[linux-2.6-block.git] / arch / s390 / kernel / vdso.c

/*
 * vdso setup for s390
 *
 *  Copyright IBM Corp. 2008
 *  Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License (version 2 only)
 * as published by the Free Software Foundation.
 */

#include <linux/module.h>
#include <linux/errno.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/smp.h>
#include <linux/stddef.h>
#include <linux/unistd.h>
#include <linux/slab.h>
#include <linux/user.h>
#include <linux/elf.h>
#include <linux/security.h>
#include <linux/bootmem.h>
#include <linux/compat.h>
#include <asm/asm-offsets.h>
#include <asm/pgtable.h>
#include <asm/processor.h>
#include <asm/mmu.h>
#include <asm/mmu_context.h>
#include <asm/sections.h>
#include <asm/vdso.h>
#include <asm/facility.h>

#ifdef CONFIG_COMPAT
extern char vdso32_start, vdso32_end;
static void *vdso32_kbase = &vdso32_start;
static unsigned int vdso32_pages;
static struct page **vdso32_pagelist;
#endif

extern char vdso64_start, vdso64_end;
static void *vdso64_kbase = &vdso64_start;
static unsigned int vdso64_pages;
static struct page **vdso64_pagelist;

/*
 * Should the kernel map a VDSO page into processes and pass its
 * address down to glibc upon exec()?
 */
unsigned int __read_mostly vdso_enabled = 1;

static int __init vdso_setup(char *s)
{
	unsigned long val;
	int rc;

	rc = 0;
	if (strncmp(s, "on", 3) == 0)
		vdso_enabled = 1;
	else if (strncmp(s, "off", 4) == 0)
		vdso_enabled = 0;
	else {
		rc = kstrtoul(s, 0, &val);
		vdso_enabled = rc ? 0 : !!val;
	}
	return !rc;
}
__setup("vdso=", vdso_setup);

/*
 * The vdso data page
 */
static union {
	struct vdso_data	data;
	u8			page[PAGE_SIZE];
} vdso_data_store __page_aligned_data;
struct vdso_data *vdso_data = &vdso_data_store.data;

/*
 * Setup vdso data page.
 */
static void vdso_init_data(struct vdso_data *vd)
{
	vd->ectg_available = test_facility(31);
}

/*
 * Allocate/free per cpu vdso data.
 */
#define SEGMENT_ORDER	2

int vdso_alloc_per_cpu(struct _lowcore *lowcore)
{
	unsigned long segment_table, page_table, page_frame;
	u32 *psal, *aste;
	int i;

	lowcore->vdso_per_cpu_data = __LC_PASTE;

	if (!vdso_enabled)
		return 0;

	segment_table = __get_free_pages(GFP_KERNEL, SEGMENT_ORDER);
	page_table = get_zeroed_page(GFP_KERNEL | GFP_DMA);
	page_frame = get_zeroed_page(GFP_KERNEL);
	if (!segment_table || !page_table || !page_frame)
		goto out;

	clear_table((unsigned long *) segment_table, _SEGMENT_ENTRY_EMPTY,
		    PAGE_SIZE << SEGMENT_ORDER);
	clear_table((unsigned long *) page_table, _PAGE_INVALID,
		    256*sizeof(unsigned long));

	*(unsigned long *) segment_table = _SEGMENT_ENTRY + page_table;
	*(unsigned long *) page_table = _PAGE_PROTECT + page_frame;

	psal = (u32 *) (page_table + 256*sizeof(unsigned long));
	aste = psal + 32;

	for (i = 4; i < 32; i += 4)
		psal[i] = 0x80000000;

	lowcore->paste[4] = (u32)(addr_t) psal;
	psal[0] = 0x02000000;
	psal[2] = (u32)(addr_t) aste;
	*(unsigned long *) (aste + 2) = segment_table +
		_ASCE_TABLE_LENGTH + _ASCE_USER_BITS + _ASCE_TYPE_SEGMENT;
	aste[4] = (u32)(addr_t) psal;
	lowcore->vdso_per_cpu_data = page_frame;

	return 0;

out:
	free_page(page_frame);
	free_page(page_table);
	free_pages(segment_table, SEGMENT_ORDER);
	return -ENOMEM;
}

void vdso_free_per_cpu(struct _lowcore *lowcore)
{
	unsigned long segment_table, page_table, page_frame;
	u32 *psal, *aste;

	if (!vdso_enabled)
		return;

	psal = (u32 *)(addr_t) lowcore->paste[4];
	aste = (u32 *)(addr_t) psal[2];
	segment_table = *(unsigned long *)(aste + 2) & PAGE_MASK;
	page_table = *(unsigned long *) segment_table;
	page_frame = *(unsigned long *) page_table;

	free_page(page_frame);
	free_page(page_table);
	free_pages(segment_table, SEGMENT_ORDER);
}

static void vdso_init_cr5(void)
{
	unsigned long cr5;

	if (!vdso_enabled)
		return;
	cr5 = offsetof(struct _lowcore, paste);
	__ctl_load(cr5, 5, 5);
}

/*
 * This is called from binfmt_elf, we create the special vma for the
 * vDSO and insert it into the mm struct tree
 */
int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
{
	struct mm_struct *mm = current->mm;
	struct page **vdso_pagelist;
	unsigned long vdso_pages;
	unsigned long vdso_base;
	int rc;

	if (!vdso_enabled)
		return 0;
	/*
	 * Only map the vdso for dynamically linked elf binaries.
	 */
	if (!uses_interp)
		return 0;

	vdso_pagelist = vdso64_pagelist;
	vdso_pages = vdso64_pages;
#ifdef CONFIG_COMPAT
	if (is_compat_task()) {
		vdso_pagelist = vdso32_pagelist;
		vdso_pages = vdso32_pages;
	}
#endif
	/*
	 * vDSO has a problem and was disabled, just don't "enable" it for
	 * the process
	 */
	if (vdso_pages == 0)
		return 0;

	current->mm->context.vdso_base = 0;

	/*
	 * pick a base address for the vDSO in process space. We try to put
	 * it at vdso_base which is the "natural" base for it, but we might
	 * fail and end up putting it elsewhere.
	 */
	down_write(&mm->mmap_sem);
	vdso_base = get_unmapped_area(NULL, 0, vdso_pages << PAGE_SHIFT, 0, 0);
	if (IS_ERR_VALUE(vdso_base)) {
		rc = vdso_base;
		goto out_up;
	}

	/*
	 * Put vDSO base into mm struct. We need to do this before calling
	 * install_special_mapping or the perf counter mmap tracking code
	 * will fail to recognise it as a vDSO (since arch_vma_name fails).
	 */
	current->mm->context.vdso_base = vdso_base;

	/*
	 * our vma flags don't have VM_WRITE so by default, the process
	 * isn't allowed to write those pages.
	 * gdb can break that with ptrace interface, and thus trigger COW
	 * on those pages but it's then your responsibility to never do that
	 * on the "data" page of the vDSO or you'll stop getting kernel
	 * updates and your nice userland gettimeofday will be totally dead.
	 * It's fine to use that for setting breakpoints in the vDSO code
	 * pages though.
	 */
	rc = install_special_mapping(mm, vdso_base, vdso_pages << PAGE_SHIFT,
				     VM_READ|VM_EXEC|
				     VM_MAYREAD|VM_MAYWRITE|VM_MAYEXEC,
				     vdso_pagelist);
	if (rc)
		current->mm->context.vdso_base = 0;
out_up:
	up_write(&mm->mmap_sem);
	return rc;
}

const char *arch_vma_name(struct vm_area_struct *vma)
{
	if (vma->vm_mm && vma->vm_start == vma->vm_mm->context.vdso_base)
		return "[vdso]";
	return NULL;
}

static int __init vdso_init(void)
{
	int i;

	if (!vdso_enabled)
		return 0;
	vdso_init_data(vdso_data);
#ifdef CONFIG_COMPAT
	/* Calculate the size of the 32 bit vDSO */
	vdso32_pages = ((&vdso32_end - &vdso32_start
			 + PAGE_SIZE - 1) >> PAGE_SHIFT) + 1;

	/* Make sure pages are in the correct state */
	vdso32_pagelist = kzalloc(sizeof(struct page *) * (vdso32_pages + 1),
				  GFP_KERNEL);
	BUG_ON(vdso32_pagelist == NULL);
	for (i = 0; i < vdso32_pages - 1; i++) {
		struct page *pg = virt_to_page(vdso32_kbase + i*PAGE_SIZE);
		ClearPageReserved(pg);
		get_page(pg);
		vdso32_pagelist[i] = pg;
	}
	vdso32_pagelist[vdso32_pages - 1] = virt_to_page(vdso_data);
	vdso32_pagelist[vdso32_pages] = NULL;
#endif

	/* Calculate the size of the 64 bit vDSO */
	vdso64_pages = ((&vdso64_end - &vdso64_start
			 + PAGE_SIZE - 1) >> PAGE_SHIFT) + 1;

	/* Make sure pages are in the correct state */
	vdso64_pagelist = kzalloc(sizeof(struct page *) * (vdso64_pages + 1),
				  GFP_KERNEL);
	BUG_ON(vdso64_pagelist == NULL);
	for (i = 0; i < vdso64_pages - 1; i++) {
		struct page *pg = virt_to_page(vdso64_kbase + i*PAGE_SIZE);
		ClearPageReserved(pg);
		get_page(pg);
		vdso64_pagelist[i] = pg;
	}
	vdso64_pagelist[vdso64_pages - 1] = virt_to_page(vdso_data);
	vdso64_pagelist[vdso64_pages] = NULL;
	if (vdso_alloc_per_cpu(&S390_lowcore))
		BUG();
	vdso_init_cr5();

	get_page(virt_to_page(vdso_data));

	smp_wmb();

	return 0;
}
early_initcall(vdso_init);
Commit	Line	Data
b020632e MS	1	/*
	2	* vdso setup for s390
	3	*
	4	* Copyright IBM Corp. 2008
	5	* Author(s): Martin Schwidefsky (schwidefsky@de.ibm.com)
	6	*
	7	* This program is free software; you can redistribute it and/or modify
	8	* it under the terms of the GNU General Public License (version 2 only)
	9	* as published by the Free Software Foundation.
	10	*/
	11
	12	#include <linux/module.h>
	13	#include <linux/errno.h>
	14	#include <linux/sched.h>
	15	#include <linux/kernel.h>
	16	#include <linux/mm.h>
	17	#include <linux/smp.h>
	18	#include <linux/stddef.h>
	19	#include <linux/unistd.h>
	20	#include <linux/slab.h>
	21	#include <linux/user.h>
	22	#include <linux/elf.h>
	23	#include <linux/security.h>
	24	#include <linux/bootmem.h>
7757591a	25	#include <linux/compat.h>
cbb870c8	26	#include <asm/asm-offsets.h>
b020632e	27	#include <asm/pgtable.h>
b020632e MS	28	#include <asm/processor.h>
	29	#include <asm/mmu.h>
	30	#include <asm/mmu_context.h>
	31	#include <asm/sections.h>
	32	#include <asm/vdso.h>
a0616cde	33	#include <asm/facility.h>
b020632e	34
5a79859a	35	#ifdef CONFIG_COMPAT
b020632e MS	36	extern char vdso32_start, vdso32_end;
	37	static void *vdso32_kbase = &vdso32_start;
	38	static unsigned int vdso32_pages;
	39	static struct page **vdso32_pagelist;
	40	#endif
	41
b020632e MS	42	extern char vdso64_start, vdso64_end;
	43	static void *vdso64_kbase = &vdso64_start;
	44	static unsigned int vdso64_pages;
	45	static struct page **vdso64_pagelist;
b020632e MS	46
	47	/*
	48	* Should the kernel map a VDSO page into processes and pass its
	49	* address down to glibc upon exec()?
	50	*/
	51	unsigned int __read_mostly vdso_enabled = 1;
	52
	53	static int __init vdso_setup(char *s)
	54	{
7aa79f94 MS	55	unsigned long val;
	56	int rc;
	57
	58	rc = 0;
	59	if (strncmp(s, "on", 3) == 0)
	60	vdso_enabled = 1;
	61	else if (strncmp(s, "off", 4) == 0)
	62	vdso_enabled = 0;
	63	else {
958d9072	64	rc = kstrtoul(s, 0, &val);
7aa79f94 MS	65	vdso_enabled = rc ? 0 : !!val;
	66	}
	67	return !rc;
b020632e MS	68	}
	69	__setup("vdso=", vdso_setup);
	70
	71	/*
	72	* The vdso data page
	73	*/
	74	static union {
	75	struct vdso_data data;
	76	u8 page[PAGE_SIZE];
abe1ee3a	77	} vdso_data_store __page_aligned_data;
b020632e MS	78	struct vdso_data *vdso_data = &vdso_data_store.data;
b020632e MS	79
c742b31c MS	80	/*
	81	* Setup vdso data page.
	82	*/
	83	static void vdso_init_data(struct vdso_data *vd)
	84	{
e258d719	85	vd->ectg_available = test_facility(31);
c742b31c MS	86	}
c742b31c MS	87
c742b31c MS	88	/*
	89	* Allocate/free per cpu vdso data.
	90	*/
c742b31c	91	#define SEGMENT_ORDER 2
c742b31c	92
8b646bd7	93	int vdso_alloc_per_cpu(struct _lowcore *lowcore)
c742b31c MS	94	{
	95	unsigned long segment_table, page_table, page_frame;
	96	u32 psal, aste;
	97	int i;
	98
	99	lowcore->vdso_per_cpu_data = __LC_PASTE;
	100
e258d719	101	if (!vdso_enabled)
c742b31c MS	102	return 0;
	103
	104	segment_table = __get_free_pages(GFP_KERNEL, SEGMENT_ORDER);
	105	page_table = get_zeroed_page(GFP_KERNEL \| GFP_DMA);
	106	page_frame = get_zeroed_page(GFP_KERNEL);
	107	if (!segment_table \|\| !page_table \|\| !page_frame)
	108	goto out;
	109
	110	clear_table((unsigned long *) segment_table, _SEGMENT_ENTRY_EMPTY,
	111	PAGE_SIZE << SEGMENT_ORDER);
e5098611	112	clear_table((unsigned long *) page_table, _PAGE_INVALID,
c742b31c MS	113	256*sizeof(unsigned long));
	114
	115	(unsigned long ) segment_table = _SEGMENT_ENTRY + page_table;
e5098611	116	(unsigned long ) page_table = _PAGE_PROTECT + page_frame;
c742b31c MS	117
	118	psal = (u32 ) (page_table + 256sizeof(unsigned long));
	119	aste = psal + 32;
	120
	121	for (i = 4; i < 32; i += 4)
	122	psal[i] = 0x80000000;
	123
	124	lowcore->paste[4] = (u32)(addr_t) psal;
06aae683	125	psal[0] = 0x02000000;
c742b31c MS	126	psal[2] = (u32)(addr_t) aste;
	127	(unsigned long ) (aste + 2) = segment_table +
	128	_ASCE_TABLE_LENGTH + _ASCE_USER_BITS + _ASCE_TYPE_SEGMENT;
	129	aste[4] = (u32)(addr_t) psal;
	130	lowcore->vdso_per_cpu_data = page_frame;
	131
c742b31c MS	132	return 0;
	133
	134	out:
	135	free_page(page_frame);
	136	free_page(page_table);
	137	free_pages(segment_table, SEGMENT_ORDER);
	138	return -ENOMEM;
	139	}
	140
8b646bd7	141	void vdso_free_per_cpu(struct _lowcore *lowcore)
c742b31c MS	142	{
	143	unsigned long segment_table, page_table, page_frame;
	144	u32 psal, aste;
	145
e258d719	146	if (!vdso_enabled)
c742b31c MS	147	return;
	148
	149	psal = (u32 *)(addr_t) lowcore->paste[4];
	150	aste = (u32 *)(addr_t) psal[2];
	151	segment_table = (unsigned long )(aste + 2) & PAGE_MASK;
	152	page_table = (unsigned long ) segment_table;
	153	page_frame = (unsigned long ) page_table;
	154
	155	free_page(page_frame);
	156	free_page(page_table);
	157	free_pages(segment_table, SEGMENT_ORDER);
	158	}
c742b31c	159
8b646bd7	160	static void vdso_init_cr5(void)
c742b31c MS	161	{
	162	unsigned long cr5;
	163
e258d719	164	if (!vdso_enabled)
8b646bd7	165	return;
c742b31c MS	166	cr5 = offsetof(struct _lowcore, paste);
	167	__ctl_load(cr5, 5, 5);
	168	}
c742b31c	169
b020632e MS	170	/*
	171	* This is called from binfmt_elf, we create the special vma for the
	172	* vDSO and insert it into the mm struct tree
	173	*/
	174	int arch_setup_additional_pages(struct linux_binprm *bprm, int uses_interp)
	175	{
	176	struct mm_struct *mm = current->mm;
	177	struct page **vdso_pagelist;
	178	unsigned long vdso_pages;
	179	unsigned long vdso_base;
	180	int rc;
	181
	182	if (!vdso_enabled)
	183	return 0;
	184	/*
	185	* Only map the vdso for dynamically linked elf binaries.
	186	*/
	187	if (!uses_interp)
	188	return 0;
	189
b020632e MS	190	vdso_pagelist = vdso64_pagelist;
	191	vdso_pages = vdso64_pages;
	192	#ifdef CONFIG_COMPAT
7757591a	193	if (is_compat_task()) {
b020632e MS	194	vdso_pagelist = vdso32_pagelist;
	195	vdso_pages = vdso32_pages;
	196	}
	197	#endif
b020632e MS	198	/*
	199	* vDSO has a problem and was disabled, just don't "enable" it for
	200	* the process
	201	*/
	202	if (vdso_pages == 0)
	203	return 0;
	204
	205	current->mm->context.vdso_base = 0;
	206
	207	/*
	208	* pick a base address for the vDSO in process space. We try to put
	209	* it at vdso_base which is the "natural" base for it, but we might
	210	* fail and end up putting it elsewhere.
	211	*/
	212	down_write(&mm->mmap_sem);
e7828bbd	213	vdso_base = get_unmapped_area(NULL, 0, vdso_pages << PAGE_SHIFT, 0, 0);
b020632e MS	214	if (IS_ERR_VALUE(vdso_base)) {
	215	rc = vdso_base;
	216	goto out_up;
	217	}
	218
930e44fb HC	219	/*
	220	* Put vDSO base into mm struct. We need to do this before calling
	221	* install_special_mapping or the perf counter mmap tracking code
	222	* will fail to recognise it as a vDSO (since arch_vma_name fails).
	223	*/
	224	current->mm->context.vdso_base = vdso_base;
	225
b020632e MS	226	/*
	227	* our vma flags don't have VM_WRITE so by default, the process
	228	* isn't allowed to write those pages.
	229	* gdb can break that with ptrace interface, and thus trigger COW
	230	* on those pages but it's then your responsibility to never do that
	231	* on the "data" page of the vDSO or you'll stop getting kernel
	232	* updates and your nice userland gettimeofday will be totally dead.
	233	* It's fine to use that for setting breakpoints in the vDSO code
909af768	234	* pages though.
b020632e MS	235	*/
	236	rc = install_special_mapping(mm, vdso_base, vdso_pages << PAGE_SHIFT,
	237	VM_READ\|VM_EXEC\|
909af768	238	VM_MAYREAD\|VM_MAYWRITE\|VM_MAYEXEC,
b020632e MS	239	vdso_pagelist);
b020632e MS	240	if (rc)
930e44fb	241	current->mm->context.vdso_base = 0;
b020632e MS	242	out_up:
	243	up_write(&mm->mmap_sem);
	244	return rc;
	245	}
	246
	247	const char arch_vma_name(struct vm_area_struct vma)
	248	{
	249	if (vma->vm_mm && vma->vm_start == vma->vm_mm->context.vdso_base)
	250	return "[vdso]";
	251	return NULL;
	252	}
	253
	254	static int __init vdso_init(void)
	255	{
	256	int i;
	257
c742b31c MS	258	if (!vdso_enabled)
	259	return 0;
	260	vdso_init_data(vdso_data);
5a79859a	261	#ifdef CONFIG_COMPAT
b020632e MS	262	/* Calculate the size of the 32 bit vDSO */
	263	vdso32_pages = ((&vdso32_end - &vdso32_start
	264	+ PAGE_SIZE - 1) >> PAGE_SHIFT) + 1;
	265
	266	/* Make sure pages are in the correct state */
	267	vdso32_pagelist = kzalloc(sizeof(struct page ) (vdso32_pages + 1),
	268	GFP_KERNEL);
	269	BUG_ON(vdso32_pagelist == NULL);
	270	for (i = 0; i < vdso32_pages - 1; i++) {
	271	struct page pg = virt_to_page(vdso32_kbase + iPAGE_SIZE);
	272	ClearPageReserved(pg);
	273	get_page(pg);
	274	vdso32_pagelist[i] = pg;
	275	}
	276	vdso32_pagelist[vdso32_pages - 1] = virt_to_page(vdso_data);
	277	vdso32_pagelist[vdso32_pages] = NULL;
	278	#endif
	279
b020632e MS	280	/* Calculate the size of the 64 bit vDSO */
	281	vdso64_pages = ((&vdso64_end - &vdso64_start
	282	+ PAGE_SIZE - 1) >> PAGE_SHIFT) + 1;
	283
	284	/* Make sure pages are in the correct state */
	285	vdso64_pagelist = kzalloc(sizeof(struct page ) (vdso64_pages + 1),
	286	GFP_KERNEL);
	287	BUG_ON(vdso64_pagelist == NULL);
	288	for (i = 0; i < vdso64_pages - 1; i++) {
	289	struct page pg = virt_to_page(vdso64_kbase + iPAGE_SIZE);
	290	ClearPageReserved(pg);
	291	get_page(pg);
	292	vdso64_pagelist[i] = pg;
	293	}
	294	vdso64_pagelist[vdso64_pages - 1] = virt_to_page(vdso_data);
	295	vdso64_pagelist[vdso64_pages] = NULL;
8b646bd7	296	if (vdso_alloc_per_cpu(&S390_lowcore))
81ffa041	297	BUG();
c742b31c	298	vdso_init_cr5();
b020632e MS	299
	300	get_page(virt_to_page(vdso_data));
	301
	302	smp_wmb();
	303
	304	return 0;
	305	}
8b646bd7	306	early_initcall(vdso_init);