[mirror_ubuntu-artful-kernel.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/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);
}

static struct property crashk_base_prop = {
	.name = "linux,crashkernel-base",
	.length = sizeof(unsigned long),
	.value = (unsigned char *)&crashk_res.start,
};

static unsigned long crashk_size;

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

static void __init export_crashk_values(void)
{
	struct device_node *node;
	struct property *prop;

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

	/* There might be existing crash kernel properties, but we can't
	 * be sure what's in them, so remove them. */
	prop = of_find_property(node, "linux,crashkernel-base", NULL);
	if (prop)
		prom_remove_property(node, prop);

	prop = of_find_property(node, "linux,crashkernel-size", NULL);
	if (prop)
		prom_remove_property(node, prop);

	if (crashk_res.start != 0) {
		prom_add_property(node, &crashk_base_prop);
		crashk_size = crashk_res.end - crashk_res.start + 1;
		prom_add_property(node, &crashk_size_prop);
	}

	of_node_put(node);
}

static int __init kexec_setup(void)
{
	export_htab_values();
	export_crashk_values();
	return 0;
}
__initcall(kexec_setup);
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
fce0d574 S	13	#include <linux/kexec.h>
	14	#include <linux/smp.h>
	15	#include <linux/thread_info.h>
	16	#include <linux/errno.h>
	17
	18	#include <asm/page.h>
	19	#include <asm/current.h>
	20	#include <asm/machdep.h>
	21	#include <asm/cacheflush.h>
	22	#include <asm/paca.h>
	23	#include <asm/mmu.h>
	24	#include <asm/sections.h> /* _end */
	25	#include <asm/prom.h>
2249ca9d	26	#include <asm/smp.h>
fce0d574	27
3d1229d6	28	int default_machine_kexec_prepare(struct kimage *image)
fce0d574 S	29	{
	30	int i;
	31	unsigned long begin, end; /* limits of segment */
	32	unsigned long low, high; /* limits of blocked memory range */
	33	struct device_node *node;
	34	unsigned long *basep;
	35	unsigned int *sizep;
	36
	37	if (!ppc_md.hpte_clear_all)
	38	return -ENOENT;
	39
	40	/*
	41	* Since we use the kernel fault handlers and paging code to
	42	* handle the virtual mode, we must make sure no destination
	43	* overlaps kernel static data or bss.
	44	*/
72414d3f	45	for (i = 0; i < image->nr_segments; i++)
fce0d574 S	46	if (image->segment[i].mem < __pa(_end))
	47	return -ETXTBSY;
	48
	49	/*
	50	* For non-LPAR, we absolutely can not overwrite the mmu hash
	51	* table, since we are still using the bolted entries in it to
	52	* do the copy. Check that here.
	53	*
	54	* It is safe if the end is below the start of the blocked
	55	* region (end <= low), or if the beginning is after the
	56	* end of the blocked region (begin >= high). Use the
	57	* boolean identity !(a \|\| b) === (!a && !b).
	58	*/
	59	if (htab_address) {
	60	low = __pa(htab_address);
337a7128	61	high = low + htab_size_bytes;
fce0d574	62
72414d3f	63	for (i = 0; i < image->nr_segments; i++) {
fce0d574 S	64	begin = image->segment[i].mem;
	65	end = begin + image->segment[i].memsz;
	66
	67	if ((begin < high) && (end > low))
	68	return -ETXTBSY;
	69	}
	70	}
	71
	72	/* We also should not overwrite the tce tables */
	73	for (node = of_find_node_by_type(NULL, "pci"); node != NULL;
	74	node = of_find_node_by_type(node, "pci")) {
	75	basep = (unsigned long *)get_property(node, "linux,tce-base",
	76	NULL);
	77	sizep = (unsigned int *)get_property(node, "linux,tce-size",
	78	NULL);
	79	if (basep == NULL \|\| sizep == NULL)
	80	continue;
	81
	82	low = *basep;
	83	high = low + (*sizep);
	84
72414d3f	85	for (i = 0; i < image->nr_segments; i++) {
fce0d574 S	86	begin = image->segment[i].mem;
	87	end = begin + image->segment[i].memsz;
	88
	89	if ((begin < high) && (end > low))
	90	return -ETXTBSY;
	91	}
	92	}
	93
	94	return 0;
	95	}
	96
fce0d574 S	97	#define IND_FLAGS (IND_DESTINATION \| IND_INDIRECTION \| IND_DONE \| IND_SOURCE)
	98
	99	static void copy_segments(unsigned long ind)
	100	{
	101	unsigned long entry;
	102	unsigned long *ptr;
	103	void *dest;
	104	void *addr;
	105
	106	/*
	107	* We rely on kexec_load to create a lists that properly
	108	* initializes these pointers before they are used.
	109	* We will still crash if the list is wrong, but at least
	110	* the compiler will be quiet.
	111	*/
	112	ptr = NULL;
	113	dest = NULL;
	114
	115	for (entry = ind; !(entry & IND_DONE); entry = *ptr++) {
	116	addr = __va(entry & PAGE_MASK);
	117
	118	switch (entry & IND_FLAGS) {
	119	case IND_DESTINATION:
	120	dest = addr;
	121	break;
	122	case IND_INDIRECTION:
	123	ptr = addr;
	124	break;
	125	case IND_SOURCE:
	126	copy_page(dest, addr);
	127	dest += PAGE_SIZE;
	128	}
	129	}
	130	}
	131
	132	void kexec_copy_flush(struct kimage *image)
	133	{
	134	long i, nr_segments = image->nr_segments;
	135	struct kexec_segment ranges[KEXEC_SEGMENT_MAX];
	136
	137	/* save the ranges on the stack to efficiently flush the icache */
	138	memcpy(ranges, image->segment, sizeof(ranges));
	139
	140	/*
	141	* After this call we may not use anything allocated in dynamic
	142	* memory, including *image.
	143	*
	144	* Only globals and the stack are allowed.
	145	*/
	146	copy_segments(image->head);
	147
	148	/*
	149	* we need to clear the icache for all dest pages sometime,
	150	* including ones that were in place on the original copy
	151	*/
	152	for (i = 0; i < nr_segments; i++)
b5666f70 ME	153	flush_icache_range((unsigned long)__va(ranges[i].mem),
b5666f70 ME	154	(unsigned long)__va(ranges[i].mem + ranges[i].memsz));
fce0d574 S	155	}
	156
	157	#ifdef CONFIG_SMP
	158
	159	/* FIXME: we should schedule this function to be called on all cpus based
	160	* on calling the interrupts, but we would like to call it off irq level
	161	* so that the interrupt controller is clean.
	162	*/
	163	void kexec_smp_down(void *arg)
	164	{
c5e24354 ME	165	if (ppc_md.kexec_cpu_down)
c5e24354 ME	166	ppc_md.kexec_cpu_down(0, 1);
fce0d574 S	167
	168	local_irq_disable();
	169	kexec_smp_wait();
	170	/* NOTREACHED */
	171	}
	172
	173	static void kexec_prepare_cpus(void)
	174	{
	175	int my_cpu, i, notified=-1;
	176
	177	smp_call_function(kexec_smp_down, NULL, 0, /* wait */0);
	178	my_cpu = get_cpu();
	179
	180	/* check the others cpus are now down (via paca hw cpu id == -1) */
	181	for (i=0; i < NR_CPUS; i++) {
	182	if (i == my_cpu)
	183	continue;
	184
	185	while (paca[i].hw_cpu_id != -1) {
b3ca8093	186	barrier();
fce0d574 S	187	if (!cpu_possible(i)) {
	188	printk("kexec: cpu %d hw_cpu_id %d is not"
	189	" possible, ignoring\n",
	190	i, paca[i].hw_cpu_id);
	191	break;
	192	}
	193	if (!cpu_online(i)) {
	194	/* Fixme: this can be spinning in
	195	* pSeries_secondary_wait with a paca
	196	* waiting for it to go online.
	197	*/
	198	printk("kexec: cpu %d hw_cpu_id %d is not"
	199	" online, ignoring\n",
	200	i, paca[i].hw_cpu_id);
	201	break;
	202	}
	203	if (i != notified) {
	204	printk( "kexec: waiting for cpu %d (physical"
	205	" %d) to go down\n",
	206	i, paca[i].hw_cpu_id);
	207	notified = i;
	208	}
	209	}
	210	}
	211
	212	/* after we tell the others to go down */
c5e24354 ME	213	if (ppc_md.kexec_cpu_down)
c5e24354 ME	214	ppc_md.kexec_cpu_down(0, 0);
fce0d574 S	215
	216	put_cpu();
	217
	218	local_irq_disable();
	219	}
	220
	221	#else /* ! SMP */
	222
	223	static void kexec_prepare_cpus(void)
	224	{
	225	/*
	226	* move the secondarys to us so that we can copy
	227	* the new kernel 0-0x100 safely
	228	*
	229	* do this if kexec in setup.c ?
75eedfed OJ	230	*
	231	* We need to release the cpus if we are ever going from an
	232	* UP to an SMP kernel.
fce0d574	233	*/
75eedfed	234	smp_release_cpus();
c5e24354 ME	235	if (ppc_md.kexec_cpu_down)
c5e24354 ME	236	ppc_md.kexec_cpu_down(0, 0);
fce0d574 S	237	local_irq_disable();
	238	}
	239
	240	#endif /* SMP */
	241
	242	/*
	243	* kexec thread structure and stack.
	244	*
	245	* We need to make sure that this is 16384-byte aligned due to the
	246	* way process stacks are handled. It also must be statically allocated
	247	* or allocated as part of the kimage, because everything else may be
	248	* overwritten when we copy the kexec image. We piggyback on the
	249	* "init_task" linker section here to statically allocate a stack.
	250	*
	251	* We could use a smaller stack if we don't care about anything using
	252	* current, but that audit has not been performed.
	253	*/
	254	union thread_union kexec_stack
	255	__attribute__((__section__(".data.init_task"))) = { };
	256
	257	/* Our assembly helper, in kexec_stub.S */
	258	extern NORET_TYPE void kexec_sequence(void *newstack, unsigned long start,
72414d3f MS	259	void image, void control,
72414d3f MS	260	void (*clear_all)(void)) ATTRIB_NORET;
fce0d574 S	261
fce0d574 S	262	/* too late to fail here */
3d1229d6	263	void default_machine_kexec(struct kimage *image)
fce0d574	264	{
fce0d574 S	265	/* prepare control code if any */
fce0d574 S	266
cc532915 ME	267	/*
	268	* If the kexec boot is the normal one, need to shutdown other cpus
	269	* into our wait loop and quiesce interrupts.
	270	* Otherwise, in the case of crashed mode (crashing_cpu >= 0),
	271	* stopping other CPUs and collecting their pt_regs is done before
	272	* using debugger IPI.
	273	*/
	274
	275	if (crashing_cpu == -1)
	276	kexec_prepare_cpus();
fce0d574 S	277
	278	/* switch to a staticly allocated stack. Based on irq stack code.
	279	* XXX: the task struct will likely be invalid once we do the copy!
	280	*/
	281	kexec_stack.thread_info.task = current_thread_info()->task;
	282	kexec_stack.thread_info.flags = 0;
	283
	284	/* Some things are best done in assembly. Finding globals with
	285	* a toc is easier in C, so pass in what we can.
	286	*/
	287	kexec_sequence(&kexec_stack, image->start, image,
	288	page_address(image->control_code_page),
	289	ppc_md.hpte_clear_all);
	290	/* NOTREACHED */
	291	}
593e537b ME	292
593e537b ME	293	/* Values we need to export to the second kernel via the device tree. */
337a7128	294	static unsigned long htab_base, kernel_end;
593e537b ME	295
	296	static struct property htab_base_prop = {
	297	.name = "linux,htab-base",
	298	.length = sizeof(unsigned long),
	299	.value = (unsigned char *)&htab_base,
	300	};
	301
	302	static struct property htab_size_prop = {
	303	.name = "linux,htab-size",
	304	.length = sizeof(unsigned long),
337a7128	305	.value = (unsigned char *)&htab_size_bytes,
593e537b ME	306	};
	307
	308	static struct property kernel_end_prop = {
	309	.name = "linux,kernel-end",
	310	.length = sizeof(unsigned long),
	311	.value = (unsigned char *)&kernel_end,
	312	};
	313
	314	static void __init export_htab_values(void)
	315	{
	316	struct device_node *node;
	317
	318	node = of_find_node_by_path("/chosen");
	319	if (!node)
	320	return;
	321
	322	kernel_end = __pa(_end);
	323	prom_add_property(node, &kernel_end_prop);
	324
	325	/* On machines with no htab htab_address is NULL */
	326	if (NULL == htab_address)
	327	goto out;
	328
	329	htab_base = __pa(htab_address);
	330	prom_add_property(node, &htab_base_prop);
593e537b ME	331	prom_add_property(node, &htab_size_prop);
	332
	333	out:
	334	of_node_put(node);
	335	}
	336
35dd5432 ME	337	static struct property crashk_base_prop = {
	338	.name = "linux,crashkernel-base",
	339	.length = sizeof(unsigned long),
	340	.value = (unsigned char *)&crashk_res.start,
	341	};
	342
	343	static unsigned long crashk_size;
	344
	345	static struct property crashk_size_prop = {
	346	.name = "linux,crashkernel-size",
	347	.length = sizeof(unsigned long),
	348	.value = (unsigned char *)&crashk_size,
	349	};
	350
	351	static void __init export_crashk_values(void)
	352	{
	353	struct device_node *node;
	354	struct property *prop;
	355
	356	node = of_find_node_by_path("/chosen");
	357	if (!node)
	358	return;
	359
	360	/* There might be existing crash kernel properties, but we can't
	361	* be sure what's in them, so remove them. */
	362	prop = of_find_property(node, "linux,crashkernel-base", NULL);
	363	if (prop)
	364	prom_remove_property(node, prop);
	365
	366	prop = of_find_property(node, "linux,crashkernel-size", NULL);
	367	if (prop)
	368	prom_remove_property(node, prop);
	369
	370	if (crashk_res.start != 0) {
	371	prom_add_property(node, &crashk_base_prop);
	372	crashk_size = crashk_res.end - crashk_res.start + 1;
	373	prom_add_property(node, &crashk_size_prop);
	374	}
	375
	376	of_node_put(node);
	377	}
	378
aa98c50d	379	static int __init kexec_setup(void)
593e537b ME	380	{
593e537b ME	381	export_htab_values();
35dd5432	382	export_crashk_values();
aa98c50d	383	return 0;
35dd5432	384	}
aa98c50d	385	__initcall(kexec_setup);