Linux-2.6.12-rc2

Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.

Let it rip!
diff --git a/drivers/scsi/scsi_transport_spi.c b/drivers/scsi/scsi_transport_spi.c
new file mode 100644
index 0000000..303d765
--- /dev/null
+++ b/drivers/scsi/scsi_transport_spi.c
@@ -0,0 +1,1020 @@
+/* 
+ *  Parallel SCSI (SPI) transport specific attributes exported to sysfs.
+ *
+ *  Copyright (c) 2003 Silicon Graphics, Inc.  All rights reserved.
+ *  Copyright (c) 2004, 2005 James Bottomley <James.Bottomley@SteelEye.com>
+ *
+ *  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; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  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
+ */
+#include <linux/ctype.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/workqueue.h>
+#include <asm/semaphore.h>
+#include <scsi/scsi.h>
+#include "scsi_priv.h"
+#include <scsi/scsi_device.h>
+#include <scsi/scsi_host.h>
+#include <scsi/scsi_request.h>
+#include <scsi/scsi_eh.h>
+#include <scsi/scsi_transport.h>
+#include <scsi/scsi_transport_spi.h>
+
+#define SPI_PRINTK(x, l, f, a...)	dev_printk(l, &(x)->dev, f , ##a)
+
+#define SPI_NUM_ATTRS 10	/* increase this if you add attributes */
+#define SPI_OTHER_ATTRS 1	/* Increase this if you add "always
+				 * on" attributes */
+#define SPI_HOST_ATTRS	1
+
+#define SPI_MAX_ECHO_BUFFER_SIZE	4096
+
+/* Private data accessors (keep these out of the header file) */
+#define spi_dv_pending(x) (((struct spi_transport_attrs *)&(x)->starget_data)->dv_pending)
+#define spi_dv_sem(x) (((struct spi_transport_attrs *)&(x)->starget_data)->dv_sem)
+
+struct spi_internal {
+	struct scsi_transport_template t;
+	struct spi_function_template *f;
+	/* The actual attributes */
+	struct class_device_attribute private_attrs[SPI_NUM_ATTRS];
+	/* The array of null terminated pointers to attributes 
+	 * needed by scsi_sysfs.c */
+	struct class_device_attribute *attrs[SPI_NUM_ATTRS + SPI_OTHER_ATTRS + 1];
+	struct class_device_attribute private_host_attrs[SPI_HOST_ATTRS];
+	struct class_device_attribute *host_attrs[SPI_HOST_ATTRS + 1];
+};
+
+#define to_spi_internal(tmpl)	container_of(tmpl, struct spi_internal, t)
+
+static const int ppr_to_ps[] = {
+	/* The PPR values 0-6 are reserved, fill them in when
+	 * the committee defines them */
+	-1,			/* 0x00 */
+	-1,			/* 0x01 */
+	-1,			/* 0x02 */
+	-1,			/* 0x03 */
+	-1,			/* 0x04 */
+	-1,			/* 0x05 */
+	-1,			/* 0x06 */
+	 3125,			/* 0x07 */
+	 6250,			/* 0x08 */
+	12500,			/* 0x09 */
+	25000,			/* 0x0a */
+	30300,			/* 0x0b */
+	50000,			/* 0x0c */
+};
+/* The PPR values at which you calculate the period in ns by multiplying
+ * by 4 */
+#define SPI_STATIC_PPR	0x0c
+
+static int sprint_frac(char *dest, int value, int denom)
+{
+	int frac = value % denom;
+	int result = sprintf(dest, "%d", value / denom);
+
+	if (frac == 0)
+		return result;
+	dest[result++] = '.';
+
+	do {
+		denom /= 10;
+		sprintf(dest + result, "%d", frac / denom);
+		result++;
+		frac %= denom;
+	} while (frac);
+
+	dest[result++] = '\0';
+	return result;
+}
+
+static struct {
+	enum spi_signal_type	value;
+	char			*name;
+} signal_types[] = {
+	{ SPI_SIGNAL_UNKNOWN, "unknown" },
+	{ SPI_SIGNAL_SE, "SE" },
+	{ SPI_SIGNAL_LVD, "LVD" },
+	{ SPI_SIGNAL_HVD, "HVD" },
+};
+
+static inline const char *spi_signal_to_string(enum spi_signal_type type)
+{
+	int i;
+
+	for (i = 0; i < sizeof(signal_types)/sizeof(signal_types[0]); i++) {
+		if (type == signal_types[i].value)
+			return signal_types[i].name;
+	}
+	return NULL;
+}
+static inline enum spi_signal_type spi_signal_to_value(const char *name)
+{
+	int i, len;
+
+	for (i = 0; i < sizeof(signal_types)/sizeof(signal_types[0]); i++) {
+		len =  strlen(signal_types[i].name);
+		if (strncmp(name, signal_types[i].name, len) == 0 &&
+		    (name[len] == '\n' || name[len] == '\0'))
+			return signal_types[i].value;
+	}
+	return SPI_SIGNAL_UNKNOWN;
+}
+
+static int spi_host_setup(struct device *dev)
+{
+	struct Scsi_Host *shost = dev_to_shost(dev);
+
+	spi_signalling(shost) = SPI_SIGNAL_UNKNOWN;
+
+	return 0;
+}
+
+static DECLARE_TRANSPORT_CLASS(spi_host_class,
+			       "spi_host",
+			       spi_host_setup,
+			       NULL,
+			       NULL);
+
+static int spi_host_match(struct attribute_container *cont,
+			  struct device *dev)
+{
+	struct Scsi_Host *shost;
+	struct spi_internal *i;
+
+	if (!scsi_is_host_device(dev))
+		return 0;
+
+	shost = dev_to_shost(dev);
+	if (!shost->transportt  || shost->transportt->host_attrs.ac.class
+	    != &spi_host_class.class)
+		return 0;
+
+	i = to_spi_internal(shost->transportt);
+	
+	return &i->t.host_attrs.ac == cont;
+}
+
+static int spi_device_configure(struct device *dev)
+{
+	struct scsi_device *sdev = to_scsi_device(dev);
+	struct scsi_target *starget = sdev->sdev_target;
+
+	/* Populate the target capability fields with the values
+	 * gleaned from the device inquiry */
+
+	spi_support_sync(starget) = scsi_device_sync(sdev);
+	spi_support_wide(starget) = scsi_device_wide(sdev);
+	spi_support_dt(starget) = scsi_device_dt(sdev);
+	spi_support_dt_only(starget) = scsi_device_dt_only(sdev);
+	spi_support_ius(starget) = scsi_device_ius(sdev);
+	spi_support_qas(starget) = scsi_device_qas(sdev);
+
+	return 0;
+}
+
+static int spi_setup_transport_attrs(struct device *dev)
+{
+	struct scsi_target *starget = to_scsi_target(dev);
+
+	spi_period(starget) = -1;	/* illegal value */
+	spi_offset(starget) = 0;	/* async */
+	spi_width(starget) = 0;	/* narrow */
+	spi_iu(starget) = 0;	/* no IU */
+	spi_dt(starget) = 0;	/* ST */
+	spi_qas(starget) = 0;
+	spi_wr_flow(starget) = 0;
+	spi_rd_strm(starget) = 0;
+	spi_rti(starget) = 0;
+	spi_pcomp_en(starget) = 0;
+	spi_dv_pending(starget) = 0;
+	spi_initial_dv(starget) = 0;
+	init_MUTEX(&spi_dv_sem(starget));
+
+	return 0;
+}
+
+#define spi_transport_show_function(field, format_string)		\
+									\
+static ssize_t								\
+show_spi_transport_##field(struct class_device *cdev, char *buf)	\
+{									\
+	struct scsi_target *starget = transport_class_to_starget(cdev);	\
+	struct Scsi_Host *shost = dev_to_shost(starget->dev.parent);	\
+	struct spi_transport_attrs *tp;					\
+	struct spi_internal *i = to_spi_internal(shost->transportt);	\
+	tp = (struct spi_transport_attrs *)&starget->starget_data;	\
+	if (i->f->get_##field)						\
+		i->f->get_##field(starget);				\
+	return snprintf(buf, 20, format_string, tp->field);		\
+}
+
+#define spi_transport_store_function(field, format_string)		\
+static ssize_t								\
+store_spi_transport_##field(struct class_device *cdev, const char *buf, \
+			    size_t count)				\
+{									\
+	int val;							\
+	struct scsi_target *starget = transport_class_to_starget(cdev);	\
+	struct Scsi_Host *shost = dev_to_shost(starget->dev.parent);	\
+	struct spi_internal *i = to_spi_internal(shost->transportt);	\
+									\
+	val = simple_strtoul(buf, NULL, 0);				\
+	i->f->set_##field(starget, val);				\
+	return count;							\
+}
+
+#define spi_transport_rd_attr(field, format_string)			\
+	spi_transport_show_function(field, format_string)		\
+	spi_transport_store_function(field, format_string)		\
+static CLASS_DEVICE_ATTR(field, S_IRUGO | S_IWUSR,			\
+			 show_spi_transport_##field,			\
+			 store_spi_transport_##field);
+
+/* The Parallel SCSI Tranport Attributes: */
+spi_transport_rd_attr(offset, "%d\n");
+spi_transport_rd_attr(width, "%d\n");
+spi_transport_rd_attr(iu, "%d\n");
+spi_transport_rd_attr(dt, "%d\n");
+spi_transport_rd_attr(qas, "%d\n");
+spi_transport_rd_attr(wr_flow, "%d\n");
+spi_transport_rd_attr(rd_strm, "%d\n");
+spi_transport_rd_attr(rti, "%d\n");
+spi_transport_rd_attr(pcomp_en, "%d\n");
+
+static ssize_t
+store_spi_revalidate(struct class_device *cdev, const char *buf, size_t count)
+{
+	struct scsi_target *starget = transport_class_to_starget(cdev);
+
+	/* FIXME: we're relying on an awful lot of device internals
+	 * here.  We really need a function to get the first available
+	 * child */
+	struct device *dev = container_of(starget->dev.children.next, struct device, node);
+	struct scsi_device *sdev = to_scsi_device(dev);
+	spi_dv_device(sdev);
+	return count;
+}
+static CLASS_DEVICE_ATTR(revalidate, S_IWUSR, NULL, store_spi_revalidate);
+
+/* Translate the period into ns according to the current spec
+ * for SDTR/PPR messages */
+static ssize_t show_spi_transport_period(struct class_device *cdev, char *buf)
+
+{
+	struct scsi_target *starget = transport_class_to_starget(cdev);
+	struct Scsi_Host *shost = dev_to_shost(starget->dev.parent);
+	struct spi_transport_attrs *tp;
+	int len, picosec;
+	struct spi_internal *i = to_spi_internal(shost->transportt);
+
+	tp = (struct spi_transport_attrs *)&starget->starget_data;
+
+	if (i->f->get_period)
+		i->f->get_period(starget);
+
+	if (tp->period < 0 || tp->period > 0xff) {
+		picosec = -1;
+	} else if (tp->period <= SPI_STATIC_PPR) {
+		picosec = ppr_to_ps[tp->period];
+	} else {
+		picosec = tp->period * 4000;
+	}
+
+	if (picosec == -1) {
+		len = sprintf(buf, "reserved");
+	} else {
+		len = sprint_frac(buf, picosec, 1000);
+	}
+
+	buf[len++] = '\n';
+	buf[len] = '\0';
+	return len;
+}
+
+static ssize_t
+store_spi_transport_period(struct class_device *cdev, const char *buf,
+			    size_t count)
+{
+	struct scsi_target *starget = transport_class_to_starget(cdev);
+	struct Scsi_Host *shost = dev_to_shost(starget->dev.parent);
+	struct spi_internal *i = to_spi_internal(shost->transportt);
+	int j, picosec, period = -1;
+	char *endp;
+
+	picosec = simple_strtoul(buf, &endp, 10) * 1000;
+	if (*endp == '.') {
+		int mult = 100;
+		do {
+			endp++;
+			if (!isdigit(*endp))
+				break;
+			picosec += (*endp - '0') * mult;
+			mult /= 10;
+		} while (mult > 0);
+	}
+
+	for (j = 0; j <= SPI_STATIC_PPR; j++) {
+		if (ppr_to_ps[j] < picosec)
+			continue;
+		period = j;
+		break;
+	}
+
+	if (period == -1)
+		period = picosec / 4000;
+
+	if (period > 0xff)
+		period = 0xff;
+
+	i->f->set_period(starget, period);
+
+	return count;
+}
+
+static CLASS_DEVICE_ATTR(period, S_IRUGO | S_IWUSR, 
+			 show_spi_transport_period,
+			 store_spi_transport_period);
+
+static ssize_t show_spi_host_signalling(struct class_device *cdev, char *buf)
+{
+	struct Scsi_Host *shost = transport_class_to_shost(cdev);
+	struct spi_internal *i = to_spi_internal(shost->transportt);
+
+	if (i->f->get_signalling)
+		i->f->get_signalling(shost);
+
+	return sprintf(buf, "%s\n", spi_signal_to_string(spi_signalling(shost)));
+}
+static ssize_t store_spi_host_signalling(struct class_device *cdev,
+					 const char *buf, size_t count)
+{
+	struct Scsi_Host *shost = transport_class_to_shost(cdev);
+	struct spi_internal *i = to_spi_internal(shost->transportt);
+	enum spi_signal_type type = spi_signal_to_value(buf);
+
+	if (type != SPI_SIGNAL_UNKNOWN)
+		i->f->set_signalling(shost, type);
+
+	return count;
+}
+static CLASS_DEVICE_ATTR(signalling, S_IRUGO | S_IWUSR,
+			 show_spi_host_signalling,
+			 store_spi_host_signalling);
+
+#define DV_SET(x, y)			\
+	if(i->f->set_##x)		\
+		i->f->set_##x(sdev->sdev_target, y)
+
+#define DV_LOOPS	3
+#define DV_TIMEOUT	(10*HZ)
+#define DV_RETRIES	3	/* should only need at most 
+				 * two cc/ua clears */
+
+enum spi_compare_returns {
+	SPI_COMPARE_SUCCESS,
+	SPI_COMPARE_FAILURE,
+	SPI_COMPARE_SKIP_TEST,
+};
+
+
+/* This is for read/write Domain Validation:  If the device supports
+ * an echo buffer, we do read/write tests to it */
+static enum spi_compare_returns
+spi_dv_device_echo_buffer(struct scsi_request *sreq, u8 *buffer,
+			  u8 *ptr, const int retries)
+{
+	struct scsi_device *sdev = sreq->sr_device;
+	int len = ptr - buffer;
+	int j, k, r;
+	unsigned int pattern = 0x0000ffff;
+
+	const char spi_write_buffer[] = {
+		WRITE_BUFFER, 0x0a, 0, 0, 0, 0, 0, len >> 8, len & 0xff, 0
+	};
+	const char spi_read_buffer[] = {
+		READ_BUFFER, 0x0a, 0, 0, 0, 0, 0, len >> 8, len & 0xff, 0
+	};
+
+	/* set up the pattern buffer.  Doesn't matter if we spill
+	 * slightly beyond since that's where the read buffer is */
+	for (j = 0; j < len; ) {
+
+		/* fill the buffer with counting (test a) */
+		for ( ; j < min(len, 32); j++)
+			buffer[j] = j;
+		k = j;
+		/* fill the buffer with alternating words of 0x0 and
+		 * 0xffff (test b) */
+		for ( ; j < min(len, k + 32); j += 2) {
+			u16 *word = (u16 *)&buffer[j];
+			
+			*word = (j & 0x02) ? 0x0000 : 0xffff;
+		}
+		k = j;
+		/* fill with crosstalk (alternating 0x5555 0xaaa)
+                 * (test c) */
+		for ( ; j < min(len, k + 32); j += 2) {
+			u16 *word = (u16 *)&buffer[j];
+
+			*word = (j & 0x02) ? 0x5555 : 0xaaaa;
+		}
+		k = j;
+		/* fill with shifting bits (test d) */
+		for ( ; j < min(len, k + 32); j += 4) {
+			u32 *word = (unsigned int *)&buffer[j];
+			u32 roll = (pattern & 0x80000000) ? 1 : 0;
+			
+			*word = pattern;
+			pattern = (pattern << 1) | roll;
+		}
+		/* don't bother with random data (test e) */
+	}
+
+	for (r = 0; r < retries; r++) {
+		sreq->sr_cmd_len = 0;	/* wait_req to fill in */
+		sreq->sr_data_direction = DMA_TO_DEVICE;
+		scsi_wait_req(sreq, spi_write_buffer, buffer, len,
+			      DV_TIMEOUT, DV_RETRIES);
+		if(sreq->sr_result || !scsi_device_online(sdev)) {
+			struct scsi_sense_hdr sshdr;
+
+			scsi_device_set_state(sdev, SDEV_QUIESCE);
+			if (scsi_request_normalize_sense(sreq, &sshdr)
+			    && sshdr.sense_key == ILLEGAL_REQUEST
+			    /* INVALID FIELD IN CDB */
+			    && sshdr.asc == 0x24 && sshdr.ascq == 0x00)
+				/* This would mean that the drive lied
+				 * to us about supporting an echo
+				 * buffer (unfortunately some Western
+				 * Digital drives do precisely this)
+				 */
+				return SPI_COMPARE_SKIP_TEST;
+
+
+			SPI_PRINTK(sdev->sdev_target, KERN_ERR, "Write Buffer failure %x\n", sreq->sr_result);
+			return SPI_COMPARE_FAILURE;
+		}
+
+		memset(ptr, 0, len);
+		sreq->sr_cmd_len = 0;	/* wait_req to fill in */
+		sreq->sr_data_direction = DMA_FROM_DEVICE;
+		scsi_wait_req(sreq, spi_read_buffer, ptr, len,
+			      DV_TIMEOUT, DV_RETRIES);
+		scsi_device_set_state(sdev, SDEV_QUIESCE);
+
+		if (memcmp(buffer, ptr, len) != 0)
+			return SPI_COMPARE_FAILURE;
+	}
+	return SPI_COMPARE_SUCCESS;
+}
+
+/* This is for the simplest form of Domain Validation: a read test
+ * on the inquiry data from the device */
+static enum spi_compare_returns
+spi_dv_device_compare_inquiry(struct scsi_request *sreq, u8 *buffer,
+			      u8 *ptr, const int retries)
+{
+	int r;
+	const int len = sreq->sr_device->inquiry_len;
+	struct scsi_device *sdev = sreq->sr_device;
+	const char spi_inquiry[] = {
+		INQUIRY, 0, 0, 0, len, 0
+	};
+
+	for (r = 0; r < retries; r++) {
+		sreq->sr_cmd_len = 0;	/* wait_req to fill in */
+		sreq->sr_data_direction = DMA_FROM_DEVICE;
+
+		memset(ptr, 0, len);
+
+		scsi_wait_req(sreq, spi_inquiry, ptr, len,
+			      DV_TIMEOUT, DV_RETRIES);
+		
+		if(sreq->sr_result || !scsi_device_online(sdev)) {
+			scsi_device_set_state(sdev, SDEV_QUIESCE);
+			return SPI_COMPARE_FAILURE;
+		}
+
+		/* If we don't have the inquiry data already, the
+		 * first read gets it */
+		if (ptr == buffer) {
+			ptr += len;
+			--r;
+			continue;
+		}
+
+		if (memcmp(buffer, ptr, len) != 0)
+			/* failure */
+			return SPI_COMPARE_FAILURE;
+	}
+	return SPI_COMPARE_SUCCESS;
+}
+
+static enum spi_compare_returns
+spi_dv_retrain(struct scsi_request *sreq, u8 *buffer, u8 *ptr,
+	       enum spi_compare_returns 
+	       (*compare_fn)(struct scsi_request *, u8 *, u8 *, int))
+{
+	struct spi_internal *i = to_spi_internal(sreq->sr_host->transportt);
+	struct scsi_device *sdev = sreq->sr_device;
+	int period = 0, prevperiod = 0; 
+	enum spi_compare_returns retval;
+
+
+	for (;;) {
+		int newperiod;
+		retval = compare_fn(sreq, buffer, ptr, DV_LOOPS);
+
+		if (retval == SPI_COMPARE_SUCCESS
+		    || retval == SPI_COMPARE_SKIP_TEST)
+			break;
+
+		/* OK, retrain, fallback */
+		if (i->f->get_period)
+			i->f->get_period(sdev->sdev_target);
+		newperiod = spi_period(sdev->sdev_target);
+		period = newperiod > period ? newperiod : period;
+		if (period < 0x0d)
+			period++;
+		else
+			period += period >> 1;
+
+		if (unlikely(period > 0xff || period == prevperiod)) {
+			/* Total failure; set to async and return */
+			SPI_PRINTK(sdev->sdev_target, KERN_ERR, "Domain Validation Failure, dropping back to Asynchronous\n");
+			DV_SET(offset, 0);
+			return SPI_COMPARE_FAILURE;
+		}
+		SPI_PRINTK(sdev->sdev_target, KERN_ERR, "Domain Validation detected failure, dropping back\n");
+		DV_SET(period, period);
+		prevperiod = period;
+	}
+	return retval;
+}
+
+static int
+spi_dv_device_get_echo_buffer(struct scsi_request *sreq, u8 *buffer)
+{
+	int l;
+
+	/* first off do a test unit ready.  This can error out 
+	 * because of reservations or some other reason.  If it
+	 * fails, the device won't let us write to the echo buffer
+	 * so just return failure */
+	
+	const char spi_test_unit_ready[] = {
+		TEST_UNIT_READY, 0, 0, 0, 0, 0
+	};
+
+	const char spi_read_buffer_descriptor[] = {
+		READ_BUFFER, 0x0b, 0, 0, 0, 0, 0, 0, 4, 0
+	};
+
+	
+	sreq->sr_cmd_len = 0;
+	sreq->sr_data_direction = DMA_NONE;
+
+	/* We send a set of three TURs to clear any outstanding 
+	 * unit attention conditions if they exist (Otherwise the
+	 * buffer tests won't be happy).  If the TUR still fails
+	 * (reservation conflict, device not ready, etc) just
+	 * skip the write tests */
+	for (l = 0; ; l++) {
+		scsi_wait_req(sreq, spi_test_unit_ready, NULL, 0,
+			      DV_TIMEOUT, DV_RETRIES);
+
+		if(sreq->sr_result) {
+			if(l >= 3)
+				return 0;
+		} else {
+			/* TUR succeeded */
+			break;
+		}
+	}
+
+	sreq->sr_cmd_len = 0;
+	sreq->sr_data_direction = DMA_FROM_DEVICE;
+
+	scsi_wait_req(sreq, spi_read_buffer_descriptor, buffer, 4,
+		      DV_TIMEOUT, DV_RETRIES);
+
+	if (sreq->sr_result)
+		/* Device has no echo buffer */
+		return 0;
+
+	return buffer[3] + ((buffer[2] & 0x1f) << 8);
+}
+
+static void
+spi_dv_device_internal(struct scsi_request *sreq, u8 *buffer)
+{
+	struct spi_internal *i = to_spi_internal(sreq->sr_host->transportt);
+	struct scsi_device *sdev = sreq->sr_device;
+	int len = sdev->inquiry_len;
+	/* first set us up for narrow async */
+	DV_SET(offset, 0);
+	DV_SET(width, 0);
+	
+	if (spi_dv_device_compare_inquiry(sreq, buffer, buffer, DV_LOOPS)
+	    != SPI_COMPARE_SUCCESS) {
+		SPI_PRINTK(sdev->sdev_target, KERN_ERR, "Domain Validation Initial Inquiry Failed\n");
+		/* FIXME: should probably offline the device here? */
+		return;
+	}
+
+	/* test width */
+	if (i->f->set_width && sdev->wdtr) {
+		i->f->set_width(sdev->sdev_target, 1);
+
+		if (spi_dv_device_compare_inquiry(sreq, buffer,
+						   buffer + len,
+						   DV_LOOPS)
+		    != SPI_COMPARE_SUCCESS) {
+			SPI_PRINTK(sdev->sdev_target, KERN_ERR, "Wide Transfers Fail\n");
+			i->f->set_width(sdev->sdev_target, 0);
+		}
+	}
+
+	if (!i->f->set_period)
+		return;
+
+	/* device can't handle synchronous */
+	if(!sdev->ppr && !sdev->sdtr)
+		return;
+
+	/* see if the device has an echo buffer.  If it does we can
+	 * do the SPI pattern write tests */
+
+	len = 0;
+	if (sdev->ppr)
+		len = spi_dv_device_get_echo_buffer(sreq, buffer);
+
+ retry:
+
+	/* now set up to the maximum */
+	DV_SET(offset, 255);
+	DV_SET(period, 1);
+
+	if (len == 0) {
+		SPI_PRINTK(sdev->sdev_target, KERN_INFO, "Domain Validation skipping write tests\n");
+		spi_dv_retrain(sreq, buffer, buffer + len,
+			       spi_dv_device_compare_inquiry);
+		return;
+	}
+
+	if (len > SPI_MAX_ECHO_BUFFER_SIZE) {
+		SPI_PRINTK(sdev->sdev_target, KERN_WARNING, "Echo buffer size %d is too big, trimming to %d\n", len, SPI_MAX_ECHO_BUFFER_SIZE);
+		len = SPI_MAX_ECHO_BUFFER_SIZE;
+	}
+
+	if (spi_dv_retrain(sreq, buffer, buffer + len,
+			   spi_dv_device_echo_buffer)
+	    == SPI_COMPARE_SKIP_TEST) {
+		/* OK, the stupid drive can't do a write echo buffer
+		 * test after all, fall back to the read tests */
+		len = 0;
+		goto retry;
+	}
+}
+
+
+/**	spi_dv_device - Do Domain Validation on the device
+ *	@sdev:		scsi device to validate
+ *
+ *	Performs the domain validation on the given device in the
+ *	current execution thread.  Since DV operations may sleep,
+ *	the current thread must have user context.  Also no SCSI
+ *	related locks that would deadlock I/O issued by the DV may
+ *	be held.
+ */
+void
+spi_dv_device(struct scsi_device *sdev)
+{
+	struct scsi_request *sreq = scsi_allocate_request(sdev, GFP_KERNEL);
+	struct scsi_target *starget = sdev->sdev_target;
+	u8 *buffer;
+	const int len = SPI_MAX_ECHO_BUFFER_SIZE*2;
+
+	if (unlikely(!sreq))
+		return;
+
+	if (unlikely(scsi_device_get(sdev)))
+		goto out_free_req;
+
+	buffer = kmalloc(len, GFP_KERNEL);
+
+	if (unlikely(!buffer))
+		goto out_put;
+
+	memset(buffer, 0, len);
+
+	/* We need to verify that the actual device will quiesce; the
+	 * later target quiesce is just a nice to have */
+	if (unlikely(scsi_device_quiesce(sdev)))
+		goto out_free;
+
+	scsi_target_quiesce(starget);
+
+	spi_dv_pending(starget) = 1;
+	down(&spi_dv_sem(starget));
+
+	SPI_PRINTK(starget, KERN_INFO, "Beginning Domain Validation\n");
+
+	spi_dv_device_internal(sreq, buffer);
+
+	SPI_PRINTK(starget, KERN_INFO, "Ending Domain Validation\n");
+
+	up(&spi_dv_sem(starget));
+	spi_dv_pending(starget) = 0;
+
+	scsi_target_resume(starget);
+
+	spi_initial_dv(starget) = 1;
+
+ out_free:
+	kfree(buffer);
+ out_put:
+	scsi_device_put(sdev);
+ out_free_req:
+	scsi_release_request(sreq);
+}
+EXPORT_SYMBOL(spi_dv_device);
+
+struct work_queue_wrapper {
+	struct work_struct	work;
+	struct scsi_device	*sdev;
+};
+
+static void
+spi_dv_device_work_wrapper(void *data)
+{
+	struct work_queue_wrapper *wqw = (struct work_queue_wrapper *)data;
+	struct scsi_device *sdev = wqw->sdev;
+
+	kfree(wqw);
+	spi_dv_device(sdev);
+	spi_dv_pending(sdev->sdev_target) = 0;
+	scsi_device_put(sdev);
+}
+
+
+/**
+ *	spi_schedule_dv_device - schedule domain validation to occur on the device
+ *	@sdev:	The device to validate
+ *
+ *	Identical to spi_dv_device() above, except that the DV will be
+ *	scheduled to occur in a workqueue later.  All memory allocations
+ *	are atomic, so may be called from any context including those holding
+ *	SCSI locks.
+ */
+void
+spi_schedule_dv_device(struct scsi_device *sdev)
+{
+	struct work_queue_wrapper *wqw =
+		kmalloc(sizeof(struct work_queue_wrapper), GFP_ATOMIC);
+
+	if (unlikely(!wqw))
+		return;
+
+	if (unlikely(spi_dv_pending(sdev->sdev_target))) {
+		kfree(wqw);
+		return;
+	}
+	/* Set pending early (dv_device doesn't check it, only sets it) */
+	spi_dv_pending(sdev->sdev_target) = 1;
+	if (unlikely(scsi_device_get(sdev))) {
+		kfree(wqw);
+		spi_dv_pending(sdev->sdev_target) = 0;
+		return;
+	}
+
+	INIT_WORK(&wqw->work, spi_dv_device_work_wrapper, wqw);
+	wqw->sdev = sdev;
+
+	schedule_work(&wqw->work);
+}
+EXPORT_SYMBOL(spi_schedule_dv_device);
+
+/**
+ * spi_display_xfer_agreement - Print the current target transfer agreement
+ * @starget: The target for which to display the agreement
+ *
+ * Each SPI port is required to maintain a transfer agreement for each
+ * other port on the bus.  This function prints a one-line summary of
+ * the current agreement; more detailed information is available in sysfs.
+ */
+void spi_display_xfer_agreement(struct scsi_target *starget)
+{
+	struct spi_transport_attrs *tp;
+	tp = (struct spi_transport_attrs *)&starget->starget_data;
+
+	if (tp->offset > 0 && tp->period > 0) {
+		unsigned int picosec, kb100;
+		char *scsi = "FAST-?";
+		char tmp[8];
+
+		if (tp->period <= SPI_STATIC_PPR) {
+			picosec = ppr_to_ps[tp->period];
+			switch (tp->period) {
+				case  7: scsi = "FAST-320"; break;
+				case  8: scsi = "FAST-160"; break;
+				case  9: scsi = "FAST-80"; break;
+				case 10:
+				case 11: scsi = "FAST-40"; break;
+				case 12: scsi = "FAST-20"; break;
+			}
+		} else {
+			picosec = tp->period * 4000;
+			if (tp->period < 25)
+				scsi = "FAST-20";
+			else if (tp->period < 50)
+				scsi = "FAST-10";
+			else
+				scsi = "FAST-5";
+		}
+
+		kb100 = (10000000 + picosec / 2) / picosec;
+		if (tp->width)
+			kb100 *= 2;
+		sprint_frac(tmp, picosec, 1000);
+
+		dev_info(&starget->dev,
+			"%s %sSCSI %d.%d MB/s %s%s%s (%s ns, offset %d)\n",
+			scsi, tp->width ? "WIDE " : "", kb100/10, kb100 % 10,
+			tp->dt ? "DT" : "ST", tp->iu ? " IU" : "",
+			tp->qas  ? " QAS" : "", tmp, tp->offset);
+	} else {
+		dev_info(&starget->dev, "%sasynchronous.\n",
+				tp->width ? "wide " : "");
+	}
+}
+EXPORT_SYMBOL(spi_display_xfer_agreement);
+
+#define SETUP_ATTRIBUTE(field)						\
+	i->private_attrs[count] = class_device_attr_##field;		\
+	if (!i->f->set_##field) {					\
+		i->private_attrs[count].attr.mode = S_IRUGO;		\
+		i->private_attrs[count].store = NULL;			\
+	}								\
+	i->attrs[count] = &i->private_attrs[count];			\
+	if (i->f->show_##field)						\
+		count++
+
+#define SETUP_HOST_ATTRIBUTE(field)					\
+	i->private_host_attrs[count] = class_device_attr_##field;	\
+	if (!i->f->set_##field) {					\
+		i->private_host_attrs[count].attr.mode = S_IRUGO;	\
+		i->private_host_attrs[count].store = NULL;		\
+	}								\
+	i->host_attrs[count] = &i->private_host_attrs[count];		\
+	count++
+
+static int spi_device_match(struct attribute_container *cont,
+			    struct device *dev)
+{
+	struct scsi_device *sdev;
+	struct Scsi_Host *shost;
+
+	if (!scsi_is_sdev_device(dev))
+		return 0;
+
+	sdev = to_scsi_device(dev);
+	shost = sdev->host;
+	if (!shost->transportt  || shost->transportt->host_attrs.ac.class
+	    != &spi_host_class.class)
+		return 0;
+	/* Note: this class has no device attributes, so it has
+	 * no per-HBA allocation and thus we don't need to distinguish
+	 * the attribute containers for the device */
+	return 1;
+}
+
+static int spi_target_match(struct attribute_container *cont,
+			    struct device *dev)
+{
+	struct Scsi_Host *shost;
+	struct spi_internal *i;
+
+	if (!scsi_is_target_device(dev))
+		return 0;
+
+	shost = dev_to_shost(dev->parent);
+	if (!shost->transportt  || shost->transportt->host_attrs.ac.class
+	    != &spi_host_class.class)
+		return 0;
+
+	i = to_spi_internal(shost->transportt);
+	
+	return &i->t.target_attrs.ac == cont;
+}
+
+static DECLARE_TRANSPORT_CLASS(spi_transport_class,
+			       "spi_transport",
+			       spi_setup_transport_attrs,
+			       NULL,
+			       NULL);
+
+static DECLARE_ANON_TRANSPORT_CLASS(spi_device_class,
+				    spi_device_match,
+				    spi_device_configure);
+
+struct scsi_transport_template *
+spi_attach_transport(struct spi_function_template *ft)
+{
+	struct spi_internal *i = kmalloc(sizeof(struct spi_internal),
+					 GFP_KERNEL);
+	int count = 0;
+	if (unlikely(!i))
+		return NULL;
+
+	memset(i, 0, sizeof(struct spi_internal));
+
+
+	i->t.target_attrs.ac.class = &spi_transport_class.class;
+	i->t.target_attrs.ac.attrs = &i->attrs[0];
+	i->t.target_attrs.ac.match = spi_target_match;
+	transport_container_register(&i->t.target_attrs);
+	i->t.target_size = sizeof(struct spi_transport_attrs);
+	i->t.host_attrs.ac.class = &spi_host_class.class;
+	i->t.host_attrs.ac.attrs = &i->host_attrs[0];
+	i->t.host_attrs.ac.match = spi_host_match;
+	transport_container_register(&i->t.host_attrs);
+	i->t.host_size = sizeof(struct spi_host_attrs);
+	i->f = ft;
+
+	SETUP_ATTRIBUTE(period);
+	SETUP_ATTRIBUTE(offset);
+	SETUP_ATTRIBUTE(width);
+	SETUP_ATTRIBUTE(iu);
+	SETUP_ATTRIBUTE(dt);
+	SETUP_ATTRIBUTE(qas);
+	SETUP_ATTRIBUTE(wr_flow);
+	SETUP_ATTRIBUTE(rd_strm);
+	SETUP_ATTRIBUTE(rti);
+	SETUP_ATTRIBUTE(pcomp_en);
+
+	/* if you add an attribute but forget to increase SPI_NUM_ATTRS
+	 * this bug will trigger */
+	BUG_ON(count > SPI_NUM_ATTRS);
+
+	i->attrs[count++] = &class_device_attr_revalidate;
+
+	i->attrs[count] = NULL;
+
+	count = 0;
+	SETUP_HOST_ATTRIBUTE(signalling);
+
+	BUG_ON(count > SPI_HOST_ATTRS);
+
+	i->host_attrs[count] = NULL;
+
+	return &i->t;
+}
+EXPORT_SYMBOL(spi_attach_transport);
+
+void spi_release_transport(struct scsi_transport_template *t)
+{
+	struct spi_internal *i = to_spi_internal(t);
+
+	transport_container_unregister(&i->t.target_attrs);
+	transport_container_unregister(&i->t.host_attrs);
+
+	kfree(i);
+}
+EXPORT_SYMBOL(spi_release_transport);
+
+static __init int spi_transport_init(void)
+{
+	int error = transport_class_register(&spi_transport_class);
+	if (error)
+		return error;
+	error = anon_transport_class_register(&spi_device_class);
+	return transport_class_register(&spi_host_class);
+}
+
+static void __exit spi_transport_exit(void)
+{
+	transport_class_unregister(&spi_transport_class);
+	anon_transport_class_unregister(&spi_device_class);
+	transport_class_unregister(&spi_host_class);
+}
+
+MODULE_AUTHOR("Martin Hicks");
+MODULE_DESCRIPTION("SPI Transport Attributes");
+MODULE_LICENSE("GPL");
+
+module_init(spi_transport_init);
+module_exit(spi_transport_exit);