[deliverable/linux.git] / arch / powerpc / kernel / machine_kexec_64.c

/*
 * PPC64 code to handle Linux booting another kernel.
 *
 * Copyright (C) 2004-2005, IBM Corp.
 *
 * Created by: Milton D Miller II
 *
 * This source code is licensed under the GNU General Public License,
 * Version 2.  See the file COPYING for more details.
 */


#include <linux/cpumask.h>
#include <linux/kexec.h>
#include <linux/smp.h>
#include <linux/thread_info.h>
#include <linux/errno.h>

#include <asm/page.h>
#include <asm/current.h>
#include <asm/machdep.h>
#include <asm/cacheflush.h>
#include <asm/paca.h>
#include <asm/mmu.h>
#include <asm/sections.h>	/* _end */
#include <asm/prom.h>
#include <asm/smp.h>

int default_machine_kexec_prepare(struct kimage *image)
{
	int i;
	unsigned long begin, end;	/* limits of segment */
	unsigned long low, high;	/* limits of blocked memory range */
	struct device_node *node;
	unsigned long *basep;
	unsigned int *sizep;

	if (!ppc_md.hpte_clear_all)
		return -ENOENT;

	/*
	 * Since we use the kernel fault handlers and paging code to
	 * handle the virtual mode, we must make sure no destination
	 * overlaps kernel static data or bss.
	 */
	for (i = 0; i < image->nr_segments; i++)
		if (image->segment[i].mem < __pa(_end))
			return -ETXTBSY;

	/*
	 * For non-LPAR, we absolutely can not overwrite the mmu hash
	 * table, since we are still using the bolted entries in it to
	 * do the copy.  Check that here.
	 *
	 * It is safe if the end is below the start of the blocked
	 * region (end <= low), or if the beginning is after the
	 * end of the blocked region (begin >= high).  Use the
	 * boolean identity !(a || b)  === (!a && !b).
	 */
	if (htab_address) {
		low = __pa(htab_address);
		high = low + htab_size_bytes;

		for (i = 0; i < image->nr_segments; i++) {
			begin = image->segment[i].mem;
			end = begin + image->segment[i].memsz;

			if ((begin < high) && (end > low))
				return -ETXTBSY;
		}
	}

	/* We also should not overwrite the tce tables */
	for (node = of_find_node_by_type(NULL, "pci"); node != NULL;
			node = of_find_node_by_type(node, "pci")) {
		basep = (unsigned long *)get_property(node, "linux,tce-base",
							NULL);
		sizep = (unsigned int *)get_property(node, "linux,tce-size",
							NULL);
		if (basep == NULL || sizep == NULL)
			continue;

		low = *basep;
		high = low + (*sizep);

		for (i = 0; i < image->nr_segments; i++) {
			begin = image->segment[i].mem;
			end = begin + image->segment[i].memsz;

			if ((begin < high) && (end > low))
				return -ETXTBSY;
		}
	}

	return 0;
}

#define IND_FLAGS (IND_DESTINATION | IND_INDIRECTION | IND_DONE | IND_SOURCE)

static void copy_segments(unsigned long ind)
{
	unsigned long entry;
	unsigned long *ptr;
	void *dest;
	void *addr;

	/*
	 * We rely on kexec_load to create a lists that properly
	 * initializes these pointers before they are used.
	 * We will still crash if the list is wrong, but at least
	 * the compiler will be quiet.
	 */
	ptr = NULL;
	dest = NULL;

	for (entry = ind; !(entry & IND_DONE); entry = *ptr++) {
		addr = __va(entry & PAGE_MASK);

		switch (entry & IND_FLAGS) {
		case IND_DESTINATION:
			dest = addr;
			break;
		case IND_INDIRECTION:
			ptr = addr;
			break;
		case IND_SOURCE:
			copy_page(dest, addr);
			dest += PAGE_SIZE;
		}
	}
}

void kexec_copy_flush(struct kimage *image)
{
	long i, nr_segments = image->nr_segments;
	struct  kexec_segment ranges[KEXEC_SEGMENT_MAX];

	/* save the ranges on the stack to efficiently flush the icache */
	memcpy(ranges, image->segment, sizeof(ranges));

	/*
	 * After this call we may not use anything allocated in dynamic
	 * memory, including *image.
	 *
	 * Only globals and the stack are allowed.
	 */
	copy_segments(image->head);

	/*
	 * we need to clear the icache for all dest pages sometime,
	 * including ones that were in place on the original copy
	 */
	for (i = 0; i < nr_segments; i++)
		flush_icache_range((unsigned long)__va(ranges[i].mem),
			(unsigned long)__va(ranges[i].mem + ranges[i].memsz));
}

#ifdef CONFIG_SMP

/* FIXME: we should schedule this function to be called on all cpus based
 * on calling the interrupts, but we would like to call it off irq level
 * so that the interrupt controller is clean.
 */
void kexec_smp_down(void *arg)
{
	if (ppc_md.kexec_cpu_down)
		ppc_md.kexec_cpu_down(0, 1);

	local_irq_disable();
	kexec_smp_wait();
	/* NOTREACHED */
}

static void kexec_prepare_cpus(void)
{
	int my_cpu, i, notified=-1;

	smp_call_function(kexec_smp_down, NULL, 0, /* wait */0);
	my_cpu = get_cpu();

	/* check the others cpus are now down (via paca hw cpu id == -1) */
	for (i=0; i < NR_CPUS; i++) {
		if (i == my_cpu)
			continue;

		while (paca[i].hw_cpu_id != -1) {
			barrier();
			if (!cpu_possible(i)) {
				printk("kexec: cpu %d hw_cpu_id %d is not"
						" possible, ignoring\n",
						i, paca[i].hw_cpu_id);
				break;
			}
			if (!cpu_online(i)) {
				/* Fixme: this can be spinning in
				 * pSeries_secondary_wait with a paca
				 * waiting for it to go online.
				 */
				printk("kexec: cpu %d hw_cpu_id %d is not"
						" online, ignoring\n",
						i, paca[i].hw_cpu_id);
				break;
			}
			if (i != notified) {
				printk( "kexec: waiting for cpu %d (physical"
						" %d) to go down\n",
						i, paca[i].hw_cpu_id);
				notified = i;
			}
		}
	}

	/* after we tell the others to go down */
	if (ppc_md.kexec_cpu_down)
		ppc_md.kexec_cpu_down(0, 0);

	put_cpu();

	local_irq_disable();
}

#else /* ! SMP */

static void kexec_prepare_cpus(void)
{
	/*
	 * move the secondarys to us so that we can copy
	 * the new kernel 0-0x100 safely
	 *
	 * do this if kexec in setup.c ?
	 *
	 * We need to release the cpus if we are ever going from an
	 * UP to an SMP kernel.
	 */
	smp_release_cpus();
	if (ppc_md.kexec_cpu_down)
		ppc_md.kexec_cpu_down(0, 0);
	local_irq_disable();
}

#endif /* SMP */

/*
 * kexec thread structure and stack.
 *
 * We need to make sure that this is 16384-byte aligned due to the
 * way process stacks are handled.  It also must be statically allocated
 * or allocated as part of the kimage, because everything else may be
 * overwritten when we copy the kexec image.  We piggyback on the
 * "init_task" linker section here to statically allocate a stack.
 *
 * We could use a smaller stack if we don't care about anything using
 * current, but that audit has not been performed.
 */
union thread_union kexec_stack
	__attribute__((__section__(".data.init_task"))) = { };

/* Our assembly helper, in kexec_stub.S */
extern NORET_TYPE void kexec_sequence(void *newstack, unsigned long start,
					void *image, void *control,
					void (*clear_all)(void)) ATTRIB_NORET;

/* too late to fail here */
void default_machine_kexec(struct kimage *image)
{
	/* prepare control code if any */

	/*
        * If the kexec boot is the normal one, need to shutdown other cpus
        * into our wait loop and quiesce interrupts.
        * Otherwise, in the case of crashed mode (crashing_cpu >= 0),
        * stopping other CPUs and collecting their pt_regs is done before
        * using debugger IPI.
        */

       if (crashing_cpu == -1)
               kexec_prepare_cpus();

	/* switch to a staticly allocated stack.  Based on irq stack code.
	 * XXX: the task struct will likely be invalid once we do the copy!
	 */
	kexec_stack.thread_info.task = current_thread_info()->task;
	kexec_stack.thread_info.flags = 0;

	/* Some things are best done in assembly.  Finding globals with
	 * a toc is easier in C, so pass in what we can.
	 */
	kexec_sequence(&kexec_stack, image->start, image,
			page_address(image->control_code_page),
			ppc_md.hpte_clear_all);
	/* NOTREACHED */
}

/* Values we need to export to the second kernel via the device tree. */
static unsigned long htab_base, kernel_end;

static struct property htab_base_prop = {
	.name = "linux,htab-base",
	.length = sizeof(unsigned long),
	.value = (unsigned char *)&htab_base,
};

static struct property htab_size_prop = {
	.name = "linux,htab-size",
	.length = sizeof(unsigned long),
	.value = (unsigned char *)&htab_size_bytes,
};

static struct property kernel_end_prop = {
	.name = "linux,kernel-end",
	.length = sizeof(unsigned long),
	.value = (unsigned char *)&kernel_end,
};

static void __init export_htab_values(void)
{
	struct device_node *node;

	node = of_find_node_by_path("/chosen");
	if (!node)
		return;

	kernel_end = __pa(_end);
	prom_add_property(node, &kernel_end_prop);

	/* On machines with no htab htab_address is NULL */
	if (NULL == htab_address)
		goto out;

	htab_base = __pa(htab_address);
	prom_add_property(node, &htab_base_prop);
	prom_add_property(node, &htab_size_prop);

 out:
	of_node_put(node);
}

void __init kexec_setup(void)
{
	export_htab_values();
}

int overlaps_crashkernel(unsigned long start, unsigned long size)
{
	return (start + size) > crashk_res.start && start <= crashk_res.end;
}
Commit	Line	Data
fce0d574	1	/*
3d1229d6	2	* PPC64 code to handle Linux booting another kernel.
fce0d574 S	3	*
	4	* Copyright (C) 2004-2005, IBM Corp.
	5	*
	6	* Created by: Milton D Miller II
	7	*
	8	* This source code is licensed under the GNU General Public License,
	9	* Version 2. See the file COPYING for more details.
	10	*/
	11
	12
	13	#include <linux/cpumask.h>
	14	#include <linux/kexec.h>
	15	#include <linux/smp.h>
	16	#include <linux/thread_info.h>
	17	#include <linux/errno.h>
	18
	19	#include <asm/page.h>
	20	#include <asm/current.h>
	21	#include <asm/machdep.h>
	22	#include <asm/cacheflush.h>
	23	#include <asm/paca.h>
	24	#include <asm/mmu.h>
	25	#include <asm/sections.h> /* _end */
	26	#include <asm/prom.h>
2249ca9d	27	#include <asm/smp.h>
fce0d574	28
3d1229d6	29	int default_machine_kexec_prepare(struct kimage *image)
fce0d574 S	30	{
	31	int i;
	32	unsigned long begin, end; /* limits of segment */
	33	unsigned long low, high; /* limits of blocked memory range */
	34	struct device_node *node;
	35	unsigned long *basep;
	36	unsigned int *sizep;
	37
	38	if (!ppc_md.hpte_clear_all)
	39	return -ENOENT;
	40
	41	/*
	42	* Since we use the kernel fault handlers and paging code to
	43	* handle the virtual mode, we must make sure no destination
	44	* overlaps kernel static data or bss.
	45	*/
72414d3f	46	for (i = 0; i < image->nr_segments; i++)
fce0d574 S	47	if (image->segment[i].mem < __pa(_end))
	48	return -ETXTBSY;
	49
	50	/*
	51	* For non-LPAR, we absolutely can not overwrite the mmu hash
	52	* table, since we are still using the bolted entries in it to
	53	* do the copy. Check that here.
	54	*
	55	* It is safe if the end is below the start of the blocked
	56	* region (end <= low), or if the beginning is after the
	57	* end of the blocked region (begin >= high). Use the
	58	* boolean identity !(a \|\| b) === (!a && !b).
	59	*/
	60	if (htab_address) {
	61	low = __pa(htab_address);
337a7128	62	high = low + htab_size_bytes;
fce0d574	63
72414d3f	64	for (i = 0; i < image->nr_segments; i++) {
fce0d574 S	65	begin = image->segment[i].mem;
	66	end = begin + image->segment[i].memsz;
	67
	68	if ((begin < high) && (end > low))
	69	return -ETXTBSY;
	70	}
	71	}
	72
	73	/* We also should not overwrite the tce tables */
	74	for (node = of_find_node_by_type(NULL, "pci"); node != NULL;
	75	node = of_find_node_by_type(node, "pci")) {
	76	basep = (unsigned long *)get_property(node, "linux,tce-base",
	77	NULL);
	78	sizep = (unsigned int *)get_property(node, "linux,tce-size",
	79	NULL);
	80	if (basep == NULL \|\| sizep == NULL)
	81	continue;
	82
	83	low = *basep;
	84	high = low + (*sizep);
	85
72414d3f	86	for (i = 0; i < image->nr_segments; i++) {
fce0d574 S	87	begin = image->segment[i].mem;
	88	end = begin + image->segment[i].memsz;
	89
	90	if ((begin < high) && (end > low))
	91	return -ETXTBSY;
	92	}
	93	}
	94
	95	return 0;
	96	}
	97
fce0d574 S	98	#define IND_FLAGS (IND_DESTINATION \| IND_INDIRECTION \| IND_DONE \| IND_SOURCE)
	99
	100	static void copy_segments(unsigned long ind)
	101	{
	102	unsigned long entry;
	103	unsigned long *ptr;
	104	void *dest;
	105	void *addr;
	106
	107	/*
	108	* We rely on kexec_load to create a lists that properly
	109	* initializes these pointers before they are used.
	110	* We will still crash if the list is wrong, but at least
	111	* the compiler will be quiet.
	112	*/
	113	ptr = NULL;
	114	dest = NULL;
	115
	116	for (entry = ind; !(entry & IND_DONE); entry = *ptr++) {
	117	addr = __va(entry & PAGE_MASK);
	118
	119	switch (entry & IND_FLAGS) {
	120	case IND_DESTINATION:
	121	dest = addr;
	122	break;
	123	case IND_INDIRECTION:
	124	ptr = addr;
	125	break;
	126	case IND_SOURCE:
	127	copy_page(dest, addr);
	128	dest += PAGE_SIZE;
	129	}
	130	}
	131	}
	132
	133	void kexec_copy_flush(struct kimage *image)
	134	{
	135	long i, nr_segments = image->nr_segments;
	136	struct kexec_segment ranges[KEXEC_SEGMENT_MAX];
	137
	138	/* save the ranges on the stack to efficiently flush the icache */
	139	memcpy(ranges, image->segment, sizeof(ranges));
	140
	141	/*
	142	* After this call we may not use anything allocated in dynamic
	143	* memory, including *image.
	144	*
	145	* Only globals and the stack are allowed.
	146	*/
	147	copy_segments(image->head);
	148
	149	/*
	150	* we need to clear the icache for all dest pages sometime,
	151	* including ones that were in place on the original copy
	152	*/
	153	for (i = 0; i < nr_segments; i++)
b5666f70 ME	154	flush_icache_range((unsigned long)__va(ranges[i].mem),
b5666f70 ME	155	(unsigned long)__va(ranges[i].mem + ranges[i].memsz));
fce0d574 S	156	}
	157
	158	#ifdef CONFIG_SMP
	159
	160	/* FIXME: we should schedule this function to be called on all cpus based
	161	* on calling the interrupts, but we would like to call it off irq level
	162	* so that the interrupt controller is clean.
	163	*/
	164	void kexec_smp_down(void *arg)
	165	{
c5e24354 ME	166	if (ppc_md.kexec_cpu_down)
c5e24354 ME	167	ppc_md.kexec_cpu_down(0, 1);
fce0d574 S	168
	169	local_irq_disable();
	170	kexec_smp_wait();
	171	/* NOTREACHED */
	172	}
	173
	174	static void kexec_prepare_cpus(void)
	175	{
	176	int my_cpu, i, notified=-1;
	177
	178	smp_call_function(kexec_smp_down, NULL, 0, /* wait */0);
	179	my_cpu = get_cpu();
	180
	181	/* check the others cpus are now down (via paca hw cpu id == -1) */
	182	for (i=0; i < NR_CPUS; i++) {
	183	if (i == my_cpu)
	184	continue;
	185
	186	while (paca[i].hw_cpu_id != -1) {
b3ca8093	187	barrier();
fce0d574 S	188	if (!cpu_possible(i)) {
	189	printk("kexec: cpu %d hw_cpu_id %d is not"
	190	" possible, ignoring\n",
	191	i, paca[i].hw_cpu_id);
	192	break;
	193	}
	194	if (!cpu_online(i)) {
	195	/* Fixme: this can be spinning in
	196	* pSeries_secondary_wait with a paca
	197	* waiting for it to go online.
	198	*/
	199	printk("kexec: cpu %d hw_cpu_id %d is not"
	200	" online, ignoring\n",
	201	i, paca[i].hw_cpu_id);
	202	break;
	203	}
	204	if (i != notified) {
	205	printk( "kexec: waiting for cpu %d (physical"
	206	" %d) to go down\n",
	207	i, paca[i].hw_cpu_id);
	208	notified = i;
	209	}
	210	}
	211	}
	212
	213	/* after we tell the others to go down */
c5e24354 ME	214	if (ppc_md.kexec_cpu_down)
c5e24354 ME	215	ppc_md.kexec_cpu_down(0, 0);
fce0d574 S	216
	217	put_cpu();
	218
	219	local_irq_disable();
	220	}
	221
	222	#else /* ! SMP */
	223
	224	static void kexec_prepare_cpus(void)
	225	{
	226	/*
	227	* move the secondarys to us so that we can copy
	228	* the new kernel 0-0x100 safely
	229	*
	230	* do this if kexec in setup.c ?
75eedfed OJ	231	*
	232	* We need to release the cpus if we are ever going from an
	233	* UP to an SMP kernel.
fce0d574	234	*/
75eedfed	235	smp_release_cpus();
c5e24354 ME	236	if (ppc_md.kexec_cpu_down)
c5e24354 ME	237	ppc_md.kexec_cpu_down(0, 0);
fce0d574 S	238	local_irq_disable();
	239	}
	240
	241	#endif /* SMP */
	242
	243	/*
	244	* kexec thread structure and stack.
	245	*
	246	* We need to make sure that this is 16384-byte aligned due to the
	247	* way process stacks are handled. It also must be statically allocated
	248	* or allocated as part of the kimage, because everything else may be
	249	* overwritten when we copy the kexec image. We piggyback on the
	250	* "init_task" linker section here to statically allocate a stack.
	251	*
	252	* We could use a smaller stack if we don't care about anything using
	253	* current, but that audit has not been performed.
	254	*/
	255	union thread_union kexec_stack
	256	__attribute__((__section__(".data.init_task"))) = { };
	257
	258	/* Our assembly helper, in kexec_stub.S */
	259	extern NORET_TYPE void kexec_sequence(void *newstack, unsigned long start,
72414d3f MS	260	void image, void control,
72414d3f MS	261	void (*clear_all)(void)) ATTRIB_NORET;
fce0d574 S	262
fce0d574 S	263	/* too late to fail here */
3d1229d6	264	void default_machine_kexec(struct kimage *image)
fce0d574	265	{
fce0d574 S	266	/* prepare control code if any */
fce0d574 S	267
cc532915 ME	268	/*
	269	* If the kexec boot is the normal one, need to shutdown other cpus
	270	* into our wait loop and quiesce interrupts.
	271	* Otherwise, in the case of crashed mode (crashing_cpu >= 0),
	272	* stopping other CPUs and collecting their pt_regs is done before
	273	* using debugger IPI.
	274	*/
	275
	276	if (crashing_cpu == -1)
	277	kexec_prepare_cpus();
fce0d574 S	278
	279	/* switch to a staticly allocated stack. Based on irq stack code.
	280	* XXX: the task struct will likely be invalid once we do the copy!
	281	*/
	282	kexec_stack.thread_info.task = current_thread_info()->task;
	283	kexec_stack.thread_info.flags = 0;
	284
	285	/* Some things are best done in assembly. Finding globals with
	286	* a toc is easier in C, so pass in what we can.
	287	*/
	288	kexec_sequence(&kexec_stack, image->start, image,
	289	page_address(image->control_code_page),
	290	ppc_md.hpte_clear_all);
	291	/* NOTREACHED */
	292	}
593e537b ME	293
593e537b ME	294	/* Values we need to export to the second kernel via the device tree. */
337a7128	295	static unsigned long htab_base, kernel_end;
593e537b ME	296
	297	static struct property htab_base_prop = {
	298	.name = "linux,htab-base",
	299	.length = sizeof(unsigned long),
	300	.value = (unsigned char *)&htab_base,
	301	};
	302
	303	static struct property htab_size_prop = {
	304	.name = "linux,htab-size",
	305	.length = sizeof(unsigned long),
337a7128	306	.value = (unsigned char *)&htab_size_bytes,
593e537b ME	307	};
	308
	309	static struct property kernel_end_prop = {
	310	.name = "linux,kernel-end",
	311	.length = sizeof(unsigned long),
	312	.value = (unsigned char *)&kernel_end,
	313	};
	314
	315	static void __init export_htab_values(void)
	316	{
	317	struct device_node *node;
	318
	319	node = of_find_node_by_path("/chosen");
	320	if (!node)
	321	return;
	322
	323	kernel_end = __pa(_end);
	324	prom_add_property(node, &kernel_end_prop);
	325
	326	/* On machines with no htab htab_address is NULL */
	327	if (NULL == htab_address)
	328	goto out;
	329
	330	htab_base = __pa(htab_address);
	331	prom_add_property(node, &htab_base_prop);
593e537b ME	332	prom_add_property(node, &htab_size_prop);
	333
	334	out:
	335	of_node_put(node);
	336	}
	337
	338	void __init kexec_setup(void)
	339	{
	340	export_htab_values();
	341	}
2babf5c2 ME	342
	343	int overlaps_crashkernel(unsigned long start, unsigned long size)
	344	{
	345	return (start + size) > crashk_res.start && start <= crashk_res.end;
	346	}