bump to 2.6.25-r1 more powersaving and bay is working now

svn path=/trunk/; revision=425

1 files changed, 917 insertions, 0 deletions

diff --git a/sys-kernel/thinkpad-sources/files/2.6.25-r1/01-disk-protect-for-2.6.25.patch b/sys-kernel/thinkpad-sources/files/2.6.25-r1/01-disk-protect-for-2.6.25.patch
new file mode 100644
index 0000000..9738753
--- /dev/null
+++ b/sys-kernel/thinkpad-sources/files/2.6.25-r1/01-disk-protect-for-2.6.25.patch
@@ -0,0 +1,917 @@
+diff -Naur a/block/blk-core.c b/block/blk-core.c
+--- a/block/blk-core.c	2008-04-13 14:04:22.000000000 +0200
++++ b/block/blk-core.c	2008-04-13 14:25:48.000000000 +0200
+@@ -320,6 +320,46 @@
+ }
+ EXPORT_SYMBOL(blk_unplug);
+ 
++/*
++ * Issue lower level unprotect function if no timers are pending.
++ */
++void blk_unfreeze_work(struct work_struct *work)
++{
++	struct request_queue *q = container_of(work, struct request_queue, unfreeze_work);
++	int pending;
++	unsigned long flags;
++
++	spin_lock_irqsave(q->queue_lock, flags);
++	pending = timer_pending(&q->unfreeze_timer);
++	spin_unlock_irqrestore(q->queue_lock, flags);
++	if (!pending)
++		q->issue_unprotect_fn(q);
++}
++
++/*
++ * Called when the queue freeze timeout expires...
++ */
++void blk_unfreeze_timeout(unsigned long data)
++{
++	struct request_queue *q = (struct request_queue *) data;
++
++	kblockd_schedule_work(&q->unfreeze_work);
++}
++
++/*
++ * The lower level driver parks and freezes the queue, and this block layer
++ *  function sets up the freeze timeout timer on return. If the queue is
++ *  already frozen then this is called to extend the timer...
++ */
++void blk_freeze_queue(struct request_queue *q, int seconds)
++{
++	/* Don't accept arbitrarily long freezes */
++	if (seconds >= q->max_unfreeze)
++		seconds = q->max_unfreeze;
++	/* set/reset the timer */
++	mod_timer(&q->unfreeze_timer, msecs_to_jiffies(seconds*1000) + jiffies);
++}
++
+ /**
+  * blk_start_queue - restart a previously stopped queue
+  * @q:    The &struct request_queue in question
+@@ -482,6 +522,7 @@
+ 	}
+ 
+ 	init_timer(&q->unplug_timer);
++	init_timer(&q->unfreeze_timer);
+ 
+ 	kobject_init(&q->kobj, &blk_queue_ktype);
+ 
+diff -Naur a/block/blk.h b/block/blk.h
+--- a/block/blk.h	2008-04-13 14:04:22.000000000 +0200
++++ b/block/blk.h	2008-04-13 14:25:48.000000000 +0200
+@@ -18,6 +18,9 @@
+ 
+ void blk_unplug_work(struct work_struct *work);
+ void blk_unplug_timeout(unsigned long data);
++void blk_unfreeze_work(struct work_struct *work);
++void blk_unfreeze_timeout(unsigned long data);
++void blk_freeze_queue(struct request_queue *q, int seconds);
+ 
+ struct io_context *current_io_context(gfp_t gfp_flags, int node);
+ 
+diff -Naur a/block/blk-settings.c b/block/blk-settings.c
+--- a/block/blk-settings.c	2008-04-13 14:04:22.000000000 +0200
++++ b/block/blk-settings.c	2008-04-13 14:25:48.000000000 +0200
+@@ -112,6 +112,16 @@
+ 	q->unplug_timer.function = blk_unplug_timeout;
+ 	q->unplug_timer.data = (unsigned long)q;
+ 
++	q->max_unfreeze = 30;
++
++	INIT_WORK(&q->unfreeze_work, blk_unfreeze_work);
++
++	q->unfreeze_timer.function = blk_unfreeze_timeout;
++	q->unfreeze_timer.data = (unsigned long)q;
++
++	/* Set protect_method to auto detection initially */
++	q->protect_method = 2;
++
+ 	/*
+ 	 * by default assume old behaviour and bounce for any highmem page
+ 	 */
+@@ -119,6 +129,18 @@
+ }
+ EXPORT_SYMBOL(blk_queue_make_request);
+ 
++void blk_queue_issue_protect_fn(struct request_queue *q, issue_protect_fn *ipf)
++{
++	q->issue_protect_fn = ipf;
++}
++EXPORT_SYMBOL(blk_queue_issue_protect_fn);
++
++void blk_queue_issue_unprotect_fn(struct request_queue *q, issue_unprotect_fn *iuf)
++{
++	q->issue_unprotect_fn = iuf;
++}
++EXPORT_SYMBOL(blk_queue_issue_unprotect_fn);
++
+ /**
+  * blk_queue_bounce_limit - set bounce buffer limit for queue
+  * @q:  the request queue for the device
+diff -Naur a/block/blk-sysfs.c b/block/blk-sysfs.c
+--- a/block/blk-sysfs.c	2008-04-13 14:04:22.000000000 +0200
++++ b/block/blk-sysfs.c	2008-04-13 14:25:48.000000000 +0200
+@@ -270,6 +270,160 @@
+ 	.release	= blk_release_queue,
+ };
+ 
++/*
++ * When reading the 'protect' attribute, we return seconds remaining
++ * before unfreeze timeout expires
++ */
++static ssize_t queue_protect_show(struct request_queue *q, char *page)
++{
++	unsigned int seconds = 0;
++
++	spin_lock_irq(q->queue_lock);
++	if (blk_queue_stopped(q) && timer_pending(&q->unfreeze_timer))
++		/*
++		 * Adding 1 in order to guarantee nonzero value until timer
++		 * has actually expired.
++		 */
++		seconds = jiffies_to_msecs(q->unfreeze_timer.expires
++					   - jiffies) / 1000 + 1;
++	spin_unlock_irq(q->queue_lock);
++	return queue_var_show(seconds, (page));
++}
++
++/*
++ * When writing the 'protect' attribute, input is the number of seconds
++ * to freeze the queue for. We call a lower level helper function to
++ * park the heads and freeze/block the queue, then we make a block layer
++ * call to setup the thaw timeout. If input is 0, then we thaw the queue.
++ */
++static ssize_t queue_protect_store(struct request_queue *q,
++				   const char *page, size_t count)
++{
++	unsigned long freeze = 0;
++
++	queue_var_store(&freeze, page, count);
++
++	if (freeze>0) {
++		/* Park and freeze */
++		if (!blk_queue_stopped(q))
++			q->issue_protect_fn(q);
++		/* set / reset the thaw timer */
++		spin_lock_irq(q->queue_lock);
++		blk_freeze_queue(q, freeze);
++		spin_unlock_irq(q->queue_lock);
++	} else {
++		spin_lock_irq(q->queue_lock);
++		freeze = del_timer(&q->unfreeze_timer);
++		spin_unlock_irq(q->queue_lock);
++		if (freeze)
++			q->issue_unprotect_fn(q);
++	}
++
++	return count;
++}
++
++static ssize_t
++queue_str_show(char *page, char *str, int status)
++{
++	ssize_t len;
++
++	if (status & 1)
++		len = sprintf(page, "[%s]", str);
++	else
++		len = sprintf(page, "%s", str);
++	if (status & 2)
++		len += sprintf(page+len, "\n");
++	else
++		len += sprintf(page+len, " ");
++	return len;
++}
++
++/*
++ * Returns current protect_method.
++ */
++static ssize_t queue_protect_method_show(struct request_queue *q, char *page)
++{
++	int len = 0;
++	int unload = q->protect_method;
++
++	len += queue_str_show(page+len, "auto", (unload & 2) >> 1);
++	len += queue_str_show(page+len, "unload", unload & 1);
++	len += queue_str_show(page+len, "standby", !unload ? 3 : 2);
++	return len;
++}
++
++/*
++ * Stores the device protect method.
++ */
++static ssize_t queue_protect_method_store(struct request_queue *q,
++				   const char *page, size_t count)
++{
++	spin_lock_irq(q->queue_lock);
++	if (!strcmp(page, "auto") || !strcmp(page, "auto\n"))
++		q->protect_method = 2;
++	else if (!strcmp(page, "unload") || !strcmp(page, "unload\n"))
++		q->protect_method = 1;
++	else if (!strcmp(page, "standby") || !strcmp(page, "standby\n"))
++		q->protect_method = 0;
++	else {
++		spin_unlock_irq(q->queue_lock);
++		return -EINVAL;
++	}
++	spin_unlock_irq(q->queue_lock);
++	return count;
++}
++
++static struct queue_sysfs_entry queue_protect_entry = {
++	.attr = { .name = "protect", .mode = S_IRUGO | S_IWUSR },
++	.show = queue_protect_show,
++	.store = queue_protect_store,
++};
++static struct queue_sysfs_entry queue_protect_method_entry = {
++	.attr = { .name = "protect_method", .mode = S_IRUGO | S_IWUSR },
++	.show = queue_protect_method_show,
++	.store = queue_protect_method_store,
++};
++
++static int blk_protect_register(struct request_queue *q)
++{
++	int error = 0;
++
++	/* check that the lower level driver has a protect handler */
++	if (!q->issue_protect_fn)
++		return 0;
++
++	/* create the attributes */
++	error = sysfs_create_file(&q->kobj, &queue_protect_entry.attr);
++	if (error) {
++		printk(KERN_ERR
++		       "blk_protect_register(): failed to create protect queue attribute!\n");
++		return error;
++	}
++
++	error = sysfs_create_file(&q->kobj, &queue_protect_method_entry.attr);
++	if (error) {
++		sysfs_remove_file(&q->kobj, &queue_protect_entry.attr);
++		printk(KERN_ERR
++		       "blk_protect_register(): failed to create protect_method attribute!\n");
++		return error;
++	}
++	kobject_get(&q->kobj);
++
++	return 0;
++}
++
++static void blk_protect_unregister(struct request_queue *q)
++{
++	/* check that the lower level driver has a protect handler */
++	if (!q->issue_protect_fn)
++		return;
++
++	/* remove the attributes */
++	sysfs_remove_file(&q->kobj, &queue_protect_method_entry.attr);
++	sysfs_remove_file(&q->kobj, &queue_protect_entry.attr);
++	kobject_put(&q->kobj);
++}
++
+ int blk_register_queue(struct gendisk *disk)
+ {
+ 	int ret;
+@@ -287,13 +441,20 @@
+ 	kobject_uevent(&q->kobj, KOBJ_ADD);
+ 
+ 	ret = elv_register_queue(q);
++	if (ret)
++		goto err;
++	ret = blk_protect_register(q);
+ 	if (ret) {
+-		kobject_uevent(&q->kobj, KOBJ_REMOVE);
+-		kobject_del(&q->kobj);
+-		return ret;
++		elv_unregister_queue(q);
++		goto err;
+ 	}
+ 
+ 	return 0;
++
++err:
++	kobject_uevent(&q->kobj, KOBJ_REMOVE);
++	kobject_del(&q->kobj);
++	return ret;
+ }
+ 
+ void blk_unregister_queue(struct gendisk *disk)
+@@ -301,6 +462,7 @@
+ 	struct request_queue *q = disk->queue;
+ 
+ 	if (q && q->request_fn) {
++		blk_protect_unregister(q);
+ 		elv_unregister_queue(q);
+ 
+ 		kobject_uevent(&q->kobj, KOBJ_REMOVE);
+diff -Naur a/Documentation/block/disk-protection.txt b/Documentation/block/disk-protection.txt
+--- a/Documentation/block/disk-protection.txt	1970-01-01 01:00:00.000000000 +0100
++++ b/Documentation/block/disk-protection.txt	2008-04-13 14:25:48.000000000 +0200
+@@ -0,0 +1,79 @@
++Hard disk protection
++====================
++
++
++Intro
++-----
++ATA/ATAPI-7 specifies the IDLE IMMEDIATE command with UNLOAD FEATURE.
++Issuing this command should cause the drive to switch to idle mode and
++unload disk heads. This feature is being used in modern laptops in
++conjunction with accelerometers and appropriate software to implement
++a shock protection facility. The idea is to stop all I/O operations on
++the internal hard drive and park its heads on the ramp when critical
++situations are anticipated. The desire to have such a feature
++available on GNU/Linux systems has been the original motivation to
++implement a generic disk parking interface in the Linux kernel.
++
++
++The interface
++-------------
++The interface works as follows: Writing an integer value to
++/sys/block/*/queue/protect will park the respective drive and freeze
++the block layer queue for the specified number of seconds. When the
++timeout expires and no further disk park request has been issued in
++the meantime, the queue is unfrozen and accumulated I/O operations are
++performed.
++
++IMPORTANT NOTE:
++Not all ATA drives implement IDLE IMMEDIATE with UNLOAD FEATURE and
++quite a few of those that do so, don't report this capability as
++described in the specs. When a disk park has been requested through
++sysfs as described above, the kernel will try to determine if the
++drive supports the UNLOAD FEATURE by default. The kernel will only
++rely on the IDLE IMMEDIATE with UNLOAD FEATURE command if it is
++convinced that this command is actually supported by the disk drive;
++otherwise, it will fall back to STANDBY IMMEDIATE. Resuming from the
++latter will take much longer and it is generally more likely to have a
++negative impact on the drive's lifetime due to the inclease of spin
++down and up cycles. If you want to use this interface in a shock
++protection framework and you know that your drive does indeed support
++the IDLE IMMEDIATE with UNLOAD FEATURE command despite not saying so,
++you can force the kernel to issue that command by doing the following
++on the command line:
++# echo -n unload > /sys/block/sda/queue/protect_method
++(replace sda by the drive identifier as appropriate).
++
++/sys/block/*/queue/protect_method accepts auto, unload and standby
++respectively. Reading from protect_method shows the available options
++surrounding the active one with brackets. When auto is active, this
++will change to whatever the kernel sees fit after the next disk park
++command has been issued.
++
++
++References
++----------
++
++There are several laptops from different brands featuring shock
++protection capabilities. As manufacturers have refused to support open
++source development of the required software components so far, Linux
++support for shock protection varies considerably between different
++hardware implementations. Ideally, this section should contain a list
++of poiters at different projects aiming at an implementation of shock
++protection on different systeems. Unfortunately, I only know of a
++single project which, although still considered experimental, is fit
++for use. Please feel free to add projects that have been the victims
++of my ignorance.
++
++- http://www.thinkwiki.org/wiki/HDAPS
++  See this page for information about Linux support of the hard disk
++  active protection syystem as implemented in IBM/Lenovo Thinkpads.
++
++
++CREDITS
++-------
++
++The patch to implement the interface described in this file has
++originally been published by Jon Escombe <lists-Xbpc2PeERmvQXOPxS62xeg@public.gmane.org>.
++
++
++05 Dec 2006, Elias Oltmanns <eo-oA28OIkTjSVZXbeN9DUtxg@public.gmane.org>
+diff -Naur a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
+--- a/drivers/ata/libata-scsi.c	2008-04-13 14:04:23.000000000 +0200
++++ b/drivers/ata/libata-scsi.c	2008-04-13 14:26:04.000000000 +0200
+@@ -831,7 +831,7 @@
+ 	 * prevent SCSI midlayer from automatically deferring
+ 	 * requests.
+ 	 */
+-	sdev->max_device_blocked = 1;
++	sdev->max_device_blocked = 2;
+ }
+ 
+ /**
+@@ -905,6 +905,38 @@
+ 	return 0;
+ }
+ 
++extern int scsi_protect_queue(struct request_queue *q, int unload);
++extern int scsi_unprotect_queue(struct request_queue *q);
++
++static int ata_scsi_issue_protect_fn(struct request_queue *q)
++{
++	struct scsi_device *sdev = q->queuedata;
++	struct ata_port *ap = ata_shost_to_port(sdev->host);
++	struct ata_device *dev = ata_scsi_find_dev(ap, sdev);
++	int unload = q->protect_method;
++	unsigned long flags;
++
++	if (!dev) {
++		printk(KERN_DEBUG "ata_scsi_issue_protect_fn(): Couldn't find ATA device to be parked.\n");
++		return -ENXIO;
++	}
++
++	if (unload == 2) {
++		unload = ata_id_has_unload(dev->id) ? 1 : 0;
++		spin_lock_irqsave(q->queue_lock, flags);
++		q->protect_method = unload;
++		spin_unlock_irqrestore(q->queue_lock, flags);
++	}
++
++	/* call scsi_protect_queue, requesting either unload or standby */
++	return scsi_protect_queue(q, unload);
++}
++
++static int ata_scsi_issue_unprotect_fn(struct request_queue *q)
++{
++	return scsi_unprotect_queue(q);
++}
++
+ /**
+  *	ata_scsi_slave_config - Set SCSI device attributes
+  *	@sdev: SCSI device to examine
+@@ -927,6 +959,10 @@
+ 
+ 	if (dev)
+ 		rc = ata_scsi_dev_config(sdev, dev);
++	blk_queue_issue_protect_fn(sdev->request_queue,
++		ata_scsi_issue_protect_fn);
++	blk_queue_issue_unprotect_fn(sdev->request_queue,
++		ata_scsi_issue_unprotect_fn);
+ 
+ 	return rc;
+ }
+@@ -3206,7 +3242,7 @@
+ 		 * Set host_blocked to 1 to prevent SCSI midlayer from
+ 		 * automatically deferring requests.
+ 		 */
+-		shost->max_host_blocked = 1;
++		shost->max_host_blocked = 2;
+ 
+ 		rc = scsi_add_host(ap->scsi_host, ap->host->dev);
+ 		if (rc)
+diff -Naur a/drivers/ide/ide-disk.c b/drivers/ide/ide-disk.c
+--- a/drivers/ide/ide-disk.c	2008-04-13 14:04:31.000000000 +0200
++++ b/drivers/ide/ide-disk.c	2008-04-13 14:25:48.000000000 +0200
+@@ -612,6 +612,148 @@
+ }
+ 
+ /*
++ * todo:
++ *  - we freeze the queue regardless of success and rely on the
++ *    ide_protect_queue function to thaw immediately if the command
++ *    failed (to be consistent with the libata handler)... should
++ *    we also inspect here?
++ */
++void ide_end_protect_rq(struct request *rq, int error)
++{
++	struct completion *waiting = rq->end_io_data;
++
++	rq->end_io_data = NULL;
++	/* spin lock already accquired */
++	if (!blk_queue_stopped(rq->q))
++		blk_stop_queue(rq->q);
++
++	complete(waiting);
++}
++
++int ide_unprotect_queue(struct request_queue *q)
++{
++	struct request rq;
++	unsigned long flags;
++	int pending, rc = 0;
++	ide_drive_t *drive = q->queuedata;
++	ide_task_t task;
++
++	if (!blk_queue_stopped(q))
++		return -EIO;
++
++	/* Are there any pending jobs on the queue? */
++	pending = ((q->rq.count[READ] > 0) || (q->rq.count[WRITE] > 0)) ? 1 : 0;
++
++	spin_lock_irqsave(q->queue_lock, flags);
++	blk_start_queue(q);
++	spin_unlock_irqrestore(q->queue_lock, flags);
++
++	/* The unload feature of the IDLE_IMMEDIATE command
++	   temporarily disables HD power management from spinning down
++	   the disk. Any other command will reenable HD pm, so, if
++	   there are no pending jobs on the queue, another
++	   CHECK_POWER_MODE1 command without the unload feature should do
++	   just fine. */
++	if (!pending) {
++		printk(KERN_DEBUG "ide_unprotect_queue(): No pending I/O, re-enabling power management..\n");
++		memset(&task, 0, sizeof(task));
++		task.tf.command = WIN_CHECKPOWERMODE1; /* CHECK_POWER_MODE1 */
++		task.tf_flags = IDE_TFLAG_TF | IDE_TFLAG_DEVICE;
++		task.data_phase = TASKFILE_NO_DATA;
++		ide_init_drive_cmd(&rq);
++		rq.cmd_type = REQ_TYPE_ATA_TASKFILE;
++		rq.special = &task;
++		rc = ide_do_drive_cmd(drive, &rq, ide_head_wait);
++	}
++
++	return rc;
++}
++
++int ide_protect_queue(struct request_queue *q, int unload)
++{
++	ide_drive_t *drive = q->queuedata;
++	struct request rq;
++	ide_task_t task;
++	int ret = 0;
++	DECLARE_COMPLETION(wait);
++
++	memset(&rq, 0, sizeof(rq));
++	memset(&task, 0, sizeof(task));
++
++	if (blk_queue_stopped(q))
++		return -EIO;
++
++	task.data_phase = TASKFILE_NO_DATA;
++	task.tf_flags = IDE_TFLAG_TF | IDE_TFLAG_DEVICE;
++	if (unload) {
++		task.tf.command = 0xe1;
++		task.tf.feature = 0x44;
++		task.tf.lbal = 0x4c;
++		task.tf.lbam = 0x4e;
++		task.tf.lbah = 0x55;
++	} else
++		task.tf.command = 0xe0;
++
++	/* Issue the park command & freeze */
++	ide_init_drive_cmd(&rq);
++
++	rq.cmd_type = REQ_TYPE_ATA_TASKFILE;
++	rq.special = &task;
++	rq.end_io_data = &wait;
++	rq.end_io = ide_end_protect_rq;
++
++	ret = ide_do_drive_cmd(drive, &rq, ide_next);
++	wait_for_completion(&wait);
++
++	if (ret) {
++		printk(KERN_DEBUG "ide_protect_queue(): Warning: head NOT parked!..\n");
++		ide_unprotect_queue(q);
++		return ret;
++	}
++
++	if (unload) {
++		if (task.tf.lbal == 0xc4)
++			printk(KERN_DEBUG "ide_protect_queue(): head parked..\n");
++		else {
++			/* error parking the head */
++			printk(KERN_DEBUG "ide_protect_queue(): head NOT parked!..\n");
++			ret = -EIO;
++			ide_unprotect_queue(q);
++		}
++	} else
++		printk(KERN_DEBUG "ide_protect_queue(): head park not requested, used standby!..\n");
++
++	return ret;
++}
++
++int idedisk_issue_protect_fn(struct request_queue *q)
++{
++	ide_drive_t *drive = q->queuedata;
++	int unload = q->protect_method;
++	unsigned long flags;
++
++	/*
++	 * Check capability of the device -
++	 *  - if "idle immediate with unload" is supported we use that, else
++	 *    we use "standby immediate" and live with spinning down the drive..
++	 *    (Word 84, bit 13 of IDENTIFY DEVICE data)
++	 */
++	if (unload == 2) {
++		unload = drive->id->cfsse & (1 << 13) ? 1 : 0;
++		spin_lock_irqsave(q->queue_lock, flags);
++		q->protect_method = unload;
++		spin_unlock_irqrestore(q->queue_lock, flags);
++	}
++
++	return ide_protect_queue(q, unload);
++}
++
++int idedisk_issue_unprotect_fn(struct request_queue *q)
++{
++	return ide_unprotect_queue(q);
++}
++
++/*
+  * This is tightly woven into the driver->do_special can not touch.
+  * DON'T do it again until a total personality rewrite is committed.
+  */
+@@ -877,6 +1019,9 @@
+ 		drive->wcache = 1;
+ 
+ 	write_cache(drive, 1);
++
++	blk_queue_issue_protect_fn(drive->queue, idedisk_issue_protect_fn);
++	blk_queue_issue_unprotect_fn(drive->queue, idedisk_issue_unprotect_fn);
+ }
+ 
+ static void ide_cacheflush_p(ide_drive_t *drive)
+diff -Naur a/drivers/ide/ide-io.c b/drivers/ide/ide-io.c
+--- a/drivers/ide/ide-io.c	2008-04-13 14:04:31.000000000 +0200
++++ b/drivers/ide/ide-io.c	2008-04-13 14:25:48.000000000 +0200
+@@ -1158,6 +1158,17 @@
+ 		}
+ 
+ 		/*
++		 * Don't accept a request when the queue is stopped (unless we
++		 * are resuming from suspend). Prevents existing queue entries
++		 * being processed after queue is stopped by the hard disk
++		 * protection mechanism...
++		 */
++		if (test_bit(QUEUE_FLAG_STOPPED, &drive->queue->queue_flags) && !blk_pm_resume_request(rq)) {
++			hwgroup->busy = 0;
++			break;
++		}
++
++		/*
+ 		 * Sanity: don't accept a request that isn't a PM request
+ 		 * if we are currently power managed. This is very important as
+ 		 * blk_stop_queue() doesn't prevent the elv_next_request()
+@@ -1651,6 +1662,9 @@
+ 		where = ELEVATOR_INSERT_FRONT;
+ 		rq->cmd_flags |= REQ_PREEMPT;
+ 	}
++	if (action == ide_next)
++		where = ELEVATOR_INSERT_FRONT;
++
+ 	__elv_add_request(drive->queue, rq, where, 0);
+ 	ide_do_request(hwgroup, IDE_NO_IRQ);
+ 	spin_unlock_irqrestore(&ide_lock, flags);
+diff -Naur a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
+--- a/drivers/scsi/scsi_lib.c	2008-04-13 14:04:50.000000000 +0200
++++ b/drivers/scsi/scsi_lib.c	2008-04-13 14:25:48.000000000 +0200
+@@ -2248,7 +2248,13 @@
+ int
+ scsi_device_quiesce(struct scsi_device *sdev)
+ {
++	int i;
+ 	int err = scsi_device_set_state(sdev, SDEV_QUIESCE);
++	for (i = 0; err && (sdev->sdev_state == SDEV_BLOCK) && (i < 100);
++	     i++) {
++		msleep_interruptible(200);
++		err = scsi_device_set_state(sdev, SDEV_QUIESCE);
++	}
+ 	if (err)
+ 		return err;
+ 
+@@ -2496,3 +2502,168 @@
+ 	kunmap_atomic(virt, KM_BIO_SRC_IRQ);
+ }
+ EXPORT_SYMBOL(scsi_kunmap_atomic_sg);
++
++/*
++ * Structure required for synchronous io completion after queue freezing
++ */
++struct scsi_protect_io_context_sync {
++	struct scsi_device *sdev;
++	int result;
++	char *sense;
++	struct completion *waiting;
++};
++
++/*
++ * scsi_protect_wait_done()
++ * Command completion handler for scsi_protect_queue().
++ *
++ * Unable to call scsi_internal_device_block() as
++ * scsi_end_request() already has the spinlock. So,
++ * we put the necessary functionality inline.
++ *
++ * todo:
++ *  - we block the queue regardless of success and rely on the
++ *    scsi_protect_queue function to unblock if the command
++ *    failed... should we also inspect here?
++ */
++static void scsi_protect_wait_done(void *data, char *sense, int result, int resid)
++{
++	struct scsi_protect_io_context_sync *siocs = data;
++	struct completion *waiting = siocs->waiting;
++	struct request_queue *q = siocs->sdev->request_queue;
++
++	siocs->waiting = NULL;
++	siocs->result = result;
++	memcpy(siocs->sense, sense, SCSI_SENSE_BUFFERSIZE);
++
++	if (!scsi_device_set_state(siocs->sdev, SDEV_BLOCK))
++		blk_stop_queue(q);
++
++	complete(waiting);
++}
++
++/*
++ * scsi_unprotect_queue()
++ *  - release the queue that was previously blocked
++ */
++int scsi_unprotect_queue(struct request_queue *q)
++{
++	struct scsi_device *sdev = q->queuedata;
++	int rc = 0, pending = 0;
++	u8 scsi_cmd[MAX_COMMAND_SIZE];
++	struct scsi_sense_hdr sshdr;
++
++	if (sdev->sdev_state != SDEV_BLOCK)
++		return -ENXIO;
++
++	/* Are there any pending jobs on the queue? */
++	pending = ((q->rq.count[READ] > 0) || (q->rq.count[WRITE] > 0)) ? 1 : 0;
++
++	rc = scsi_internal_device_unblock(sdev);
++	if (rc)
++		return rc;
++
++	if (!pending) {
++		printk(KERN_DEBUG "scsi_unprotect_queue(): No pending I/O, re-enabling power management..\n");
++
++		memset(scsi_cmd, 0, sizeof(scsi_cmd));
++		scsi_cmd[0]  = ATA_16;
++		scsi_cmd[1]  = (3 << 1); /* Non-data */
++		/* scsi_cmd[2] is already 0 -- no off.line, cc, or data xfer */
++		scsi_cmd[14] = 0xe5; /* CHECK_POWER_MODE1 */
++
++		/* Good values for timeout and retries?  Values below
++   		   from scsi_ioctl_send_command() for default case... */
++		if (scsi_execute_req(sdev, scsi_cmd, DMA_NONE, NULL, 0, &sshdr,
++		   		     (10*HZ), 5))
++			rc = -EIO;
++	}
++	return rc;
++}
++EXPORT_SYMBOL_GPL(scsi_unprotect_queue);
++
++/*
++ * scsi_protect_queue()
++ *  - build and issue the park/standby command..
++ *  - queue is blocked during command completion handler
++ */
++int scsi_protect_queue(struct request_queue *q, int unload)
++{
++	struct scsi_protect_io_context_sync siocs;
++	struct scsi_device *sdev = q->queuedata;
++	int rc = 0;
++	u8 args[7];
++	u8 scsi_cmd[MAX_COMMAND_SIZE];
++	unsigned char sense[SCSI_SENSE_BUFFERSIZE];
++	unsigned char *desc;
++	DECLARE_COMPLETION_ONSTACK(wait);
++
++	if (sdev->sdev_state != SDEV_RUNNING)
++		return -ENXIO;
++
++	memset(args, 0, sizeof(args));
++	memset(sense, 0, sizeof(sense));
++
++	if (unload) {
++		args[0] = 0xe1;
++		args[1] = 0x44;
++		args[3] = 0x4c;
++		args[4] = 0x4e;
++		args[5] = 0x55;
++	} else
++		args[0] = 0xe0;
++
++	memset(scsi_cmd, 0, sizeof(scsi_cmd));
++	scsi_cmd[0]  = ATA_16;
++	scsi_cmd[1]  = (3 << 1); /* Non-data */
++	scsi_cmd[2]  = 0x20;     /* no off.line, or data xfer, request cc */
++	scsi_cmd[4]  = args[1];
++	scsi_cmd[6]  = args[2];
++	scsi_cmd[8]  = args[3];
++	scsi_cmd[10] = args[4];
++	scsi_cmd[12] = args[5];
++	scsi_cmd[14] = args[0];
++	siocs.sdev = sdev;
++	siocs.sense = sense;
++	siocs.waiting = &wait;
++
++	rc = scsi_execute_async(sdev, scsi_cmd, COMMAND_SIZE(scsi_cmd[0]),
++				DMA_NONE, NULL, 0, 0, (10*HZ), 5,
++				&siocs, &scsi_protect_wait_done, GFP_NOWAIT);
++	if (rc)
++		goto out;
++	wait_for_completion(&wait);
++
++	if (siocs.result != ((DRIVER_SENSE << 24) + SAM_STAT_CHECK_CONDITION)) {
++		printk(KERN_DEBUG "scsi_protect_queue(): head NOT parked!..\n");
++		scsi_unprotect_queue(q);		/* just in case we still managed to block */
++		rc = -EIO;
++		goto out;
++	}
++
++	desc = sense + 8;
++
++	/* Retrieve data from check condition */
++	args[1] = desc[3];
++	args[2] = desc[5];
++	args[3] = desc[7];
++	args[4] = desc[9];
++	args[5] = desc[11];
++	args[0] = desc[13];
++
++	if (unload) {
++		if (args[3] == 0xc4)
++			printk(KERN_DEBUG "scsi_protect_queue(): head parked..\n");
++		else {
++			/* error parking the head */
++			printk(KERN_DEBUG "scsi_protect_queue(): head NOT parked!..\n");
++			rc = -EIO;
++			scsi_unprotect_queue(q);
++		}
++	} else
++		printk(KERN_DEBUG "scsi_protect_queue(): head park not requested, used standby!..\n");
++
++out:
++	return rc;
++}
++EXPORT_SYMBOL_GPL(scsi_protect_queue);
+diff -Naur a/include/linux/ata.h b/include/linux/ata.h
+--- a/include/linux/ata.h	2008-04-13 14:04:53.000000000 +0200
++++ b/include/linux/ata.h	2008-04-13 14:25:48.000000000 +0200
+@@ -459,6 +459,18 @@
+ 
+ #define ata_id_cdb_intr(id)	(((id)[0] & 0x60) == 0x20)
+ 
++static inline int ata_id_has_unload(const u16 *id)
++{
++	/* ATA-7 specifies two places to indicate unload feature support.
++	 * Since I don't really understand the difference, I'll just check
++	 * both and only return zero if none of them indicates otherwise. */
++	if ((id[84] & 0xC000) == 0x4000 && id[84] & (1 << 13))
++		return id[84] & (1 << 13);
++	if ((id[87] & 0xC000) == 0x4000)
++		return id[87] & (1 << 13);
++	return 0;
++}
++
+ static inline bool ata_id_has_hipm(const u16 *id)
+ {
+ 	u16 val = id[76];
+diff -Naur a/include/linux/blkdev.h b/include/linux/blkdev.h
+--- a/include/linux/blkdev.h	2008-04-13 14:04:53.000000000 +0200
++++ b/include/linux/blkdev.h	2008-04-13 14:25:48.000000000 +0200
+@@ -260,6 +260,8 @@
+ typedef void (prepare_flush_fn) (struct request_queue *, struct request *);
+ typedef void (softirq_done_fn)(struct request *);
+ typedef int (dma_drain_needed_fn)(struct request *);
++typedef int (issue_protect_fn) (struct request_queue *);
++typedef int (issue_unprotect_fn) (struct request_queue *);
+ 
+ enum blk_queue_state {
+ 	Queue_down,
+@@ -297,6 +299,8 @@
+ 	prepare_flush_fn	*prepare_flush_fn;
+ 	softirq_done_fn		*softirq_done_fn;
+ 	dma_drain_needed_fn	*dma_drain_needed;
++	issue_protect_fn	*issue_protect_fn;
++	issue_unprotect_fn	*issue_unprotect_fn;
+ 
+ 	/*
+ 	 * Dispatch queue sorting
+@@ -312,6 +316,14 @@
+ 	unsigned long		unplug_delay;	/* After this many jiffies */
+ 	struct work_struct	unplug_work;
+ 
++	/*
++	 * Auto-unfreeze state
++	 */
++	struct timer_list	unfreeze_timer;
++	int			max_unfreeze;	/* At most this many seconds */
++	struct work_struct	unfreeze_work;
++	int			protect_method;
++
+ 	struct backing_dev_info	backing_dev_info;
+ 
+ 	/*
+@@ -718,6 +730,8 @@
+ extern unsigned blk_ordered_cur_seq(struct request_queue *);
+ extern unsigned blk_ordered_req_seq(struct request *);
+ extern void blk_ordered_complete_seq(struct request_queue *, unsigned, int);
++extern void blk_queue_issue_protect_fn(struct request_queue *, issue_protect_fn *);
++extern void blk_queue_issue_unprotect_fn(struct request_queue *, issue_unprotect_fn *);
+ 
+ extern int blk_rq_map_sg(struct request_queue *, struct request *, struct scatterlist *);
+ extern void blk_dump_rq_flags(struct request *, char *);
+diff -Naur a/include/linux/ide.h b/include/linux/ide.h
+--- a/include/linux/ide.h	2008-04-13 14:04:53.000000000 +0200
++++ b/include/linux/ide.h	2008-04-13 14:25:48.000000000 +0200
+@@ -837,6 +837,7 @@
+  */
+ typedef enum {
+ 	ide_wait,	/* insert rq at end of list, and wait for it */
++	ide_next,	/* insert rq immediately after current request */
+ 	ide_preempt,	/* insert rq in front of current request */
+ 	ide_head_wait,	/* insert rq in front of current request and wait for it */
+ 	ide_end		/* insert rq at end of list, but don't wait for it */