[mirror_ubuntu-jammy-kernel.git] / fs / proc / kcore.c

// SPDX-License-Identifier: GPL-2.0
/*
 *	fs/proc/kcore.c kernel ELF core dumper
 *
 *	Modelled on fs/exec.c:aout_core_dump()
 *	Jeremy Fitzhardinge <jeremy@sw.oz.au>
 *	ELF version written by David Howells <David.Howells@nexor.co.uk>
 *	Modified and incorporated into 2.3.x by Tigran Aivazian <tigran@veritas.com>
 *	Support to dump vmalloc'd areas (ELF only), Tigran Aivazian <tigran@veritas.com>
 *	Safe accesses to vmalloc/direct-mapped discontiguous areas, Kanoj Sarcar <kanoj@sgi.com>
 */

#include <linux/crash_core.h>
#include <linux/mm.h>
#include <linux/proc_fs.h>
#include <linux/kcore.h>
#include <linux/user.h>
#include <linux/capability.h>
#include <linux/elf.h>
#include <linux/elfcore.h>
#include <linux/notifier.h>
#include <linux/vmalloc.h>
#include <linux/highmem.h>
#include <linux/printk.h>
#include <linux/memblock.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/uaccess.h>
#include <asm/io.h>
#include <linux/list.h>
#include <linux/ioport.h>
#include <linux/memory.h>
#include <linux/sched/task.h>
#include <linux/security.h>
#include <asm/sections.h>
#include "internal.h"

#define CORE_STR "CORE"

#ifndef ELF_CORE_EFLAGS
#define ELF_CORE_EFLAGS	0
#endif

static struct proc_dir_entry *proc_root_kcore;


#ifndef kc_vaddr_to_offset
#define	kc_vaddr_to_offset(v) ((v) - PAGE_OFFSET)
#endif
#ifndef	kc_offset_to_vaddr
#define	kc_offset_to_vaddr(o) ((o) + PAGE_OFFSET)
#endif

static LIST_HEAD(kclist_head);
static DECLARE_RWSEM(kclist_lock);
static int kcore_need_update = 1;

/*
 * Returns > 0 for RAM pages, 0 for non-RAM pages, < 0 on error
 * Same as oldmem_pfn_is_ram in vmcore
 */
static int (*mem_pfn_is_ram)(unsigned long pfn);

int __init register_mem_pfn_is_ram(int (*fn)(unsigned long pfn))
{
	if (mem_pfn_is_ram)
		return -EBUSY;
	mem_pfn_is_ram = fn;
	return 0;
}

static int pfn_is_ram(unsigned long pfn)
{
	if (mem_pfn_is_ram)
		return mem_pfn_is_ram(pfn);
	else
		return 1;
}

/* This doesn't grab kclist_lock, so it should only be used at init time. */
void __init kclist_add(struct kcore_list *new, void *addr, size_t size,
		       int type)
{
	new->addr = (unsigned long)addr;
	new->size = size;
	new->type = type;

	list_add_tail(&new->list, &kclist_head);
}

static size_t get_kcore_size(int *nphdr, size_t *phdrs_len, size_t *notes_len,
			     size_t *data_offset)
{
	size_t try, size;
	struct kcore_list *m;

	*nphdr = 1; /* PT_NOTE */
	size = 0;

	list_for_each_entry(m, &kclist_head, list) {
		try = kc_vaddr_to_offset((size_t)m->addr + m->size);
		if (try > size)
			size = try;
		*nphdr = *nphdr + 1;
	}

	*phdrs_len = *nphdr * sizeof(struct elf_phdr);
	*notes_len = (4 * sizeof(struct elf_note) +
		      3 * ALIGN(sizeof(CORE_STR), 4) +
		      VMCOREINFO_NOTE_NAME_BYTES +
		      ALIGN(sizeof(struct elf_prstatus), 4) +
		      ALIGN(sizeof(struct elf_prpsinfo), 4) +
		      ALIGN(arch_task_struct_size, 4) +
		      ALIGN(vmcoreinfo_size, 4));
	*data_offset = PAGE_ALIGN(sizeof(struct elfhdr) + *phdrs_len +
				  *notes_len);
	return *data_offset + size;
}

#ifdef CONFIG_HIGHMEM
/*
 * If no highmem, we can assume [0...max_low_pfn) continuous range of memory
 * because memory hole is not as big as !HIGHMEM case.
 * (HIGHMEM is special because part of memory is _invisible_ from the kernel.)
 */
static int kcore_ram_list(struct list_head *head)
{
	struct kcore_list *ent;

	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
	if (!ent)
		return -ENOMEM;
	ent->addr = (unsigned long)__va(0);
	ent->size = max_low_pfn << PAGE_SHIFT;
	ent->type = KCORE_RAM;
	list_add(&ent->list, head);
	return 0;
}

#else /* !CONFIG_HIGHMEM */

#ifdef CONFIG_SPARSEMEM_VMEMMAP
/* calculate vmemmap's address from given system ram pfn and register it */
static int
get_sparsemem_vmemmap_info(struct kcore_list *ent, struct list_head *head)
{
	unsigned long pfn = __pa(ent->addr) >> PAGE_SHIFT;
	unsigned long nr_pages = ent->size >> PAGE_SHIFT;
	unsigned long start, end;
	struct kcore_list *vmm, *tmp;


	start = ((unsigned long)pfn_to_page(pfn)) & PAGE_MASK;
	end = ((unsigned long)pfn_to_page(pfn + nr_pages)) - 1;
	end = PAGE_ALIGN(end);
	/* overlap check (because we have to align page */
	list_for_each_entry(tmp, head, list) {
		if (tmp->type != KCORE_VMEMMAP)
			continue;
		if (start < tmp->addr + tmp->size)
			if (end > tmp->addr)
				end = tmp->addr;
	}
	if (start < end) {
		vmm = kmalloc(sizeof(*vmm), GFP_KERNEL);
		if (!vmm)
			return 0;
		vmm->addr = start;
		vmm->size = end - start;
		vmm->type = KCORE_VMEMMAP;
		list_add_tail(&vmm->list, head);
	}
	return 1;

}
#else
static int
get_sparsemem_vmemmap_info(struct kcore_list *ent, struct list_head *head)
{
	return 1;
}

#endif

static int
kclist_add_private(unsigned long pfn, unsigned long nr_pages, void *arg)
{
	struct list_head *head = (struct list_head *)arg;
	struct kcore_list *ent;
	struct page *p;

	if (!pfn_valid(pfn))
		return 1;

	p = pfn_to_page(pfn);
	if (!memmap_valid_within(pfn, p, page_zone(p)))
		return 1;

	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
	if (!ent)
		return -ENOMEM;
	ent->addr = (unsigned long)page_to_virt(p);
	ent->size = nr_pages << PAGE_SHIFT;

	if (!virt_addr_valid(ent->addr))
		goto free_out;

	/* cut not-mapped area. ....from ppc-32 code. */
	if (ULONG_MAX - ent->addr < ent->size)
		ent->size = ULONG_MAX - ent->addr;

	/*
	 * We've already checked virt_addr_valid so we know this address
	 * is a valid pointer, therefore we can check against it to determine
	 * if we need to trim
	 */
	if (VMALLOC_START > ent->addr) {
		if (VMALLOC_START - ent->addr < ent->size)
			ent->size = VMALLOC_START - ent->addr;
	}

	ent->type = KCORE_RAM;
	list_add_tail(&ent->list, head);

	if (!get_sparsemem_vmemmap_info(ent, head)) {
		list_del(&ent->list);
		goto free_out;
	}

	return 0;
free_out:
	kfree(ent);
	return 1;
}

static int kcore_ram_list(struct list_head *list)
{
	int nid, ret;
	unsigned long end_pfn;

	/* Not inialized....update now */
	/* find out "max pfn" */
	end_pfn = 0;
	for_each_node_state(nid, N_MEMORY) {
		unsigned long node_end;
		node_end = node_end_pfn(nid);
		if (end_pfn < node_end)
			end_pfn = node_end;
	}
	/* scan 0 to max_pfn */
	ret = walk_system_ram_range(0, end_pfn, list, kclist_add_private);
	if (ret)
		return -ENOMEM;
	return 0;
}
#endif /* CONFIG_HIGHMEM */

static int kcore_update_ram(void)
{
	LIST_HEAD(list);
	LIST_HEAD(garbage);
	int nphdr;
	size_t phdrs_len, notes_len, data_offset;
	struct kcore_list *tmp, *pos;
	int ret = 0;

	down_write(&kclist_lock);
	if (!xchg(&kcore_need_update, 0))
		goto out;

	ret = kcore_ram_list(&list);
	if (ret) {
		/* Couldn't get the RAM list, try again next time. */
		WRITE_ONCE(kcore_need_update, 1);
		list_splice_tail(&list, &garbage);
		goto out;
	}

	list_for_each_entry_safe(pos, tmp, &kclist_head, list) {
		if (pos->type == KCORE_RAM || pos->type == KCORE_VMEMMAP)
			list_move(&pos->list, &garbage);
	}
	list_splice_tail(&list, &kclist_head);

	proc_root_kcore->size = get_kcore_size(&nphdr, &phdrs_len, &notes_len,
					       &data_offset);

out:
	up_write(&kclist_lock);
	list_for_each_entry_safe(pos, tmp, &garbage, list) {
		list_del(&pos->list);
		kfree(pos);
	}
	return ret;
}

static void append_kcore_note(char *notes, size_t *i, const char *name,
			      unsigned int type, const void *desc,
			      size_t descsz)
{
	struct elf_note *note = (struct elf_note *)&notes[*i];

	note->n_namesz = strlen(name) + 1;
	note->n_descsz = descsz;
	note->n_type = type;
	*i += sizeof(*note);
	memcpy(&notes[*i], name, note->n_namesz);
	*i = ALIGN(*i + note->n_namesz, 4);
	memcpy(&notes[*i], desc, descsz);
	*i = ALIGN(*i + descsz, 4);
}

static ssize_t
read_kcore(struct file *file, char __user *buffer, size_t buflen, loff_t *fpos)
{
	char *buf = file->private_data;
	size_t phdrs_offset, notes_offset, data_offset;
	size_t phdrs_len, notes_len;
	struct kcore_list *m;
	size_t tsz;
	int nphdr;
	unsigned long start;
	size_t orig_buflen = buflen;
	int ret = 0;

	down_read(&kclist_lock);

	get_kcore_size(&nphdr, &phdrs_len, &notes_len, &data_offset);
	phdrs_offset = sizeof(struct elfhdr);
	notes_offset = phdrs_offset + phdrs_len;

	/* ELF file header. */
	if (buflen && *fpos < sizeof(struct elfhdr)) {
		struct elfhdr ehdr = {
			.e_ident = {
				[EI_MAG0] = ELFMAG0,
				[EI_MAG1] = ELFMAG1,
				[EI_MAG2] = ELFMAG2,
				[EI_MAG3] = ELFMAG3,
				[EI_CLASS] = ELF_CLASS,
				[EI_DATA] = ELF_DATA,
				[EI_VERSION] = EV_CURRENT,
				[EI_OSABI] = ELF_OSABI,
			},
			.e_type = ET_CORE,
			.e_machine = ELF_ARCH,
			.e_version = EV_CURRENT,
			.e_phoff = sizeof(struct elfhdr),
			.e_flags = ELF_CORE_EFLAGS,
			.e_ehsize = sizeof(struct elfhdr),
			.e_phentsize = sizeof(struct elf_phdr),
			.e_phnum = nphdr,
		};

		tsz = min_t(size_t, buflen, sizeof(struct elfhdr) - *fpos);
		if (copy_to_user(buffer, (char *)&ehdr + *fpos, tsz)) {
			ret = -EFAULT;
			goto out;
		}

		buffer += tsz;
		buflen -= tsz;
		*fpos += tsz;
	}

	/* ELF program headers. */
	if (buflen && *fpos < phdrs_offset + phdrs_len) {
		struct elf_phdr *phdrs, *phdr;

		phdrs = kzalloc(phdrs_len, GFP_KERNEL);
		if (!phdrs) {
			ret = -ENOMEM;
			goto out;
		}

		phdrs[0].p_type = PT_NOTE;
		phdrs[0].p_offset = notes_offset;
		phdrs[0].p_filesz = notes_len;

		phdr = &phdrs[1];
		list_for_each_entry(m, &kclist_head, list) {
			phdr->p_type = PT_LOAD;
			phdr->p_flags = PF_R | PF_W | PF_X;
			phdr->p_offset = kc_vaddr_to_offset(m->addr) + data_offset;
			if (m->type == KCORE_REMAP)
				phdr->p_vaddr = (size_t)m->vaddr;
			else
				phdr->p_vaddr = (size_t)m->addr;
			if (m->type == KCORE_RAM || m->type == KCORE_REMAP)
				phdr->p_paddr = __pa(m->addr);
			else if (m->type == KCORE_TEXT)
				phdr->p_paddr = __pa_symbol(m->addr);
			else
				phdr->p_paddr = (elf_addr_t)-1;
			phdr->p_filesz = phdr->p_memsz = m->size;
			phdr->p_align = PAGE_SIZE;
			phdr++;
		}

		tsz = min_t(size_t, buflen, phdrs_offset + phdrs_len - *fpos);
		if (copy_to_user(buffer, (char *)phdrs + *fpos - phdrs_offset,
				 tsz)) {
			kfree(phdrs);
			ret = -EFAULT;
			goto out;
		}
		kfree(phdrs);

		buffer += tsz;
		buflen -= tsz;
		*fpos += tsz;
	}

	/* ELF note segment. */
	if (buflen && *fpos < notes_offset + notes_len) {
		struct elf_prstatus prstatus = {};
		struct elf_prpsinfo prpsinfo = {
			.pr_sname = 'R',
			.pr_fname = "vmlinux",
		};
		char *notes;
		size_t i = 0;

		strlcpy(prpsinfo.pr_psargs, saved_command_line,
			sizeof(prpsinfo.pr_psargs));

		notes = kzalloc(notes_len, GFP_KERNEL);
		if (!notes) {
			ret = -ENOMEM;
			goto out;
		}

		append_kcore_note(notes, &i, CORE_STR, NT_PRSTATUS, &prstatus,
				  sizeof(prstatus));
		append_kcore_note(notes, &i, CORE_STR, NT_PRPSINFO, &prpsinfo,
				  sizeof(prpsinfo));
		append_kcore_note(notes, &i, CORE_STR, NT_TASKSTRUCT, current,
				  arch_task_struct_size);
		/*
		 * vmcoreinfo_size is mostly constant after init time, but it
		 * can be changed by crash_save_vmcoreinfo(). Racing here with a
		 * panic on another CPU before the machine goes down is insanely
		 * unlikely, but it's better to not leave potential buffer
		 * overflows lying around, regardless.
		 */
		append_kcore_note(notes, &i, VMCOREINFO_NOTE_NAME, 0,
				  vmcoreinfo_data,
				  min(vmcoreinfo_size, notes_len - i));

		tsz = min_t(size_t, buflen, notes_offset + notes_len - *fpos);
		if (copy_to_user(buffer, notes + *fpos - notes_offset, tsz)) {
			kfree(notes);
			ret = -EFAULT;
			goto out;
		}
		kfree(notes);

		buffer += tsz;
		buflen -= tsz;
		*fpos += tsz;
	}

	/*
	 * Check to see if our file offset matches with any of
	 * the addresses in the elf_phdr on our list.
	 */
	start = kc_offset_to_vaddr(*fpos - data_offset);
	if ((tsz = (PAGE_SIZE - (start & ~PAGE_MASK))) > buflen)
		tsz = buflen;

	m = NULL;
	while (buflen) {
		/*
		 * If this is the first iteration or the address is not within
		 * the previous entry, search for a matching entry.
		 */
		if (!m || start < m->addr || start >= m->addr + m->size) {
			list_for_each_entry(m, &kclist_head, list) {
				if (start >= m->addr &&
				    start < m->addr + m->size)
					break;
			}
		}

		if (&m->list == &kclist_head) {
			if (clear_user(buffer, tsz)) {
				ret = -EFAULT;
				goto out;
			}
			m = NULL;	/* skip the list anchor */
		} else if (!pfn_is_ram(__pa(start) >> PAGE_SHIFT)) {
			if (clear_user(buffer, tsz)) {
				ret = -EFAULT;
				goto out;
			}
		} else if (m->type == KCORE_VMALLOC) {
			vread(buf, (char *)start, tsz);
			/* we have to zero-fill user buffer even if no read */
			if (copy_to_user(buffer, buf, tsz)) {
				ret = -EFAULT;
				goto out;
			}
		} else if (m->type == KCORE_USER) {
			/* User page is handled prior to normal kernel page: */
			if (copy_to_user(buffer, (char *)start, tsz)) {
				ret = -EFAULT;
				goto out;
			}
		} else {
			if (kern_addr_valid(start)) {
				/*
				 * Using bounce buffer to bypass the
				 * hardened user copy kernel text checks.
				 */
				if (probe_kernel_read(buf, (void *) start, tsz)) {
					if (clear_user(buffer, tsz)) {
						ret = -EFAULT;
						goto out;
					}
				} else {
					if (copy_to_user(buffer, buf, tsz)) {
						ret = -EFAULT;
						goto out;
					}
				}
			} else {
				if (clear_user(buffer, tsz)) {
					ret = -EFAULT;
					goto out;
				}
			}
		}
		buflen -= tsz;
		*fpos += tsz;
		buffer += tsz;
		start += tsz;
		tsz = (buflen > PAGE_SIZE ? PAGE_SIZE : buflen);
	}

out:
	up_read(&kclist_lock);
	if (ret)
		return ret;
	return orig_buflen - buflen;
}

static int open_kcore(struct inode *inode, struct file *filp)
{
	int ret = security_locked_down(LOCKDOWN_KCORE);

	if (!capable(CAP_SYS_RAWIO))
		return -EPERM;

	if (ret)
		return ret;

	filp->private_data = kmalloc(PAGE_SIZE, GFP_KERNEL);
	if (!filp->private_data)
		return -ENOMEM;

	if (kcore_need_update)
		kcore_update_ram();
	if (i_size_read(inode) != proc_root_kcore->size) {
		inode_lock(inode);
		i_size_write(inode, proc_root_kcore->size);
		inode_unlock(inode);
	}
	return 0;
}

static int release_kcore(struct inode *inode, struct file *file)
{
	kfree(file->private_data);
	return 0;
}

static const struct proc_ops kcore_proc_ops = {
	.proc_read	= read_kcore,
	.proc_open	= open_kcore,
	.proc_release	= release_kcore,
	.proc_lseek	= default_llseek,
};

/* just remember that we have to update kcore */
static int __meminit kcore_callback(struct notifier_block *self,
				    unsigned long action, void *arg)
{
	switch (action) {
	case MEM_ONLINE:
	case MEM_OFFLINE:
		kcore_need_update = 1;
		break;
	}
	return NOTIFY_OK;
}

static struct notifier_block kcore_callback_nb __meminitdata = {
	.notifier_call = kcore_callback,
	.priority = 0,
};

static struct kcore_list kcore_vmalloc;

#ifdef CONFIG_ARCH_PROC_KCORE_TEXT
static struct kcore_list kcore_text;
/*
 * If defined, special segment is used for mapping kernel text instead of
 * direct-map area. We need to create special TEXT section.
 */
static void __init proc_kcore_text_init(void)
{
	kclist_add(&kcore_text, _text, _end - _text, KCORE_TEXT);
}
#else
static void __init proc_kcore_text_init(void)
{
}
#endif

#if defined(CONFIG_MODULES) && defined(MODULES_VADDR)
/*
 * MODULES_VADDR has no intersection with VMALLOC_ADDR.
 */
static struct kcore_list kcore_modules;
static void __init add_modules_range(void)
{
	if (MODULES_VADDR != VMALLOC_START && MODULES_END != VMALLOC_END) {
		kclist_add(&kcore_modules, (void *)MODULES_VADDR,
			MODULES_END - MODULES_VADDR, KCORE_VMALLOC);
	}
}
#else
static void __init add_modules_range(void)
{
}
#endif

static int __init proc_kcore_init(void)
{
	proc_root_kcore = proc_create("kcore", S_IRUSR, NULL, &kcore_proc_ops);
	if (!proc_root_kcore) {
		pr_err("couldn't create /proc/kcore\n");
		return 0; /* Always returns 0. */
	}
	/* Store text area if it's special */
	proc_kcore_text_init();
	/* Store vmalloc area */
	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START,
		VMALLOC_END - VMALLOC_START, KCORE_VMALLOC);
	add_modules_range();
	/* Store direct-map area from physical memory map */
	kcore_update_ram();
	register_hotmemory_notifier(&kcore_callback_nb);

	return 0;
}
fs_initcall(proc_kcore_init);
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/*
	3	* fs/proc/kcore.c kernel ELF core dumper
	4	*
	5	* Modelled on fs/exec.c:aout_core_dump()
	6	* Jeremy Fitzhardinge <jeremy@sw.oz.au>
	7	* ELF version written by David Howells <David.Howells@nexor.co.uk>
	8	* Modified and incorporated into 2.3.x by Tigran Aivazian <tigran@veritas.com>
	9	* Support to dump vmalloc'd areas (ELF only), Tigran Aivazian <tigran@veritas.com>
	10	* Safe accesses to vmalloc/direct-mapped discontiguous areas, Kanoj Sarcar <kanoj@sgi.com>
	11	*/
	12
23c85094	13	#include <linux/crash_core.h>
1da177e4 LT	14	#include <linux/mm.h>
1da177e4 LT	15	#include <linux/proc_fs.h>
2f96b8c1	16	#include <linux/kcore.h>
1da177e4	17	#include <linux/user.h>
16f7e0fe	18	#include <linux/capability.h>
1da177e4 LT	19	#include <linux/elf.h>
1da177e4 LT	20	#include <linux/elfcore.h>
3c743a7f	21	#include <linux/notifier.h>
1da177e4 LT	22	#include <linux/vmalloc.h>
1da177e4 LT	23	#include <linux/highmem.h>
87ebdc00	24	#include <linux/printk.h>
57c8a661	25	#include <linux/memblock.h>
1da177e4	26	#include <linux/init.h>
5a0e3ad6	27	#include <linux/slab.h>
7c0f6ba6	28	#include <linux/uaccess.h>
1da177e4	29	#include <asm/io.h>
2ef43ec7	30	#include <linux/list.h>
3089aa1b	31	#include <linux/ioport.h>
3089aa1b	32	#include <linux/memory.h>
29930025	33	#include <linux/sched/task.h>
02e935bf	34	#include <linux/security.h>
9492587c	35	#include <asm/sections.h>
59d8053f	36	#include "internal.h"
1da177e4	37
36027604	38	#define CORE_STR "CORE"
1da177e4	39
79885b22 EI	40	#ifndef ELF_CORE_EFLAGS
	41	#define ELF_CORE_EFLAGS 0
	42	#endif
	43
97ce5d6d AD	44	static struct proc_dir_entry *proc_root_kcore;
97ce5d6d AD	45
1da177e4 LT	46
	47	#ifndef kc_vaddr_to_offset
	48	#define kc_vaddr_to_offset(v) ((v) - PAGE_OFFSET)
	49	#endif
	50	#ifndef kc_offset_to_vaddr
	51	#define kc_offset_to_vaddr(o) ((o) + PAGE_OFFSET)
	52	#endif
	53
2ef43ec7	54	static LIST_HEAD(kclist_head);
0b172f84	55	static DECLARE_RWSEM(kclist_lock);
3089aa1b	56	static int kcore_need_update = 1;
1da177e4	57
ffc8599a KS	58	/*
	59	* Returns > 0 for RAM pages, 0 for non-RAM pages, < 0 on error
	60	* Same as oldmem_pfn_is_ram in vmcore
	61	*/
	62	static int (*mem_pfn_is_ram)(unsigned long pfn);
	63
	64	int __init register_mem_pfn_is_ram(int (*fn)(unsigned long pfn))
	65	{
	66	if (mem_pfn_is_ram)
	67	return -EBUSY;
	68	mem_pfn_is_ram = fn;
	69	return 0;
	70	}
	71
	72	static int pfn_is_ram(unsigned long pfn)
	73	{
	74	if (mem_pfn_is_ram)
	75	return mem_pfn_is_ram(pfn);
	76	else
	77	return 1;
	78	}
	79
a8dd9c4d OS	80	/* This doesn't grab kclist_lock, so it should only be used at init time. */
	81	void __init kclist_add(struct kcore_list new, void addr, size_t size,
	82	int type)
1da177e4 LT	83	{
	84	new->addr = (unsigned long)addr;
	85	new->size = size;
c30bb2a2	86	new->type = type;
1da177e4	87
2ef43ec7	88	list_add_tail(&new->list, &kclist_head);
1da177e4 LT	89	}
1da177e4 LT	90
37e949bd OS	91	static size_t get_kcore_size(int nphdr, size_t phdrs_len, size_t *notes_len,
37e949bd OS	92	size_t *data_offset)
1da177e4 LT	93	{
	94	size_t try, size;
	95	struct kcore_list *m;
	96
	97	nphdr = 1; / PT_NOTE */
	98	size = 0;
	99
2ef43ec7	100	list_for_each_entry(m, &kclist_head, list) {
1da177e4 LT	101	try = kc_vaddr_to_offset((size_t)m->addr + m->size);
	102	if (try > size)
	103	size = try;
	104	nphdr = nphdr + 1;
	105	}
37e949bd OS	106
37e949bd OS	107	phdrs_len = nphdr * sizeof(struct elf_phdr);
23c85094 OS	108	notes_len = (4 sizeof(struct elf_note) +
	109	3 * ALIGN(sizeof(CORE_STR), 4) +
	110	VMCOREINFO_NOTE_NAME_BYTES +
37e949bd OS	111	ALIGN(sizeof(struct elf_prstatus), 4) +
37e949bd OS	112	ALIGN(sizeof(struct elf_prpsinfo), 4) +
23c85094 OS	113	ALIGN(arch_task_struct_size, 4) +
23c85094 OS	114	ALIGN(vmcoreinfo_size, 4));
37e949bd OS	115	data_offset = PAGE_ALIGN(sizeof(struct elfhdr) + phdrs_len +
	116	*notes_len);
	117	return *data_offset + size;
1da177e4 LT	118	}
1da177e4 LT	119
3089aa1b KH	120	#ifdef CONFIG_HIGHMEM
	121	/*
	122	* If no highmem, we can assume [0...max_low_pfn) continuous range of memory
	123	* because memory hole is not as big as !HIGHMEM case.
	124	* (HIGHMEM is special because part of memory is _invisible_ from the kernel.)
	125	*/
b66fb005	126	static int kcore_ram_list(struct list_head *head)
3089aa1b	127	{
3089aa1b	128	struct kcore_list *ent;
3089aa1b KH	129
	130	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
	131	if (!ent)
	132	return -ENOMEM;
	133	ent->addr = (unsigned long)__va(0);
	134	ent->size = max_low_pfn << PAGE_SHIFT;
	135	ent->type = KCORE_RAM;
b66fb005 OS	136	list_add(&ent->list, head);
b66fb005 OS	137	return 0;
3089aa1b KH	138	}
	139
	140	#else /* !CONFIG_HIGHMEM */
	141
26562c59 KH	142	#ifdef CONFIG_SPARSEMEM_VMEMMAP
26562c59 KH	143	/* calculate vmemmap's address from given system ram pfn and register it */
b908243c DH	144	static int
b908243c DH	145	get_sparsemem_vmemmap_info(struct kcore_list ent, struct list_head head)
26562c59 KH	146	{
	147	unsigned long pfn = __pa(ent->addr) >> PAGE_SHIFT;
	148	unsigned long nr_pages = ent->size >> PAGE_SHIFT;
	149	unsigned long start, end;
	150	struct kcore_list vmm, tmp;
	151
	152
	153	start = ((unsigned long)pfn_to_page(pfn)) & PAGE_MASK;
	154	end = ((unsigned long)pfn_to_page(pfn + nr_pages)) - 1;
108a8a11	155	end = PAGE_ALIGN(end);
26562c59 KH	156	/* overlap check (because we have to align page */
	157	list_for_each_entry(tmp, head, list) {
	158	if (tmp->type != KCORE_VMEMMAP)
	159	continue;
	160	if (start < tmp->addr + tmp->size)
	161	if (end > tmp->addr)
	162	end = tmp->addr;
	163	}
	164	if (start < end) {
	165	vmm = kmalloc(sizeof(*vmm), GFP_KERNEL);
	166	if (!vmm)
	167	return 0;
	168	vmm->addr = start;
	169	vmm->size = end - start;
	170	vmm->type = KCORE_VMEMMAP;
	171	list_add_tail(&vmm->list, head);
	172	}
	173	return 1;
	174
	175	}
	176	#else
b908243c DH	177	static int
b908243c DH	178	get_sparsemem_vmemmap_info(struct kcore_list ent, struct list_head head)
26562c59 KH	179	{
	180	return 1;
	181	}
	182
	183	#endif
	184
3089aa1b KH	185	static int
	186	kclist_add_private(unsigned long pfn, unsigned long nr_pages, void *arg)
	187	{
	188	struct list_head head = (struct list_head )arg;
	189	struct kcore_list *ent;
3955333d LA	190	struct page *p;
	191
	192	if (!pfn_valid(pfn))
	193	return 1;
	194
	195	p = pfn_to_page(pfn);
	196	if (!memmap_valid_within(pfn, p, page_zone(p)))
	197	return 1;
3089aa1b KH	198
	199	ent = kmalloc(sizeof(*ent), GFP_KERNEL);
	200	if (!ent)
	201	return -ENOMEM;
3955333d	202	ent->addr = (unsigned long)page_to_virt(p);
3089aa1b KH	203	ent->size = nr_pages << PAGE_SHIFT;
3089aa1b KH	204
3955333d	205	if (!virt_addr_valid(ent->addr))
3089aa1b KH	206	goto free_out;
	207
	208	/* cut not-mapped area. ....from ppc-32 code. */
	209	if (ULONG_MAX - ent->addr < ent->size)
	210	ent->size = ULONG_MAX - ent->addr;
	211
3955333d LA	212	/*
	213	* We've already checked virt_addr_valid so we know this address
	214	* is a valid pointer, therefore we can check against it to determine
	215	* if we need to trim
	216	*/
	217	if (VMALLOC_START > ent->addr) {
3089aa1b KH	218	if (VMALLOC_START - ent->addr < ent->size)
	219	ent->size = VMALLOC_START - ent->addr;
	220	}
	221
	222	ent->type = KCORE_RAM;
	223	list_add_tail(&ent->list, head);
26562c59 KH	224
	225	if (!get_sparsemem_vmemmap_info(ent, head)) {
	226	list_del(&ent->list);
	227	goto free_out;
	228	}
	229
3089aa1b KH	230	return 0;
	231	free_out:
	232	kfree(ent);
	233	return 1;
	234	}
	235
b66fb005	236	static int kcore_ram_list(struct list_head *list)
3089aa1b KH	237	{
	238	int nid, ret;
	239	unsigned long end_pfn;
3089aa1b KH	240
	241	/* Not inialized....update now */
	242	/* find out "max pfn" */
	243	end_pfn = 0;
4ff1b2c2	244	for_each_node_state(nid, N_MEMORY) {
3089aa1b	245	unsigned long node_end;
83285c72	246	node_end = node_end_pfn(nid);
3089aa1b KH	247	if (end_pfn < node_end)
	248	end_pfn = node_end;
	249	}
	250	/* scan 0 to max_pfn */
b66fb005 OS	251	ret = walk_system_ram_range(0, end_pfn, list, kclist_add_private);
b66fb005 OS	252	if (ret)
3089aa1b	253	return -ENOMEM;
b66fb005 OS	254	return 0;
	255	}
	256	#endif /* CONFIG_HIGHMEM */
	257
	258	static int kcore_update_ram(void)
	259	{
	260	LIST_HEAD(list);
	261	LIST_HEAD(garbage);
	262	int nphdr;
37e949bd	263	size_t phdrs_len, notes_len, data_offset;
b66fb005 OS	264	struct kcore_list tmp, pos;
	265	int ret = 0;
	266
	267	down_write(&kclist_lock);
	268	if (!xchg(&kcore_need_update, 0))
	269	goto out;
	270
	271	ret = kcore_ram_list(&list);
	272	if (ret) {
	273	/* Couldn't get the RAM list, try again next time. */
	274	WRITE_ONCE(kcore_need_update, 1);
	275	list_splice_tail(&list, &garbage);
	276	goto out;
	277	}
	278
	279	list_for_each_entry_safe(pos, tmp, &kclist_head, list) {
	280	if (pos->type == KCORE_RAM \|\| pos->type == KCORE_VMEMMAP)
	281	list_move(&pos->list, &garbage);
	282	}
	283	list_splice_tail(&list, &kclist_head);
	284
37e949bd OS	285	proc_root_kcore->size = get_kcore_size(&nphdr, &phdrs_len, &notes_len,
37e949bd OS	286	&data_offset);
b66fb005 OS	287
	288	out:
	289	up_write(&kclist_lock);
	290	list_for_each_entry_safe(pos, tmp, &garbage, list) {
	291	list_del(&pos->list);
	292	kfree(pos);
3089aa1b	293	}
3089aa1b KH	294	return ret;
3089aa1b KH	295	}
1da177e4	296
37e949bd OS	297	static void append_kcore_note(char notes, size_t i, const char *name,
	298	unsigned int type, const void *desc,
	299	size_t descsz)
1da177e4	300	{
37e949bd OS	301	struct elf_note note = (struct elf_note )&notes[*i];
	302
	303	note->n_namesz = strlen(name) + 1;
	304	note->n_descsz = descsz;
	305	note->n_type = type;
	306	i += sizeof(note);
	307	memcpy(&notes[*i], name, note->n_namesz);
	308	i = ALIGN(i + note->n_namesz, 4);
	309	memcpy(&notes[*i], desc, descsz);
	310	i = ALIGN(i + descsz, 4);
	311	}
1da177e4	312
1da177e4 LT	313	static ssize_t
	314	read_kcore(struct file file, char __user buffer, size_t buflen, loff_t *fpos)
	315	{
f5beeb18	316	char *buf = file->private_data;
37e949bd OS	317	size_t phdrs_offset, notes_offset, data_offset;
	318	size_t phdrs_len, notes_len;
	319	struct kcore_list *m;
	320	size_t tsz;
1da177e4 LT	321	int nphdr;
1da177e4 LT	322	unsigned long start;
3673fb08 OS	323	size_t orig_buflen = buflen;
3673fb08 OS	324	int ret = 0;
1da177e4	325
0b172f84	326	down_read(&kclist_lock);
678ad5d8	327
37e949bd OS	328	get_kcore_size(&nphdr, &phdrs_len, &notes_len, &data_offset);
	329	phdrs_offset = sizeof(struct elfhdr);
	330	notes_offset = phdrs_offset + phdrs_len;
	331
	332	/* ELF file header. */
	333	if (buflen && *fpos < sizeof(struct elfhdr)) {
	334	struct elfhdr ehdr = {
	335	.e_ident = {
	336	[EI_MAG0] = ELFMAG0,
	337	[EI_MAG1] = ELFMAG1,
	338	[EI_MAG2] = ELFMAG2,
	339	[EI_MAG3] = ELFMAG3,
	340	[EI_CLASS] = ELF_CLASS,
	341	[EI_DATA] = ELF_DATA,
	342	[EI_VERSION] = EV_CURRENT,
	343	[EI_OSABI] = ELF_OSABI,
	344	},
	345	.e_type = ET_CORE,
	346	.e_machine = ELF_ARCH,
	347	.e_version = EV_CURRENT,
	348	.e_phoff = sizeof(struct elfhdr),
	349	.e_flags = ELF_CORE_EFLAGS,
	350	.e_ehsize = sizeof(struct elfhdr),
	351	.e_phentsize = sizeof(struct elf_phdr),
	352	.e_phnum = nphdr,
	353	};
	354
	355	tsz = min_t(size_t, buflen, sizeof(struct elfhdr) - *fpos);
	356	if (copy_to_user(buffer, (char )&ehdr + fpos, tsz)) {
	357	ret = -EFAULT;
	358	goto out;
	359	}
1da177e4	360
37e949bd OS	361	buffer += tsz;
	362	buflen -= tsz;
	363	*fpos += tsz;
	364	}
1da177e4	365
37e949bd OS	366	/* ELF program headers. */
	367	if (buflen && *fpos < phdrs_offset + phdrs_len) {
	368	struct elf_phdr phdrs, phdr;
1da177e4	369
37e949bd OS	370	phdrs = kzalloc(phdrs_len, GFP_KERNEL);
37e949bd OS	371	if (!phdrs) {
3673fb08 OS	372	ret = -ENOMEM;
3673fb08 OS	373	goto out;
1da177e4	374	}
37e949bd OS	375
	376	phdrs[0].p_type = PT_NOTE;
	377	phdrs[0].p_offset = notes_offset;
	378	phdrs[0].p_filesz = notes_len;
	379
	380	phdr = &phdrs[1];
	381	list_for_each_entry(m, &kclist_head, list) {
	382	phdr->p_type = PT_LOAD;
	383	phdr->p_flags = PF_R \| PF_W \| PF_X;
	384	phdr->p_offset = kc_vaddr_to_offset(m->addr) + data_offset;
d207ea8e LT	385	if (m->type == KCORE_REMAP)
	386	phdr->p_vaddr = (size_t)m->vaddr;
	387	else
	388	phdr->p_vaddr = (size_t)m->addr;
	389	if (m->type == KCORE_RAM \|\| m->type == KCORE_REMAP)
37e949bd OS	390	phdr->p_paddr = __pa(m->addr);
	391	else if (m->type == KCORE_TEXT)
	392	phdr->p_paddr = __pa_symbol(m->addr);
	393	else
	394	phdr->p_paddr = (elf_addr_t)-1;
	395	phdr->p_filesz = phdr->p_memsz = m->size;
	396	phdr->p_align = PAGE_SIZE;
	397	phdr++;
	398	}
	399
	400	tsz = min_t(size_t, buflen, phdrs_offset + phdrs_len - *fpos);
	401	if (copy_to_user(buffer, (char )phdrs + fpos - phdrs_offset,
	402	tsz)) {
	403	kfree(phdrs);
3673fb08 OS	404	ret = -EFAULT;
3673fb08 OS	405	goto out;
1da177e4	406	}
37e949bd OS	407	kfree(phdrs);
	408
	409	buffer += tsz;
1da177e4 LT	410	buflen -= tsz;
1da177e4 LT	411	*fpos += tsz;
37e949bd OS	412	}
	413
	414	/* ELF note segment. */
	415	if (buflen && *fpos < notes_offset + notes_len) {
	416	struct elf_prstatus prstatus = {};
	417	struct elf_prpsinfo prpsinfo = {
	418	.pr_sname = 'R',
	419	.pr_fname = "vmlinux",
	420	};
	421	char *notes;
	422	size_t i = 0;
	423
	424	strlcpy(prpsinfo.pr_psargs, saved_command_line,
	425	sizeof(prpsinfo.pr_psargs));
	426
	427	notes = kzalloc(notes_len, GFP_KERNEL);
	428	if (!notes) {
	429	ret = -ENOMEM;
	430	goto out;
	431	}
	432
	433	append_kcore_note(notes, &i, CORE_STR, NT_PRSTATUS, &prstatus,
	434	sizeof(prstatus));
	435	append_kcore_note(notes, &i, CORE_STR, NT_PRPSINFO, &prpsinfo,
	436	sizeof(prpsinfo));
	437	append_kcore_note(notes, &i, CORE_STR, NT_TASKSTRUCT, current,
	438	arch_task_struct_size);
23c85094 OS	439	/*
	440	* vmcoreinfo_size is mostly constant after init time, but it
	441	* can be changed by crash_save_vmcoreinfo(). Racing here with a
	442	* panic on another CPU before the machine goes down is insanely
	443	* unlikely, but it's better to not leave potential buffer
	444	* overflows lying around, regardless.
	445	*/
	446	append_kcore_note(notes, &i, VMCOREINFO_NOTE_NAME, 0,
	447	vmcoreinfo_data,
	448	min(vmcoreinfo_size, notes_len - i));
1da177e4	449
37e949bd OS	450	tsz = min_t(size_t, buflen, notes_offset + notes_len - *fpos);
	451	if (copy_to_user(buffer, notes + *fpos - notes_offset, tsz)) {
	452	kfree(notes);
	453	ret = -EFAULT;
3673fb08	454	goto out;
37e949bd OS	455	}
	456	kfree(notes);
	457
	458	buffer += tsz;
	459	buflen -= tsz;
	460	*fpos += tsz;
3673fb08	461	}
1da177e4 LT	462
	463	/*
	464	* Check to see if our file offset matches with any of
	465	* the addresses in the elf_phdr on our list.
	466	*/
37e949bd	467	start = kc_offset_to_vaddr(*fpos - data_offset);
1da177e4 LT	468	if ((tsz = (PAGE_SIZE - (start & ~PAGE_MASK))) > buflen)
1da177e4 LT	469	tsz = buflen;
1da177e4	470
bf991c22	471	m = NULL;
37e949bd	472	while (buflen) {
bf991c22 OS	473	/*
	474	* If this is the first iteration or the address is not within
	475	* the previous entry, search for a matching entry.
	476	*/
	477	if (!m \|\| start < m->addr \|\| start >= m->addr + m->size) {
	478	list_for_each_entry(m, &kclist_head, list) {
	479	if (start >= m->addr &&
	480	start < m->addr + m->size)
	481	break;
	482	}
1da177e4	483	}
1da177e4	484
4fd2c20d	485	if (&m->list == &kclist_head) {
3673fb08 OS	486	if (clear_user(buffer, tsz)) {
	487	ret = -EFAULT;
	488	goto out;
	489	}
a1b3d2f2	490	m = NULL; /* skip the list anchor */
ffc8599a KS	491	} else if (!pfn_is_ram(__pa(start) >> PAGE_SHIFT)) {
	492	if (clear_user(buffer, tsz)) {
	493	ret = -EFAULT;
	494	goto out;
	495	}
737326aa	496	} else if (m->type == KCORE_VMALLOC) {
f5beeb18	497	vread(buf, (char *)start, tsz);
73d7c33e	498	/* we have to zero-fill user buffer even if no read */
3673fb08 OS	499	if (copy_to_user(buffer, buf, tsz)) {
	500	ret = -EFAULT;
	501	goto out;
	502	}
595dd46e JZ	503	} else if (m->type == KCORE_USER) {
595dd46e JZ	504	/* User page is handled prior to normal kernel page: */
3673fb08 OS	505	if (copy_to_user(buffer, (char *)start, tsz)) {
	506	ret = -EFAULT;
	507	goto out;
	508	}
1da177e4 LT	509	} else {
1da177e4 LT	510	if (kern_addr_valid(start)) {
df04abfd JO	511	/*
	512	* Using bounce buffer to bypass the
	513	* hardened user copy kernel text checks.
	514	*/
d0290bc2	515	if (probe_kernel_read(buf, (void *) start, tsz)) {
3673fb08 OS	516	if (clear_user(buffer, tsz)) {
	517	ret = -EFAULT;
	518	goto out;
	519	}
d0290bc2	520	} else {
3673fb08 OS	521	if (copy_to_user(buffer, buf, tsz)) {
	522	ret = -EFAULT;
	523	goto out;
	524	}
1da177e4 LT	525	}
1da177e4 LT	526	} else {
3673fb08 OS	527	if (clear_user(buffer, tsz)) {
	528	ret = -EFAULT;
	529	goto out;
	530	}
1da177e4 LT	531	}
	532	}
	533	buflen -= tsz;
	534	*fpos += tsz;
	535	buffer += tsz;
1da177e4 LT	536	start += tsz;
	537	tsz = (buflen > PAGE_SIZE ? PAGE_SIZE : buflen);
	538	}
	539
3673fb08 OS	540	out:
	541	up_read(&kclist_lock);
	542	if (ret)
	543	return ret;
	544	return orig_buflen - buflen;
1da177e4	545	}
97ce5d6d	546
3089aa1b KH	547	static int open_kcore(struct inode inode, struct file filp)
3089aa1b KH	548	{
02e935bf DH	549	int ret = security_locked_down(LOCKDOWN_KCORE);
02e935bf DH	550
3089aa1b KH	551	if (!capable(CAP_SYS_RAWIO))
3089aa1b KH	552	return -EPERM;
f5beeb18	553
b602614a MG	554	if (ret)
	555	return ret;
	556
f5beeb18 JO	557	filp->private_data = kmalloc(PAGE_SIZE, GFP_KERNEL);
	558	if (!filp->private_data)
	559	return -ENOMEM;
	560
3089aa1b KH	561	if (kcore_need_update)
3089aa1b KH	562	kcore_update_ram();
0d4c36a9	563	if (i_size_read(inode) != proc_root_kcore->size) {
5955102c	564	inode_lock(inode);
0d4c36a9	565	i_size_write(inode, proc_root_kcore->size);
5955102c	566	inode_unlock(inode);
0d4c36a9	567	}
3089aa1b KH	568	return 0;
	569	}
	570
f5beeb18 JO	571	static int release_kcore(struct inode inode, struct file file)
	572	{
	573	kfree(file->private_data);
	574	return 0;
	575	}
3089aa1b	576
97a32539 AD	577	static const struct proc_ops kcore_proc_ops = {
	578	.proc_read = read_kcore,
	579	.proc_open = open_kcore,
	580	.proc_release = release_kcore,
	581	.proc_lseek = default_llseek,
3089aa1b KH	582	};
3089aa1b KH	583
3089aa1b KH	584	/* just remember that we have to update kcore */
	585	static int __meminit kcore_callback(struct notifier_block *self,
	586	unsigned long action, void *arg)
	587	{
	588	switch (action) {
	589	case MEM_ONLINE:
	590	case MEM_OFFLINE:
3089aa1b	591	kcore_need_update = 1;
bf531831	592	break;
3089aa1b KH	593	}
	594	return NOTIFY_OK;
	595	}
3089aa1b	596
3c743a7f AM	597	static struct notifier_block kcore_callback_nb __meminitdata = {
	598	.notifier_call = kcore_callback,
	599	.priority = 0,
	600	};
3089aa1b	601
a0614da8 KH	602	static struct kcore_list kcore_vmalloc;
a0614da8 KH	603
9492587c KH	604	#ifdef CONFIG_ARCH_PROC_KCORE_TEXT
	605	static struct kcore_list kcore_text;
	606	/*
	607	* If defined, special segment is used for mapping kernel text instead of
	608	* direct-map area. We need to create special TEXT section.
	609	*/
	610	static void __init proc_kcore_text_init(void)
	611	{
36e15263	612	kclist_add(&kcore_text, _text, _end - _text, KCORE_TEXT);
9492587c KH	613	}
	614	#else
	615	static void __init proc_kcore_text_init(void)
	616	{
	617	}
	618	#endif
	619
81ac3ad9 KH	620	#if defined(CONFIG_MODULES) && defined(MODULES_VADDR)
	621	/*
	622	* MODULES_VADDR has no intersection with VMALLOC_ADDR.
	623	*/
eebf3648	624	static struct kcore_list kcore_modules;
81ac3ad9 KH	625	static void __init add_modules_range(void)
81ac3ad9 KH	626	{
bf3e2692 BH	627	if (MODULES_VADDR != VMALLOC_START && MODULES_END != VMALLOC_END) {
bf3e2692 BH	628	kclist_add(&kcore_modules, (void *)MODULES_VADDR,
81ac3ad9	629	MODULES_END - MODULES_VADDR, KCORE_VMALLOC);
bf3e2692	630	}
81ac3ad9 KH	631	}
	632	#else
	633	static void __init add_modules_range(void)
	634	{
	635	}
	636	#endif
	637
97ce5d6d AD	638	static int __init proc_kcore_init(void)
97ce5d6d AD	639	{
97a32539	640	proc_root_kcore = proc_create("kcore", S_IRUSR, NULL, &kcore_proc_ops);
90396f96	641	if (!proc_root_kcore) {
87ebdc00	642	pr_err("couldn't create /proc/kcore\n");
90396f96 KH	643	return 0; /* Always returns 0. */
90396f96 KH	644	}
3089aa1b	645	/* Store text area if it's special */
9492587c	646	proc_kcore_text_init();
3089aa1b	647	/* Store vmalloc area */
a0614da8 KH	648	kclist_add(&kcore_vmalloc, (void *)VMALLOC_START,
a0614da8 KH	649	VMALLOC_END - VMALLOC_START, KCORE_VMALLOC);
81ac3ad9	650	add_modules_range();
3089aa1b KH	651	/* Store direct-map area from physical memory map */
3089aa1b KH	652	kcore_update_ram();
3c743a7f	653	register_hotmemory_notifier(&kcore_callback_nb);
3089aa1b	654
97ce5d6d AD	655	return 0;
97ce5d6d AD	656	}
abaf3787	657	fs_initcall(proc_kcore_init);