eeh_event.c 4.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156
  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. #include <linux/delay.h>
  21. #include <linux/list.h>
  22. #include <linux/mutex.h>
  23. #include <linux/pci.h>
  24. #include <linux/workqueue.h>
  25. #include <asm/eeh_event.h>
  26. #include <asm/ppc-pci.h>
  27. /** Overview:
  28. * EEH error states may be detected within exception handlers;
  29. * however, the recovery processing needs to occur asynchronously
  30. * in a normal kernel context and not an interrupt context.
  31. * This pair of routines creates an event and queues it onto a
  32. * work-queue, where a worker thread can drive recovery.
  33. */
  34. /* EEH event workqueue setup. */
  35. static DEFINE_SPINLOCK(eeh_eventlist_lock);
  36. LIST_HEAD(eeh_eventlist);
  37. static void eeh_thread_launcher(struct work_struct *);
  38. DECLARE_WORK(eeh_event_wq, eeh_thread_launcher);
  39. /* Serialize reset sequences for a given pci device */
  40. DEFINE_MUTEX(eeh_event_mutex);
  41. /**
  42. * eeh_event_handler - dispatch EEH events.
  43. * @dummy - unused
  44. *
  45. * The detection of a frozen slot can occur inside an interrupt,
  46. * where it can be hard to do anything about it. The goal of this
  47. * routine is to pull these detection events out of the context
  48. * of the interrupt handler, and re-dispatch them for processing
  49. * at a later time in a normal context.
  50. */
  51. static int eeh_event_handler(void * dummy)
  52. {
  53. unsigned long flags;
  54. struct eeh_event *event;
  55. struct pci_dn *pdn;
  56. daemonize ("eehd");
  57. set_current_state(TASK_INTERRUPTIBLE);
  58. spin_lock_irqsave(&eeh_eventlist_lock, flags);
  59. event = NULL;
  60. /* Unqueue the event, get ready to process. */
  61. if (!list_empty(&eeh_eventlist)) {
  62. event = list_entry(eeh_eventlist.next, struct eeh_event, list);
  63. list_del(&event->list);
  64. }
  65. spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
  66. if (event == NULL)
  67. return 0;
  68. /* Serialize processing of EEH events */
  69. mutex_lock(&eeh_event_mutex);
  70. eeh_mark_slot(event->dn, EEH_MODE_RECOVERING);
  71. printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n",
  72. pci_name(event->dev));
  73. pdn = handle_eeh_events(event);
  74. eeh_clear_slot(event->dn, EEH_MODE_RECOVERING);
  75. pci_dev_put(event->dev);
  76. kfree(event);
  77. mutex_unlock(&eeh_event_mutex);
  78. /* If there are no new errors after an hour, clear the counter. */
  79. if (pdn && pdn->eeh_freeze_count>0) {
  80. msleep_interruptible (3600*1000);
  81. if (pdn->eeh_freeze_count>0)
  82. pdn->eeh_freeze_count--;
  83. }
  84. return 0;
  85. }
  86. /**
  87. * eeh_thread_launcher
  88. * @dummy - unused
  89. */
  90. static void eeh_thread_launcher(struct work_struct *dummy)
  91. {
  92. if (kernel_thread(eeh_event_handler, NULL, CLONE_KERNEL) < 0)
  93. printk(KERN_ERR "Failed to start EEH daemon\n");
  94. }
  95. /**
  96. * eeh_send_failure_event - generate a PCI error event
  97. * @dev pci device
  98. *
  99. * This routine can be called within an interrupt context;
  100. * the actual event will be delivered in a normal context
  101. * (from a workqueue).
  102. */
  103. int eeh_send_failure_event (struct device_node *dn,
  104. struct pci_dev *dev)
  105. {
  106. unsigned long flags;
  107. struct eeh_event *event;
  108. const char *location;
  109. if (!mem_init_done) {
  110. printk(KERN_ERR "EEH: event during early boot not handled\n");
  111. location = of_get_property(dn, "ibm,loc-code", NULL);
  112. printk(KERN_ERR "EEH: device node = %s\n", dn->full_name);
  113. printk(KERN_ERR "EEH: PCI location = %s\n", location);
  114. return 1;
  115. }
  116. event = kmalloc(sizeof(*event), GFP_ATOMIC);
  117. if (event == NULL) {
  118. printk (KERN_ERR "EEH: out of memory, event not handled\n");
  119. return 1;
  120. }
  121. if (dev)
  122. pci_dev_get(dev);
  123. event->dn = dn;
  124. event->dev = dev;
  125. /* We may or may not be called in an interrupt context */
  126. spin_lock_irqsave(&eeh_eventlist_lock, flags);
  127. list_add(&event->list, &eeh_eventlist);
  128. spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
  129. schedule_work(&eeh_event_wq);
  130. return 0;
  131. }
  132. /********************** END OF FILE ******************************/