[mirror_ubuntu-eoan-kernel.git] / arch / powerpc / mm / init_64.c

// SPDX-License-Identifier: GPL-2.0-or-later
/*
 *  PowerPC version
 *    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
 *
 *  Dave Engebretsen <engebret@us.ibm.com>
 *      Rework for PPC64 port.
 */

#undef DEBUG

#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/types.h>
#include <linux/mman.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/stddef.h>
#include <linux/vmalloc.h>
#include <linux/init.h>
#include <linux/delay.h>
#include <linux/highmem.h>
#include <linux/idr.h>
#include <linux/nodemask.h>
#include <linux/module.h>
#include <linux/poison.h>
#include <linux/memblock.h>
#include <linux/hugetlb.h>
#include <linux/slab.h>
#include <linux/of_fdt.h>
#include <linux/libfdt.h>
#include <linux/memremap.h>

#include <asm/pgalloc.h>
#include <asm/page.h>
#include <asm/prom.h>
#include <asm/rtas.h>
#include <asm/io.h>
#include <asm/mmu_context.h>
#include <asm/pgtable.h>
#include <asm/mmu.h>
#include <linux/uaccess.h>
#include <asm/smp.h>
#include <asm/machdep.h>
#include <asm/tlb.h>
#include <asm/eeh.h>
#include <asm/processor.h>
#include <asm/mmzone.h>
#include <asm/cputable.h>
#include <asm/sections.h>
#include <asm/iommu.h>
#include <asm/vdso.h>

#include <mm/mmu_decl.h>

phys_addr_t memstart_addr = ~0;
EXPORT_SYMBOL_GPL(memstart_addr);
phys_addr_t kernstart_addr;
EXPORT_SYMBOL_GPL(kernstart_addr);

#ifdef CONFIG_SPARSEMEM_VMEMMAP
/*
 * Given an address within the vmemmap, determine the pfn of the page that
 * represents the start of the section it is within.  Note that we have to
 * do this by hand as the proffered address may not be correctly aligned.
 * Subtraction of non-aligned pointers produces undefined results.
 */
static unsigned long __meminit vmemmap_section_start(unsigned long page)
{
	unsigned long offset = page - ((unsigned long)(vmemmap));

	/* Return the pfn of the start of the section. */
	return (offset / sizeof(struct page)) & PAGE_SECTION_MASK;
}

/*
 * Check if this vmemmap page is already initialised.  If any section
 * which overlaps this vmemmap page is initialised then this page is
 * initialised already.
 */
static int __meminit vmemmap_populated(unsigned long start, int page_size)
{
	unsigned long end = start + page_size;
	start = (unsigned long)(pfn_to_page(vmemmap_section_start(start)));

	for (; start < end; start += (PAGES_PER_SECTION * sizeof(struct page)))
		if (pfn_valid(page_to_pfn((struct page *)start)))
			return 1;

	return 0;
}

/*
 * vmemmap virtual address space management does not have a traditonal page
 * table to track which virtual struct pages are backed by physical mapping.
 * The virtual to physical mappings are tracked in a simple linked list
 * format. 'vmemmap_list' maintains the entire vmemmap physical mapping at
 * all times where as the 'next' list maintains the available
 * vmemmap_backing structures which have been deleted from the
 * 'vmemmap_global' list during system runtime (memory hotplug remove
 * operation). The freed 'vmemmap_backing' structures are reused later when
 * new requests come in without allocating fresh memory. This pointer also
 * tracks the allocated 'vmemmap_backing' structures as we allocate one
 * full page memory at a time when we dont have any.
 */
struct vmemmap_backing *vmemmap_list;
static struct vmemmap_backing *next;

/*
 * The same pointer 'next' tracks individual chunks inside the allocated
 * full page during the boot time and again tracks the freeed nodes during
 * runtime. It is racy but it does not happen as they are separated by the
 * boot process. Will create problem if some how we have memory hotplug
 * operation during boot !!
 */
static int num_left;
static int num_freed;

static __meminit struct vmemmap_backing * vmemmap_list_alloc(int node)
{
	struct vmemmap_backing *vmem_back;
	/* get from freed entries first */
	if (num_freed) {
		num_freed--;
		vmem_back = next;
		next = next->list;

		return vmem_back;
	}

	/* allocate a page when required and hand out chunks */
	if (!num_left) {
		next = vmemmap_alloc_block(PAGE_SIZE, node);
		if (unlikely(!next)) {
			WARN_ON(1);
			return NULL;
		}
		num_left = PAGE_SIZE / sizeof(struct vmemmap_backing);
	}

	num_left--;

	return next++;
}

static __meminit void vmemmap_list_populate(unsigned long phys,
					    unsigned long start,
					    int node)
{
	struct vmemmap_backing *vmem_back;

	vmem_back = vmemmap_list_alloc(node);
	if (unlikely(!vmem_back)) {
		WARN_ON(1);
		return;
	}

	vmem_back->phys = phys;
	vmem_back->virt_addr = start;
	vmem_back->list = vmemmap_list;

	vmemmap_list = vmem_back;
}

int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node,
		struct vmem_altmap *altmap)
{
	unsigned long page_size = 1 << mmu_psize_defs[mmu_vmemmap_psize].shift;

	/* Align to the page size of the linear mapping. */
	start = _ALIGN_DOWN(start, page_size);

	pr_debug("vmemmap_populate %lx..%lx, node %d\n", start, end, node);

	for (; start < end; start += page_size) {
		void *p = NULL;
		int rc;

		if (vmemmap_populated(start, page_size))
			continue;

		/*
		 * Allocate from the altmap first if we have one. This may
		 * fail due to alignment issues when using 16MB hugepages, so
		 * fall back to system memory if the altmap allocation fail.
		 */
		if (altmap)
			p = altmap_alloc_block_buf(page_size, altmap);
		if (!p)
			p = vmemmap_alloc_block_buf(page_size, node);
		if (!p)
			return -ENOMEM;

		vmemmap_list_populate(__pa(p), start, node);

		pr_debug("      * %016lx..%016lx allocated at %p\n",
			 start, start + page_size, p);

		rc = vmemmap_create_mapping(start, page_size, __pa(p));
		if (rc < 0) {
			pr_warn("%s: Unable to create vmemmap mapping: %d\n",
				__func__, rc);
			return -EFAULT;
		}
	}

	return 0;
}

#ifdef CONFIG_MEMORY_HOTPLUG
static unsigned long vmemmap_list_free(unsigned long start)
{
	struct vmemmap_backing *vmem_back, *vmem_back_prev;

	vmem_back_prev = vmem_back = vmemmap_list;

	/* look for it with prev pointer recorded */
	for (; vmem_back; vmem_back = vmem_back->list) {
		if (vmem_back->virt_addr == start)
			break;
		vmem_back_prev = vmem_back;
	}

	if (unlikely(!vmem_back)) {
		WARN_ON(1);
		return 0;
	}

	/* remove it from vmemmap_list */
	if (vmem_back == vmemmap_list) /* remove head */
		vmemmap_list = vmem_back->list;
	else
		vmem_back_prev->list = vmem_back->list;

	/* next point to this freed entry */
	vmem_back->list = next;
	next = vmem_back;
	num_freed++;

	return vmem_back->phys;
}

void __ref vmemmap_free(unsigned long start, unsigned long end,
		struct vmem_altmap *altmap)
{
	unsigned long page_size = 1 << mmu_psize_defs[mmu_vmemmap_psize].shift;
	unsigned long page_order = get_order(page_size);
	unsigned long alt_start = ~0, alt_end = ~0;
	unsigned long base_pfn;

	start = _ALIGN_DOWN(start, page_size);
	if (altmap) {
		alt_start = altmap->base_pfn;
		alt_end = altmap->base_pfn + altmap->reserve +
			  altmap->free + altmap->alloc + altmap->align;
	}

	pr_debug("vmemmap_free %lx...%lx\n", start, end);

	for (; start < end; start += page_size) {
		unsigned long nr_pages, addr;
		struct page *page;

		/*
		 * the section has already be marked as invalid, so
		 * vmemmap_populated() true means some other sections still
		 * in this page, so skip it.
		 */
		if (vmemmap_populated(start, page_size))
			continue;

		addr = vmemmap_list_free(start);
		if (!addr)
			continue;

		page = pfn_to_page(addr >> PAGE_SHIFT);
		nr_pages = 1 << page_order;
		base_pfn = PHYS_PFN(addr);

		if (base_pfn >= alt_start && base_pfn < alt_end) {
			vmem_altmap_free(altmap, nr_pages);
		} else if (PageReserved(page)) {
			/* allocated from bootmem */
			if (page_size < PAGE_SIZE) {
				/*
				 * this shouldn't happen, but if it is
				 * the case, leave the memory there
				 */
				WARN_ON_ONCE(1);
			} else {
				while (nr_pages--)
					free_reserved_page(page++);
			}
		} else {
			free_pages((unsigned long)(__va(addr)), page_order);
		}

		vmemmap_remove_mapping(start, page_size);
	}
}
#endif
void register_page_bootmem_memmap(unsigned long section_nr,
				  struct page *start_page, unsigned long size)
{
}

#endif /* CONFIG_SPARSEMEM_VMEMMAP */

#ifdef CONFIG_PPC_BOOK3S_64
static bool disable_radix = !IS_ENABLED(CONFIG_PPC_RADIX_MMU_DEFAULT);

static int __init parse_disable_radix(char *p)
{
	bool val;

	if (!p)
		val = true;
	else if (kstrtobool(p, &val))
		return -EINVAL;

	disable_radix = val;

	return 0;
}
early_param("disable_radix", parse_disable_radix);

/*
 * If we're running under a hypervisor, we need to check the contents of
 * /chosen/ibm,architecture-vec-5 to see if the hypervisor is willing to do
 * radix.  If not, we clear the radix feature bit so we fall back to hash.
 */
static void __init early_check_vec5(void)
{
	unsigned long root, chosen;
	int size;
	const u8 *vec5;
	u8 mmu_supported;

	root = of_get_flat_dt_root();
	chosen = of_get_flat_dt_subnode_by_name(root, "chosen");
	if (chosen == -FDT_ERR_NOTFOUND) {
		cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
		return;
	}
	vec5 = of_get_flat_dt_prop(chosen, "ibm,architecture-vec-5", &size);
	if (!vec5) {
		cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
		return;
	}
	if (size <= OV5_INDX(OV5_MMU_SUPPORT)) {
		cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
		return;
	}

	/* Check for supported configuration */
	mmu_supported = vec5[OV5_INDX(OV5_MMU_SUPPORT)] &
			OV5_FEAT(OV5_MMU_SUPPORT);
	if (mmu_supported == OV5_FEAT(OV5_MMU_RADIX)) {
		/* Hypervisor only supports radix - check enabled && GTSE */
		if (!early_radix_enabled()) {
			pr_warn("WARNING: Ignoring cmdline option disable_radix\n");
		}
		if (!(vec5[OV5_INDX(OV5_RADIX_GTSE)] &
						OV5_FEAT(OV5_RADIX_GTSE))) {
			pr_warn("WARNING: Hypervisor doesn't support RADIX with GTSE\n");
		}
		/* Do radix anyway - the hypervisor said we had to */
		cur_cpu_spec->mmu_features |= MMU_FTR_TYPE_RADIX;
	} else if (mmu_supported == OV5_FEAT(OV5_MMU_HASH)) {
		/* Hypervisor only supports hash - disable radix */
		cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
	}
}

void __init mmu_early_init_devtree(void)
{
	/* Disable radix mode based on kernel command line. */
	if (disable_radix)
		cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;

	/*
	 * Check /chosen/ibm,architecture-vec-5 if running as a guest.
	 * When running bare-metal, we can use radix if we like
	 * even though the ibm,architecture-vec-5 property created by
	 * skiboot doesn't have the necessary bits set.
	 */
	if (!(mfmsr() & MSR_HV))
		early_check_vec5();

	if (early_radix_enabled())
		radix__early_init_devtree();
	else
		hash__early_init_devtree();
}
#endif /* CONFIG_PPC_BOOK3S_64 */
Commit	Line	Data
	1	// SPDX-License-Identifier: GPL-2.0-or-later
	2	/*
	3	* PowerPC version
	4	* Copyright (C) 1995-1996 Gary Thomas (gdt@linuxppc.org)
	5	*
	6	* Modifications by Paul Mackerras (PowerMac) (paulus@cs.anu.edu.au)
	7	* and Cort Dougan (PReP) (cort@cs.nmt.edu)
	8	* Copyright (C) 1996 Paul Mackerras
	9	*
	10	* Derived from "arch/i386/mm/init.c"
	11	* Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds
	12	*
	13	* Dave Engebretsen <engebret@us.ibm.com>
	14	* Rework for PPC64 port.
	15	*/
	16
	17	#undef DEBUG
	18
	19	#include <linux/signal.h>
	20	#include <linux/sched.h>
	21	#include <linux/kernel.h>
	22	#include <linux/errno.h>
	23	#include <linux/string.h>
	24	#include <linux/types.h>
	25	#include <linux/mman.h>
	26	#include <linux/mm.h>
	27	#include <linux/swap.h>
	28	#include <linux/stddef.h>
	29	#include <linux/vmalloc.h>
	30	#include <linux/init.h>
	31	#include <linux/delay.h>
	32	#include <linux/highmem.h>
	33	#include <linux/idr.h>
	34	#include <linux/nodemask.h>
	35	#include <linux/module.h>
	36	#include <linux/poison.h>
	37	#include <linux/memblock.h>
	38	#include <linux/hugetlb.h>
	39	#include <linux/slab.h>
	40	#include <linux/of_fdt.h>
	41	#include <linux/libfdt.h>
	42	#include <linux/memremap.h>
	43
	44	#include <asm/pgalloc.h>
	45	#include <asm/page.h>
	46	#include <asm/prom.h>
	47	#include <asm/rtas.h>
	48	#include <asm/io.h>
	49	#include <asm/mmu_context.h>
	50	#include <asm/pgtable.h>
	51	#include <asm/mmu.h>
	52	#include <linux/uaccess.h>
	53	#include <asm/smp.h>
	54	#include <asm/machdep.h>
	55	#include <asm/tlb.h>
	56	#include <asm/eeh.h>
	57	#include <asm/processor.h>
	58	#include <asm/mmzone.h>
	59	#include <asm/cputable.h>
	60	#include <asm/sections.h>
	61	#include <asm/iommu.h>
	62	#include <asm/vdso.h>
	63
	64	#include <mm/mmu_decl.h>
	65
	66	phys_addr_t memstart_addr = ~0;
	67	EXPORT_SYMBOL_GPL(memstart_addr);
	68	phys_addr_t kernstart_addr;
	69	EXPORT_SYMBOL_GPL(kernstart_addr);
	70
	71	#ifdef CONFIG_SPARSEMEM_VMEMMAP
	72	/*
	73	* Given an address within the vmemmap, determine the pfn of the page that
	74	* represents the start of the section it is within. Note that we have to
	75	* do this by hand as the proffered address may not be correctly aligned.
	76	* Subtraction of non-aligned pointers produces undefined results.
	77	*/
	78	static unsigned long __meminit vmemmap_section_start(unsigned long page)
	79	{
	80	unsigned long offset = page - ((unsigned long)(vmemmap));
	81
	82	/* Return the pfn of the start of the section. */
	83	return (offset / sizeof(struct page)) & PAGE_SECTION_MASK;
	84	}
	85
	86	/*
	87	* Check if this vmemmap page is already initialised. If any section
	88	* which overlaps this vmemmap page is initialised then this page is
	89	* initialised already.
	90	*/
	91	static int __meminit vmemmap_populated(unsigned long start, int page_size)
	92	{
	93	unsigned long end = start + page_size;
	94	start = (unsigned long)(pfn_to_page(vmemmap_section_start(start)));
	95
	96	for (; start < end; start += (PAGES_PER_SECTION * sizeof(struct page)))
	97	if (pfn_valid(page_to_pfn((struct page *)start)))
	98	return 1;
	99
	100	return 0;
	101	}
	102
	103	/*
	104	* vmemmap virtual address space management does not have a traditonal page
	105	* table to track which virtual struct pages are backed by physical mapping.
	106	* The virtual to physical mappings are tracked in a simple linked list
	107	* format. 'vmemmap_list' maintains the entire vmemmap physical mapping at
	108	* all times where as the 'next' list maintains the available
	109	* vmemmap_backing structures which have been deleted from the
	110	* 'vmemmap_global' list during system runtime (memory hotplug remove
	111	* operation). The freed 'vmemmap_backing' structures are reused later when
	112	* new requests come in without allocating fresh memory. This pointer also
	113	* tracks the allocated 'vmemmap_backing' structures as we allocate one
	114	* full page memory at a time when we dont have any.
	115	*/
	116	struct vmemmap_backing *vmemmap_list;
	117	static struct vmemmap_backing *next;
	118
	119	/*
	120	* The same pointer 'next' tracks individual chunks inside the allocated
	121	* full page during the boot time and again tracks the freeed nodes during
	122	* runtime. It is racy but it does not happen as they are separated by the
	123	* boot process. Will create problem if some how we have memory hotplug
	124	* operation during boot !!
	125	*/
	126	static int num_left;
	127	static int num_freed;
	128
	129	static __meminit struct vmemmap_backing * vmemmap_list_alloc(int node)
	130	{
	131	struct vmemmap_backing *vmem_back;
	132	/* get from freed entries first */
	133	if (num_freed) {
	134	num_freed--;
	135	vmem_back = next;
	136	next = next->list;
	137
	138	return vmem_back;
	139	}
	140
	141	/* allocate a page when required and hand out chunks */
	142	if (!num_left) {
	143	next = vmemmap_alloc_block(PAGE_SIZE, node);
	144	if (unlikely(!next)) {
	145	WARN_ON(1);
	146	return NULL;
	147	}
	148	num_left = PAGE_SIZE / sizeof(struct vmemmap_backing);
	149	}
	150
	151	num_left--;
	152
	153	return next++;
	154	}
	155
	156	static __meminit void vmemmap_list_populate(unsigned long phys,
	157	unsigned long start,
	158	int node)
	159	{
	160	struct vmemmap_backing *vmem_back;
	161
	162	vmem_back = vmemmap_list_alloc(node);
	163	if (unlikely(!vmem_back)) {
	164	WARN_ON(1);
	165	return;
	166	}
	167
	168	vmem_back->phys = phys;
	169	vmem_back->virt_addr = start;
	170	vmem_back->list = vmemmap_list;
	171
	172	vmemmap_list = vmem_back;
	173	}
	174
	175	int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node,
	176	struct vmem_altmap *altmap)
	177	{
	178	unsigned long page_size = 1 << mmu_psize_defs[mmu_vmemmap_psize].shift;
	179
	180	/* Align to the page size of the linear mapping. */
	181	start = _ALIGN_DOWN(start, page_size);
	182
	183	pr_debug("vmemmap_populate %lx..%lx, node %d\n", start, end, node);
	184
	185	for (; start < end; start += page_size) {
	186	void *p = NULL;
	187	int rc;
	188
	189	if (vmemmap_populated(start, page_size))
	190	continue;
	191
	192	/*
	193	* Allocate from the altmap first if we have one. This may
	194	* fail due to alignment issues when using 16MB hugepages, so
	195	* fall back to system memory if the altmap allocation fail.
	196	*/
	197	if (altmap)
	198	p = altmap_alloc_block_buf(page_size, altmap);
	199	if (!p)
	200	p = vmemmap_alloc_block_buf(page_size, node);
	201	if (!p)
	202	return -ENOMEM;
	203
	204	vmemmap_list_populate(__pa(p), start, node);
	205
	206	pr_debug(" * %016lx..%016lx allocated at %p\n",
	207	start, start + page_size, p);
	208
	209	rc = vmemmap_create_mapping(start, page_size, __pa(p));
	210	if (rc < 0) {
	211	pr_warn("%s: Unable to create vmemmap mapping: %d\n",
	212	__func__, rc);
	213	return -EFAULT;
	214	}
	215	}
	216
	217	return 0;
	218	}
	219
	220	#ifdef CONFIG_MEMORY_HOTPLUG
	221	static unsigned long vmemmap_list_free(unsigned long start)
	222	{
	223	struct vmemmap_backing vmem_back, vmem_back_prev;
	224
	225	vmem_back_prev = vmem_back = vmemmap_list;
	226
	227	/* look for it with prev pointer recorded */
	228	for (; vmem_back; vmem_back = vmem_back->list) {
	229	if (vmem_back->virt_addr == start)
	230	break;
	231	vmem_back_prev = vmem_back;
	232	}
	233
	234	if (unlikely(!vmem_back)) {
	235	WARN_ON(1);
	236	return 0;
	237	}
	238
	239	/* remove it from vmemmap_list */
	240	if (vmem_back == vmemmap_list) /* remove head */
	241	vmemmap_list = vmem_back->list;
	242	else
	243	vmem_back_prev->list = vmem_back->list;
	244
	245	/* next point to this freed entry */
	246	vmem_back->list = next;
	247	next = vmem_back;
	248	num_freed++;
	249
	250	return vmem_back->phys;
	251	}
	252
	253	void __ref vmemmap_free(unsigned long start, unsigned long end,
	254	struct vmem_altmap *altmap)
	255	{
	256	unsigned long page_size = 1 << mmu_psize_defs[mmu_vmemmap_psize].shift;
	257	unsigned long page_order = get_order(page_size);
	258	unsigned long alt_start = ~0, alt_end = ~0;
	259	unsigned long base_pfn;
	260
	261	start = _ALIGN_DOWN(start, page_size);
	262	if (altmap) {
	263	alt_start = altmap->base_pfn;
	264	alt_end = altmap->base_pfn + altmap->reserve +
	265	altmap->free + altmap->alloc + altmap->align;
	266	}
	267
	268	pr_debug("vmemmap_free %lx...%lx\n", start, end);
	269
	270	for (; start < end; start += page_size) {
	271	unsigned long nr_pages, addr;
	272	struct page *page;
	273
	274	/*
	275	* the section has already be marked as invalid, so
	276	* vmemmap_populated() true means some other sections still
	277	* in this page, so skip it.
	278	*/
	279	if (vmemmap_populated(start, page_size))
	280	continue;
	281
	282	addr = vmemmap_list_free(start);
	283	if (!addr)
	284	continue;
	285
	286	page = pfn_to_page(addr >> PAGE_SHIFT);
	287	nr_pages = 1 << page_order;
	288	base_pfn = PHYS_PFN(addr);
	289
	290	if (base_pfn >= alt_start && base_pfn < alt_end) {
	291	vmem_altmap_free(altmap, nr_pages);
	292	} else if (PageReserved(page)) {
	293	/* allocated from bootmem */
	294	if (page_size < PAGE_SIZE) {
	295	/*
	296	* this shouldn't happen, but if it is
	297	* the case, leave the memory there
	298	*/
	299	WARN_ON_ONCE(1);
	300	} else {
	301	while (nr_pages--)
	302	free_reserved_page(page++);
	303	}
	304	} else {
	305	free_pages((unsigned long)(__va(addr)), page_order);
	306	}
	307
	308	vmemmap_remove_mapping(start, page_size);
	309	}
	310	}
	311	#endif
	312	void register_page_bootmem_memmap(unsigned long section_nr,
	313	struct page *start_page, unsigned long size)
	314	{
	315	}
	316
	317	#endif /* CONFIG_SPARSEMEM_VMEMMAP */
	318
	319	#ifdef CONFIG_PPC_BOOK3S_64
	320	static bool disable_radix = !IS_ENABLED(CONFIG_PPC_RADIX_MMU_DEFAULT);
	321
	322	static int __init parse_disable_radix(char *p)
	323	{
	324	bool val;
	325
	326	if (!p)
	327	val = true;
	328	else if (kstrtobool(p, &val))
	329	return -EINVAL;
	330
	331	disable_radix = val;
	332
	333	return 0;
	334	}
	335	early_param("disable_radix", parse_disable_radix);
	336
	337	/*
	338	* If we're running under a hypervisor, we need to check the contents of
	339	* /chosen/ibm,architecture-vec-5 to see if the hypervisor is willing to do
	340	* radix. If not, we clear the radix feature bit so we fall back to hash.
	341	*/
	342	static void __init early_check_vec5(void)
	343	{
	344	unsigned long root, chosen;
	345	int size;
	346	const u8 *vec5;
	347	u8 mmu_supported;
	348
	349	root = of_get_flat_dt_root();
	350	chosen = of_get_flat_dt_subnode_by_name(root, "chosen");
	351	if (chosen == -FDT_ERR_NOTFOUND) {
	352	cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
	353	return;
	354	}
	355	vec5 = of_get_flat_dt_prop(chosen, "ibm,architecture-vec-5", &size);
	356	if (!vec5) {
	357	cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
	358	return;
	359	}
	360	if (size <= OV5_INDX(OV5_MMU_SUPPORT)) {
	361	cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
	362	return;
	363	}
	364
	365	/* Check for supported configuration */
	366	mmu_supported = vec5[OV5_INDX(OV5_MMU_SUPPORT)] &
	367	OV5_FEAT(OV5_MMU_SUPPORT);
	368	if (mmu_supported == OV5_FEAT(OV5_MMU_RADIX)) {
	369	/* Hypervisor only supports radix - check enabled && GTSE */
	370	if (!early_radix_enabled()) {
	371	pr_warn("WARNING: Ignoring cmdline option disable_radix\n");
	372	}
	373	if (!(vec5[OV5_INDX(OV5_RADIX_GTSE)] &
	374	OV5_FEAT(OV5_RADIX_GTSE))) {
	375	pr_warn("WARNING: Hypervisor doesn't support RADIX with GTSE\n");
	376	}
	377	/* Do radix anyway - the hypervisor said we had to */
	378	cur_cpu_spec->mmu_features \|= MMU_FTR_TYPE_RADIX;
	379	} else if (mmu_supported == OV5_FEAT(OV5_MMU_HASH)) {
	380	/* Hypervisor only supports hash - disable radix */
	381	cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
	382	}
	383	}
	384
	385	void __init mmu_early_init_devtree(void)
	386	{
	387	/* Disable radix mode based on kernel command line. */
	388	if (disable_radix)
	389	cur_cpu_spec->mmu_features &= ~MMU_FTR_TYPE_RADIX;
	390
	391	/*
	392	* Check /chosen/ibm,architecture-vec-5 if running as a guest.
	393	* When running bare-metal, we can use radix if we like
	394	* even though the ibm,architecture-vec-5 property created by
	395	* skiboot doesn't have the necessary bits set.
	396	*/
	397	if (!(mfmsr() & MSR_HV))
	398	early_check_vec5();
	399
	400	if (early_radix_enabled())
	401	radix__early_init_devtree();
	402	else
	403	hash__early_init_devtree();
	404	}
	405	#endif /* CONFIG_PPC_BOOK3S_64 */