cppc_cpufreq.c 13 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0-only
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 * CPPC (Collaborative Processor Performance Control) driver for
 * interfacing with the CPUfreq layer and governors. See
 * cppc_acpi.c for CPPC specific methods.
 *
 * (C) Copyright 2014, 2015 Linaro Ltd.
 * Author: Ashwin Chaugule <ashwin.chaugule@linaro.org>
 */

#define pr_fmt(fmt)	"CPPC Cpufreq:"	fmt

#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/delay.h>
#include <linux/cpu.h>
#include <linux/cpufreq.h>
18
#include <linux/dmi.h>
19
#include <linux/time.h>
20 21
#include <linux/vmalloc.h>

22 23
#include <asm/unaligned.h>

24 25
#include <acpi/cppc_acpi.h>

26 27 28 29 30 31
/* Minimum struct length needed for the DMI processor entry we want */
#define DMI_ENTRY_PROCESSOR_MIN_LENGTH	48

/* Offest in the DMI processor structure for the max frequency */
#define DMI_PROCESSOR_MAX_SPEED  0x14

32 33 34 35 36 37 38
/*
 * These structs contain information parsed from per CPU
 * ACPI _CPC structures.
 * e.g. For each CPU the highest, lowest supported
 * performance capabilities, desired performance level
 * requested etc.
 */
39
static struct cppc_cpudata **all_cpu_data;
40
static bool boost_supported;
41

42
struct cppc_workaround_oem_info {
43
	char oem_id[ACPI_OEM_ID_SIZE + 1];
44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
	char oem_table_id[ACPI_OEM_TABLE_ID_SIZE + 1];
	u32 oem_revision;
};

static struct cppc_workaround_oem_info wa_info[] = {
	{
		.oem_id		= "HISI  ",
		.oem_table_id	= "HIP07   ",
		.oem_revision	= 0,
	}, {
		.oem_id		= "HISI  ",
		.oem_table_id	= "HIP08   ",
		.oem_revision	= 0,
	}
};

60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89
/* Callback function used to retrieve the max frequency from DMI */
static void cppc_find_dmi_mhz(const struct dmi_header *dm, void *private)
{
	const u8 *dmi_data = (const u8 *)dm;
	u16 *mhz = (u16 *)private;

	if (dm->type == DMI_ENTRY_PROCESSOR &&
	    dm->length >= DMI_ENTRY_PROCESSOR_MIN_LENGTH) {
		u16 val = (u16)get_unaligned((const u16 *)
				(dmi_data + DMI_PROCESSOR_MAX_SPEED));
		*mhz = val > *mhz ? val : *mhz;
	}
}

/* Look up the max frequency in DMI */
static u64 cppc_get_dmi_max_khz(void)
{
	u16 mhz = 0;

	dmi_walk(cppc_find_dmi_mhz, &mhz);

	/*
	 * Real stupid fallback value, just in case there is no
	 * actual value set.
	 */
	mhz = mhz ? mhz : 1;

	return (1000 * mhz);
}

90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147
/*
 * If CPPC lowest_freq and nominal_freq registers are exposed then we can
 * use them to convert perf to freq and vice versa
 *
 * If the perf/freq point lies between Nominal and Lowest, we can treat
 * (Low perf, Low freq) and (Nom Perf, Nom freq) as 2D co-ordinates of a line
 * and extrapolate the rest
 * For perf/freq > Nominal, we use the ratio perf:freq at Nominal for conversion
 */
static unsigned int cppc_cpufreq_perf_to_khz(struct cppc_cpudata *cpu,
					unsigned int perf)
{
	static u64 max_khz;
	struct cppc_perf_caps *caps = &cpu->perf_caps;
	u64 mul, div;

	if (caps->lowest_freq && caps->nominal_freq) {
		if (perf >= caps->nominal_perf) {
			mul = caps->nominal_freq;
			div = caps->nominal_perf;
		} else {
			mul = caps->nominal_freq - caps->lowest_freq;
			div = caps->nominal_perf - caps->lowest_perf;
		}
	} else {
		if (!max_khz)
			max_khz = cppc_get_dmi_max_khz();
		mul = max_khz;
		div = cpu->perf_caps.highest_perf;
	}
	return (u64)perf * mul / div;
}

static unsigned int cppc_cpufreq_khz_to_perf(struct cppc_cpudata *cpu,
					unsigned int freq)
{
	static u64 max_khz;
	struct cppc_perf_caps *caps = &cpu->perf_caps;
	u64  mul, div;

	if (caps->lowest_freq && caps->nominal_freq) {
		if (freq >= caps->nominal_freq) {
			mul = caps->nominal_perf;
			div = caps->nominal_freq;
		} else {
			mul = caps->lowest_perf;
			div = caps->lowest_freq;
		}
	} else {
		if (!max_khz)
			max_khz = cppc_get_dmi_max_khz();
		mul = cpu->perf_caps.highest_perf;
		div = max_khz;
	}

	return (u64)freq * mul / div;
}

148 149 150 151
static int cppc_cpufreq_set_target(struct cpufreq_policy *policy,
		unsigned int target_freq,
		unsigned int relation)
{
152
	struct cppc_cpudata *cpu;
153
	struct cpufreq_freqs freqs;
154
	u32 desired_perf;
155 156 157 158
	int ret = 0;

	cpu = all_cpu_data[policy->cpu];

159
	desired_perf = cppc_cpufreq_khz_to_perf(cpu, target_freq);
160 161 162 163 164
	/* Return if it is exactly the same perf */
	if (desired_perf == cpu->perf_ctrls.desired_perf)
		return ret;

	cpu->perf_ctrls.desired_perf = desired_perf;
165 166 167 168 169 170 171 172 173 174 175 176 177 178
	freqs.old = policy->cur;
	freqs.new = target_freq;

	cpufreq_freq_transition_begin(policy, &freqs);
	ret = cppc_set_perf(cpu->cpu, &cpu->perf_ctrls);
	cpufreq_freq_transition_end(policy, &freqs, ret != 0);

	if (ret)
		pr_debug("Failed to set target on CPU:%d. ret:%d\n",
				cpu->cpu, ret);

	return ret;
}

179
static int cppc_verify_policy(struct cpufreq_policy_data *policy)
180 181 182 183 184 185 186 187
{
	cpufreq_verify_within_cpu_limits(policy);
	return 0;
}

static void cppc_cpufreq_stop_cpu(struct cpufreq_policy *policy)
{
	int cpu_num = policy->cpu;
188
	struct cppc_cpudata *cpu = all_cpu_data[cpu_num];
189 190 191 192 193 194 195 196 197 198
	int ret;

	cpu->perf_ctrls.desired_perf = cpu->perf_caps.lowest_perf;

	ret = cppc_set_perf(cpu_num, &cpu->perf_ctrls);
	if (ret)
		pr_debug("Err setting perf value:%d on CPU:%d. ret:%d\n",
				cpu->perf_caps.lowest_perf, cpu_num, ret);
}

199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241
/*
 * The PCC subspace describes the rate at which platform can accept commands
 * on the shared PCC channel (including READs which do not count towards freq
 * trasition requests), so ideally we need to use the PCC values as a fallback
 * if we don't have a platform specific transition_delay_us
 */
#ifdef CONFIG_ARM64
#include <asm/cputype.h>

static unsigned int cppc_cpufreq_get_transition_delay_us(int cpu)
{
	unsigned long implementor = read_cpuid_implementor();
	unsigned long part_num = read_cpuid_part_number();
	unsigned int delay_us = 0;

	switch (implementor) {
	case ARM_CPU_IMP_QCOM:
		switch (part_num) {
		case QCOM_CPU_PART_FALKOR_V1:
		case QCOM_CPU_PART_FALKOR:
			delay_us = 10000;
			break;
		default:
			delay_us = cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
			break;
		}
		break;
	default:
		delay_us = cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
		break;
	}

	return delay_us;
}

#else

static unsigned int cppc_cpufreq_get_transition_delay_us(int cpu)
{
	return cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
}
#endif

242 243
static int cppc_cpufreq_cpu_init(struct cpufreq_policy *policy)
{
244
	struct cppc_cpudata *cpu;
245 246 247 248 249 250 251 252 253 254 255 256 257 258
	unsigned int cpu_num = policy->cpu;
	int ret = 0;

	cpu = all_cpu_data[policy->cpu];

	cpu->cpu = cpu_num;
	ret = cppc_get_perf_caps(policy->cpu, &cpu->perf_caps);

	if (ret) {
		pr_debug("Err reading CPU%d perf capabilities. ret:%d\n",
				cpu_num, ret);
		return ret;
	}

259 260 261
	/* Convert the lowest and nominal freq from MHz to KHz */
	cpu->perf_caps.lowest_freq *= 1000;
	cpu->perf_caps.nominal_freq *= 1000;
262

263 264 265 266
	/*
	 * Set min to lowest nonlinear perf to avoid any efficiency penalty (see
	 * Section 8.4.7.1.1.5 of ACPI 6.1 spec)
	 */
267
	policy->min = cppc_cpufreq_perf_to_khz(cpu, cpu->perf_caps.lowest_nonlinear_perf);
268
	policy->max = cppc_cpufreq_perf_to_khz(cpu, cpu->perf_caps.nominal_perf);
269 270 271 272 273 274

	/*
	 * Set cpuinfo.min_freq to Lowest to make the full range of performance
	 * available if userspace wants to use any perf between lowest & lowest
	 * nonlinear perf
	 */
275
	policy->cpuinfo.min_freq = cppc_cpufreq_perf_to_khz(cpu, cpu->perf_caps.lowest_perf);
276
	policy->cpuinfo.max_freq = cppc_cpufreq_perf_to_khz(cpu, cpu->perf_caps.nominal_perf);
277

278
	policy->transition_delay_us = cppc_cpufreq_get_transition_delay_us(cpu_num);
279
	policy->shared_type = cpu->shared_type;
280

281 282 283
	if (policy->shared_type == CPUFREQ_SHARED_TYPE_ANY) {
		int i;

284
		cpumask_copy(policy->cpus, cpu->shared_cpu_map);
285 286 287 288 289 290 291 292 293

		for_each_cpu(i, policy->cpus) {
			if (unlikely(i == policy->cpu))
				continue;

			memcpy(&all_cpu_data[i]->perf_caps, &cpu->perf_caps,
			       sizeof(cpu->perf_caps));
		}
	} else if (policy->shared_type == CPUFREQ_SHARED_TYPE_ALL) {
294 295 296 297 298 299 300
		/* Support only SW_ANY for now. */
		pr_debug("Unsupported CPU co-ord type\n");
		return -EFAULT;
	}

	cpu->cur_policy = policy;

301 302 303 304 305 306 307
	/*
	 * If 'highest_perf' is greater than 'nominal_perf', we assume CPU Boost
	 * is supported.
	 */
	if (cpu->perf_caps.highest_perf > cpu->perf_caps.nominal_perf)
		boost_supported = true;

308
	/* Set policy->cur to max now. The governors will adjust later. */
309 310
	policy->cur = cppc_cpufreq_perf_to_khz(cpu,
					cpu->perf_caps.highest_perf);
311
	cpu->perf_ctrls.desired_perf = cpu->perf_caps.highest_perf;
312 313 314 315 316 317 318 319 320

	ret = cppc_set_perf(cpu_num, &cpu->perf_ctrls);
	if (ret)
		pr_debug("Err setting perf value:%d on CPU:%d. ret:%d\n",
				cpu->perf_caps.highest_perf, cpu_num, ret);

	return ret;
}

321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371
static inline u64 get_delta(u64 t1, u64 t0)
{
	if (t1 > t0 || t0 > ~(u32)0)
		return t1 - t0;

	return (u32)t1 - (u32)t0;
}

static int cppc_get_rate_from_fbctrs(struct cppc_cpudata *cpu,
				     struct cppc_perf_fb_ctrs fb_ctrs_t0,
				     struct cppc_perf_fb_ctrs fb_ctrs_t1)
{
	u64 delta_reference, delta_delivered;
	u64 reference_perf, delivered_perf;

	reference_perf = fb_ctrs_t0.reference_perf;

	delta_reference = get_delta(fb_ctrs_t1.reference,
				    fb_ctrs_t0.reference);
	delta_delivered = get_delta(fb_ctrs_t1.delivered,
				    fb_ctrs_t0.delivered);

	/* Check to avoid divide-by zero */
	if (delta_reference || delta_delivered)
		delivered_perf = (reference_perf * delta_delivered) /
					delta_reference;
	else
		delivered_perf = cpu->perf_ctrls.desired_perf;

	return cppc_cpufreq_perf_to_khz(cpu, delivered_perf);
}

static unsigned int cppc_cpufreq_get_rate(unsigned int cpunum)
{
	struct cppc_perf_fb_ctrs fb_ctrs_t0 = {0}, fb_ctrs_t1 = {0};
	struct cppc_cpudata *cpu = all_cpu_data[cpunum];
	int ret;

	ret = cppc_get_perf_ctrs(cpunum, &fb_ctrs_t0);
	if (ret)
		return ret;

	udelay(2); /* 2usec delay between sampling */

	ret = cppc_get_perf_ctrs(cpunum, &fb_ctrs_t1);
	if (ret)
		return ret;

	return cppc_get_rate_from_fbctrs(cpu, fb_ctrs_t0, fb_ctrs_t1);
}

372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397
static int cppc_cpufreq_set_boost(struct cpufreq_policy *policy, int state)
{
	struct cppc_cpudata *cpudata;
	int ret;

	if (!boost_supported) {
		pr_err("BOOST not supported by CPU or firmware\n");
		return -EINVAL;
	}

	cpudata = all_cpu_data[policy->cpu];
	if (state)
		policy->max = cppc_cpufreq_perf_to_khz(cpudata,
					cpudata->perf_caps.highest_perf);
	else
		policy->max = cppc_cpufreq_perf_to_khz(cpudata,
					cpudata->perf_caps.nominal_perf);
	policy->cpuinfo.max_freq = policy->max;

	ret = freq_qos_update_request(policy->max_freq_req, policy->max);
	if (ret < 0)
		return ret;

	return 0;
}

398 399 400 401
static struct cpufreq_driver cppc_cpufreq_driver = {
	.flags = CPUFREQ_CONST_LOOPS,
	.verify = cppc_verify_policy,
	.target = cppc_cpufreq_set_target,
402
	.get = cppc_cpufreq_get_rate,
403 404
	.init = cppc_cpufreq_cpu_init,
	.stop_cpu = cppc_cpufreq_stop_cpu,
405
	.set_boost = cppc_cpufreq_set_boost,
406 407 408
	.name = "cppc_cpufreq",
};

409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450
/*
 * HISI platform does not support delivered performance counter and
 * reference performance counter. It can calculate the performance using the
 * platform specific mechanism. We reuse the desired performance register to
 * store the real performance calculated by the platform.
 */
static unsigned int hisi_cppc_cpufreq_get_rate(unsigned int cpunum)
{
	struct cppc_cpudata *cpudata = all_cpu_data[cpunum];
	u64 desired_perf;
	int ret;

	ret = cppc_get_desired_perf(cpunum, &desired_perf);
	if (ret < 0)
		return -EIO;

	return cppc_cpufreq_perf_to_khz(cpudata, desired_perf);
}

static void cppc_check_hisi_workaround(void)
{
	struct acpi_table_header *tbl;
	acpi_status status = AE_OK;
	int i;

	status = acpi_get_table(ACPI_SIG_PCCT, 0, &tbl);
	if (ACPI_FAILURE(status) || !tbl)
		return;

	for (i = 0; i < ARRAY_SIZE(wa_info); i++) {
		if (!memcmp(wa_info[i].oem_id, tbl->oem_id, ACPI_OEM_ID_SIZE) &&
		    !memcmp(wa_info[i].oem_table_id, tbl->oem_table_id, ACPI_OEM_TABLE_ID_SIZE) &&
		    wa_info[i].oem_revision == tbl->oem_revision) {
			/* Overwrite the get() callback */
			cppc_cpufreq_driver.get = hisi_cppc_cpufreq_get_rate;
			break;
		}
	}

	acpi_put_table(tbl);
}

451 452 453
static int __init cppc_cpufreq_init(void)
{
	int i, ret = 0;
454
	struct cppc_cpudata *cpu;
455 456 457 458

	if (acpi_disabled)
		return -ENODEV;

459 460
	all_cpu_data = kcalloc(num_possible_cpus(), sizeof(void *),
			       GFP_KERNEL);
461 462 463 464
	if (!all_cpu_data)
		return -ENOMEM;

	for_each_possible_cpu(i) {
465
		all_cpu_data[i] = kzalloc(sizeof(struct cppc_cpudata), GFP_KERNEL);
466 467 468 469 470 471 472 473 474 475 476 477 478 479
		if (!all_cpu_data[i])
			goto out;

		cpu = all_cpu_data[i];
		if (!zalloc_cpumask_var(&cpu->shared_cpu_map, GFP_KERNEL))
			goto out;
	}

	ret = acpi_get_psd_map(all_cpu_data);
	if (ret) {
		pr_debug("Error parsing PSD data. Aborting cpufreq registration.\n");
		goto out;
	}

480 481
	cppc_check_hisi_workaround();

482 483 484 485 486 487 488
	ret = cpufreq_register_driver(&cppc_cpufreq_driver);
	if (ret)
		goto out;

	return ret;

out:
489 490 491 492 493 494 495
	for_each_possible_cpu(i) {
		cpu = all_cpu_data[i];
		if (!cpu)
			break;
		free_cpumask_var(cpu->shared_cpu_map);
		kfree(cpu);
	}
496 497 498 499 500

	kfree(all_cpu_data);
	return -ENODEV;
}

501 502
static void __exit cppc_cpufreq_exit(void)
{
503
	struct cppc_cpudata *cpu;
504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521
	int i;

	cpufreq_unregister_driver(&cppc_cpufreq_driver);

	for_each_possible_cpu(i) {
		cpu = all_cpu_data[i];
		free_cpumask_var(cpu->shared_cpu_map);
		kfree(cpu);
	}

	kfree(all_cpu_data);
}

module_exit(cppc_cpufreq_exit);
MODULE_AUTHOR("Ashwin Chaugule");
MODULE_DESCRIPTION("CPUFreq driver based on the ACPI CPPC v5.0+ spec");
MODULE_LICENSE("GPL");

522
late_initcall(cppc_cpufreq_init);
523

524
static const struct acpi_device_id cppc_acpi_ids[] __used = {
525 526 527 528 529
	{ACPI_PROCESSOR_DEVICE_HID, },
	{}
};

MODULE_DEVICE_TABLE(acpi, cppc_acpi_ids);