Merge branch 'intelfb-patches' of master.kernel.org:/pub/scm/linux/kernel/git/airlied...

[linux.git] / arch / ia64 / kernel / mca_drv.c

diff --git a/arch/ia64/kernel/mca_drv.c b/arch/ia64/kernel/mca_drv.c
index 37c88eb55873792250d74ac22455ee5cdcab93af..a45009d2bc90149e5199a97c9eff86a8d552f5b0 100644 (file)
--- a/arch/ia64/kernel/mca_drv.c
+++ b/arch/ia64/kernel/mca_drv.c
@@ -8,7 +8,6 @@
  * Copyright (C) 2005 Keith Owens <[email protected]>
  * Copyright (C) 2006 Russ Anderson <[email protected]>
  */
-#include <linux/config.h>
 #include <linux/types.h>
 #include <linux/init.h>
 #include <linux/sched.h>
@@ -62,6 +61,11 @@ typedef enum {
        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
  */
@@ -71,6 +75,34 @@ static struct {
        int          max_idx; /* Maximum index of section pointer list pool */
 } slidx_pool;
 
+static int
+fatal_mca(const char *fmt, ...)
+{
+       va_list args;
+       char buf[256];
+
+       va_start(args, fmt);
+       vsnprintf(buf, sizeof(buf), fmt, args);
+       va_end(args);
+       ia64_mca_printk(KERN_ALERT "MCA: %s\n", buf);
+
+       return MCA_NOT_RECOVERED;
+}
+
+static int
+mca_recovered(const char *fmt, ...)
+{
+       va_list args;
+       char buf[256];
+
+       va_start(args, fmt);
+       vsnprintf(buf, sizeof(buf), fmt, args);
+       va_end(args);
+       ia64_mca_printk(KERN_INFO "MCA: %s\n", buf);
+
+       return MCA_RECOVERED;
+}
+
 /**
  * mca_page_isolate - isolate a poisoned page in order not to use it later
  * @paddr:     poisoned memory location
@@ -124,6 +156,7 @@ mca_page_isolate(unsigned long paddr)
 void
 mca_handler_bh(unsigned long paddr, void *iip, unsigned long ipsr)
 {
+       ia64_mlogbuf_dump();
        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,
@@ -424,7 +457,7 @@ recover_from_read_error(slidx_table_t *slidx,
 
        /* Is target address valid? */
        if (!pbci->tv)
-               return 0;
+               return fatal_mca("target address not valid");
 
        /*
         * cpu read or memory-mapped io read
@@ -442,7 +475,7 @@ recover_from_read_error(slidx_table_t *slidx,
 
        /* Is minstate valid? */
        if (!peidx_bottom(peidx) || !(peidx_bottom(peidx)->valid.minstate))
-               return 0;
+               return fatal_mca("minstate not valid");
        psr1 =(struct ia64_psr *)&(peidx_minstate_area(peidx)->pmsa_ipsr);
        psr2 =(struct ia64_psr *)&(peidx_minstate_area(peidx)->pmsa_xpsr);
 
@@ -476,12 +509,14 @@ recover_from_read_error(slidx_table_t *slidx,
                        psr2->bn  = 1;
                        psr2->i  = 0;
 
-                       return 1;
+                       return mca_recovered("user memory corruption. "
+                               "kill affected process - recovered.");
                }
 
        }
 
-       return 0;
+       return fatal_mca("kernel context not recovered, iip 0x%lx\n",
+                        pmsa->pmsa_iip);
 }
 
 /**
@@ -567,13 +602,13 @@ recover_from_processor_error(int platform, slidx_table_t *slidx,
         * The machine check is corrected.
         */
        if (psp->cm == 1)
-               return 1;
+               return mca_recovered("machine check is already corrected.");
 
        /*
         * The error was not contained.  Software must be reset.
         */
        if (psp->us || psp->ci == 0)
-               return 0;
+               return fatal_mca("error not contained");
 
        /*
         * The cache check and bus check bits have four possible states
@@ -584,20 +619,22 @@ recover_from_processor_error(int platform, slidx_table_t *slidx,
         *    1  1      Memory error, attempt recovery
         */
        if (psp->bc == 0 || pbci == NULL)
-               return 0;
+               return fatal_mca("No bus check");
 
        /*
         * Sorry, we cannot handle so many.
         */
        if (peidx_bus_check_num(peidx) > 1)
-               return 0;
+               return fatal_mca("Too many bus checks");
        /*
         * Well, here is only one bus error.
         */
-       if (pbci->ib || pbci->cc)
-               return 0;
+       if (pbci->ib)
+               return fatal_mca("Internal Bus error");
+       if (pbci->cc)
+               return fatal_mca("Cache-cache error");
        if (pbci->eb && pbci->bsi > 0)
-               return 0;
+               return fatal_mca("External bus check fatal status");
 
        /*
         * This is a local MCA and estimated as recoverble external bus error.
@@ -609,7 +646,7 @@ recover_from_processor_error(int platform, slidx_table_t *slidx,
        /*
         * On account of strange SAL error record, we cannot recover.
         */
-       return 0;
+       return fatal_mca("Strange SAL record");
 }
 
 /**
@@ -638,12 +675,10 @@ mca_try_to_recover(void *rec, struct ia64_sal_os_state *sos)
 
         /* Now, OS can recover when there is one processor error section */
        if (n_proc_err > 1)
-               return 0;
-       else if (n_proc_err == 0) {
-               /* Weird SAL record ... We need not to recover */
-
-               return 1;
-       }
+               return fatal_mca("Too Many Errors");
+       else if (n_proc_err == 0)
+               /* Weird SAL record ... We can't do anything */
+               return fatal_mca("Weird SAL record");
 
        /* Make index of processor error section */
        mca_make_peidx((sal_log_processor_info_t*)
@@ -654,7 +689,7 @@ mca_try_to_recover(void *rec, struct ia64_sal_os_state *sos)
 
        /* Check whether MCA is global or not */
        if (is_mca_global(&peidx, &pbci, sos))
-               return 0;
+               return fatal_mca("global MCA");
        
        /* Try to recover a processor error */
        return recover_from_processor_error(platform_err, &slidx, &peidx,