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 * $Id$ 33 */ 34 35 #include <linux/jiffies.h> 36 #include <linux/timer.h> 37 #include <linux/workqueue.h> 38 39 #include "mthca_dev.h" 40 41 enum { 42 MTHCA_CATAS_POLL_INTERVAL = 5 * HZ, 43 44 MTHCA_CATAS_TYPE_INTERNAL = 0, 45 MTHCA_CATAS_TYPE_UPLINK = 3, 46 MTHCA_CATAS_TYPE_DDR = 4, 47 MTHCA_CATAS_TYPE_PARITY = 5, 48 }; 49 50 static DEFINE_SPINLOCK(catas_lock); 51 52 static LIST_HEAD(catas_list); 53 static struct workqueue_struct *catas_wq; 54 static struct work_struct catas_work; 55 56 static int catas_reset_disable; 57 module_param_named(catas_reset_disable, catas_reset_disable, int, 0644); 58 MODULE_PARM_DESC(catas_reset_disable, "disable reset on catastrophic event if nonzero"); 59 60 static void catas_reset(void *work_ptr) 61 { 62 struct mthca_dev *dev, *tmpdev; 63 LIST_HEAD(tlist); 64 int ret; 65 66 mutex_lock(&mthca_device_mutex); 67 68 spin_lock_irq(&catas_lock); 69 list_splice_init(&catas_list, &tlist); 70 spin_unlock_irq(&catas_lock); 71 72 list_for_each_entry_safe(dev, tmpdev, &tlist, catas_err.list) { 73 ret = __mthca_restart_one(dev->pdev); 74 if (ret) 75 mthca_err(dev, "Reset failed (%d)\n", ret); 76 else 77 mthca_dbg(dev, "Reset succeeded\n"); 78 } 79 80 mutex_unlock(&mthca_device_mutex); 81 } 82 83 static void handle_catas(struct mthca_dev *dev) 84 { 85 struct ib_event event; 86 unsigned long flags; 87 const char *type; 88 int i; 89 90 event.device = &dev->ib_dev; 91 event.event = IB_EVENT_DEVICE_FATAL; 92 event.element.port_num = 0; 93 94 ib_dispatch_event(&event); 95 96 switch (swab32(readl(dev->catas_err.map)) >> 24) { 97 case MTHCA_CATAS_TYPE_INTERNAL: 98 type = "internal error"; 99 break; 100 case MTHCA_CATAS_TYPE_UPLINK: 101 type = "uplink bus error"; 102 break; 103 case MTHCA_CATAS_TYPE_DDR: 104 type = "DDR data error"; 105 break; 106 case MTHCA_CATAS_TYPE_PARITY: 107 type = "internal parity error"; 108 break; 109 default: 110 type = "unknown error"; 111 break; 112 } 113 114 mthca_err(dev, "Catastrophic error detected: %s\n", type); 115 for (i = 0; i < dev->catas_err.size; ++i) 116 mthca_err(dev, " buf[%02x]: %08x\n", 117 i, swab32(readl(dev->catas_err.map + i))); 118 119 if (catas_reset_disable) 120 return; 121 122 spin_lock_irqsave(&catas_lock, flags); 123 list_add(&dev->catas_err.list, &catas_list); 124 queue_work(catas_wq, &catas_work); 125 spin_unlock_irqrestore(&catas_lock, flags); 126 } 127 128 static void poll_catas(unsigned long dev_ptr) 129 { 130 struct mthca_dev *dev = (struct mthca_dev *) dev_ptr; 131 unsigned long flags; 132 int i; 133 134 for (i = 0; i < dev->catas_err.size; ++i) 135 if (readl(dev->catas_err.map + i)) { 136 handle_catas(dev); 137 return; 138 } 139 140 spin_lock_irqsave(&catas_lock, flags); 141 if (!dev->catas_err.stop) 142 mod_timer(&dev->catas_err.timer, 143 jiffies + MTHCA_CATAS_POLL_INTERVAL); 144 spin_unlock_irqrestore(&catas_lock, flags); 145 146 return; 147 } 148 149 void mthca_start_catas_poll(struct mthca_dev *dev) 150 { 151 unsigned long addr; 152 153 init_timer(&dev->catas_err.timer); 154 dev->catas_err.stop = 0; 155 dev->catas_err.map = NULL; 156 157 addr = pci_resource_start(dev->pdev, 0) + 158 ((pci_resource_len(dev->pdev, 0) - 1) & 159 dev->catas_err.addr); 160 161 if (!request_mem_region(addr, dev->catas_err.size * 4, 162 DRV_NAME)) { 163 mthca_warn(dev, "couldn't request catastrophic error region " 164 "at 0x%lx/0x%x\n", addr, dev->catas_err.size * 4); 165 return; 166 } 167 168 dev->catas_err.map = ioremap(addr, dev->catas_err.size * 4); 169 if (!dev->catas_err.map) { 170 mthca_warn(dev, "couldn't map catastrophic error region " 171 "at 0x%lx/0x%x\n", addr, dev->catas_err.size * 4); 172 release_mem_region(addr, dev->catas_err.size * 4); 173 return; 174 } 175 176 dev->catas_err.timer.data = (unsigned long) dev; 177 dev->catas_err.timer.function = poll_catas; 178 dev->catas_err.timer.expires = jiffies + MTHCA_CATAS_POLL_INTERVAL; 179 INIT_LIST_HEAD(&dev->catas_err.list); 180 add_timer(&dev->catas_err.timer); 181 } 182 183 void mthca_stop_catas_poll(struct mthca_dev *dev) 184 { 185 spin_lock_irq(&catas_lock); 186 dev->catas_err.stop = 1; 187 spin_unlock_irq(&catas_lock); 188 189 del_timer_sync(&dev->catas_err.timer); 190 191 if (dev->catas_err.map) { 192 iounmap(dev->catas_err.map); 193 release_mem_region(pci_resource_start(dev->pdev, 0) + 194 ((pci_resource_len(dev->pdev, 0) - 1) & 195 dev->catas_err.addr), 196 dev->catas_err.size * 4); 197 } 198 199 spin_lock_irq(&catas_lock); 200 list_del(&dev->catas_err.list); 201 spin_unlock_irq(&catas_lock); 202 } 203 204 int __init mthca_catas_init(void) 205 { 206 INIT_WORK(&catas_work, catas_reset, NULL); 207 208 catas_wq = create_singlethread_workqueue("mthca_catas"); 209 if (!catas_wq) 210 return -ENOMEM; 211 212 return 0; 213 } 214 215 void mthca_catas_cleanup(void) 216 { 217 destroy_workqueue(catas_wq); 218 } 219