release/src-rt-6.x.4708/linux/linux-2.6.36/arch/powerpc/platforms/pseries/eeh_event.c

   1 /*
   2  * eeh_event.c
   3  *
   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.
   8  *
   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.
  13  *
  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  * Copyright (c) 2005 Linas Vepstas <linas@linas.org>
  19  */
  20
  21 #include <linux/delay.h>
  22 #include <linux/list.h>
  23 #include <linux/mutex.h>
  24 #include <linux/pci.h>
  25 #include <linux/slab.h>
  26 #include <linux/workqueue.h>
  27 #include <asm/eeh_event.h>
  28 #include <asm/ppc-pci.h>
  29
  30 /** Overview:
  31  *  EEH error states may be detected within exception handlers;
  32  *  however, the recovery processing needs to occur asynchronously
  33  *  in a normal kernel context and not an interrupt context.
  34  *  This pair of routines creates an event and queues it onto a
  35  *  work-queue, where a worker thread can drive recovery.
  36  */
  37
  38 /* EEH event workqueue setup. */
  39 static DEFINE_SPINLOCK(eeh_eventlist_lock);
  40 LIST_HEAD(eeh_eventlist);
  41 static void eeh_thread_launcher(struct work_struct *);
  42 DECLARE_WORK(eeh_event_wq, eeh_thread_launcher);
  43
  44 /* Serialize reset sequences for a given pci device */
  45 DEFINE_MUTEX(eeh_event_mutex);
  46
  47 /**
  48  * eeh_event_handler - dispatch EEH events.
  49  * @dummy - unused
  50  *
  51  * The detection of a frozen slot can occur inside an interrupt,
  52  * where it can be hard to do anything about it.  The goal of this
  53  * routine is to pull these detection events out of the context
  54  * of the interrupt handler, and re-dispatch them for processing
  55  * at a later time in a normal context.
  56  */
  57 static int eeh_event_handler(void * dummy)
  58 {
  59         unsigned long flags;
  60         struct eeh_event        *event;
  61         struct pci_dn *pdn;
  62
  63         daemonize ("eehd");
  64         set_current_state(TASK_INTERRUPTIBLE);
  65
  66         spin_lock_irqsave(&eeh_eventlist_lock, flags);
  67         event = NULL;
  68
  69         /* Unqueue the event, get ready to process. */
  70         if (!list_empty(&eeh_eventlist)) {
  71                 event = list_entry(eeh_eventlist.next, struct eeh_event, list);
  72                 list_del(&event->list);
  73         }
  74         spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
  75
  76         if (event == NULL)
  77                 return 0;
  78
  79         /* Serialize processing of EEH events */
  80         mutex_lock(&eeh_event_mutex);
  81         eeh_mark_slot(event->dn, EEH_MODE_RECOVERING);
  82
  83         printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n",
  84                eeh_pci_name(event->dev));
  85
  86         pdn = handle_eeh_events(event);
  87
  88         eeh_clear_slot(event->dn, EEH_MODE_RECOVERING);
  89         pci_dev_put(event->dev);
  90         kfree(event);
  91         mutex_unlock(&eeh_event_mutex);
  92
  93         /* If there are no new errors after an hour, clear the counter. */
  94         if (pdn && pdn->eeh_freeze_count>0) {
  95                 msleep_interruptible (3600*1000);
  96                 if (pdn->eeh_freeze_count>0)
  97                         pdn->eeh_freeze_count--;
  98         }
  99
 100         return 0;
 101 }
 102
 103 /**
 104  * eeh_thread_launcher
 105  * @dummy - unused
 106  */
 107 static void eeh_thread_launcher(struct work_struct *dummy)
 108 {
 109         if (kernel_thread(eeh_event_handler, NULL, CLONE_KERNEL) < 0)
 110                 printk(KERN_ERR "Failed to start EEH daemon\n");
 111 }
 112
 113 /**
 114  * eeh_send_failure_event - generate a PCI error event
 115  * @dev pci device
 116  *
 117  * This routine can be called within an interrupt context;
 118  * the actual event will be delivered in a normal context
 119  * (from a workqueue).
 120  */
 121 int eeh_send_failure_event (struct device_node *dn,
 122                             struct pci_dev *dev)
 123 {
 124         unsigned long flags;
 125         struct eeh_event *event;
 126         const char *location;
 127
 128         if (!mem_init_done) {
 129                 printk(KERN_ERR "EEH: event during early boot not handled\n");
 130                 location = of_get_property(dn, "ibm,loc-code", NULL);
 131                 printk(KERN_ERR "EEH: device node = %s\n", dn->full_name);
 132                 printk(KERN_ERR "EEH: PCI location = %s\n", location);
 133                 return 1;
 134         }
 135         event = kmalloc(sizeof(*event), GFP_ATOMIC);
 136         if (event == NULL) {
 137                 printk (KERN_ERR "EEH: out of memory, event not handled\n");
 138                 return 1;
 139         }
 140
 141         if (dev)
 142                 pci_dev_get(dev);
 143
 144         event->dn = dn;
 145         event->dev = dev;
 146
 147         /* We may or may not be called in an interrupt context */
 148         spin_lock_irqsave(&eeh_eventlist_lock, flags);
 149         list_add(&event->list, &eeh_eventlist);
 150         spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
 151
 152         schedule_work(&eeh_event_wq);
 153
 154         return 0;
 155 }
 156
 157 /********************** END OF FILE ******************************/