tlb_nohash.c 18.0 KB
Newer Older
1 2 3 4 5 6 7 8 9
/*
 * This file contains the routines for TLB flushing.
 * On machines where the MMU does not use a hash table to store virtual to
 * physical translations (ie, SW loaded TLBs or Book3E compilant processors,
 * this does -not- include 603 however which shares the implementation with
 * hash based processors)
 *
 *  -- BenH
 *
10 11
 * Copyright 2008,2009 Ben Herrenschmidt <benh@kernel.crashing.org>
 *                     IBM Corp.
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
 *
 *  Derived from arch/ppc/mm/init.c:
 *    Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
 *
 *  Modifications by Paul Mackerras (PowerMac) (paulus@cs.anu.edu.au)
 *  and Cort Dougan (PReP) (cort@cs.nmt.edu)
 *    Copyright (C) 1996 Paul Mackerras
 *
 *  Derived from "arch/i386/mm/init.c"
 *    Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 *
 *  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.
 *
 */

#include <linux/kernel.h>
31
#include <linux/export.h>
32 33 34 35 36 37
#include <linux/mm.h>
#include <linux/init.h>
#include <linux/highmem.h>
#include <linux/pagemap.h>
#include <linux/preempt.h>
#include <linux/spinlock.h>
Y
Yinghai Lu 已提交
38
#include <linux/memblock.h>
39
#include <linux/of_fdt.h>
B
Becky Bruce 已提交
40
#include <linux/hugetlb.h>
41 42 43

#include <asm/tlbflush.h>
#include <asm/tlb.h>
44
#include <asm/code-patching.h>
B
Becky Bruce 已提交
45
#include <asm/hugetlb.h>
46
#include <asm/paca.h>
47 48 49

#include "mmu_decl.h"

B
Becky Bruce 已提交
50 51 52 53 54 55
/*
 * This struct lists the sw-supported page sizes.  The hardawre MMU may support
 * other sizes not listed here.   The .ind field is only used on MMUs that have
 * indirect page table entries.
 */
#ifdef CONFIG_PPC_BOOK3E_MMU
56
#ifdef CONFIG_PPC_FSL_BOOK3E
B
Becky Bruce 已提交
57 58 59 60 61
struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT] = {
	[MMU_PAGE_4K] = {
		.shift	= 12,
		.enc	= BOOK3E_PAGESZ_4K,
	},
62 63 64 65
	[MMU_PAGE_2M] = {
		.shift	= 21,
		.enc	= BOOK3E_PAGESZ_2M,
	},
B
Becky Bruce 已提交
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
	[MMU_PAGE_4M] = {
		.shift	= 22,
		.enc	= BOOK3E_PAGESZ_4M,
	},
	[MMU_PAGE_16M] = {
		.shift	= 24,
		.enc	= BOOK3E_PAGESZ_16M,
	},
	[MMU_PAGE_64M] = {
		.shift	= 26,
		.enc	= BOOK3E_PAGESZ_64M,
	},
	[MMU_PAGE_256M] = {
		.shift	= 28,
		.enc	= BOOK3E_PAGESZ_256M,
	},
	[MMU_PAGE_1G] = {
		.shift	= 30,
		.enc	= BOOK3E_PAGESZ_1GB,
	},
};
#else
88 89 90
struct mmu_psize_def mmu_psize_defs[MMU_PAGE_COUNT] = {
	[MMU_PAGE_4K] = {
		.shift	= 12,
91
		.ind	= 20,
92 93 94 95 96 97 98 99
		.enc	= BOOK3E_PAGESZ_4K,
	},
	[MMU_PAGE_16K] = {
		.shift	= 14,
		.enc	= BOOK3E_PAGESZ_16K,
	},
	[MMU_PAGE_64K] = {
		.shift	= 16,
100
		.ind	= 28,
101 102 103 104 105 106 107 108
		.enc	= BOOK3E_PAGESZ_64K,
	},
	[MMU_PAGE_1M] = {
		.shift	= 20,
		.enc	= BOOK3E_PAGESZ_1M,
	},
	[MMU_PAGE_16M] = {
		.shift	= 24,
109
		.ind	= 36,
110 111 112 113 114 115 116 117 118 119 120
		.enc	= BOOK3E_PAGESZ_16M,
	},
	[MMU_PAGE_256M] = {
		.shift	= 28,
		.enc	= BOOK3E_PAGESZ_256M,
	},
	[MMU_PAGE_1G] = {
		.shift	= 30,
		.enc	= BOOK3E_PAGESZ_1GB,
	},
};
B
Becky Bruce 已提交
121 122
#endif /* CONFIG_FSL_BOOKE */

123 124 125 126 127 128 129 130 131 132
static inline int mmu_get_tsize(int psize)
{
	return mmu_psize_defs[psize].enc;
}
#else
static inline int mmu_get_tsize(int psize)
{
	/* This isn't used on !Book3E for now */
	return 0;
}
B
Becky Bruce 已提交
133
#endif /* CONFIG_PPC_BOOK3E_MMU */
134 135 136 137 138 139 140 141 142

/* The variables below are currently only used on 64-bit Book3E
 * though this will probably be made common with other nohash
 * implementations at some point
 */
#ifdef CONFIG_PPC64

int mmu_linear_psize;		/* Page size used for the linear mapping */
int mmu_pte_psize;		/* Page size used for PTE pages */
143
int mmu_vmemmap_psize;		/* Page size used for the virtual mem map */
144
int book3e_htw_mode;		/* HW tablewalk?  Value is PPC_HTW_* */
145 146
unsigned long linear_map_top;	/* Top of linear mapping */

147 148 149 150 151 152 153 154 155

/*
 * Number of bytes to add to SPRN_SPRG_TLB_EXFRAME on crit/mcheck/debug
 * exceptions.  This is used for bolted and e6500 TLB miss handlers which
 * do not modify this SPRG in the TLB miss code; for other TLB miss handlers,
 * this is set to zero.
 */
int extlb_level_exc;

156 157
#endif /* CONFIG_PPC64 */

158 159 160 161 162 163
#ifdef CONFIG_PPC_FSL_BOOK3E
/* next_tlbcam_idx is used to round-robin tlbcam entry assignment */
DEFINE_PER_CPU(int, next_tlbcam_idx);
EXPORT_PER_CPU_SYMBOL(next_tlbcam_idx);
#endif

164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190
/*
 * Base TLB flushing operations:
 *
 *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
 *  - flush_tlb_page(vma, vmaddr) flushes one page
 *  - flush_tlb_range(vma, start, end) flushes a range of pages
 *  - flush_tlb_kernel_range(start, end) flushes kernel pages
 *
 *  - local_* variants of page and mm only apply to the current
 *    processor
 */

/*
 * These are the base non-SMP variants of page and mm flushing
 */
void local_flush_tlb_mm(struct mm_struct *mm)
{
	unsigned int pid;

	preempt_disable();
	pid = mm->context.id;
	if (pid != MMU_NO_CONTEXT)
		_tlbil_pid(pid);
	preempt_enable();
}
EXPORT_SYMBOL(local_flush_tlb_mm);

191 192
void __local_flush_tlb_page(struct mm_struct *mm, unsigned long vmaddr,
			    int tsize, int ind)
193 194 195 196
{
	unsigned int pid;

	preempt_disable();
197
	pid = mm ? mm->context.id : 0;
198
	if (pid != MMU_NO_CONTEXT)
199
		_tlbil_va(vmaddr, pid, tsize, ind);
200 201 202
	preempt_enable();
}

203 204 205
void local_flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr)
{
	__local_flush_tlb_page(vma ? vma->vm_mm : NULL, vmaddr,
206
			       mmu_get_tsize(mmu_virtual_psize), 0);
207 208
}
EXPORT_SYMBOL(local_flush_tlb_page);
209 210 211 212 213 214

/*
 * And here are the SMP non-local implementations
 */
#ifdef CONFIG_SMP

215
static DEFINE_RAW_SPINLOCK(tlbivax_lock);
216

217 218 219 220 221 222
static int mm_is_core_local(struct mm_struct *mm)
{
	return cpumask_subset(mm_cpumask(mm),
			      topology_thread_cpumask(smp_processor_id()));
}

223 224 225
struct tlb_flush_param {
	unsigned long addr;
	unsigned int pid;
226 227
	unsigned int tsize;
	unsigned int ind;
228 229 230 231 232 233 234 235 236 237 238 239 240
};

static void do_flush_tlb_mm_ipi(void *param)
{
	struct tlb_flush_param *p = param;

	_tlbil_pid(p ? p->pid : 0);
}

static void do_flush_tlb_page_ipi(void *param)
{
	struct tlb_flush_param *p = param;

241
	_tlbil_va(p->addr, p->pid, p->tsize, p->ind);
242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268
}


/* Note on invalidations and PID:
 *
 * We snapshot the PID with preempt disabled. At this point, it can still
 * change either because:
 * - our context is being stolen (PID -> NO_CONTEXT) on another CPU
 * - we are invaliating some target that isn't currently running here
 *   and is concurrently acquiring a new PID on another CPU
 * - some other CPU is re-acquiring a lost PID for this mm
 * etc...
 *
 * However, this shouldn't be a problem as we only guarantee
 * invalidation of TLB entries present prior to this call, so we
 * don't care about the PID changing, and invalidating a stale PID
 * is generally harmless.
 */

void flush_tlb_mm(struct mm_struct *mm)
{
	unsigned int pid;

	preempt_disable();
	pid = mm->context.id;
	if (unlikely(pid == MMU_NO_CONTEXT))
		goto no_context;
269
	if (!mm_is_core_local(mm)) {
270
		struct tlb_flush_param p = { .pid = pid };
271 272 273
		/* Ignores smp_processor_id() even if set. */
		smp_call_function_many(mm_cpumask(mm),
				       do_flush_tlb_mm_ipi, &p, 1);
274 275 276 277 278 279 280
	}
	_tlbil_pid(pid);
 no_context:
	preempt_enable();
}
EXPORT_SYMBOL(flush_tlb_mm);

281 282
void __flush_tlb_page(struct mm_struct *mm, unsigned long vmaddr,
		      int tsize, int ind)
283
{
284
	struct cpumask *cpu_mask;
285 286 287
	unsigned int pid;

	preempt_disable();
288
	pid = mm ? mm->context.id : 0;
289 290
	if (unlikely(pid == MMU_NO_CONTEXT))
		goto bail;
291
	cpu_mask = mm_cpumask(mm);
292
	if (!mm_is_core_local(mm)) {
293 294 295 296
		/* If broadcast tlbivax is supported, use it */
		if (mmu_has_feature(MMU_FTR_USE_TLBIVAX_BCAST)) {
			int lock = mmu_has_feature(MMU_FTR_LOCK_BCAST_INVAL);
			if (lock)
297
				raw_spin_lock(&tlbivax_lock);
298
			_tlbivax_bcast(vmaddr, pid, tsize, ind);
299
			if (lock)
300
				raw_spin_unlock(&tlbivax_lock);
301 302
			goto bail;
		} else {
303 304 305 306 307 308
			struct tlb_flush_param p = {
				.pid = pid,
				.addr = vmaddr,
				.tsize = tsize,
				.ind = ind,
			};
309 310
			/* Ignores smp_processor_id() even if set in cpu_mask */
			smp_call_function_many(cpu_mask,
311 312 313
					       do_flush_tlb_page_ipi, &p, 1);
		}
	}
314
	_tlbil_va(vmaddr, pid, tsize, ind);
315 316 317
 bail:
	preempt_enable();
}
318 319 320

void flush_tlb_page(struct vm_area_struct *vma, unsigned long vmaddr)
{
B
Becky Bruce 已提交
321
#ifdef CONFIG_HUGETLB_PAGE
322
	if (vma && is_vm_hugetlb_page(vma))
B
Becky Bruce 已提交
323 324 325
		flush_hugetlb_page(vma, vmaddr);
#endif

326
	__flush_tlb_page(vma ? vma->vm_mm : NULL, vmaddr,
327
			 mmu_get_tsize(mmu_virtual_psize), 0);
328
}
329 330 331 332
EXPORT_SYMBOL(flush_tlb_page);

#endif /* CONFIG_SMP */

333 334 335 336 337 338 339 340 341 342 343
#ifdef CONFIG_PPC_47x
void __init early_init_mmu_47x(void)
{
#ifdef CONFIG_SMP
	unsigned long root = of_get_flat_dt_root();
	if (of_get_flat_dt_prop(root, "cooperative-partition", NULL))
		mmu_clear_feature(MMU_FTR_USE_TLBIVAX_BCAST);
#endif /* CONFIG_SMP */
}
#endif /* CONFIG_PPC_47x */

344 345 346 347 348 349 350 351 352 353
/*
 * Flush kernel TLB entries in the given range
 */
void flush_tlb_kernel_range(unsigned long start, unsigned long end)
{
#ifdef CONFIG_SMP
	preempt_disable();
	smp_call_function(do_flush_tlb_mm_ipi, NULL, 1);
	_tlbil_pid(0);
	preempt_enable();
354
#else
355
	_tlbil_pid(0);
356
#endif
357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372
}
EXPORT_SYMBOL(flush_tlb_kernel_range);

/*
 * Currently, for range flushing, we just do a full mm flush. This should
 * be optimized based on a threshold on the size of the range, since
 * some implementation can stack multiple tlbivax before a tlbsync but
 * for now, we keep it that way
 */
void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
		     unsigned long end)

{
	flush_tlb_mm(vma->vm_mm);
}
EXPORT_SYMBOL(flush_tlb_range);
373 374 375 376 377

void tlb_flush(struct mmu_gather *tlb)
{
	flush_tlb_mm(tlb->mm);
}
378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393

/*
 * Below are functions specific to the 64-bit variant of Book3E though that
 * may change in the future
 */

#ifdef CONFIG_PPC64

/*
 * Handling of virtual linear page tables or indirect TLB entries
 * flushing when PTE pages are freed
 */
void tlb_flush_pgtable(struct mmu_gather *tlb, unsigned long address)
{
	int tsize = mmu_psize_defs[mmu_pte_psize].enc;

394
	if (book3e_htw_mode != PPC_HTW_NONE) {
395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421
		unsigned long start = address & PMD_MASK;
		unsigned long end = address + PMD_SIZE;
		unsigned long size = 1UL << mmu_psize_defs[mmu_pte_psize].shift;

		/* This isn't the most optimal, ideally we would factor out the
		 * while preempt & CPU mask mucking around, or even the IPI but
		 * it will do for now
		 */
		while (start < end) {
			__flush_tlb_page(tlb->mm, start, tsize, 1);
			start += size;
		}
	} else {
		unsigned long rmask = 0xf000000000000000ul;
		unsigned long rid = (address & rmask) | 0x1000000000000000ul;
		unsigned long vpte = address & ~rmask;

#ifdef CONFIG_PPC_64K_PAGES
		vpte = (vpte >> (PAGE_SHIFT - 4)) & ~0xfffful;
#else
		vpte = (vpte >> (PAGE_SHIFT - 3)) & ~0xffful;
#endif
		vpte |= rid;
		__flush_tlb_page(tlb->mm, vpte, tsize, 0);
	}
}

422 423
static void setup_page_sizes(void)
{
424 425 426
	unsigned int tlb0cfg;
	unsigned int tlb0ps;
	unsigned int eptcfg;
427 428
	int i, psize;

429 430
#ifdef CONFIG_PPC_FSL_BOOK3E
	unsigned int mmucfg = mfspr(SPRN_MMUCFG);
431
	int fsl_mmu = mmu_has_feature(MMU_FTR_TYPE_FSL_E);
432

433
	if (fsl_mmu && (mmucfg & MMUCFG_MAVN) == MMUCFG_MAVN_V1) {
434 435 436 437 438 439 440 441 442 443 444 445 446
		unsigned int tlb1cfg = mfspr(SPRN_TLB1CFG);
		unsigned int min_pg, max_pg;

		min_pg = (tlb1cfg & TLBnCFG_MINSIZE) >> TLBnCFG_MINSIZE_SHIFT;
		max_pg = (tlb1cfg & TLBnCFG_MAXSIZE) >> TLBnCFG_MAXSIZE_SHIFT;

		for (psize = 0; psize < MMU_PAGE_COUNT; ++psize) {
			struct mmu_psize_def *def;
			unsigned int shift;

			def = &mmu_psize_defs[psize];
			shift = def->shift;

447
			if (shift == 0 || shift & 1)
448 449 450 451 452 453 454 455 456
				continue;

			/* adjust to be in terms of 4^shift Kb */
			shift = (shift - 10) >> 1;

			if ((shift >= min_pg) && (shift <= max_pg))
				def->flags |= MMU_PAGE_SIZE_DIRECT;
		}

457
		goto out;
458
	}
459 460

	if (fsl_mmu && (mmucfg & MMUCFG_MAVN) == MMUCFG_MAVN_V2) {
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
		u32 tlb1cfg, tlb1ps;

		tlb0cfg = mfspr(SPRN_TLB0CFG);
		tlb1cfg = mfspr(SPRN_TLB1CFG);
		tlb1ps = mfspr(SPRN_TLB1PS);
		eptcfg = mfspr(SPRN_EPTCFG);

		if ((tlb1cfg & TLBnCFG_IND) && (tlb0cfg & TLBnCFG_PT))
			book3e_htw_mode = PPC_HTW_E6500;

		/*
		 * We expect 4K subpage size and unrestricted indirect size.
		 * The lack of a restriction on indirect size is a Freescale
		 * extension, indicated by PSn = 0 but SPSn != 0.
		 */
		if (eptcfg != 2)
			book3e_htw_mode = PPC_HTW_NONE;
478 479 480 481 482 483

		for (psize = 0; psize < MMU_PAGE_COUNT; ++psize) {
			struct mmu_psize_def *def = &mmu_psize_defs[psize];

			if (tlb1ps & (1U << (def->shift - 10))) {
				def->flags |= MMU_PAGE_SIZE_DIRECT;
484 485 486

				if (book3e_htw_mode && psize == MMU_PAGE_2M)
					def->flags |= MMU_PAGE_SIZE_INDIRECT;
487 488 489
			}
		}

490
		goto out;
491
	}
492 493 494 495 496 497
#endif

	tlb0cfg = mfspr(SPRN_TLB0CFG);
	tlb0ps = mfspr(SPRN_TLB0PS);
	eptcfg = mfspr(SPRN_EPTCFG);

498 499 500 501 502 503 504 505 506
	/* Look for supported direct sizes */
	for (psize = 0; psize < MMU_PAGE_COUNT; ++psize) {
		struct mmu_psize_def *def = &mmu_psize_defs[psize];

		if (tlb0ps & (1U << (def->shift - 10)))
			def->flags |= MMU_PAGE_SIZE_DIRECT;
	}

	/* Indirect page sizes supported ? */
507 508 509 510 511
	if ((tlb0cfg & TLBnCFG_IND) == 0 ||
	    (tlb0cfg & TLBnCFG_PT) == 0)
		goto out;

	book3e_htw_mode = PPC_HTW_IBM;
512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536

	/* Now, we only deal with one IND page size for each
	 * direct size. Hopefully all implementations today are
	 * unambiguous, but we might want to be careful in the
	 * future.
	 */
	for (i = 0; i < 3; i++) {
		unsigned int ps, sps;

		sps = eptcfg & 0x1f;
		eptcfg >>= 5;
		ps = eptcfg & 0x1f;
		eptcfg >>= 5;
		if (!ps || !sps)
			continue;
		for (psize = 0; psize < MMU_PAGE_COUNT; psize++) {
			struct mmu_psize_def *def = &mmu_psize_defs[psize];

			if (ps == (def->shift - 10))
				def->flags |= MMU_PAGE_SIZE_INDIRECT;
			if (sps == (def->shift - 10))
				def->ind = ps + 10;
		}
	}

537
out:
538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556
	/* Cleanup array and print summary */
	pr_info("MMU: Supported page sizes\n");
	for (psize = 0; psize < MMU_PAGE_COUNT; ++psize) {
		struct mmu_psize_def *def = &mmu_psize_defs[psize];
		const char *__page_type_names[] = {
			"unsupported",
			"direct",
			"indirect",
			"direct & indirect"
		};
		if (def->flags == 0) {
			def->shift = 0;	
			continue;
		}
		pr_info("  %8ld KB as %s\n", 1ul << (def->shift - 10),
			__page_type_names[def->flags & 0x3]);
	}
}

557 558
static void setup_mmu_htw(void)
{
559 560 561 562 563 564 565
	/*
	 * If we want to use HW tablewalk, enable it by patching the TLB miss
	 * handlers to branch to the one dedicated to it.
	 */

	switch (book3e_htw_mode) {
	case PPC_HTW_IBM:
566 567
		patch_exception(0x1c0, exc_data_tlb_miss_htw_book3e);
		patch_exception(0x1e0, exc_instruction_tlb_miss_htw_book3e);
568
		break;
569
#ifdef CONFIG_PPC_FSL_BOOK3E
570
	case PPC_HTW_E6500:
571
		extlb_level_exc = EX_TLB_SIZE;
572 573 574
		patch_exception(0x1c0, exc_data_tlb_miss_e6500_book3e);
		patch_exception(0x1e0, exc_instruction_tlb_miss_e6500_book3e);
		break;
575
#endif
576
	}
577
	pr_info("MMU: Book3E HW tablewalk %s\n",
578
		book3e_htw_mode != PPC_HTW_NONE ? "enabled" : "not supported");
579 580 581 582 583 584 585
}

/*
 * Early initialization of the MMU TLB code
 */
static void __early_init_mmu(int boot_cpu)
{
586 587 588
	unsigned int mas4;

	/* XXX This will have to be decided at runtime, but right
589 590 591 592
	 * now our boot and TLB miss code hard wires it. Ideally
	 * we should find out a suitable page size and patch the
	 * TLB miss code (either that or use the PACA to store
	 * the value we want)
593 594 595
	 */
	mmu_linear_psize = MMU_PAGE_1G;

596 597 598 599 600
	/* XXX This should be decided at runtime based on supported
	 * page sizes in the TLB, but for now let's assume 16M is
	 * always there and a good fit (which it probably is)
	 */
	mmu_vmemmap_psize = MMU_PAGE_16M;
601 602 603 604 605 606 607

	/* XXX This code only checks for TLB 0 capabilities and doesn't
	 *     check what page size combos are supported by the HW. It
	 *     also doesn't handle the case where a separate array holds
	 *     the IND entries from the array loaded by the PT.
	 */
	if (boot_cpu) {
608 609
		/* Look for supported page sizes */
		setup_page_sizes();
610

611 612
		/* Look for HW tablewalk support */
		setup_mmu_htw();
613 614 615 616 617
	}

	/* Set MAS4 based on page table setting */

	mas4 = 0x4 << MAS4_WIMGED_SHIFT;
618 619 620 621 622 623 624 625 626 627
	switch (book3e_htw_mode) {
	case PPC_HTW_E6500:
		mas4 |= MAS4_INDD;
		mas4 |= BOOK3E_PAGESZ_2M << MAS4_TSIZED_SHIFT;
		mas4 |= MAS4_TLBSELD(1);
		mmu_pte_psize = MMU_PAGE_2M;
		break;

	case PPC_HTW_IBM:
		mas4 |= MAS4_INDD;
628 629 630 631 632 633 634
#ifdef CONFIG_PPC_64K_PAGES
		mas4 |=	BOOK3E_PAGESZ_256M << MAS4_TSIZED_SHIFT;
		mmu_pte_psize = MMU_PAGE_256M;
#else
		mas4 |=	BOOK3E_PAGESZ_1M << MAS4_TSIZED_SHIFT;
		mmu_pte_psize = MMU_PAGE_1M;
#endif
635 636 637
		break;

	case PPC_HTW_NONE:
638 639 640 641 642 643
#ifdef CONFIG_PPC_64K_PAGES
		mas4 |=	BOOK3E_PAGESZ_64K << MAS4_TSIZED_SHIFT;
#else
		mas4 |=	BOOK3E_PAGESZ_4K << MAS4_TSIZED_SHIFT;
#endif
		mmu_pte_psize = mmu_virtual_psize;
644
		break;
645 646 647 648 649 650
	}
	mtspr(SPRN_MAS4, mas4);

	/* Set the global containing the top of the linear mapping
	 * for use by the TLB miss code
	 */
Y
Yinghai Lu 已提交
651
	linear_map_top = memblock_end_of_DRAM();
652

653 654 655 656 657 658 659 660 661 662
#ifdef CONFIG_PPC_FSL_BOOK3E
	if (mmu_has_feature(MMU_FTR_TYPE_FSL_E)) {
		unsigned int num_cams;

		/* use a quarter of the TLBCAM for bolted linear map */
		num_cams = (mfspr(SPRN_TLB1CFG) & TLBnCFG_N_ENTRY) / 4;
		linear_map_top = map_mem_in_cams(linear_map_top, num_cams);

		/* limit memory so we dont have linear faults */
		memblock_enforce_memory_limit(linear_map_top);
663

664
		if (book3e_htw_mode == PPC_HTW_NONE) {
665
			extlb_level_exc = EX_TLB_SIZE;
666 667 668 669
			patch_exception(0x1c0, exc_data_tlb_miss_bolted_book3e);
			patch_exception(0x1e0,
				exc_instruction_tlb_miss_bolted_book3e);
		}
670 671 672
	}
#endif

673 674 675 676
	/* A sync won't hurt us after mucking around with
	 * the MMU configuration
	 */
	mb();
677 678

	memblock_set_current_limit(linear_map_top);
679 680 681 682 683 684 685
}

void __init early_init_mmu(void)
{
	__early_init_mmu(1);
}

686
void early_init_mmu_secondary(void)
687 688 689 690
{
	__early_init_mmu(0);
}

691 692 693
void setup_initial_memory_limit(phys_addr_t first_memblock_base,
				phys_addr_t first_memblock_size)
{
694
	/* On non-FSL Embedded 64-bit, we adjust the RMA size to match
695 696
	 * the bolted TLB entry. We know for now that only 1G
	 * entries are supported though that may eventually
697 698 699 700 701 702 703 704
	 * change.
	 *
	 * on FSL Embedded 64-bit, we adjust the RMA size to match the
	 * first bolted TLB entry size.  We still limit max to 1G even if
	 * the TLB could cover more.  This is due to what the early init
	 * code is setup to do.
	 *
	 * We crop it to the size of the first MEMBLOCK to
705 706
	 * avoid going over total available memory just in case...
	 */
707 708 709 710 711 712 713 714 715
#ifdef CONFIG_PPC_FSL_BOOK3E
	if (mmu_has_feature(MMU_FTR_TYPE_FSL_E)) {
		unsigned long linear_sz;
		linear_sz = calc_cam_sz(first_memblock_size, PAGE_OFFSET,
					first_memblock_base);
		ppc64_rma_size = min_t(u64, linear_sz, 0x40000000);
	} else
#endif
		ppc64_rma_size = min_t(u64, first_memblock_size, 0x40000000);
716 717

	/* Finally limit subsequent allocations */
718
	memblock_set_current_limit(first_memblock_base + ppc64_rma_size);
719
}
720 721 722 723 724 725 726
#else /* ! CONFIG_PPC64 */
void __init early_init_mmu(void)
{
#ifdef CONFIG_PPC_47x
	early_init_mmu_47x();
#endif
}
727
#endif /* CONFIG_PPC64 */