[linux-2.6-block.git] / arch / loongarch / kernel / vdso.c

// SPDX-License-Identifier: GPL-2.0
/*
 * Author: Huacai Chen <chenhuacai@loongson.cn>
 * Copyright (C) 2020-2022 Loongson Technology Corporation Limited
 */

#include <linux/binfmts.h>
#include <linux/elf.h>
#include <linux/err.h>
#include <linux/init.h>
#include <linux/ioport.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/random.h>
#include <linux/sched.h>
#include <linux/slab.h>
#include <linux/time_namespace.h>
#include <linux/timekeeper_internal.h>

#include <asm/page.h>
#include <asm/vdso.h>
#include <vdso/helpers.h>
#include <vdso/vsyscall.h>
#include <vdso/datapage.h>
#include <generated/vdso-offsets.h>

extern char vdso_start[], vdso_end[];

/* Kernel-provided data used by the VDSO. */
static union vdso_data_store generic_vdso_data __page_aligned_data;

static union {
	u8 page[LOONGARCH_VDSO_DATA_SIZE];
	struct loongarch_vdso_data vdata;
} loongarch_vdso_data __page_aligned_data;

static struct page *vdso_pages[] = { NULL };
struct vdso_data *vdso_data = generic_vdso_data.data;
struct vdso_pcpu_data *vdso_pdata = loongarch_vdso_data.vdata.pdata;

static int vdso_mremap(const struct vm_special_mapping *sm, struct vm_area_struct *new_vma)
{
	current->mm->context.vdso = (void *)(new_vma->vm_start);

	return 0;
}

static vm_fault_t vvar_fault(const struct vm_special_mapping *sm,
			     struct vm_area_struct *vma, struct vm_fault *vmf)
{
	unsigned long pfn;
	struct page *timens_page = find_timens_vvar_page(vma);

	switch (vmf->pgoff) {
	case VVAR_GENERIC_PAGE_OFFSET:
		if (!timens_page)
			pfn = sym_to_pfn(vdso_data);
		else
			pfn = page_to_pfn(timens_page);
		break;
#ifdef CONFIG_TIME_NS
	case VVAR_TIMENS_PAGE_OFFSET:
		/*
		 * If a task belongs to a time namespace then a namespace specific
		 * VVAR is mapped with the VVAR_GENERIC_PAGE_OFFSET and the real
		 * VVAR page is mapped with the VVAR_TIMENS_PAGE_OFFSET offset.
		 * See also the comment near timens_setup_vdso_data().
		 */
		if (!timens_page)
			return VM_FAULT_SIGBUS;
		else
			pfn = sym_to_pfn(vdso_data);
		break;
#endif /* CONFIG_TIME_NS */
	case VVAR_LOONGARCH_PAGES_START ... VVAR_LOONGARCH_PAGES_END:
		pfn = sym_to_pfn(&loongarch_vdso_data) + vmf->pgoff - VVAR_LOONGARCH_PAGES_START;
		break;
	default:
		return VM_FAULT_SIGBUS;
	}

	return vmf_insert_pfn(vma, vmf->address, pfn);
}

struct loongarch_vdso_info vdso_info = {
	.vdso = vdso_start,
	.size = PAGE_SIZE,
	.code_mapping = {
		.name = "[vdso]",
		.pages = vdso_pages,
		.mremap = vdso_mremap,
	},
	.data_mapping = {
		.name = "[vvar]",
		.fault = vvar_fault,
	},
	.offset_sigreturn = vdso_offset_sigreturn,
};

static int __init init_vdso(void)
{
	unsigned long i, cpu, pfn;

	BUG_ON(!PAGE_ALIGNED(vdso_info.vdso));
	BUG_ON(!PAGE_ALIGNED(vdso_info.size));

	for_each_possible_cpu(cpu)
		vdso_pdata[cpu].node = cpu_to_node(cpu);

	pfn = __phys_to_pfn(__pa_symbol(vdso_info.vdso));
	for (i = 0; i < vdso_info.size / PAGE_SIZE; i++)
		vdso_info.code_mapping.pages[i] = pfn_to_page(pfn + i);

	return 0;
}
subsys_initcall(init_vdso);

#ifdef CONFIG_TIME_NS
struct vdso_data *arch_get_vdso_data(void *vvar_page)
{
	return (struct vdso_data *)(vvar_page);
}

/*
 * The vvar mapping contains data for a specific time namespace, so when a
 * task changes namespace we must unmap its vvar data for the old namespace.
 * Subsequent faults will map in data for the new namespace.
 *
 * For more details see timens_setup_vdso_data().
 */
int vdso_join_timens(struct task_struct *task, struct time_namespace *ns)
{
	struct mm_struct *mm = task->mm;
	struct vm_area_struct *vma;

	VMA_ITERATOR(vmi, mm, 0);

	mmap_read_lock(mm);
	for_each_vma(vmi, vma) {
		if (vma_is_special_mapping(vma, &vdso_info.data_mapping))
			zap_vma_pages(vma);
	}
	mmap_read_unlock(mm);

	return 0;
}
#endif

static unsigned long vdso_base(void)
{
	unsigned long base = STACK_TOP;

	if (current->flags & PF_RANDOMIZE) {
		base += get_random_u32_below(VDSO_RANDOMIZE_SIZE);
		base = PAGE_ALIGN(base);
	}

	return base;
}

int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
{
	int ret;
	unsigned long size, data_addr, vdso_addr;
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma;
	struct loongarch_vdso_info *info = current->thread.vdso;

	if (mmap_write_lock_killable(mm))
		return -EINTR;

	/*
	 * Determine total area size. This includes the VDSO data itself
	 * and the data pages.
	 */
	size = VVAR_SIZE + info->size;

	data_addr = get_unmapped_area(NULL, vdso_base(), size, 0, 0);
	if (IS_ERR_VALUE(data_addr)) {
		ret = data_addr;
		goto out;
	}

	vma = _install_special_mapping(mm, data_addr, VVAR_SIZE,
				       VM_READ | VM_MAYREAD | VM_PFNMAP,
				       &info->data_mapping);
	if (IS_ERR(vma)) {
		ret = PTR_ERR(vma);
		goto out;
	}

	vdso_addr = data_addr + VVAR_SIZE;
	vma = _install_special_mapping(mm, vdso_addr, info->size,
				       VM_READ | VM_EXEC | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC,
				       &info->code_mapping);
	if (IS_ERR(vma)) {
		ret = PTR_ERR(vma);
		goto out;
	}

	mm->context.vdso = (void *)vdso_addr;
	ret = 0;

out:
	mmap_write_unlock(mm);
	return ret;
}
Commit	Line	Data
c6b99bed HC	1	// SPDX-License-Identifier: GPL-2.0
	2	/*
	3	* Author: Huacai Chen <chenhuacai@loongson.cn>
	4	* Copyright (C) 2020-2022 Loongson Technology Corporation Limited
	5	*/
	6
	7	#include <linux/binfmts.h>
	8	#include <linux/elf.h>
	9	#include <linux/err.h>
	10	#include <linux/init.h>
	11	#include <linux/ioport.h>
	12	#include <linux/kernel.h>
	13	#include <linux/mm.h>
	14	#include <linux/random.h>
	15	#include <linux/sched.h>
	16	#include <linux/slab.h>
aa5e65dc	17	#include <linux/time_namespace.h>
c6b99bed HC	18	#include <linux/timekeeper_internal.h>
	19
	20	#include <asm/page.h>
	21	#include <asm/vdso.h>
	22	#include <vdso/helpers.h>
	23	#include <vdso/vsyscall.h>
8d87d2cd	24	#include <vdso/datapage.h>
c6b99bed HC	25	#include <generated/vdso-offsets.h>
	26
	27	extern char vdso_start[], vdso_end[];
	28
	29	/* Kernel-provided data used by the VDSO. */
8d87d2cd	30	static union vdso_data_store generic_vdso_data __page_aligned_data;
aa5e65dc TY	31
	32	static union {
	33	u8 page[LOONGARCH_VDSO_DATA_SIZE];
dce6098b	34	struct loongarch_vdso_data vdata;
c6b99bed	35	} loongarch_vdso_data __page_aligned_data;
dce6098b	36
c6b99bed	37	static struct page *vdso_pages[] = { NULL };
aa5e65dc	38	struct vdso_data *vdso_data = generic_vdso_data.data;
dce6098b	39	struct vdso_pcpu_data *vdso_pdata = loongarch_vdso_data.vdata.pdata;
c6b99bed HC	40
	41	static int vdso_mremap(const struct vm_special_mapping sm, struct vm_area_struct new_vma)
	42	{
	43	current->mm->context.vdso = (void *)(new_vma->vm_start);
	44
	45	return 0;
	46	}
	47
aa5e65dc TY	48	static vm_fault_t vvar_fault(const struct vm_special_mapping *sm,
	49	struct vm_area_struct vma, struct vm_fault vmf)
	50	{
	51	unsigned long pfn;
	52	struct page *timens_page = find_timens_vvar_page(vma);
	53
	54	switch (vmf->pgoff) {
	55	case VVAR_GENERIC_PAGE_OFFSET:
	56	if (!timens_page)
	57	pfn = sym_to_pfn(vdso_data);
	58	else
	59	pfn = page_to_pfn(timens_page);
	60	break;
	61	#ifdef CONFIG_TIME_NS
	62	case VVAR_TIMENS_PAGE_OFFSET:
	63	/*
	64	* If a task belongs to a time namespace then a namespace specific
	65	* VVAR is mapped with the VVAR_GENERIC_PAGE_OFFSET and the real
	66	* VVAR page is mapped with the VVAR_TIMENS_PAGE_OFFSET offset.
	67	* See also the comment near timens_setup_vdso_data().
	68	*/
	69	if (!timens_page)
	70	return VM_FAULT_SIGBUS;
	71	else
	72	pfn = sym_to_pfn(vdso_data);
	73	break;
	74	#endif /* CONFIG_TIME_NS */
	75	case VVAR_LOONGARCH_PAGES_START ... VVAR_LOONGARCH_PAGES_END:
	76	pfn = sym_to_pfn(&loongarch_vdso_data) + vmf->pgoff - VVAR_LOONGARCH_PAGES_START;
	77	break;
	78	default:
	79	return VM_FAULT_SIGBUS;
	80	}
	81
	82	return vmf_insert_pfn(vma, vmf->address, pfn);
	83	}
	84
c6b99bed HC	85	struct loongarch_vdso_info vdso_info = {
	86	.vdso = vdso_start,
	87	.size = PAGE_SIZE,
	88	.code_mapping = {
	89	.name = "[vdso]",
	90	.pages = vdso_pages,
	91	.mremap = vdso_mremap,
	92	},
	93	.data_mapping = {
	94	.name = "[vvar]",
aa5e65dc	95	.fault = vvar_fault,
c6b99bed HC	96	},
	97	.offset_sigreturn = vdso_offset_sigreturn,
	98	};
	99
	100	static int __init init_vdso(void)
	101	{
dce6098b	102	unsigned long i, cpu, pfn;
c6b99bed HC	103
	104	BUG_ON(!PAGE_ALIGNED(vdso_info.vdso));
	105	BUG_ON(!PAGE_ALIGNED(vdso_info.size));
	106
dce6098b HC	107	for_each_possible_cpu(cpu)
	108	vdso_pdata[cpu].node = cpu_to_node(cpu);
	109
c6b99bed HC	110	pfn = __phys_to_pfn(__pa_symbol(vdso_info.vdso));
	111	for (i = 0; i < vdso_info.size / PAGE_SIZE; i++)
	112	vdso_info.code_mapping.pages[i] = pfn_to_page(pfn + i);
	113
	114	return 0;
	115	}
	116	subsys_initcall(init_vdso);
	117
aa5e65dc TY	118	#ifdef CONFIG_TIME_NS
	119	struct vdso_data arch_get_vdso_data(void vvar_page)
	120	{
	121	return (struct vdso_data *)(vvar_page);
	122	}
	123
	124	/*
	125	* The vvar mapping contains data for a specific time namespace, so when a
	126	* task changes namespace we must unmap its vvar data for the old namespace.
	127	* Subsequent faults will map in data for the new namespace.
	128	*
	129	* For more details see timens_setup_vdso_data().
	130	*/
	131	int vdso_join_timens(struct task_struct task, struct time_namespace ns)
	132	{
	133	struct mm_struct *mm = task->mm;
	134	struct vm_area_struct *vma;
	135
	136	VMA_ITERATOR(vmi, mm, 0);
	137
	138	mmap_read_lock(mm);
	139	for_each_vma(vmi, vma) {
	140	if (vma_is_special_mapping(vma, &vdso_info.data_mapping))
	141	zap_vma_pages(vma);
	142	}
	143	mmap_read_unlock(mm);
	144
	145	return 0;
	146	}
	147	#endif
	148
c6b99bed HC	149	static unsigned long vdso_base(void)
	150	{
	151	unsigned long base = STACK_TOP;
	152
	153	if (current->flags & PF_RANDOMIZE) {
8032bf12	154	base += get_random_u32_below(VDSO_RANDOMIZE_SIZE);
c6b99bed HC	155	base = PAGE_ALIGN(base);
	156	}
	157
	158	return base;
	159	}
	160
	161	int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
	162	{
	163	int ret;
aa5e65dc	164	unsigned long size, data_addr, vdso_addr;
c6b99bed HC	165	struct mm_struct *mm = current->mm;
	166	struct vm_area_struct *vma;
	167	struct loongarch_vdso_info *info = current->thread.vdso;
	168
	169	if (mmap_write_lock_killable(mm))
	170	return -EINTR;
	171
	172	/*
	173	* Determine total area size. This includes the VDSO data itself
dce6098b	174	* and the data pages.
c6b99bed	175	*/
aa5e65dc	176	size = VVAR_SIZE + info->size;
c6b99bed HC	177
	178	data_addr = get_unmapped_area(NULL, vdso_base(), size, 0, 0);
	179	if (IS_ERR_VALUE(data_addr)) {
	180	ret = data_addr;
	181	goto out;
	182	}
c6b99bed	183
aa5e65dc TY	184	vma = _install_special_mapping(mm, data_addr, VVAR_SIZE,
aa5e65dc TY	185	VM_READ \| VM_MAYREAD \| VM_PFNMAP,
c6b99bed HC	186	&info->data_mapping);
	187	if (IS_ERR(vma)) {
	188	ret = PTR_ERR(vma);
	189	goto out;
	190	}
	191
aa5e65dc	192	vdso_addr = data_addr + VVAR_SIZE;
c6b99bed HC	193	vma = _install_special_mapping(mm, vdso_addr, info->size,
	194	VM_READ \| VM_EXEC \| VM_MAYREAD \| VM_MAYWRITE \| VM_MAYEXEC,
	195	&info->code_mapping);
	196	if (IS_ERR(vma)) {
	197	ret = PTR_ERR(vma);
	198	goto out;
	199	}
	200
	201	mm->context.vdso = (void *)vdso_addr;
	202	ret = 0;
	203
	204	out:
	205	mmap_write_unlock(mm);
	206	return ret;
	207	}