* Copyright (C) Hidetoshi Seto (seto.hidetoshi@jp.fujitsu.com)
* Copyright (C) 2005 Silicon Graphics, Inc
* Copyright (C) 2005 Keith Owens <kaos@sgi.com>
+ * Copyright (C) 2006 Russ Anderson <rja@sgi.com>
*/
#include <linux/config.h>
#include <linux/types.h>
ISOLATE_NONE
} isolate_status_t;
+typedef enum {
+ MCA_NOT_RECOVERED = 0,
+ MCA_RECOVERED = 1
+} recovery_status_t;
+
/*
* This pool keeps pointers to the section part of SAL error record
*/
int max_idx; /* Maximum index of section pointer list pool */
} slidx_pool;
+static int
+fatal_mca(const char *fmt, ...)
+{
+ va_list args;
+
+ va_start(args, fmt);
+ vprintk(fmt, args);
+ va_end(args);
+
+ return MCA_NOT_RECOVERED;
+}
+
/**
* mca_page_isolate - isolate a poisoned page in order not to use it later
* @paddr: poisoned memory location
*/
void
-mca_handler_bh(unsigned long paddr)
+mca_handler_bh(unsigned long paddr, void *iip, unsigned long ipsr)
{
- printk(KERN_DEBUG "OS_MCA: process [pid: %d](%s) encounters MCA.\n",
- current->pid, current->comm);
+ printk(KERN_ERR "OS_MCA: process [cpu %d, pid: %d, uid: %d, "
+ "iip: %p, psr: 0x%lx,paddr: 0x%lx](%s) encounters MCA.\n",
+ raw_smp_processor_id(), current->pid, current->uid,
+ iip, ipsr, paddr, current->comm);
spin_lock(&mca_bh_lock);
switch (mca_page_isolate(paddr)) {
printk(KERN_DEBUG "Page isolation: ( %lx ) success.\n", paddr);
break;
case ISOLATE_NG:
- printk(KERN_DEBUG "Page isolation: ( %lx ) failure.\n", paddr);
+ printk(KERN_CRIT "Page isolation: ( %lx ) failure.\n", paddr);
break;
default:
break;
/* Is target address valid? */
if (!pbci->tv)
- return 0;
+ return fatal_mca(KERN_ALERT "MCA: target address not valid\n");
/*
* cpu read or memory-mapped io read
* the process not have any locks of kernel.
*/
+ /* Is minstate valid? */
+ if (!peidx_bottom(peidx) || !(peidx_bottom(peidx)->valid.minstate))
+ return fatal_mca(KERN_ALERT "MCA: minstate not valid\n");
psr1 =(struct ia64_psr *)&(peidx_minstate_area(peidx)->pmsa_ipsr);
+ psr2 =(struct ia64_psr *)&(peidx_minstate_area(peidx)->pmsa_xpsr);
/*
* Check the privilege level of interrupted context.
* If it is user-mode, then terminate affected process.
*/
- if (psr1->cpl != 0) {
+
+ pmsa = sos->pal_min_state;
+ if (psr1->cpl != 0 ||
+ ((psr2->cpl != 0) && mca_recover_range(pmsa->pmsa_iip))) {
smei = peidx_bus_check(peidx, 0);
if (smei->valid.target_identifier) {
/*
* setup for resume to bottom half of MCA,
* "mca_handler_bhhook"
*/
- pmsa = sos->pal_min_state;
- /* pass to bhhook as 1st argument (gr8) */
+ /* pass to bhhook as argument (gr8, ...) */
pmsa->pmsa_gr[8-1] = smei->target_identifier;
+ pmsa->pmsa_gr[9-1] = pmsa->pmsa_iip;
+ pmsa->pmsa_gr[10-1] = pmsa->pmsa_ipsr;
/* set interrupted return address (but no use) */
pmsa->pmsa_br0 = pmsa->pmsa_iip;
/* change resume address to bottom half */
psr2 = (struct ia64_psr *)&pmsa->pmsa_ipsr;
psr2->cpl = 0;
psr2->ri = 0;
+ psr2->bn = 1;
psr2->i = 0;
- return 1;
+ return MCA_RECOVERED;
}
}
- return 0;
+ return fatal_mca(KERN_ALERT "MCA: kernel context not recovered,"
+ " iip 0x%lx\n", pmsa->pmsa_iip);
}
/**
* The machine check is corrected.
*/
if (psp->cm == 1)
- return 1;
+ return MCA_RECOVERED;
/*
* The error was not contained. Software must be reset.
*/
if (psp->us || psp->ci == 0)
- return 0;
+ return fatal_mca(KERN_ALERT "MCA: error not contained\n");
/*
- * If there is no bus error, record is weird but we need not to recover.
+ * The cache check and bus check bits have four possible states
+ * cc bc
+ * 0 0 Weird record, not recovered
+ * 1 0 Cache error, not recovered
+ * 0 1 I/O error, attempt recovery
+ * 1 1 Memory error, attempt recovery
*/
if (psp->bc == 0 || pbci == NULL)
- return 1;
+ return fatal_mca(KERN_ALERT "MCA: No bus check\n");
/*
* Sorry, we cannot handle so many.
*/
if (peidx_bus_check_num(peidx) > 1)
- return 0;
+ return fatal_mca(KERN_ALERT "MCA: Too many bus checks\n");
/*
* Well, here is only one bus error.
*/
- if (pbci->ib || pbci->cc)
- return 0;
+ if (pbci->ib)
+ return fatal_mca(KERN_ALERT "MCA: Internal Bus error\n");
+ if (pbci->cc)
+ return fatal_mca(KERN_ALERT "MCA: Cache-cache error\n");
if (pbci->eb && pbci->bsi > 0)
- return 0;
+ return fatal_mca(KERN_ALERT "MCA: External bus check fatal status\n");
/*
* This is a local MCA and estimated as recoverble external bus error.
/*
* On account of strange SAL error record, we cannot recover.
*/
- return 0;
+ return fatal_mca(KERN_ALERT "MCA: Strange SAL record\n");
}
/**
/* Now, OS can recover when there is one processor error section */
if (n_proc_err > 1)
- return 0;
- else if (n_proc_err == 0) {
+ return fatal_mca(KERN_ALERT "MCA: Too Many Errors\n");
+ else if (n_proc_err == 0)
/* Weird SAL record ... We need not to recover */
-
- return 1;
- }
+ return fatal_mca(KERN_ALERT "MCA: Weird SAL record\n");
/* Make index of processor error section */
mca_make_peidx((sal_log_processor_info_t*)
/* Check whether MCA is global or not */
if (is_mca_global(&peidx, &pbci, sos))
- return 0;
+ return fatal_mca(KERN_ALERT "MCA: global MCA\n");
/* Try to recover a processor error */
return recover_from_processor_error(platform_err, &slidx, &peidx,