[GitHub/mt8127/android_kernel_alcatel_ttab.git] / arch / powerpc / platforms / pseries / ras.c

/*
 * Copyright (C) 2001 Dave Engebretsen 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.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
 */

#include <linux/sched.h>
#include <linux/interrupt.h>
#include <linux/irq.h>
#include <linux/of.h>
#include <linux/fs.h>
#include <linux/reboot.h>

#include <asm/machdep.h>
#include <asm/rtas.h>
#include <asm/firmware.h>

#include "pseries.h"

static unsigned char ras_log_buf[RTAS_ERROR_LOG_MAX];
static DEFINE_SPINLOCK(ras_log_buf_lock);

static char global_mce_data_buf[RTAS_ERROR_LOG_MAX];
static DEFINE_PER_CPU(__u64, mce_data_buf);

static int ras_check_exception_token;

#define EPOW_SENSOR_TOKEN	9
#define EPOW_SENSOR_INDEX	0

static irqreturn_t ras_epow_interrupt(int irq, void *dev_id);
static irqreturn_t ras_error_interrupt(int irq, void *dev_id);


/*
 * Initialize handlers for the set of interrupts caused by hardware errors
 * and power system events.
 */
static int __init init_ras_IRQ(void)
{
	struct device_node *np;

	ras_check_exception_token = rtas_token("check-exception");

	/* Internal Errors */
	np = of_find_node_by_path("/event-sources/internal-errors");
	if (np != NULL) {
		request_event_sources_irqs(np, ras_error_interrupt,
					   "RAS_ERROR");
		of_node_put(np);
	}

	/* EPOW Events */
	np = of_find_node_by_path("/event-sources/epow-events");
	if (np != NULL) {
		request_event_sources_irqs(np, ras_epow_interrupt, "RAS_EPOW");
		of_node_put(np);
	}

	return 0;
}
subsys_initcall(init_ras_IRQ);

#define EPOW_SHUTDOWN_NORMAL				1
#define EPOW_SHUTDOWN_ON_UPS				2
#define EPOW_SHUTDOWN_LOSS_OF_CRITICAL_FUNCTIONS	3
#define EPOW_SHUTDOWN_AMBIENT_TEMPERATURE_TOO_HIGH	4

static void handle_system_shutdown(char event_modifier)
{
	switch (event_modifier) {
	case EPOW_SHUTDOWN_NORMAL:
		pr_emerg("Firmware initiated power off");
		orderly_poweroff(1);
		break;

	case EPOW_SHUTDOWN_ON_UPS:
		pr_emerg("Loss of power reported by firmware, system is "
			"running on UPS/battery");
		break;

	case EPOW_SHUTDOWN_LOSS_OF_CRITICAL_FUNCTIONS:
		pr_emerg("Loss of system critical functions reported by "
			"firmware");
		pr_emerg("Check RTAS error log for details");
		orderly_poweroff(1);
		break;

	case EPOW_SHUTDOWN_AMBIENT_TEMPERATURE_TOO_HIGH:
		pr_emerg("Ambient temperature too high reported by firmware");
		pr_emerg("Check RTAS error log for details");
		orderly_poweroff(1);
		break;

	default:
		pr_err("Unknown power/cooling shutdown event (modifier %d)",
			event_modifier);
	}
}

struct epow_errorlog {
	unsigned char sensor_value;
	unsigned char event_modifier;
	unsigned char extended_modifier;
	unsigned char reserved;
	unsigned char platform_reason;
};

#define EPOW_RESET			0
#define EPOW_WARN_COOLING		1
#define EPOW_WARN_POWER			2
#define EPOW_SYSTEM_SHUTDOWN		3
#define EPOW_SYSTEM_HALT		4
#define EPOW_MAIN_ENCLOSURE		5
#define EPOW_POWER_OFF			7

void rtas_parse_epow_errlog(struct rtas_error_log *log)
{
	struct pseries_errorlog *pseries_log;
	struct epow_errorlog *epow_log;
	char action_code;
	char modifier;

	pseries_log = get_pseries_errorlog(log, PSERIES_ELOG_SECT_ID_EPOW);
	if (pseries_log == NULL)
		return;

	epow_log = (struct epow_errorlog *)pseries_log->data;
	action_code = epow_log->sensor_value & 0xF;	/* bottom 4 bits */
	modifier = epow_log->event_modifier & 0xF;	/* bottom 4 bits */

	switch (action_code) {
	case EPOW_RESET:
		pr_err("Non critical power or cooling issue cleared");
		break;

	case EPOW_WARN_COOLING:
		pr_err("Non critical cooling issue reported by firmware");
		pr_err("Check RTAS error log for details");
		break;

	case EPOW_WARN_POWER:
		pr_err("Non critical power issue reported by firmware");
		pr_err("Check RTAS error log for details");
		break;

	case EPOW_SYSTEM_SHUTDOWN:
		handle_system_shutdown(epow_log->event_modifier);
		break;

	case EPOW_SYSTEM_HALT:
		pr_emerg("Firmware initiated power off");
		orderly_poweroff(1);
		break;

	case EPOW_MAIN_ENCLOSURE:
	case EPOW_POWER_OFF:
		pr_emerg("Critical power/cooling issue reported by firmware");
		pr_emerg("Check RTAS error log for details");
		pr_emerg("Immediate power off");
		emergency_sync();
		kernel_power_off();
		break;

	default:
		pr_err("Unknown power/cooling event (action code %d)",
			action_code);
	}
}

/* Handle environmental and power warning (EPOW) interrupts. */
static irqreturn_t ras_epow_interrupt(int irq, void *dev_id)
{
	int status;
	int state;
	int critical;

	status = rtas_get_sensor(EPOW_SENSOR_TOKEN, EPOW_SENSOR_INDEX, &state);

	if (state > 3)
		critical = 1;		/* Time Critical */
	else
		critical = 0;

	spin_lock(&ras_log_buf_lock);

	status = rtas_call(ras_check_exception_token, 6, 1, NULL,
			   RTAS_VECTOR_EXTERNAL_INTERRUPT,
			   virq_to_hw(irq),
			   RTAS_EPOW_WARNING,
			   critical, __pa(&ras_log_buf),
				rtas_get_error_log_max());

	log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, 0);

	rtas_parse_epow_errlog((struct rtas_error_log *)ras_log_buf);

	spin_unlock(&ras_log_buf_lock);
	return IRQ_HANDLED;
}

/*
 * Handle hardware error interrupts.
 *
 * RTAS check-exception is called to collect data on the exception.  If
 * the error is deemed recoverable, we log a warning and return.
 * For nonrecoverable errors, an error is logged and we stop all processing
 * as quickly as possible in order to prevent propagation of the failure.
 */
static irqreturn_t ras_error_interrupt(int irq, void *dev_id)
{
	struct rtas_error_log *rtas_elog;
	int status;
	int fatal;

	spin_lock(&ras_log_buf_lock);

	status = rtas_call(ras_check_exception_token, 6, 1, NULL,
			   RTAS_VECTOR_EXTERNAL_INTERRUPT,
			   virq_to_hw(irq),
			   RTAS_INTERNAL_ERROR, 1 /* Time Critical */,
			   __pa(&ras_log_buf),
				rtas_get_error_log_max());

	rtas_elog = (struct rtas_error_log *)ras_log_buf;

	if ((status == 0) && (rtas_elog->severity >= RTAS_SEVERITY_ERROR_SYNC))
		fatal = 1;
	else
		fatal = 0;

	/* format and print the extended information */
	log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, fatal);

	if (fatal) {
		pr_emerg("Fatal hardware error reported by firmware");
		pr_emerg("Check RTAS error log for details");
		pr_emerg("Immediate power off");
		emergency_sync();
		kernel_power_off();
	} else {
		pr_err("Recoverable hardware error reported by firmware");
	}

	spin_unlock(&ras_log_buf_lock);
	return IRQ_HANDLED;
}

/*
 * Some versions of FWNMI place the buffer inside the 4kB page starting at
 * 0x7000. Other versions place it inside the rtas buffer. We check both.
 */
#define VALID_FWNMI_BUFFER(A) \
	((((A) >= 0x7000) && ((A) < 0x7ff0)) || \
	(((A) >= rtas.base) && ((A) < (rtas.base + rtas.size - 16))))

/*
 * Get the error information for errors coming through the
 * FWNMI vectors.  The pt_regs' r3 will be updated to reflect
 * the actual r3 if possible, and a ptr to the error log entry
 * will be returned if found.
 *
 * If the RTAS error is not of the extended type, then we put it in a per
 * cpu 64bit buffer. If it is the extended type we use global_mce_data_buf.
 *
 * The global_mce_data_buf does not have any locks or protection around it,
 * if a second machine check comes in, or a system reset is done
 * before we have logged the error, then we will get corruption in the
 * error log.  This is preferable over holding off on calling
 * ibm,nmi-interlock which would result in us checkstopping if a
 * second machine check did come in.
 */
static struct rtas_error_log *fwnmi_get_errinfo(struct pt_regs *regs)
{
	unsigned long *savep;
	struct rtas_error_log *h, *errhdr = NULL;

	if (!VALID_FWNMI_BUFFER(regs->gpr[3])) {
		printk(KERN_ERR "FWNMI: corrupt r3 0x%016lx\n", regs->gpr[3]);
		return NULL;
	}

	savep = __va(regs->gpr[3]);
	regs->gpr[3] = savep[0];	/* restore original r3 */

	/* If it isn't an extended log we can use the per cpu 64bit buffer */
	h = (struct rtas_error_log *)&savep[1];
	if (!h->extended) {
		memcpy(&__get_cpu_var(mce_data_buf), h, sizeof(__u64));
		errhdr = (struct rtas_error_log *)&__get_cpu_var(mce_data_buf);
	} else {
		int len;

		len = max_t(int, 8+h->extended_log_length, RTAS_ERROR_LOG_MAX);
		memset(global_mce_data_buf, 0, RTAS_ERROR_LOG_MAX);
		memcpy(global_mce_data_buf, h, len);
		errhdr = (struct rtas_error_log *)global_mce_data_buf;
	}

	return errhdr;
}

/* Call this when done with the data returned by FWNMI_get_errinfo.
 * It will release the saved data area for other CPUs in the
 * partition to receive FWNMI errors.
 */
static void fwnmi_release_errinfo(void)
{
	int ret = rtas_call(rtas_token("ibm,nmi-interlock"), 0, 1, NULL);
	if (ret != 0)
		printk(KERN_ERR "FWNMI: nmi-interlock failed: %d\n", ret);
}

int pSeries_system_reset_exception(struct pt_regs *regs)
{
	if (fwnmi_active) {
		struct rtas_error_log *errhdr = fwnmi_get_errinfo(regs);
		if (errhdr) {
			/* XXX Should look at FWNMI information */
		}
		fwnmi_release_errinfo();
	}
	return 0; /* need to perform reset */
}

/*
 * See if we can recover from a machine check exception.
 * This is only called on power4 (or above) and only via
 * the Firmware Non-Maskable Interrupts (fwnmi) handler
 * which provides the error analysis for us.
 *
 * Return 1 if corrected (or delivered a signal).
 * Return 0 if there is nothing we can do.
 */
static int recover_mce(struct pt_regs *regs, struct rtas_error_log *err)
{
	int recovered = 0;

	if (!(regs->msr & MSR_RI)) {
		/* If MSR_RI isn't set, we cannot recover */
		recovered = 0;

	} else if (err->disposition == RTAS_DISP_FULLY_RECOVERED) {
		/* Platform corrected itself */
		recovered = 1;

	} else if (err->disposition == RTAS_DISP_LIMITED_RECOVERY) {
		/* Platform corrected itself but could be degraded */
		printk(KERN_ERR "MCE: limited recovery, system may "
		       "be degraded\n");
		recovered = 1;

	} else if (user_mode(regs) && !is_global_init(current) &&
		   err->severity == RTAS_SEVERITY_ERROR_SYNC) {

		/*
		 * If we received a synchronous error when in userspace
		 * kill the task. Firmware may report details of the fail
		 * asynchronously, so we can't rely on the target and type
		 * fields being valid here.
		 */
		printk(KERN_ERR "MCE: uncorrectable error, killing task "
		       "%s:%d\n", current->comm, current->pid);

		_exception(SIGBUS, regs, BUS_MCEERR_AR, regs->nip);
		recovered = 1;
	}

	log_error((char *)err, ERR_TYPE_RTAS_LOG, 0);

	return recovered;
}

/*
 * Handle a machine check.
 *
 * Note that on Power 4 and beyond Firmware Non-Maskable Interrupts (fwnmi)
 * should be present.  If so the handler which called us tells us if the
 * error was recovered (never true if RI=0).
 *
 * On hardware prior to Power 4 these exceptions were asynchronous which
 * means we can't tell exactly where it occurred and so we can't recover.
 */
int pSeries_machine_check_exception(struct pt_regs *regs)
{
	struct rtas_error_log *errp;

	if (fwnmi_active) {
		errp = fwnmi_get_errinfo(regs);
		fwnmi_release_errinfo();
		if (errp && recover_mce(regs, errp))
			return 1;
	}

	return 0;
}
Commit	Line	Data
1da177e4	1	/*
1da177e4	2	* Copyright (C) 2001 Dave Engebretsen IBM Corporation
d9953105	3	*
1da177e4 LT	4	* This program is free software; you can redistribute it and/or modify
	5	* it under the terms of the GNU General Public License as published by
	6	* the Free Software Foundation; either version 2 of the License, or
	7	* (at your option) any later version.
d9953105	8	*
1da177e4 LT	9	* This program is distributed in the hope that it will be useful,
	10	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	11	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	12	* GNU General Public License for more details.
d9953105	13	*
1da177e4 LT	14	* You should have received a copy of the GNU General Public License
	15	* along with this program; if not, write to the Free Software
	16	* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
	17	*/
	18
1da177e4	19	#include <linux/sched.h>
1da177e4	20	#include <linux/interrupt.h>
1da177e4	21	#include <linux/irq.h>
90128997	22	#include <linux/of.h>
55fc0c56 AB	23	#include <linux/fs.h>
55fc0c56 AB	24	#include <linux/reboot.h>
1da177e4	25
1da177e4 LT	26	#include <asm/machdep.h>
1da177e4 LT	27	#include <asm/rtas.h>
8c4f1f29	28	#include <asm/firmware.h>
1da177e4	29
577830b0	30	#include "pseries.h"
c902be71	31
1da177e4 LT	32	static unsigned char ras_log_buf[RTAS_ERROR_LOG_MAX];
	33	static DEFINE_SPINLOCK(ras_log_buf_lock);
	34
d368514c AB	35	static char global_mce_data_buf[RTAS_ERROR_LOG_MAX];
d368514c AB	36	static DEFINE_PER_CPU(__u64, mce_data_buf);
1da177e4	37
1da177e4 LT	38	static int ras_check_exception_token;
	39
	40	#define EPOW_SENSOR_TOKEN 9
	41	#define EPOW_SENSOR_INDEX 0
1da177e4	42
7d12e780 DH	43	static irqreturn_t ras_epow_interrupt(int irq, void *dev_id);
7d12e780 DH	44	static irqreturn_t ras_error_interrupt(int irq, void *dev_id);
1da177e4	45
0ebfff14	46
1da177e4 LT	47	/*
	48	* Initialize handlers for the set of interrupts caused by hardware errors
	49	* and power system events.
	50	*/
	51	static int __init init_ras_IRQ(void)
	52	{
	53	struct device_node *np;
	54
1da177e4 LT	55	ras_check_exception_token = rtas_token("check-exception");
	56
	57	/* Internal Errors */
	58	np = of_find_node_by_path("/event-sources/internal-errors");
	59	if (np != NULL) {
32c96f77 MN	60	request_event_sources_irqs(np, ras_error_interrupt,
32c96f77 MN	61	"RAS_ERROR");
1da177e4 LT	62	of_node_put(np);
	63	}
	64
	65	/* EPOW Events */
	66	np = of_find_node_by_path("/event-sources/epow-events");
	67	if (np != NULL) {
32c96f77	68	request_event_sources_irqs(np, ras_epow_interrupt, "RAS_EPOW");
1da177e4 LT	69	of_node_put(np);
	70	}
	71
69ed3324	72	return 0;
1da177e4	73	}
55fc0c56	74	subsys_initcall(init_ras_IRQ);
1da177e4	75
55fc0c56 AB	76	#define EPOW_SHUTDOWN_NORMAL 1
	77	#define EPOW_SHUTDOWN_ON_UPS 2
	78	#define EPOW_SHUTDOWN_LOSS_OF_CRITICAL_FUNCTIONS 3
	79	#define EPOW_SHUTDOWN_AMBIENT_TEMPERATURE_TOO_HIGH 4
	80
	81	static void handle_system_shutdown(char event_modifier)
	82	{
	83	switch (event_modifier) {
	84	case EPOW_SHUTDOWN_NORMAL:
	85	pr_emerg("Firmware initiated power off");
	86	orderly_poweroff(1);
	87	break;
	88
	89	case EPOW_SHUTDOWN_ON_UPS:
	90	pr_emerg("Loss of power reported by firmware, system is "
	91	"running on UPS/battery");
	92	break;
	93
	94	case EPOW_SHUTDOWN_LOSS_OF_CRITICAL_FUNCTIONS:
	95	pr_emerg("Loss of system critical functions reported by "
	96	"firmware");
	97	pr_emerg("Check RTAS error log for details");
	98	orderly_poweroff(1);
	99	break;
	100
	101	case EPOW_SHUTDOWN_AMBIENT_TEMPERATURE_TOO_HIGH:
	102	pr_emerg("Ambient temperature too high reported by firmware");
	103	pr_emerg("Check RTAS error log for details");
	104	orderly_poweroff(1);
	105	break;
	106
	107	default:
	108	pr_err("Unknown power/cooling shutdown event (modifier %d)",
	109	event_modifier);
	110	}
	111	}
	112
	113	struct epow_errorlog {
	114	unsigned char sensor_value;
	115	unsigned char event_modifier;
	116	unsigned char extended_modifier;
	117	unsigned char reserved;
	118	unsigned char platform_reason;
	119	};
	120
	121	#define EPOW_RESET 0
	122	#define EPOW_WARN_COOLING 1
	123	#define EPOW_WARN_POWER 2
	124	#define EPOW_SYSTEM_SHUTDOWN 3
	125	#define EPOW_SYSTEM_HALT 4
	126	#define EPOW_MAIN_ENCLOSURE 5
	127	#define EPOW_POWER_OFF 7
	128
	129	void rtas_parse_epow_errlog(struct rtas_error_log *log)
	130	{
	131	struct pseries_errorlog *pseries_log;
	132	struct epow_errorlog *epow_log;
	133	char action_code;
	134	char modifier;
	135
	136	pseries_log = get_pseries_errorlog(log, PSERIES_ELOG_SECT_ID_EPOW);
	137	if (pseries_log == NULL)
	138	return;
	139
140	epow_log = (struct epow_errorlog *)pseries_log->data;
141	action_code = epow_log->sensor_value & 0xF; /* bottom 4 bits */
142	modifier = epow_log->event_modifier & 0xF; /* bottom 4 bits */
143
144	switch (action_code) {
145	case EPOW_RESET:
146	pr_err("Non critical power or cooling issue cleared");
147	break;
148
149	case EPOW_WARN_COOLING:
150	pr_err("Non critical cooling issue reported by firmware");
151	pr_err("Check RTAS error log for details");
152	break;
153
154	case EPOW_WARN_POWER:
155	pr_err("Non critical power issue reported by firmware");
156	pr_err("Check RTAS error log for details");
157	break;
158
159	case EPOW_SYSTEM_SHUTDOWN:
160	handle_system_shutdown(epow_log->event_modifier);
161	break;
162
163	case EPOW_SYSTEM_HALT:
164	pr_emerg("Firmware initiated power off");
165	orderly_poweroff(1);
166	break;
167
168	case EPOW_MAIN_ENCLOSURE:
169	case EPOW_POWER_OFF:
170	pr_emerg("Critical power/cooling issue reported by firmware");
171	pr_emerg("Check RTAS error log for details");
172	pr_emerg("Immediate power off");
173	emergency_sync();
174	kernel_power_off();
175	break;
176
177	default:
178	pr_err("Unknown power/cooling event (action code %d)",
179	action_code);
180	}
181	}
182
183	/* Handle environmental and power warning (EPOW) interrupts. */
7d12e780	184	static irqreturn_t ras_epow_interrupt(int irq, void *dev_id)
1da177e4	185	{
55fc0c56 AB	186	int status;
55fc0c56 AB	187	int state;
1da177e4 LT	188	int critical;
1da177e4 LT	189
587f83e8	190	status = rtas_get_sensor(EPOW_SENSOR_TOKEN, EPOW_SENSOR_INDEX, &state);
1da177e4 LT	191
1da177e4 LT	192	if (state > 3)
55fc0c56	193	critical = 1; /* Time Critical */
1da177e4 LT	194	else
	195	critical = 0;
	196
	197	spin_lock(&ras_log_buf_lock);
	198
	199	status = rtas_call(ras_check_exception_token, 6, 1, NULL,
b08e281b	200	RTAS_VECTOR_EXTERNAL_INTERRUPT,
476eb491	201	virq_to_hw(irq),
6f43747f	202	RTAS_EPOW_WARNING,
1da177e4 LT	203	critical, __pa(&ras_log_buf),
	204	rtas_get_error_log_max());
	205
1da177e4 LT	206	log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, 0);
1da177e4 LT	207
55fc0c56 AB	208	rtas_parse_epow_errlog((struct rtas_error_log *)ras_log_buf);
55fc0c56 AB	209
1da177e4 LT	210	spin_unlock(&ras_log_buf_lock);
	211	return IRQ_HANDLED;
	212	}
	213
	214	/*
	215	* Handle hardware error interrupts.
	216	*
	217	* RTAS check-exception is called to collect data on the exception. If
	218	* the error is deemed recoverable, we log a warning and return.
	219	* For nonrecoverable errors, an error is logged and we stop all processing
	220	* as quickly as possible in order to prevent propagation of the failure.
	221	*/
7d12e780	222	static irqreturn_t ras_error_interrupt(int irq, void *dev_id)
1da177e4 LT	223	{
1da177e4 LT	224	struct rtas_error_log *rtas_elog;
cc8b5263	225	int status;
1da177e4 LT	226	int fatal;
	227
	228	spin_lock(&ras_log_buf_lock);
	229
	230	status = rtas_call(ras_check_exception_token, 6, 1, NULL,
b08e281b	231	RTAS_VECTOR_EXTERNAL_INTERRUPT,
476eb491	232	virq_to_hw(irq),
cc8b5263	233	RTAS_INTERNAL_ERROR, 1 /* Time Critical */,
1da177e4 LT	234	__pa(&ras_log_buf),
	235	rtas_get_error_log_max());
	236
	237	rtas_elog = (struct rtas_error_log *)ras_log_buf;
	238
	239	if ((status == 0) && (rtas_elog->severity >= RTAS_SEVERITY_ERROR_SYNC))
	240	fatal = 1;
	241	else
	242	fatal = 0;
	243
	244	/* format and print the extended information */
	245	log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, fatal);
	246
	247	if (fatal) {
cc8b5263 AB	248	pr_emerg("Fatal hardware error reported by firmware");
	249	pr_emerg("Check RTAS error log for details");
	250	pr_emerg("Immediate power off");
	251	emergency_sync();
	252	kernel_power_off();
1da177e4	253	} else {
cc8b5263	254	pr_err("Recoverable hardware error reported by firmware");
1da177e4 LT	255	}
	256
	257	spin_unlock(&ras_log_buf_lock);
	258	return IRQ_HANDLED;
	259	}
	260
d368514c AB	261	/*
	262	* Some versions of FWNMI place the buffer inside the 4kB page starting at
	263	* 0x7000. Other versions place it inside the rtas buffer. We check both.
	264	*/
	265	#define VALID_FWNMI_BUFFER(A) \
	266	((((A) >= 0x7000) && ((A) < 0x7ff0)) \|\| \
	267	(((A) >= rtas.base) && ((A) < (rtas.base + rtas.size - 16))))
	268
	269	/*
	270	* Get the error information for errors coming through the
1da177e4 LT	271	* FWNMI vectors. The pt_regs' r3 will be updated to reflect
	272	* the actual r3 if possible, and a ptr to the error log entry
	273	* will be returned if found.
	274	*
d368514c AB	275	* If the RTAS error is not of the extended type, then we put it in a per
	276	* cpu 64bit buffer. If it is the extended type we use global_mce_data_buf.
	277	*
	278	* The global_mce_data_buf does not have any locks or protection around it,
1da177e4 LT	279	* if a second machine check comes in, or a system reset is done
	280	* before we have logged the error, then we will get corruption in the
	281	* error log. This is preferable over holding off on calling
	282	* ibm,nmi-interlock which would result in us checkstopping if a
	283	* second machine check did come in.
	284	*/
	285	static struct rtas_error_log fwnmi_get_errinfo(struct pt_regs regs)
	286	{
1da177e4	287	unsigned long *savep;
d368514c	288	struct rtas_error_log h, errhdr = NULL;
1da177e4	289
d368514c	290	if (!VALID_FWNMI_BUFFER(regs->gpr[3])) {
f0e939ae	291	printk(KERN_ERR "FWNMI: corrupt r3 0x%016lx\n", regs->gpr[3]);
d368514c AB	292	return NULL;
	293	}
	294
	295	savep = __va(regs->gpr[3]);
	296	regs->gpr[3] = savep[0]; /* restore original r3 */
	297
	298	/* If it isn't an extended log we can use the per cpu 64bit buffer */
	299	h = (struct rtas_error_log *)&savep[1];
	300	if (!h->extended) {
	301	memcpy(&__get_cpu_var(mce_data_buf), h, sizeof(__u64));
	302	errhdr = (struct rtas_error_log *)&__get_cpu_var(mce_data_buf);
1da177e4	303	} else {
d368514c AB	304	int len;
	305
	306	len = max_t(int, 8+h->extended_log_length, RTAS_ERROR_LOG_MAX);
	307	memset(global_mce_data_buf, 0, RTAS_ERROR_LOG_MAX);
	308	memcpy(global_mce_data_buf, h, len);
	309	errhdr = (struct rtas_error_log *)global_mce_data_buf;
1da177e4	310	}
d368514c	311
1da177e4 LT	312	return errhdr;
	313	}
	314
	315	/* Call this when done with the data returned by FWNMI_get_errinfo.
	316	* It will release the saved data area for other CPUs in the
	317	* partition to receive FWNMI errors.
	318	*/
	319	static void fwnmi_release_errinfo(void)
	320	{
	321	int ret = rtas_call(rtas_token("ibm,nmi-interlock"), 0, 1, NULL);
	322	if (ret != 0)
d368514c	323	printk(KERN_ERR "FWNMI: nmi-interlock failed: %d\n", ret);
1da177e4 LT	324	}
1da177e4 LT	325
c902be71	326	int pSeries_system_reset_exception(struct pt_regs *regs)
1da177e4 LT	327	{
	328	if (fwnmi_active) {
	329	struct rtas_error_log *errhdr = fwnmi_get_errinfo(regs);
	330	if (errhdr) {
	331	/* XXX Should look at FWNMI information */
	332	}
	333	fwnmi_release_errinfo();
	334	}
c902be71	335	return 0; /* need to perform reset */
1da177e4 LT	336	}
	337
	338	/*
	339	* See if we can recover from a machine check exception.
	340	* This is only called on power4 (or above) and only via
	341	* the Firmware Non-Maskable Interrupts (fwnmi) handler
	342	* which provides the error analysis for us.
	343	*
	344	* Return 1 if corrected (or delivered a signal).
	345	* Return 0 if there is nothing we can do.
	346	*/
d47d1d8a	347	static int recover_mce(struct pt_regs regs, struct rtas_error_log err)
1da177e4	348	{
d47d1d8a	349	int recovered = 0;
1da177e4	350
d47d1d8a AB	351	if (!(regs->msr & MSR_RI)) {
	352	/* If MSR_RI isn't set, we cannot recover */
	353	recovered = 0;
	354
	355	} else if (err->disposition == RTAS_DISP_FULLY_RECOVERED) {
1da177e4	356	/* Platform corrected itself */
d47d1d8a AB	357	recovered = 1;
	358
	359	} else if (err->disposition == RTAS_DISP_LIMITED_RECOVERY) {
	360	/* Platform corrected itself but could be degraded */
	361	printk(KERN_ERR "MCE: limited recovery, system may "
	362	"be degraded\n");
	363	recovered = 1;
	364
	365	} else if (user_mode(regs) && !is_global_init(current) &&
	366	err->severity == RTAS_SEVERITY_ERROR_SYNC) {
	367
	368	/*
	369	* If we received a synchronous error when in userspace
	370	* kill the task. Firmware may report details of the fail
	371	* asynchronously, so we can't rely on the target and type
	372	* fields being valid here.
	373	*/
	374	printk(KERN_ERR "MCE: uncorrectable error, killing task "
	375	"%s:%d\n", current->comm, current->pid);
	376
	377	_exception(SIGBUS, regs, BUS_MCEERR_AR, regs->nip);
	378	recovered = 1;
1da177e4 LT	379	}
1da177e4 LT	380
3f9793e6	381	log_error((char *)err, ERR_TYPE_RTAS_LOG, 0);
1da177e4	382
d47d1d8a	383	return recovered;
1da177e4 LT	384	}
	385
	386	/*
	387	* Handle a machine check.
	388	*
	389	* Note that on Power 4 and beyond Firmware Non-Maskable Interrupts (fwnmi)
	390	* should be present. If so the handler which called us tells us if the
	391	* error was recovered (never true if RI=0).
	392	*
	393	* On hardware prior to Power 4 these exceptions were asynchronous which
	394	* means we can't tell exactly where it occurred and so we can't recover.
	395	*/
	396	int pSeries_machine_check_exception(struct pt_regs *regs)
	397	{
	398	struct rtas_error_log *errp;
	399
	400	if (fwnmi_active) {
	401	errp = fwnmi_get_errinfo(regs);
	402	fwnmi_release_errinfo();
	403	if (errp && recover_mce(regs, errp))
	404	return 1;
	405	}
	406
	407	return 0;
	408	}