diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2011-11-05 22:32:53 (GMT) |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2011-11-05 22:32:53 (GMT) |
commit | cd3f07d1e6e59fade92a8edaff9315cc534199cf (patch) | |
tree | a2f30bccdfad1826a81c098c6766c2a4345fc171 /drivers/scsi/hpsa.c | |
parent | dc822e57ee05856291a8c9324d2309076ee5f5e2 (diff) | |
parent | f7c9c6bb14f3104608a3a83cadea10a6943d2804 (diff) | |
download | linux-cd3f07d1e6e59fade92a8edaff9315cc534199cf.tar.xz |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (45 commits)
[SCSI] Fix block queue and elevator memory leak in scsi_alloc_sdev
[SCSI] scsi_dh_alua: Fix the time inteval for alua rtpg commands
[SCSI] scsi_transport_iscsi: Fix documentation os parameter
[SCSI] mv_sas: OCZ RevoDrive3 & zDrive R4 support
[SCSI] libfc: improve flogi retries to avoid lport stuck
[SCSI] libfc: avoid exchanges collision during lport reset
[SCSI] libfc: fix checking FC_TYPE_BLS
[SCSI] edd: Treat "XPRS" host bus type the same as "PCI"
[SCSI] isci: overriding max_concurr_spinup oem parameter by max(oem, user)
[SCSI] isci: revert bcn filtering
[SCSI] isci: Fix hard reset timeout conditions.
[SCSI] isci: No need to manage the pending reset bit on pending requests.
[SCSI] isci: Remove redundant isci_request.ttype field.
[SCSI] isci: Fix task management for SMP, SATA and on dev remove.
[SCSI] isci: No task_done callbacks in error handler paths.
[SCSI] isci: Handle task request timeouts correctly.
[SCSI] isci: Fix tag leak in tasks and terminated requests.
[SCSI] isci: Immediately fail I/O to removed devices.
[SCSI] isci: Lookup device references through requests in completions.
[SCSI] ipr: add definitions for additional adapter
...
Diffstat (limited to 'drivers/scsi/hpsa.c')
-rw-r--r-- | drivers/scsi/hpsa.c | 221 |
1 files changed, 200 insertions, 21 deletions
diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c index bbdc9f9..e76107b 100644 --- a/drivers/scsi/hpsa.c +++ b/drivers/scsi/hpsa.c @@ -48,6 +48,7 @@ #include <linux/bitmap.h> #include <linux/atomic.h> #include <linux/kthread.h> +#include <linux/jiffies.h> #include "hpsa_cmd.h" #include "hpsa.h" @@ -127,6 +128,10 @@ static struct board_type products[] = { static int number_of_controllers; +static struct list_head hpsa_ctlr_list = LIST_HEAD_INIT(hpsa_ctlr_list); +static spinlock_t lockup_detector_lock; +static struct task_struct *hpsa_lockup_detector; + static irqreturn_t do_hpsa_intr_intx(int irq, void *dev_id); static irqreturn_t do_hpsa_intr_msi(int irq, void *dev_id); static int hpsa_ioctl(struct scsi_device *dev, int cmd, void *arg); @@ -484,6 +489,7 @@ static struct scsi_host_template hpsa_driver_template = { #endif .sdev_attrs = hpsa_sdev_attrs, .shost_attrs = hpsa_shost_attrs, + .max_sectors = 8192, }; @@ -566,16 +572,16 @@ static int hpsa_find_target_lun(struct ctlr_info *h, * assumes h->devlock is held */ int i, found = 0; - DECLARE_BITMAP(lun_taken, HPSA_MAX_SCSI_DEVS_PER_HBA); + DECLARE_BITMAP(lun_taken, HPSA_MAX_DEVICES); - memset(&lun_taken[0], 0, HPSA_MAX_SCSI_DEVS_PER_HBA >> 3); + memset(&lun_taken[0], 0, HPSA_MAX_DEVICES >> 3); for (i = 0; i < h->ndevices; i++) { if (h->dev[i]->bus == bus && h->dev[i]->target != -1) set_bit(h->dev[i]->target, lun_taken); } - for (i = 0; i < HPSA_MAX_SCSI_DEVS_PER_HBA; i++) { + for (i = 0; i < HPSA_MAX_DEVICES; i++) { if (!test_bit(i, lun_taken)) { /* *bus = 1; */ *target = i; @@ -598,7 +604,7 @@ static int hpsa_scsi_add_entry(struct ctlr_info *h, int hostno, unsigned char addr1[8], addr2[8]; struct hpsa_scsi_dev_t *sd; - if (n >= HPSA_MAX_SCSI_DEVS_PER_HBA) { + if (n >= HPSA_MAX_DEVICES) { dev_err(&h->pdev->dev, "too many devices, some will be " "inaccessible.\n"); return -1; @@ -673,7 +679,7 @@ static void hpsa_scsi_replace_entry(struct ctlr_info *h, int hostno, struct hpsa_scsi_dev_t *removed[], int *nremoved) { /* assumes h->devlock is held */ - BUG_ON(entry < 0 || entry >= HPSA_MAX_SCSI_DEVS_PER_HBA); + BUG_ON(entry < 0 || entry >= HPSA_MAX_DEVICES); removed[*nremoved] = h->dev[entry]; (*nremoved)++; @@ -702,7 +708,7 @@ static void hpsa_scsi_remove_entry(struct ctlr_info *h, int hostno, int entry, int i; struct hpsa_scsi_dev_t *sd; - BUG_ON(entry < 0 || entry >= HPSA_MAX_SCSI_DEVS_PER_HBA); + BUG_ON(entry < 0 || entry >= HPSA_MAX_DEVICES); sd = h->dev[entry]; removed[*nremoved] = h->dev[entry]; @@ -814,10 +820,8 @@ static void adjust_hpsa_scsi_table(struct ctlr_info *h, int hostno, int nadded, nremoved; struct Scsi_Host *sh = NULL; - added = kzalloc(sizeof(*added) * HPSA_MAX_SCSI_DEVS_PER_HBA, - GFP_KERNEL); - removed = kzalloc(sizeof(*removed) * HPSA_MAX_SCSI_DEVS_PER_HBA, - GFP_KERNEL); + added = kzalloc(sizeof(*added) * HPSA_MAX_DEVICES, GFP_KERNEL); + removed = kzalloc(sizeof(*removed) * HPSA_MAX_DEVICES, GFP_KERNEL); if (!added || !removed) { dev_warn(&h->pdev->dev, "out of memory in " @@ -1338,6 +1342,22 @@ static inline void hpsa_scsi_do_simple_cmd_core(struct ctlr_info *h, wait_for_completion(&wait); } +static void hpsa_scsi_do_simple_cmd_core_if_no_lockup(struct ctlr_info *h, + struct CommandList *c) +{ + unsigned long flags; + + /* If controller lockup detected, fake a hardware error. */ + spin_lock_irqsave(&h->lock, flags); + if (unlikely(h->lockup_detected)) { + spin_unlock_irqrestore(&h->lock, flags); + c->err_info->CommandStatus = CMD_HARDWARE_ERR; + } else { + spin_unlock_irqrestore(&h->lock, flags); + hpsa_scsi_do_simple_cmd_core(h, c); + } +} + static void hpsa_scsi_do_simple_cmd_with_retry(struct ctlr_info *h, struct CommandList *c, int data_direction) { @@ -1735,7 +1755,6 @@ static int add_msa2xxx_enclosure_device(struct ctlr_info *h, if (is_scsi_rev_5(h)) return 0; /* p1210m doesn't need to do this. */ -#define MAX_MSA2XXX_ENCLOSURES 32 if (*nmsa2xxx_enclosures >= MAX_MSA2XXX_ENCLOSURES) { dev_warn(&h->pdev->dev, "Maximum number of MSA2XXX " "enclosures exceeded. Check your hardware " @@ -1846,8 +1865,7 @@ static void hpsa_update_scsi_devices(struct ctlr_info *h, int hostno) int raid_ctlr_position; DECLARE_BITMAP(lunzerobits, HPSA_MAX_TARGETS_PER_CTLR); - currentsd = kzalloc(sizeof(*currentsd) * HPSA_MAX_SCSI_DEVS_PER_HBA, - GFP_KERNEL); + currentsd = kzalloc(sizeof(*currentsd) * HPSA_MAX_DEVICES, GFP_KERNEL); physdev_list = kzalloc(reportlunsize, GFP_KERNEL); logdev_list = kzalloc(reportlunsize, GFP_KERNEL); tmpdevice = kzalloc(sizeof(*tmpdevice), GFP_KERNEL); @@ -1870,6 +1888,13 @@ static void hpsa_update_scsi_devices(struct ctlr_info *h, int hostno) /* Allocate the per device structures */ for (i = 0; i < ndevs_to_allocate; i++) { + if (i >= HPSA_MAX_DEVICES) { + dev_warn(&h->pdev->dev, "maximum devices (%d) exceeded." + " %d devices ignored.\n", HPSA_MAX_DEVICES, + ndevs_to_allocate - HPSA_MAX_DEVICES); + break; + } + currentsd[i] = kzalloc(sizeof(*currentsd[i]), GFP_KERNEL); if (!currentsd[i]) { dev_warn(&h->pdev->dev, "out of memory at %s:%d\n", @@ -1956,7 +1981,7 @@ static void hpsa_update_scsi_devices(struct ctlr_info *h, int hostno) default: break; } - if (ncurrent >= HPSA_MAX_SCSI_DEVS_PER_HBA) + if (ncurrent >= HPSA_MAX_DEVICES) break; } adjust_hpsa_scsi_table(h, hostno, currentsd, ncurrent); @@ -2048,8 +2073,14 @@ static int hpsa_scsi_queue_command_lck(struct scsi_cmnd *cmd, } memcpy(scsi3addr, dev->scsi3addr, sizeof(scsi3addr)); - /* Need a lock as this is being allocated from the pool */ spin_lock_irqsave(&h->lock, flags); + if (unlikely(h->lockup_detected)) { + spin_unlock_irqrestore(&h->lock, flags); + cmd->result = DID_ERROR << 16; + done(cmd); + return 0; + } + /* Need a lock as this is being allocated from the pool */ c = cmd_alloc(h); spin_unlock_irqrestore(&h->lock, flags); if (c == NULL) { /* trouble... */ @@ -2601,7 +2632,7 @@ static int hpsa_passthru_ioctl(struct ctlr_info *h, void __user *argp) c->SG[0].Len = iocommand.buf_size; c->SG[0].Ext = 0; /* we are not chaining*/ } - hpsa_scsi_do_simple_cmd_core(h, c); + hpsa_scsi_do_simple_cmd_core_if_no_lockup(h, c); if (iocommand.buf_size > 0) hpsa_pci_unmap(h->pdev, c, 1, PCI_DMA_BIDIRECTIONAL); check_ioctl_unit_attention(h, c); @@ -2724,7 +2755,7 @@ static int hpsa_big_passthru_ioctl(struct ctlr_info *h, void __user *argp) c->SG[i].Ext = 0; } } - hpsa_scsi_do_simple_cmd_core(h, c); + hpsa_scsi_do_simple_cmd_core_if_no_lockup(h, c); if (sg_used) hpsa_pci_unmap(h->pdev, c, sg_used, PCI_DMA_BIDIRECTIONAL); check_ioctl_unit_attention(h, c); @@ -2872,6 +2903,8 @@ static void fill_cmd(struct CommandList *c, u8 cmd, struct ctlr_info *h, c->Request.Timeout = 0; c->Request.CDB[0] = BMIC_WRITE; c->Request.CDB[6] = BMIC_CACHE_FLUSH; + c->Request.CDB[7] = (size >> 8) & 0xFF; + c->Request.CDB[8] = size & 0xFF; break; case TEST_UNIT_READY: c->Request.CDBLen = 6; @@ -3091,6 +3124,7 @@ static irqreturn_t hpsa_intx_discard_completions(int irq, void *dev_id) if (interrupt_not_for_us(h)) return IRQ_NONE; spin_lock_irqsave(&h->lock, flags); + h->last_intr_timestamp = get_jiffies_64(); while (interrupt_pending(h)) { raw_tag = get_next_completion(h); while (raw_tag != FIFO_EMPTY) @@ -3110,6 +3144,7 @@ static irqreturn_t hpsa_msix_discard_completions(int irq, void *dev_id) return IRQ_NONE; spin_lock_irqsave(&h->lock, flags); + h->last_intr_timestamp = get_jiffies_64(); raw_tag = get_next_completion(h); while (raw_tag != FIFO_EMPTY) raw_tag = next_command(h); @@ -3126,6 +3161,7 @@ static irqreturn_t do_hpsa_intr_intx(int irq, void *dev_id) if (interrupt_not_for_us(h)) return IRQ_NONE; spin_lock_irqsave(&h->lock, flags); + h->last_intr_timestamp = get_jiffies_64(); while (interrupt_pending(h)) { raw_tag = get_next_completion(h); while (raw_tag != FIFO_EMPTY) { @@ -3146,6 +3182,7 @@ static irqreturn_t do_hpsa_intr_msi(int irq, void *dev_id) u32 raw_tag; spin_lock_irqsave(&h->lock, flags); + h->last_intr_timestamp = get_jiffies_64(); raw_tag = get_next_completion(h); while (raw_tag != FIFO_EMPTY) { if (hpsa_tag_contains_index(raw_tag)) @@ -4090,6 +4127,149 @@ static void hpsa_undo_allocations_after_kdump_soft_reset(struct ctlr_info *h) kfree(h); } +static void remove_ctlr_from_lockup_detector_list(struct ctlr_info *h) +{ + assert_spin_locked(&lockup_detector_lock); + if (!hpsa_lockup_detector) + return; + if (h->lockup_detected) + return; /* already stopped the lockup detector */ + list_del(&h->lockup_list); +} + +/* Called when controller lockup detected. */ +static void fail_all_cmds_on_list(struct ctlr_info *h, struct list_head *list) +{ + struct CommandList *c = NULL; + + assert_spin_locked(&h->lock); + /* Mark all outstanding commands as failed and complete them. */ + while (!list_empty(list)) { + c = list_entry(list->next, struct CommandList, list); + c->err_info->CommandStatus = CMD_HARDWARE_ERR; + finish_cmd(c, c->Header.Tag.lower); + } +} + +static void controller_lockup_detected(struct ctlr_info *h) +{ + unsigned long flags; + + assert_spin_locked(&lockup_detector_lock); + remove_ctlr_from_lockup_detector_list(h); + h->access.set_intr_mask(h, HPSA_INTR_OFF); + spin_lock_irqsave(&h->lock, flags); + h->lockup_detected = readl(h->vaddr + SA5_SCRATCHPAD_OFFSET); + spin_unlock_irqrestore(&h->lock, flags); + dev_warn(&h->pdev->dev, "Controller lockup detected: 0x%08x\n", + h->lockup_detected); + pci_disable_device(h->pdev); + spin_lock_irqsave(&h->lock, flags); + fail_all_cmds_on_list(h, &h->cmpQ); + fail_all_cmds_on_list(h, &h->reqQ); + spin_unlock_irqrestore(&h->lock, flags); +} + +#define HEARTBEAT_SAMPLE_INTERVAL (10 * HZ) +#define HEARTBEAT_CHECK_MINIMUM_INTERVAL (HEARTBEAT_SAMPLE_INTERVAL / 2) + +static void detect_controller_lockup(struct ctlr_info *h) +{ + u64 now; + u32 heartbeat; + unsigned long flags; + + assert_spin_locked(&lockup_detector_lock); + now = get_jiffies_64(); + /* If we've received an interrupt recently, we're ok. */ + if (time_after64(h->last_intr_timestamp + + (HEARTBEAT_CHECK_MINIMUM_INTERVAL), now)) + return; + + /* + * If we've already checked the heartbeat recently, we're ok. + * This could happen if someone sends us a signal. We + * otherwise don't care about signals in this thread. + */ + if (time_after64(h->last_heartbeat_timestamp + + (HEARTBEAT_CHECK_MINIMUM_INTERVAL), now)) + return; + + /* If heartbeat has not changed since we last looked, we're not ok. */ + spin_lock_irqsave(&h->lock, flags); + heartbeat = readl(&h->cfgtable->HeartBeat); + spin_unlock_irqrestore(&h->lock, flags); + if (h->last_heartbeat == heartbeat) { + controller_lockup_detected(h); + return; + } + + /* We're ok. */ + h->last_heartbeat = heartbeat; + h->last_heartbeat_timestamp = now; +} + +static int detect_controller_lockup_thread(void *notused) +{ + struct ctlr_info *h; + unsigned long flags; + + while (1) { + struct list_head *this, *tmp; + + schedule_timeout_interruptible(HEARTBEAT_SAMPLE_INTERVAL); + if (kthread_should_stop()) + break; + spin_lock_irqsave(&lockup_detector_lock, flags); + list_for_each_safe(this, tmp, &hpsa_ctlr_list) { + h = list_entry(this, struct ctlr_info, lockup_list); + detect_controller_lockup(h); + } + spin_unlock_irqrestore(&lockup_detector_lock, flags); + } + return 0; +} + +static void add_ctlr_to_lockup_detector_list(struct ctlr_info *h) +{ + unsigned long flags; + + spin_lock_irqsave(&lockup_detector_lock, flags); + list_add_tail(&h->lockup_list, &hpsa_ctlr_list); + spin_unlock_irqrestore(&lockup_detector_lock, flags); +} + +static void start_controller_lockup_detector(struct ctlr_info *h) +{ + /* Start the lockup detector thread if not already started */ + if (!hpsa_lockup_detector) { + spin_lock_init(&lockup_detector_lock); + hpsa_lockup_detector = + kthread_run(detect_controller_lockup_thread, + NULL, "hpsa"); + } + if (!hpsa_lockup_detector) { + dev_warn(&h->pdev->dev, + "Could not start lockup detector thread\n"); + return; + } + add_ctlr_to_lockup_detector_list(h); +} + +static void stop_controller_lockup_detector(struct ctlr_info *h) +{ + unsigned long flags; + + spin_lock_irqsave(&lockup_detector_lock, flags); + remove_ctlr_from_lockup_detector_list(h); + /* If the list of ctlr's to monitor is empty, stop the thread */ + if (list_empty(&hpsa_ctlr_list)) { + kthread_stop(hpsa_lockup_detector); + hpsa_lockup_detector = NULL; + } + spin_unlock_irqrestore(&lockup_detector_lock, flags); +} + static int __devinit hpsa_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) { @@ -4127,7 +4307,6 @@ reinit_after_soft_reset: return -ENOMEM; h->pdev = pdev; - h->busy_initializing = 1; h->intr_mode = hpsa_simple_mode ? SIMPLE_MODE_INT : PERF_MODE_INT; INIT_LIST_HEAD(&h->cmpQ); INIT_LIST_HEAD(&h->reqQ); @@ -4236,7 +4415,7 @@ reinit_after_soft_reset: hpsa_hba_inquiry(h); hpsa_register_scsi(h); /* hook ourselves into SCSI subsystem */ - h->busy_initializing = 0; + start_controller_lockup_detector(h); return 1; clean4: @@ -4245,7 +4424,6 @@ clean4: free_irq(h->intr[h->intr_mode], h); clean2: clean1: - h->busy_initializing = 0; kfree(h); return rc; } @@ -4300,10 +4478,11 @@ static void __devexit hpsa_remove_one(struct pci_dev *pdev) struct ctlr_info *h; if (pci_get_drvdata(pdev) == NULL) { - dev_err(&pdev->dev, "unable to remove device \n"); + dev_err(&pdev->dev, "unable to remove device\n"); return; } h = pci_get_drvdata(pdev); + stop_controller_lockup_detector(h); hpsa_unregister_scsi(h); /* unhook from SCSI subsystem */ hpsa_shutdown(pdev); iounmap(h->vaddr); |