1/* 2 * Copyright (c) 2007 Cisco Systems, Inc. All rights reserved. 3 * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved. 4 * 5 * This software is available to you under a choice of one of two 6 * licenses. You may choose to be licensed under the terms of the GNU 7 * General Public License (GPL) Version 2, available from the file 8 * COPYING in the main directory of this source tree, or the 9 * OpenIB.org BSD license below: 10 * 11 * Redistribution and use in source and binary forms, with or 12 * without modification, are permitted provided that the following 13 * conditions are met: 14 * 15 * - Redistributions of source code must retain the above 16 * copyright notice, this list of conditions and the following 17 * disclaimer. 18 * 19 * - Redistributions in binary form must reproduce the above 20 * copyright notice, this list of conditions and the following 21 * disclaimer in the documentation and/or other materials 22 * provided with the distribution. 23 * 24 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 25 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 26 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND 27 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS 28 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN 29 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN 30 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 31 * SOFTWARE. 32 */ 33 34#include <linux/workqueue.h> 35#include <linux/module.h> 36 37#include "mlx4.h" 38 39#define MLX4_CATAS_POLL_INTERVAL (5 * HZ) 40 41 42static DEFINE_SPINLOCK(catas_lock); 43 44static LIST_HEAD(catas_list); 45static struct work_struct catas_work; 46 47static int internal_err_reset = 1; 48module_param(internal_err_reset, int, 0644); 49MODULE_PARM_DESC(internal_err_reset, 50 "Reset device on internal errors if non-zero" 51 " (default 1, in SRIOV mode default is 0)"); 52 53static void dump_err_buf(struct mlx4_dev *dev) 54{ 55 struct mlx4_priv *priv = mlx4_priv(dev); 56 57 int i; 58 59 mlx4_err(dev, "Internal error detected:\n"); 60 for (i = 0; i < priv->fw.catas_size; ++i) 61 mlx4_err(dev, " buf[%02x]: %08x\n", 62 i, swab32(readl(priv->catas_err.map + i))); 63} 64 65static void poll_catas(unsigned long dev_ptr) 66{ 67 struct mlx4_dev *dev = (struct mlx4_dev *) dev_ptr; 68 struct mlx4_priv *priv = mlx4_priv(dev); 69 70 if (readl(priv->catas_err.map)) { 71 /* If the device is off-line, we cannot try to recover it */ 72 if (pci_channel_offline(dev->pdev)) 73 mod_timer(&priv->catas_err.timer, 74 round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL)); 75 else { 76 dump_err_buf(dev); 77 mlx4_dispatch_event(dev, MLX4_DEV_EVENT_CATASTROPHIC_ERROR, 0); 78 79 if (internal_err_reset) { 80 spin_lock(&catas_lock); 81 list_add(&priv->catas_err.list, &catas_list); 82 spin_unlock(&catas_lock); 83 84 queue_work(mlx4_wq, &catas_work); 85 } 86 } 87 } else 88 mod_timer(&priv->catas_err.timer, 89 round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL)); 90} 91 92static void catas_reset(struct work_struct *work) 93{ 94 struct mlx4_priv *priv, *tmppriv; 95 struct mlx4_dev *dev; 96 97 LIST_HEAD(tlist); 98 int ret; 99 100 spin_lock_irq(&catas_lock); 101 list_splice_init(&catas_list, &tlist); 102 spin_unlock_irq(&catas_lock); 103 104 list_for_each_entry_safe(priv, tmppriv, &tlist, catas_err.list) { 105 struct pci_dev *pdev = priv->dev.pdev; 106 107 /* If the device is off-line, we cannot reset it */ 108 if (pci_channel_offline(pdev)) 109 continue; 110 111 ret = mlx4_restart_one(priv->dev.pdev); 112 /* 'priv' now is not valid */ 113 if (ret) 114 pr_err("mlx4 %s: Reset failed (%d)\n", 115 pci_name(pdev), ret); 116 else { 117 dev = pci_get_drvdata(pdev); 118 mlx4_dbg(dev, "Reset succeeded\n"); 119 } 120 } 121} 122 123void mlx4_start_catas_poll(struct mlx4_dev *dev) 124{ 125 struct mlx4_priv *priv = mlx4_priv(dev); 126 phys_addr_t addr; 127 128 /*If we are in SRIOV the default of the module param must be 0*/ 129 if (mlx4_is_mfunc(dev)) 130 internal_err_reset = 0; 131 132 INIT_LIST_HEAD(&priv->catas_err.list); 133 init_timer(&priv->catas_err.timer); 134 priv->catas_err.map = NULL; 135 136 addr = pci_resource_start(dev->pdev, priv->fw.catas_bar) + 137 priv->fw.catas_offset; 138 139 priv->catas_err.map = ioremap(addr, priv->fw.catas_size * 4); 140 if (!priv->catas_err.map) { 141 mlx4_warn(dev, "Failed to map internal error buffer at 0x%llx\n", 142 (unsigned long long) addr); 143 return; 144 } 145 146 priv->catas_err.timer.data = (unsigned long) dev; 147 priv->catas_err.timer.function = poll_catas; 148 priv->catas_err.timer.expires = 149 round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL); 150 add_timer(&priv->catas_err.timer); 151} 152 153void mlx4_stop_catas_poll(struct mlx4_dev *dev) 154{ 155 struct mlx4_priv *priv = mlx4_priv(dev); 156 157 del_timer_sync(&priv->catas_err.timer); 158 159 if (priv->catas_err.map) 160 iounmap(priv->catas_err.map); 161 162 spin_lock_irq(&catas_lock); 163 list_del(&priv->catas_err.list); 164 spin_unlock_irq(&catas_lock); 165} 166 167void __init mlx4_catas_init(void) 168{ 169 INIT_WORK(&catas_work, catas_reset); 170} 171