u-boot/drivers/clk/kendryte/pll.c
Sean Anderson 019ef9a3f3 clk: Add K210 pll support
This pll code is primarily based on the code from the kendryte standalone
sdk in lib/drivers/sysctl.c. k210_pll_calc_config is roughly analogous to
the algorithm used to set the pll frequency, but it has been completely
rewritten to be fixed-point based.

Signed-off-by: Sean Anderson <seanga2@gmail.com>
CC: Lukasz Majewski <lukma@denx.de>
2020-07-01 15:01:21 +08:00

601 lines
15 KiB
C

// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2019-20 Sean Anderson <seanga2@gmail.com>
*/
#define LOG_CATEGORY UCLASS_CLK
#include <kendryte/pll.h>
#include <asm/io.h>
/* For DIV_ROUND_DOWN_ULL, defined in linux/kernel.h */
#include <div64.h>
#include <dt-bindings/clock/k210-sysctl.h>
#include <linux/bitfield.h>
#include <linux/clk-provider.h>
#include <linux/delay.h>
#include <linux/err.h>
#include <log.h>
#include <serial.h>
#define CLK_K210_PLL "k210_clk_pll"
#ifdef CONFIG_CLK_K210_SET_RATE
static int k210_pll_enable(struct clk *clk);
static int k210_pll_disable(struct clk *clk);
/*
* The PLL included with the Kendryte K210 appears to be a True Circuits, Inc.
* General-Purpose PLL. The logical layout of the PLL with internal feedback is
* approximately the following:
*
* +---------------+
* |reference clock|
* +---------------+
* |
* v
* +--+
* |/r|
* +--+
* |
* v
* +-------------+
* |divided clock|
* +-------------+
* |
* v
* +--------------+
* |phase detector|<---+
* +--------------+ |
* | |
* v +--------------+
* +---+ |feedback clock|
* |VCO| +--------------+
* +---+ ^
* | +--+ |
* +--->|/f|---+
* | +--+
* v
* +---+
* |/od|
* +---+
* |
* v
* +------+
* |output|
* +------+
*
* The k210 PLLs have three factors: r, f, and od. Because of the feedback mode,
* the effect of the division by f is to multiply the input frequency. The
* equation for the output rate is
* rate = (rate_in * f) / (r * od).
* Moving knowns to one side of the equation, we get
* rate / rate_in = f / (r * od)
* Rearranging slightly,
* abs_error = abs((rate / rate_in) - (f / (r * od))).
* To get relative, error, we divide by the expected ratio
* error = abs((rate / rate_in) - (f / (r * od))) / (rate / rate_in).
* Simplifying,
* error = abs(1 - f / (r * od)) / (rate / rate_in)
* error = abs(1 - (f * rate_in) / (r * od * rate))
* Using the constants ratio = rate / rate_in and inv_ratio = rate_in / rate,
* error = abs((f * inv_ratio) / (r * od) - 1)
* This is the error used in evaluating parameters.
*
* r and od are four bits each, while f is six bits. Because r and od are
* multiplied together, instead of the full 256 values possible if both bits
* were used fully, there are only 97 distinct products. Combined with f, there
* are 6208 theoretical settings for the PLL. However, most of these settings
* can be ruled out immediately because they do not have the correct ratio.
*
* In addition to the constraint of approximating the desired ratio, parameters
* must also keep internal pll frequencies within acceptable ranges. The divided
* clock's minimum and maximum frequencies have a ratio of around 128. This
* leaves fairly substantial room to work with, especially since the only
* affected parameter is r. The VCO's minimum and maximum frequency have a ratio
* of 5, which is considerably more restrictive.
*
* The r and od factors are stored in a table. This is to make it easy to find
* the next-largest product. Some products have multiple factorizations, but
* only when one factor has at least a 2.5x ratio to the factors of the other
* factorization. This is because any smaller ratio would not make a difference
* when ensuring the VCO's frequency is within spec.
*
* Throughout the calculation function, fixed point arithmetic is used. Because
* the range of rate and rate_in may be up to 1.75 GHz, or around 2^30, 64-bit
* 32.32 fixed-point numbers are used to represent ratios. In general, to
* implement division, the numerator is first multiplied by 2^32. This gives a
* result where the whole number part is in the upper 32 bits, and the fraction
* is in the lower 32 bits.
*
* In general, rounding is done to the closest integer. This helps find the best
* approximation for the ratio. Rounding in one direction (e.g down) could cause
* the function to miss a better ratio with one of the parameters increased by
* one.
*/
/*
* The factors table was generated with the following python code:
*
* def p(x, y):
* return (1.0*x/y > 2.5) or (1.0*y/x > 2.5)
*
* factors = {}
* for i in range(1, 17):
* for j in range(1, 17):
* fs = factors.get(i*j) or []
* if fs == [] or all([
* (p(i, x) and p(i, y)) or (p(j, x) and p(j, y))
* for (x, y) in fs]):
* fs.append((i, j))
* factors[i*j] = fs
*
* for k, l in sorted(factors.items()):
* for v in l:
* print("PACK(%s, %s)," % v)
*/
#define PACK(r, od) (((((r) - 1) & 0xF) << 4) | (((od) - 1) & 0xF))
#define UNPACK_R(val) ((((val) >> 4) & 0xF) + 1)
#define UNPACK_OD(val) (((val) & 0xF) + 1)
static const u8 factors[] = {
PACK(1, 1),
PACK(1, 2),
PACK(1, 3),
PACK(1, 4),
PACK(1, 5),
PACK(1, 6),
PACK(1, 7),
PACK(1, 8),
PACK(1, 9),
PACK(3, 3),
PACK(1, 10),
PACK(1, 11),
PACK(1, 12),
PACK(3, 4),
PACK(1, 13),
PACK(1, 14),
PACK(1, 15),
PACK(3, 5),
PACK(1, 16),
PACK(4, 4),
PACK(2, 9),
PACK(2, 10),
PACK(3, 7),
PACK(2, 11),
PACK(2, 12),
PACK(5, 5),
PACK(2, 13),
PACK(3, 9),
PACK(2, 14),
PACK(2, 15),
PACK(2, 16),
PACK(3, 11),
PACK(5, 7),
PACK(3, 12),
PACK(3, 13),
PACK(4, 10),
PACK(3, 14),
PACK(4, 11),
PACK(3, 15),
PACK(3, 16),
PACK(7, 7),
PACK(5, 10),
PACK(4, 13),
PACK(6, 9),
PACK(5, 11),
PACK(4, 14),
PACK(4, 15),
PACK(7, 9),
PACK(4, 16),
PACK(5, 13),
PACK(6, 11),
PACK(5, 14),
PACK(6, 12),
PACK(5, 15),
PACK(7, 11),
PACK(6, 13),
PACK(5, 16),
PACK(9, 9),
PACK(6, 14),
PACK(8, 11),
PACK(6, 15),
PACK(7, 13),
PACK(6, 16),
PACK(7, 14),
PACK(9, 11),
PACK(10, 10),
PACK(8, 13),
PACK(7, 15),
PACK(9, 12),
PACK(10, 11),
PACK(7, 16),
PACK(9, 13),
PACK(8, 15),
PACK(11, 11),
PACK(9, 14),
PACK(8, 16),
PACK(10, 13),
PACK(11, 12),
PACK(9, 15),
PACK(10, 14),
PACK(11, 13),
PACK(9, 16),
PACK(10, 15),
PACK(11, 14),
PACK(12, 13),
PACK(10, 16),
PACK(11, 15),
PACK(12, 14),
PACK(13, 13),
PACK(11, 16),
PACK(12, 15),
PACK(13, 14),
PACK(12, 16),
PACK(13, 15),
PACK(14, 14),
PACK(13, 16),
PACK(14, 15),
PACK(14, 16),
PACK(15, 15),
PACK(15, 16),
PACK(16, 16),
};
TEST_STATIC int k210_pll_calc_config(u32 rate, u32 rate_in,
struct k210_pll_config *best)
{
int i;
s64 error, best_error;
u64 ratio, inv_ratio; /* fixed point 32.32 ratio of the rates */
u64 max_r;
u64 r, f, od;
/*
* Can't go over 1.75 GHz or under 21.25 MHz due to limitations on the
* VCO frequency. These are not the same limits as below because od can
* reduce the output frequency by 16.
*/
if (rate > 1750000000 || rate < 21250000)
return -EINVAL;
/* Similar restrictions on the input rate */
if (rate_in > 1750000000 || rate_in < 13300000)
return -EINVAL;
ratio = DIV_ROUND_CLOSEST_ULL((u64)rate << 32, rate_in);
inv_ratio = DIV_ROUND_CLOSEST_ULL((u64)rate_in << 32, rate);
/* Can't increase by more than 64 or reduce by more than 256 */
if (rate > rate_in && ratio > (64ULL << 32))
return -EINVAL;
else if (rate <= rate_in && inv_ratio > (256ULL << 32))
return -EINVAL;
/*
* The divided clock (rate_in / r) must stay between 1.75 GHz and 13.3
* MHz. There is no minimum, since the only way to get a higher input
* clock than 26 MHz is to use a clock generated by a PLL. Because PLLs
* cannot output frequencies greater than 1.75 GHz, the minimum would
* never be greater than one.
*/
max_r = DIV_ROUND_DOWN_ULL(rate_in, 13300000);
/* Variables get immediately incremented, so start at -1th iteration */
i = -1;
f = 0;
r = 0;
od = 0;
best_error = S64_MAX;
error = best_error;
/* do-while here so we always try at least one ratio */
do {
/*
* Whether we swapped r and od while enforcing frequency limits
*/
bool swapped = false;
u64 last_od = od;
u64 last_r = r;
/*
* Try the next largest value for f (or r and od) and
* recalculate the other parameters based on that
*/
if (rate > rate_in) {
/*
* Skip factors of the same product if we already tried
* out that product
*/
do {
i++;
r = UNPACK_R(factors[i]);
od = UNPACK_OD(factors[i]);
} while (i + 1 < ARRAY_SIZE(factors) &&
r * od == last_r * last_od);
/* Round close */
f = (r * od * ratio + BIT(31)) >> 32;
if (f > 64)
f = 64;
} else {
u64 tmp = ++f * inv_ratio;
bool round_up = !!(tmp & BIT(31));
u32 goal = (tmp >> 32) + round_up;
u32 err, last_err;
/* Get the next r/od pair in factors */
while (r * od < goal && i + 1 < ARRAY_SIZE(factors)) {
i++;
r = UNPACK_R(factors[i]);
od = UNPACK_OD(factors[i]);
}
/*
* This is a case of double rounding. If we rounded up
* above, we need to round down (in cases of ties) here.
* This prevents off-by-one errors resulting from
* choosing X+2 over X when X.Y rounds up to X+1 and
* there is no r * od = X+1. For the converse, when X.Y
* is rounded down to X, we should choose X+1 over X-1.
*/
err = abs(r * od - goal);
last_err = abs(last_r * last_od - goal);
if (last_err < err || (round_up && last_err == err)) {
i--;
r = last_r;
od = last_od;
}
}
/*
* Enforce limits on internal clock frequencies. If we
* aren't in spec, try swapping r and od. If everything is
* in-spec, calculate the relative error.
*/
while (true) {
/*
* Whether the intermediate frequencies are out-of-spec
*/
bool out_of_spec = false;
if (r > max_r) {
out_of_spec = true;
} else {
/*
* There is no way to only divide once; we need
* to examine the frequency with and without the
* effect of od.
*/
u64 vco = DIV_ROUND_CLOSEST_ULL(rate_in * f, r);
if (vco > 1750000000 || vco < 340000000)
out_of_spec = true;
}
if (out_of_spec) {
if (!swapped) {
u64 tmp = r;
r = od;
od = tmp;
swapped = true;
continue;
} else {
/*
* Try looking ahead to see if there are
* additional factors for the same
* product.
*/
if (i + 1 < ARRAY_SIZE(factors)) {
u64 new_r, new_od;
i++;
new_r = UNPACK_R(factors[i]);
new_od = UNPACK_OD(factors[i]);
if (r * od == new_r * new_od) {
r = new_r;
od = new_od;
swapped = false;
continue;
}
i--;
}
break;
}
}
error = DIV_ROUND_CLOSEST_ULL(f * inv_ratio, r * od);
/* The lower 16 bits are spurious */
error = abs((error - BIT(32))) >> 16;
if (error < best_error) {
best->r = r;
best->f = f;
best->od = od;
best_error = error;
}
break;
}
} while (f < 64 && i + 1 < ARRAY_SIZE(factors) && error != 0);
if (best_error == S64_MAX)
return -EINVAL;
log_debug("best error %lld\n", best_error);
return 0;
}
static ulong k210_pll_set_rate(struct clk *clk, ulong rate)
{
int err;
long long rate_in = clk_get_parent_rate(clk);
struct k210_pll_config config = {};
struct k210_pll *pll = to_k210_pll(clk);
u32 reg;
if (rate_in < 0)
return rate_in;
log_debug("Calculating parameters with rate=%lu and rate_in=%lld\n",
rate, rate_in);
err = k210_pll_calc_config(rate, rate_in, &config);
if (err)
return err;
log_debug("Got r=%u f=%u od=%u\n", config.r, config.f, config.od);
/*
* Don't use clk_disable as it might not actually disable the pll due to
* refcounting
*/
k210_pll_disable(clk);
reg = readl(pll->reg);
reg &= ~K210_PLL_CLKR
& ~K210_PLL_CLKF
& ~K210_PLL_CLKOD
& ~K210_PLL_BWADJ;
reg |= FIELD_PREP(K210_PLL_CLKR, config.r - 1)
| FIELD_PREP(K210_PLL_CLKF, config.f - 1)
| FIELD_PREP(K210_PLL_CLKOD, config.od - 1)
| FIELD_PREP(K210_PLL_BWADJ, config.f - 1);
writel(reg, pll->reg);
err = k210_pll_enable(clk);
if (err)
return err;
serial_setbrg();
return clk_get_rate(clk);
}
#endif /* CONFIG_CLK_K210_SET_RATE */
static ulong k210_pll_get_rate(struct clk *clk)
{
long long rate_in = clk_get_parent_rate(clk);
struct k210_pll *pll = to_k210_pll(clk);
u64 r, f, od;
u32 reg = readl(pll->reg);
if (rate_in < 0 || (reg & K210_PLL_BYPASS))
return rate_in;
if (!(reg & K210_PLL_PWRD))
return 0;
r = FIELD_GET(K210_PLL_CLKR, reg) + 1;
f = FIELD_GET(K210_PLL_CLKF, reg) + 1;
od = FIELD_GET(K210_PLL_CLKOD, reg) + 1;
return DIV_ROUND_DOWN_ULL(((u64)rate_in) * f, r * od);
}
/*
* Wait for the PLL to be locked. If the PLL is not locked, try clearing the
* slip before retrying
*/
static void k210_pll_waitfor_lock(struct k210_pll *pll)
{
u32 mask = GENMASK(pll->width - 1, 0) << pll->shift;
while (true) {
u32 reg = readl(pll->lock);
if ((reg & mask) == mask)
break;
reg |= BIT(pll->shift + K210_PLL_CLEAR_SLIP);
writel(reg, pll->lock);
}
}
/* Adapted from sysctl_pll_enable */
static int k210_pll_enable(struct clk *clk)
{
struct k210_pll *pll = to_k210_pll(clk);
u32 reg = readl(pll->reg);
if ((reg | K210_PLL_PWRD) && !(reg | K210_PLL_RESET))
return 0;
reg |= K210_PLL_PWRD;
writel(reg, pll->reg);
/* Ensure reset is low before asserting it */
reg &= ~K210_PLL_RESET;
writel(reg, pll->reg);
reg |= K210_PLL_RESET;
writel(reg, pll->reg);
nop();
nop();
reg &= ~K210_PLL_RESET;
writel(reg, pll->reg);
k210_pll_waitfor_lock(pll);
reg &= ~K210_PLL_BYPASS;
writel(reg, pll->reg);
return 0;
}
static int k210_pll_disable(struct clk *clk)
{
struct k210_pll *pll = to_k210_pll(clk);
u32 reg = readl(pll->reg);
/*
* Bypassing before powering off is important so child clocks don't stop
* working. This is especially important for pll0, the indirect parent
* of the cpu clock.
*/
reg |= K210_PLL_BYPASS;
writel(reg, pll->reg);
reg &= ~K210_PLL_PWRD;
writel(reg, pll->reg);
return 0;
}
const struct clk_ops k210_pll_ops = {
.get_rate = k210_pll_get_rate,
#ifdef CONFIG_CLK_K210_SET_RATE
.set_rate = k210_pll_set_rate,
#endif
.enable = k210_pll_enable,
.disable = k210_pll_disable,
};
struct clk *k210_register_pll_struct(const char *name, const char *parent_name,
struct k210_pll *pll)
{
int ret;
struct clk *clk = &pll->clk;
ret = clk_register(clk, CLK_K210_PLL, name, parent_name);
if (ret)
return ERR_PTR(ret);
return clk;
}
struct clk *k210_register_pll(const char *name, const char *parent_name,
void __iomem *reg, void __iomem *lock, u8 shift,
u8 width)
{
struct clk *clk;
struct k210_pll *pll;
pll = kzalloc(sizeof(*pll), GFP_KERNEL);
if (!pll)
return ERR_PTR(-ENOMEM);
pll->reg = reg;
pll->lock = lock;
pll->shift = shift;
pll->width = width;
clk = k210_register_pll_struct(name, parent_name, pll);
if (IS_ERR(clk))
kfree(pll);
return clk;
}
U_BOOT_DRIVER(k210_pll) = {
.name = CLK_K210_PLL,
.id = UCLASS_CLK,
.ops = &k210_pll_ops,
};