ftrace: Synchronize the initialization of ftrace_opsgrafted

There's some internal state that ftrace_ops needs to have set, but since
it can be declared outside of the ftrace.c code, it calls
ftrace_ops_init() on the ops in every global function. The issue is that
if two tasks call it on the same ops at the same time it is possible to
have the initialization of one corrupt the initialization of the other
call.

Create a ops_mutex to use to synchronize every initialization of the
ftrace_ops. The mutex is taken within checking the ftrace_ops flag that
states it was initializied but the flag is checked again after the mutex
has been taken. Checking first outside the mutex allows it to shortcut
having to take the mutex. But then the check needs to be done again after
the mute is taken in case of races.

Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260902095501.6b59af20@gandalf.local.home
Fixes: f04f24fb7e48d ("ftrace, kprobes: Fix a deadlock on ftrace_regex_lock")
Reported-by: sashiko-bot@kernel.org
Close: https://lore.kernel.org/all/20260829025528.49A831F000E9@smtp.kernel.org/
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>

1 files changed, 955 insertions, 0 deletions

diff --git a/drivers/clk/davinci/pll.c b/drivers/clk/davinci/pll.c
new file mode 100644
index 000000000..f73b0ac5d
--- /dev/null
+++ b/drivers/clk/davinci/pll.c
@@ -0,0 +1,955 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * PLL clock driver for TI Davinci SoCs
+ *
+ * Copyright (C) 2018 David Lechner <david@lechnology.com>
+ *
+ * Based on arch/arm/mach-davinci/clock.c
+ * Copyright (C) 2006-2007 Texas Instruments.
+ * Copyright (C) 2008-2009 Deep Root Systems, LLC
+ */
+
+#include <linux/clk-provider.h>
+#include <linux/clk.h>
+#include <linux/clk/davinci.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/mfd/syscon.h>
+#include <linux/notifier.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/property.h>
+#include <linux/regmap.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include "pll.h"
+
+#define MAX_NAME_SIZE	20
+#define OSCIN_CLK_NAME	"oscin"
+
+#define REVID		0x000
+#define PLLCTL		0x100
+#define OCSEL		0x104
+#define PLLSECCTL	0x108
+#define PLLM		0x110
+#define PREDIV		0x114
+#define PLLDIV1		0x118
+#define PLLDIV2		0x11c
+#define PLLDIV3		0x120
+#define OSCDIV		0x124
+#define POSTDIV		0x128
+#define BPDIV		0x12c
+#define PLLCMD		0x138
+#define PLLSTAT		0x13c
+#define ALNCTL		0x140
+#define DCHANGE		0x144
+#define CKEN		0x148
+#define CKSTAT		0x14c
+#define SYSTAT		0x150
+#define PLLDIV4		0x160
+#define PLLDIV5		0x164
+#define PLLDIV6		0x168
+#define PLLDIV7		0x16c
+#define PLLDIV8		0x170
+#define PLLDIV9		0x174
+
+#define PLLCTL_PLLEN		BIT(0)
+#define PLLCTL_PLLPWRDN		BIT(1)
+#define PLLCTL_PLLRST		BIT(3)
+#define PLLCTL_PLLDIS		BIT(4)
+#define PLLCTL_PLLENSRC		BIT(5)
+#define PLLCTL_CLKMODE		BIT(8)
+
+/* shared by most *DIV registers */
+#define DIV_RATIO_SHIFT		0
+#define DIV_RATIO_WIDTH		5
+#define DIV_ENABLE_SHIFT	15
+
+#define PLLCMD_GOSET		BIT(0)
+#define PLLSTAT_GOSTAT		BIT(0)
+
+#define CKEN_OBSCLK_SHIFT	1
+#define CKEN_AUXEN_SHIFT	0
+
+/*
+ * OMAP-L138 system reference guide recommends a wait for 4 OSCIN/CLKIN
+ * cycles to ensure that the PLLC has switched to bypass mode. Delay of 1us
+ * ensures we are good for all > 4MHz OSCIN/CLKIN inputs. Typically the input
+ * is ~25MHz. Units are micro seconds.
+ */
+#define PLL_BYPASS_TIME		1
+
+/* From OMAP-L138 datasheet table 6-4. Units are micro seconds */
+#define PLL_RESET_TIME		1
+
+/*
+ * From OMAP-L138 datasheet table 6-4; assuming prediv = 1, sqrt(pllm) = 4
+ * Units are micro seconds.
+ */
+#define PLL_LOCK_TIME		20
+
+/**
+ * struct davinci_pll_clk - Main PLL clock (aka PLLOUT)
+ * @hw: clk_hw for the pll
+ * @base: Base memory address
+ * @pllm_min: The minimum allowable PLLM[PLLM] value
+ * @pllm_max: The maximum allowable PLLM[PLLM] value
+ * @pllm_mask: Bitmask for PLLM[PLLM] value
+ */
+struct davinci_pll_clk {
+	struct clk_hw hw;
+	void __iomem *base;
+	u32 pllm_min;
+	u32 pllm_max;
+	u32 pllm_mask;
+};
+
+#define to_davinci_pll_clk(_hw) \
+	container_of((_hw), struct davinci_pll_clk, hw)
+
+static unsigned long davinci_pll_recalc_rate(struct clk_hw *hw,
+					     unsigned long parent_rate)
+{
+	struct davinci_pll_clk *pll = to_davinci_pll_clk(hw);
+	unsigned long rate = parent_rate;
+	u32 mult;
+
+	mult = readl(pll->base + PLLM) & pll->pllm_mask;
+	rate *= mult + 1;
+
+	return rate;
+}
+
+static int davinci_pll_determine_rate(struct clk_hw *hw,
+				      struct clk_rate_request *req)
+{
+	struct davinci_pll_clk *pll = to_davinci_pll_clk(hw);
+	struct clk_hw *parent = req->best_parent_hw;
+	unsigned long parent_rate = req->best_parent_rate;
+	unsigned long rate = req->rate;
+	unsigned long best_rate, r;
+	u32 mult;
+
+	/* there is a limited range of valid outputs (see datasheet) */
+	if (rate < req->min_rate)
+		return -EINVAL;
+
+	rate = min(rate, req->max_rate);
+	mult = rate / parent_rate;
+	best_rate = parent_rate * mult;
+
+	/* easy case when there is no PREDIV */
+	if (!(clk_hw_get_flags(hw) & CLK_SET_RATE_PARENT)) {
+		if (best_rate < req->min_rate)
+			return -EINVAL;
+
+		if (mult < pll->pllm_min || mult > pll->pllm_max)
+			return -EINVAL;
+
+		req->rate = best_rate;
+
+		return 0;
+	}
+
+	/* see if the PREDIV clock can help us */
+	best_rate = 0;
+
+	for (mult = pll->pllm_min; mult <= pll->pllm_max; mult++) {
+		parent_rate = clk_hw_round_rate(parent, rate / mult);
+		r = parent_rate * mult;
+		if (r < req->min_rate)
+			continue;
+		if (r > rate || r > req->max_rate)
+			break;
+		if (r > best_rate) {
+			best_rate = r;
+			req->rate = best_rate;
+			req->best_parent_rate = parent_rate;
+			if (best_rate == rate)
+				break;
+		}
+	}
+
+	return 0;
+}
+
+static int davinci_pll_set_rate(struct clk_hw *hw, unsigned long rate,
+				unsigned long parent_rate)
+{
+	struct davinci_pll_clk *pll = to_davinci_pll_clk(hw);
+	u32 mult;
+
+	mult = rate / parent_rate;
+	writel(mult - 1, pll->base + PLLM);
+
+	return 0;
+}
+
+#ifdef CONFIG_DEBUG_FS
+static void davinci_pll_debug_init(struct clk_hw *hw, struct dentry *dentry);
+#else
+#define davinci_pll_debug_init NULL
+#endif
+
+static const struct clk_ops davinci_pll_ops = {
+	.recalc_rate	= davinci_pll_recalc_rate,
+	.determine_rate	= davinci_pll_determine_rate,
+	.set_rate	= davinci_pll_set_rate,
+	.debug_init	= davinci_pll_debug_init,
+};
+
+/* PLLM works differently on DM365 */
+static unsigned long dm365_pll_recalc_rate(struct clk_hw *hw,
+					   unsigned long parent_rate)
+{
+	struct davinci_pll_clk *pll = to_davinci_pll_clk(hw);
+	unsigned long rate = parent_rate;
+	u32 mult;
+
+	mult = readl(pll->base + PLLM) & pll->pllm_mask;
+	rate *= mult * 2;
+
+	return rate;
+}
+
+static const struct clk_ops dm365_pll_ops = {
+	.recalc_rate	= dm365_pll_recalc_rate,
+	.debug_init	= davinci_pll_debug_init,
+};
+
+/**
+ * davinci_pll_div_register - common *DIV clock implementation
+ * @dev: The PLL platform device or NULL
+ * @name: the clock name
+ * @parent_name: the parent clock name
+ * @reg: the *DIV register
+ * @fixed: if true, the divider is a fixed value
+ * @flags: bitmap of CLK_* flags from clock-provider.h
+ */
+static struct clk *davinci_pll_div_register(struct device *dev,
+					    const char *name,
+					    const char *parent_name,
+					    void __iomem *reg,
+					    bool fixed, u32 flags)
+{
+	const char * const *parent_names = parent_name ? &parent_name : NULL;
+	int num_parents = parent_name ? 1 : 0;
+	const struct clk_ops *divider_ops = &clk_divider_ops;
+	struct clk_gate *gate;
+	struct clk_divider *divider;
+	struct clk *clk;
+	int ret;
+
+	gate = kzalloc_obj(*gate);
+	if (!gate)
+		return ERR_PTR(-ENOMEM);
+
+	gate->reg = reg;
+	gate->bit_idx = DIV_ENABLE_SHIFT;
+
+	divider = kzalloc_obj(*divider);
+	if (!divider) {
+		ret = -ENOMEM;
+		goto err_free_gate;
+	}
+
+	divider->reg = reg;
+	divider->shift = DIV_RATIO_SHIFT;
+	divider->width = DIV_RATIO_WIDTH;
+
+	if (fixed) {
+		divider->flags |= CLK_DIVIDER_READ_ONLY;
+		divider_ops = &clk_divider_ro_ops;
+	}
+
+	clk = clk_register_composite(dev, name, parent_names, num_parents,
+				     NULL, NULL, &divider->hw, divider_ops,
+				     &gate->hw, &clk_gate_ops, flags);
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto err_free_divider;
+	}
+
+	return clk;
+
+err_free_divider:
+	kfree(divider);
+err_free_gate:
+	kfree(gate);
+
+	return ERR_PTR(ret);
+}
+
+struct davinci_pllen_clk {
+	struct clk_hw hw;
+	void __iomem *base;
+};
+
+#define to_davinci_pllen_clk(_hw) \
+	container_of((_hw), struct davinci_pllen_clk, hw)
+
+static const struct clk_ops davinci_pllen_ops = {
+	/* this clocks just uses the clock notification feature */
+};
+
+/*
+ * The PLL has to be switched into bypass mode while we are chaning the rate,
+ * so we do that on the PLLEN clock since it is the end of the line. This will
+ * switch to bypass before any of the parent clocks (PREDIV, PLL, POSTDIV) are
+ * changed and will switch back to the PLL after the changes have been made.
+ */
+static int davinci_pllen_rate_change(struct notifier_block *nb,
+				     unsigned long flags, void *data)
+{
+	struct clk_notifier_data *cnd = data;
+	struct clk_hw *hw = __clk_get_hw(cnd->clk);
+	struct davinci_pllen_clk *pll = to_davinci_pllen_clk(hw);
+	u32 ctrl;
+
+	ctrl = readl(pll->base + PLLCTL);
+
+	if (flags == PRE_RATE_CHANGE) {
+		/* Switch the PLL to bypass mode */
+		ctrl &= ~(PLLCTL_PLLENSRC | PLLCTL_PLLEN);
+		writel(ctrl, pll->base + PLLCTL);
+
+		udelay(PLL_BYPASS_TIME);
+
+		/* Reset and enable PLL */
+		ctrl &= ~(PLLCTL_PLLRST | PLLCTL_PLLDIS);
+		writel(ctrl, pll->base + PLLCTL);
+	} else {
+		udelay(PLL_RESET_TIME);
+
+		/* Bring PLL out of reset */
+		ctrl |= PLLCTL_PLLRST;
+		writel(ctrl, pll->base + PLLCTL);
+
+		udelay(PLL_LOCK_TIME);
+
+		/* Remove PLL from bypass mode */
+		ctrl |= PLLCTL_PLLEN;
+		writel(ctrl, pll->base + PLLCTL);
+	}
+
+	return NOTIFY_OK;
+}
+
+static struct notifier_block davinci_pllen_notifier = {
+	.notifier_call = davinci_pllen_rate_change,
+};
+
+/**
+ * davinci_pll_clk_register - Register a PLL clock
+ * @dev: The PLL platform device or NULL
+ * @info: The device-specific clock info
+ * @parent_name: The parent clock name
+ * @base: The PLL's memory region
+ * @cfgchip: CFGCHIP syscon regmap for info->unlock_reg or NULL
+ *
+ * This creates a series of clocks that represent the PLL.
+ *
+ *     OSCIN > [PREDIV >] PLL > [POSTDIV >] PLLEN
+ *
+ * - OSCIN is the parent clock (on secondary PLL, may come from primary PLL)
+ * - PREDIV and POSTDIV are optional (depends on the PLL controller)
+ * - PLL is the PLL output (aka PLLOUT)
+ * - PLLEN is the bypass multiplexer
+ *
+ * Returns: The PLLOUT clock or a negative error code.
+ */
+struct clk *davinci_pll_clk_register(struct device *dev,
+				     const struct davinci_pll_clk_info *info,
+				     const char *parent_name,
+				     void __iomem *base,
+				     struct regmap *cfgchip)
+{
+	char prediv_name[MAX_NAME_SIZE];
+	char pllout_name[MAX_NAME_SIZE];
+	char postdiv_name[MAX_NAME_SIZE];
+	char pllen_name[MAX_NAME_SIZE];
+	struct clk_init_data init;
+	struct davinci_pll_clk *pllout;
+	struct davinci_pllen_clk *pllen;
+	struct clk *oscin_clk = NULL;
+	struct clk *prediv_clk = NULL;
+	struct clk *pllout_clk;
+	struct clk *postdiv_clk = NULL;
+	struct clk *pllen_clk;
+	int ret;
+
+	if (info->flags & PLL_HAS_CLKMODE) {
+		/*
+		 * If a PLL has PLLCTL[CLKMODE], then it is the primary PLL.
+		 * We register a clock named "oscin" that serves as the internal
+		 * "input clock" domain shared by both PLLs (if there are 2)
+		 * and will be the parent clock to the AUXCLK, SYSCLKBP and
+		 * OBSCLK domains. NB: The various TRMs use "OSCIN" to mean
+		 * a number of different things. In this driver we use it to
+		 * mean the signal after the PLLCTL[CLKMODE] switch.
+		 */
+		oscin_clk = clk_register_fixed_factor(dev, OSCIN_CLK_NAME,
+						      parent_name, 0, 1, 1);
+		if (IS_ERR(oscin_clk))
+			return oscin_clk;
+
+		parent_name = OSCIN_CLK_NAME;
+	}
+
+	if (info->flags & PLL_HAS_PREDIV) {
+		bool fixed = info->flags & PLL_PREDIV_FIXED_DIV;
+		u32 flags = 0;
+
+		snprintf(prediv_name, MAX_NAME_SIZE, "%s_prediv", info->name);
+
+		if (info->flags & PLL_PREDIV_ALWAYS_ENABLED)
+			flags |= CLK_IS_CRITICAL;
+
+		/* Some? DM355 chips don't correctly report the PREDIV value */
+		if (info->flags & PLL_PREDIV_FIXED8)
+			prediv_clk = clk_register_fixed_factor(dev, prediv_name,
+							parent_name, flags, 1, 8);
+		else
+			prediv_clk = davinci_pll_div_register(dev, prediv_name,
+				parent_name, base + PREDIV, fixed, flags);
+		if (IS_ERR(prediv_clk)) {
+			ret = PTR_ERR(prediv_clk);
+			goto err_unregister_oscin;
+		}
+
+		parent_name = prediv_name;
+	}
+
+	/* Unlock writing to PLL registers */
+	if (info->unlock_reg) {
+		if (IS_ERR_OR_NULL(cfgchip))
+			dev_warn(dev, "Failed to get CFGCHIP (%ld)\n",
+				 PTR_ERR(cfgchip));
+		else
+			regmap_write_bits(cfgchip, info->unlock_reg,
+					  info->unlock_mask, 0);
+	}
+
+	pllout = kzalloc_obj(*pllout);
+	if (!pllout) {
+		ret = -ENOMEM;
+		goto err_unregister_prediv;
+	}
+
+	snprintf(pllout_name, MAX_NAME_SIZE, "%s_pllout", info->name);
+
+	init.name = pllout_name;
+	if (info->flags & PLL_PLLM_2X)
+		init.ops = &dm365_pll_ops;
+	else
+		init.ops = &davinci_pll_ops;
+	init.parent_names = &parent_name;
+	init.num_parents = 1;
+	init.flags = 0;
+
+	if (info->flags & PLL_HAS_PREDIV)
+		init.flags |= CLK_SET_RATE_PARENT;
+
+	pllout->hw.init = &init;
+	pllout->base = base;
+	pllout->pllm_mask = info->pllm_mask;
+	pllout->pllm_min = info->pllm_min;
+	pllout->pllm_max = info->pllm_max;
+
+	pllout_clk = clk_register(dev, &pllout->hw);
+	if (IS_ERR(pllout_clk)) {
+		ret = PTR_ERR(pllout_clk);
+		goto err_free_pllout;
+	}
+
+	clk_hw_set_rate_range(&pllout->hw, info->pllout_min_rate,
+			      info->pllout_max_rate);
+
+	parent_name = pllout_name;
+
+	if (info->flags & PLL_HAS_POSTDIV) {
+		bool fixed = info->flags & PLL_POSTDIV_FIXED_DIV;
+		u32 flags = CLK_SET_RATE_PARENT;
+
+		snprintf(postdiv_name, MAX_NAME_SIZE, "%s_postdiv", info->name);
+
+		if (info->flags & PLL_POSTDIV_ALWAYS_ENABLED)
+			flags |= CLK_IS_CRITICAL;
+
+		postdiv_clk = davinci_pll_div_register(dev, postdiv_name,
+				parent_name, base + POSTDIV, fixed, flags);
+		if (IS_ERR(postdiv_clk)) {
+			ret = PTR_ERR(postdiv_clk);
+			goto err_unregister_pllout;
+		}
+
+		parent_name = postdiv_name;
+	}
+
+	pllen = kzalloc_obj(*pllen);
+	if (!pllen) {
+		ret = -ENOMEM;
+		goto err_unregister_postdiv;
+	}
+
+	snprintf(pllen_name, MAX_NAME_SIZE, "%s_pllen", info->name);
+
+	init.name = pllen_name;
+	init.ops = &davinci_pllen_ops;
+	init.parent_names = &parent_name;
+	init.num_parents = 1;
+	init.flags = CLK_SET_RATE_PARENT;
+
+	pllen->hw.init = &init;
+	pllen->base = base;
+
+	pllen_clk = clk_register(dev, &pllen->hw);
+	if (IS_ERR(pllen_clk)) {
+		ret = PTR_ERR(pllen_clk);
+		goto err_free_pllen;
+	}
+
+	clk_notifier_register(pllen_clk, &davinci_pllen_notifier);
+
+	return pllout_clk;
+
+err_free_pllen:
+	kfree(pllen);
+err_unregister_postdiv:
+	clk_unregister(postdiv_clk);
+err_unregister_pllout:
+	clk_unregister(pllout_clk);
+err_free_pllout:
+	kfree(pllout);
+err_unregister_prediv:
+	clk_unregister(prediv_clk);
+err_unregister_oscin:
+	clk_unregister(oscin_clk);
+
+	return ERR_PTR(ret);
+}
+
+/**
+ * davinci_pll_auxclk_register - Register bypass clock (AUXCLK)
+ * @dev: The PLL platform device or NULL
+ * @name: The clock name
+ * @base: The PLL memory region
+ */
+struct clk *davinci_pll_auxclk_register(struct device *dev,
+					const char *name,
+					void __iomem *base)
+{
+	return clk_register_gate(dev, name, OSCIN_CLK_NAME, 0, base + CKEN,
+				 CKEN_AUXEN_SHIFT, 0, NULL);
+}
+
+/**
+ * davinci_pll_sysclkbp_clk_register - Register bypass divider clock (SYSCLKBP)
+ * @dev: The PLL platform device or NULL
+ * @name: The clock name
+ * @base: The PLL memory region
+ */
+struct clk *davinci_pll_sysclkbp_clk_register(struct device *dev,
+					      const char *name,
+					      void __iomem *base)
+{
+	return clk_register_divider(dev, name, OSCIN_CLK_NAME, 0, base + BPDIV,
+				    DIV_RATIO_SHIFT, DIV_RATIO_WIDTH,
+				    CLK_DIVIDER_READ_ONLY, NULL);
+}
+
+/**
+ * davinci_pll_obsclk_register - Register oscillator divider clock (OBSCLK)
+ * @dev: The PLL platform device or NULL
+ * @info: The clock info
+ * @base: The PLL memory region
+ */
+struct clk *
+davinci_pll_obsclk_register(struct device *dev,
+			    const struct davinci_pll_obsclk_info *info,
+			    void __iomem *base)
+{
+	struct clk_mux *mux;
+	struct clk_gate *gate;
+	struct clk_divider *divider;
+	struct clk *clk;
+	u32 oscdiv;
+	int ret;
+
+	mux = kzalloc_obj(*mux);
+	if (!mux)
+		return ERR_PTR(-ENOMEM);
+
+	mux->reg = base + OCSEL;
+	mux->table = info->table;
+	mux->mask = info->ocsrc_mask;
+
+	gate = kzalloc_obj(*gate);
+	if (!gate) {
+		ret = -ENOMEM;
+		goto err_free_mux;
+	}
+
+	gate->reg = base + CKEN;
+	gate->bit_idx = CKEN_OBSCLK_SHIFT;
+
+	divider = kzalloc_obj(*divider);
+	if (!divider) {
+		ret = -ENOMEM;
+		goto err_free_gate;
+	}
+
+	divider->reg = base + OSCDIV;
+	divider->shift = DIV_RATIO_SHIFT;
+	divider->width = DIV_RATIO_WIDTH;
+
+	/* make sure divider is enabled just in case bootloader disabled it */
+	oscdiv = readl(base + OSCDIV);
+	oscdiv |= BIT(DIV_ENABLE_SHIFT);
+	writel(oscdiv, base + OSCDIV);
+
+	clk = clk_register_composite(dev, info->name, info->parent_names,
+				     info->num_parents,
+				     &mux->hw, &clk_mux_ops,
+				     &divider->hw, &clk_divider_ops,
+				     &gate->hw, &clk_gate_ops, 0);
+
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto err_free_divider;
+	}
+
+	return clk;
+
+err_free_divider:
+	kfree(divider);
+err_free_gate:
+	kfree(gate);
+err_free_mux:
+	kfree(mux);
+
+	return ERR_PTR(ret);
+}
+
+/* The PLL SYSCLKn clocks have a mechanism for synchronizing rate changes. */
+static int davinci_pll_sysclk_rate_change(struct notifier_block *nb,
+					  unsigned long flags, void *data)
+{
+	struct clk_notifier_data *cnd = data;
+	struct clk_hw *hw = __clk_get_hw(clk_get_parent(cnd->clk));
+	struct davinci_pllen_clk *pll = to_davinci_pllen_clk(hw);
+	u32 pllcmd, pllstat;
+
+	switch (flags) {
+	case POST_RATE_CHANGE:
+		/* apply the changes */
+		pllcmd = readl(pll->base + PLLCMD);
+		pllcmd |= PLLCMD_GOSET;
+		writel(pllcmd, pll->base + PLLCMD);
+		fallthrough;
+	case PRE_RATE_CHANGE:
+		/* Wait until for outstanding changes to take effect */
+		do {
+			pllstat = readl(pll->base + PLLSTAT);
+		} while (pllstat & PLLSTAT_GOSTAT);
+		break;
+	}
+
+	return NOTIFY_OK;
+}
+
+static struct notifier_block davinci_pll_sysclk_notifier = {
+	.notifier_call = davinci_pll_sysclk_rate_change,
+};
+
+/**
+ * davinci_pll_sysclk_register - Register divider clocks (SYSCLKn)
+ * @dev: The PLL platform device or NULL
+ * @info: The clock info
+ * @base: The PLL memory region
+ */
+struct clk *
+davinci_pll_sysclk_register(struct device *dev,
+			    const struct davinci_pll_sysclk_info *info,
+			    void __iomem *base)
+{
+	const struct clk_ops *divider_ops = &clk_divider_ops;
+	struct clk_gate *gate;
+	struct clk_divider *divider;
+	struct clk *clk;
+	u32 reg;
+	u32 flags = 0;
+	int ret;
+
+	/* PLLDIVn registers are not entirely consecutive */
+	if (info->id < 4)
+		reg = PLLDIV1 + 4 * (info->id - 1);
+	else
+		reg = PLLDIV4 + 4 * (info->id - 4);
+
+	gate = kzalloc_obj(*gate);
+	if (!gate)
+		return ERR_PTR(-ENOMEM);
+
+	gate->reg = base + reg;
+	gate->bit_idx = DIV_ENABLE_SHIFT;
+
+	divider = kzalloc_obj(*divider);
+	if (!divider) {
+		ret = -ENOMEM;
+		goto err_free_gate;
+	}
+
+	divider->reg = base + reg;
+	divider->shift = DIV_RATIO_SHIFT;
+	divider->width = info->ratio_width;
+	divider->flags = 0;
+
+	if (info->flags & SYSCLK_FIXED_DIV) {
+		divider->flags |= CLK_DIVIDER_READ_ONLY;
+		divider_ops = &clk_divider_ro_ops;
+	}
+
+	/* Only the ARM clock can change the parent PLL rate */
+	if (info->flags & SYSCLK_ARM_RATE)
+		flags |= CLK_SET_RATE_PARENT;
+
+	if (info->flags & SYSCLK_ALWAYS_ENABLED)
+		flags |= CLK_IS_CRITICAL;
+
+	clk = clk_register_composite(dev, info->name, &info->parent_name, 1,
+				     NULL, NULL, &divider->hw, divider_ops,
+				     &gate->hw, &clk_gate_ops, flags);
+	if (IS_ERR(clk)) {
+		ret = PTR_ERR(clk);
+		goto err_free_divider;
+	}
+
+	clk_notifier_register(clk, &davinci_pll_sysclk_notifier);
+
+	return clk;
+
+err_free_divider:
+	kfree(divider);
+err_free_gate:
+	kfree(gate);
+
+	return ERR_PTR(ret);
+}
+
+int of_davinci_pll_init(struct device *dev, struct device_node *node,
+			const struct davinci_pll_clk_info *info,
+			const struct davinci_pll_obsclk_info *obsclk_info,
+			const struct davinci_pll_sysclk_info **div_info,
+			u8 max_sysclk_id,
+			void __iomem *base,
+			struct regmap *cfgchip)
+{
+	struct device_node *child;
+	const char *parent_name;
+	struct clk *clk;
+
+	if (info->flags & PLL_HAS_CLKMODE)
+		parent_name = of_clk_get_parent_name(node, 0);
+	else
+		parent_name = OSCIN_CLK_NAME;
+
+	clk = davinci_pll_clk_register(dev, info, parent_name, base, cfgchip);
+	if (IS_ERR(clk)) {
+		dev_err(dev, "failed to register %s\n", info->name);
+		return PTR_ERR(clk);
+	}
+
+	child = of_get_available_child_by_name(node, "pllout");
+	if (child) {
+		of_clk_add_provider(child, of_clk_src_simple_get, clk);
+		of_node_put(child);
+	}
+
+	child = of_get_available_child_by_name(node, "sysclk");
+	if (child) {
+		struct clk_onecell_data *clk_data;
+		struct clk **clks;
+		int n_clks =  max_sysclk_id + 1;
+		int i;
+
+		clk_data = kzalloc_obj(*clk_data);
+		if (!clk_data) {
+			of_node_put(child);
+			return -ENOMEM;
+		}
+
+		clks = kmalloc_objs(*clks, n_clks);
+		if (!clks) {
+			kfree(clk_data);
+			of_node_put(child);
+			return -ENOMEM;
+		}
+
+		clk_data->clks = clks;
+		clk_data->clk_num = n_clks;
+
+		for (i = 0; i < n_clks; i++)
+			clks[i] = ERR_PTR(-ENOENT);
+
+		for (; *div_info; div_info++) {
+			clk = davinci_pll_sysclk_register(dev, *div_info, base);
+			if (IS_ERR(clk))
+				dev_warn(dev, "failed to register %s (%ld)\n",
+					 (*div_info)->name, PTR_ERR(clk));
+			else
+				clks[(*div_info)->id] = clk;
+		}
+		of_clk_add_provider(child, of_clk_src_onecell_get, clk_data);
+		of_node_put(child);
+	}
+
+	child = of_get_available_child_by_name(node, "auxclk");
+	if (child) {
+		char child_name[MAX_NAME_SIZE];
+
+		snprintf(child_name, MAX_NAME_SIZE, "%s_auxclk", info->name);
+
+		clk = davinci_pll_auxclk_register(dev, child_name, base);
+		if (IS_ERR(clk))
+			dev_warn(dev, "failed to register %s (%ld)\n",
+				 child_name, PTR_ERR(clk));
+		else
+			of_clk_add_provider(child, of_clk_src_simple_get, clk);
+
+		of_node_put(child);
+	}
+
+	child = of_get_available_child_by_name(node, "obsclk");
+	if (child) {
+		if (obsclk_info)
+			clk = davinci_pll_obsclk_register(dev, obsclk_info, base);
+		else
+			clk = ERR_PTR(-EINVAL);
+
+		if (IS_ERR(clk))
+			dev_warn(dev, "failed to register obsclk (%ld)\n",
+				 PTR_ERR(clk));
+		else
+			of_clk_add_provider(child, of_clk_src_simple_get, clk);
+		of_node_put(child);
+	}
+
+	return 0;
+}
+
+/* needed in early boot for clocksource/clockevent */
+CLK_OF_DECLARE(da850_pll0, "ti,da850-pll0", of_da850_pll0_init);
+
+static const struct of_device_id davinci_pll_of_match[] = {
+	{ .compatible = "ti,da850-pll1", .data = of_da850_pll1_init },
+	{ }
+};
+
+static const struct platform_device_id davinci_pll_id_table[] = {
+	{ .name = "da850-pll0",  .driver_data = (kernel_ulong_t)da850_pll0_init  },
+	{ .name = "da850-pll1",  .driver_data = (kernel_ulong_t)da850_pll1_init  },
+	{ }
+};
+
+typedef int (*davinci_pll_init)(struct device *dev, void __iomem *base,
+				struct regmap *cfgchip);
+
+static int davinci_pll_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	davinci_pll_init pll_init = NULL;
+	struct regmap *cfgchip;
+	void __iomem *base;
+
+	pll_init = device_get_match_data(dev);
+	if (!pll_init && pdev->id_entry)
+		pll_init = (void *)pdev->id_entry->driver_data;
+
+	if (!pll_init) {
+		dev_err(dev, "unable to find driver data\n");
+		return -EINVAL;
+	}
+
+	cfgchip = syscon_regmap_lookup_by_compatible("ti,da830-cfgchip");
+
+	base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	return pll_init(dev, base, cfgchip);
+}
+
+static struct platform_driver davinci_pll_driver = {
+	.probe		= davinci_pll_probe,
+	.driver		= {
+		.name		= "davinci-pll-clk",
+		.of_match_table	= davinci_pll_of_match,
+	},
+	.id_table	= davinci_pll_id_table,
+};
+
+static int __init davinci_pll_driver_init(void)
+{
+	return platform_driver_register(&davinci_pll_driver);
+}
+
+/* has to be postcore_initcall because PSC devices depend on PLL parent clocks */
+postcore_initcall(davinci_pll_driver_init);
+
+#ifdef CONFIG_DEBUG_FS
+#include <linux/debugfs.h>
+
+#define DEBUG_REG(n)	\
+{			\
+	.name	= #n,	\
+	.offset	= n,	\
+}
+
+static const struct debugfs_reg32 davinci_pll_regs[] = {
+	DEBUG_REG(REVID),
+	DEBUG_REG(PLLCTL),
+	DEBUG_REG(OCSEL),
+	DEBUG_REG(PLLSECCTL),
+	DEBUG_REG(PLLM),
+	DEBUG_REG(PREDIV),
+	DEBUG_REG(PLLDIV1),
+	DEBUG_REG(PLLDIV2),
+	DEBUG_REG(PLLDIV3),
+	DEBUG_REG(OSCDIV),
+	DEBUG_REG(POSTDIV),
+	DEBUG_REG(BPDIV),
+	DEBUG_REG(PLLCMD),
+	DEBUG_REG(PLLSTAT),
+	DEBUG_REG(ALNCTL),
+	DEBUG_REG(DCHANGE),
+	DEBUG_REG(CKEN),
+	DEBUG_REG(CKSTAT),
+	DEBUG_REG(SYSTAT),
+	DEBUG_REG(PLLDIV4),
+	DEBUG_REG(PLLDIV5),
+	DEBUG_REG(PLLDIV6),
+	DEBUG_REG(PLLDIV7),
+	DEBUG_REG(PLLDIV8),
+	DEBUG_REG(PLLDIV9),
+};
+
+static void davinci_pll_debug_init(struct clk_hw *hw, struct dentry *dentry)
+{
+	struct davinci_pll_clk *pll = to_davinci_pll_clk(hw);
+	struct debugfs_regset32 *regset;
+
+	regset = kzalloc_obj(*regset);
+	if (!regset)
+		return;
+
+	regset->regs = davinci_pll_regs;
+	regset->nregs = ARRAY_SIZE(davinci_pll_regs);
+	regset->base = pll->base;
+
+	debugfs_create_regset32("registers", 0400, dentry, regset);
+}
+#endif