setup_64.c 21.5 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12
/*
 * 
 * Common boot and setup code.
 *
 * Copyright (C) 2001 PPC64 Team, IBM Corp
 *
 *      This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

13
#include <linux/export.h>
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
#include <linux/string.h>
#include <linux/sched.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/reboot.h>
#include <linux/delay.h>
#include <linux/initrd.h>
#include <linux/seq_file.h>
#include <linux/ioport.h>
#include <linux/console.h>
#include <linux/utsname.h>
#include <linux/tty.h>
#include <linux/root_dev.h>
#include <linux/notifier.h>
#include <linux/cpu.h>
#include <linux/unistd.h>
#include <linux/serial.h>
#include <linux/serial_8250.h>
32
#include <linux/bootmem.h>
33
#include <linux/pci.h>
34
#include <linux/lockdep.h>
Y
Yinghai Lu 已提交
35
#include <linux/memblock.h>
36
#include <linux/memory.h>
37
#include <linux/nmi.h>
38

39
#include <asm/io.h>
40
#include <asm/kdump.h>
41 42 43 44 45 46 47 48 49
#include <asm/prom.h>
#include <asm/processor.h>
#include <asm/pgtable.h>
#include <asm/smp.h>
#include <asm/elf.h>
#include <asm/machdep.h>
#include <asm/paca.h>
#include <asm/time.h>
#include <asm/cputable.h>
50
#include <asm/dt_cpu_ftrs.h>
51 52 53 54 55 56 57 58 59 60 61
#include <asm/sections.h>
#include <asm/btext.h>
#include <asm/nvram.h>
#include <asm/setup.h>
#include <asm/rtas.h>
#include <asm/iommu.h>
#include <asm/serial.h>
#include <asm/cache.h>
#include <asm/page.h>
#include <asm/mmu.h>
#include <asm/firmware.h>
P
Paul Mackerras 已提交
62
#include <asm/xmon.h>
D
David Gibson 已提交
63
#include <asm/udbg.h>
64
#include <asm/kexec.h>
65
#include <asm/code-patching.h>
66
#include <asm/livepatch.h>
67
#include <asm/opal.h>
68
#include <asm/cputhreads.h>
69
#include <asm/hw_irq.h>
70

71 72
#include "setup.h"

73 74 75 76 77 78
#ifdef DEBUG
#define DBG(fmt...) udbg_printf(fmt)
#else
#define DBG(fmt...)
#endif

79
int spinning_secondaries;
80 81
u64 ppc64_pft_size;

82
struct ppc64_caches ppc64_caches = {
83 84 85 86 87 88 89 90
	.l1d = {
		.block_size = 0x40,
		.log_block_size = 6,
	},
	.l1i = {
		.block_size = 0x40,
		.log_block_size = 6
	},
91
};
92 93
EXPORT_SYMBOL_GPL(ppc64_caches);

94
#if defined(CONFIG_PPC_BOOK3E) && defined(CONFIG_SMP)
95
void __init setup_tlb_core_data(void)
96 97 98
{
	int cpu;

99 100
	BUILD_BUG_ON(offsetof(struct tlb_core_data, lock) != 0);

101 102 103
	for_each_possible_cpu(cpu) {
		int first = cpu_first_thread_sibling(cpu);

104 105 106 107 108 109 110 111
		/*
		 * If we boot via kdump on a non-primary thread,
		 * make sure we point at the thread that actually
		 * set up this TLB.
		 */
		if (cpu_first_thread_sibling(boot_cpuid) == first)
			first = boot_cpuid;

112 113 114 115 116 117
		paca[cpu].tcd_ptr = &paca[first].tcd;

		/*
		 * If we have threads, we need either tlbsrx.
		 * or e6500 tablewalk mode, or else TLB handlers
		 * will be racy and could produce duplicate entries.
118
		 * Should we panic instead?
119
		 */
120 121 122 123
		WARN_ONCE(smt_enabled_at_boot >= 2 &&
			  !mmu_has_feature(MMU_FTR_USE_TLBRSRV) &&
			  book3e_htw_mode != PPC_HTW_E6500,
			  "%s: unsupported MMU configuration\n", __func__);
124 125 126 127
	}
}
#endif

128 129
#ifdef CONFIG_SMP

130
static char *smt_enabled_cmdline;
131 132

/* Look for ibm,smt-enabled OF option */
133
void __init check_smt_enabled(void)
134 135
{
	struct device_node *dn;
136
	const char *smt_option;
137

138 139
	/* Default to enabling all threads */
	smt_enabled_at_boot = threads_per_core;
140

141 142 143 144 145 146 147
	/* Allow the command line to overrule the OF option */
	if (smt_enabled_cmdline) {
		if (!strcmp(smt_enabled_cmdline, "on"))
			smt_enabled_at_boot = threads_per_core;
		else if (!strcmp(smt_enabled_cmdline, "off"))
			smt_enabled_at_boot = 0;
		else {
148
			int smt;
149 150
			int rc;

151
			rc = kstrtoint(smt_enabled_cmdline, 10, &smt);
152 153
			if (!rc)
				smt_enabled_at_boot =
154
					min(threads_per_core, smt);
155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171
		}
	} else {
		dn = of_find_node_by_path("/options");
		if (dn) {
			smt_option = of_get_property(dn, "ibm,smt-enabled",
						     NULL);

			if (smt_option) {
				if (!strcmp(smt_option, "on"))
					smt_enabled_at_boot = threads_per_core;
				else if (!strcmp(smt_option, "off"))
					smt_enabled_at_boot = 0;
			}

			of_node_put(dn);
		}
	}
172 173 174 175 176
}

/* Look for smt-enabled= cmdline option */
static int __init early_smt_enabled(char *p)
{
177
	smt_enabled_cmdline = p;
178 179 180 181 182 183
	return 0;
}
early_param("smt-enabled", early_smt_enabled);

#endif /* CONFIG_SMP */

184
/** Fix up paca fields required for the boot cpu */
185
static void __init fixup_boot_paca(void)
186 187 188 189 190
{
	/* The boot cpu is started */
	get_paca()->cpu_start = 1;
	/* Allow percpu accesses to work until we setup percpu data */
	get_paca()->data_offset = 0;
191
	/* Mark interrupts disabled in PACA */
192
	soft_enabled_set(IRQS_DISABLED);
193 194
}

195
static void __init configure_exceptions(void)
196
{
197
	/*
198 199
	 * Setup the trampolines from the lowmem exception vectors
	 * to the kdump kernel when not using a relocatable kernel.
200
	 */
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223
	setup_kdump_trampoline();

	/* Under a PAPR hypervisor, we need hypercalls */
	if (firmware_has_feature(FW_FEATURE_SET_MODE)) {
		/* Enable AIL if possible */
		pseries_enable_reloc_on_exc();

		/*
		 * Tell the hypervisor that we want our exceptions to
		 * be taken in little endian mode.
		 *
		 * We don't call this for big endian as our calling convention
		 * makes us always enter in BE, and the call may fail under
		 * some circumstances with kdump.
		 */
#ifdef __LITTLE_ENDIAN__
		pseries_little_endian_exceptions();
#endif
	} else {
		/* Set endian mode using OPAL */
		if (firmware_has_feature(FW_FEATURE_OPAL))
			opal_configure_cores();

224
		/* AIL on native is done in cpu_ready_for_interrupts() */
225 226 227
	}
}

228 229
static void cpu_ready_for_interrupts(void)
{
230 231 232 233 234 235 236
	/*
	 * Enable AIL if supported, and we are in hypervisor mode. This
	 * is called once for every processor.
	 *
	 * If we are not in hypervisor mode the job is done once for
	 * the whole partition in configure_exceptions().
	 */
237 238
	if (cpu_has_feature(CPU_FTR_HVMODE) &&
	    cpu_has_feature(CPU_FTR_ARCH_207S)) {
239 240 241 242
		unsigned long lpcr = mfspr(SPRN_LPCR);
		mtspr(SPRN_LPCR, lpcr | LPCR_AIL_3);
	}

243 244 245 246 247 248 249 250 251
	/*
	 * Fixup HFSCR:TM based on CPU features. The bit is set by our
	 * early asm init because at that point we haven't updated our
	 * CPU features from firmware and device-tree. Here we have,
	 * so let's do it.
	 */
	if (cpu_has_feature(CPU_FTR_HVMODE) && !cpu_has_feature(CPU_FTR_TM_COMP))
		mtspr(SPRN_HFSCR, mfspr(SPRN_HFSCR) & ~HFSCR_TM);

252 253 254 255
	/* Set IR and DR in PACA MSR */
	get_paca()->kernel_msr = MSR_KERNEL;
}

256 257 258 259 260 261
/*
 * Early initialization entry point. This is called by head.S
 * with MMU translation disabled. We rely on the "feature" of
 * the CPU that ignores the top 2 bits of the address in real
 * mode so we can access kernel globals normally provided we
 * only toy with things in the RMO region. From here, we do
Y
Yinghai Lu 已提交
262
 * some early parsing of the device-tree to setup out MEMBLOCK
263 264 265 266 267 268 269 270 271 272 273 274 275 276
 * data structures, and allocate & initialize the hash table
 * and segment tables so we can start running with translation
 * enabled.
 *
 * It is this function which will call the probe() callback of
 * the various platform types and copy the matching one to the
 * global ppc_md structure. Your platform can eventually do
 * some very early initializations from the probe() routine, but
 * this is not recommended, be very careful as, for example, the
 * device-tree is not accessible via normal means at this point.
 */

void __init early_setup(unsigned long dt_ptr)
{
277 278
	static __initdata struct paca_struct boot_paca;

279 280
	/* -------- printk is _NOT_ safe to use here ! ------- */

281 282 283 284
	/* Try new device tree based feature discovery ... */
	if (!dt_cpu_ftrs_init(__va(dt_ptr)))
		/* Otherwise use the old style CPU table */
		identify_cpu(0, mfspr(SPRN_PVR));
285

286
	/* Assume we're on cpu 0 for now. Don't write to the paca yet! */
287 288
	initialise_paca(&boot_paca, 0);
	setup_paca(&boot_paca);
289
	fixup_boot_paca();
290

291 292
	/* -------- printk is now safe to use ------- */

293 294 295
	/* Enable early debugging if any specified (see udbg.h) */
	udbg_early_init();

296
 	DBG(" -> early_setup(), dt_ptr: 0x%lx\n", dt_ptr);
297 298

	/*
299 300 301
	 * Do early initialization using the flattened device
	 * tree, such as retrieving the physical memory map or
	 * calculating/retrieving the hash table size.
302 303 304
	 */
	early_init_devtree(__va(dt_ptr));

305
	/* Now we know the logical id of our boot cpu, setup the paca. */
306
	setup_paca(&paca[boot_cpuid]);
307
	fixup_boot_paca();
308

309
	/*
310 311
	 * Configure exception handlers. This include setting up trampolines
	 * if needed, setting exception endian mode, etc...
312
	 */
313
	configure_exceptions();
314

315 316
	/* Apply all the dynamic patching */
	apply_feature_fixups();
317
	setup_feature_keys();
318

319 320 321
	/* Initialize the hash table or TLB handling */
	early_init_mmu();

322 323 324 325 326 327 328
	/*
	 * After firmware and early platform setup code has set things up,
	 * we note the SPR values for configurable control/performance
	 * registers, and use those as initial defaults.
	 */
	record_spr_defaults();

329 330 331
	/*
	 * At this point, we can let interrupts switch to virtual mode
	 * (the MMU has been setup), so adjust the MSR in the PACA to
332
	 * have IR and DR set and enable AIL if it exists
333
	 */
334
	cpu_ready_for_interrupts();
335

336
	DBG(" <- early_setup()\n");
337 338 339 340 341 342 343 344 345 346 347 348

#ifdef CONFIG_PPC_EARLY_DEBUG_BOOTX
	/*
	 * This needs to be done *last* (after the above DBG() even)
	 *
	 * Right after we return from this function, we turn on the MMU
	 * which means the real-mode access trick that btext does will
	 * no longer work, it needs to switch to using a real MMU
	 * mapping. This call will ensure that it does
	 */
	btext_map();
#endif /* CONFIG_PPC_EARLY_DEBUG_BOOTX */
349 350
}

351 352 353
#ifdef CONFIG_SMP
void early_setup_secondary(void)
{
354
	/* Mark interrupts disabled in PACA */
355
	soft_enabled_set(IRQS_DISABLED);
356

357 358
	/* Initialize the hash table or TLB handling */
	early_init_mmu_secondary();
359 360 361 362 363 364

	/*
	 * At this point, we can let interrupts switch to virtual mode
	 * (the MMU has been setup), so adjust the MSR in the PACA to
	 * have IR and DR set.
	 */
365
	cpu_ready_for_interrupts();
366 367 368
}

#endif /* CONFIG_SMP */
369

370
#if defined(CONFIG_SMP) || defined(CONFIG_KEXEC_CORE)
371 372
static bool use_spinloop(void)
{
373 374 375 376 377 378 379 380
	if (IS_ENABLED(CONFIG_PPC_BOOK3S)) {
		/*
		 * See comments in head_64.S -- not all platforms insert
		 * secondaries at __secondary_hold and wait at the spin
		 * loop.
		 */
		if (firmware_has_feature(FW_FEATURE_OPAL))
			return false;
381
		return true;
382
	}
383 384 385 386 387 388 389 390

	/*
	 * When book3e boots from kexec, the ePAPR spin table does
	 * not get used.
	 */
	return of_property_read_bool(of_chosen, "linux,booted-from-kexec");
}

391 392
void smp_release_cpus(void)
{
393
	unsigned long *ptr;
394
	int i;
395

396 397 398
	if (!use_spinloop())
		return;

399 400 401 402 403 404
	DBG(" -> smp_release_cpus()\n");

	/* All secondary cpus are spinning on a common spinloop, release them
	 * all now so they can start to spin on their individual paca
	 * spinloops. For non SMP kernels, the secondary cpus never get out
	 * of the common spinloop.
405
	 */
406

407 408
	ptr  = (unsigned long *)((unsigned long)&__secondary_hold_spinloop
			- PHYSICAL_START);
409
	*ptr = ppc_function_entry(generic_secondary_smp_init);
410 411 412 413 414

	/* And wait a bit for them to catch up */
	for (i = 0; i < 100000; i++) {
		mb();
		HMT_low();
415
		if (spinning_secondaries == 0)
416 417 418
			break;
		udelay(1);
	}
419
	DBG("spinning_secondaries = %d\n", spinning_secondaries);
420 421 422

	DBG(" <- smp_release_cpus()\n");
}
423
#endif /* CONFIG_SMP || CONFIG_KEXEC_CORE */
424

425
/*
426 427
 * Initialize some remaining members of the ppc64_caches and systemcfg
 * structures
428 429 430 431
 * (at least until we get rid of them completely). This is mostly some
 * cache informations about the CPU that will be used by cache flush
 * routines and/or provided to userland
 */
432 433 434 435 436 437 438 439 440

static void init_cache_info(struct ppc_cache_info *info, u32 size, u32 lsize,
			    u32 bsize, u32 sets)
{
	info->size = size;
	info->sets = sets;
	info->line_size = lsize;
	info->block_size = bsize;
	info->log_block_size = __ilog2(bsize);
441 442 443 444
	if (bsize)
		info->blocks_per_page = PAGE_SIZE / bsize;
	else
		info->blocks_per_page = 0;
445 446 447 448 449

	if (sets == 0)
		info->assoc = 0xffff;
	else
		info->assoc = size / (sets * lsize);
450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508
}

static bool __init parse_cache_info(struct device_node *np,
				    bool icache,
				    struct ppc_cache_info *info)
{
	static const char *ipropnames[] __initdata = {
		"i-cache-size",
		"i-cache-sets",
		"i-cache-block-size",
		"i-cache-line-size",
	};
	static const char *dpropnames[] __initdata = {
		"d-cache-size",
		"d-cache-sets",
		"d-cache-block-size",
		"d-cache-line-size",
	};
	const char **propnames = icache ? ipropnames : dpropnames;
	const __be32 *sizep, *lsizep, *bsizep, *setsp;
	u32 size, lsize, bsize, sets;
	bool success = true;

	size = 0;
	sets = -1u;
	lsize = bsize = cur_cpu_spec->dcache_bsize;
	sizep = of_get_property(np, propnames[0], NULL);
	if (sizep != NULL)
		size = be32_to_cpu(*sizep);
	setsp = of_get_property(np, propnames[1], NULL);
	if (setsp != NULL)
		sets = be32_to_cpu(*setsp);
	bsizep = of_get_property(np, propnames[2], NULL);
	lsizep = of_get_property(np, propnames[3], NULL);
	if (bsizep == NULL)
		bsizep = lsizep;
	if (lsizep != NULL)
		lsize = be32_to_cpu(*lsizep);
	if (bsizep != NULL)
		bsize = be32_to_cpu(*bsizep);
	if (sizep == NULL || bsizep == NULL || lsizep == NULL)
		success = false;

	/*
	 * OF is weird .. it represents fully associative caches
	 * as "1 way" which doesn't make much sense and doesn't
	 * leave room for direct mapped. We'll assume that 0
	 * in OF means direct mapped for that reason.
	 */
	if (sets == 1)
		sets = 0;
	else if (sets == 0)
		sets = 1;

	init_cache_info(info, size, lsize, bsize, sets);

	return success;
}

509
void __init initialize_cache_info(void)
510
{
511 512
	struct device_node *cpu = NULL, *l2, *l3 = NULL;
	u32 pvr;
513 514 515

	DBG(" -> initialize_cache_info()\n");

516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
	/*
	 * All shipping POWER8 machines have a firmware bug that
	 * puts incorrect information in the device-tree. This will
	 * be (hopefully) fixed for future chips but for now hard
	 * code the values if we are running on one of these
	 */
	pvr = PVR_VER(mfspr(SPRN_PVR));
	if (pvr == PVR_POWER8 || pvr == PVR_POWER8E ||
	    pvr == PVR_POWER8NVL) {
						/* size    lsize   blk  sets */
		init_cache_info(&ppc64_caches.l1i, 0x8000,   128,  128, 32);
		init_cache_info(&ppc64_caches.l1d, 0x10000,  128,  128, 64);
		init_cache_info(&ppc64_caches.l2,  0x80000,  128,  0,   512);
		init_cache_info(&ppc64_caches.l3,  0x800000, 128,  0,   8192);
	} else
		cpu = of_find_node_by_type(NULL, "cpu");
532

533 534 535 536
	/*
	 * We're assuming *all* of the CPUs have the same
	 * d-cache and i-cache sizes... -Peter
	 */
537 538
	if (cpu) {
		if (!parse_cache_info(cpu, false, &ppc64_caches.l1d))
539 540
			DBG("Argh, can't find dcache properties !\n");

541
		if (!parse_cache_info(cpu, true, &ppc64_caches.l1i))
542
			DBG("Argh, can't find icache properties !\n");
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558

		/*
		 * Try to find the L2 and L3 if any. Assume they are
		 * unified and use the D-side properties.
		 */
		l2 = of_find_next_cache_node(cpu);
		of_node_put(cpu);
		if (l2) {
			parse_cache_info(l2, false, &ppc64_caches.l2);
			l3 = of_find_next_cache_node(l2);
			of_node_put(l2);
		}
		if (l3) {
			parse_cache_info(l3, false, &ppc64_caches.l3);
			of_node_put(l3);
		}
559 560
	}

561
	/* For use by binfmt_elf */
562 563
	dcache_bsize = ppc64_caches.l1d.block_size;
	icache_bsize = ppc64_caches.l1i.block_size;
564

565 566 567
	cur_cpu_spec->dcache_bsize = dcache_bsize;
	cur_cpu_spec->icache_bsize = icache_bsize;

568 569 570
	DBG(" <- initialize_cache_info()\n");
}

571 572 573 574 575 576 577 578
/*
 * This returns the limit below which memory accesses to the linear
 * mapping are guarnateed not to cause an architectural exception (e.g.,
 * TLB or SLB miss fault).
 *
 * This is used to allocate PACAs and various interrupt stacks that
 * that are accessed early in interrupt handlers that must not cause
 * re-entrant interrupts.
579
 */
580
__init u64 ppc64_bolted_size(void)
581
{
582 583
#ifdef CONFIG_PPC_BOOK3E
	/* Freescale BookE bolts the entire linear mapping */
584 585
	/* XXX: BookE ppc64_rma_limit setup seems to disagree? */
	if (early_mmu_has_feature(MMU_FTR_TYPE_FSL_E))
586 587 588 589
		return linear_map_top;
	/* Other BookE, we assume the first GB is bolted */
	return 1ul << 30;
#else
590
	/* BookS radix, does not take faults on linear mapping */
591 592 593
	if (early_radix_enabled())
		return ULONG_MAX;

594 595
	/* BookS hash, the first segment is bolted */
	if (early_mmu_has_feature(MMU_FTR_1T_SEGMENT))
596 597
		return 1UL << SID_SHIFT_1T;
	return 1UL << SID_SHIFT;
598
#endif
599 600
}

601
void __init irqstack_early_init(void)
602
{
603
	u64 limit = ppc64_bolted_size();
604 605 606
	unsigned int i;

	/*
607
	 * Interrupt stacks must be in the first segment since we
608 609
	 * cannot afford to take SLB misses on them. They are not
	 * accessed in realmode.
610
	 */
611
	for_each_possible_cpu(i) {
612
		softirq_ctx[i] = (struct thread_info *)
Y
Yinghai Lu 已提交
613
			__va(memblock_alloc_base(THREAD_SIZE,
614
					    THREAD_SIZE, limit));
615
		hardirq_ctx[i] = (struct thread_info *)
Y
Yinghai Lu 已提交
616
			__va(memblock_alloc_base(THREAD_SIZE,
617
					    THREAD_SIZE, limit));
618 619 620
	}
}

621
#ifdef CONFIG_PPC_BOOK3E
622
void __init exc_lvl_early_init(void)
623 624
{
	unsigned int i;
625
	unsigned long sp;
626 627

	for_each_possible_cpu(i) {
628 629 630 631 632 633 634 635 636 637 638
		sp = memblock_alloc(THREAD_SIZE, THREAD_SIZE);
		critirq_ctx[i] = (struct thread_info *)__va(sp);
		paca[i].crit_kstack = __va(sp + THREAD_SIZE);

		sp = memblock_alloc(THREAD_SIZE, THREAD_SIZE);
		dbgirq_ctx[i] = (struct thread_info *)__va(sp);
		paca[i].dbg_kstack = __va(sp + THREAD_SIZE);

		sp = memblock_alloc(THREAD_SIZE, THREAD_SIZE);
		mcheckirq_ctx[i] = (struct thread_info *)__va(sp);
		paca[i].mc_kstack = __va(sp + THREAD_SIZE);
639
	}
640 641

	if (cpu_has_feature(CPU_FTR_DEBUG_LVL_EXC))
642
		patch_exception(0x040, exc_debug_debug_book3e);
643 644 645
}
#endif

646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663
/*
 * Emergency stacks are used for a range of things, from asynchronous
 * NMIs (system reset, machine check) to synchronous, process context.
 * We set preempt_count to zero, even though that isn't necessarily correct. To
 * get the right value we'd need to copy it from the previous thread_info, but
 * doing that might fault causing more problems.
 * TODO: what to do with accounting?
 */
static void emerg_stack_init_thread_info(struct thread_info *ti, int cpu)
{
	ti->task = NULL;
	ti->cpu = cpu;
	ti->preempt_count = 0;
	ti->local_flags = 0;
	ti->flags = 0;
	klp_init_thread_info(ti);
}

664 665
/*
 * Stack space used when we detect a bad kernel stack pointer, and
666 667
 * early in SMP boots before relocation is enabled. Exclusive emergency
 * stack for machine checks.
668
 */
669
void __init emergency_stack_init(void)
670
{
671
	u64 limit;
672 673 674 675 676 677 678 679
	unsigned int i;

	/*
	 * Emergency stacks must be under 256MB, we cannot afford to take
	 * SLB misses on them. The ABI also requires them to be 128-byte
	 * aligned.
	 *
	 * Since we use these as temporary stacks during secondary CPU
680 681 682
	 * bringup, machine check, system reset, and HMI, we need to get
	 * at them in real mode. This means they must also be within the RMO
	 * region.
683 684 685 686
	 *
	 * The IRQ stacks allocated elsewhere in this file are zeroed and
	 * initialized in kernel/irq.c. These are initialized here in order
	 * to have emergency stacks available as early as possible.
687
	 */
688
	limit = min(ppc64_bolted_size(), ppc64_rma_size);
689

690
	for_each_possible_cpu(i) {
691 692
		struct thread_info *ti;
		ti = __va(memblock_alloc_base(THREAD_SIZE, THREAD_SIZE, limit));
693 694
		memset(ti, 0, THREAD_SIZE);
		emerg_stack_init_thread_info(ti, i);
695
		paca[i].emergency_sp = (void *)ti + THREAD_SIZE;
696 697

#ifdef CONFIG_PPC_BOOK3S_64
698 699
		/* emergency stack for NMI exception handling. */
		ti = __va(memblock_alloc_base(THREAD_SIZE, THREAD_SIZE, limit));
700 701
		memset(ti, 0, THREAD_SIZE);
		emerg_stack_init_thread_info(ti, i);
702 703
		paca[i].nmi_emergency_sp = (void *)ti + THREAD_SIZE;

704
		/* emergency stack for machine check exception handling. */
705
		ti = __va(memblock_alloc_base(THREAD_SIZE, THREAD_SIZE, limit));
706 707
		memset(ti, 0, THREAD_SIZE);
		emerg_stack_init_thread_info(ti, i);
708
		paca[i].mc_emergency_sp = (void *)ti + THREAD_SIZE;
709
#endif
710
	}
711 712
}

713
#ifdef CONFIG_SMP
714 715 716
#define PCPU_DYN_SIZE		()

static void * __init pcpu_fc_alloc(unsigned int cpu, size_t size, size_t align)
717
{
718
	return __alloc_bootmem_node(NODE_DATA(early_cpu_to_node(cpu)), size, align,
719 720
				    __pa(MAX_DMA_ADDRESS));
}
721

722 723 724 725
static void __init pcpu_fc_free(void *ptr, size_t size)
{
	free_bootmem(__pa(ptr), size);
}
726

727 728
static int pcpu_cpu_distance(unsigned int from, unsigned int to)
{
729
	if (early_cpu_to_node(from) == early_cpu_to_node(to))
730 731 732 733 734
		return LOCAL_DISTANCE;
	else
		return REMOTE_DISTANCE;
}

735 736 737
unsigned long __per_cpu_offset[NR_CPUS] __read_mostly;
EXPORT_SYMBOL(__per_cpu_offset);

738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761
void __init setup_per_cpu_areas(void)
{
	const size_t dyn_size = PERCPU_MODULE_RESERVE + PERCPU_DYNAMIC_RESERVE;
	size_t atom_size;
	unsigned long delta;
	unsigned int cpu;
	int rc;

	/*
	 * Linear mapping is one of 4K, 1M and 16M.  For 4K, no need
	 * to group units.  For larger mappings, use 1M atom which
	 * should be large enough to contain a number of units.
	 */
	if (mmu_linear_psize == MMU_PAGE_4K)
		atom_size = PAGE_SIZE;
	else
		atom_size = 1 << 20;

	rc = pcpu_embed_first_chunk(0, dyn_size, atom_size, pcpu_cpu_distance,
				    pcpu_fc_alloc, pcpu_fc_free);
	if (rc < 0)
		panic("cannot initialize percpu area (err=%d)", rc);

	delta = (unsigned long)pcpu_base_addr - (unsigned long)__per_cpu_start;
762 763 764 765
	for_each_possible_cpu(cpu) {
                __per_cpu_offset[cpu] = delta + pcpu_unit_offsets[cpu];
		paca[cpu].data_offset = __per_cpu_offset[cpu];
	}
766 767
}
#endif
768

769 770 771 772 773 774 775 776 777
#ifdef CONFIG_MEMORY_HOTPLUG_SPARSE
unsigned long memory_block_size_bytes(void)
{
	if (ppc_md.memory_block_size)
		return ppc_md.memory_block_size();

	return MIN_MEMORY_BLOCK_SIZE;
}
#endif
778

779
#if defined(CONFIG_PPC_INDIRECT_PIO) || defined(CONFIG_PPC_INDIRECT_MMIO)
780 781
struct ppc_pci_io ppc_pci_io;
EXPORT_SYMBOL(ppc_pci_io);
782
#endif
783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810

#ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF
u64 hw_nmi_get_sample_period(int watchdog_thresh)
{
	return ppc_proc_freq * watchdog_thresh;
}
#endif

/*
 * The perf based hardlockup detector breaks PMU event based branches, so
 * disable it by default. Book3S has a soft-nmi hardlockup detector based
 * on the decrementer interrupt, so it does not suffer from this problem.
 *
 * It is likely to get false positives in VM guests, so disable it there
 * by default too.
 */
static int __init disable_hardlockup_detector(void)
{
#ifdef CONFIG_HARDLOCKUP_DETECTOR_PERF
	hardlockup_detector_disable();
#else
	if (firmware_has_feature(FW_FEATURE_LPAR))
		hardlockup_detector_disable();
#endif

	return 0;
}
early_initcall(disable_hardlockup_detector);