |= (1 << 16);	  /* Enhanced SpeedStep Enable */
-	msr_write(MSR_IA32_MISC_ENABLE, msr);
-
-	/* Disable thermal interrupts */
-	msr.lo = 0;
-	msr.hi = 0;
-	msr_write(MSR_IA32_THERM_INTERRUPT, msr);
-
-	/* Enable package critical interrupt only */
-	msr.lo = 1 << 4;
-	msr.hi = 0;
-	msr_write(MSR_IA32_PACKAGE_THERM_INTERRUPT, msr);
-}
-
-static void configure_thermal_target(struct udevice *dev)
-{
-	int tcc_offset;
-	msr_t msr;
-
-	tcc_offset = fdtdec_get_int(gd->fdt_blob, dev_of_offset(dev),
-				    "intel,tcc-offset", 0);
-
-	/* Set TCC activaiton offset if supported */
-	msr = msr_read(MSR_PLATFORM_INFO);
-	if ((msr.lo & (1 << 30)) && tcc_offset) {
-		msr = msr_read(MSR_TEMPERATURE_TARGET);
-		msr.lo &= ~(0xf << 24); /* Bits 27:24 */
-		msr.lo |= (tcc_offset & 0xf) << 24;
-		msr_write(MSR_TEMPERATURE_TARGET, msr);
-	}
-}
-
-static void configure_dca_cap(void)
-{
-	struct cpuid_result cpuid_regs;
-	msr_t msr;
-
-	/* Check feature flag in CPUID.(EAX=1):ECX[18]==1 */
-	cpuid_regs = cpuid(1);
-	if (cpuid_regs.ecx & (1 << 18)) {
-		msr = msr_read(MSR_IA32_PLATFORM_DCA_CAP);
-		msr.lo |= 1;
-		msr_write(MSR_IA32_PLATFORM_DCA_CAP, msr);
-	}
-}
-
-static void set_energy_perf_bias(u8 policy)
-{
-	msr_t msr;
-	int ecx;
-
-	/* Determine if energy efficient policy is supported */
-	ecx = cpuid_ecx(0x6);
-	if (!(ecx & (1 << 3)))
-		return;
-
-	/* Energy Policy is bits 3:0 */
-	msr = msr_read(MSR_IA32_ENERGY_PERFORMANCE_BIAS);
-	msr.lo &= ~0xf;
-	msr.lo |= policy & 0xf;
-	msr_write(MSR_IA32_ENERGY_PERFORMANCE_BIAS, msr);
-
-	debug("cpu: energy policy set to %u\n", policy);
-}
-
-/* All CPUs including BSP will run the following function */
-static void cpu_core_init(struct udevice *dev)
-{
-	/* Clear out pending MCEs */
-	configure_mca();
-
-	/* Enable the local cpu apics */
-	enable_lapic_tpr();
-
-	/* Configure C States */
-	configure_c_states();
-
-	/* Configure Enhanced SpeedStep and Thermal Sensors */
-	configure_misc();
-
-	/* Thermal throttle activation offset */
-	configure_thermal_target(dev);
-
-	/* Enable Direct Cache Access */
-	configure_dca_cap();
-
-	/* Set energy policy */
-	set_energy_perf_bias(ENERGY_POLICY_NORMAL);
-
-	/* Enable Turbo */
-	turbo_enable();
-}
-
-/*
- * Configure processor power limits if possible
- * This must be done AFTER set of BIOS_RESET_CPL
- */
-void cpu_set_power_limits(int power_limit_1_time)
-{
-	msr_t msr;
-	msr_t limit;
-	unsigned power_unit;
-	unsigned tdp, min_power, max_power, max_time;
-	u8 power_limit_1_val;
-
-	msr = msr_read(MSR_PLATFORM_INFO);
-	if (power_limit_1_time > ARRAY_SIZE(power_limit_time_sec_to_msr))
-		power_limit_1_time = 28;
-
-	if (!(msr.lo & PLATFORM_INFO_SET_TDP))
-		return;
-
-	/* Get units */
-	msr = msr_read(MSR_PKG_POWER_SKU_UNIT);
-	power_unit = 2 << ((msr.lo & 0xf) - 1);
-
-	/* Get power defaults for this SKU */
-	msr = msr_read(MSR_PKG_POWER_SKU);
-	tdp = msr.lo & 0x7fff;
-	min_power = (msr.lo >> 16) & 0x7fff;
-	max_power = msr.hi & 0x7fff;
-	max_time = (msr.hi >> 16) & 0x7f;
-
-	debug("CPU TDP: %u Watts\n", tdp / power_unit);
-
-	if (power_limit_time_msr_to_sec[max_time] > power_limit_1_time)
-		power_limit_1_time = power_limit_time_msr_to_sec[max_time];
-
-	if (min_power > 0 && tdp < min_power)
-		tdp = min_power;
-
-	if (max_power > 0 && tdp > max_power)
-		tdp = max_power;
-
-	power_limit_1_val = power_limit_time_sec_to_msr[power_limit_1_time];
-
-	/* Set long term power limit to TDP */
-	limit.lo = 0;
-	limit.lo |= tdp & PKG_POWER_LIMIT_MASK;
-	limit.lo |= PKG_POWER_LIMIT_EN;
-	limit.lo |= (power_limit_1_val & PKG_POWER_LIMIT_TIME_MASK) <<
-		PKG_POWER_LIMIT_TIME_SHIFT;
-
-	/* Set short term power limit to 1.25 * TDP */
-	limit.hi = 0;
-	limit.hi |= ((tdp * 125) / 100) & PKG_POWER_LIMIT_MASK;
-	limit.hi |= PKG_POWER_LIMIT_EN;
-	/* Power limit 2 time is only programmable on server SKU */
-
-	msr_write(MSR_PKG_POWER_LIMIT, limit);
-
-	/* Set power limit values in MCHBAR as well */
-	writel(limit.lo, MCHBAR_REG(MCH_PKG_POWER_LIMIT_LO));
-	writel(limit.hi, MCHBAR_REG(MCH_PKG_POWER_LIMIT_HI));
-
-	/* Set DDR RAPL power limit by copying from MMIO to MSR */
-	msr.lo = readl(MCHBAR_REG(MCH_DDR_POWER_LIMIT_LO));
-	msr.hi = readl(MCHBAR_REG(MCH_DDR_POWER_LIMIT_HI));
-	msr_write(MSR_DDR_RAPL_LIMIT, msr);
-
-	/* Use nominal TDP values for CPUs with configurable TDP */
-	if (cpu_config_tdp_levels()) {
-		msr = msr_read(MSR_CONFIG_TDP_NOMINAL);
-		limit.hi = 0;
-		limit.lo = msr.lo & 0xff;
-		msr_write(MSR_TURBO_ACTIVATION_RATIO, limit);
-	}
-}
-
-static int broadwell_get_info(struct udevice *dev, struct cpu_info *info)
-{
-	msr_t msr;
-
-	msr = msr_read(IA32_PERF_CTL);
-	info->cpu_freq = ((msr.lo >> 8) & 0xff) * BROADWELL_BCLK * 1000000;
-	info->features = 1 << CPU_FEAT_L1_CACHE | 1 << CPU_FEAT_MMU |
-		1 << CPU_FEAT_UCODE | 1 << CPU_FEAT_DEVICE_ID;
-
-	return 0;
-}
-
-static int broadwell_get_count(struct udevice *dev)
+void board_debug_uart_init(void)
 {
-	return 4;
-}
+	struct udevice *bus = NULL;
 
-static int cpu_x86_broadwell_probe(struct udevice *dev)
-{
-	if (dev->seq == 0) {
-		cpu_core_init(dev);
-		return broadwell_init(dev);
-	}
+	/* com1 / com2 decode range */
+	pci_x86_write_config(bus, PCH_DEV_LPC, LPC_IO_DEC, 1 << 4, PCI_SIZE_16);
 
-	return 0;
+	pci_x86_write_config(bus, PCH_DEV_LPC, LPC_EN, COMA_LPC_EN,
+			     PCI_SIZE_16);
 }
-
-static const struct cpu_ops cpu_x86_broadwell_ops = {
-	.get_desc	= cpu_x86_get_desc,
-	.get_info	= broadwell_get_info,
-	.get_count	= broadwell_get_count,
-	.get_vendor	= cpu_x86_get_vendor,
-};
-
-static const struct udevice_id cpu_x86_broadwell_ids[] = {
-	{ .compatible = "intel,core-i3-gen5" },
-	{ }
-};
-
-U_BOOT_DRIVER(cpu_x86_broadwell_drv) = {
-	.name		= "cpu_x86_broadwell",
-	.id		= UCLASS_CPU,
-	.of_match	= cpu_x86_broadwell_ids,
-	.bind		= cpu_x86_bind,
-	.probe		= cpu_x86_broadwell_probe,
-	.ops		= &cpu_x86_broadwell_ops,
-	.priv_auto_alloc_size	= sizeof(struct cpu_broadwell_priv),
-	.flags		= DM_FLAG_PRE_RELOC,
-};
diff --git a/arch/x86/cpu/broadwell/cpu_from_spl.c b/arch/x86/cpu/broadwell/cpu_from_spl.c
new file mode 100644
index 0000000000..c3d4a8d547
--- /dev/null
+++ b/arch/x86/cpu/broadwell/cpu_from_spl.c
@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * (C) Copyright 2016 Google, Inc
+ * Written by Simon Glass <sjg@chromium.org>
+ */
+
+#include <common.h>
+#include <bloblist.h>
+#include <debug_uart.h>
+#include <handoff.h>
+#include <asm/mtrr.h>
+
+int misc_init_r(void)
+{
+	return 0;
+}
+
+int dram_init(void)
+{
+	struct spl_handoff *ho;
+
+	ho = bloblist_find(BLOBLISTT_SPL_HANDOFF, sizeof(*ho));
+	if (!ho)
+		return log_msg_ret("Missing SPL hand-off info", -ENOENT);
+	handoff_load_dram_size(ho);
+#ifdef CONFIG_TPL
+	/* TODO(sjg@chromium.org): MTRR cannot be adjusted without a hang */
+	mtrr_add_request(MTRR_TYPE_WRBACK, 0, 2ULL << 30);
+#else
+	mtrr_add_request(MTRR_TYPE_WRBACK, 0, gd->ram_size);
+	mtrr_commit(true);
+#endif
+
+	return 0;
+}
+
+int checkcpu(void)
+{
+	return 0;
+}
+
+int print_cpuinfo(void)
+{
+	return 0;
+}
+
+void board_debug_uart_init(void)
+{
+}
+
+int dram_init_banksize(void)
+{
+#ifdef CONFIG_NR_DRAM_BANKS
+	struct spl_handoff *ho;
+
+	ho = bloblist_find(BLOBLISTT_SPL_HANDOFF, sizeof(*ho));
+	if (!ho)
+		return log_msg_ret("Missing SPL hand-off info", -ENOENT);
+	handoff_load_dram_banks(ho);
+#endif
+
+	return 0;
+}
diff --git a/arch/x86/cpu/broadwell/cpu_full.c b/arch/x86/cpu/broadwell/cpu_full.c
new file mode 100644
index 0000000000..c1db184549
--- /dev/null
+++ b/arch/x86/cpu/broadwell/cpu_full.c
@@ -0,0 +1,694 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 Google, Inc
+ *
+ * Based on code from coreboot src/soc/intel/broadwell/cpu.c
+ */
+
+#include <common.h>
+#include <dm.h>
+#include <cpu.h>
+#include <asm/cpu.h>
+#include <asm/cpu_x86.h>
+#include <asm/cpu_common.h>
+#include <asm/intel_regs.h>
+#include <asm/msr.h>
+#include <asm/post.h>
+#include <asm/turbo.h>
+#include <asm/arch/cpu.h>
+#include <asm/arch/pch.h>
+#include <asm/arch/rcb.h>
+
+struct cpu_broadwell_priv {
+	bool ht_disabled;
+};
+
+/* Convert time in seconds to POWER_LIMIT_1_TIME MSR value */
+static const u8 power_limit_time_sec_to_msr[] = {
+	[0]   = 0x00,
+	[1]   = 0x0a,
+	[2]   = 0x0b,
+	[3]   = 0x4b,
+	[4]   = 0x0c,
+	[5]   = 0x2c,
+	[6]   = 0x4c,
+	[7]   = 0x6c,
+	[8]   = 0x0d,
+	[10]  = 0x2d,
+	[12]  = 0x4d,
+	[14]  = 0x6d,
+	[16]  = 0x0e,
+	[20]  = 0x2e,
+	[24]  = 0x4e,
+	[28]  = 0x6e,
+	[32]  = 0x0f,
+	[40]  = 0x2f,
+	[48]  = 0x4f,
+	[56]  = 0x6f,
+	[64]  = 0x10,
+	[80]  = 0x30,
+	[96]  = 0x50,
+	[112] = 0x70,
+	[128] = 0x11,
+};
+
+/* Convert POWER_LIMIT_1_TIME MSR value to seconds */
+static const u8 power_limit_time_msr_to_sec[] = {
+	[0x00] = 0,
+	[0x0a] = 1,
+	[0x0b] = 2,
+	[0x4b] = 3,
+	[0x0c] = 4,
+	[0x2c] = 5,
+	[0x4c] = 6,
+	[0x6c] = 7,
+	[0x0d] = 8,
+	[0x2d] = 10,
+	[0x4d] = 12,
+	[0x6d] = 14,
+	[0x0e] = 16,
+	[0x2e] = 20,
+	[0x4e] = 24,
+	[0x6e] = 28,
+	[0x0f] = 32,
+	[0x2f] = 40,
+	[0x4f] = 48,
+	[0x6f] = 56,
+	[0x10] = 64,
+	[0x30] = 80,
+	[0x50] = 96,
+	[0x70] = 112,
+	[0x11] = 128,
+};
+
+/*
+ * The core 100MHz BLCK is disabled in deeper c-states. One needs to calibrate
+ * the 100MHz BCLCK against the 24MHz BLCK to restore the clocks properly
+ * when a core is woken up
+ */
+static int pcode_ready(void)
+{
+	int wait_count;
+	const int delay_step = 10;
+
+	wait_count = 0;
+	do {
+		if (!(readl(MCHBAR_REG(BIOS_MAILBOX_INTERFACE)) &
+				MAILBOX_RUN_BUSY))
+			return 0;
+		wait_count += delay_step;
+		udelay(delay_step);
+	} while (wait_count < 1000);
+
+	return -ETIMEDOUT;
+}
+
+static u32 pcode_mailbox_read(u32 command)
+{
+	int ret;
+
+	ret = pcode_ready();
+	if (ret) {
+		debug("PCODE: mailbox timeout on wait ready\n");
+		return ret;
+	}
+
+	/* Send command and start transaction */
+	writel(command | MAILBOX_RUN_BUSY, MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
+
+	ret = pcode_ready();
+	if (ret) {
+		debug("PCODE: mailbox timeout on completion\n");
+		return ret;
+	}
+
+	/* Read mailbox */
+	return readl(MCHBAR_REG(BIOS_MAILBOX_DATA));
+}
+
+static int pcode_mailbox_write(u32 command, u32 data)
+{
+	int ret;
+
+	ret = pcode_ready();
+	if (ret) {
+		debug("PCODE: mailbox timeout on wait ready\n");
+		return ret;
+	}
+
+	writel(data, MCHBAR_REG(BIOS_MAILBOX_DATA));
+
+	/* Send command and start transaction */
+	writel(command | MAILBOX_RUN_BUSY, MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
+
+	ret = pcode_ready();
+	if (ret) {
+		debug("PCODE: mailbox timeout on completion\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+/* @dev is the CPU device */
+static void initialize_vr_config(struct udevice *dev)
+{
+	int ramp, min_vid;
+	msr_t msr;
+
+	debug("Initializing VR config\n");
+
+	/* Configure VR_CURRENT_CONFIG */
+	msr = msr_read(MSR_VR_CURRENT_CONFIG);
+	/*
+	 * Preserve bits 63 and 62. Bit 62 is PSI4 enable, but it is only valid
+	 * on ULT systems
+	 */
+	msr.hi &= 0xc0000000;
+	msr.hi |= (0x01 << (52 - 32)); /* PSI3 threshold -  1A */
+	msr.hi |= (0x05 << (42 - 32)); /* PSI2 threshold -  5A */
+	msr.hi |= (0x14 << (32 - 32)); /* PSI1 threshold - 20A */
+	msr.hi |= (1 <<  (62 - 32)); /* Enable PSI4 */
+	/* Leave the max instantaneous current limit (12:0) to default */
+	msr_write(MSR_VR_CURRENT_CONFIG, msr);
+
+	/* Configure VR_MISC_CONFIG MSR */
+	msr = msr_read(MSR_VR_MISC_CONFIG);
+	/* Set the IOUT_SLOPE scalar applied to dIout in U10.1.9 format */
+	msr.hi &= ~(0x3ff << (40 - 32));
+	msr.hi |= (0x200 << (40 - 32)); /* 1.0 */
+	/* Set IOUT_OFFSET to 0 */
+	msr.hi &= ~0xff;
+	/* Set entry ramp rate to slow */
+	msr.hi &= ~(1 << (51 - 32));
+	/* Enable decay mode on C-state entry */
+	msr.hi |= (1 << (52 - 32));
+	/* Set the slow ramp rate */
+	msr.hi &= ~(0x3 << (53 - 32));
+	/* Configure the C-state exit ramp rate */
+	ramp = fdtdec_get_int(gd->fdt_blob, dev_of_offset(dev),
+			      "intel,slow-ramp", -1);
+	if (ramp != -1) {
+		/* Configured slow ramp rate */
+		msr.hi |= ((ramp & 0x3) << (53 - 32));
+		/* Set exit ramp rate to slow */
+		msr.hi &= ~(1 << (50 - 32));
+	} else {
+		/* Fast ramp rate / 4 */
+		msr.hi |= (0x01 << (53 - 32));
+		/* Set exit ramp rate to fast */
+		msr.hi |= (1 << (50 - 32));
+	}
+	/* Set MIN_VID (31:24) to allow CPU to have full control */
+	msr.lo &= ~0xff000000;
+	min_vid = fdtdec_get_int(gd->fdt_blob, dev_of_offset(dev),
+				 "intel,min-vid", 0);
+	msr.lo |= (min_vid & 0xff) << 24;
+	msr_write(MSR_VR_MISC_CONFIG, msr);
+
+	/*  Configure VR_MISC_CONFIG2 MSR */
+	msr = msr_read(MSR_VR_MISC_CONFIG2);
+	msr.lo &= ~0xffff;
+	/*
+	 * Allow CPU to control minimum voltage completely (15:8) and
+	 * set the fast ramp voltage in 10mV steps
+	 */
+	if (cpu_get_family_model() == BROADWELL_FAMILY_ULT)
+		msr.lo |= 0x006a; /* 1.56V */
+	else
+		msr.lo |= 0x006f; /* 1.60V */
+	msr_write(MSR_VR_MISC_CONFIG2, msr);
+
+	/* Set C9/C10 VCC Min */
+	pcode_mailbox_write(MAILBOX_BIOS_CMD_WRITE_C9C10_VOLTAGE, 0x1f1f);
+}
+
+static int calibrate_24mhz_bclk(void)
+{
+	int err_code;
+	int ret;
+
+	ret = pcode_ready();
+	if (ret)
+		return ret;
+
+	/* A non-zero value initiates the PCODE calibration */
+	writel(~0, MCHBAR_REG(BIOS_MAILBOX_DATA));
+	writel(MAILBOX_RUN_BUSY | MAILBOX_BIOS_CMD_FSM_MEASURE_INTVL,
+	       MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
+
+	ret = pcode_ready();
+	if (ret)
+		return ret;
+
+	err_code = readl(MCHBAR_REG(BIOS_MAILBOX_INTERFACE)) & 0xff;
+
+	debug("PCODE: 24MHz BLCK calibration response: %d\n", err_code);
+
+	/* Read the calibrated value */
+	writel(MAILBOX_RUN_BUSY | MAILBOX_BIOS_CMD_READ_CALIBRATION,
+	       MCHBAR_REG(BIOS_MAILBOX_INTERFACE));
+
+	ret = pcode_ready();
+	if (ret)
+		return ret;
+
+	debug("PCODE: 24MHz BLCK calibration value: 0x%08x\n",
+	      readl(MCHBAR_REG(BIOS_MAILBOX_DATA)));
+
+	return 0;
+}
+
+static void configure_pch_power_sharing(void)
+{
+	u32 pch_power, pch_power_ext, pmsync, pmsync2;
+	int i;
+
+	/* Read PCH Power levels from PCODE */
+	pch_power = pcode_mailbox_read(MAILBOX_BIOS_CMD_READ_PCH_POWER);
+	pch_power_ext = pcode_mailbox_read(MAILBOX_BIOS_CMD_READ_PCH_POWER_EXT);
+
+	debug("PCH Power: PCODE Levels 0x%08x 0x%08x\n", pch_power,
+	      pch_power_ext);
+
+	pmsync = readl(RCB_REG(PMSYNC_CONFIG));
+	pmsync2 = readl(RCB_REG(PMSYNC_CONFIG2));
+
+	/*
+	 * Program PMSYNC_TPR_CONFIG PCH power limit values
+	 *  pmsync[0:4]   = mailbox[0:5]
+	 *  pmsync[8:12]  = mailbox[6:11]
+	 *  pmsync[16:20] = mailbox[12:17]
+	 */
+	for (i = 0; i < 3; i++) {
+		u32 level = pch_power & 0x3f;
+
+		pch_power >>= 6;
+		pmsync &= ~(0x1f << (i * 8));
+		pmsync |= (level & 0x1f) << (i * 8);
+	}
+	writel(pmsync, RCB_REG(PMSYNC_CONFIG));
+
+	/*
+	 * Program PMSYNC_TPR_CONFIG2 Extended PCH power limit values
+	 *  pmsync2[0:4]   = mailbox[23:18]
+	 *  pmsync2[8:12]  = mailbox_ext[6:11]
+	 *  pmsync2[16:20] = mailbox_ext[12:17]
+	 *  pmsync2[24:28] = mailbox_ext[18:22]
+	 */
+	pmsync2 &= ~0x1f;
+	pmsync2 |= pch_power & 0x1f;
+
+	for (i = 1; i < 4; i++) {
+		u32 level = pch_power_ext & 0x3f;
+
+		pch_power_ext >>= 6;
+		pmsync2 &= ~(0x1f << (i * 8));
+		pmsync2 |= (level & 0x1f) << (i * 8);
+	}
+	writel(pmsync2, RCB_REG(PMSYNC_CONFIG2));
+}
+
+static int bsp_init_before_ap_bringup(struct udevice *dev)
+{
+	int ret;
+
+	initialize_vr_config(dev);
+	ret = calibrate_24mhz_bclk();
+	if (ret)
+		return ret;
+	configure_pch_power_sharing();
+
+	return 0;
+}
+
+static int cpu_config_tdp_levels(void)
+{
+	msr_t platform_info;
+
+	/* Bits 34:33 indicate how many levels supported */
+	platform_info = msr_read(MSR_PLATFORM_INFO);
+	return (platform_info.hi >> 1) & 3;
+}
+
+static void set_max_ratio(void)
+{
+	msr_t msr, perf_ctl;
+
+	perf_ctl.hi = 0;
+
+	/* Check for configurable TDP option */
+	if (turbo_get_state() == TURBO_ENABLED) {
+		msr = msr_read(MSR_NHM_TURBO_RATIO_LIMIT);
+		perf_ctl.lo = (msr.lo & 0xff) << 8;
+	} else if (cpu_config_tdp_levels()) {
+		/* Set to nominal TDP ratio */
+		msr = msr_read(MSR_CONFIG_TDP_NOMINAL);
+		perf_ctl.lo = (msr.lo & 0xff) << 8;
+	} else {
+		/* Platform Info bits 15:8 give max ratio */
+		msr = msr_read(MSR_PLATFORM_INFO);
+		perf_ctl.lo = msr.lo & 0xff00;
+	}
+	msr_write(IA32_PERF_CTL, perf_ctl);
+
+	debug("cpu: frequency set to %d\n",
+	      ((perf_ctl.lo >> 8) & 0xff) * CPU_BCLK);
+}
+
+int broadwell_init(struct udevice *dev)
+{
+	struct cpu_broadwell_priv *priv = dev_get_priv(dev);
+	int num_threads;
+	int num_cores;
+	msr_t msr;
+	int ret;
+
+	msr = msr_read(CORE_THREAD_COUNT_MSR);
+	num_threads = (msr.lo >> 0) & 0xffff;
+	num_cores = (msr.lo >> 16) & 0xffff;
+	debug("CPU has %u cores, %u threads enabled\n", num_cores,
+	      num_threads);
+
+	priv->ht_disabled = num_threads == num_cores;
+
+	ret = bsp_init_before_ap_bringup(dev);
+	if (ret)
+		return ret;
+
+	set_max_ratio();
+
+	return ret;
+}
+
+static void configure_mca(void)
+{
+	msr_t msr;
+	const unsigned int mcg_cap_msr = 0x179;
+	int i;
+	int num_banks;
+
+	msr = msr_read(mcg_cap_msr);
+	num_banks = msr.lo & 0xff;
+	msr.lo = 0;
+	msr.hi = 0;
+	/*
+	 * TODO(adurbin): This should only be done on a cold boot. Also, some
+	 * of these banks are core vs package scope. For now every CPU clears
+	 * every bank
+	 */
+	for (i = 0; i < num_banks; i++)
+		msr_write(MSR_IA32_MC0_STATUS + (i * 4), msr);
+}
+
+static void enable_lapic_tpr(void)
+{
+	msr_t msr;
+
+	msr = msr_read(MSR_PIC_MSG_CONTROL);
+	msr.lo &= ~(1 << 10);	/* Enable APIC TPR updates */
+	msr_write(MSR_PIC_MSG_CONTROL, msr);
+}
+
+static void configure_c_states(void)
+{
+	msr_t msr;
+
+	msr = msr_read(MSR_PMG_CST_CONFIG_CONTROL);
+	msr.lo |= (1 << 31);	/* Timed MWAIT Enable */
+	msr.lo |= (1 << 30);	/* Package c-state Undemotion Enable */
+	msr.lo |= (1 << 29);	/* Package c-state Demotion Enable */
+	msr.lo |= (1 << 28);	/* C1 Auto Undemotion Enable */
+	msr.lo |= (1 << 27);	/* C3 Auto Undemotion Enable */
+	msr.lo |= (1 << 26);	/* C1 Auto Demotion Enable */
+	msr.lo |= (1 << 25);	/* C3 Auto Demotion Enable */
+	msr.lo &= ~(1 << 10);	/* Disable IO MWAIT redirection */
+	/* The deepest package c-state defaults to factory-configured value */
+	msr_write(MSR_PMG_CST_CONFIG_CONTROL, msr);
+
+	msr = msr_read(MSR_MISC_PWR_MGMT);
+	msr.lo &= ~(1 << 0);	/* Enable P-state HW_ALL coordination */
+	msr_write(MSR_MISC_PWR_MGMT, msr);
+
+	msr = msr_read(MSR_POWER_CTL);
+	msr.lo |= (1 << 18);	/* Enable Energy Perf Bias MSR 0x1b0 */
+	msr.lo |= (1 << 1);	/* C1E Enable */
+	msr.lo |= (1 << 0);	/* Bi-directional PROCHOT# */
+	msr_write(MSR_POWER_CTL, msr);
+
+	/* C-state Interrupt Response Latency Control 0 - package C3 latency */
+	msr.hi = 0;
+	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_0_LIMIT;
+	msr_write(MSR_C_STATE_LATENCY_CONTROL_0, msr);
+
+	/* C-state Interrupt Response Latency Control 1 */
+	msr.hi = 0;
+	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_1_LIMIT;
+	msr_write(MSR_C_STATE_LATENCY_CONTROL_1, msr);
+
+	/* C-state Interrupt Response Latency Control 2 - package C6/C7 short */
+	msr.hi = 0;
+	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_2_LIMIT;
+	msr_write(MSR_C_STATE_LATENCY_CONTROL_2, msr);
+
+	/* C-state Interrupt Response Latency Control 3 - package C8 */
+	msr.hi = 0;
+	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_3_LIMIT;
+	msr_write(MSR_C_STATE_LATENCY_CONTROL_3, msr);
+
+	/* C-state Interrupt Response Latency Control 4 - package C9 */
+	msr.hi = 0;
+	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_4_LIMIT;
+	msr_write(MSR_C_STATE_LATENCY_CONTROL_4, msr);
+
+	/* C-state Interrupt Response Latency Control 5 - package C10 */
+	msr.hi = 0;
+	msr.lo = IRTL_VALID | IRTL_1024_NS | C_STATE_LATENCY_CONTROL_5_LIMIT;
+	msr_write(MSR_C_STATE_LATENCY_CONTROL_5, msr);
+}
+
+static void configure_misc(void)
+{
+	msr_t msr;
+
+	msr = msr_read(MSR_IA32_MISC_ENABLE);
+	msr.lo |= (1 << 0);	  /* Fast String enable */
+	msr.lo |= (1 << 3);	  /* TM1/TM2/EMTTM enable */
+	msr.lo |= (1 << 16);	  /* Enhanced SpeedStep Enable */
+	msr_write(MSR_IA32_MISC_ENABLE, msr);
+
+	/* Disable thermal interrupts */
+	msr.lo = 0;
+	msr.hi = 0;
+	msr_write(MSR_IA32_THERM_INTERRUPT, msr);
+
+	/* Enable package critical interrupt only */
+	msr.lo = 1 << 4;
+	msr.hi = 0;
+	msr_write(MSR_IA32_PACKAGE_THERM_INTERRUPT, msr);
+}
+
+static void configure_thermal_target(struct udevice *dev)
+{
+	int tcc_offset;
+	msr_t msr;
+
+	tcc_offset = fdtdec_get_int(gd->fdt_blob, dev_of_offset(dev),
+				    "intel,tcc-offset", 0);
+
+	/* Set TCC activaiton offset if supported */
+	msr = msr_read(MSR_PLATFORM_INFO);
+	if ((msr.lo & (1 << 30)) && tcc_offset) {
+		msr = msr_read(MSR_TEMPERATURE_TARGET);
+		msr.lo &= ~(0xf << 24); /* Bits 27:24 */
+		msr.lo |= (tcc_offset & 0xf) << 24;
+		msr_write(MSR_TEMPERATURE_TARGET, msr);
+	}
+}
+
+static void configure_dca_cap(void)
+{
+	struct cpuid_result cpuid_regs;
+	msr_t msr;
+
+	/* Check feature flag in CPUID.(EAX=1):ECX[18]==1 */
+	cpuid_regs = cpuid(1);
+	if (cpuid_regs.ecx & (1 << 18)) {
+		msr = msr_read(MSR_IA32_PLATFORM_DCA_CAP);
+		msr.lo |= 1;
+		msr_write(MSR_IA32_PLATFORM_DCA_CAP, msr);
+	}
+}
+
+static void set_energy_perf_bias(u8 policy)
+{
+	msr_t msr;
+	int ecx;
+
+	/* Determine if energy efficient policy is supported */
+	ecx = cpuid_ecx(0x6);
+	if (!(ecx & (1 << 3)))
+		return;
+
+	/* Energy Policy is bits 3:0 */
+	msr = msr_read(MSR_IA32_ENERGY_PERFORMANCE_BIAS);
+	msr.lo &= ~0xf;
+	msr.lo |= policy & 0xf;
+	msr_write(MSR_IA32_ENERGY_PERFORMANCE_BIAS, msr);
+
+	debug("cpu: energy policy set to %u\n", policy);
+}
+
+/* All CPUs including BSP will run the following function */
+static void cpu_core_init(struct udevice *dev)
+{
+	/* Clear out pending MCEs */
+	configure_mca();
+
+	/* Enable the local cpu apics */
+	enable_lapic_tpr();
+
+	/* Configure C States */
+	configure_c_states();
+
+	/* Configure Enhanced SpeedStep and Thermal Sensors */
+	configure_misc();
+
+	/* Thermal throttle activation offset */
+	configure_thermal_target(dev);
+
+	/* Enable Direct Cache Access */
+	configure_dca_cap();
+
+	/* Set energy policy */
+	set_energy_perf_bias(ENERGY_POLICY_NORMAL);
+
+	/* Enable Turbo */
+	turbo_enable();
+}
+
+/*
+ * Configure processor power limits if possible
+ * This must be done AFTER set of BIOS_RESET_CPL
+ */
+void cpu_set_power_limits(int power_limit_1_time)
+{
+	msr_t msr;
+	msr_t limit;
+	uint power_unit;
+	uint tdp, min_power, max_power, max_time;
+	u8 power_limit_1_val;
+
+	msr = msr_read(MSR_PLATFORM_INFO);
+	if (power_limit_1_time > ARRAY_SIZE(power_limit_time_sec_to_msr))
+		power_limit_1_time = 28;
+
+	if (!(msr.lo & PLATFORM_INFO_SET_TDP))
+		return;
+
+	/* Get units */
+	msr = msr_read(MSR_PKG_POWER_SKU_UNIT);
+	power_unit = 2 << ((msr.lo & 0xf) - 1);
+
+	/* Get power defaults for this SKU */
+	msr = msr_read(MSR_PKG_POWER_SKU);
+	tdp = msr.lo & 0x7fff;
+	min_power = (msr.lo >> 16) & 0x7fff;
+	max_power = msr.hi & 0x7fff;
+	max_time = (msr.hi >> 16) & 0x7f;
+
+	debug("CPU TDP: %u Watts\n", tdp / power_unit);
+
+	if (power_limit_time_msr_to_sec[max_time] > power_limit_1_time)
+		power_limit_1_time = power_limit_time_msr_to_sec[max_time];
+
+	if (min_power > 0 && tdp < min_power)
+		tdp = min_power;
+
+	if (max_power > 0 && tdp > max_power)
+		tdp = max_power;
+
+	power_limit_1_val = power_limit_time_sec_to_msr[power_limit_1_time];
+
+	/* Set long term power limit to TDP */
+	limit.lo = 0;
+	limit.lo |= tdp & PKG_POWER_LIMIT_MASK;
+	limit.lo |= PKG_POWER_LIMIT_EN;
+	limit.lo |= (power_limit_1_val & PKG_POWER_LIMIT_TIME_MASK) <<
+		PKG_POWER_LIMIT_TIME_SHIFT;
+
+	/* Set short term power limit to 1.25 * TDP */
+	limit.hi = 0;
+	limit.hi |= ((tdp * 125) / 100) & PKG_POWER_LIMIT_MASK;
+	limit.hi |= PKG_POWER_LIMIT_EN;
+	/* Power limit 2 time is only programmable on server SKU */
+
+	msr_write(MSR_PKG_POWER_LIMIT, limit);
+
+	/* Set power limit values in MCHBAR as well */
+	writel(limit.lo, MCHBAR_REG(MCH_PKG_POWER_LIMIT_LO));
+	writel(limit.hi, MCHBAR_REG(MCH_PKG_POWER_LIMIT_HI));
+
+	/* Set DDR RAPL power limit by copying from MMIO to MSR */
+	msr.lo = readl(MCHBAR_REG(MCH_DDR_POWER_LIMIT_LO));
+	msr.hi = readl(MCHBAR_REG(MCH_DDR_POWER_LIMIT_HI));
+	msr_write(MSR_DDR_RAPL_LIMIT, msr);
+
+	/* Use nominal TDP values for CPUs with configurable TDP */
+	if (cpu_config_tdp_levels()) {
+		msr = msr_read(MSR_CONFIG_TDP_NOMINAL);
+		limit.hi = 0;
+		limit.lo = msr.lo & 0xff;
+		msr_write(MSR_TURBO_ACTIVATION_RATIO, limit);
+	}
+}
+
+static int broadwell_get_info(struct udevice *dev, struct cpu_info *info)
+{
+	msr_t msr;
+
+	msr = msr_read(IA32_PERF_CTL);
+	info->cpu_freq = ((msr.lo >> 8) & 0xff) * BROADWELL_BCLK * 1000000;
+	info->features = 1 << CPU_FEAT_L1_CACHE | 1 << CPU_FEAT_MMU |
+		1 << CPU_FEAT_UCODE | 1 << CPU_FEAT_DEVICE_ID;
+
+	return 0;
+}
+
+static int broadwell_get_count(struct udevice *dev)
+{
+	return 4;
+}
+
+static int cpu_x86_broadwell_probe(struct udevice *dev)
+{
+	if (dev->seq == 0) {
+		cpu_core_init(dev);
+		return broadwell_init(dev);
+	}
+
+	return 0;
+}
+
+static const struct cpu_ops cpu_x86_broadwell_ops = {
+	.get_desc	= cpu_x86_get_desc,
+	.get_info	= broadwell_get_info,
+	.get_count	= broadwell_get_count,
+	.get_vendor	= cpu_x86_get_vendor,
+};
+
+static const struct udevice_id cpu_x86_broadwell_ids[] = {
+	{ .compatible = "intel,core-i3-gen5" },
+	{ }
+};
+
+U_BOOT_DRIVER(cpu_x86_broadwell_drv) = {
+	.name		= "cpu_x86_broadwell",
+	.id		= UCLASS_CPU,
+	.of_match	= cpu_x86_broadwell_ids,
+	.bind		= cpu_x86_bind,
+	.probe		= cpu_x86_broadwell_probe,
+	.ops		= &cpu_x86_broadwell_ops,
+	.priv_auto_alloc_size	= sizeof(struct cpu_broadwell_priv),
+	.flags		= DM_FLAG_PRE_RELOC,
+};
diff --git a/arch/x86/cpu/broadwell/northbridge.c b/arch/x86/cpu/broadwell/northbridge.c
index 3055880bb7..4bcab78556 100644
--- a/arch/x86/cpu/broadwell/northbridge.c
+++ b/arch/x86/cpu/broadwell/northbridge.c
@@ -6,8 +6,108 @@
 #include <common.h>
 #include <dm.h>
 #include <asm/io.h>
+#include <asm/mrc_common.h>
 #include <asm/arch/iomap.h>
 #include <asm/arch/pch.h>
+#include <asm/arch/pei_data.h>
+
+__weak asmlinkage void sdram_console_tx_byte(unsigned char byte)
+{
+#ifdef DEBUG
+	putc(byte);
+#endif
+}
+
+void broadwell_fill_pei_data(struct pei_data *pei_data)
+{
+	pei_data->pei_version = PEI_VERSION;
+	pei_data->board_type = BOARD_TYPE_ULT;
+	pei_data->pciexbar = MCFG_BASE_ADDRESS;
+	pei_data->smbusbar = SMBUS_BASE_ADDRESS;
+	pei_data->ehcibar = EARLY_EHCI_BAR;
+	pei_data->xhcibar = EARLY_XHCI_BAR;
+	pei_data->gttbar = EARLY_GTT_BAR;
+	pei_data->pmbase = ACPI_BASE_ADDRESS;
+	pei_data->gpiobase = GPIO_BASE_ADDRESS;
+	pei_data->tseg_size = CONFIG_SMM_TSEG_SIZE;
+	pei_data->temp_mmio_base = EARLY_TEMP_MMIO;
+	pei_data->tx_byte = sdram_console_tx_byte;
+	pei_data->ddr_refresh_2x = 1;
+}
+
+static void pei_data_usb2_port(struct pei_data *pei_data, int port, uint length,
+			       uint enable, uint oc_pin, uint location)
+{
+	pei_data->usb2_ports[port].length   = length;
+	pei_data->usb2_ports[port].enable   = enable;
+	pei_data->usb2_ports[port].oc_pin   = oc_pin;
+	pei_data->usb2_ports[port].location = location;
+}
+
+static void pei_data_usb3_port(struct pei_data *pei_data, int port, uint enable,
+			       uint oc_pin, uint fixed_eq)
+{
+	pei_data->usb3_ports[port].enable   = enable;
+	pei_data->usb3_ports[port].oc_pin   = oc_pin;
+	pei_data->usb3_ports[port].fixed_eq = fixed_eq;
+}
+
+void mainboard_fill_pei_data(struct pei_data *pei_data)
+{
+	/* DQ byte map for Samus board */
+	const u8 dq_map[2][6][2] = {
+		{ { 0x0F, 0xF0 }, { 0x00, 0xF0 }, { 0x0F, 0xF0 },
+		  { 0x0F, 0x00 }, { 0xFF, 0x00 }, { 0xFF, 0x00 } },
+		{ { 0x0F, 0xF0 }, { 0x00, 0xF0 }, { 0x0F, 0xF0 },
+		  { 0x0F, 0x00 }, { 0xFF, 0x00 }, { 0xFF, 0x00 } } };
+	/* DQS CPU<>DRAM map for Samus board */
+	const u8 dqs_map[2][8] = {
+		{ 2, 0, 1, 3, 6, 4, 7, 5 },
+		{ 2, 1, 0, 3, 6, 5, 4, 7 } };
+
+	pei_data->ec_present = 1;
+
+	/* One installed DIMM per channel */
+	pei_data->dimm_channel0_disabled = 2;
+	pei_data->dimm_channel1_disabled = 2;
+
+	memcpy(pei_data->dq_map, dq_map, sizeof(dq_map));
+	memcpy(pei_data->dqs_map, dqs_map, sizeof(dqs_map));
+
+	/* P0: HOST PORT */
+	pei_data_usb2_port(pei_data, 0, 0x0080, 1, 0,
+			   USB_PORT_BACK_PANEL);
+	/* P1: HOST PORT */
+	pei_data_usb2_port(pei_data, 1, 0x0080, 1, 1,
+			   USB_PORT_BACK_PANEL);
+	/* P2: RAIDEN */
+	pei_data_usb2_port(pei_data, 2, 0x0080, 1, USB_OC_PIN_SKIP,
+			   USB_PORT_BACK_PANEL);
+	/* P3: SD CARD */
+	pei_data_usb2_port(pei_data, 3, 0x0040, 1, USB_OC_PIN_SKIP,
+			   USB_PORT_INTERNAL);
+	/* P4: RAIDEN */
+	pei_data_usb2_port(pei_data, 4, 0x0080, 1, USB_OC_PIN_SKIP,
+			   USB_PORT_BACK_PANEL);
+	/* P5: WWAN (Disabled) */
+	pei_data_usb2_port(pei_data, 5, 0x0000, 0, USB_OC_PIN_SKIP,
+			   USB_PORT_SKIP);
+	/* P6: CAMERA */
+	pei_data_usb2_port(pei_data, 6, 0x0040, 1, USB_OC_PIN_SKIP,
+			   USB_PORT_INTERNAL);
+	/* P7: BT */
+	pei_data_usb2_port(pei_data, 7, 0x0040, 1, USB_OC_PIN_SKIP,
+			   USB_PORT_INTERNAL);
+
+	/* P1: HOST PORT */
+	pei_data_usb3_port(pei_data, 0, 1, 0, 0);
+	/* P2: HOST PORT */
+	pei_data_usb3_port(pei_data, 1, 1, 1, 0);
+	/* P3: RAIDEN */
+	pei_data_usb3_port(pei_data, 2, 1, USB_OC_PIN_SKIP, 0);
+	/* P4: RAIDEN */
+	pei_data_usb3_port(pei_data, 3, 1, USB_OC_PIN_SKIP, 0);
+}
 
 static int broadwell_northbridge_early_init(struct udevice *dev)
 {
diff --git a/arch/x86/cpu/broadwell/pch.c b/arch/x86/cpu/broadwell/pch.c
index 73d3d3b515..a48945adf1 100644
--- a/arch/x86/cpu/broadwell/pch.c
+++ b/arch/x86/cpu/broadwell/pch.c
@@ -599,10 +599,16 @@ static int broadwell_pch_init(struct udevice *dev)
 
 static int broadwell_pch_probe(struct udevice *dev)
 {
-	if (!(gd->flags & GD_FLG_RELOC))
-		return broadwell_pch_early_init(dev);
-	else
+	if (CONFIG_IS_ENABLED(X86_32BIT_INIT)) {
+		if (!(gd->flags & GD_FLG_RELOC))
+			return broadwell_pch_early_init(dev);
+		else
+			return broadwell_pch_init(dev);
+	} else if (IS_ENABLED(CONFIG_SPL) && !IS_ENABLED(CONFIG_SPL_BUILD)) {
 		return broadwell_pch_init(dev);
+	} else {
+		return 0;
+	}
 }
 
 static int broadwell_pch_get_spi_base(struct udevice *dev, ulong *sbasep)
@@ -630,10 +636,35 @@ static int broadwell_get_gpio_base(struct udevice *dev, u32 *gbasep)
 	return 0;
 }
 
+static int broadwell_ioctl(struct udevice *dev, enum pch_req_t req, void *data,
+			   int size)
+{
+	switch (req) {
+	case PCH_REQ_PMBASE_INFO: {
+		struct pch_pmbase_info *pm = data;
+		int ret;
+
+		/* Find the base address of the powermanagement registers */
+		ret = dm_pci_read_config16(dev, 0x40, &pm->base);
+		if (ret)
+			return ret;
+		pm->base &= 0xfffe;
+		pm->gpio0_en_ofs = GPE0_EN(0);
+		pm->pm1_sts_ofs = PM1_STS;
+		pm->pm1_cnt_ofs = PM1_CNT;
+
+		return 0;
+	}
+	default:
+		return -ENOSYS;
+	}
+}
+
 static const struct pch_ops broadwell_pch_ops = {
 	.get_spi_base	= broadwell_pch_get_spi_base,
 	.set_spi_protect = broadwell_set_spi_protect,
 	.get_gpio_base	= broadwell_get_gpio_base,
+	.ioctl		= broadwell_ioctl,
 };
 
 static const struct udevice_id broadwell_pch_ids[] = {
diff --git a/arch/x86/cpu/broadwell/sdram.c b/arch/x86/cpu/broadwell/sdram.c
index 03a35bcf73..b31d78c092 100644
--- a/arch/x86/cpu/broadwell/sdram.c
+++ b/arch/x86/cpu/broadwell/sdram.c
@@ -34,99 +34,6 @@ int dram_init_banksize(void)
 	return 0;
 }
 
-void broadwell_fill_pei_data(struct pei_data *pei_data)
-{
-	pei_data->pei_version = PEI_VERSION;
-	pei_data->board_type = BOARD_TYPE_ULT;
-	pei_data->pciexbar = MCFG_BASE_ADDRESS;
-	pei_data->smbusbar = SMBUS_BASE_ADDRESS;
-	pei_data->ehcibar = EARLY_EHCI_BAR;
-	pei_data->xhcibar = EARLY_XHCI_BAR;
-	pei_data->gttbar = EARLY_GTT_BAR;
-	pei_data->pmbase = ACPI_BASE_ADDRESS;
-	pei_data->gpiobase = GPIO_BASE_ADDRESS;
-	pei_data->tseg_size = CONFIG_SMM_TSEG_SIZE;
-	pei_data->temp_mmio_base = EARLY_TEMP_MMIO;
-	pei_data->tx_byte = sdram_console_tx_byte;
-	pei_data->ddr_refresh_2x = 1;
-}
-
-static inline void pei_data_usb2_port(struct pei_data *pei_data, int port,
-				      uint16_t length, uint8_t enable,
-				      uint8_t oc_pin, uint8_t location)
-{
-	pei_data->usb2_ports[port].length   = length;
-	pei_data->usb2_ports[port].enable   = enable;
-	pei_data->usb2_ports[port].oc_pin   = oc_pin;
-	pei_data->usb2_ports[port].location = location;
-}
-
-static inline void pei_data_usb3_port(struct pei_data *pei_data, int port,
-				      uint8_t enable, uint8_t oc_pin,
-				      uint8_t fixed_eq)
-{
-	pei_data->usb3_ports[port].enable   = enable;
-	pei_data->usb3_ports[port].oc_pin   = oc_pin;
-	pei_data->usb3_ports[port].fixed_eq = fixed_eq;
-}
-
-void mainboard_fill_pei_data(struct pei_data *pei_data)
-{
-	/* DQ byte map for Samus board */
-	const u8 dq_map[2][6][2] = {
-		{ { 0x0F, 0xF0 }, { 0x00, 0xF0 }, { 0x0F, 0xF0 },
-		  { 0x0F, 0x00 }, { 0xFF, 0x00 }, { 0xFF, 0x00 } },
-		{ { 0x0F, 0xF0 }, { 0x00, 0xF0 }, { 0x0F, 0xF0 },
-		  { 0x0F, 0x00 }, { 0xFF, 0x00 }, { 0xFF, 0x00 } } };
-	/* DQS CPU<>DRAM map for Samus board */
-	const u8 dqs_map[2][8] = {
-		{ 2, 0, 1, 3, 6, 4, 7, 5 },
-		{ 2, 1, 0, 3, 6, 5, 4, 7 } };
-
-	pei_data->ec_present = 1;
-
-	/* One installed DIMM per channel */
-	pei_data->dimm_channel0_disabled = 2;
-	pei_data->dimm_channel1_disabled = 2;
-
-	memcpy(pei_data->dq_map, dq_map, sizeof(dq_map));
-	memcpy(pei_data->dqs_map, dqs_map, sizeof(dqs_map));
-
-	/* P0: HOST PORT */
-	pei_data_usb2_port(pei_data, 0, 0x0080, 1, 0,
-			   USB_PORT_BACK_PANEL);
-	/* P1: HOST PORT */
-	pei_data_usb2_port(pei_data, 1, 0x0080, 1, 1,
-			   USB_PORT_BACK_PANEL);
-	/* P2: RAIDEN */
-	pei_data_usb2_port(pei_data, 2, 0x0080, 1, USB_OC_PIN_SKIP,
-			   USB_PORT_BACK_PANEL);
-	/* P3: SD CARD */
-	pei_data_usb2_port(pei_data, 3, 0x0040, 1, USB_OC_PIN_SKIP,
-			   USB_PORT_INTERNAL);
-	/* P4: RAIDEN */
-	pei_data_usb2_port(pei_data, 4, 0x0080, 1, USB_OC_PIN_SKIP,
-			   USB_PORT_BACK_PANEL);
-	/* P5: WWAN (Disabled) */
-	pei_data_usb2_port(pei_data, 5, 0x0000, 0, USB_OC_PIN_SKIP,
-			   USB_PORT_SKIP);
-	/* P6: CAMERA */
-	pei_data_usb2_port(pei_data, 6, 0x0040, 1, USB_OC_PIN_SKIP,
-			   USB_PORT_INTERNAL);
-	/* P7: BT */
-	pei_data_usb2_port(pei_data, 7, 0x0040, 1, USB_OC_PIN_SKIP,
-			   USB_PORT_INTERNAL);
-
-	/* P1: HOST PORT */
-	pei_data_usb3_port(pei_data, 0, 1, 0, 0);
-	/* P2: HOST PORT */
-	pei_data_usb3_port(pei_data, 1, 1, 1, 0);
-	/* P3: RAIDEN */
-	pei_data_usb3_port(pei_data, 2, 1, USB_OC_PIN_SKIP, 0);
-	/* P4: RAIDEN */
-	pei_data_usb3_port(pei_data, 3, 1, USB_OC_PIN_SKIP, 0);
-}
-
 static unsigned long get_top_of_ram(struct udevice *dev)
 {
 	/*
@@ -204,16 +111,18 @@ int dram_init(void)
 
 	/* Print ME state before MRC */
 	ret = syscon_get_by_driver_data(X86_SYSCON_ME, &me_dev);
-	if (ret)
+	if (ret) {
+		debug("Cannot get ME (err=%d)\n", ret);
 		return ret;
+	}
 	intel_me_status(me_dev);
 
 	/* Save ME HSIO version */
-	ret = uclass_first_device(UCLASS_PCH, &pch_dev);
-	if (ret)
+	ret = uclass_first_device_err(UCLASS_PCH, &pch_dev);
+	if (ret) {
+		debug("Cannot get PCH (err=%d)\n", ret);
 		return ret;
-	if (!pch_dev)
-		return -ENODEV;
+	}
 	power_state_get(pch_dev, &ps);
 
 	intel_me_hsio_version(me_dev, &ps.hsio_version, &ps.hsio_checksum);
@@ -221,15 +130,17 @@ int dram_init(void)
 	broadwell_fill_pei_data(pei_data);
 	mainboard_fill_pei_data(pei_data);
 
-	ret = uclass_first_device(UCLASS_NORTHBRIDGE, &dev);
-	if (ret)
+	ret = uclass_first_device_err(UCLASS_NORTHBRIDGE, &dev);
+	if (ret) {
+		debug("Cannot get Northbridge (err=%d)\n", ret);
 		return ret;
-	if (!dev)
-		return -ENODEV;
+	}
 	size = 256;
 	ret = mrc_locate_spd(dev, size, &spd_data);
-	if (ret)
+	if (ret) {
+		debug("Cannot locate SPD (err=%d)\n", ret);
 		return ret;
+	}
 	memcpy(pei_data->spd_data[0][0], spd_data, size);
 	memcpy(pei_data->spd_data[1][0], spd_data, size);
 
@@ -239,13 +150,17 @@ int dram_init(void)
 
 	debug("PEI version %#x\n", pei_data->pei_version);
 	ret = mrc_common_init(dev, pei_data, true);
-	if (ret)
+	if (ret) {
+		debug("mrc_common_init() failed(err=%d)\n", ret);
 		return ret;
+	}
 	debug("Memory init done\n");
 
 	ret = sdram_find(dev);
-	if (ret)
+	if (ret) {
+		debug("sdram_find() failed (err=%d)\n", ret);
 		return ret;
+	}
 	gd->ram_size = gd->arch.meminfo.total_32bit_memory;
 	debug("RAM size %llx\n", (unsigned long long)gd->ram_size);
 
@@ -279,17 +194,6 @@ int misc_init_r(void)
 	return 0;
 }
 
-void board_debug_uart_init(void)
-{
-	struct udevice *bus = NULL;
-
-	/* com1 / com2 decode range */
-	pci_x86_write_config(bus, PCH_DEV_LPC, LPC_IO_DEC, 1 << 4, PCI_SIZE_16);
-
-	pci_x86_write_config(bus, PCH_DEV_LPC, LPC_EN, COMA_LPC_EN,
-			     PCI_SIZE_16);
-}
-
 static const struct udevice_id broadwell_syscon_ids[] = {
 	{ .compatible = "intel,me", .data = X86_SYSCON_ME },
 	{ }
diff --git a/arch/x86/cpu/i386/cpu.c b/arch/x86/cpu/i386/cpu.c
index 3bde44ebf5..90b546e741 100644
--- a/arch/x86/cpu/i386/cpu.c
+++ b/arch/x86/cpu/i386/cpu.c
@@ -309,21 +309,22 @@ u32 cpu_get_stepping(void)
 	return gd->arch.x86_mask;
 }
 
-int x86_cpu_init_f(void)
+/* initialise FPU, reset EM, set MP and NE */
+static void setup_cpu_features(void)
 {
 	const u32 em_rst = ~X86_CR0_EM;
 	const u32 mp_ne_set = X86_CR0_MP | X86_CR0_NE;
 
-	if (ll_boot_init()) {
-		/* initialize FPU, reset EM, set MP and NE */
-		asm ("fninit\n" \
-		"movl %%cr0, %%eax\n" \
-		"andl %0, %%eax\n" \
-		"orl  %1, %%eax\n" \
-		"movl %%eax, %%cr0\n" \
-		: : "i" (em_rst), "i" (mp_ne_set) : "eax");
-	}
+	asm ("fninit\n" \
+	"movl %%cr0, %%eax\n" \
+	"andl %0, %%eax\n" \
+	"orl  %1, %%eax\n" \
+	"movl %%eax, %%cr0\n" \
+	: : "i" (em_rst), "i" (mp_ne_set) : "eax");
+}
 
+static void setup_identity(void)
+{
 	/* identify CPU via cpuid and store the decoded info into gd->arch */
 	if (has_cpuid()) {
 		struct cpu_device_id cpu;
@@ -339,46 +340,70 @@ int x86_cpu_init_f(void)
 
 		gd->arch.has_mtrr = has_mtrr();
 	}
-	/* Don't allow PCI region 3 to use memory in the 2-4GB memory hole */
+}
+
+/* Don't allow PCI region 3 to use memory in the 2-4GB memory hole */
+static void setup_pci_ram_top(void)
+{
 	gd->pci_ram_top = 0x80000000U;
+}
+
+static void setup_mtrr(void)
+{
+	u64 mtrr_cap;
 
 	/* Configure fixed range MTRRs for some legacy regions */
-	if (gd->arch.has_mtrr) {
-		u64 mtrr_cap;
-
-		mtrr_cap = native_read_msr(MTRR_CAP_MSR);
-		if (mtrr_cap & MTRR_CAP_FIX) {
-			/* Mark the VGA RAM area as uncacheable */
-			native_write_msr(MTRR_FIX_16K_A0000_MSR,
-					 MTRR_FIX_TYPE(MTRR_TYPE_UNCACHEABLE),
-					 MTRR_FIX_TYPE(MTRR_TYPE_UNCACHEABLE));
-
-			/*
-			 * Mark the PCI ROM area as cacheable to improve ROM
-			 * execution performance.
-			 */
-			native_write_msr(MTRR_FIX_4K_C0000_MSR,
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
-			native_write_msr(MTRR_FIX_4K_C8000_MSR,
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
-			native_write_msr(MTRR_FIX_4K_D0000_MSR,
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
-			native_write_msr(MTRR_FIX_4K_D8000_MSR,
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
-					 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
-
-			/* Enable the fixed range MTRRs */
-			msr_setbits_64(MTRR_DEF_TYPE_MSR, MTRR_DEF_TYPE_FIX_EN);
-		}
+	if (!gd->arch.has_mtrr)
+		return;
+
+	mtrr_cap = native_read_msr(MTRR_CAP_MSR);
+	if (mtrr_cap & MTRR_CAP_FIX) {
+		/* Mark the VGA RAM area as uncacheable */
+		native_write_msr(MTRR_FIX_16K_A0000_MSR,
+				 MTRR_FIX_TYPE(MTRR_TYPE_UNCACHEABLE),
+				 MTRR_FIX_TYPE(MTRR_TYPE_UNCACHEABLE));
+
+		/*
+		 * Mark the PCI ROM area as cacheable to improve ROM
+		 * execution performance.
+		 */
+		native_write_msr(MTRR_FIX_4K_C0000_MSR,
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
+		native_write_msr(MTRR_FIX_4K_C8000_MSR,
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
+		native_write_msr(MTRR_FIX_4K_D0000_MSR,
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
+		native_write_msr(MTRR_FIX_4K_D8000_MSR,
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK),
+				 MTRR_FIX_TYPE(MTRR_TYPE_WRBACK));
+
+		/* Enable the fixed range MTRRs */
+		msr_setbits_64(MTRR_DEF_TYPE_MSR, MTRR_DEF_TYPE_FIX_EN);
 	}
+}
+
+int x86_cpu_init_f(void)
+{
+	if (ll_boot_init())
+		setup_cpu_features();
+	setup_identity();
+	setup_mtrr();
+	setup_pci_ram_top();
 
-#ifdef CONFIG_I8254_TIMER
 	/* Set up the i8254 timer if required */
-	i8254_init();
-#endif
+	if (IS_ENABLED(CONFIG_I8254_TIMER))
+		i8254_init();
+
+	return 0;
+}
+
+int x86_cpu_reinit_f(void)
+{
+	setup_identity();
+	setup_pci_ram_top();
 
 	return 0;
 }
diff --git a/arch/x86/cpu/i386/interrupt.c b/arch/x86/cpu/i386/interrupt.c
index 1ea415b876..47df3172b7 100644
--- a/arch/x86/cpu/i386/interrupt.c
+++ b/arch/x86/cpu/i386/interrupt.c
@@ -37,7 +37,7 @@ static char *exceptions[] = {
 	"Overflow",
 	"BOUND Range Exceeded",
 	"Invalid Opcode (Undefined Opcode)",
-	"Device Not Avaiable (No Math Coprocessor)",
+	"Device Not Available (No Math Coprocessor)",
 	"Double Fault",
 	"Coprocessor Segment Overrun",
 	"Invalid TSS",
diff --git a/arch/x86/cpu/intel_common/Makefile b/arch/x86/cpu/intel_common/Makefile
index bf798c287f..07f27c29ec 100644
--- a/arch/x86/cpu/intel_common/Makefile
+++ b/arch/x86/cpu/intel_common/Makefile
@@ -3,14 +3,23 @@
 # Copyright (c) 2016 Google, Inc
 
 ifdef CONFIG_HAVE_MRC
-obj-$(CONFIG_$(SPL_)X86_32BIT_INIT) += car.o
-obj-$(CONFIG_$(SPL_)X86_32BIT_INIT) += me_status.o
-obj-$(CONFIG_$(SPL_)X86_32BIT_INIT) += report_platform.o
-obj-$(CONFIG_$(SPL_)X86_32BIT_INIT) += mrc.o
+obj-$(CONFIG_$(SPL_TPL_)X86_16BIT_INIT) += car.o
+obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += me_status.o
+obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += report_platform.o
+obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += mrc.o
 endif
 obj-y += cpu.o
 obj-y += lpc.o
 ifndef CONFIG_TARGET_EFI_APP
+obj-$(CONFIG_$(SPL_TPL_)X86_32BIT_INIT) += microcode.o
+ifndef CONFIG_$(SPL_)X86_64
 obj-y += microcode.o
 endif
+endif
 obj-y += pch.o
+
+ifdef CONFIG_SPL
+ifndef CONFIG_SPL_BUILD
+obj-y += cpu_from_spl.o
+endif
+endif
diff --git a/arch/x86/cpu/intel_common/car.S b/arch/x86/cpu/intel_common/car.S
index 52a77bb2d1..00308dbdef 100644
--- a/arch/x86/cpu/intel_common/car.S
+++ b/arch/x86/cpu/intel_common/car.S
@@ -235,7 +235,7 @@ mtrr_table_end:
 
 	.align 4
 _dt_ucode_base_size:
-	/* These next two fields are filled in by ifdtool */
+	/* These next two fields are filled in by binman */
 .globl ucode_base
 ucode_base:	/* Declared in microcode.h */
 	.long	0			/* microcode base */
diff --git a/arch/x86/cpu/intel_common/cpu_from_spl.c b/arch/x86/cpu/intel_common/cpu_from_spl.c
new file mode 100644
index 0000000000..a6233c75ce
--- /dev/null
+++ b/arch/x86/cpu/intel_common/cpu_from_spl.c
@@ -0,0 +1,27 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 Google, Inc
+ */
+
+#include <common.h>
+#include <dm.h>
+#include <errno.h>
+#include <asm/cpu_common.h>
+#include <asm/intel_regs.h>
+#include <asm/lapic.h>
+#include <asm/lpc_common.h>
+#include <asm/msr.h>
+#include <asm/mtrr.h>
+#include <asm/post.h>
+#include <asm/microcode.h>
+
+DECLARE_GLOBAL_DATA_PTR;
+
+int arch_cpu_init(void)
+{
+	int ret;
+
+	ret = x86_cpu_reinit_f();
+
+	return ret;
+}
diff --git a/arch/x86/cpu/ivybridge/bd82x6x.c b/arch/x86/cpu/ivybridge/bd82x6x.c
index ed9bce6416..1cb6cecda9 100644
--- a/arch/x86/cpu/ivybridge/bd82x6x.c
+++ b/arch/x86/cpu/ivybridge/bd82x6x.c
@@ -229,6 +229,21 @@ static int bd82x6x_ioctl(struct udevice *dev, enum pch_req_t req, void *data,
 			return -ENOENT;
 
 		return val & RCBA_AUDIO_CONFIG_MASK;
+	case PCH_REQ_PMBASE_INFO: {
+		struct pch_pmbase_info *pm = data;
+		int ret;
+
+		/* Find the base address of the powermanagement registers */
+		ret = dm_pci_read_config16(dev, 0x40, &pm->base);
+		if (ret)
+			return ret;
+		pm->base &= 0xfffe;
+		pm->gpio0_en_ofs = GPE0_EN;
+		pm->pm1_sts_ofs = PM1_STS;
+		pm->pm1_cnt_ofs = PM1_CNT;
+
+		return 0;
+	}
 	default:
 		return -ENOSYS;
 	}
diff --git a/arch/x86/cpu/mp_init.c b/arch/x86/cpu/mp_init.c
index ea64c2ee57..fefbf8f728 100644
--- a/arch/x86/cpu/mp_init.c
+++ b/arch/x86/cpu/mp_init.c
@@ -322,7 +322,7 @@ static int start_aps(int ap_count, atomic_t *num_aps)
 	if (sipi_vector > max_vector_loc) {
 		printf("SIPI vector too large! 0x%08x\n",
 		       sipi_vector);
-		return -1;
+		return -ENOSPC;
 	}
 
 	debug("Attempting to start %d APs\n", ap_count);
@@ -364,7 +364,7 @@ static int start_aps(int ap_count, atomic_t *num_aps)
 	if (wait_for_aps(num_aps, ap_count, 10000, 50)) {
 		debug("Not all APs checked in: %d/%d\n",
 		      atomic_read(num_aps), ap_count);
-		return -1;
+		return -EIO;
 	}
 
 	return 0;
@@ -387,7 +387,7 @@ static int bsp_do_flight_plan(struct udevice *cpu, struct mp_params *mp_params)
 			if (wait_for_aps(&rec->cpus_entered, num_aps,
 					 timeout_us, step_us)) {
 				debug("MP record %d timeout\n", i);
-				ret = -1;
+				ret = -ETIMEDOUT;
 			}
 		}
 
@@ -508,7 +508,7 @@ int mp_init(struct mp_params *p)
 
 	if (p == NULL || p->flight_plan == NULL || p->num_records < 1) {
 		printf("Invalid MP parameters\n");
-		return -1;
+		return -EINVAL;
 	}
 
 	num_cpus = cpu_get_count(cpu);
@@ -531,7 +531,7 @@ int mp_init(struct mp_params *p)
 	/* Load the SIPI vector */
 	ret = load_sipi_vector(&ap_count, num_cpus);
 	if (ap_count == NULL)
-		return -1;
+		return -ENOENT;
 
 	/*
 	 * Make sure SIPI data hits RAM so the APs that come up will see
diff --git a/arch/x86/cpu/start.S b/arch/x86/cpu/start.S
index 30fa7def46..4a82add76b 100644
--- a/arch/x86/cpu/start.S
+++ b/arch/x86/cpu/start.S
@@ -190,6 +190,19 @@ board_init_f_r_trampoline:
 	/* Re-enter U-Boot by calling board_init_f_r() */
 	call	board_init_f_r
 
+#ifdef CONFIG_TPL
+.globl jump_to_spl
+.type jump_to_spl, @function
+jump_to_spl:
+	/* Reset stack to the top of CAR space */
+	movl	$(CONFIG_SYS_CAR_ADDR + CONFIG_SYS_CAR_SIZE - 4), %esp
+#ifdef CONFIG_DCACHE_RAM_MRC_VAR_SIZE
+	subl	$CONFIG_DCACHE_RAM_MRC_VAR_SIZE, %esp
+#endif
+
+	jmp	*%eax
+#endif
+
 die:
 	hlt
 	jmp	die
diff --git a/arch/x86/cpu/start64.S b/arch/x86/cpu/start64.S
index a78a3316b6..7be834788b 100644
--- a/arch/x86/cpu/start64.S
+++ b/arch/x86/cpu/start64.S
@@ -2,7 +2,7 @@
 /*
  * 64-bit x86 Startup Code
  *
- * (C) Copyright 216 Google, Inc
+ * Copyright 2019 Google, Inc
  * Written by Simon Glass <sjg@chromium.org>
  */
 
diff --git a/arch/x86/cpu/start_from_spl.S b/arch/x86/cpu/start_from_spl.S
new file mode 100644
index 0000000000..4d4e5d0758
--- /dev/null
+++ b/arch/x86/cpu/start_from_spl.S
@@ -0,0 +1,71 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * 32-bit x86 Startup Code when running from SPL
+ *
+ * Copyright 2018 Google, Inc
+ * Written by Simon Glass <sjg@chromium.org>
+ */
+
+#include <config.h>
+
+.section .text.start
+.code32
+.globl _start
+.type _start, @function
+_start:
+	/* Set up memory using the existing stack */
+	movl	$(CONFIG_SYS_CAR_ADDR + CONFIG_SYS_CAR_SIZE - 4), %eax
+#ifdef CONFIG_DCACHE_RAM_MRC_VAR_SIZE
+	subl	$CONFIG_DCACHE_RAM_MRC_VAR_SIZE, %eax
+#endif
+	/*
+	 * We don't subject CONFIG_DCACHE_RAM_MRC_VAR_SIZE since memory is
+	 * already set up. This has the happy side-effect of putting gd in a
+	 * new place separate from SPL, so the memset() in
+	 * board_init_f_init_reserve() does not cause any problems (otherwise
+	 * it would zero out the gd and crash)
+	 */
+	call	board_init_f_alloc_reserve
+	mov	%eax, %esp
+
+	call	board_init_f_init_reserve
+
+	xorl	%eax, %eax
+	call	board_init_f
+	call	board_init_f_r
+
+	/* Should not return here */
+	jmp	.
+
+.globl board_init_f_r_trampoline
+.type board_init_f_r_trampoline, @function
+board_init_f_r_trampoline:
+	/*
+	 * SPL has been executed and SDRAM has been initialised, U-Boot code
+	 * has been copied into RAM, BSS has been cleared and relocation
+	 * adjustments have been made. It is now time to jump into the in-RAM
+	 * copy of U-Boot
+	 *
+	 * %eax = Address of top of new stack
+	 */
+
+	/* Stack grows down from top of SDRAM */
+	movl	%eax, %esp
+
+	/* Re-enter U-Boot by calling board_init_f_r() */
+	call	board_init_f_r
+
+die:
+	hlt
+	jmp	die
+	hlt
+
+	.align 4
+_dt_ucode_base_size:
+	/* These next two fields are filled in by binman */
+.globl ucode_base
+ucode_base:	/* Declared in microcode.h */
+	.long	0			/* microcode base */
+.globl ucode_size
+ucode_size:	/* Declared in microcode.h */
+	.long	0			/* microcode size */
diff --git a/arch/x86/cpu/start_from_tpl.S b/arch/x86/cpu/start_from_tpl.S
new file mode 100644
index 0000000000..44b5363a68
--- /dev/null
+++ b/arch/x86/cpu/start_from_tpl.S
@@ -0,0 +1,49 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * 32-bit x86 Startup Code when running from TPL
+ *
+ * Copyright 2018 Google, Inc
+ * Written by Simon Glass <sjg@chromium.org>
+ */
+
+#include <config.h>
+
+.section .text.start
+.code32
+.globl _start
+.type _start, @function
+_start:
+	/* Set up memory using the existing stack */
+	mov	%esp, %eax
+	call	board_init_f_alloc_reserve
+	mov	%eax, %esp
+
+	call	board_init_f_init_reserve
+
+	xorl	%eax, %eax
+	call	board_init_f
+	call	board_init_f_r
+
+	/* Should not return here */
+	jmp	.
+
+.globl board_init_f_r_trampoline
+.type board_init_f_r_trampoline, @function
+board_init_f_r_trampoline:
+	/*
+	 * TPL has been executed: SDRAM has been initialised, BSS has been
+	 * cleared.
+	 *
+	 * %eax = Address of top of new stack
+	 */
+
+	/* Stack grows down from top of SDRAM */
+	movl	%eax, %esp
+
+	/* Re-enter SPL by calling board_init_f_r() */
+	call	board_init_f_r
+
+die:
+	hlt
+	jmp	die
+	hlt
diff --git a/arch/x86/cpu/u-boot-spl.lds b/arch/x86/cpu/u-boot-spl.lds
index 4e656dc4e5..f20c0b810d 100644
--- a/arch/x86/cpu/u-boot-spl.lds
+++ b/arch/x86/cpu/u-boot-spl.lds
@@ -54,7 +54,7 @@ SECTIONS
 	/DISCARD/ : { *(.interp*) }
 	/DISCARD/ : { *(.gnu*) }
 
-#ifdef CONFIG_SPL_X86_16BIT_INIT
+#if defined(CONFIG_SPL_X86_16BIT_INIT) || defined(CONFIG_TPL_X86_16BIT_INIT)
 	/*
 	 * The following expressions place the 16-bit Real-Mode code and
 	 * Reset Vector at the end of the Flash ROM
diff --git a/arch/x86/cpu/x86_64/cpu.c b/arch/x86/cpu/x86_64/cpu.c
index 6c063e8200..42abb23a9e 100644
--- a/arch/x86/cpu/x86_64/cpu.c
+++ b/arch/x86/cpu/x86_64/cpu.c
@@ -61,3 +61,8 @@ int print_cpuinfo(void)
 {
 	return 0;
 }
+
+int x86_cpu_reinit_f(void)
+{
+	return 0;
+}