projects / deliverable / linux.git / blame
| Commit | Line | Data |
|---|---|---|
| 172ca926 LV |
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/list.h> | |
| 22 | #include <linux/pci.h> | |
| 23 | #include <asm/eeh_event.h> | |
| 24 | ||
| 25 | /** Overview: | |
| 26 | * EEH error states may be detected within exception handlers; | |
| 27 | * however, the recovery processing needs to occur asynchronously | |
| 28 | * in a normal kernel context and not an interrupt context. | |
| 29 | * This pair of routines creates an event and queues it onto a | |
| 30 | * work-queue, where a worker thread can drive recovery. | |
| 31 | */ | |
| 32 | ||
| 33 | /* EEH event workqueue setup. */ | |
| 34 | static spinlock_t eeh_eventlist_lock = SPIN_LOCK_UNLOCKED; | |
| 35 | LIST_HEAD(eeh_eventlist); | |
| 36 | static void eeh_thread_launcher(void *); | |
| 37 | DECLARE_WORK(eeh_event_wq, eeh_thread_launcher, NULL); | |
| 38 | ||
| 39 | /** | |
| 40 | * eeh_panic - call panic() for an eeh event that cannot be handled. | |
| 41 | * The philosophy of this routine is that it is better to panic and | |
| 42 | * halt the OS than it is to risk possible data corruption by | |
| 43 | * oblivious device drivers that don't know better. | |
| 44 | * | |
| 45 | * @dev pci device that had an eeh event | |
| 46 | * @reset_state current reset state of the device slot | |
| 47 | */ | |
| 48 | static void eeh_panic(struct pci_dev *dev, int reset_state) | |
| 49 | { | |
| 50 | /* | |
| 51 | * Since the panic_on_oops sysctl is used to halt the system | |
| 52 | * in light of potential corruption, we can use it here. | |
| 53 | */ | |
| 54 | if (panic_on_oops) { | |
| 55 | panic("EEH: MMIO failure (%d) on device:%s\n", reset_state, | |
| 56 | pci_name(dev)); | |
| 57 | } | |
| 58 | else { | |
| 59 | printk(KERN_INFO "EEH: Ignored MMIO failure (%d) on device:%s\n", | |
| 60 | reset_state, pci_name(dev)); | |
| 61 | } | |
| 62 | } | |
| 63 | ||
| 64 | /** | |
| 65 | * eeh_event_handler - dispatch EEH events. The detection of a frozen | |
| 66 | * slot can occur inside an interrupt, where it can be hard to do | |
| 67 | * anything about it. The goal of this routine is to pull these | |
| 68 | * detection events out of the context of the interrupt handler, and | |
| 69 | * re-dispatch them for processing at a later time in a normal context. | |
| 70 | * | |
| 71 | * @dummy - unused | |
| 72 | */ | |
| 73 | static int eeh_event_handler(void * dummy) | |
| 74 | { | |
| 75 | unsigned long flags; | |
| 76 | struct eeh_event *event; | |
| 77 | ||
| 78 | daemonize ("eehd"); | |
| 79 | ||
| 80 | while (1) { | |
| 81 | set_current_state(TASK_INTERRUPTIBLE); | |
| 82 | ||
| 83 | spin_lock_irqsave(&eeh_eventlist_lock, flags); | |
| 84 | event = NULL; | |
| 85 | if (!list_empty(&eeh_eventlist)) { | |
| 86 | event = list_entry(eeh_eventlist.next, struct eeh_event, list); | |
| 87 | list_del(&event->list); | |
| 88 | } | |
| 89 | spin_unlock_irqrestore(&eeh_eventlist_lock, flags); | |
| 90 | if (event == NULL) | |
| 91 | break; | |
| 92 | ||
| 93 | printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n", | |
| 94 | pci_name(event->dev)); | |
| 95 | ||
| 96 | eeh_panic (event->dev, event->state); | |
| 97 | ||
| 98 | kfree(event); | |
| 99 | } | |
| 100 | ||
| 101 | return 0; | |
| 102 | } | |
| 103 | ||
| 104 | /** | |
| 105 | * eeh_thread_launcher | |
| 106 | * | |
| 107 | * @dummy - unused | |
| 108 | */ | |
| 109 | static void eeh_thread_launcher(void *dummy) | |
| 110 | { | |
| 111 | if (kernel_thread(eeh_event_handler, NULL, CLONE_KERNEL) < 0) | |
| 112 | printk(KERN_ERR "Failed to start EEH daemon\n"); | |
| 113 | } | |
| 114 | ||
| 115 | /** | |
| 116 | * eeh_send_failure_event - generate a PCI error event | |
| 117 | * @dev pci device | |
| 118 | * | |
| 119 | * This routine can be called within an interrupt context; | |
| 120 | * the actual event will be delivered in a normal context | |
| 121 | * (from a workqueue). | |
| 122 | */ | |
| 123 | int eeh_send_failure_event (struct device_node *dn, | |
| 124 | struct pci_dev *dev, | |
| 125 | int state, | |
| 126 | int time_unavail) | |
| 127 | { | |
| 128 | unsigned long flags; | |
| 129 | struct eeh_event *event; | |
| 130 | ||
| 131 | event = kmalloc(sizeof(*event), GFP_ATOMIC); | |
| 132 | if (event == NULL) { | |
| 133 | printk (KERN_ERR "EEH: out of memory, event not handled\n"); | |
| 134 | return 1; | |
| 135 | } | |
| 136 | ||
| 137 | if (dev) | |
| 138 | pci_dev_get(dev); | |
| 139 | ||
| 140 | event->dn = dn; | |
| 141 | event->dev = dev; | |
| 142 | event->state = state; | |
| 143 | event->time_unavail = time_unavail; | |
| 144 | ||
| 145 | /* We may or may not be called in an interrupt context */ | |
| 146 | spin_lock_irqsave(&eeh_eventlist_lock, flags); | |
| 147 | list_add(&event->list, &eeh_eventlist); | |
| 148 | spin_unlock_irqrestore(&eeh_eventlist_lock, flags); | |
| 149 | ||
| 150 | schedule_work(&eeh_event_wq); | |
| 151 | ||
| 152 | return 0; | |
| 153 | } | |
| 154 | ||
| 155 | /********************** END OF FILE ******************************/ |