[linux-2.6-block.git] / arch / powerpc / platforms / cell / ras.c

/*
 * Copyright 2006-2008, IBM Corporation.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version
 * 2 of the License, or (at your option) any later version.
 */

#undef DEBUG

#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/smp.h>
#include <linux/reboot.h>
#include <linux/kexec.h>
#include <linux/crash_dump.h>

#include <asm/kexec.h>
#include <asm/reg.h>
#include <asm/io.h>
#include <asm/prom.h>
#include <asm/machdep.h>
#include <asm/rtas.h>
#include <asm/cell-regs.h>

#include "ras.h"


static void dump_fir(int cpu)
{
	struct cbe_pmd_regs __iomem *pregs = cbe_get_cpu_pmd_regs(cpu);
	struct cbe_iic_regs __iomem *iregs = cbe_get_cpu_iic_regs(cpu);

	if (pregs == NULL)
		return;

	/* Todo: do some nicer parsing of bits and based on them go down
	 * to other sub-units FIRs and not only IIC
	 */
	printk(KERN_ERR "Global Checkstop FIR    : 0x%016llx\n",
	       in_be64(&pregs->checkstop_fir));
	printk(KERN_ERR "Global Recoverable FIR  : 0x%016llx\n",
	       in_be64(&pregs->checkstop_fir));
	printk(KERN_ERR "Global MachineCheck FIR : 0x%016llx\n",
	       in_be64(&pregs->spec_att_mchk_fir));

	if (iregs == NULL)
		return;
	printk(KERN_ERR "IOC FIR                 : 0x%016llx\n",
	       in_be64(&iregs->ioc_fir));

}

void cbe_system_error_exception(struct pt_regs *regs)
{
	int cpu = smp_processor_id();

	printk(KERN_ERR "System Error Interrupt on CPU %d !\n", cpu);
	dump_fir(cpu);
	dump_stack();
}

void cbe_maintenance_exception(struct pt_regs *regs)
{
	int cpu = smp_processor_id();

	/*
	 * Nothing implemented for the maintenance interrupt at this point
	 */

	printk(KERN_ERR "Unhandled Maintenance interrupt on CPU %d !\n", cpu);
	dump_stack();
}

void cbe_thermal_exception(struct pt_regs *regs)
{
	int cpu = smp_processor_id();

	/*
	 * Nothing implemented for the thermal interrupt at this point
	 */

	printk(KERN_ERR "Unhandled Thermal interrupt on CPU %d !\n", cpu);
	dump_stack();
}

static int cbe_machine_check_handler(struct pt_regs *regs)
{
	int cpu = smp_processor_id();

	printk(KERN_ERR "Machine Check Interrupt on CPU %d !\n", cpu);
	dump_fir(cpu);

	/* No recovery from this code now, lets continue */
	return 0;
}

struct ptcal_area {
	struct list_head list;
	int nid;
	int order;
	struct page *pages;
};

static LIST_HEAD(ptcal_list);

static int ptcal_start_tok, ptcal_stop_tok;

static int __init cbe_ptcal_enable_on_node(int nid, int order)
{
	struct ptcal_area *area;
	int ret = -ENOMEM;
	unsigned long addr;

	if (is_kdump_kernel())
		rtas_call(ptcal_stop_tok, 1, 1, NULL, nid);

	area = kmalloc(sizeof(*area), GFP_KERNEL);
	if (!area)
		goto out_err;

	area->nid = nid;
	area->order = order;
	area->pages = alloc_pages_node(area->nid, GFP_KERNEL | GFP_THISNODE,
					area->order);

	if (!area->pages) {
		printk(KERN_WARNING "%s: no page on node %d\n",
			__func__, area->nid);
		goto out_free_area;
	}

	/*
	 * We move the ptcal area to the middle of the allocated
	 * page, in order to avoid prefetches in memcpy and similar
	 * functions stepping on it.
	 */
	addr = __pa(page_address(area->pages)) + (PAGE_SIZE >> 1);
	printk(KERN_DEBUG "%s: enabling PTCAL on node %d address=0x%016lx\n",
			__func__, area->nid, addr);

	ret = -EIO;
	if (rtas_call(ptcal_start_tok, 3, 1, NULL, area->nid,
				(unsigned int)(addr >> 32),
				(unsigned int)(addr & 0xffffffff))) {
		printk(KERN_ERR "%s: error enabling PTCAL on node %d!\n",
				__func__, nid);
		goto out_free_pages;
	}

	list_add(&area->list, &ptcal_list);

	return 0;

out_free_pages:
	__free_pages(area->pages, area->order);
out_free_area:
	kfree(area);
out_err:
	return ret;
}

static int __init cbe_ptcal_enable(void)
{
	const u32 *size;
	struct device_node *np;
	int order, found_mic = 0;

	np = of_find_node_by_path("/rtas");
	if (!np)
		return -ENODEV;

	size = of_get_property(np, "ibm,cbe-ptcal-size", NULL);
	if (!size)
		return -ENODEV;

	pr_debug("%s: enabling PTCAL, size = 0x%x\n", __func__, *size);
	order = get_order(*size);
	of_node_put(np);

	/* support for malta device trees, with be@/mic@ nodes */
	for_each_node_by_type(np, "mic-tm") {
		cbe_ptcal_enable_on_node(of_node_to_nid(np), order);
		found_mic = 1;
	}

	if (found_mic)
		return 0;

	/* support for older device tree - use cpu nodes */
	for_each_node_by_type(np, "cpu") {
		const u32 *nid = of_get_property(np, "node-id", NULL);
		if (!nid) {
			printk(KERN_ERR "%s: node %s is missing node-id?\n",
					__func__, np->full_name);
			continue;
		}
		cbe_ptcal_enable_on_node(*nid, order);
		found_mic = 1;
	}

	return found_mic ? 0 : -ENODEV;
}

static int cbe_ptcal_disable(void)
{
	struct ptcal_area *area, *tmp;
	int ret = 0;

	pr_debug("%s: disabling PTCAL\n", __func__);

	list_for_each_entry_safe(area, tmp, &ptcal_list, list) {
		/* disable ptcal on this node */
		if (rtas_call(ptcal_stop_tok, 1, 1, NULL, area->nid)) {
			printk(KERN_ERR "%s: error disabling PTCAL "
					"on node %d!\n", __func__,
					area->nid);
			ret = -EIO;
			continue;
		}

		/* ensure we can access the PTCAL area */
		memset(page_address(area->pages), 0,
				1 << (area->order + PAGE_SHIFT));

		/* clean up */
		list_del(&area->list);
		__free_pages(area->pages, area->order);
		kfree(area);
	}

	return ret;
}

static int cbe_ptcal_notify_reboot(struct notifier_block *nb,
		unsigned long code, void *data)
{
	return cbe_ptcal_disable();
}

static void cbe_ptcal_crash_shutdown(void)
{
	cbe_ptcal_disable();
}

static struct notifier_block cbe_ptcal_reboot_notifier = {
	.notifier_call = cbe_ptcal_notify_reboot
};

#ifdef CONFIG_PPC_IBM_CELL_RESETBUTTON
static int sysreset_hack;

static int __init cbe_sysreset_init(void)
{
	struct cbe_pmd_regs __iomem *regs;

	sysreset_hack = machine_is_compatible("IBM,CBPLUS-1.0");
	if (!sysreset_hack)
		return 0;

	regs = cbe_get_cpu_pmd_regs(0);
	if (!regs)
		return 0;

	/* Enable JTAG system-reset hack */
	out_be32(&regs->fir_mode_reg,
		in_be32(&regs->fir_mode_reg) |
		CBE_PMD_FIR_MODE_M8);

	return 0;
}
device_initcall(cbe_sysreset_init);

int cbe_sysreset_hack(void)
{
	struct cbe_pmd_regs __iomem *regs;

	/*
	 * The BMC can inject user triggered system reset exceptions,
	 * but cannot set the system reset reason in srr1,
	 * so check an extra register here.
	 */
	if (sysreset_hack && (smp_processor_id() == 0)) {
		regs = cbe_get_cpu_pmd_regs(0);
		if (!regs)
			return 0;
		if (in_be64(&regs->ras_esc_0) & 0x0000ffff) {
			out_be64(&regs->ras_esc_0, 0);
			return 0;
		}
	}
	return 1;
}
#endif /* CONFIG_PPC_IBM_CELL_RESETBUTTON */

int __init cbe_ptcal_init(void)
{
	int ret;
	ptcal_start_tok = rtas_token("ibm,cbe-start-ptcal");
	ptcal_stop_tok = rtas_token("ibm,cbe-stop-ptcal");

	if (ptcal_start_tok == RTAS_UNKNOWN_SERVICE
			|| ptcal_stop_tok == RTAS_UNKNOWN_SERVICE)
		return -ENODEV;

	ret = register_reboot_notifier(&cbe_ptcal_reboot_notifier);
	if (ret)
		goto out1;

	ret = crash_shutdown_register(&cbe_ptcal_crash_shutdown);
	if (ret)
		goto out2;

	return cbe_ptcal_enable();

out2:
	unregister_reboot_notifier(&cbe_ptcal_reboot_notifier);
out1:
	printk(KERN_ERR "Can't disable PTCAL, so not enabling\n");
	return ret;
}

arch_initcall(cbe_ptcal_init);

void __init cbe_ras_init(void)
{
	unsigned long hid0;

	/*
	 * Enable System Error & thermal interrupts and wakeup conditions
	 */

	hid0 = mfspr(SPRN_HID0);
	hid0 |= HID0_CBE_THERM_INT_EN | HID0_CBE_THERM_WAKEUP |
		HID0_CBE_SYSERR_INT_EN | HID0_CBE_SYSERR_WAKEUP;
	mtspr(SPRN_HID0, hid0);
	mb();

	/*
	 * Install machine check handler. Leave setting of precise mode to
	 * what the firmware did for now
	 */
	ppc_md.machine_check_exception = cbe_machine_check_handler;
	mb();

	/*
	 * For now, we assume that IOC_FIR is already set to forward some
	 * error conditions to the System Error handler. If that is not true
	 * then it will have to be fixed up here.
	 */
}
Commit	Line	Data
bdb226ba ME	1	/*
	2	* Copyright 2006-2008, IBM Corporation.
	3	*
	4	* This program is free software; you can redistribute it and/or
	5	* modify it under the terms of the GNU General Public License
	6	* as published by the Free Software Foundation; either version
	7	* 2 of the License, or (at your option) any later version.
	8	*/
	9
	10	#undef DEBUG
acf7d768	11
acf7d768 BH	12	#include <linux/types.h>
	13	#include <linux/kernel.h>
	14	#include <linux/smp.h>
150f7e3c	15	#include <linux/reboot.h>
62a8bd6c MM	16	#include <linux/kexec.h>
62a8bd6c MM	17	#include <linux/crash_dump.h>
acf7d768	18
9c8b4aff	19	#include <asm/kexec.h>
acf7d768 BH	20	#include <asm/reg.h>
	21	#include <asm/io.h>
	22	#include <asm/prom.h>
	23	#include <asm/machdep.h>
150f7e3c	24	#include <asm/rtas.h>
eef686a0	25	#include <asm/cell-regs.h>
acf7d768 BH	26
acf7d768 BH	27	#include "ras.h"
acf7d768 BH	28
	29
	30	static void dump_fir(int cpu)
	31	{
	32	struct cbe_pmd_regs __iomem *pregs = cbe_get_cpu_pmd_regs(cpu);
	33	struct cbe_iic_regs __iomem *iregs = cbe_get_cpu_iic_regs(cpu);
	34
	35	if (pregs == NULL)
	36	return;
	37
	38	/* Todo: do some nicer parsing of bits and based on them go down
	39	* to other sub-units FIRs and not only IIC
	40	*/
fe333321	41	printk(KERN_ERR "Global Checkstop FIR : 0x%016llx\n",
acf7d768	42	in_be64(&pregs->checkstop_fir));
fe333321	43	printk(KERN_ERR "Global Recoverable FIR : 0x%016llx\n",
acf7d768	44	in_be64(&pregs->checkstop_fir));
fe333321	45	printk(KERN_ERR "Global MachineCheck FIR : 0x%016llx\n",
acf7d768 BH	46	in_be64(&pregs->spec_att_mchk_fir));
	47
	48	if (iregs == NULL)
	49	return;
fe333321	50	printk(KERN_ERR "IOC FIR : 0x%016llx\n",
acf7d768 BH	51	in_be64(&iregs->ioc_fir));
	52
	53	}
	54
	55	void cbe_system_error_exception(struct pt_regs *regs)
	56	{
	57	int cpu = smp_processor_id();
	58
	59	printk(KERN_ERR "System Error Interrupt on CPU %d !\n", cpu);
	60	dump_fir(cpu);
	61	dump_stack();
	62	}
	63
	64	void cbe_maintenance_exception(struct pt_regs *regs)
	65	{
	66	int cpu = smp_processor_id();
	67
	68	/*
	69	* Nothing implemented for the maintenance interrupt at this point
	70	*/
	71
	72	printk(KERN_ERR "Unhandled Maintenance interrupt on CPU %d !\n", cpu);
	73	dump_stack();
	74	}
	75
	76	void cbe_thermal_exception(struct pt_regs *regs)
	77	{
	78	int cpu = smp_processor_id();
	79
	80	/*
	81	* Nothing implemented for the thermal interrupt at this point
	82	*/
	83
	84	printk(KERN_ERR "Unhandled Thermal interrupt on CPU %d !\n", cpu);
	85	dump_stack();
	86	}
	87
	88	static int cbe_machine_check_handler(struct pt_regs *regs)
	89	{
	90	int cpu = smp_processor_id();
	91
	92	printk(KERN_ERR "Machine Check Interrupt on CPU %d !\n", cpu);
	93	dump_fir(cpu);
	94
	95	/* No recovery from this code now, lets continue */
	96	return 0;
	97	}
	98
150f7e3c JK	99	struct ptcal_area {
	100	struct list_head list;
	101	int nid;
	102	int order;
	103	struct page *pages;
	104	};
	105
	106	static LIST_HEAD(ptcal_list);
	107
	108	static int ptcal_start_tok, ptcal_stop_tok;
	109
	110	static int __init cbe_ptcal_enable_on_node(int nid, int order)
	111	{
	112	struct ptcal_area *area;
	113	int ret = -ENOMEM;
	114	unsigned long addr;
	115
62a8bd6c	116	if (is_kdump_kernel())
54622f10	117	rtas_call(ptcal_stop_tok, 1, 1, NULL, nid);
150f7e3c JK	118
	119	area = kmalloc(sizeof(*area), GFP_KERNEL);
	120	if (!area)
	121	goto out_err;
	122
	123	area->nid = nid;
	124	area->order = order;
37cd8ed9 GS	125	area->pages = alloc_pages_node(area->nid, GFP_KERNEL \| GFP_THISNODE,
37cd8ed9 GS	126	area->order);
150f7e3c	127
37cd8ed9 GS	128	if (!area->pages) {
	129	printk(KERN_WARNING "%s: no page on node %d\n",
	130	__func__, area->nid);
150f7e3c	131	goto out_free_area;
37cd8ed9	132	}
150f7e3c	133
37cd8ed9 GS	134	/*
	135	* We move the ptcal area to the middle of the allocated
	136	* page, in order to avoid prefetches in memcpy and similar
	137	* functions stepping on it.
	138	*/
	139	addr = __pa(page_address(area->pages)) + (PAGE_SIZE >> 1);
	140	printk(KERN_DEBUG "%s: enabling PTCAL on node %d address=0x%016lx\n",
	141	__func__, area->nid, addr);
150f7e3c JK	142
	143	ret = -EIO;
	144	if (rtas_call(ptcal_start_tok, 3, 1, NULL, area->nid,
	145	(unsigned int)(addr >> 32),
	146	(unsigned int)(addr & 0xffffffff))) {
	147	printk(KERN_ERR "%s: error enabling PTCAL on node %d!\n",
e48b1b45	148	__func__, nid);
150f7e3c JK	149	goto out_free_pages;
	150	}
	151
	152	list_add(&area->list, &ptcal_list);
	153
	154	return 0;
	155
	156	out_free_pages:
	157	__free_pages(area->pages, area->order);
	158	out_free_area:
	159	kfree(area);
	160	out_err:
	161	return ret;
	162	}
	163
	164	static int __init cbe_ptcal_enable(void)
	165	{
	166	const u32 *size;
	167	struct device_node *np;
	168	int order, found_mic = 0;
	169
	170	np = of_find_node_by_path("/rtas");
	171	if (!np)
	172	return -ENODEV;
	173
12d371a6	174	size = of_get_property(np, "ibm,cbe-ptcal-size", NULL);
150f7e3c JK	175	if (!size)
	176	return -ENODEV;
	177
e48b1b45	178	pr_debug("%s: enabling PTCAL, size = 0x%x\n", __func__, *size);
150f7e3c JK	179	order = get_order(*size);
	180	of_node_put(np);
	181
	182	/* support for malta device trees, with be@/mic@ nodes */
	183	for_each_node_by_type(np, "mic-tm") {
	184	cbe_ptcal_enable_on_node(of_node_to_nid(np), order);
	185	found_mic = 1;
	186	}
	187
	188	if (found_mic)
	189	return 0;
	190
	191	/* support for older device tree - use cpu nodes */
	192	for_each_node_by_type(np, "cpu") {
12d371a6	193	const u32 *nid = of_get_property(np, "node-id", NULL);
150f7e3c JK	194	if (!nid) {
150f7e3c JK	195	printk(KERN_ERR "%s: node %s is missing node-id?\n",
e48b1b45	196	__func__, np->full_name);
150f7e3c JK	197	continue;
	198	}
	199	cbe_ptcal_enable_on_node(*nid, order);
	200	found_mic = 1;
	201	}
	202
	203	return found_mic ? 0 : -ENODEV;
	204	}
	205
	206	static int cbe_ptcal_disable(void)
	207	{
	208	struct ptcal_area area, tmp;
	209	int ret = 0;
	210
e48b1b45	211	pr_debug("%s: disabling PTCAL\n", __func__);
150f7e3c JK	212
	213	list_for_each_entry_safe(area, tmp, &ptcal_list, list) {
	214	/* disable ptcal on this node */
	215	if (rtas_call(ptcal_stop_tok, 1, 1, NULL, area->nid)) {
	216	printk(KERN_ERR "%s: error disabling PTCAL "
e48b1b45	217	"on node %d!\n", __func__,
150f7e3c JK	218	area->nid);
	219	ret = -EIO;
	220	continue;
	221	}
	222
	223	/* ensure we can access the PTCAL area */
	224	memset(page_address(area->pages), 0,
	225	1 << (area->order + PAGE_SHIFT));
	226
	227	/* clean up */
	228	list_del(&area->list);
	229	__free_pages(area->pages, area->order);
	230	kfree(area);
	231	}
	232
	233	return ret;
	234	}
	235
	236	static int cbe_ptcal_notify_reboot(struct notifier_block *nb,
	237	unsigned long code, void *data)
	238	{
	239	return cbe_ptcal_disable();
	240	}
	241
5acb0807 AB	242	static void cbe_ptcal_crash_shutdown(void)
	243	{
	244	cbe_ptcal_disable();
	245	}
	246
150f7e3c JK	247	static struct notifier_block cbe_ptcal_reboot_notifier = {
	248	.notifier_call = cbe_ptcal_notify_reboot
	249	};
	250
70694a8b CK	251	#ifdef CONFIG_PPC_IBM_CELL_RESETBUTTON
	252	static int sysreset_hack;
	253
	254	static int __init cbe_sysreset_init(void)
	255	{
	256	struct cbe_pmd_regs __iomem *regs;
	257
	258	sysreset_hack = machine_is_compatible("IBM,CBPLUS-1.0");
	259	if (!sysreset_hack)
	260	return 0;
	261
	262	regs = cbe_get_cpu_pmd_regs(0);
	263	if (!regs)
	264	return 0;
	265
	266	/* Enable JTAG system-reset hack */
	267	out_be32(&regs->fir_mode_reg,
	268	in_be32(&regs->fir_mode_reg) \|
	269	CBE_PMD_FIR_MODE_M8);
	270
	271	return 0;
	272	}
	273	device_initcall(cbe_sysreset_init);
	274
	275	int cbe_sysreset_hack(void)
	276	{
	277	struct cbe_pmd_regs __iomem *regs;
	278
	279	/*
	280	* The BMC can inject user triggered system reset exceptions,
	281	* but cannot set the system reset reason in srr1,
	282	* so check an extra register here.
	283	*/
	284	if (sysreset_hack && (smp_processor_id() == 0)) {
	285	regs = cbe_get_cpu_pmd_regs(0);
	286	if (!regs)
	287	return 0;
	288	if (in_be64(&regs->ras_esc_0) & 0x0000ffff) {
	289	out_be64(&regs->ras_esc_0, 0);
	290	return 0;
	291	}
	292	}
	293	return 1;
	294	}
	295	#endif /* CONFIG_PPC_IBM_CELL_RESETBUTTON */
	296
150f7e3c JK	297	int __init cbe_ptcal_init(void)
	298	{
	299	int ret;
	300	ptcal_start_tok = rtas_token("ibm,cbe-start-ptcal");
	301	ptcal_stop_tok = rtas_token("ibm,cbe-stop-ptcal");
	302
	303	if (ptcal_start_tok == RTAS_UNKNOWN_SERVICE
	304	\|\| ptcal_stop_tok == RTAS_UNKNOWN_SERVICE)
	305	return -ENODEV;
	306
	307	ret = register_reboot_notifier(&cbe_ptcal_reboot_notifier);
5acb0807 AB	308	if (ret)
	309	goto out1;
	310
	311	ret = crash_shutdown_register(&cbe_ptcal_crash_shutdown);
	312	if (ret)
	313	goto out2;
150f7e3c JK	314
150f7e3c JK	315	return cbe_ptcal_enable();
5acb0807 AB	316
	317	out2:
	318	unregister_reboot_notifier(&cbe_ptcal_reboot_notifier);
	319	out1:
	320	printk(KERN_ERR "Can't disable PTCAL, so not enabling\n");
	321	return ret;
150f7e3c JK	322	}
	323
	324	arch_initcall(cbe_ptcal_init);
	325
acf7d768 BH	326	void __init cbe_ras_init(void)
	327	{
	328	unsigned long hid0;
	329
	330	/*
	331	* Enable System Error & thermal interrupts and wakeup conditions
	332	*/
	333
	334	hid0 = mfspr(SPRN_HID0);
	335	hid0 \|= HID0_CBE_THERM_INT_EN \| HID0_CBE_THERM_WAKEUP \|
	336	HID0_CBE_SYSERR_INT_EN \| HID0_CBE_SYSERR_WAKEUP;
	337	mtspr(SPRN_HID0, hid0);
	338	mb();
	339
	340	/*
	341	* Install machine check handler. Leave setting of precise mode to
	342	* what the firmware did for now
	343	*/
	344	ppc_md.machine_check_exception = cbe_machine_check_handler;
	345	mb();
	346
	347	/*
	348	* For now, we assume that IOC_FIR is already set to forward some
	349	* error conditions to the System Error handler. If that is not true
	350	* then it will have to be fixed up here.
	351	*/
	352	}