/*
 *    Disk Array driver for HP Smart Array controllers.
 *    (C) Copyright 2000, 2007 Hewlett-Packard Development Company, L.P.
 *
 *    This program is free software; you can redistribute it and/or modify
 *    it under the terms of the GNU General Public License as published by
 *    the Free Software Foundation; version 2 of the License.
 *
 *    This program is distributed in the hope that it will be useful,
 *    but WITHOUT ANY WARRANTY; without even the implied warranty of
 *    MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 *    General Public License for more details.
 *
 *    You should have received a copy of the GNU General Public License
 *    along with this program; if not, write to the Free Software
 *    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
 *    02111-1307, USA.
 *
 *    Questions/Comments/Bugfixes to iss_storagedev@hp.com
 *
 */

#include <linux/module.h>
#include <linux/interrupt.h>
#include <linux/types.h>
#include <linux/pci.h>
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/delay.h>
#include <linux/major.h>
#include <linux/fs.h>
#include <linux/bio.h>
#include <linux/blkpg.h>
#include <linux/timer.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/init.h>
#include <linux/hdreg.h>
#include <linux/spinlock.h>
#include <linux/compat.h>
#include <linux/blktrace_api.h>
#include <asm/uaccess.h>
#include <asm/io.h>

#include <linux/dma-mapping.h>
#include <linux/blkdev.h>
#include <linux/genhd.h>
#include <linux/completion.h>
#include <scsi/scsi.h>
#include <scsi/sg.h>
#include <scsi/scsi_ioctl.h>
#include <linux/cdrom.h>
#include <linux/scatterlist.h>
#include <linux/kthread.h>

#define CCISS_DRIVER_VERSION(maj,min,submin) ((maj<<16)|(min<<8)|(submin))
#define DRIVER_NAME "HP CISS Driver (v 3.6.20)"
#define DRIVER_VERSION CCISS_DRIVER_VERSION(3, 6, 20)

/* Embedded module documentation macros - see modules.h */
MODULE_AUTHOR("Hewlett-Packard Company");
MODULE_DESCRIPTION("Driver for HP Smart Array Controllers");
MODULE_SUPPORTED_DEVICE("HP SA5i SA5i+ SA532 SA5300 SA5312 SA641 SA642 SA6400"
                        " SA6i P600 P800 P400 P400i E200 E200i E500 P700m"
                        " Smart Array G2 Series SAS/SATA Controllers");
MODULE_VERSION("3.6.20");
MODULE_LICENSE("GPL");

#include "cciss_cmd.h"
#include "cciss.h"
#include <linux/cciss_ioctl.h>

/* define the PCI info for the cards we can control */
static const struct pci_device_id cciss_pci_device_id[] = {
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISS,  0x0E11, 0x4070},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSB, 0x0E11, 0x4080},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSB, 0x0E11, 0x4082},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSB, 0x0E11, 0x4083},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSC, 0x0E11, 0x4091},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSC, 0x0E11, 0x409A},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSC, 0x0E11, 0x409B},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSC, 0x0E11, 0x409C},
        {PCI_VENDOR_ID_COMPAQ, PCI_DEVICE_ID_COMPAQ_CISSC, 0x0E11, 0x409D},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSA,     0x103C, 0x3225},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSC,     0x103C, 0x3223},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSC,     0x103C, 0x3234},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSC,     0x103C, 0x3235},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSD,     0x103C, 0x3211},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSD,     0x103C, 0x3212},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSD,     0x103C, 0x3213},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSD,     0x103C, 0x3214},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSD,     0x103C, 0x3215},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSC,     0x103C, 0x3237},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSC,     0x103C, 0x323D},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSE,     0x103C, 0x3241},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSE,     0x103C, 0x3243},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSE,     0x103C, 0x3245},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSE,     0x103C, 0x3247},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSE,     0x103C, 0x3249},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSE,     0x103C, 0x324A},
        {PCI_VENDOR_ID_HP,     PCI_DEVICE_ID_HP_CISSE,     0x103C, 0x324B},
        {PCI_VENDOR_ID_HP,     PCI_ANY_ID,      PCI_ANY_ID, PCI_ANY_ID,
                PCI_CLASS_STORAGE_RAID << 8, 0xffff << 8, 0},
        {0,}
};

MODULE_DEVICE_TABLE(pci, cciss_pci_device_id);

/*  board_id = Subsystem Device ID & Vendor ID
 *  product = Marketing Name for the board
 *  access = Address of the struct of function pointers
 */
static struct board_type products[] = {
        {0x40700E11, "Smart Array 5300", &SA5_access},
        {0x40800E11, "Smart Array 5i", &SA5B_access},
        {0x40820E11, "Smart Array 532", &SA5B_access},
        {0x40830E11, "Smart Array 5312", &SA5B_access},
        {0x409A0E11, "Smart Array 641", &SA5_access},
        {0x409B0E11, "Smart Array 642", &SA5_access},
        {0x409C0E11, "Smart Array 6400", &SA5_access},
        {0x409D0E11, "Smart Array 6400 EM", &SA5_access},
        {0x40910E11, "Smart Array 6i", &SA5_access},
        {0x3225103C, "Smart Array P600", &SA5_access},
        {0x3223103C, "Smart Array P800", &SA5_access},
        {0x3234103C, "Smart Array P400", &SA5_access},
        {0x3235103C, "Smart Array P400i", &SA5_access},
        {0x3211103C, "Smart Array E200i", &SA5_access},
        {0x3212103C, "Smart Array E200", &SA5_access},
        {0x3213103C, "Smart Array E200i", &SA5_access},
        {0x3214103C, "Smart Array E200i", &SA5_access},
        {0x3215103C, "Smart Array E200i", &SA5_access},
        {0x3237103C, "Smart Array E500", &SA5_access},
        {0x323D103C, "Smart Array P700m", &SA5_access},
        {0x3241103C, "Smart Array P212", &SA5_access},
        {0x3243103C, "Smart Array P410", &SA5_access},
        {0x3245103C, "Smart Array P410i", &SA5_access},
        {0x3247103C, "Smart Array P411", &SA5_access},
        {0x3249103C, "Smart Array P812", &SA5_access},
        {0x324A103C, "Smart Array P712m", &SA5_access},
        {0x324B103C, "Smart Array P711m", &SA5_access},
        {0xFFFF103C, "Unknown Smart Array", &SA5_access},
};

/* How long to wait (in milliseconds) for board to go into simple mode */
#define MAX_CONFIG_WAIT 30000
#define MAX_IOCTL_CONFIG_WAIT 1000

/*define how many times we will try a command because of bus resets */
#define MAX_CMD_RETRIES 3

#define MAX_CTLR        32

/* Originally cciss driver only supports 8 major numbers */
#define MAX_CTLR_ORIG   8

static ctlr_info_t *hba[MAX_CTLR];

static void do_cciss_request(struct request_queue *q);
static irqreturn_t do_cciss_intr(int irq, void *dev_id);
static int cciss_open(struct block_device *bdev, fmode_t mode);
static int cciss_release(struct gendisk *disk, fmode_t mode);
static int cciss_ioctl(struct block_device *bdev, fmode_t mode,
                       unsigned int cmd, unsigned long arg);
static int cciss_getgeo(struct block_device *bdev, struct hd_geometry *geo);

static int cciss_revalidate(struct gendisk *disk);
static int rebuild_lun_table(ctlr_info_t *h, int first_time);
static int deregister_disk(ctlr_info_t *h, int drv_index,
                           int clear_all);

static void cciss_read_capacity(int ctlr, int logvol, int withirq,
                        sector_t *total_size, unsigned int *block_size);
static void cciss_read_capacity_16(int ctlr, int logvol, int withirq,
                        sector_t *total_size, unsigned int *block_size);
static void cciss_geometry_inquiry(int ctlr, int logvol,
                        int withirq, sector_t total_size,
                        unsigned int block_size, InquiryData_struct *inq_buff,
                                   drive_info_struct *drv);
static void __devinit cciss_interrupt_mode(ctlr_info_t *, struct pci_dev *,
                                           __u32);
static void start_io(ctlr_info_t *h);
static int sendcmd(__u8 cmd, int ctlr, void *buff, size_t size,
                   unsigned int use_unit_num, unsigned int log_unit,
                   __u8 page_code, unsigned char *scsi3addr, int cmd_type);
static int sendcmd_withirq(__u8 cmd, int ctlr, void *buff, size_t size,
                           unsigned int use_unit_num, unsigned int log_unit,
                           __u8 page_code, int cmd_type);

static void fail_all_cmds(unsigned long ctlr);
static int scan_thread(void *data);
static int check_for_unit_attention(ctlr_info_t *h, CommandList_struct *c);

#ifdef CONFIG_PROC_FS
static void cciss_procinit(int i);
#else
static void cciss_procinit(int i)
{
}
#endif                          /* CONFIG_PROC_FS */

#ifdef CONFIG_COMPAT
static int cciss_compat_ioctl(struct block_device *, fmode_t,
                              unsigned, unsigned long);
#endif

static struct block_device_operations cciss_fops = {
        .owner = THIS_MODULE,
        .open = cciss_open,
        .release = cciss_release,
        .locked_ioctl = cciss_ioctl,
        .getgeo = cciss_getgeo,
#ifdef CONFIG_COMPAT
        .compat_ioctl = cciss_compat_ioctl,
#endif
        .revalidate_disk = cciss_revalidate,
};

/*
 * Enqueuing and dequeuing functions for cmdlists.
 */
static inline void addQ(struct hlist_head *list, CommandList_struct *c)
{
        hlist_add_head(&c->list, list);
}

static inline void removeQ(CommandList_struct *c)
{
        if (WARN_ON(hlist_unhashed(&c->list)))
                return;

        hlist_del_init(&c->list);
}

#include "cciss_scsi.c"         /* For SCSI tape support */

#define RAID_UNKNOWN 6

#ifdef CONFIG_PROC_FS

/*
 * Report information about this controller.
 */
#define ENG_GIG 1000000000
#define ENG_GIG_FACTOR (ENG_GIG/512)
#define ENGAGE_SCSI     "engage scsi"
static const char *raid_label[] = { "0", "4", "1(1+0)", "5", "5+1", "ADG",
        "UNKNOWN"
};

static struct proc_dir_entry *proc_cciss;

static void cciss_seq_show_header(struct seq_file *seq)
{
        ctlr_info_t *h = seq->private;

        seq_printf(seq, "%s: HP %s Controller\n"
                "Board ID: 0x%08lx\n"
                "Firmware Version: %c%c%c%c\n"
                "IRQ: %d\n"
                "Logical drives: %d\n"
                "Current Q depth: %d\n"
                "Current # commands on controller: %d\n"
                "Max Q depth since init: %d\n"
                "Max # commands on controller since init: %d\n"
                "Max SG entries since init: %d\n",
                h->devname,
                h->product_name,
                (unsigned long)h->board_id,
                h->firm_ver[0], h->firm_ver[1], h->firm_ver[2],
                h->firm_ver[3], (unsigned int)h->intr[SIMPLE_MODE_INT],
                h->num_luns,
                h->Qdepth, h->commands_outstanding,
                h->maxQsinceinit, h->max_outstanding, h->maxSG);

#ifdef CONFIG_CISS_SCSI_TAPE
        cciss_seq_tape_report(seq, h->ctlr);
#endif /* CONFIG_CISS_SCSI_TAPE */
}

static void *cciss_seq_start(struct seq_file *seq, loff_t *pos)
{
        ctlr_info_t *h = seq->private;
        unsigned ctlr = h->ctlr;
        unsigned long flags;

        /* prevent displaying bogus info during configuration
         * or deconfiguration of a logical volume
         */
        spin_lock_irqsave(CCISS_LOCK(ctlr), flags);
        if (h->busy_configuring) {
                spin_unlock_irqrestore(CCISS_LOCK(ctlr), flags);
                return ERR_PTR(-EBUSY);
        }
        h->busy_configuring = 1;
        spin_unlock_irqrestore(CCISS_LOCK(ctlr), flags);

        if (*pos == 0)
                cciss_seq_show_header(seq);

        return pos;
}

static int cciss_seq_show(struct seq_file *seq, void *v)
{
        sector_t vol_sz, vol_sz_frac;
        ctlr_info_t *h = seq->private;
        unsigned ctlr = h->ctlr;
        loff_t *pos = v;
        drive_info_struct *drv = &h->drv[*pos];

        if (*pos > h->highest_lun)
                return 0;

        if (drv->heads == 0)
                return 0;

        vol_sz = drv->nr_blocks;
        vol_sz_frac = sector_div(vol_sz, ENG_GIG_FACTOR);
        vol_sz_frac *= 100;
        sector_div(vol_sz_frac, ENG_GIG_FACTOR);

        if (drv->raid_level > 5)
                drv->raid_level = RAID_UNKNOWN;
        seq_printf(seq, "cciss/c%dd%d:"
                        "\t%4u.%02uGB\tRAID %s\n",
                        ctlr, (int) *pos, (int)vol_sz, (int)vol_sz_frac,
                        raid_label[drv->raid_level]);
        return 0;
}

static void *cciss_seq_next(struct seq_file *seq, void *v, loff_t *pos)
{
        ctlr_info_t *h = seq->private;

        if (*pos > h->highest_lun)
                return NULL;
        *pos += 1;

        return pos;
}

static void cciss_seq_stop(struct seq_file *seq, void *v)
{
        ctlr_info_t *h = seq->private;

        /* Only reset h->busy_configuring if we succeeded in setting
         * it during cciss_seq_start. */
        if (v == ERR_PTR(-EBUSY))
                return;

        h->busy_configuring = 0;
}

static struct seq_operations cciss_seq_ops = {
        .start = cciss_seq_start,
        .show  = cciss_seq_show,
        .next  = cciss_seq_next,
        .stop  = cciss_seq_stop,
};

static int cciss_seq_open(struct inode *inode, struct file *file)
{
        int ret = seq_open(file, &cciss_seq_ops);
        struct seq_file *seq = file->private_data;

        if (!ret)
                seq->private = PDE(inode)->data;

        return ret;
}

static ssize_t
cciss_proc_write(struct file *file, const char __user *buf,
                 size_t length, loff_t *ppos)
{
        int err;
        char *buffer;

#ifndef CONFIG_CISS_SCSI_TAPE
        return -EINVAL;
#endif

        if (!buf || length > PAGE_SIZE - 1)
                return -EINVAL;

        buffer = (char *)__get_free_page(GFP_KERNEL);
        if (!buffer)
                return -ENOMEM;

        err = -EFAULT;
        if (copy_from_user(buffer, buf, length))
                goto out;
        buffer[length] = '\0';

#ifdef CONFIG_CISS_SCSI_TAPE
        if (strncmp(ENGAGE_SCSI, buffer, sizeof ENGAGE_SCSI - 1) == 0) {
                struct seq_file *seq = file->private_data;
                ctlr_info_t *h = seq->private;
                int rc;

                rc = cciss_engage_scsi(h->ctlr);
                if (rc != 0)
                        err = -rc;
                else
                        err = length;
        } else
#endif /* CONFIG_CISS_SCSI_TAPE */
                err = -EINVAL;
        /* might be nice to have "disengage" too, but it's not
           safely possible. (only 1 module use count, lock issues.) */

out:
        free_page((unsigned long)buffer);
        return err;
}

static struct file_operations cciss_proc_fops = {
        .owner   = THIS_MODULE,
        .open    = cciss_seq_open,
        .read    = seq_read,
        .llseek  = seq_lseek,
        .release = seq_release,
        .write   = cciss_proc_write,
};

static void __devinit cciss_procinit(int i)
{
        struct proc_dir_entry *pde;

        if (proc_cciss == NULL)
                proc_cciss = proc_mkdir("driver/cciss", NULL);
        if (!proc_cciss)
                return;
        pde = proc_create_data(hba[i]->devname, S_IWUSR | S_IRUSR | S_IRGRP |
                                        S_IROTH, proc_cciss,
                                        &cciss_proc_fops, hba[i]);
}
#endif                          /* CONFIG_PROC_FS */

/*
 * For operations that cannot sleep, a command block is allocated at init,
 * and managed by cmd_alloc() and cmd_free() using a simple bitmap to track
 * which ones are free or in use.  For operations that can wait for kmalloc
 * to possible sleep, this routine can be called with get_from_pool set to 0.
 * cmd_free() MUST be called with a got_from_pool set to 0 if cmd_alloc was.
 */
static CommandList_struct *cmd_alloc(ctlr_info_t *h, int get_from_pool)
{
        CommandList_struct *c;
        int i;
        u64bit temp64;
        dma_addr_t cmd_dma_handle, err_dma_handle;

        if (!get_from_pool) {
                c = (CommandList_struct *) pci_alloc_consistent(h->pdev,
                        sizeof(CommandList_struct), &cmd_dma_handle);
                if (c == NULL)
                        return NULL;
                memset(c, 0, sizeof(CommandList_struct));

                c->cmdindex = -1;

                c->err_info = (ErrorInfo_struct *)
                    pci_alloc_consistent(h->pdev, sizeof(ErrorInfo_struct),
                            &err_dma_handle);

                if (c->err_info == NULL) {
                        pci_free_consistent(h->pdev,
                                sizeof(CommandList_struct), c, cmd_dma_handle);
                        return NULL;
                }
                memset(c->err_info, 0, sizeof(ErrorInfo_struct));
        } else {                /* get it out of the controllers pool */

                do {
                        i = find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds);
                        if (i == h->nr_cmds)
                                return NULL;
                } while (test_and_set_bit
                         (i & (BITS_PER_LONG - 1),
                          h->cmd_pool_bits + (i / BITS_PER_LONG)) != 0);
#ifdef CCISS_DEBUG
                printk(KERN_DEBUG "cciss: using command buffer %d\n", i);
#endif
                c = h->cmd_pool + i;
                memset(c, 0, sizeof(CommandList_struct));
                cmd_dma_handle = h->cmd_pool_dhandle
                    + i * sizeof(CommandList_struct);
                c->err_info = h->errinfo_pool + i;
                memset(c->err_info, 0, sizeof(ErrorInfo_struct));
                err_dma_handle = h->errinfo_pool_dhandle
                    + i * sizeof(ErrorInfo_struct);
                h->nr_allocs++;

                c->cmdindex = i;
        }

        INIT_HLIST_NODE(&c->list);
        c->busaddr = (__u32) cmd_dma_handle;
        temp64.val = (__u64) err_dma_handle;
        c->ErrDesc.Addr.lower = temp64.val32.lower;
        c->ErrDesc.Addr.upper = temp64.val32.upper;
        c->ErrDesc.Len = sizeof(ErrorInfo_struct);

        c->ctlr = h->ctlr;
        return c;
}

/*
 * Frees a command block that was previously allocated with cmd_alloc().
 */
static void cmd_free(ctlr_info_t *h, CommandList_struct *c, int got_from_pool)
{
        int i;
        u64bit temp64;

        if (!got_from_pool) {
                temp64.val32.lower = c->ErrDesc.Addr.lower;
                temp64.val32.upper = c->ErrDesc.Addr.upper;
                pci_free_consistent(h->pdev, sizeof(ErrorInfo_struct),
                                    c->err_info, (dma_addr_t) temp64.val);
                pci_free_consistent(h->pdev, sizeof(CommandList_struct),
                                    c, (dma_addr_t) c->busaddr);
        } else {
                i = c - h->cmd_pool;
                clear_bit(i & (BITS_PER_LONG - 1),
                          h->cmd_pool_bits + (i / BITS_PER_LONG));
                h->nr_frees++;
        }
}

static inline ctlr_info_t *get_host(struct gendisk *disk)
{
        return disk->queue->queuedata;
}

static inline drive_info_struct *get_drv(struct gendisk *disk)
{
        return disk->private_data;
}

/*
 * Open.  Make sure the device is really there.
 */
static int cciss_open(struct block_device *bdev, fmode_t mode)
{
        ctlr_info_t *host = get_host(bdev->bd_disk);
        drive_info_struct *drv = get_drv(bdev->bd_disk);

#ifdef CCISS_DEBUG
        printk(KERN_DEBUG "cciss_open %s\n", bdev->bd_disk->disk_name);
#endif                          /* CCISS_DEBUG */

        if (host->busy_initializing || drv->busy_configuring)
                return -EBUSY;
        /*
         * Root is allowed to open raw volume zero even if it's not configured
         * so array config can still work. Root is also allowed to open any
         * volume that has a LUN ID, so it can issue IOCTL to reread the
         * disk information.  I don't think I really like this
         * but I'm already using way to many device nodes to claim another one
         * for "raw controller".
         */
        if (drv->heads == 0) {
                if (MINOR(bdev->bd_dev) != 0) { /* not node 0? */
                        /* if not node 0 make sure it is a partition = 0 */
                        if (MINOR(bdev->bd_dev) & 0x0f) {
                                return -ENXIO;
                                /* if it is, make sure we have a LUN ID */
                        } else if (drv->LunID == 0) {
                                return -ENXIO;
                        }
                }
                if (!capable(CAP_SYS_ADMIN))
                        return -EPERM;
        }
        drv->usage_count++;
        host->usage_count++;
        return 0;
}

/*
 * Close.  Sync first.
 */
static int cciss_release(struct gendisk *disk, fmode_t mode)
{
        ctlr_info_t *host = get_host(disk);
        drive_info_struct *drv = get_drv(disk);

#ifdef CCISS_DEBUG
        printk(KERN_DEBUG "cciss_release %s\n", disk->disk_name);
#endif                          /* CCISS_DEBUG */

        drv->usage_count--;
        host->usage_count--;
        return 0;
}

#ifdef CONFIG_COMPAT

static int do_ioctl(struct block_device *bdev, fmode_t mode,
                    unsigned cmd, unsigned long arg)
{
        int ret;
        lock_kernel();
        ret = cciss_ioctl(bdev, mode, cmd, arg);
        unlock_kernel();
        return ret;
}

static int cciss_ioctl32_passthru(struct block_device *bdev, fmode_t mode,
                                  unsigned cmd, unsigned long arg);
static int cciss_ioctl32_big_passthru(struct block_device *bdev, fmode_t mode,
                                      unsigned cmd, unsigned long arg);

static int cciss_compat_ioctl(struct block_device *bdev, fmode_t mode,
                              unsigned cmd, unsigned long arg)
{
        switch (cmd) {
        case CCISS_GETPCIINFO:
        case CCISS_GETINTINFO:
        case CCISS_SETINTINFO:
        case CCISS_GETNODENAME:
        case CCISS_SETNODENAME:
        case CCISS_GETHEARTBEAT:
        case CCISS_GETBUSTYPES:
        case CCISS_GETFIRMVER:
        case CCISS_GETDRIVVER:
        case CCISS_REVALIDVOLS:
        case CCISS_DEREGDISK:
        case CCISS_REGNEWDISK:
        case CCISS_REGNEWD:
        case CCISS_RESCANDISK:
        case CCISS_GETLUNINFO:
                return do_ioctl(bdev, mode, cmd, arg);

        case CCISS_PASSTHRU32:
                return cciss_ioctl32_passthru(bdev, mode, cmd, arg);
        case CCISS_BIG_PASSTHRU32:
                return cciss_ioctl32_big_passthru(bdev, mode, cmd, arg);

        default:
                return -ENOIOCTLCMD;
        }
}

static int cciss_ioctl32_passthru(struct block_device *bdev, fmode_t mode,
                                  unsigned cmd, unsigned long arg)
{
        IOCTL32_Command_struct __user *arg32 =
            (IOCTL32_Command_struct __user *) arg;
        IOCTL_Command_struct arg64;
        IOCTL_Command_struct __user *p = compat_alloc_user_space(sizeof(arg64));
        int err;
        u32 cp;

        err = 0;
        err |=
            copy_from_user(&arg64.LUN_info, &arg32->LUN_info,
                           sizeof(arg64.LUN_info));
        err |=
            copy_from_user(&arg64.Request, &arg32->Request,
                           sizeof(arg64.Request));
        err |=
            copy_from_user(&arg64.error_info, &arg32->error_info,
                           sizeof(arg64.error_info));
        err |= get_user(arg64.buf_size, &arg32->buf_size);
        err |= get_user(cp, &arg32->buf);
        arg64.buf = compat_ptr(cp);
        err |= copy_to_user(p, &arg64, sizeof(arg64));

        if (err)
                return -EFAULT;

        err = do_ioctl(bdev, mode, CCISS_PASSTHRU, (unsigned long)p);
        if (err)
                return err;
        err |=
            copy_in_user(&arg32->error_info, &p->error_info,
                         sizeof(arg32->error_info));
        if (err)
                return -EFAULT;
        return err;
}

static int cciss_ioctl32_big_passthru(struct block_device *bdev, fmode_t mode,
                                      unsigned cmd, unsigned long arg)
{
        BIG_IOCTL32_Command_struct __user *arg32 =
            (BIG_IOCTL32_Command_struct __user *) arg;
        BIG_IOCTL_Command_struct arg64;
        BIG_IOCTL_Command_struct __user *p =
            compat_alloc_user_space(sizeof(arg64));
        int err;
        u32 cp;

        err = 0;
        err |=
            copy_from_user(&arg64.LUN_info, &arg32->LUN_info,
                           sizeof(arg64.LUN_info));
        err |=
            copy_from_user(&arg64.Request, &arg32->Request,
                           sizeof(arg64.Request));
        err |=
            copy_from_user(&arg64.error_info, &arg32->error_info,
                           sizeof(arg64.error_info));
        err |= get_user(arg64.buf_size, &arg32->buf_size);
        err |= get_user(arg64.malloc_size, &arg32->malloc_size);
        err |= get_user(cp, &arg32->buf);
        arg64.buf = compat_ptr(cp);
        err |= copy_to_user(p, &arg64, sizeof(arg64));

        if (err)
                return -EFAULT;

        err = do_ioctl(bdev, mode, CCISS_BIG_PASSTHRU, (unsigned long)p);
        if (err)
                return err;
        err |=
            copy_in_user(&arg32->error_info, &p->error_info,
                         sizeof(arg32->error_info));
        if (err)
                return -EFAULT;
        return err;
}
#endif

static int cciss_getgeo(struct block_device *bdev, struct hd_geometry *geo)
{
        drive_info_struct *drv = get_drv(bdev->bd_disk);

        if (!drv->cylinders)
                return -ENXIO;

        geo->heads = drv->heads;
        geo->sectors = drv->sectors;
        geo->cylinders = drv->cylinders;
        return 0;
}

static void check_ioctl_unit_attention(ctlr_info_t *host, CommandList_struct *c)
{
        if (c->err_info->CommandStatus == CMD_TARGET_STATUS &&
                        c->err_info->ScsiStatus != SAM_STAT_CHECK_CONDITION)
                (void)check_for_unit_attention(host, c);
}
/*
 * ioctl
 */
static int cciss_ioctl(struct block_device *bdev, fmode_t mode,
                       unsigned int cmd, unsigned long arg)
{
        struct gendisk *disk = bdev->bd_disk;
        ctlr_info_t *host = get_host(disk);
        drive_info_struct *drv = get_drv(disk);
        int ctlr = host->ctlr;
        void __user *argp = (void __user *)arg;

#ifdef CCISS_DEBUG
        printk(KERN_DEBUG "cciss_ioctl: Called with cmd=%x %lx\n", cmd, arg);
#endif                          /* CCISS_DEBUG */

        switch (cmd) {
        case CCISS_GETPCIINFO:
                {
                        cciss_pci_info_struct pciinfo;

                        if (!arg)
                                return -EINVAL;
                        pciinfo.domain = pci_domain_nr(host->pdev->bus);
                        pciinfo.bus = host->pdev->bus->number;
                        pciinfo.dev_fn = host->pdev->devfn;
                        pciinfo.board_id = host->board_id;
                        if (copy_to_user
                            (argp, &pciinfo, sizeof(cciss_pci_info_struct)))
                                return -EFAULT;
                        return 0;
                }
        case CCISS_GETINTINFO:
                {
                        cciss_coalint_struct intinfo;
                        if (!arg)
                                return -EINVAL;
                        intinfo.delay =
                            readl(&host->cfgtable->HostWrite.CoalIntDelay);
                        intinfo.count =
                            readl(&host->cfgtable->HostWrite.CoalIntCount);
                        if (copy_to_user
                            (argp, &intinfo, sizeof(cciss_coalint_struct)))
                                return -EFAULT;
                        return 0;
                }
        case CCISS_SETINTINFO:
                {
                        cciss_coalint_struct intinfo;
                        unsigned long flags;
                        int i;

                        if (!arg)
                                return -EINVAL;
                        if (!capable(CAP_SYS_ADMIN))
                                return -EPERM;
                        if (copy_from_user
                            (&intinfo, argp, sizeof(cciss_coalint_struct)))
                                return -EFAULT;
                        if ((intinfo.delay == 0) && (intinfo.count == 0))
                        {
//                      printk("cciss_ioctl: delay and count cannot be 0\n");
                                return -EINVAL;
                        }
                        spin_lock_irqsave(CCISS_LOCK(ctlr), flags);
                        /* Update the field, and then ring the doorbell */
                        writel(intinfo.delay,
                               &(host->cfgtable->HostWrite.CoalIntDelay));
                        writel(intinfo.count,
                               &(host->cfgtable->HostWrite.CoalIntCount));
                        writel(CFGTBL_ChangeReq, host->vaddr + SA5_DOORBELL);

                        for (i = 0; i < MAX_IOCTL_CONFIG_WAIT; i++) {
                                if (!(readl(host->vaddr + SA5_DOORBELL)
                                      & CFGTBL_ChangeReq))
                                        break;
                                /* delay and try again */
                                udelay(1000);
                        }
                        spin_unlock_irqrestore(CCISS_LOCK(ctlr), flags);
                        if (i >= MAX_IOCTL_CONFIG_WAIT)
                                return -EAGAIN;
                        return 0;
                }
        case CCISS_GETNODENAME:
                {
                        NodeName_type NodeName;
                        int i;

                        if (!arg)
                                return -EINVAL;
                        for (i = 0; i < 16; i++)
                                NodeName[i] =
                                    readb(&host->cfgtable->ServerName[i]);
                        if (copy_to_user(argp, NodeName, sizeof(NodeName_type)))
                                return -EFAULT;
                        return 0;
                }
        case CCISS_SETNODENAME:
                {
                        NodeName_type NodeName;
                        unsigned long flags;
                        int i;

                        if (!arg)
                                return -EINVAL;
                        if (!capable(CAP_SYS_ADMIN))
                                return -EPERM;

                        if (copy_from_user
                            (NodeName, argp, sizeof(NodeName_type)))
                                return -EFAULT;

                        spin_lock_irqsave(CCISS_LOCK(ctlr), flags);

                        /* Update the field, and then ring the doorbell */
                        for (i = 0; i < 16; i++)
                                writeb(NodeName[i],
                                       &host->cfgtable->ServerName[i]);

                        writel(CFGTBL_ChangeReq, host->vaddr + SA5_DOORBELL);

                        for (i = 0; i < MAX_IOCTL_CONFIG_WAIT; i++) {
                                if (!(readl(host->vaddr + SA5_DOORBELL)
                                      & CFGTBL_ChangeReq))
                                        break;
                                /* delay and try again */
                                udelay(1000);
                        }
                        spin_unlock_irqrestore(CCISS_LOCK(ctlr), flags);
                        if (i >= MAX_IOCTL_CONFIG_WAIT)
                                return -EAGAIN;
                        return 0;
                }

        case CCISS_GETHEARTBEAT:
                {
                        Heartbeat_type heartbeat;

                        if (!arg)
                                return -EINVAL;
                        heartbeat = readl(&host->cfgtable->HeartBeat);
                        if (copy_to_user
                            (argp, &heartbeat, sizeof(Heartbeat_type)))
                                return -EFAULT;
                        return 0;
                }
        case CCISS_GETBUSTYPES:
                {
                        BusTypes_type BusTypes;

                        if (!arg)
                                return -EINVAL;
                        BusTypes = readl(&host->cfgtable->BusTypes);
                        if (copy_to_user
                            (argp, &BusTypes, sizeof(BusTypes_type)))
                                return -EFAULT;
                        return 0;
                }
        case CCISS_GETFIRMVER:
                {
                        FirmwareVer_type firmware;

                        if (!arg)
                                return -EINVAL;
                        memcpy(firmware, host->firm_ver, 4);

                        if (copy_to_user
                            (argp, firmware, sizeof(FirmwareVer_type)))
                                return -EFAULT;
                        return 0;
                }
        case CCISS_GETDRIVVER:
                {
                        DriverVer_type DriverVer = DRIVER_VERSION;

                        if (!arg)
                                return -EINVAL;

                        if (copy_to_user
                            (argp, &DriverVer, sizeof(DriverVer_type)))
                                return -EFAULT;
                        return 0;
                }

        case CCISS_DEREGDISK:
        case CCISS_REGNEWD:
        case CCISS_REVALIDVOLS:
                return rebuild_lun_table(host, 0);

        case CCISS_GETLUNINFO:{
                        LogvolInfo_struct luninfo;

                        luninfo.LunID = drv->LunID;
                        luninfo.num_opens = drv->usage_count;
                        luninfo.num_parts = 0;
                        if (copy_to_user(argp, &luninfo,
                                         sizeof(LogvolInfo_struct)))
                                return -EFAULT;
                        return 0;
                }
        case CCISS_PASSTHRU:
                {
                        IOCTL_Command_struct iocommand;
                        CommandList_struct *c;
                        char *buff = NULL;
                        u64bit temp64;
                        unsigned long flags;
                        DECLARE_COMPLETION_ONSTACK(wait);

                        if (!arg)
                                return -EINVAL;

                        if (!capable(CAP_SYS_RAWIO))
                                return -EPERM;

                        if (copy_from_user
                            (&iocommand, argp, sizeof(IOCTL_Command_struct)))
                                return -EFAULT;
                        if ((iocommand.buf_size < 1) &&
                            (iocommand.Request.Type.Direction != XFER_NONE)) {
                                return -EINVAL;
                        }
#if 0                           /* 'buf_size' member is 16-bits, and always smaller than kmalloc limit */
                        /* Check kmalloc limits */
                        if (iocommand.buf_size > 128000)
                                return -EINVAL;
#endif
                        if (iocommand.buf_size > 0) {
                                buff = kmalloc(iocommand.buf_size, GFP_KERNEL);
                                if (buff == NULL)
                                        return -EFAULT;
                        }
                        if (iocommand.Request.Type.Direction == XFER_WRITE) {
                                /* Copy the data into the buffer we created */
                                if (copy_from_user
                                    (buff, iocommand.buf, iocommand.buf_size)) {
                                        kfree(buff);
                                        return -EFAULT;
                                }
                        } else {
                                memset(buff, 0, iocommand.buf_size);
                        }
                        if ((c = cmd_alloc(host, 0)) == NULL) {
                                kfree(buff);
                                return -ENOMEM;
                        }
                        // Fill in the command type
                        c->cmd_type = CMD_IOCTL_PEND;
                        // Fill in Command Header
                        c->Header.ReplyQueue = 0;       // unused in simple mode
                        if (iocommand.buf_size > 0)     // buffer to fill
                        {
                                c->Header.SGList = 1;

                                c->Header.SGTotal = 1;
                        } else  // no buffers to fill
                        {
                                c->Header.SGList = 0;
                                c->Header.SGTotal = 0;
                        }
                        c->Header.LUN = iocommand.LUN_info;
                        c->Header.Tag.lower = c->busaddr;       // use the kernel address the cmd block for tag

                        // Fill in Request block
                        c->Request = iocommand.Request;

                        // Fill in the scatter gather information
                        if (iocommand.buf_size > 0) {
                                temp64.val = pci_map_single(host->pdev, buff,
                                        iocommand.buf_size,
                                        PCI_DMA_BIDIRECTIONAL);
                                c->SG[0].Addr.lower = temp64.val32.lower;
                                c->SG[0].Addr.upper = temp64.val32.upper;
                                c->SG[0].Len = iocommand.buf_size;
                                c->SG[0].Ext = 0;       // we are not chaining
                        }
                        c->waiting = &wait;

                        /* Put the request on the tail of the request queue */
                        spin_lock_irqsave(CCISS_LOCK(ctlr), flags);
                        addQ(&host->reqQ, c);
                        host->Qdepth++;
                        start_io(host);
                        spin_unlock_irqrestore(CCISS_LOCK(ctlr), flags);

                        wait_for_completion(&wait);

                        /* unlock the buffers from DMA */
                        temp64.val32.lower = c->SG[0].Addr.lower;
                        temp64.val32.upper = c->SG[0].Addr.upper;
                        pci_unmap_single(host->pdev, (dma_addr_t) temp64.val,
                                         iocommand.buf_size,
                                         PCI_DMA_BIDIRECTIONAL);

                        check_ioctl_unit_attention(host, c);

                        /* Copy the error information out */
                        iocommand.error_info = *(c->err_info);
                        if (copy_to_user
                            (argp, &iocommand, sizeof(IOCTL_Command_struct))) {
                                kfree(buff);
                                cmd_free(host, c, 0);
                                return -EFAULT;
                        }

                        if (iocommand.Request.Type.Direction == XFER_READ) {
                                /* Copy the data out of the buffer we created */
                                if (copy_to_user
                                    (iocommand.buf, buff, iocommand.buf_size)) {
                                        kfree(buff);
                                        cmd_free(host, c, 0);
                                        return -EFAULT;
                                }
                        }
                        kfree(buff);
                        cmd_free(host, c, 0);
                        return 0;
                }
        case CCISS_BIG_PASSTHRU:{
                        BIG_IOCTL_Command_struct *ioc;
                        CommandList_struct *c;
                        unsigned char **buff = NULL;
                        int *buff_size = NULL;
                        u64bit temp64;
                        unsigned long flags;
                        BYTE sg_used = 0;
                        int status = 0;
                        int i;
                        DECLARE_COMPLETION_ONSTACK(wait);
                        __u32 left;
                        __u32 sz;
                        BYTE __user *data_ptr;

                        if (!arg)
                                return -EINVAL;
                        if (!capable(CAP_SYS_RAWIO))
                                return -EPERM;
                        ioc = (BIG_IOCTL_Command_struct *)
                            kmalloc(sizeof(*ioc), GFP_KERNEL);
                        if (!ioc) {
                                status = -ENOMEM;
                                goto cleanup1;
                        }
                        if (copy_from_user(ioc, argp, sizeof(*ioc))) {
                                status = -EFAULT;
                                goto cleanup1;
                        }
                        if ((ioc->buf_size < 1) &&
                            (ioc->Request.Type.Direction != XFER_NONE)) {
                                status = -EINVAL;
                                goto cleanup1;
                        }
                        /* Check kmalloc limits  using all SGs */
                        if (ioc->malloc_size > MAX_KMALLOC_SIZE) {
                                status = -EINVAL;
                                goto cleanup1;
                        }
                        if (ioc->buf_size > ioc->malloc_size * MAXSGENTRIES) {
                                status = -EINVAL;
                                goto cleanup1;
                        }
                        buff =
                            kzalloc(MAXSGENTRIES * sizeof(char *), GFP_KERNEL);
                        if (!buff) {
                                status = -ENOMEM;
                                goto cleanup1;
                        }
                        buff_size = kmalloc(MAXSGENTRIES * sizeof(int),
                                                   GFP_KERNEL);
                        if (!buff_size) {
                                status = -ENOMEM;
                                goto cleanup1;
                        }
                        left = ioc->buf_size;
                        data_ptr = ioc->buf;
                        while (left) {
                                sz = (left >
                                      ioc->malloc_size) ? ioc->
                                    malloc_size : left;
                                buff_size[sg_used] = sz;
                                buff[sg_used] = kmalloc(sz, GFP_KERNEL);
                                if (buff[sg_used] == NULL) {
                                        status = -ENOMEM;
                                        goto cleanup1;
                                }
                                if (ioc->Request.Type.Direction == XFER_WRITE) {
                                        if (copy_from_user
                                            (buff[sg_used], data_ptr, sz)) {
                                                status = -EFAULT;
                                                goto cleanup1;
                                        }
                                } else {
                                        memset(buff[sg_used], 0, sz);
                                }
                                left -= sz;
                                data_ptr += sz;
                                sg_used++;
                        }
                        if ((c = cmd_alloc(host, 0)) == NULL) {
                                status = -ENOMEM;
                                goto cleanup1;
                        }
                        c->cmd_type = CMD_IOCTL_PEND;
                        c->Header.ReplyQueue = 0;

                        if (ioc->buf_size > 0) {
                                c->Header.SGList = sg_used;
                                c->Header.SGTotal = sg_used;
                        } else {
                                c->Header.SGList = 0;
                                c->Header.SGTotal = 0;
                        }
                        c->Header.LUN = ioc->LUN_info;
                        c->Header.Tag.lower = c->busaddr;

                        c->Request = ioc->Request;
                        if (ioc->buf_size > 0) {
                                int i;
                                for (i = 0; i < sg_used; i++) {
                                        temp64.val =
                                            pci_map_single(host->pdev, buff[i],
                                                    buff_size[i],
                                                    PCI_DMA_BIDIRECTIONAL);
                                        c->SG[i].Addr.lower =
                                            temp64.val32.lower;
                                        c->SG[i].Addr.upper =
                                            temp64.val32.upper;
                                        c->SG[i].Len = buff_size[i];
                                        c->SG[i].Ext = 0;       /* we are not chaining */
                                }
                        }
                        c->waiting = &wait;
                        /* Put the request on the tail of the request queue */
                        spin_lock_irqsave(CCISS_LOCK(ctlr), flags);
                        addQ(&host->reqQ, c);
                        host->Qdepth++;
                        start_io(host);
                        spin_unlock_irqrestore(CCISS_LOCK(ctlr), flags);
                        wait_for_completion(&wait);
                        /* unlock the buffers from DMA */
                        for (i = 0; i < sg_used; i++) {
                                temp64.val32.lower = c->SG[i].Addr.lower;
                                temp64.val32.upper = c->SG[i].Addr.upper;
                                pci_unmap_single(host->pdev,
                                        (dma_addr_t) temp64.val, buff_size[i],
                                        PCI_DMA_BIDIRECTIONAL);
                        }
                        check_ioctl_unit_attention(host, c);
                        /* Copy the error information out */
                        ioc->error_info = *(c->err_info);
                        if (copy_to_user(argp, ioc, sizeof(*ioc))) {
                                cmd_free(host, c, 0);
                                status = -EFAULT;
                                goto cleanup1;
                        }
                        if (ioc->Request.Type.Direction == XFER_READ) {
                                /* Copy the data out of the buffer we created */
                                BYTE __user *ptr = ioc->buf;
                                for (i = 0; i < sg_used; i++) {
                                        if (copy_to_user
                                            (ptr, buff[i], buff_size[i])) {
                                                cmd_free(host, c, 0);
                                                status = -EFAULT;
                                                goto cleanup1;
                                        }
                                        ptr += buff_size[i];
                                }
                        }
                        cmd_free(host, c, 0);
                        status = 0;
                      cleanup1:
                        if (buff) {
                                for (i = 0; i < sg_used; i++)
                                        kfree(buff[i]);
                                kfree(buff);
                        }
                        kfree(buff_size);
                        kfree(ioc);
                        return status;
                }

        /* scsi_cmd_ioctl handles these, below, though some are not */
        /* very meaningful for cciss.  SG_IO is the main one people want. */

        case SG_GET_VERSION_NUM:
        case SG_SET_TIMEOUT:
        case SG_GET_TIMEOUT:
        case SG_GET_RESERVED_SIZE:
        case SG_SET_RESERVED_SIZE:
        case SG_EMULATED_HOST:
        case SG_IO:
        case SCSI_IOCTL_SEND_COMMAND:
                return scsi_cmd_ioctl(disk->queue, disk, mode, cmd, argp);

        /* scsi_cmd_ioctl would normally handle these, below, but */
        /* they aren't a good fit for cciss, as CD-ROMs are */
        /* not supported, and we don't have any bus/target/lun */
        /* which we present to the kernel. */

        case CDROM_SEND_PACKET:
        case CDROMCLOSETRAY:
        case CDROMEJECT:
        case SCSI_IOCTL_GET_IDLUN:
        case SCSI_IOCTL_GET_BUS_NUMBER:
        default:
                return -ENOTTY;
        }
}

static void cciss_check_queues(ctlr_info_t *h)
{
        int start_queue = h->next_to_run;
        int i;

        /* check to see if we have maxed out the number of commands that can
         * be placed on the queue.  If so then exit.  We do this check here
         * in case the interrupt we serviced was from an ioctl and did not
         * free any new commands.
         */
        if ((find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds)) == h->nr_cmds)
                return;

        /* We have room on the queue for more commands.  Now we need to queue
         * them up.  We will also keep track of the next queue to run so
         * that every queue gets a chance to be started first.
         */
        for (i = 0; i < h->highest_lun + 1; i++) {
                int curr_queue = (start_queue + i) % (h->highest_lun + 1);
                /* make sure the disk has been added and the drive is real
                 * because this can be called from the middle of init_one.
                 */
                if (!(h->drv[curr_queue].queue) || !(h->drv[curr_queue].heads))
                        continue;
                blk_start_queue(h->gendisk[curr_queue]->queue);

                /* check to see if we have maxed out the number of commands
                 * that can be placed on the queue.
                 */
                if ((find_first_zero_bit(h->cmd_pool_bits, h->nr_cmds)) == h->nr_cmds) {
                        if (curr_queue == start_queue) {
                                h->next_to_run =
                                    (start_queue + 1) % (h->highest_lun + 1);
                                break;
                        } else {
                                h->next_to_run = curr_queue;
                                break;
                        }
                }
        }
}

static void cciss_softirq_done(struct request *rq)
{
        CommandList_struct *cmd = rq->completion_data;
        ctlr_info_t *h = hba[cmd->ctlr];
        unsigned int nr_bytes;
        unsigned long flags;
        u64bit temp64;
        int i, ddir;

        if (cmd->Request.Type.Direction == XFER_READ)
                ddir = PCI_DMA_FROMDEVICE;
        else
                ddir = PCI_DMA_TODEVICE;

        /* command did not need to be retried */
        /* unmap the DMA mapping for all the scatter gather elements */
        for (i = 0; i < cmd->Header.SGList; i++) {
                temp64.val32.lower = cmd->SG[i].Addr.lower;
                temp64.val32.upper = cmd->SG[i].Addr.upper;
                pci_unmap_page(h->pdev, temp64.val, cmd->SG[i].Len, ddir);
        }

#ifdef CCISS_DEBUG
        printk("Done with %p\n", rq);
#endif                          /* CCISS_DEBUG */

        /*
         * Store the full size and set the residual count for pc requests
         */
        nr_bytes = blk_rq_bytes(rq);
        if (blk_pc_request(rq))
                rq->data_len = cmd->err_info->ResidualCnt;

        if (blk_end_request(rq, (rq->errors == 0) ? 0 : -EIO, nr_bytes))
                BUG();

        spin_lock_irqsave(&h->lock, flags);
        cmd_free(h, cmd, 1);
        cciss_check_queues(h);
        spin_unlock_irqrestore(&h->lock, flags);
}

/* This function gets the serial number of a logical drive via
 * inquiry page 0x83.  Serial no. is 16 bytes.  If the serial
 * number cannot be had, for whatever reason, 16 bytes of 0xff
 * are returned instead.
 */
static void cciss_get_serial_no(int ctlr, int logvol, int withirq,
                                unsigned char *serial_no, int buflen)
{
#define PAGE_83_INQ_BYTES 64
        int rc;
        unsigned char *buf;

        if (buflen > 16)
                buflen = 16;
        memset(serial_no, 0xff, buflen);
        buf = kzalloc(PAGE_83_INQ_BYTES, GFP_KERNEL);
        if (!buf)
                return;
        memset(serial_no, 0, buflen);
        if (withirq)
                rc = sendcmd_withirq(CISS_INQUIRY, ctlr, buf,
                        PAGE_83_INQ_BYTES, 1, logvol, 0x83, TYPE_CMD);
        else
                rc = sendcmd(CISS_INQUIRY, ctlr, buf,
                        PAGE_83_INQ_BYTES, 1, logvol, 0x83, NULL, TYPE_CMD);
        if (rc == IO_OK)
                memcpy(serial_no, &buf[8], buflen);
        kfree(buf);
        return;
}

static void cciss_add_disk(ctlr_info_t *h, struct gendisk *disk,
                                int drv_index)
{
        disk->queue = blk_init_queue(do_cciss_request, &h->lock);
        sprintf(disk->disk_name, "cciss/c%dd%d", h->ctlr, drv_index);
        disk->major = h->major;
        disk->first_minor = drv_index << NWD_SHIFT;
        disk->fops = &cciss_fops;
        disk->private_data = &h->drv[drv_index];
        disk->driverfs_dev = &h->pdev->dev;

        /* Set up queue information */
        blk_queue_bounce_limit(disk->queue, h->pdev->dma_mask);

        /* This is a hardware imposed limit. */
        blk_queue_max_hw_segments(disk->queue, MAXSGENTRIES);

        /* This is a limit in the driver and could be eliminated. */
        blk_queue_max_phys_segments(disk->queue, MAXSGENTRIES);

        blk_queue_max_sectors(disk->queue, h->cciss_max_sectors);

        blk_queue_softirq_done(disk->queue, cciss_softirq_done);

        disk->queue->queuedata = h;

        blk_queue_hardsect_size(disk->queue,
                                h->drv[drv_index].block_size);

        /* Make sure all queue data is written out before */
        /* setting h->drv[drv_index].queue, as setting this */
        /* allows the interrupt handler to start the queue */
        wmb();
        h->drv[drv_index].queue = disk->queue;
        add_disk(disk);
}

/* This function will check the usage_count of the drive to be updated/added.
 * If the usage_count is zero and it is a heretofore unknown drive, or,
 * the drive's capacity, geometry, or serial number has changed,
 * then the drive information will be updated and the disk will be
 * re-registered with the kernel.  If these conditions don't hold,
 * then it will be left alone for the next reboot.  The exception to this
 * is disk 0 which will always be left registered with the kernel since it
 * is also the controller node.  Any changes to disk 0 will show up on
 * the next reboot.
 */
static void cciss_update_drive_info(int ctlr, int drv_index, int first_time)
{
        ctlr_info_t *h = hba[ctlr];
        struct gendisk *disk;
        InquiryData_struct *inq_buff = NULL;
        unsigned int block_size;
        sector_t total_size;
        unsigned long flags = 0;
        int ret = 0;
        drive_info_struct *drvinfo;
        int was_only_controller_node;

        /* Get information about the disk and modify the driver structure */
        inq_buff = kmalloc(sizeof(InquiryData_struct), GFP_KERNEL);
        drvinfo = kmalloc(sizeof(*drvinfo), GFP_KERNEL);
        if (inq_buff == NULL || drvinfo == NULL)
                goto mem_msg;

        /* See if we're trying to update the "controller node"
         * this will happen the when the first logical drive gets
         * created by ACU.
         */
        was_only_controller_node = (drv_index == 0 &&
                                h->drv[0].raid_level == -1);

        /* testing to see if 16-byte CDBs are already being used */
        if (h->cciss_read == CCISS_READ_16) {
                cciss_read_capacity_16(h->ctlr, drv_index, 1,
                        &total_size, &block_size);

        } else {
                cciss_read_capacity(ctlr, drv_index, 1,
                                    &total_size, &block_size);

                /* if read_capacity returns all F's this volume is >2TB */
                /* in size so we switch to 16-byte CDB's for all */
                /* read/write ops */
                if (total_size == 0xFFFFFFFFULL) {
                        cciss_read_capacity_16(ctlr, drv_index, 1,
                        &total_size, &block_size);
                        h->cciss_read = CCISS_READ_16;
                        h->cciss_write = CCISS_WRITE_16;
                } else {
                        h->cciss_read = CCISS_READ_10;
                        h->cciss_write = CCISS_WRITE_10;
                }
        }

        cciss_geometry_inquiry(ctlr, drv_index, 1, total_size, block_size,
                               inq_buff, drvinfo);
        drvinfo->block_size = block_size;
        drvinfo->nr_blocks = total_size + 1;

        cciss_get_serial_no(ctlr, drv_index, 1, drvinfo->serial_no,
                        sizeof(drvinfo->serial_no));

        /* Is it the same disk we already know, and nothing's changed? */
        if (h->drv[drv_index].raid_level != -1 &&
                ((memcmp(drvinfo->serial_no,
                                h->drv[drv_index].serial_no, 16) == 0) &&
                drvinfo->block_size == h->drv[drv_index].block_size &&
                drvinfo->nr_blocks == h->drv[drv_index].nr_blocks &&
                drvinfo->heads == h->drv[drv_index].heads &&
                drvinfo->sectors == h->drv[drv_index].sectors &&
                drvinfo->cylinders == h->drv[drv_index].cylinders))
                        /* The disk is unchanged, nothing to update */
                        goto freeret;

        /* If we get here it's not the same disk, or something's changed,
         * so we need to * deregister it, and re-register it, if it's not
         * in use.
         * If the disk already exists then deregister it before proceeding
         * (unless it's the first disk (for the controller node).
         */
        if (h->drv[drv_index].raid_level != -1 && drv_index != 0) {
                printk(KERN_WARNING "disk %d has changed.\n", drv_index);
                spin_lock_irqsave(CCISS_LOCK(h->ctlr), flags);
                h->drv[drv_index].busy_configuring = 1;
                spin_unlock_irqrestore(CCISS_LOCK(h->ctlr), flags);

                /* deregister_disk sets h->drv[drv_index].queue = NULL
                 * which keeps the interrupt handler from starting
                 * the queue.
                 */
                ret = deregister_disk(h, drv_index, 0);
                h->drv[drv_index].busy_configuring = 0;
        }

        /* If the disk is in use return */
        if (ret)
                goto freeret;

        /* Save the new information from cciss_geometry_inquiry
         * and serial number inquiry.
         */
        h->drv[drv_index].block_size = drvinfo->block_size;
        h->drv[drv_index].nr_blocks = drvinfo->nr_blocks;
        h->drv[drv_index].heads = drvinfo->heads;
        h->drv[drv_index].sectors = drvinfo->sectors;
        h->drv[drv_index].cylinders = drvinfo->cylinders;
        h->drv[drv_index].raid_level = drvinfo->raid_level;
        memcpy(h->drv[drv_index].serial_no, drvinfo->serial_no, 16);

        ++h->num_luns;
        disk = h->gendisk[drv_index];
        set_capacity(disk, h->drv[drv_index].nr_blocks);

        /* If it's not disk 0 (drv_index != 0)
         * or if it was disk 0, but there was previously
         * no actual corresponding configured logical drive
         * (raid_leve == -1) then we want to update the
         * logical drive's information.
         */
        if (drv_index || first_time)
                cciss_add_disk(h, disk, drv_index);

freeret:
        kfree(inq_buff);
        kfree(drvinfo);
        return;
mem_msg:
        printk(KERN_ERR "cciss: out of memory\n");
        goto freeret;
}

/* This function will find the first index of the controllers drive array
 * that has a -1 for the raid_level and will return that index.  This is
 * where new drives will be added.  If the index to be returned is greater
 * than the highest_lun index for the controller then highest_lun is set
 * to this new index.  If there are no available indexes then -1 is returned.
 * "controller_node" is used to know if this is a real logical drive, or just
 * the controller node, which determines if this counts towards highest_lun.
 */
static int cciss_find_free_drive_index(int ctlr, int controller_node)
{
        int i;

        for (i = 0; i < CISS_MAX_LUN; i++) {
                if (hba[ctlr]->drv[i].raid_level == -1) {
                        if (i > hba[ctlr]->highest_lun)
                                if (!controller_node)
                                        hba[ctlr]->highest_lun = i;
                        return i;
                }
        }
        return -1;
}

/* cciss_add_gendisk finds a free hba[]->drv structure
 * and allocates a gendisk if needed, and sets the lunid
 * in the drvinfo structure.   It returns the index into
 * the ->drv[] array, or -1 if none are free.
 * is_controller_node indicates whether highest_lun should
 * count this disk, or if it's only being added to provide
 * a means to talk to the controller in case no logical
 * drives have yet been configured.
 */
static int cciss_add_gendisk(ctlr_info_t *h, __u32 lunid, int controller_node)
{
        int drv_index;

        drv_index = cciss_find_free_drive_index(h->ctlr, controller_node);
        if (drv_index == -1)
                return -1;
        /*Check if the gendisk needs to be allocated */
        if (!h->gendisk[drv_index]) {
                h->gendisk[drv_index] =
                        alloc_disk(1 << NWD_SHIFT);
                if (!h->gendisk[drv_index]) {
                        printk(KERN_ERR "cciss%d: could not "
                                "allocate a new disk %d\n",
                                h->ctlr, drv_index);
                        return -1;
                }
        }
        h->drv[drv_index].LunID = lunid;

        /* Don't need to mark this busy because nobody */
        /* else knows about this disk yet to contend */
        /* for access to it. */
        h->drv[drv_index].busy_configuring = 0;
        wmb();
        return drv_index;
}

/* This is for the special case of a controller which
 * has no logical drives.  In this case, we still need
 * to register a disk so the controller can be accessed
 * by the Array Config Utility.
 */
static void cciss_add_controller_node(ctlr_info_t *h)
{
        struct gendisk *disk;
        int drv_index;

        if (h->gendisk[0] != NULL) /* already did this? Then bail. */
                return;

        drv_index = cciss_add_gendisk(h, 0, 1);
        if (drv_index == -1) {
                printk(KERN_WARNING "cciss%d: could not "
                        "add disk 0.\n", h->ctlr);
                return;
        }
        h->drv[drv_index].block_size = 512;
        h->drv[drv_index].nr_blocks = 0;
        h->drv[drv_index].heads = 0;
        h->drv[drv_index].sectors = 0;
        h->drv[drv_index].cylinders = 0;
        h->drv[drv_index].raid_level = -1;
        memset(h->drv[drv_index].serial_no, 0, 16);
        disk = h->gendisk[drv_index];
        cciss_add_disk(h, disk, drv_index);
}

/* This function will add and remove logical drives from the Logical
 * drive array of the controller and maintain persistency of ordering
 * so that mount points are preserved until the next reboot.  This allows
 * for the removal of logical drives in the middle of the drive array
 * without a re-ordering of those drives.
 * INPUT
 * h            = The controller to perform the operations on
 */
static int rebuild_lun_table(ctlr_info_t *h, int first_time)
{
        int ctlr = h->ctlr;
        int num_luns;
        ReportLunData_struct *ld_buff = NULL;
        int return_code;
        int listlength = 0;
        int i;
        int drv_found;
        int drv_index = 0;
        __u32 lunid = 0;
        unsigned long flags;

        if (!capable(CAP_SYS_RAWIO))
                return -EPERM;

        /* Set busy_configuring flag for this operation */
        spin_lock_irqsave(CCISS_LOCK(h->ctlr), flags);
        if (h->busy_configuring) {
                spin_unlock_irqrestore(CCISS_LOCK(h->ctlr), flags);
                return -EBUSY;
        }
        h->busy_configuring = 1;
        spin_unlock_irqrestore(CCISS_LOCK(h->ctlr), flags);

        ld_buff = kzalloc(sizeof(ReportLunData_struct), GFP_KERNEL);
        if (ld_buff == NULL)
                goto mem_msg;

        return_code = sendcmd_withirq(CISS_REPORT_LOG, ctlr, ld_buff,
                                      sizeof(ReportLunData_struct), 0,
                                      0, 0, TYPE_CMD);

        if (return_code == IO_OK)
                listlength = be32_to_cpu(*(__be32 *) ld_buff->LUNListLength);
        else {  /* reading number of logical volumes failed */
                printk(KERN_WARNING "cciss: report logical volume"
                       " command failed\n");
                listlength = 0;
                goto freeret;
        }

        num_luns = listlength / 8;      /* 8 bytes per entry */
        if (num_luns > CISS_MAX_LUN) {
                num_luns = CISS_MAX_LUN;
                printk(KERN_WARNING "cciss: more luns configured"
                       " on controller than can be handled by"
                       " this driver.\n");
        }

        if (num_luns == 0)
                cciss_add_controller_node(h);

        /* Compare controller drive array to driver's drive array
         * to see if any drives are missing on the controller due
         * to action of Array Config Utility (user deletes drive)
         * and deregister logical drives which have disappeared.
         */
        for (i = 0; i <= h->highest_lun; i++) {
                int j;
                drv_found = 0;

                /* skip holes in the array from already deleted drives */
                if (h->drv[i].raid_level == -1)
                        continue;

                for (j = 0; j < num_luns; j++) {
                        memcpy(&lunid, &ld_buff->LUN[j][0], 4);
                        lunid = le32_to_cpu(lunid);
                        if (h->drv[i].LunID == lunid) {
                                drv_found = 1;
                                break;
                        }
                }
                if (!drv_found) {
                        /* Deregister it from the OS, it's gone. */
                        spin_lock_irqsave(CCISS_LOCK(h->ctlr), flags);
                        h->drv[i].busy_configuring = 1;
                        spin_unlock_irqrestore(CCISS_LOCK(h->ctlr), flags);
                        return_code = deregister_disk(h, i, 1);
                        h->drv[i].busy_configuring = 0;
                }
        }

        /* Compare controller drive array to driver's drive array.
         * Check for updates in the drive information and any new drives
         * on the controller due to ACU adding logical drives, or changing
         * a logical drive's size, etc.  Reregister any new/changed drives
         */
        for (i = 0; i < num_luns; i++) {
                int j;

                drv_found = 0;

                memcpy(&lunid, &ld_buff->LUN[i][0], 4);
                lunid = le32_to_cpu(lunid);

                /* Find if the LUN is already in the drive array
                 * of the driver.  If so then update its info
                 * if not in use.  If it does not exist then find
                 * the first free index and add it.
                 */
                for (j = 0; j <= h->highest_lun; j++) {
                        if (h->drv[j].raid_level != -1 &&
                                h->drv[j].LunID == lunid) {
                                drv_index = j;
                                drv_found = 1;
                                break;
                        }
                }

                /* check if the drive was found already in the array */
                if (!drv_found) {
                        drv_index = cciss_add_gendisk(h, lunid, 0);
                        if (drv_index == -1)
                                goto freeret;
                }
                cciss_update_drive_info(ctlr, drv_index, first_time);
        }               /* end for */

freeret:
        kfree(ld_buff);
        h->busy_configuring = 0;
        /* We return -1 here to tell the ACU that we have registered/updated
         * all of the drives that we can and to keep it from calling us
         * additional times.
         */
        return -1;
mem_msg:
        printk(KERN_ERR "cciss: out of memory\n");
        h->busy_configuring = 0;
        goto freeret;
}

/* This function will deregister the disk and it's queue from the
 * kernel.  It must be called with the controller lock held and the
 * drv structures busy_configuring flag set.  It's parameters are:
 *
 * disk = This is the disk to be deregistered
 * drv  = This is the drive_info_struct associated with the disk to be
 *        deregistered.  It contains information about the disk used
 *        by the driver.
 * clear_all = This flag determines whether or not the disk information
 *             is going to be completely cleared out and the highest_lun
 *             reset.  Sometimes we want to clear out information about
 *             the disk in preparation for re-adding it.  In this case
 *             the highest_lun should be left unchanged and the LunID
 *             should not be cleared.
*/
static int deregister_disk(ctlr_info_t *h, int drv_index,
                           int clear_all)
{
        int i;
        struct gendisk *disk;
        drive_info_struct *drv;

        if (!capable(CAP_SYS_RAWIO))
                return -EPERM;

        drv = &h->drv[drv_index];
        disk = h->gendisk[drv_index];

        /* make sure logical volume is NOT is use */
        if (clear_all || (h->gendisk[0] == disk)) {
                if (drv->usage_count > 1)
                        return -EBUSY;
        } else if (drv->usage_count > 0)
                return -EBUSY;

        /* invalidate the devices and deregister the disk.  If it is disk
         * zero do not deregister it but just zero out it's values.  This
         * allows us to delete disk zero but keep the controller registered.
         */
        if (h->gendisk[0] != disk) {
                struct request_queue *q = disk->queue;
                if (disk->flags & GENHD_FL_UP)
                        del_gendisk(disk);
                if (q) {
                        blk_cleanup_queue(q);
                        /* Set drv->queue to NULL so that we do not try
                         * to call blk_start_queue on this queue in the
                         * interrupt handler
                         */
                        drv->queue = NULL;
                }
                /* If clear_all is set then we are deleting the logical
                 * drive, not just refreshing its info.  For drives
                 * other than disk 0 we will call put_disk.  We do not
                 * do this for disk 0 as we need it to be able to
                 * configure the controller.
                 */
                if (clear_all){
                        /* This isn't pretty, but we need to find the
                         * disk in our array and NULL our the pointer.
                         * This is so that we will call alloc_disk if
                         * this index is used again later.
                         */
                        for (i=0; i < CISS_MAX_LUN; i++){
                                if (h->gendisk[i] == disk) {
                                        h->gendisk[i] = NULL;
                                        break;
                                }
                        }
                        put_disk(disk);
                }
        } else {
                set_capacity(disk, 0);
        }

        --h->num_luns;
        /* zero out the disk size info */
        drv->nr_blocks = 0;
        drv->block_size = 0;
        drv->heads = 0;
        drv->sectors = 0;
        drv->cylinders = 0;
        drv->raid_level = -1;   /* This can be used as a flag variable to
                                 * indicate that this element of the drive
                                 * array is free.
                                 */

        if (clear_all) {
                /* check to see if it was the last disk */
                if (drv == h->drv + h->highest_lun) {
                        /* if so, find the new hightest lun */
                        int i, newhighest = -1;
                        for (i = 0; i <= h->highest_lun; i++) {
                                /* if the disk has size > 0, it is available */
                                if (h->drv[i].heads)
                                        newhighest = i;
                        }
                        h->highest_lun = newhighest;
                }

                drv->LunID = 0;
        }
        return 0;
}

static int fill_cmd(CommandList_struct *c, __u8 cmd, int ctlr, void *buff, size_t size, unsigned int use_unit_num,      /* 0: address the controller,
                                                                                                                           1: address logical volume log_unit,
                                                                                                                           2: periph device address is scsi3addr */
                    unsigned int log_unit, __u8 page_code,
                    unsigned char *scsi3addr, int cmd_type)
{
        ctlr_info_t *h = hba[ctlr];
        u64bit buff_dma_handle;
        int status = IO_OK;

        c->cmd_type = CMD_IOCTL_PEND;
        c->Header.ReplyQueue = 0;
        if (buff != NULL) {
                c->Header.SGList = 1;
                c->Header.SGTotal = 1;
        } else {
                c->Header.SGList = 0;
                c->Header.SGTotal = 0;
        }
        c->Header.Tag.lower = c->busaddr;

        c->Request.Type.Type = cmd_type;
        if (cmd_type == TYPE_CMD) {
                switch (cmd) {
                case CISS_INQUIRY:
                        /* If the logical unit number is 0 then, this is going
                           to controller so It's a physical command
                           mode = 0 target = 0.  So we have nothing to write.
                           otherwise, if use_unit_num == 1,
                           mode = 1(volume set addressing) target = LUNID
                           otherwise, if use_unit_num == 2,
                           mode = 0(periph dev addr) target = scsi3addr */
                        if (use_unit_num == 1) {
                                c->Header.LUN.LogDev.VolId =
                                    h->drv[log_unit].LunID;
                                c->Header.LUN.LogDev.Mode = 1;
                        } else if (use_unit_num == 2) {
                                memcpy(c->Header.LUN.LunAddrBytes, scsi3addr,
                                       8);
                                c->Header.LUN.LogDev.Mode = 0;
                        }
                        /* are we trying to read a vital product page */
                        if (page_code != 0) {
                                c->Request.CDB[1] = 0x01;
                                c->Request.CDB[2] = page_code;
                        }
                        c->Request.CDBLen = 6;
                        c->Request.Type.Attribute = ATTR_SIMPLE;
                        c->Request.Type.Direction = XFER_READ;
                        c->Request.Timeout = 0;
                        c->Request.CDB[0] = CISS_INQUIRY;
                        c->Request.CDB[4] = size & 0xFF;
                        break;
                case CISS_REPORT_LOG:
                case CISS_REPORT_PHYS:
                        /* Talking to controller so It's a physical command
                           mode = 00 target = 0.  Nothing to write.
                         */
                        c->Request.CDBLen = 12;
                        c->Request.Type.Attribute = ATTR_SIMPLE;
                        c->Request.Type.Direction = XFER_READ;
                        c->Request.Timeout = 0;
                        c->Request.CDB[0] = cmd;
                        c->Request.CDB[6] = (size >> 24) & 0xFF;        //MSB
                        c->Request.CDB[7] = (size >> 16) & 0xFF;
                        c->Request.CDB[8] = (size >> 8) & 0xFF;
                        c->Request.CDB[9] = size & 0xFF;
                        break;

                case CCISS_READ_CAPACITY:
                        c->Header.LUN.LogDev.VolId = h->drv[log_unit].LunID;
                        c->Header.LUN.LogDev.Mode = 1;
                        c->Request.CDBLen = 10;
                        c->Request.Type.Attribute = ATTR_SIMPLE;
                        c->Request.Type.Direction = XFER_READ;
                        c->Request.Timeout = 0;
                        c->Request.CDB[0] = cmd;
                        break;
                case CCISS_READ_CAPACITY_16:
                        c->Header.LUN.LogDev.VolId = h->drv[log_unit].LunID;
                        c->Header.LUN.LogDev.Mode = 1;
                        c->Request.CDBLen = 16;
                        c->Request.Type.Attribute = ATTR_SIMPLE;
                        c->Request.Type.Direction = XFER_READ;
                        c->Request.Timeout = 0;
                        c->Request.CDB[0] = cmd;
                        c->Request.CDB[1] = 0x10;
                        c->Request.CDB[10] = (size >> 24) & 0xFF;
                        c->Request.CDB[11] = (size >> 16) & 0xFF;
                        c->Request.CDB[12] = (size >> 8) & 0xFF;
                        c->Request.CDB[13] = size & 0xFF;
                        c->Request.Timeout = 0;
                        c->Request.CDB[0] = cmd;
                        break;
                case CCISS_CACHE_FLUSH:
                        c->Request.CDBLen = 12;
                        c->Request.Type.Attribute = ATTR_SIMPLE;
                        c->Request.Type.Direction = XFER_WRITE;
                        c->Request.Timeout = 0;
                        c->Request.CDB[0] = BMIC_WRITE;
                        c->Request.CDB[6] = BMIC_CACHE_FLUSH;
                        break;
                default:
                        printk(KERN_WARNING
                               "cciss%d:  Unknown Command 0x%c\n", ctlr, cmd);
                        return IO_ERROR;
                }
        } else if (cmd_type == TYPE_MSG) {
                switch (cmd) {
                case 0: /* ABORT message */
                        c->Request.CDBLen = 12;
                        c->Request.Type.Attribute = ATTR_SIMPLE;
                        c->Request.Type.Direction = XFER_WRITE;
                        c->Request.Timeout = 0;
                        c->Request.CDB[0] = cmd;        /* abort */
                        c->Request.CDB[1] = 0;  /* abort a command */
                        /* buff contains the tag of the command to abort */
                        memcpy(&c->Request.CDB[4], buff, 8);
                        break;
                case 1: /* RESET message */
                        c->Request.CDBLen = 12;