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.
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.
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
18 * Copyright (c) 2005 Linas Vepstas <linas@linas.org>
21 #include <linux/list.h>
22 #include <linux/pci.h>
23 #include <asm/eeh_event.h>
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.
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);
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.
45 * @dev pci device that had an eeh event
46 * @reset_state current reset state of the device slot
48 static void eeh_panic(struct pci_dev *dev, int reset_state)
51 * Since the panic_on_oops sysctl is used to halt the system
52 * in light of potential corruption, we can use it here.
55 panic("EEH: MMIO failure (%d) on device:%s\n", reset_state,
59 printk(KERN_INFO "EEH: Ignored MMIO failure (%d) on device:%s\n",
60 reset_state, pci_name(dev));
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.
73 static int eeh_event_handler(void * dummy)
76 struct eeh_event *event;
81 set_current_state(TASK_INTERRUPTIBLE);
83 spin_lock_irqsave(&eeh_eventlist_lock, flags);
85 if (!list_empty(&eeh_eventlist)) {
86 event = list_entry(eeh_eventlist.next, struct eeh_event, list);
87 list_del(&event->list);
89 spin_unlock_irqrestore(&eeh_eventlist_lock, flags);
93 printk(KERN_INFO "EEH: Detected PCI bus error on device %s\n",
94 pci_name(event->dev));
96 eeh_panic (event->dev, event->state);
105 * eeh_thread_launcher
109 static void eeh_thread_launcher(void *dummy)
111 if (kernel_thread(eeh_event_handler, NULL, CLONE_KERNEL) < 0)
112 printk(KERN_ERR "Failed to start EEH daemon\n");
116 * eeh_send_failure_event - generate a PCI error event
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).
123 int eeh_send_failure_event (struct device_node *dn,
129 struct eeh_event *event;
131 event = kmalloc(sizeof(*event), GFP_ATOMIC);
133 printk (KERN_ERR "EEH: out of memory, event not handled\n");
142 event->state = state;
143 event->time_unavail = time_unavail;
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);
150 schedule_work(&eeh_event_wq);
155 /********************** END OF FILE ******************************/