[mirror_ubuntu-zesty-kernel.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;
}
machine_subsys_initcall(pseries, 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(true);
		break;

	case EPOW_SHUTDOWN_ON_UPS:
		pr_emerg("Loss of power reported by firmware, system is "
			"running on UPS/battery");
		pr_emerg("Check RTAS error log for details");
		orderly_poweroff(true);
		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(true);
		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(true);
		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

static 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(true);
		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_fast(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_error_severity(rtas_elog) >= 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;

	/* Mask top two bits */
	regs->gpr[3] &= ~(0x3UL << 62);

	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 (!rtas_error_extended(h)) {
		memcpy(this_cpu_ptr(&mce_data_buf), h, sizeof(__u64));
		errhdr = (struct rtas_error_log *)this_cpu_ptr(&mce_data_buf);
	} else {
		int len, error_log_length;

		error_log_length = 8 + rtas_error_extended_log_length(h);
		len = max_t(int, error_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;
	int disposition = rtas_error_disposition(err);

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

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

	} else if (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) &&
		   rtas_error_severity(err) == 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	}
8e83e905	74	machine_subsys_initcall(pseries, 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");
1b7e0cbe	86	orderly_poweroff(true);
55fc0c56 AB	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");
79872e35 AK	92	pr_emerg("Check RTAS error log for details");
79872e35 AK	93	orderly_poweroff(true);
55fc0c56 AB	94	break;
	95
	96	case EPOW_SHUTDOWN_LOSS_OF_CRITICAL_FUNCTIONS:
	97	pr_emerg("Loss of system critical functions reported by "
	98	"firmware");
	99	pr_emerg("Check RTAS error log for details");
1b7e0cbe	100	orderly_poweroff(true);
55fc0c56 AB	101	break;
	102
	103	case EPOW_SHUTDOWN_AMBIENT_TEMPERATURE_TOO_HIGH:
	104	pr_emerg("Ambient temperature too high reported by firmware");
	105	pr_emerg("Check RTAS error log for details");
1b7e0cbe	106	orderly_poweroff(true);
55fc0c56 AB	107	break;
	108
	109	default:
	110	pr_err("Unknown power/cooling shutdown event (modifier %d)",
	111	event_modifier);
	112	}
	113	}
	114
	115	struct epow_errorlog {
	116	unsigned char sensor_value;
	117	unsigned char event_modifier;
	118	unsigned char extended_modifier;
	119	unsigned char reserved;
	120	unsigned char platform_reason;
	121	};
	122
	123	#define EPOW_RESET 0
	124	#define EPOW_WARN_COOLING 1
	125	#define EPOW_WARN_POWER 2
	126	#define EPOW_SYSTEM_SHUTDOWN 3
	127	#define EPOW_SYSTEM_HALT 4
	128	#define EPOW_MAIN_ENCLOSURE 5
	129	#define EPOW_POWER_OFF 7
	130
e51df2c1	131	static void rtas_parse_epow_errlog(struct rtas_error_log *log)
55fc0c56 AB	132	{
	133	struct pseries_errorlog *pseries_log;
	134	struct epow_errorlog *epow_log;
	135	char action_code;
	136	char modifier;
	137
	138	pseries_log = get_pseries_errorlog(log, PSERIES_ELOG_SECT_ID_EPOW);
	139	if (pseries_log == NULL)
	140	return;
	141
	142	epow_log = (struct epow_errorlog *)pseries_log->data;
	143	action_code = epow_log->sensor_value & 0xF; /* bottom 4 bits */
	144	modifier = epow_log->event_modifier & 0xF; /* bottom 4 bits */
	145
	146	switch (action_code) {
	147	case EPOW_RESET:
	148	pr_err("Non critical power or cooling issue cleared");
	149	break;
	150
	151	case EPOW_WARN_COOLING:
	152	pr_err("Non critical cooling issue reported by firmware");
	153	pr_err("Check RTAS error log for details");
	154	break;
	155
	156	case EPOW_WARN_POWER:
	157	pr_err("Non critical power issue reported by firmware");
	158	pr_err("Check RTAS error log for details");
	159	break;
	160
	161	case EPOW_SYSTEM_SHUTDOWN:
	162	handle_system_shutdown(epow_log->event_modifier);
	163	break;
	164
	165	case EPOW_SYSTEM_HALT:
	166	pr_emerg("Firmware initiated power off");
1b7e0cbe	167	orderly_poweroff(true);
55fc0c56 AB	168	break;
	169
	170	case EPOW_MAIN_ENCLOSURE:
	171	case EPOW_POWER_OFF:
	172	pr_emerg("Critical power/cooling issue reported by firmware");
	173	pr_emerg("Check RTAS error log for details");
	174	pr_emerg("Immediate power off");
	175	emergency_sync();
	176	kernel_power_off();
	177	break;
	178
	179	default:
	180	pr_err("Unknown power/cooling event (action code %d)",
	181	action_code);
	182	}
	183	}
	184
	185	/* Handle environmental and power warning (EPOW) interrupts. */
7d12e780	186	static irqreturn_t ras_epow_interrupt(int irq, void *dev_id)
1da177e4	187	{
55fc0c56 AB	188	int status;
55fc0c56 AB	189	int state;
1da177e4 LT	190	int critical;
1da177e4 LT	191
1c2cb594 TH	192	status = rtas_get_sensor_fast(EPOW_SENSOR_TOKEN, EPOW_SENSOR_INDEX,
1c2cb594 TH	193	&state);
1da177e4 LT	194
1da177e4 LT	195	if (state > 3)
55fc0c56	196	critical = 1; /* Time Critical */
1da177e4 LT	197	else
	198	critical = 0;
	199
	200	spin_lock(&ras_log_buf_lock);
	201
	202	status = rtas_call(ras_check_exception_token, 6, 1, NULL,
b08e281b	203	RTAS_VECTOR_EXTERNAL_INTERRUPT,
476eb491	204	virq_to_hw(irq),
6f43747f	205	RTAS_EPOW_WARNING,
1da177e4 LT	206	critical, __pa(&ras_log_buf),
	207	rtas_get_error_log_max());
	208
1da177e4 LT	209	log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, 0);
1da177e4 LT	210
55fc0c56 AB	211	rtas_parse_epow_errlog((struct rtas_error_log *)ras_log_buf);
55fc0c56 AB	212
1da177e4 LT	213	spin_unlock(&ras_log_buf_lock);
	214	return IRQ_HANDLED;
	215	}
	216
	217	/*
	218	* Handle hardware error interrupts.
	219	*
	220	* RTAS check-exception is called to collect data on the exception. If
	221	* the error is deemed recoverable, we log a warning and return.
	222	* For nonrecoverable errors, an error is logged and we stop all processing
	223	* as quickly as possible in order to prevent propagation of the failure.
	224	*/
7d12e780	225	static irqreturn_t ras_error_interrupt(int irq, void *dev_id)
1da177e4 LT	226	{
1da177e4 LT	227	struct rtas_error_log *rtas_elog;
cc8b5263	228	int status;
1da177e4 LT	229	int fatal;
	230
	231	spin_lock(&ras_log_buf_lock);
	232
	233	status = rtas_call(ras_check_exception_token, 6, 1, NULL,
b08e281b	234	RTAS_VECTOR_EXTERNAL_INTERRUPT,
476eb491	235	virq_to_hw(irq),
cc8b5263	236	RTAS_INTERNAL_ERROR, 1 /* Time Critical */,
1da177e4 LT	237	__pa(&ras_log_buf),
	238	rtas_get_error_log_max());
	239
	240	rtas_elog = (struct rtas_error_log *)ras_log_buf;
	241
a08a53ea GK	242	if (status == 0 &&
a08a53ea GK	243	rtas_error_severity(rtas_elog) >= RTAS_SEVERITY_ERROR_SYNC)
1da177e4 LT	244	fatal = 1;
	245	else
	246	fatal = 0;
	247
	248	/* format and print the extended information */
	249	log_error(ras_log_buf, ERR_TYPE_RTAS_LOG, fatal);
	250
	251	if (fatal) {
cc8b5263 AB	252	pr_emerg("Fatal hardware error reported by firmware");
	253	pr_emerg("Check RTAS error log for details");
	254	pr_emerg("Immediate power off");
	255	emergency_sync();
	256	kernel_power_off();
1da177e4	257	} else {
cc8b5263	258	pr_err("Recoverable hardware error reported by firmware");
1da177e4 LT	259	}
	260
	261	spin_unlock(&ras_log_buf_lock);
	262	return IRQ_HANDLED;
	263	}
	264
d368514c AB	265	/*
	266	* Some versions of FWNMI place the buffer inside the 4kB page starting at
	267	* 0x7000. Other versions place it inside the rtas buffer. We check both.
	268	*/
	269	#define VALID_FWNMI_BUFFER(A) \
	270	((((A) >= 0x7000) && ((A) < 0x7ff0)) \|\| \
	271	(((A) >= rtas.base) && ((A) < (rtas.base + rtas.size - 16))))
	272
	273	/*
	274	* Get the error information for errors coming through the
1da177e4 LT	275	* FWNMI vectors. The pt_regs' r3 will be updated to reflect
	276	* the actual r3 if possible, and a ptr to the error log entry
	277	* will be returned if found.
	278	*
d368514c AB	279	* If the RTAS error is not of the extended type, then we put it in a per
	280	* cpu 64bit buffer. If it is the extended type we use global_mce_data_buf.
	281	*
	282	* The global_mce_data_buf does not have any locks or protection around it,
1da177e4 LT	283	* if a second machine check comes in, or a system reset is done
	284	* before we have logged the error, then we will get corruption in the
	285	* error log. This is preferable over holding off on calling
	286	* ibm,nmi-interlock which would result in us checkstopping if a
	287	* second machine check did come in.
	288	*/
	289	static struct rtas_error_log fwnmi_get_errinfo(struct pt_regs regs)
	290	{
1da177e4	291	unsigned long *savep;
d368514c	292	struct rtas_error_log h, errhdr = NULL;
1da177e4	293
ee1dd1e3 MS	294	/* Mask top two bits */
	295	regs->gpr[3] &= ~(0x3UL << 62);
	296
d368514c	297	if (!VALID_FWNMI_BUFFER(regs->gpr[3])) {
f0e939ae	298	printk(KERN_ERR "FWNMI: corrupt r3 0x%016lx\n", regs->gpr[3]);
d368514c AB	299	return NULL;
	300	}
	301
	302	savep = __va(regs->gpr[3]);
	303	regs->gpr[3] = savep[0]; /* restore original r3 */
	304
	305	/* If it isn't an extended log we can use the per cpu 64bit buffer */
	306	h = (struct rtas_error_log *)&savep[1];
a08a53ea	307	if (!rtas_error_extended(h)) {
69111bac CL	308	memcpy(this_cpu_ptr(&mce_data_buf), h, sizeof(__u64));
69111bac CL	309	errhdr = (struct rtas_error_log *)this_cpu_ptr(&mce_data_buf);
1da177e4	310	} else {
a08a53ea	311	int len, error_log_length;
d368514c	312
a08a53ea GK	313	error_log_length = 8 + rtas_error_extended_log_length(h);
a08a53ea GK	314	len = max_t(int, error_log_length, RTAS_ERROR_LOG_MAX);
d368514c AB	315	memset(global_mce_data_buf, 0, RTAS_ERROR_LOG_MAX);
	316	memcpy(global_mce_data_buf, h, len);
	317	errhdr = (struct rtas_error_log *)global_mce_data_buf;
1da177e4	318	}
d368514c	319
1da177e4 LT	320	return errhdr;
	321	}
	322
	323	/* Call this when done with the data returned by FWNMI_get_errinfo.
	324	* It will release the saved data area for other CPUs in the
	325	* partition to receive FWNMI errors.
	326	*/
	327	static void fwnmi_release_errinfo(void)
	328	{
	329	int ret = rtas_call(rtas_token("ibm,nmi-interlock"), 0, 1, NULL);
	330	if (ret != 0)
d368514c	331	printk(KERN_ERR "FWNMI: nmi-interlock failed: %d\n", ret);
1da177e4 LT	332	}
1da177e4 LT	333
c902be71	334	int pSeries_system_reset_exception(struct pt_regs *regs)
1da177e4 LT	335	{
	336	if (fwnmi_active) {
	337	struct rtas_error_log *errhdr = fwnmi_get_errinfo(regs);
	338	if (errhdr) {
	339	/* XXX Should look at FWNMI information */
	340	}
	341	fwnmi_release_errinfo();
	342	}
c902be71	343	return 0; /* need to perform reset */
1da177e4 LT	344	}
	345
	346	/*
	347	* See if we can recover from a machine check exception.
	348	* This is only called on power4 (or above) and only via
	349	* the Firmware Non-Maskable Interrupts (fwnmi) handler
	350	* which provides the error analysis for us.
	351	*
	352	* Return 1 if corrected (or delivered a signal).
	353	* Return 0 if there is nothing we can do.
	354	*/
d47d1d8a	355	static int recover_mce(struct pt_regs regs, struct rtas_error_log err)
1da177e4	356	{
d47d1d8a	357	int recovered = 0;
a08a53ea	358	int disposition = rtas_error_disposition(err);
1da177e4	359
d47d1d8a AB	360	if (!(regs->msr & MSR_RI)) {
	361	/* If MSR_RI isn't set, we cannot recover */
	362	recovered = 0;
	363
a08a53ea	364	} else if (disposition == RTAS_DISP_FULLY_RECOVERED) {
1da177e4	365	/* Platform corrected itself */
d47d1d8a AB	366	recovered = 1;
d47d1d8a AB	367
a08a53ea	368	} else if (disposition == RTAS_DISP_LIMITED_RECOVERY) {
d47d1d8a AB	369	/* Platform corrected itself but could be degraded */
	370	printk(KERN_ERR "MCE: limited recovery, system may "
	371	"be degraded\n");
	372	recovered = 1;
	373
	374	} else if (user_mode(regs) && !is_global_init(current) &&
a08a53ea	375	rtas_error_severity(err) == RTAS_SEVERITY_ERROR_SYNC) {
d47d1d8a AB	376
	377	/*
	378	* If we received a synchronous error when in userspace
	379	* kill the task. Firmware may report details of the fail
	380	* asynchronously, so we can't rely on the target and type
	381	* fields being valid here.
	382	*/
	383	printk(KERN_ERR "MCE: uncorrectable error, killing task "
	384	"%s:%d\n", current->comm, current->pid);
	385
	386	_exception(SIGBUS, regs, BUS_MCEERR_AR, regs->nip);
	387	recovered = 1;
1da177e4 LT	388	}
1da177e4 LT	389
3f9793e6	390	log_error((char *)err, ERR_TYPE_RTAS_LOG, 0);
1da177e4	391
d47d1d8a	392	return recovered;
1da177e4 LT	393	}
	394
	395	/*
	396	* Handle a machine check.
	397	*
	398	* Note that on Power 4 and beyond Firmware Non-Maskable Interrupts (fwnmi)
	399	* should be present. If so the handler which called us tells us if the
	400	* error was recovered (never true if RI=0).
	401	*
	402	* On hardware prior to Power 4 these exceptions were asynchronous which
	403	* means we can't tell exactly where it occurred and so we can't recover.
	404	*/
	405	int pSeries_machine_check_exception(struct pt_regs *regs)
	406	{
	407	struct rtas_error_log *errp;
	408
	409	if (fwnmi_active) {
	410	errp = fwnmi_get_errinfo(regs);
	411	fwnmi_release_errinfo();
	412	if (errp && recover_mce(regs, errp))
	413	return 1;
	414	}
	415
	416	return 0;
	417	}