1 /* 2 * Copyright (c) 2005 Cisco Systems. All rights reserved. 3 * 4 * This software is available to you under a choice of one of two 5 * licenses. You may choose to be licensed under the terms of the GNU 6 * General Public License (GPL) Version 2, available from the file 7 * COPYING in the main directory of this source tree, or the 8 * OpenIB.org BSD license below: 9 * 10 * Redistribution and use in source and binary forms, with or 11 * without modification, are permitted provided that the following 12 * conditions are met: 13 * 14 * - Redistributions of source code must retain the above 15 * copyright notice, this list of conditions and the following 16 * disclaimer. 17 * 18 * - Redistributions in binary form must reproduce the above 19 * copyright notice, this list of conditions and the following 20 * disclaimer in the documentation and/or other materials 21 * provided with the distribution. 22 * 23 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 24 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 25 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND 26 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS 27 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN 28 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN 29 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 30 * SOFTWARE. 31 */ 32 33 #include <linux/jiffies.h> 34 #include <linux/timer.h> 35 #include <linux/workqueue.h> 36 37 #include "mthca_dev.h" 38 39 enum { 40 MTHCA_CATAS_POLL_INTERVAL = 5 * HZ, 41 42 MTHCA_CATAS_TYPE_INTERNAL = 0, 43 MTHCA_CATAS_TYPE_UPLINK = 3, 44 MTHCA_CATAS_TYPE_DDR = 4, 45 MTHCA_CATAS_TYPE_PARITY = 5, 46 }; 47 48 static DEFINE_SPINLOCK(catas_lock); 49 50 static LIST_HEAD(catas_list); 51 static struct workqueue_struct *catas_wq; 52 static struct work_struct catas_work; 53 54 static int catas_reset_disable; 55 module_param_named(catas_reset_disable, catas_reset_disable, int, 0644); 56 MODULE_PARM_DESC(catas_reset_disable, "disable reset on catastrophic event if nonzero"); 57 58 static void catas_reset(struct work_struct *work) 59 { 60 struct mthca_dev *dev, *tmpdev; 61 LIST_HEAD(tlist); 62 int ret; 63 64 mutex_lock(&mthca_device_mutex); 65 66 spin_lock_irq(&catas_lock); 67 list_splice_init(&catas_list, &tlist); 68 spin_unlock_irq(&catas_lock); 69 70 list_for_each_entry_safe(dev, tmpdev, &tlist, catas_err.list) { 71 ret = __mthca_restart_one(dev->pdev); 72 if (ret) 73 mthca_err(dev, "Reset failed (%d)\n", ret); 74 else 75 mthca_dbg(dev, "Reset succeeded\n"); 76 } 77 78 mutex_unlock(&mthca_device_mutex); 79 } 80 81 static void handle_catas(struct mthca_dev *dev) 82 { 83 struct ib_event event; 84 unsigned long flags; 85 const char *type; 86 int i; 87 88 event.device = &dev->ib_dev; 89 event.event = IB_EVENT_DEVICE_FATAL; 90 event.element.port_num = 0; 91 92 ib_dispatch_event(&event); 93 94 switch (swab32(readl(dev->catas_err.map)) >> 24) { 95 case MTHCA_CATAS_TYPE_INTERNAL: 96 type = "internal error"; 97 break; 98 case MTHCA_CATAS_TYPE_UPLINK: 99 type = "uplink bus error"; 100 break; 101 case MTHCA_CATAS_TYPE_DDR: 102 type = "DDR data error"; 103 break; 104 case MTHCA_CATAS_TYPE_PARITY: 105 type = "internal parity error"; 106 break; 107 default: 108 type = "unknown error"; 109 break; 110 } 111 112 mthca_err(dev, "Catastrophic error detected: %s\n", type); 113 for (i = 0; i < dev->catas_err.size; ++i) 114 mthca_err(dev, " buf[%02x]: %08x\n", 115 i, swab32(readl(dev->catas_err.map + i))); 116 117 if (catas_reset_disable) 118 return; 119 120 spin_lock_irqsave(&catas_lock, flags); 121 list_add(&dev->catas_err.list, &catas_list); 122 queue_work(catas_wq, &catas_work); 123 spin_unlock_irqrestore(&catas_lock, flags); 124 } 125 126 static void poll_catas(unsigned long dev_ptr) 127 { 128 struct mthca_dev *dev = (struct mthca_dev *) dev_ptr; 129 int i; 130 131 for (i = 0; i < dev->catas_err.size; ++i) 132 if (readl(dev->catas_err.map + i)) { 133 handle_catas(dev); 134 return; 135 } 136 137 mod_timer(&dev->catas_err.timer, 138 round_jiffies(jiffies + MTHCA_CATAS_POLL_INTERVAL)); 139 } 140 141 void mthca_start_catas_poll(struct mthca_dev *dev) 142 { 143 unsigned long addr; 144 145 init_timer(&dev->catas_err.timer); 146 dev->catas_err.map = NULL; 147 148 addr = pci_resource_start(dev->pdev, 0) + 149 ((pci_resource_len(dev->pdev, 0) - 1) & 150 dev->catas_err.addr); 151 152 dev->catas_err.map = ioremap(addr, dev->catas_err.size * 4); 153 if (!dev->catas_err.map) { 154 mthca_warn(dev, "couldn't map catastrophic error region " 155 "at 0x%lx/0x%x\n", addr, dev->catas_err.size * 4); 156 return; 157 } 158 159 dev->catas_err.timer.data = (unsigned long) dev; 160 dev->catas_err.timer.function = poll_catas; 161 dev->catas_err.timer.expires = jiffies + MTHCA_CATAS_POLL_INTERVAL; 162 INIT_LIST_HEAD(&dev->catas_err.list); 163 add_timer(&dev->catas_err.timer); 164 } 165 166 void mthca_stop_catas_poll(struct mthca_dev *dev) 167 { 168 del_timer_sync(&dev->catas_err.timer); 169 170 if (dev->catas_err.map) 171 iounmap(dev->catas_err.map); 172 173 spin_lock_irq(&catas_lock); 174 list_del(&dev->catas_err.list); 175 spin_unlock_irq(&catas_lock); 176 } 177 178 int __init mthca_catas_init(void) 179 { 180 INIT_WORK(&catas_work, catas_reset); 181 182 catas_wq = create_singlethread_workqueue("mthca_catas"); 183 if (!catas_wq) 184 return -ENOMEM; 185 186 return 0; 187 } 188 189 void mthca_catas_cleanup(void) 190 { 191 destroy_workqueue(catas_wq); 192 } 193