1 files changed, 232 insertions, 95 deletions

diff --git a/drivers/block/mtip32xx/mtip32xx.c b/drivers/block/mtip32xx/mtip32xx.c
index 92250af84e7d..32c678028e53 100644
--- a/drivers/block/mtip32xx/mtip32xx.c
+++ b/drivers/block/mtip32xx/mtip32xx.c
@@ -81,12 +81,17 @@
 /* Device instance number, incremented each time a device is probed. */
 static int instance;
 
+struct list_head online_list;
+struct list_head removing_list;
+spinlock_t dev_lock;
+
 /*
  * Global variable used to hold the major block device number
  * allocated in mtip_init().
  */
 static int mtip_major;
 static struct dentry *dfs_parent;
+static struct dentry *dfs_device_status;
 
 static u32 cpu_use[NR_CPUS];
 
@@ -243,40 +248,31 @@ static inline void release_slot(struct mtip_port *port, int tag)
 /*
  * Reset the HBA (without sleeping)
  *
- * Just like hba_reset, except does not call sleep, so can be
- * run from interrupt/tasklet context.
- *
  * @dd Pointer to the driver data structure.
  *
  * return value
  *      0       The reset was successful.
  *      -1      The HBA Reset bit did not clear.
  */
-static int hba_reset_nosleep(struct driver_data *dd)
+static int mtip_hba_reset(struct driver_data *dd)
 {
         unsigned long timeout;
 
-        /* Chip quirk: quiesce any chip function */
-        mdelay(10);
-
         /* Set the reset bit */
         writel(HOST_RESET, dd->mmio + HOST_CTL);
 
         /* Flush */
         readl(dd->mmio + HOST_CTL);
 
-        /*
-         * Wait 10ms then spin for up to 1 second
-         * waiting for reset acknowledgement
-         */
-        timeout = jiffies + msecs_to_jiffies(1000);
-        mdelay(10);
-        while ((readl(dd->mmio + HOST_CTL) & HOST_RESET)
-                 && time_before(jiffies, timeout))
-                mdelay(1);
+        /* Spin for up to 2 seconds, waiting for reset acknowledgement */
+        timeout = jiffies + msecs_to_jiffies(2000);
+        do {
+                mdelay(10);
+                if (test_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag))
+                        return -1;
 
-        if (test_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag))
-                return -1;
+        } while ((readl(dd->mmio + HOST_CTL) & HOST_RESET)
+                 && time_before(jiffies, timeout));
 
         if (readl(dd->mmio + HOST_CTL) & HOST_RESET)
                 return -1;
@@ -481,7 +477,7 @@ static void mtip_restart_port(struct mtip_port *port)
                 dev_warn(&port->dd->pdev->dev,
                         "PxCMD.CR not clear, escalating reset\n");
 
-                if (hba_reset_nosleep(port->dd))
+                if (mtip_hba_reset(port->dd))
                         dev_err(&port->dd->pdev->dev,
                                 "HBA reset escalation failed.\n");
 
@@ -527,6 +523,26 @@ static void mtip_restart_port(struct mtip_port *port)
 
 }
 
+static int mtip_device_reset(struct driver_data *dd)
+{
+        int rv = 0;
+
+        if (mtip_check_surprise_removal(dd->pdev))
+                return 0;
+
+        if (mtip_hba_reset(dd) < 0)
+                rv = -EFAULT;
+
+        mdelay(1);
+        mtip_init_port(dd->port);
+        mtip_start_port(dd->port);
+
+        /* Enable interrupts on the HBA. */
+        writel(readl(dd->mmio + HOST_CTL) | HOST_IRQ_EN,
+                                        dd->mmio + HOST_CTL);
+        return rv;
+}
+
 /*
  * Helper function for tag logging
  */
@@ -632,7 +648,7 @@ static void mtip_timeout_function(unsigned long int data)
         if (cmdto_cnt) {
                 print_tags(port->dd, "timed out", tagaccum, cmdto_cnt);
                 if (!test_bit(MTIP_PF_IC_ACTIVE_BIT, &port->flags)) {
-                        mtip_restart_port(port);
+                        mtip_device_reset(port->dd);
                         wake_up_interruptible(&port->svc_wait);
                 }
                 clear_bit(MTIP_PF_EH_ACTIVE_BIT, &port->flags);
@@ -1283,11 +1299,11 @@ static int mtip_exec_internal_command(struct mtip_port *port,
         int rv = 0, ready2go = 1;
         struct mtip_cmd *int_cmd = &port->commands[MTIP_TAG_INTERNAL];
         unsigned long to;
+        struct driver_data *dd = port->dd;
 
         /* Make sure the buffer is 8 byte aligned. This is asic specific. */
         if (buffer & 0x00000007) {
-                dev_err(&port->dd->pdev->dev,
-                        "SG buffer is not 8 byte aligned\n");
+                dev_err(&dd->pdev->dev, "SG buffer is not 8 byte aligned\n");
                 return -EFAULT;
         }
 
@@ -1300,23 +1316,21 @@ static int mtip_exec_internal_command(struct mtip_port *port,
                 mdelay(100);
         } while (time_before(jiffies, to));
         if (!ready2go) {
-                dev_warn(&port->dd->pdev->dev,
+                dev_warn(&dd->pdev->dev,
                         "Internal cmd active. new cmd [%02X]\n", fis->command);
                 return -EBUSY;
         }
         set_bit(MTIP_PF_IC_ACTIVE_BIT, &port->flags);
         port->ic_pause_timer = 0;
 
-        if (fis->command == ATA_CMD_SEC_ERASE_UNIT)
-                clear_bit(MTIP_PF_SE_ACTIVE_BIT, &port->flags);
-        else if (fis->command == ATA_CMD_DOWNLOAD_MICRO)
-                clear_bit(MTIP_PF_DM_ACTIVE_BIT, &port->flags);
+        clear_bit(MTIP_PF_SE_ACTIVE_BIT, &port->flags);
+        clear_bit(MTIP_PF_DM_ACTIVE_BIT, &port->flags);
 
         if (atomic == GFP_KERNEL) {
                 if (fis->command != ATA_CMD_STANDBYNOW1) {
                         /* wait for io to complete if non atomic */
                         if (mtip_quiesce_io(port, 5000) < 0) {
-                                dev_warn(&port->dd->pdev->dev,
+                                dev_warn(&dd->pdev->dev,
                                         "Failed to quiesce IO\n");
                                 release_slot(port, MTIP_TAG_INTERNAL);
                                 clear_bit(MTIP_PF_IC_ACTIVE_BIT, &port->flags);
@@ -1361,58 +1375,84 @@ static int mtip_exec_internal_command(struct mtip_port *port,
         /* Issue the command to the hardware */
         mtip_issue_non_ncq_command(port, MTIP_TAG_INTERNAL);
 
-        /* Poll if atomic, wait_for_completion otherwise */
         if (atomic == GFP_KERNEL) {
                 /* Wait for the command to complete or timeout. */
-                if (wait_for_completion_timeout(
+                if (wait_for_completion_interruptible_timeout(
                                 &wait,
-                                msecs_to_jiffies(timeout)) == 0) {
-                        dev_err(&port->dd->pdev->dev,
-                                "Internal command did not complete [%d] "
-                                "within timeout of  %lu ms\n",
-                                atomic, timeout);
-                        if (mtip_check_surprise_removal(port->dd->pdev) ||
+                                msecs_to_jiffies(timeout)) <= 0) {
+                        if (rv == -ERESTARTSYS) { /* interrupted */
+                                dev_err(&dd->pdev->dev,
+                                        "Internal command [%02X] was interrupted after %lu ms\n",
+                                        fis->command, timeout);
+                                rv = -EINTR;
+                                goto exec_ic_exit;
+                        } else if (rv == 0) /* timeout */
+                                dev_err(&dd->pdev->dev,
+                                        "Internal command did not complete [%02X] within timeout of  %lu ms\n",
+                                        fis->command, timeout);
+                        else
+                                dev_err(&dd->pdev->dev,
+                                        "Internal command [%02X] wait returned code [%d] after %lu ms - unhandled\n",
+                                        fis->command, rv, timeout);
+
+                        if (mtip_check_surprise_removal(dd->pdev) ||
                                 test_bit(MTIP_DDF_REMOVE_PENDING_BIT,
-                                                &port->dd->dd_flag)) {
+                                                &dd->dd_flag)) {
+                                dev_err(&dd->pdev->dev,
+                                        "Internal command [%02X] wait returned due to SR\n",
+                                        fis->command);
                                 rv = -ENXIO;
                                 goto exec_ic_exit;
                         }
+                        mtip_device_reset(dd); /* recover from timeout issue */
                         rv = -EAGAIN;
+                        goto exec_ic_exit;
                 }
         } else {
+                u32 hba_stat, port_stat;
+
                 /* Spin for <timeout> checking if command still outstanding */
                 timeout = jiffies + msecs_to_jiffies(timeout);
                 while ((readl(port->cmd_issue[MTIP_TAG_INTERNAL])
                                 & (1 << MTIP_TAG_INTERNAL))
                                 && time_before(jiffies, timeout)) {
-                        if (mtip_check_surprise_removal(port->dd->pdev)) {
+                        if (mtip_check_surprise_removal(dd->pdev)) {
                                 rv = -ENXIO;
                                 goto exec_ic_exit;
                         }
                         if ((fis->command != ATA_CMD_STANDBYNOW1) &&
                                 test_bit(MTIP_DDF_REMOVE_PENDING_BIT,
-                                                &port->dd->dd_flag)) {
+                                                &dd->dd_flag)) {
                                 rv = -ENXIO;
                                 goto exec_ic_exit;
                         }
-                        if (readl(port->mmio + PORT_IRQ_STAT) & PORT_IRQ_ERR) {
-                                atomic_inc(&int_cmd->active); /* error */
-                                break;
+                        port_stat = readl(port->mmio + PORT_IRQ_STAT);
+                        if (!port_stat)
+                                continue;
+
+                        if (port_stat & PORT_IRQ_ERR) {
+                                dev_err(&dd->pdev->dev,
+                                        "Internal command [%02X] failed\n",
+                                        fis->command);
+                                mtip_device_reset(dd);
+                                rv = -EIO;
+                                goto exec_ic_exit;
+                        } else {
+                                writel(port_stat, port->mmio + PORT_IRQ_STAT);
+                                hba_stat = readl(dd->mmio + HOST_IRQ_STAT);
+                                if (hba_stat)
+                                        writel(hba_stat,
+                                                dd->mmio + HOST_IRQ_STAT);
                         }
+                        break;
                 }
         }
 
-        if (atomic_read(&int_cmd->active) > 1) {
-                dev_err(&port->dd->pdev->dev,
-                        "Internal command [%02X] failed\n", fis->command);
-                rv = -EIO;
-        }
         if (readl(port->cmd_issue[MTIP_TAG_INTERNAL])
                         & (1 << MTIP_TAG_INTERNAL)) {
                 rv = -ENXIO;
-                if (!test_bit(MTIP_DDF_REMOVE_PENDING_BIT,
-                                        &port->dd->dd_flag)) {
-                        mtip_restart_port(port);
+                if (!test_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag)) {
+                        mtip_device_reset(dd);
                         rv = -EAGAIN;
                 }
         }
@@ -1724,7 +1764,8 @@ static int mtip_get_smart_attr(struct mtip_port *port, unsigned int id,
  *      -EINVAL         Invalid parameters passed in, trim not supported
  *      -EIO            Error submitting trim request to hw
  */
-static int mtip_send_trim(struct driver_data *dd, unsigned int lba, unsigned int len)
+static int mtip_send_trim(struct driver_data *dd, unsigned int lba,
+                                unsigned int len)
 {
         int i, rv = 0;
         u64 tlba, tlen, sect_left;
@@ -1811,45 +1852,6 @@ static bool mtip_hw_get_capacity(struct driver_data *dd, sector_t *sectors)
 }
 
 /*
- * Reset the HBA.
- *
- * Resets the HBA by setting the HBA Reset bit in the Global
- * HBA Control register. After setting the HBA Reset bit the
- * function waits for 1 second before reading the HBA Reset
- * bit to make sure it has cleared. If HBA Reset is not clear
- * an error is returned. Cannot be used in non-blockable
- * context.
- *
- * @dd Pointer to the driver data structure.
- *
- * return value
- *      0  The reset was successful.
- *      -1 The HBA Reset bit did not clear.
- */
-static int mtip_hba_reset(struct driver_data *dd)
-{
-        mtip_deinit_port(dd->port);
-
-        /* Set the reset bit */
-        writel(HOST_RESET, dd->mmio + HOST_CTL);
-
-        /* Flush */
-        readl(dd->mmio + HOST_CTL);
-
-        /* Wait for reset to clear */
-        ssleep(1);
-
-        /* Check the bit has cleared */
-        if (readl(dd->mmio + HOST_CTL) & HOST_RESET) {
-                dev_err(&dd->pdev->dev,
-                        "Reset bit did not clear.\n");
-                return -1;
-        }
-
-        return 0;
-}
-
-/*
  * Display the identify command data.
  *
  * @port Pointer to the port data structure.
@@ -2710,6 +2712,100 @@ static ssize_t mtip_hw_show_status(struct device *dev,
 
 static DEVICE_ATTR(status, S_IRUGO, mtip_hw_show_status, NULL);
 
+/* debugsfs entries */
+
+static ssize_t show_device_status(struct device_driver *drv, char *buf)
+{
+        int size = 0;
+        struct driver_data *dd, *tmp;
+        unsigned long flags;
+        char id_buf[42];
+        u16 status = 0;
+
+        spin_lock_irqsave(&dev_lock, flags);
+        size += sprintf(&buf[size], "Devices Present:\n");
+        list_for_each_entry_safe(dd, tmp, &online_list, online_list) {
+                if (dd->pdev) {
+                        if (dd->port &&
+                            dd->port->identify &&
+                            dd->port->identify_valid) {
+                                strlcpy(id_buf,
+                                        (char *) (dd->port->identify + 10), 21);
+                                status = *(dd->port->identify + 141);
+                        } else {
+                                memset(id_buf, 0, 42);
+                                status = 0;
+                        }
+
+                        if (dd->port &&
+                            test_bit(MTIP_PF_REBUILD_BIT, &dd->port->flags)) {
+                                size += sprintf(&buf[size],
+                                        " device %s %s (ftl rebuild %d %%)\n",
+                                        dev_name(&dd->pdev->dev),
+                                        id_buf,
+                                        status);
+                        } else {
+                                size += sprintf(&buf[size],
+                                        " device %s %s\n",
+                                        dev_name(&dd->pdev->dev),
+                                        id_buf);
+                        }
+                }
+        }
+
+        size += sprintf(&buf[size], "Devices Being Removed:\n");
+        list_for_each_entry_safe(dd, tmp, &removing_list, remove_list) {
+                if (dd->pdev) {
+                        if (dd->port &&
+                            dd->port->identify &&
+                            dd->port->identify_valid) {
+                                strlcpy(id_buf,
+                                        (char *) (dd->port->identify+10), 21);
+                                status = *(dd->port->identify + 141);
+                        } else {
+                                memset(id_buf, 0, 42);
+                                status = 0;
+                        }
+
+                        if (dd->port &&
+                            test_bit(MTIP_PF_REBUILD_BIT, &dd->port->flags)) {
+                                size += sprintf(&buf[size],
+                                        " device %s %s (ftl rebuild %d %%)\n",
+                                        dev_name(&dd->pdev->dev),
+                                        id_buf,
+                                        status);
+                        } else {
+                                size += sprintf(&buf[size],
+                                        " device %s %s\n",
+                                        dev_name(&dd->pdev->dev),
+                                        id_buf);
+                        }
+                }
+        }
+        spin_unlock_irqrestore(&dev_lock, flags);
+
+        return size;
+}
+
+static ssize_t mtip_hw_read_device_status(struct file *f, char __user *ubuf,
+                                                size_t len, loff_t *offset)
+{
+        int size = *offset;
+        char buf[MTIP_DFS_MAX_BUF_SIZE];
+
+        if (!len || *offset)
+                return 0;
+
+        size += show_device_status(NULL, buf);
+
+        *offset = size <= len ? size : len;
+        size = copy_to_user(ubuf, buf, *offset);
+        if (size)
+                return -EFAULT;
+
+        return *offset;
+}
+
 static ssize_t mtip_hw_read_registers(struct file *f, char __user *ubuf,
                                   size_t len, loff_t *offset)
 {
@@ -2804,6 +2900,13 @@ static ssize_t mtip_hw_read_flags(struct file *f, char __user *ubuf,
         return *offset;
 }
 
+static const struct file_operations mtip_device_status_fops = {
+        .owner  = THIS_MODULE,
+        .open   = simple_open,
+        .read   = mtip_hw_read_device_status,
+        .llseek = no_llseek,
+};
+
 static const struct file_operations mtip_regs_fops = {
         .owner  = THIS_MODULE,
         .open   = simple_open,
@@ -4161,6 +4264,7 @@ static int mtip_pci_probe(struct pci_dev *pdev,
         const struct cpumask *node_mask;
         int cpu, i = 0, j = 0;
         int my_node = NUMA_NO_NODE;
+        unsigned long flags;
 
         /* Allocate memory for this devices private data. */
         my_node = pcibus_to_node(pdev->bus);
@@ -4218,6 +4322,9 @@ static int mtip_pci_probe(struct pci_dev *pdev,
         dd->pdev        = pdev;
         dd->numa_node   = my_node;
 
+        INIT_LIST_HEAD(&dd->online_list);
+        INIT_LIST_HEAD(&dd->remove_list);
+
         memset(dd->workq_name, 0, 32);
         snprintf(dd->workq_name, 31, "mtipq%d", dd->instance);
 
@@ -4305,6 +4412,14 @@ static int mtip_pci_probe(struct pci_dev *pdev,
         instance++;
         if (rv != MTIP_FTL_REBUILD_MAGIC)
                 set_bit(MTIP_DDF_INIT_DONE_BIT, &dd->dd_flag);
+        else
+                rv = 0; /* device in rebuild state, return 0 from probe */
+
+        /* Add to online list even if in ftl rebuild */
+        spin_lock_irqsave(&dev_lock, flags);
+        list_add(&dd->online_list, &online_list);
+        spin_unlock_irqrestore(&dev_lock, flags);
+
         goto done;
 
 block_initialize_err:
@@ -4338,9 +4453,15 @@ static void mtip_pci_remove(struct pci_dev *pdev)
 {
         struct driver_data *dd = pci_get_drvdata(pdev);
         int counter = 0;
+        unsigned long flags;
 
         set_bit(MTIP_DDF_REMOVE_PENDING_BIT, &dd->dd_flag);
 
+        spin_lock_irqsave(&dev_lock, flags);
+        list_del_init(&dd->online_list);
+        list_add(&dd->remove_list, &removing_list);
+        spin_unlock_irqrestore(&dev_lock, flags);
+
         if (mtip_check_surprise_removal(pdev)) {
                 while (!test_bit(MTIP_DDF_CLEANUP_BIT, &dd->dd_flag)) {
                         counter++;
@@ -4366,6 +4487,10 @@ static void mtip_pci_remove(struct pci_dev *pdev)
 
         pci_disable_msi(pdev);
 
+        spin_lock_irqsave(&dev_lock, flags);
+        list_del_init(&dd->remove_list);
+        spin_unlock_irqrestore(&dev_lock, flags);
+
         kfree(dd);
         pcim_iounmap_regions(pdev, 1 << MTIP_ABAR);
 }
@@ -4513,6 +4638,11 @@ static int __init mtip_init(void)
 
         pr_info(MTIP_DRV_NAME " Version " MTIP_DRV_VERSION "\n");
 
+        spin_lock_init(&dev_lock);
+
+        INIT_LIST_HEAD(&online_list);
+        INIT_LIST_HEAD(&removing_list);
+
         /* Allocate a major block device number to use with this driver. */
         error = register_blkdev(0, MTIP_DRV_NAME);
         if (error <= 0) {
@@ -4522,11 +4652,18 @@ static int __init mtip_init(void)
         }
         mtip_major = error;
 
-        if (!dfs_parent) {
-                dfs_parent = debugfs_create_dir("rssd", NULL);
-                if (IS_ERR_OR_NULL(dfs_parent)) {
-                        pr_warn("Error creating debugfs parent\n");
-                        dfs_parent = NULL;
+        dfs_parent = debugfs_create_dir("rssd", NULL);
+        if (IS_ERR_OR_NULL(dfs_parent)) {
+                pr_warn("Error creating debugfs parent\n");
+                dfs_parent = NULL;
+        }
+        if (dfs_parent) {
+                dfs_device_status = debugfs_create_file("device_status",
+                                        S_IRUGO, dfs_parent, NULL,
+                                        &mtip_device_status_fops);
+                if (IS_ERR_OR_NULL(dfs_device_status)) {
+                        pr_err("Error creating device_status node\n");
+                        dfs_device_status = NULL;
                 }
         }