[deliverable/linux.git] / arch / x86 / xen / setup.c

/*
 * Machine specific setup for xen
 *
 * Jeremy Fitzhardinge <jeremy@xensource.com>, XenSource Inc, 2007
 */

#include <linux/module.h>
#include <linux/sched.h>
#include <linux/mm.h>
#include <linux/pm.h>
#include <linux/memblock.h>

#include <asm/elf.h>
#include <asm/vdso.h>
#include <asm/e820.h>
#include <asm/setup.h>
#include <asm/acpi.h>
#include <asm/xen/hypervisor.h>
#include <asm/xen/hypercall.h>

#include <xen/xen.h>
#include <xen/page.h>
#include <xen/interface/callback.h>
#include <xen/interface/memory.h>
#include <xen/interface/physdev.h>
#include <xen/interface/memory.h>
#include <xen/features.h>

#include "xen-ops.h"
#include "vdso.h"

/* These are code, but not functions.  Defined in entry.S */
extern const char xen_hypervisor_callback[];
extern const char xen_failsafe_callback[];
extern void xen_sysenter_target(void);
extern void xen_syscall_target(void);
extern void xen_syscall32_target(void);

/* Amount of extra memory space we add to the e820 ranges */
phys_addr_t xen_extra_mem_start, xen_extra_mem_size;

/* 
 * The maximum amount of extra memory compared to the base size.  The
 * main scaling factor is the size of struct page.  At extreme ratios
 * of base:extra, all the base memory can be filled with page
 * structures for the extra memory, leaving no space for anything
 * else.
 * 
 * 10x seems like a reasonable balance between scaling flexibility and
 * leaving a practically usable system.
 */
#define EXTRA_MEM_RATIO		(10)

static __init void xen_add_extra_mem(unsigned long pages)
{
	u64 size = (u64)pages * PAGE_SIZE;
	u64 extra_start = xen_extra_mem_start + xen_extra_mem_size;

	if (!pages)
		return;

	e820_add_region(extra_start, size, E820_RAM);
	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);

	memblock_x86_reserve_range(extra_start, extra_start + size, "XEN EXTRA");

	xen_extra_mem_size += size;

	xen_max_p2m_pfn = PFN_DOWN(extra_start + size);
}

static unsigned long __init xen_release_chunk(phys_addr_t start_addr,
					      phys_addr_t end_addr)
{
	struct xen_memory_reservation reservation = {
		.address_bits = 0,
		.extent_order = 0,
		.domid        = DOMID_SELF
	};
	unsigned long start, end;
	unsigned long len = 0;
	unsigned long pfn;
	int ret;

	start = PFN_UP(start_addr);
	end = PFN_DOWN(end_addr);

	if (end <= start)
		return 0;

	printk(KERN_INFO "xen_release_chunk: looking at area pfn %lx-%lx: ",
	       start, end);
	for(pfn = start; pfn < end; pfn++) {
		unsigned long mfn = pfn_to_mfn(pfn);

		/* Make sure pfn exists to start with */
		if (mfn == INVALID_P2M_ENTRY || mfn_to_pfn(mfn) != pfn)
			continue;

		set_xen_guest_handle(reservation.extent_start, &mfn);
		reservation.nr_extents = 1;

		ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
					   &reservation);
		WARN(ret != 1, "Failed to release memory %lx-%lx err=%d\n",
		     start, end, ret);
		if (ret == 1) {
			set_phys_to_machine(pfn, INVALID_P2M_ENTRY);
			len++;
		}
	}
	printk(KERN_CONT "%ld pages freed\n", len);

	return len;
}

static unsigned long __init xen_return_unused_memory(unsigned long max_pfn,
						     const struct e820map *e820)
{
	phys_addr_t max_addr = PFN_PHYS(max_pfn);
	phys_addr_t last_end = 0;
	unsigned long released = 0;
	int i;

	for (i = 0; i < e820->nr_map && last_end < max_addr; i++) {
		phys_addr_t end = e820->map[i].addr;
		end = min(max_addr, end);

		released += xen_release_chunk(last_end, end);
		last_end = e820->map[i].addr + e820->map[i].size;
	}

	if (last_end < max_addr)
		released += xen_release_chunk(last_end, max_addr);

	printk(KERN_INFO "released %ld pages of unused memory\n", released);
	return released;
}

/**
 * machine_specific_memory_setup - Hook for machine specific memory setup.
 **/
char * __init xen_memory_setup(void)
{
	static struct e820entry map[E820MAX] __initdata;

	unsigned long max_pfn = xen_start_info->nr_pages;
	unsigned long long mem_end;
	int rc;
	struct xen_memory_map memmap;
	unsigned long extra_pages = 0;
	unsigned long extra_limit;
	int i;
	int op;

	max_pfn = min(MAX_DOMAIN_PAGES, max_pfn);
	mem_end = PFN_PHYS(max_pfn);

	memmap.nr_entries = E820MAX;
	set_xen_guest_handle(memmap.buffer, map);

	op = xen_initial_domain() ?
		XENMEM_machine_memory_map :
		XENMEM_memory_map;
	rc = HYPERVISOR_memory_op(op, &memmap);
	if (rc == -ENOSYS) {
		memmap.nr_entries = 1;
		map[0].addr = 0ULL;
		map[0].size = mem_end;
		/* 8MB slack (to balance backend allocations). */
		map[0].size += 8ULL << 20;
		map[0].type = E820_RAM;
		rc = 0;
	}
	BUG_ON(rc);

	e820.nr_map = 0;
	xen_extra_mem_start = mem_end;
	for (i = 0; i < memmap.nr_entries; i++) {
		unsigned long long end = map[i].addr + map[i].size;

		if (map[i].type == E820_RAM) {
			if (map[i].addr < mem_end && end > mem_end) {
				/* Truncate region to max_mem. */
				u64 delta = end - mem_end;

				map[i].size -= delta;
				extra_pages += PFN_DOWN(delta);

				end = mem_end;
			}
		}

		if (end > xen_extra_mem_start)
			xen_extra_mem_start = end;

		/* If region is non-RAM or below mem_end, add what remains */
		if ((map[i].type != E820_RAM || map[i].addr < mem_end) &&
		    map[i].size > 0)
			e820_add_region(map[i].addr, map[i].size, map[i].type);
	}

	/*
	 * Even though this is normal, usable memory under Xen, reserve
	 * ISA memory anyway because too many things think they can poke
	 * about in there.
	 *
	 * In a dom0 kernel, this region is identity mapped with the
	 * hardware ISA area, so it really is out of bounds.
	 */
	e820_add_region(ISA_START_ADDRESS, ISA_END_ADDRESS - ISA_START_ADDRESS,
			E820_RESERVED);

	/*
	 * Reserve Xen bits:
	 *  - mfn_list
	 *  - xen_start_info
	 * See comment above "struct start_info" in <xen/interface/xen.h>
	 */
	memblock_x86_reserve_range(__pa(xen_start_info->mfn_list),
		      __pa(xen_start_info->pt_base),
			"XEN START INFO");

	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);

	extra_pages += xen_return_unused_memory(xen_start_info->nr_pages, &e820);

	/*
	 * Clamp the amount of extra memory to a EXTRA_MEM_RATIO
	 * factor the base size.  On non-highmem systems, the base
	 * size is the full initial memory allocation; on highmem it
	 * is limited to the max size of lowmem, so that it doesn't
	 * get completely filled.
	 *
	 * In principle there could be a problem in lowmem systems if
	 * the initial memory is also very large with respect to
	 * lowmem, but we won't try to deal with that here.
	 */
	extra_limit = min(EXTRA_MEM_RATIO * min(max_pfn, PFN_DOWN(MAXMEM)),
			  max_pfn + extra_pages);

	if (extra_limit >= max_pfn)
		extra_pages = extra_limit - max_pfn;
	else
		extra_pages = 0;

	if (!xen_initial_domain())
		xen_add_extra_mem(extra_pages);

	return "Xen";
}

static void xen_idle(void)
{
	local_irq_disable();

	if (need_resched())
		local_irq_enable();
	else {
		current_thread_info()->status &= ~TS_POLLING;
		smp_mb__after_clear_bit();
		safe_halt();
		current_thread_info()->status |= TS_POLLING;
	}
}

/*
 * Set the bit indicating "nosegneg" library variants should be used.
 * We only need to bother in pure 32-bit mode; compat 32-bit processes
 * can have un-truncated segments, so wrapping around is allowed.
 */
static void __init fiddle_vdso(void)
{
#ifdef CONFIG_X86_32
	u32 *mask;
	mask = VDSO32_SYMBOL(&vdso32_int80_start, NOTE_MASK);
	*mask |= 1 << VDSO_NOTE_NONEGSEG_BIT;
	mask = VDSO32_SYMBOL(&vdso32_sysenter_start, NOTE_MASK);
	*mask |= 1 << VDSO_NOTE_NONEGSEG_BIT;
#endif
}

static __cpuinit int register_callback(unsigned type, const void *func)
{
	struct callback_register callback = {
		.type = type,
		.address = XEN_CALLBACK(__KERNEL_CS, func),
		.flags = CALLBACKF_mask_events,
	};

	return HYPERVISOR_callback_op(CALLBACKOP_register, &callback);
}

void __cpuinit xen_enable_sysenter(void)
{
	int ret;
	unsigned sysenter_feature;

#ifdef CONFIG_X86_32
	sysenter_feature = X86_FEATURE_SEP;
#else
	sysenter_feature = X86_FEATURE_SYSENTER32;
#endif

	if (!boot_cpu_has(sysenter_feature))
		return;

	ret = register_callback(CALLBACKTYPE_sysenter, xen_sysenter_target);
	if(ret != 0)
		setup_clear_cpu_cap(sysenter_feature);
}

void __cpuinit xen_enable_syscall(void)
{
#ifdef CONFIG_X86_64
	int ret;

	ret = register_callback(CALLBACKTYPE_syscall, xen_syscall_target);
	if (ret != 0) {
		printk(KERN_ERR "Failed to set syscall callback: %d\n", ret);
		/* Pretty fatal; 64-bit userspace has no other
		   mechanism for syscalls. */
	}

	if (boot_cpu_has(X86_FEATURE_SYSCALL32)) {
		ret = register_callback(CALLBACKTYPE_syscall32,
					xen_syscall32_target);
		if (ret != 0)
			setup_clear_cpu_cap(X86_FEATURE_SYSCALL32);
	}
#endif /* CONFIG_X86_64 */
}

void __init xen_arch_setup(void)
{
	struct physdev_set_iopl set_iopl;
	int rc;

	xen_panic_handler_init();

	HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_4gb_segments);
	HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_writable_pagetables);

	if (!xen_feature(XENFEAT_auto_translated_physmap))
		HYPERVISOR_vm_assist(VMASST_CMD_enable,
				     VMASST_TYPE_pae_extended_cr3);

	if (register_callback(CALLBACKTYPE_event, xen_hypervisor_callback) ||
	    register_callback(CALLBACKTYPE_failsafe, xen_failsafe_callback))
		BUG();

	xen_enable_sysenter();
	xen_enable_syscall();

	set_iopl.iopl = 1;
	rc = HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl);
	if (rc != 0)
		printk(KERN_INFO "physdev_op failed %d\n", rc);

#ifdef CONFIG_ACPI
	if (!(xen_start_info->flags & SIF_INITDOMAIN)) {
		printk(KERN_INFO "ACPI in unprivileged domain disabled\n");
		disable_acpi();
	}
#endif

	memcpy(boot_command_line, xen_start_info->cmd_line,
	       MAX_GUEST_CMDLINE > COMMAND_LINE_SIZE ?
	       COMMAND_LINE_SIZE : MAX_GUEST_CMDLINE);

	pm_idle = xen_idle;

	fiddle_vdso();
}
Commit	Line	Data
5ead97c8 JF	1	/*
	2	* Machine specific setup for xen
	3	*
	4	* Jeremy Fitzhardinge <jeremy@xensource.com>, XenSource Inc, 2007
	5	*/
	6
	7	#include <linux/module.h>
	8	#include <linux/sched.h>
	9	#include <linux/mm.h>
	10	#include <linux/pm.h>
a9ce6bc1	11	#include <linux/memblock.h>
5ead97c8 JF	12
5ead97c8 JF	13	#include <asm/elf.h>
6c3652ef	14	#include <asm/vdso.h>
5ead97c8 JF	15	#include <asm/e820.h>
5ead97c8 JF	16	#include <asm/setup.h>
b792c755	17	#include <asm/acpi.h>
5ead97c8 JF	18	#include <asm/xen/hypervisor.h>
	19	#include <asm/xen/hypercall.h>
	20
45263cb0	21	#include <xen/xen.h>
8006ec3e	22	#include <xen/page.h>
e2a81baf	23	#include <xen/interface/callback.h>
35ae11fd	24	#include <xen/interface/memory.h>
5ead97c8	25	#include <xen/interface/physdev.h>
093d7b46	26	#include <xen/interface/memory.h>
5ead97c8 JF	27	#include <xen/features.h>
	28
	29	#include "xen-ops.h"
d2eea68e	30	#include "vdso.h"
5ead97c8 JF	31
	32	/* These are code, but not functions. Defined in entry.S */
	33	extern const char xen_hypervisor_callback[];
	34	extern const char xen_failsafe_callback[];
f63c2f24 T	35	extern void xen_sysenter_target(void);
	36	extern void xen_syscall_target(void);
	37	extern void xen_syscall32_target(void);
5ead97c8	38
42ee1471 JF	39	/* Amount of extra memory space we add to the e820 ranges */
	40	phys_addr_t xen_extra_mem_start, xen_extra_mem_size;
	41
698bb8d1 JF	42	/*
	43	* The maximum amount of extra memory compared to the base size. The
	44	* main scaling factor is the size of struct page. At extreme ratios
	45	* of base:extra, all the base memory can be filled with page
	46	* structures for the extra memory, leaving no space for anything
	47	* else.
	48	*
	49	* 10x seems like a reasonable balance between scaling flexibility and
	50	* leaving a practically usable system.
	51	*/
	52	#define EXTRA_MEM_RATIO (10)
	53
42ee1471 JF	54	static __init void xen_add_extra_mem(unsigned long pages)
	55	{
	56	u64 size = (u64)pages * PAGE_SIZE;
3654581e	57	u64 extra_start = xen_extra_mem_start + xen_extra_mem_size;
42ee1471 JF	58
	59	if (!pages)
	60	return;
	61
3654581e	62	e820_add_region(extra_start, size, E820_RAM);
42ee1471 JF	63	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
42ee1471 JF	64
520045db	65	memblock_x86_reserve_range(extra_start, extra_start + size, "XEN EXTRA");
42ee1471 JF	66
42ee1471 JF	67	xen_extra_mem_size += size;
2f7acb20	68
3654581e	69	xen_max_p2m_pfn = PFN_DOWN(extra_start + size);
42ee1471 JF	70	}
42ee1471 JF	71
f89e048e JF	72	static unsigned long __init xen_release_chunk(phys_addr_t start_addr,
f89e048e JF	73	phys_addr_t end_addr)
093d7b46 MR	74	{
	75	struct xen_memory_reservation reservation = {
	76	.address_bits = 0,
	77	.extent_order = 0,
	78	.domid = DOMID_SELF
	79	};
093d7b46	80	unsigned long start, end;
f89e048e	81	unsigned long len = 0;
093d7b46 MR	82	unsigned long pfn;
	83	int ret;
	84
	85	start = PFN_UP(start_addr);
f89e048e	86	end = PFN_DOWN(end_addr);
093d7b46 MR	87
	88	if (end <= start)
	89	return 0;
	90
f89e048e JF	91	printk(KERN_INFO "xen_release_chunk: looking at area pfn %lx-%lx: ",
	92	start, end);
	93	for(pfn = start; pfn < end; pfn++) {
	94	unsigned long mfn = pfn_to_mfn(pfn);
	95
	96	/* Make sure pfn exists to start with */
	97	if (mfn == INVALID_P2M_ENTRY \|\| mfn_to_pfn(mfn) != pfn)
	98	continue;
	99
	100	set_xen_guest_handle(reservation.extent_start, &mfn);
	101	reservation.nr_extents = 1;
	102
	103	ret = HYPERVISOR_memory_op(XENMEM_decrease_reservation,
	104	&reservation);
	105	WARN(ret != 1, "Failed to release memory %lx-%lx err=%d\n",
	106	start, end, ret);
	107	if (ret == 1) {
	108	set_phys_to_machine(pfn, INVALID_P2M_ENTRY);
	109	len++;
	110	}
	111	}
	112	printk(KERN_CONT "%ld pages freed\n", len);
093d7b46 MR	113
	114	return len;
	115	}
	116
f89e048e JF	117	static unsigned long __init xen_return_unused_memory(unsigned long max_pfn,
f89e048e JF	118	const struct e820map *e820)
093d7b46	119	{
f89e048e JF	120	phys_addr_t max_addr = PFN_PHYS(max_pfn);
f89e048e JF	121	phys_addr_t last_end = 0;
093d7b46 MR	122	unsigned long released = 0;
	123	int i;
	124
f89e048e JF	125	for (i = 0; i < e820->nr_map && last_end < max_addr; i++) {
	126	phys_addr_t end = e820->map[i].addr;
	127	end = min(max_addr, end);
	128
	129	released += xen_release_chunk(last_end, end);
093d7b46 MR	130	last_end = e820->map[i].addr + e820->map[i].size;
	131	}
	132
f89e048e JF	133	if (last_end < max_addr)
f89e048e JF	134	released += xen_release_chunk(last_end, max_addr);
093d7b46 MR	135
	136	printk(KERN_INFO "released %ld pages of unused memory\n", released);
	137	return released;
	138	}
5ead97c8 JF	139
	140	/**
	141	* machine_specific_memory_setup - Hook for machine specific memory setup.
	142	**/
5ead97c8 JF	143	char * __init xen_memory_setup(void)
5ead97c8 JF	144	{
35ae11fd IC	145	static struct e820entry map[E820MAX] __initdata;
35ae11fd IC	146
5ead97c8	147	unsigned long max_pfn = xen_start_info->nr_pages;
35ae11fd IC	148	unsigned long long mem_end;
	149	int rc;
	150	struct xen_memory_map memmap;
42ee1471	151	unsigned long extra_pages = 0;
698bb8d1	152	unsigned long extra_limit;
35ae11fd	153	int i;
9e9a5fcb	154	int op;
5ead97c8	155
8006ec3e	156	max_pfn = min(MAX_DOMAIN_PAGES, max_pfn);
35ae11fd IC	157	mem_end = PFN_PHYS(max_pfn);
	158
	159	memmap.nr_entries = E820MAX;
	160	set_xen_guest_handle(memmap.buffer, map);
	161
9e9a5fcb IC	162	op = xen_initial_domain() ?
	163	XENMEM_machine_memory_map :
	164	XENMEM_memory_map;
	165	rc = HYPERVISOR_memory_op(op, &memmap);
35ae11fd IC	166	if (rc == -ENOSYS) {
	167	memmap.nr_entries = 1;
	168	map[0].addr = 0ULL;
	169	map[0].size = mem_end;
	170	/* 8MB slack (to balance backend allocations). */
	171	map[0].size += 8ULL << 20;
	172	map[0].type = E820_RAM;
	173	rc = 0;
	174	}
	175	BUG_ON(rc);
8006ec3e	176
5ead97c8	177	e820.nr_map = 0;
42ee1471	178	xen_extra_mem_start = mem_end;
35ae11fd IC	179	for (i = 0; i < memmap.nr_entries; i++) {
35ae11fd IC	180	unsigned long long end = map[i].addr + map[i].size;
be5bf9fa	181
35ae11fd	182	if (map[i].type == E820_RAM) {
3654581e	183	if (map[i].addr < mem_end && end > mem_end) {
35ae11fd	184	/* Truncate region to max_mem. */
3654581e	185	u64 delta = end - mem_end;
42ee1471	186
3654581e JF	187	map[i].size -= delta;
	188	extra_pages += PFN_DOWN(delta);
	189
	190	end = mem_end;
35ae11fd	191	}
3654581e JF	192	}
	193
	194	if (end > xen_extra_mem_start)
36bc251b	195	xen_extra_mem_start = end;
b5b43ced	196
3654581e	197	/* If region is non-RAM or below mem_end, add what remains */
b5b43ced JF	198	if ((map[i].type != E820_RAM \|\| map[i].addr < mem_end) &&
b5b43ced JF	199	map[i].size > 0)
35ae11fd IC	200	e820_add_region(map[i].addr, map[i].size, map[i].type);
35ae11fd IC	201	}
b792c755 JF	202
	203	/*
	204	* Even though this is normal, usable memory under Xen, reserve
	205	* ISA memory anyway because too many things think they can poke
	206	* about in there.
4ec5387c JQ	207	*
	208	* In a dom0 kernel, this region is identity mapped with the
	209	* hardware ISA area, so it really is out of bounds.
b792c755 JF	210	*/
	211	e820_add_region(ISA_START_ADDRESS, ISA_END_ADDRESS - ISA_START_ADDRESS,
	212	E820_RESERVED);
5ead97c8	213
be5bf9fa JF	214	/*
	215	* Reserve Xen bits:
	216	* - mfn_list
	217	* - xen_start_info
	218	* See comment above "struct start_info" in <xen/interface/xen.h>
	219	*/
a9ce6bc1	220	memblock_x86_reserve_range(__pa(xen_start_info->mfn_list),
6b2e8523 JF	221	__pa(xen_start_info->pt_base),
6b2e8523 JF	222	"XEN START INFO");
be5bf9fa JF	223
	224	sanitize_e820_map(e820.map, ARRAY_SIZE(e820.map), &e820.nr_map);
	225
42ee1471 JF	226	extra_pages += xen_return_unused_memory(xen_start_info->nr_pages, &e820);
42ee1471 JF	227
698bb8d1 JF	228	/*
	229	* Clamp the amount of extra memory to a EXTRA_MEM_RATIO
	230	* factor the base size. On non-highmem systems, the base
	231	* size is the full initial memory allocation; on highmem it
	232	* is limited to the max size of lowmem, so that it doesn't
	233	* get completely filled.
	234	*
	235	* In principle there could be a problem in lowmem systems if
	236	* the initial memory is also very large with respect to
	237	* lowmem, but we won't try to deal with that here.
	238	*/
	239	extra_limit = min(EXTRA_MEM_RATIO * min(max_pfn, PFN_DOWN(MAXMEM)),
	240	max_pfn + extra_pages);
	241
	242	if (extra_limit >= max_pfn)
	243	extra_pages = extra_limit - max_pfn;
	244	else
	245	extra_pages = 0;
	246
9e9a5fcb IC	247	if (!xen_initial_domain())
9e9a5fcb IC	248	xen_add_extra_mem(extra_pages);
093d7b46	249
5ead97c8 JF	250	return "Xen";
	251	}
	252
	253	static void xen_idle(void)
	254	{
	255	local_irq_disable();
	256
	257	if (need_resched())
	258	local_irq_enable();
	259	else {
	260	current_thread_info()->status &= ~TS_POLLING;
	261	smp_mb__after_clear_bit();
	262	safe_halt();
	263	current_thread_info()->status \|= TS_POLLING;
	264	}
	265	}
	266
d2eea68e RM	267	/*
d2eea68e RM	268	* Set the bit indicating "nosegneg" library variants should be used.
6a52e4b1 JF	269	* We only need to bother in pure 32-bit mode; compat 32-bit processes
6a52e4b1 JF	270	* can have un-truncated segments, so wrapping around is allowed.
d2eea68e	271	*/
08b6d290	272	static void __init fiddle_vdso(void)
d2eea68e	273	{
6a52e4b1 JF	274	#ifdef CONFIG_X86_32
	275	u32 *mask;
	276	mask = VDSO32_SYMBOL(&vdso32_int80_start, NOTE_MASK);
	277	*mask \|= 1 << VDSO_NOTE_NONEGSEG_BIT;
	278	mask = VDSO32_SYMBOL(&vdso32_sysenter_start, NOTE_MASK);
d2eea68e	279	*mask \|= 1 << VDSO_NOTE_NONEGSEG_BIT;
6fcac6d3	280	#endif
d2eea68e RM	281	}
d2eea68e RM	282
88459d4c	283	static __cpuinit int register_callback(unsigned type, const void *func)
e2a81baf	284	{
88459d4c JF	285	struct callback_register callback = {
	286	.type = type,
	287	.address = XEN_CALLBACK(__KERNEL_CS, func),
e2a81baf JF	288	.flags = CALLBACKF_mask_events,
	289	};
	290
88459d4c JF	291	return HYPERVISOR_callback_op(CALLBACKOP_register, &callback);
	292	}
	293
	294	void __cpuinit xen_enable_sysenter(void)
	295	{
6fcac6d3	296	int ret;
62541c37	297	unsigned sysenter_feature;
6fcac6d3 JF	298
6fcac6d3 JF	299	#ifdef CONFIG_X86_32
62541c37	300	sysenter_feature = X86_FEATURE_SEP;
6fcac6d3	301	#else
62541c37	302	sysenter_feature = X86_FEATURE_SYSENTER32;
6fcac6d3	303	#endif
88459d4c	304
62541c37 JF	305	if (!boot_cpu_has(sysenter_feature))
	306	return;
	307
6fcac6d3	308	ret = register_callback(CALLBACKTYPE_sysenter, xen_sysenter_target);
62541c37 JF	309	if(ret != 0)
62541c37 JF	310	setup_clear_cpu_cap(sysenter_feature);
e2a81baf JF	311	}
e2a81baf JF	312
6fcac6d3 JF	313	void __cpuinit xen_enable_syscall(void)
	314	{
	315	#ifdef CONFIG_X86_64
6fcac6d3	316	int ret;
6fcac6d3 JF	317
	318	ret = register_callback(CALLBACKTYPE_syscall, xen_syscall_target);
	319	if (ret != 0) {
d5303b81	320	printk(KERN_ERR "Failed to set syscall callback: %d\n", ret);
62541c37 JF	321	/* Pretty fatal; 64-bit userspace has no other
	322	mechanism for syscalls. */
	323	}
	324
	325	if (boot_cpu_has(X86_FEATURE_SYSCALL32)) {
6fcac6d3 JF	326	ret = register_callback(CALLBACKTYPE_syscall32,
6fcac6d3 JF	327	xen_syscall32_target);
d5303b81	328	if (ret != 0)
62541c37	329	setup_clear_cpu_cap(X86_FEATURE_SYSCALL32);
6fcac6d3 JF	330	}
	331	#endif /* CONFIG_X86_64 */
	332	}
	333
5ead97c8 JF	334	void __init xen_arch_setup(void)
	335	{
	336	struct physdev_set_iopl set_iopl;
	337	int rc;
	338
f09f6d19 DD	339	xen_panic_handler_init();
f09f6d19 DD	340
5ead97c8 JF	341	HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_4gb_segments);
	342	HYPERVISOR_vm_assist(VMASST_CMD_enable, VMASST_TYPE_writable_pagetables);
	343
	344	if (!xen_feature(XENFEAT_auto_translated_physmap))
f63c2f24 T	345	HYPERVISOR_vm_assist(VMASST_CMD_enable,
f63c2f24 T	346	VMASST_TYPE_pae_extended_cr3);
5ead97c8	347
88459d4c JF	348	if (register_callback(CALLBACKTYPE_event, xen_hypervisor_callback) \|\|
	349	register_callback(CALLBACKTYPE_failsafe, xen_failsafe_callback))
	350	BUG();
5ead97c8	351
e2a81baf	352	xen_enable_sysenter();
6fcac6d3	353	xen_enable_syscall();
e2a81baf	354
5ead97c8 JF	355	set_iopl.iopl = 1;
	356	rc = HYPERVISOR_physdev_op(PHYSDEVOP_set_iopl, &set_iopl);
	357	if (rc != 0)
	358	printk(KERN_INFO "physdev_op failed %d\n", rc);
	359
	360	#ifdef CONFIG_ACPI
	361	if (!(xen_start_info->flags & SIF_INITDOMAIN)) {
	362	printk(KERN_INFO "ACPI in unprivileged domain disabled\n");
	363	disable_acpi();
	364	}
	365	#endif
	366
	367	memcpy(boot_command_line, xen_start_info->cmd_line,
	368	MAX_GUEST_CMDLINE > COMMAND_LINE_SIZE ?
	369	COMMAND_LINE_SIZE : MAX_GUEST_CMDLINE);
	370
	371	pm_idle = xen_idle;
f87e4cac	372
d2eea68e	373	fiddle_vdso();
5ead97c8	374	}