Merge tag 'for-linus-20130331' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:
 "Alright, this time from 10K up in the air.

  Collection of fixes that have been queued up since the merge window
  opened, hence postponed until later in the cycle.  The pull request
  contains:

   - A bunch of fixes for the xen blk front/back driver.

   - A round of fixes for the new IBM RamSan driver, fixing various
     nasty issues.

   - Fixes for multiple drives from Wei Yongjun, bad handling of return
     values and wrong pointer math.

   - A fix for loop properly killing partitions when being detached."

* tag 'for-linus-20130331' of git://git.kernel.dk/linux-block: (25 commits)
  mg_disk: fix error return code in mg_probe()
  rsxx: remove unused variable
  rsxx: enable error return of rsxx_eeh_save_issued_dmas()
  block: removes dynamic allocation on stack
  Block: blk-flush: Fixed indent code style
  cciss: fix invalid use of sizeof in cciss_find_cfgtables()
  loop: cleanup partitions when detaching loop device
  loop: fix error return code in loop_add()
  mtip32xx: fix error return code in mtip_pci_probe()
  xen-blkfront: remove frame list from blk_shadow
  xen-blkfront: pre-allocate pages for requests
  xen-blkback: don't store dev_bus_addr
  xen-blkfront: switch from llist to list
  xen-blkback: fix foreach_grant_safe to handle empty lists
  xen-blkfront: replace kmalloc and then memcpy with kmemdup
  xen-blkback: fix dispatch_rw_block_io() error path
  rsxx: fix missing unlock on error return in rsxx_eeh_remap_dmas()
  Adding in EEH support to the IBM FlashSystem 70/80 device driver
  block: IBM RamSan 70/80 error message bug fix.
  block: IBM RamSan 70/80 branding changes.
  ...

MAINTAINERS

@@ -3242,6 +3242,12 @@ F:	Documentation/firmware_class/
 F:	drivers/base/firmware*.c
 F:	include/linux/firmware.h
 
+FLASHSYSTEM DRIVER (IBM FlashSystem 70/80 PCI SSD Flash Card)
+M:	Joshua Morris <josh.h.morris@us.ibm.com>
+M:	Philip Kelleher <pjk1939@linux.vnet.ibm.com>
+S:	Maintained
+F:	drivers/block/rsxx/
+
 FLOPPY DRIVER
 M:	Jiri Kosina <jkosina@suse.cz>
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/jikos/floppy.git
@@ -6551,12 +6557,6 @@ S:	Maintained
 F:	Documentation/blockdev/ramdisk.txt
 F:	drivers/block/brd.c
 
-RAMSAM DRIVER (IBM RamSan 70/80 PCI SSD Flash Card)
-M:	Joshua Morris <josh.h.morris@us.ibm.com>
-M:	Philip Kelleher <pjk1939@linux.vnet.ibm.com>
-S:	Maintained
-F:	drivers/block/rsxx/
-
 RANDOM NUMBER DRIVER
 M:	Theodore Ts'o" <tytso@mit.edu>
 S:	Maintained

block/blk-flush.c

@@ -444,7 +444,7 @@ int blkdev_issue_flush(struct block_device *bdev, gfp_t gfp_mask,
 	 * copied from blk_rq_pos(rq).
 	 */
 	if (error_sector)
-               *error_sector = bio->bi_sector;
+		*error_sector = bio->bi_sector;
 
 	if (!bio_flagged(bio, BIO_UPTODATE))
 		ret = -EIO;

block/partition-generic.c

@@ -257,6 +257,7 @@ void delete_partition(struct gendisk *disk, int partno)
 
 	hd_struct_put(part);
 }
+EXPORT_SYMBOL(delete_partition);
 
 static ssize_t whole_disk_show(struct device *dev,
 			       struct device_attribute *attr, char *buf)

drivers/block/Kconfig

@@ -532,11 +532,11 @@ config BLK_DEV_RBD
 	  If unsure, say N.
 
 config BLK_DEV_RSXX
-	tristate "RamSam PCIe Flash SSD Device Driver"
+	tristate "IBM FlashSystem 70/80 PCIe SSD Device Driver"
 	depends on PCI
 	help
 	  Device driver for IBM's high speed PCIe SSD
-	  storage devices: RamSan-70 and RamSan-80.
+	  storage devices: FlashSystem-70 and FlashSystem-80.
 
 	  To compile this driver as a module, choose M here: the
 	  module will be called rsxx.

drivers/block/cciss.c

@@ -4206,7 +4206,7 @@ static int cciss_find_cfgtables(ctlr_info_t *h)
 	if (rc)
 		return rc;
 	h->cfgtable = remap_pci_mem(pci_resource_start(h->pdev,
-		cfg_base_addr_index) + cfg_offset, sizeof(h->cfgtable));
+		cfg_base_addr_index) + cfg_offset, sizeof(*h->cfgtable));
 	if (!h->cfgtable)
 		return -ENOMEM;
 	rc = write_driver_ver_to_cfgtable(h->cfgtable);

drivers/block/loop.c

@@ -1044,12 +1044,29 @@ static int loop_clr_fd(struct loop_device *lo)
 	lo->lo_state = Lo_unbound;
 	/* This is safe: open() is still holding a reference. */
 	module_put(THIS_MODULE);
-	if (lo->lo_flags & LO_FLAGS_PARTSCAN && bdev)
-		ioctl_by_bdev(bdev, BLKRRPART, 0);
 	lo->lo_flags = 0;
 	if (!part_shift)
 		lo->lo_disk->flags |= GENHD_FL_NO_PART_SCAN;
 	mutex_unlock(&lo->lo_ctl_mutex);
+
+	/*
+	 * Remove all partitions, since BLKRRPART won't remove user
+	 * added partitions when max_part=0
+	 */
+	if (bdev) {
+		struct disk_part_iter piter;
+		struct hd_struct *part;
+
+		mutex_lock_nested(&bdev->bd_mutex, 1);
+		invalidate_partition(bdev->bd_disk, 0);
+		disk_part_iter_init(&piter, bdev->bd_disk,
+					DISK_PITER_INCL_EMPTY);
+		while ((part = disk_part_iter_next(&piter)))
+			delete_partition(bdev->bd_disk, part->partno);
+		disk_part_iter_exit(&piter);
+		mutex_unlock(&bdev->bd_mutex);
+	}
+
 	/*
 	 * Need not hold lo_ctl_mutex to fput backing file.
 	 * Calling fput holding lo_ctl_mutex triggers a circular
@@ -1623,6 +1640,7 @@ static int loop_add(struct loop_device **l, int i)
 		goto out_free_dev;
 	i = err;
 
+	err = -ENOMEM;
 	lo->lo_queue = blk_alloc_queue(GFP_KERNEL);
 	if (!lo->lo_queue)
 		goto out_free_dev;

drivers/block/mg_disk.c

@@ -890,8 +890,10 @@ static int mg_probe(struct platform_device *plat_dev)
 	gpio_direction_output(host->rst, 1);
 
 	/* reset out pin */
-	if (!(prv_data->dev_attr & MG_DEV_MASK))
+	if (!(prv_data->dev_attr & MG_DEV_MASK)) {
+		err = -EINVAL;
 		goto probe_err_3a;
+	}
 
 	if (prv_data->dev_attr != MG_BOOT_DEV) {
 		rsc = platform_get_resource_byname(plat_dev, IORESOURCE_IO,

drivers/block/mtip32xx/mtip32xx.c

@@ -4224,6 +4224,7 @@ static int mtip_pci_probe(struct pci_dev *pdev,
 	dd->isr_workq = create_workqueue(dd->workq_name);
 	if (!dd->isr_workq) {
 		dev_warn(&pdev->dev, "Can't create wq %d\n", dd->instance);
+		rv = -ENOMEM;
 		goto block_initialize_err;
 	}
 
@@ -4282,7 +4283,8 @@ static int mtip_pci_probe(struct pci_dev *pdev,
 	INIT_WORK(&dd->work[7].work, mtip_workq_sdbf7);
 
 	pci_set_master(pdev);
-	if (pci_enable_msi(pdev)) {
+	rv = pci_enable_msi(pdev);
+	if (rv) {
 		dev_warn(&pdev->dev,
 			"Unable to enable MSI interrupt.\n");
 		goto block_initialize_err;

drivers/block/rsxx/Makefile

@@ -1,2 +1,2 @@
 obj-$(CONFIG_BLK_DEV_RSXX) += rsxx.o
-rsxx-y := config.o core.o cregs.o dev.o dma.o
+rsxx-objs := config.o core.o cregs.o dev.o dma.o

drivers/block/rsxx/config.c

@@ -29,15 +29,13 @@
 #include "rsxx_priv.h"
 #include "rsxx_cfg.h"
 
-static void initialize_config(void *config)
+static void initialize_config(struct rsxx_card_cfg *cfg)
 {
-	struct rsxx_card_cfg *cfg = config;
-
 	cfg->hdr.version = RSXX_CFG_VERSION;
 
 	cfg->data.block_size        = RSXX_HW_BLK_SIZE;
 	cfg->data.stripe_size       = RSXX_HW_BLK_SIZE;
-	cfg->data.vendor_id         = RSXX_VENDOR_ID_TMS_IBM;
+	cfg->data.vendor_id         = RSXX_VENDOR_ID_IBM;
 	cfg->data.cache_order       = (-1);
 	cfg->data.intr_coal.mode    = RSXX_INTR_COAL_DISABLED;
 	cfg->data.intr_coal.count   = 0;
@@ -181,7 +179,7 @@ int rsxx_load_config(struct rsxx_cardinfo *card)
 	} else {
 		dev_info(CARD_TO_DEV(card),
 			"Initializing card configuration.\n");
-		initialize_config(card);
+		initialize_config(&card->config);
 		st = rsxx_save_config(card);
 		if (st)
 			return st;

drivers/block/rsxx/core.c

@@ -30,6 +30,7 @@
 #include <linux/reboot.h>
 #include <linux/slab.h>
 #include <linux/bitops.h>
+#include <linux/delay.h>
 
 #include <linux/genhd.h>
 #include <linux/idr.h>
@@ -39,8 +40,8 @@
 
 #define NO_LEGACY 0
 
-MODULE_DESCRIPTION("IBM RamSan PCIe Flash SSD Device Driver");
-MODULE_AUTHOR("IBM <support@ramsan.com>");
+MODULE_DESCRIPTION("IBM FlashSystem 70/80 PCIe SSD Device Driver");
+MODULE_AUTHOR("Joshua Morris/Philip Kelleher, IBM");
 MODULE_LICENSE("GPL");
 MODULE_VERSION(DRIVER_VERSION);
 
@@ -52,6 +53,13 @@ static DEFINE_IDA(rsxx_disk_ida);
 static DEFINE_SPINLOCK(rsxx_ida_lock);
 
 /*----------------- Interrupt Control & Handling -------------------*/
+
+static void rsxx_mask_interrupts(struct rsxx_cardinfo *card)
+{
+	card->isr_mask = 0;
+	card->ier_mask = 0;
+}
+
 static void __enable_intr(unsigned int *mask, unsigned int intr)
 {
 	*mask |= intr;
@@ -71,7 +79,8 @@ static void __disable_intr(unsigned int *mask, unsigned int intr)
  */
 void rsxx_enable_ier(struct rsxx_cardinfo *card, unsigned int intr)
 {
-	if (unlikely(card->halt))
+	if (unlikely(card->halt) ||
+	    unlikely(card->eeh_state))
 		return;
 
 	__enable_intr(&card->ier_mask, intr);
@@ -80,6 +89,9 @@ void rsxx_enable_ier(struct rsxx_cardinfo *card, unsigned int intr)
 
 void rsxx_disable_ier(struct rsxx_cardinfo *card, unsigned int intr)
 {
+	if (unlikely(card->eeh_state))
+		return;
+
 	__disable_intr(&card->ier_mask, intr);
 	iowrite32(card->ier_mask, card->regmap + IER);
 }
@@ -87,7 +99,8 @@ void rsxx_disable_ier(struct rsxx_cardinfo *card, unsigned int intr)
 void rsxx_enable_ier_and_isr(struct rsxx_cardinfo *card,
 				 unsigned int intr)
 {
-	if (unlikely(card->halt))
+	if (unlikely(card->halt) ||
+	    unlikely(card->eeh_state))
 		return;
 
 	__enable_intr(&card->isr_mask, intr);
@@ -97,6 +110,9 @@ void rsxx_enable_ier_and_isr(struct rsxx_cardinfo *card,
 void rsxx_disable_ier_and_isr(struct rsxx_cardinfo *card,
 				  unsigned int intr)
 {
+	if (unlikely(card->eeh_state))
+		return;
+
 	__disable_intr(&card->isr_mask, intr);
 	__disable_intr(&card->ier_mask, intr);
 	iowrite32(card->ier_mask, card->regmap + IER);
@@ -115,6 +131,9 @@ static irqreturn_t rsxx_isr(int irq, void *pdata)
 	do {
 		reread_isr = 0;
 
+		if (unlikely(card->eeh_state))
+			break;
+
 		isr = ioread32(card->regmap + ISR);
 		if (isr == 0xffffffff) {
 			/*
@@ -161,9 +180,9 @@ static irqreturn_t rsxx_isr(int irq, void *pdata)
 }
 
 /*----------------- Card Event Handler -------------------*/
-static char *rsxx_card_state_to_str(unsigned int state)
+static const char * const rsxx_card_state_to_str(unsigned int state)
 {
-	static char *state_strings[] = {
+	static const char * const state_strings[] = {
 		"Unknown", "Shutdown", "Starting", "Formatting",
 		"Uninitialized", "Good", "Shutting Down",
 		"Fault", "Read Only Fault", "dStroying"
@@ -304,6 +323,192 @@ static int card_shutdown(struct rsxx_cardinfo *card)
 	return 0;
 }
 
+static int rsxx_eeh_frozen(struct pci_dev *dev)
+{
+	struct rsxx_cardinfo *card = pci_get_drvdata(dev);
+	int i;
+	int st;
+
+	dev_warn(&dev->dev, "IBM FlashSystem PCI: preparing for slot reset.\n");
+
+	card->eeh_state = 1;
+	rsxx_mask_interrupts(card);
+
+	/*
+	 * We need to guarantee that the write for eeh_state and masking
+	 * interrupts does not become reordered. This will prevent a possible
+	 * race condition with the EEH code.
+	 */
+	wmb();
+
+	pci_disable_device(dev);
+
+	st = rsxx_eeh_save_issued_dmas(card);
+	if (st)
+		return st;
+
+	rsxx_eeh_save_issued_creg(card);
+
+	for (i = 0; i < card->n_targets; i++) {
+		if (card->ctrl[i].status.buf)
+			pci_free_consistent(card->dev, STATUS_BUFFER_SIZE8,
+					    card->ctrl[i].status.buf,
+					    card->ctrl[i].status.dma_addr);
+		if (card->ctrl[i].cmd.buf)
+			pci_free_consistent(card->dev, COMMAND_BUFFER_SIZE8,
+					    card->ctrl[i].cmd.buf,
+					    card->ctrl[i].cmd.dma_addr);
+	}
+
+	return 0;
+}
+
+static void rsxx_eeh_failure(struct pci_dev *dev)
+{
+	struct rsxx_cardinfo *card = pci_get_drvdata(dev);
+	int i;
+
+	dev_err(&dev->dev, "IBM FlashSystem PCI: disabling failed card.\n");
+
+	card->eeh_state = 1;
+
+	for (i = 0; i < card->n_targets; i++)
+		del_timer_sync(&card->ctrl[i].activity_timer);
+
+	rsxx_eeh_cancel_dmas(card);
+}
+
+static int rsxx_eeh_fifo_flush_poll(struct rsxx_cardinfo *card)
+{
+	unsigned int status;
+	int iter = 0;
+
+	/* We need to wait for the hardware to reset */
+	while (iter++ < 10) {
+		status = ioread32(card->regmap + PCI_RECONFIG);
+
+		if (status & RSXX_FLUSH_BUSY) {
+			ssleep(1);
+			continue;
+		}
+
+		if (status & RSXX_FLUSH_TIMEOUT)
+			dev_warn(CARD_TO_DEV(card), "HW: flash controller timeout\n");
+		return 0;
+	}
+
+	/* Hardware failed resetting itself. */
+	return -1;
+}
+
+static pci_ers_result_t rsxx_error_detected(struct pci_dev *dev,
+					    enum pci_channel_state error)
+{
+	int st;
+
+	if (dev->revision < RSXX_EEH_SUPPORT)
+		return PCI_ERS_RESULT_NONE;
+
+	if (error == pci_channel_io_perm_failure) {
+		rsxx_eeh_failure(dev);
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+
+	st = rsxx_eeh_frozen(dev);
+	if (st) {
+		dev_err(&dev->dev, "Slot reset setup failed\n");
+		rsxx_eeh_failure(dev);
+		return PCI_ERS_RESULT_DISCONNECT;
+	}
+
+	return PCI_ERS_RESULT_NEED_RESET;
+}
+
+static pci_ers_result_t rsxx_slot_reset(struct pci_dev *dev)
+{
+	struct rsxx_cardinfo *card = pci_get_drvdata(dev);
+	unsigned long flags;
+	int i;
+	int st;
+
+	dev_warn(&dev->dev,
+		"IBM FlashSystem PCI: recovering from slot reset.\n");
+
+	st = pci_enable_device(dev);
+	if (st)
+		goto failed_hw_setup;
+
+	pci_set_master(dev);
+
+	st = rsxx_eeh_fifo_flush_poll(card);
+	if (st)
+		goto failed_hw_setup;
+
+	rsxx_dma_queue_reset(card);
+
+	for (i = 0; i < card->n_targets; i++) {
+		st = rsxx_hw_buffers_init(dev, &card->ctrl[i]);
+		if (st)
+			goto failed_hw_buffers_init;
+	}
+
+	if (card->config_valid)
+		rsxx_dma_configure(card);
+
+	/* Clears the ISR register from spurious interrupts */
+	st = ioread32(card->regmap + ISR);
+
+	card->eeh_state = 0;
+
+	st = rsxx_eeh_remap_dmas(card);
+	if (st)
+		goto failed_remap_dmas;
+
+	spin_lock_irqsave(&card->irq_lock, flags);
+	if (card->n_targets & RSXX_MAX_TARGETS)
+		rsxx_enable_ier_and_isr(card, CR_INTR_ALL_G);
+	else
+		rsxx_enable_ier_and_isr(card, CR_INTR_ALL_C);
+	spin_unlock_irqrestore(&card->irq_lock, flags);
+
+	rsxx_kick_creg_queue(card);
+
+	for (i = 0; i < card->n_targets; i++) {
+		spin_lock(&card->ctrl[i].queue_lock);
+		if (list_empty(&card->ctrl[i].queue)) {
+			spin_unlock(&card->ctrl[i].queue_lock);
+			continue;
+		}
+		spin_unlock(&card->ctrl[i].queue_lock);
+
+		queue_work(card->ctrl[i].issue_wq,
+				&card->ctrl[i].issue_dma_work);
+	}
+
+	dev_info(&dev->dev, "IBM FlashSystem PCI: recovery complete.\n");
+
+	return PCI_ERS_RESULT_RECOVERED;
+
+failed_hw_buffers_init:
+failed_remap_dmas:
+	for (i = 0; i < card->n_targets; i++) {
+		if (card->ctrl[i].status.buf)
+			pci_free_consistent(card->dev,
+					STATUS_BUFFER_SIZE8,
+					card->ctrl[i].status.buf,
+					card->ctrl[i].status.dma_addr);
+		if (card->ctrl[i].cmd.buf)
+			pci_free_consistent(card->dev,
+					COMMAND_BUFFER_SIZE8,
+					card->ctrl[i].cmd.buf,
+					card->ctrl[i].cmd.dma_addr);
+	}
+failed_hw_setup:
+	rsxx_eeh_failure(dev);
+	return PCI_ERS_RESULT_DISCONNECT;
+
+}
+
 /*----------------- Driver Initialization & Setup -------------------*/
 /* Returns:   0 if the driver is compatible with the device
 	     -1 if the driver is NOT compatible with the device */
@@ -383,6 +588,7 @@ static int rsxx_pci_probe(struct pci_dev *dev,
 
 	spin_lock_init(&card->irq_lock);
 	card->halt = 0;
+	card->eeh_state = 0;
 
 	spin_lock_irq(&card->irq_lock);
 	rsxx_disable_ier_and_isr(card, CR_INTR_ALL);
@@ -538,9 +744,6 @@ static void rsxx_pci_remove(struct pci_dev *dev)
 	rsxx_disable_ier_and_isr(card, CR_INTR_EVENT);
 	spin_unlock_irqrestore(&card->irq_lock, flags);
 
-	/* Prevent work_structs from re-queuing themselves. */
-	card->halt = 1;
-
 	cancel_work_sync(&card->event_work);
 
 	rsxx_destroy_dev(card);
@@ -549,6 +752,10 @@ static void rsxx_pci_remove(struct pci_dev *dev)
 	spin_lock_irqsave(&card->irq_lock, flags);
 	rsxx_disable_ier_and_isr(card, CR_INTR_ALL);
 	spin_unlock_irqrestore(&card->irq_lock, flags);
+
+	/* Prevent work_structs from re-queuing themselves. */
+	card->halt = 1;
+
 	free_irq(dev->irq, card);
 
 	if (!force_legacy)
@@ -592,11 +799,14 @@ static void rsxx_pci_shutdown(struct pci_dev *dev)
 	card_shutdown(card);
 }
 
+static const struct pci_error_handlers rsxx_err_handler = {
+	.error_detected = rsxx_error_detected,
+	.slot_reset     = rsxx_slot_reset,
+};
+
 static DEFINE_PCI_DEVICE_TABLE(rsxx_pci_ids) = {
-	{PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS70_FLASH)},
-	{PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS70D_FLASH)},
-	{PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS80_FLASH)},
-	{PCI_DEVICE(PCI_VENDOR_ID_TMS_IBM, PCI_DEVICE_ID_RS81_FLASH)},
+	{PCI_DEVICE(PCI_VENDOR_ID_IBM, PCI_DEVICE_ID_FS70_FLASH)},
+	{PCI_DEVICE(PCI_VENDOR_ID_IBM, PCI_DEVICE_ID_FS80_FLASH)},
 	{0,},
 };
 
@@ -609,6 +819,7 @@ static struct pci_driver rsxx_pci_driver = {
 	.remove		= rsxx_pci_remove,
 	.suspend	= rsxx_pci_suspend,
 	.shutdown	= rsxx_pci_shutdown,
+	.err_handler    = &rsxx_err_handler,
 };
 
 static int __init rsxx_core_init(void)

drivers/block/rsxx/cregs.c

@@ -58,7 +58,7 @@ static struct kmem_cache *creg_cmd_pool;
 #error Unknown endianess!!! Aborting...
 #endif
 
-static void copy_to_creg_data(struct rsxx_cardinfo *card,
+static int copy_to_creg_data(struct rsxx_cardinfo *card,
 			      int cnt8,
 			      void *buf,
 			      unsigned int stream)
@@ -66,6 +66,9 @@ static void copy_to_creg_data(struct rsxx_cardinfo *card,
 	int i = 0;
 	u32 *data = buf;
 
+	if (unlikely(card->eeh_state))
+		return -EIO;
+
 	for (i = 0; cnt8 > 0; i++, cnt8 -= 4) {
 		/*
 		 * Firmware implementation makes it necessary to byte swap on
@@ -76,10 +79,12 @@ static void copy_to_creg_data(struct rsxx_cardinfo *card,
 		else
 			iowrite32(data[i], card->regmap + CREG_DATA(i));
 	}
+
+	return 0;
 }
 
 
-static void copy_from_creg_data(struct rsxx_cardinfo *card,
+static int copy_from_creg_data(struct rsxx_cardinfo *card,
 				int cnt8,
 				void *buf,
 				unsigned int stream)
@@ -87,6 +92,9 @@ static void copy_from_creg_data(struct rsxx_cardinfo *card,
 	int i = 0;
 	u32 *data = buf;
 
+	if (unlikely(card->eeh_state))
+		return -EIO;
+
 	for (i = 0; cnt8 > 0; i++, cnt8 -= 4) {
 		/*
 		 * Firmware implementation makes it necessary to byte swap on
@@ -97,41 +105,31 @@ static void copy_from_creg_data(struct rsxx_cardinfo *card,
 		else
 			data[i] = ioread32(card->regmap + CREG_DATA(i));
 	}
-}
-
-static struct creg_cmd *pop_active_cmd(struct rsxx_cardinfo *card)
-{
-	struct creg_cmd *cmd;
 
-	/*
-	 * Spin lock is needed because this can be called in atomic/interrupt
-	 * context.
-	 */
-	spin_lock_bh(&card->creg_ctrl.lock);
-	cmd = card->creg_ctrl.active_cmd;
-	card->creg_ctrl.active_cmd = NULL;
-	spin_unlock_bh(&card->creg_ctrl.lock);
-
-	return cmd;
+	return 0;
 }
 
 static void creg_issue_cmd(struct rsxx_cardinfo *card, struct creg_cmd *cmd)
 {
+	int st;
+
+	if (unlikely(card->eeh_state))
+		return;
+
 	iowrite32(cmd->addr, card->regmap + CREG_ADD);
 	iowrite32(cmd->cnt8, card->regmap + CREG_CNT);
 
 	if (cmd->op == CREG_OP_WRITE) {
-		if (cmd->buf)
-			copy_to_creg_data(card, cmd->cnt8,
-					  cmd->buf, cmd->stream);
+		if (cmd->buf) {
+			st = copy_to_creg_data(card, cmd->cnt8,
+					       cmd->buf, cmd->stream);
+			if (st)
+				return;
+		}
 	}
 
-	/*
-	 * Data copy must complete before initiating the command. This is
-	 * needed for weakly ordered processors (i.e. PowerPC), so that all
-	 * neccessary registers are written before we kick the hardware.
-	 */
-	wmb();
+	if (unlikely(card->eeh_state))
+		return;
 
 	/* Setting the valid bit will kick off the command. */
 	iowrite32(cmd->op, card->regmap + CREG_CMD);
@@ -196,11 +194,11 @@ static int creg_queue_cmd(struct rsxx_cardinfo *card,
 	cmd->cb_private = cb_private;
 	cmd->status	= 0;
 
-	spin_lock(&card->creg_ctrl.lock);
+	spin_lock_bh(&card->creg_ctrl.lock);
 	list_add_tail(&cmd->list, &card->creg_ctrl.queue);
 	card->creg_ctrl.q_depth++;
 	creg_kick_queue(card);
-	spin_unlock(&card->creg_ctrl.lock);
+	spin_unlock_bh(&card->creg_ctrl.lock);
 
 	return 0;
 }
@@ -210,7 +208,11 @@ static void creg_cmd_timed_out(unsigned long data)
 	struct rsxx_cardinfo *card = (struct rsxx_cardinfo *) data;
 	struct creg_cmd *cmd;
 
-	cmd = pop_active_cmd(card);
+	spin_lock(&card->creg_ctrl.lock);
+	cmd = card->creg_ctrl.active_cmd;
+	card->creg_ctrl.active_cmd = NULL;
+	spin_unlock(&card->creg_ctrl.lock);
+
 	if (cmd == NULL) {
 		card->creg_ctrl.creg_stats.creg_timeout++;
 		dev_warn(CARD_TO_DEV(card),
@@ -247,7 +249,11 @@ static void creg_cmd_done(struct work_struct *work)
 	if (del_timer_sync(&card->creg_ctrl.cmd_timer) == 0)
 		card->creg_ctrl.creg_stats.failed_cancel_timer++;
 
-	cmd = pop_active_cmd(card);
+	spin_lock_bh(&card->creg_ctrl.lock);
+	cmd = card->creg_ctrl.active_cmd;
+	card->creg_ctrl.active_cmd = NULL;
+	spin_unlock_bh(&card->creg_ctrl.lock);
+
 	if (cmd == NULL) {
 		dev_err(CARD_TO_DEV(card),
 			"Spurious creg interrupt!\n");
@@ -287,7 +293,7 @@ static void creg_cmd_done(struct work_struct *work)
 			goto creg_done;
 		}
 
-		copy_from_creg_data(card, cnt8, cmd->buf, cmd->stream);
+		st = copy_from_creg_data(card, cnt8, cmd->buf, cmd->stream);
 	}
 
 creg_done:
@@ -296,10 +302,10 @@ creg_done:
 
 	kmem_cache_free(creg_cmd_pool, cmd);
 
-	spin_lock(&card->creg_ctrl.lock);
+	spin_lock_bh(&card->creg_ctrl.lock);
 	card->creg_ctrl.active = 0;
 	creg_kick_queue(card);
-	spin_unlock(&card->creg_ctrl.lock);
+	spin_unlock_bh(&card->creg_ctrl.lock);
 }
 
 static void creg_reset(struct rsxx_cardinfo *card)
@@ -324,7 +330,7 @@ static void creg_reset(struct rsxx_cardinfo *card)
 		"Resetting creg interface for recovery\n");
 
 	/* Cancel outstanding commands */
-	spin_lock(&card->creg_ctrl.lock);
+	spin_lock_bh(&card->creg_ctrl.lock);
 	list_for_each_entry_safe(cmd, tmp, &card->creg_ctrl.queue, list) {
 		list_del(&cmd->list);
 		card->creg_ctrl.q_depth--;
@@ -345,7 +351,7 @@ static void creg_reset(struct rsxx_cardinfo *card)
 
 		card->creg_ctrl.active = 0;
 	}
-	spin_unlock(&card->creg_ctrl.lock);
+	spin_unlock_bh(&card->creg_ctrl.lock);
 
 	card->creg_ctrl.reset = 0;
 	spin_lock_irqsave(&card->irq_lock, flags);
@@ -399,12 +405,12 @@ static int __issue_creg_rw(struct rsxx_cardinfo *card,
 		return st;
 
 	/*
-	 * This timeout is neccessary for unresponsive hardware. The additional
+	 * This timeout is necessary for unresponsive hardware. The additional
 	 * 20 seconds to used to guarantee that each cregs requests has time to
 	 * complete.
 	 */
-	timeout = msecs_to_jiffies((CREG_TIMEOUT_MSEC *
-				card->creg_ctrl.q_depth) + 20000);
+	timeout = msecs_to_jiffies(CREG_TIMEOUT_MSEC *
+				   card->creg_ctrl.q_depth + 20000);
 
 	/*
 	 * The creg interface is guaranteed to complete. It has a timeout
@@ -690,6 +696,32 @@ int rsxx_reg_access(struct rsxx_cardinfo *card,
 	return 0;
 }
 
+void rsxx_eeh_save_issued_creg(struct rsxx_cardinfo *card)
+{
+	struct creg_cmd *cmd = NULL;
+
+	cmd = card->creg_ctrl.active_cmd;
+	card->creg_ctrl.active_cmd = NULL;
+
+	if (cmd) {
+		del_timer_sync(&card->creg_ctrl.cmd_timer);
+
+		spin_lock_bh(&card->creg_ctrl.lock);
+		list_add(&cmd->list, &card->creg_ctrl.queue);
+		card->creg_ctrl.q_depth++;
+		card->creg_ctrl.active = 0;
+		spin_unlock_bh(&card->creg_ctrl.lock);
+	}
+}
+
+void rsxx_kick_creg_queue(struct rsxx_cardinfo *card)
+{
+	spin_lock_bh(&card->creg_ctrl.lock);
+	if (!list_empty(&card->creg_ctrl.queue))
+		creg_kick_queue(card);
+	spin_unlock_bh(&card->creg_ctrl.lock);
+}
+
 /*------------ Initialization & Setup --------------*/
 int rsxx_creg_setup(struct rsxx_cardinfo *card)
 {
@@ -712,7 +744,7 @@ void rsxx_creg_destroy(struct rsxx_cardinfo *card)
 	int cnt = 0;
 
 	/* Cancel outstanding commands */
-	spin_lock(&card->creg_ctrl.lock);
+	spin_lock_bh(&card->creg_ctrl.lock);
 	list_for_each_entry_safe(cmd, tmp, &card->creg_ctrl.queue, list) {
 		list_del(&cmd->list);
 		if (cmd->cb)
@@ -737,7 +769,7 @@ void rsxx_creg_destroy(struct rsxx_cardinfo *card)
 			"Canceled active creg command\n");
 		kmem_cache_free(creg_cmd_pool, cmd);
 	}
-	spin_unlock(&card->creg_ctrl.lock);
+	spin_unlock_bh(&card->creg_ctrl.lock);
 
 	cancel_work_sync(&card->creg_ctrl.done_work);
 }

drivers/block/rsxx/dma.c

@@ -28,7 +28,7 @@
 struct rsxx_dma {
 	struct list_head	 list;
 	u8			 cmd;
-	unsigned int		 laddr;     /* Logical address on the ramsan */
+	unsigned int		 laddr;     /* Logical address */
 	struct {
 		u32		 off;
 		u32		 cnt;
@@ -81,9 +81,6 @@ enum rsxx_hw_status {
 	HW_STATUS_FAULT		= 0x08,
 };
 
-#define STATUS_BUFFER_SIZE8     4096
-#define COMMAND_BUFFER_SIZE8    4096
-
 static struct kmem_cache *rsxx_dma_pool;
 
 struct dma_tracker {
@@ -122,7 +119,7 @@ static unsigned int rsxx_get_dma_tgt(struct rsxx_cardinfo *card, u64 addr8)
 	return tgt;
 }
 
-static void rsxx_dma_queue_reset(struct rsxx_cardinfo *card)
+void rsxx_dma_queue_reset(struct rsxx_cardinfo *card)
 {
 	/* Reset all DMA Command/Status Queues */
 	iowrite32(DMA_QUEUE_RESET, card->regmap + RESET);
@@ -210,7 +207,8 @@ static void dma_intr_coal_auto_tune(struct rsxx_cardinfo *card)
 	u32 q_depth = 0;
 	u32 intr_coal;
 
-	if (card->config.data.intr_coal.mode != RSXX_INTR_COAL_AUTO_TUNE)
+	if (card->config.data.intr_coal.mode != RSXX_INTR_COAL_AUTO_TUNE ||
+	    unlikely(card->eeh_state))
 		return;
 
 	for (i = 0; i < card->n_targets; i++)
@@ -223,31 +221,26 @@ static void dma_intr_coal_auto_tune(struct rsxx_cardinfo *card)
 }
 
 /*----------------- RSXX DMA Handling -------------------*/
-static void rsxx_complete_dma(struct rsxx_cardinfo *card,
+static void rsxx_complete_dma(struct rsxx_dma_ctrl *ctrl,
 				  struct rsxx_dma *dma,
 				  unsigned int status)
 {
 	if (status & DMA_SW_ERR)
-		printk_ratelimited(KERN_ERR
-				   "SW Error in DMA(cmd x%02x, laddr x%08x)\n",
-				   dma->cmd, dma->laddr);
+		ctrl->stats.dma_sw_err++;
 	if (status & DMA_HW_FAULT)
-		printk_ratelimited(KERN_ERR
-				   "HW Fault in DMA(cmd x%02x, laddr x%08x)\n",
-				   dma->cmd, dma->laddr);
+		ctrl->stats.dma_hw_fault++;
 	if (status & DMA_CANCELLED)
-		printk_ratelimited(KERN_ERR
-				   "DMA Cancelled(cmd x%02x, laddr x%08x)\n",
-				   dma->cmd, dma->laddr);
+		ctrl->stats.dma_cancelled++;
 
 	if (dma->dma_addr)
-		pci_unmap_page(card->dev, dma->dma_addr, get_dma_size(dma),
+		pci_unmap_page(ctrl->card->dev, dma->dma_addr,
+			       get_dma_size(dma),
 			       dma->cmd == HW_CMD_BLK_WRITE ?
 					   PCI_DMA_TODEVICE :
 					   PCI_DMA_FROMDEVICE);
 
 	if (dma->cb)
-		dma->cb(card, dma->cb_data, status ? 1 : 0);
+		dma->cb(ctrl->card, dma->cb_data, status ? 1 : 0);
 
 	kmem_cache_free(rsxx_dma_pool, dma);
 }
@@ -330,14 +323,15 @@ static void rsxx_handle_dma_error(struct rsxx_dma_ctrl *ctrl,
 	if (requeue_cmd)
 		rsxx_requeue_dma(ctrl, dma);
 	else
-		rsxx_complete_dma(ctrl->card, dma, status);
+		rsxx_complete_dma(ctrl, dma, status);
 }
 
 static void dma_engine_stalled(unsigned long data)
 {
 	struct rsxx_dma_ctrl *ctrl = (struct rsxx_dma_ctrl *)data;
 
-	if (atomic_read(&ctrl->stats.hw_q_depth) == 0)
+	if (atomic_read(&ctrl->stats.hw_q_depth) == 0 ||
+	    unlikely(ctrl->card->eeh_state))
 		return;
 
 	if (ctrl->cmd.idx != ioread32(ctrl->regmap + SW_CMD_IDX)) {
@@ -369,7 +363,8 @@ static void rsxx_issue_dmas(struct work_struct *work)
 	ctrl = container_of(work, struct rsxx_dma_ctrl, issue_dma_work);
 	hw_cmd_buf = ctrl->cmd.buf;
 
-	if (unlikely(ctrl->card->halt))
+	if (unlikely(ctrl->card->halt) ||
+	    unlikely(ctrl->card->eeh_state))
 		return;
 
 	while (1) {
@@ -397,7 +392,7 @@ static void rsxx_issue_dmas(struct work_struct *work)
 		 */
 		if (unlikely(ctrl->card->dma_fault)) {
 			push_tracker(ctrl->trackers, tag);
-			rsxx_complete_dma(ctrl->card, dma, DMA_CANCELLED);
+			rsxx_complete_dma(ctrl, dma, DMA_CANCELLED);
 			continue;
 		}
 
@@ -432,19 +427,15 @@ static void rsxx_issue_dmas(struct work_struct *work)
 
 	/* Let HW know we've queued commands. */
 	if (cmds_pending) {
-		/*
-		 * We must guarantee that the CPU writes to 'ctrl->cmd.buf'
-		 * (which is in PCI-consistent system-memory) from the loop
-		 * above make it into the coherency domain before the
-		 * following PIO "trigger" updating the cmd.idx.  A WMB is
-		 * sufficient. We need not explicitly CPU cache-flush since
-		 * the memory is a PCI-consistent (ie; coherent) mapping.
-		 */
-		wmb();
-
 		atomic_add(cmds_pending, &ctrl->stats.hw_q_depth);
 		mod_timer(&ctrl->activity_timer,
 			  jiffies + DMA_ACTIVITY_TIMEOUT);
+
+		if (unlikely(ctrl->card->eeh_state)) {
+			del_timer_sync(&ctrl->activity_timer);
+			return;
+		}
+
 		iowrite32(ctrl->cmd.idx, ctrl->regmap + SW_CMD_IDX);
 	}
 }
@@ -463,7 +454,8 @@ static void rsxx_dma_done(struct work_struct *work)
 	hw_st_buf = ctrl->status.buf;
 
 	if (unlikely(ctrl->card->halt) ||
-	    unlikely(ctrl->card->dma_fault))
+	    unlikely(ctrl->card->dma_fault) ||
+	    unlikely(ctrl->card->eeh_state))
 		return;
 
 	count = le16_to_cpu(hw_st_buf[ctrl->status.idx].count);
@@ -508,7 +500,7 @@ static void rsxx_dma_done(struct work_struct *work)
 		if (status)
 			rsxx_handle_dma_error(ctrl, dma, status);
 		else
-			rsxx_complete_dma(ctrl->card, dma, 0);
+			rsxx_complete_dma(ctrl, dma, 0);
 
 		push_tracker(ctrl->trackers, tag);
 
@@ -727,20 +719,54 @@ bvec_err:
 
 
 /*----------------- DMA Engine Initialization & Setup -------------------*/
+int rsxx_hw_buffers_init(struct pci_dev *dev, struct rsxx_dma_ctrl *ctrl)
+{
+	ctrl->status.buf = pci_alloc_consistent(dev, STATUS_BUFFER_SIZE8,
+				&ctrl->status.dma_addr);
+	ctrl->cmd.buf = pci_alloc_consistent(dev, COMMAND_BUFFER_SIZE8,
+				&ctrl->cmd.dma_addr);
+	if (ctrl->status.buf == NULL || ctrl->cmd.buf == NULL)
+		return -ENOMEM;
+
+	memset(ctrl->status.buf, 0xac, STATUS_BUFFER_SIZE8);
+	iowrite32(lower_32_bits(ctrl->status.dma_addr),
+		ctrl->regmap + SB_ADD_LO);
+	iowrite32(upper_32_bits(ctrl->status.dma_addr),
+		ctrl->regmap + SB_ADD_HI);
+
+	memset(ctrl->cmd.buf, 0x83, COMMAND_BUFFER_SIZE8);
+	iowrite32(lower_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_LO);
+	iowrite32(upper_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_HI);
+
+	ctrl->status.idx = ioread32(ctrl->regmap + HW_STATUS_CNT);
+	if (ctrl->status.idx > RSXX_MAX_OUTSTANDING_CMDS) {
+		dev_crit(&dev->dev, "Failed reading status cnt x%x\n",
+			ctrl->status.idx);
+		return -EINVAL;
+	}
+	iowrite32(ctrl->status.idx, ctrl->regmap + HW_STATUS_CNT);
+	iowrite32(ctrl->status.idx, ctrl->regmap + SW_STATUS_CNT);
+
+	ctrl->cmd.idx = ioread32(ctrl->regmap + HW_CMD_IDX);
+	if (ctrl->cmd.idx > RSXX_MAX_OUTSTANDING_CMDS) {
+		dev_crit(&dev->dev, "Failed reading cmd cnt x%x\n",
+			ctrl->status.idx);
+		return -EINVAL;
+	}
+	iowrite32(ctrl->cmd.idx, ctrl->regmap + HW_CMD_IDX);
+	iowrite32(ctrl->cmd.idx, ctrl->regmap + SW_CMD_IDX);
+
+	return 0;
+}
+
 static int rsxx_dma_ctrl_init(struct pci_dev *dev,
 				  struct rsxx_dma_ctrl *ctrl)
 {
 	int i;
+	int st;
 
 	memset(&ctrl->stats, 0, sizeof(ctrl->stats));
 
-	ctrl->status.buf = pci_alloc_consistent(dev, STATUS_BUFFER_SIZE8,
-						&ctrl->status.dma_addr);
-	ctrl->cmd.buf = pci_alloc_consistent(dev, COMMAND_BUFFER_SIZE8,
-					     &ctrl->cmd.dma_addr);
-	if (ctrl->status.buf == NULL || ctrl->cmd.buf == NULL)
-		return -ENOMEM;
-
 	ctrl->trackers = vmalloc(DMA_TRACKER_LIST_SIZE8);
 	if (!ctrl->trackers)
 		return -ENOMEM;
@@ -770,35 +796,9 @@ static int rsxx_dma_ctrl_init(struct pci_dev *dev,
 	INIT_WORK(&ctrl->issue_dma_work, rsxx_issue_dmas);
 	INIT_WORK(&ctrl->dma_done_work, rsxx_dma_done);
 
-	memset(ctrl->status.buf, 0xac, STATUS_BUFFER_SIZE8);
-	iowrite32(lower_32_bits(ctrl->status.dma_addr),
-		  ctrl->regmap + SB_ADD_LO);
-	iowrite32(upper_32_bits(ctrl->status.dma_addr),
-		  ctrl->regmap + SB_ADD_HI);
-
-	memset(ctrl->cmd.buf, 0x83, COMMAND_BUFFER_SIZE8);
-	iowrite32(lower_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_LO);
-	iowrite32(upper_32_bits(ctrl->cmd.dma_addr), ctrl->regmap + CB_ADD_HI);
-
-	ctrl->status.idx = ioread32(ctrl->regmap + HW_STATUS_CNT);
-	if (ctrl->status.idx > RSXX_MAX_OUTSTANDING_CMDS) {
-		dev_crit(&dev->dev, "Failed reading status cnt x%x\n",
-			 ctrl->status.idx);
-		return -EINVAL;
-	}
-	iowrite32(ctrl->status.idx, ctrl->regmap + HW_STATUS_CNT);
-	iowrite32(ctrl->status.idx, ctrl->regmap + SW_STATUS_CNT);
-
-	ctrl->cmd.idx = ioread32(ctrl->regmap + HW_CMD_IDX);
-	if (ctrl->cmd.idx > RSXX_MAX_OUTSTANDING_CMDS) {
-		dev_crit(&dev->dev, "Failed reading cmd cnt x%x\n",
-			 ctrl->status.idx);
-		return -EINVAL;
-	}
-	iowrite32(ctrl->cmd.idx, ctrl->regmap + HW_CMD_IDX);
-	iowrite32(ctrl->cmd.idx, ctrl->regmap + SW_CMD_IDX);
-
-	wmb();
+	st = rsxx_hw_buffers_init(dev, ctrl);
+	if (st)
+		return st;
 
 	return 0;
 }
@@ -834,7 +834,7 @@ static int rsxx_dma_stripe_setup(struct rsxx_cardinfo *card,
 	return 0;
 }
 
-static int rsxx_dma_configure(struct rsxx_cardinfo *card)
+int rsxx_dma_configure(struct rsxx_cardinfo *card)
 {
 	u32 intr_coal;
 
@@ -980,6 +980,103 @@ void rsxx_dma_destroy(struct rsxx_cardinfo *card)
 	}
 }
 
+int rsxx_eeh_save_issued_dmas(struct rsxx_cardinfo *card)
+{
+	int i;
+	int j;
+	int cnt;
+	struct rsxx_dma *dma;
+	struct list_head *issued_dmas;
+
+	issued_dmas = kzalloc(sizeof(*issued_dmas) * card->n_targets,
+			      GFP_KERNEL);
+	if (!issued_dmas)
+		return -ENOMEM;
+
+	for (i = 0; i < card->n_targets; i++) {
+		INIT_LIST_HEAD(&issued_dmas[i]);
+		cnt = 0;
+		for (j = 0; j < RSXX_MAX_OUTSTANDING_CMDS; j++) {
+			dma = get_tracker_dma(card->ctrl[i].trackers, j);
+			if (dma == NULL)
+				continue;
+
+			if (dma->cmd == HW_CMD_BLK_WRITE)
+				card->ctrl[i].stats.writes_issued--;
+			else if (dma->cmd == HW_CMD_BLK_DISCARD)
+				card->ctrl[i].stats.discards_issued--;
+			else
+				card->ctrl[i].stats.reads_issued--;
+
+			list_add_tail(&dma->list, &issued_dmas[i]);
+			push_tracker(card->ctrl[i].trackers, j);
+			cnt++;
+		}
+
+		spin_lock(&card->ctrl[i].queue_lock);
+		list_splice(&issued_dmas[i], &card->ctrl[i].queue);
+
+		atomic_sub(cnt, &card->ctrl[i].stats.hw_q_depth);
+		card->ctrl[i].stats.sw_q_depth += cnt;
+		card->ctrl[i].e_cnt = 0;
+
+		list_for_each_entry(dma, &card->ctrl[i].queue, list) {
+			if (dma->dma_addr)
+				pci_unmap_page(card->dev, dma->dma_addr,
+					       get_dma_size(dma),
+					       dma->cmd == HW_CMD_BLK_WRITE ?
+					       PCI_DMA_TODEVICE :
+					       PCI_DMA_FROMDEVICE);
+		}
+		spin_unlock(&card->ctrl[i].queue_lock);
+	}
+
+	kfree(issued_dmas);
+
+	return 0;
+}
+
+void rsxx_eeh_cancel_dmas(struct rsxx_cardinfo *card)
+{
+	struct rsxx_dma *dma;
+	struct rsxx_dma *tmp;
+	int i;
+
+	for (i = 0; i < card->n_targets; i++) {
+		spin_lock(&card->ctrl[i].queue_lock);
+		list_for_each_entry_safe(dma, tmp, &card->ctrl[i].queue, list) {
+			list_del(&dma->list);
+
+			rsxx_complete_dma(&card->ctrl[i], dma, DMA_CANCELLED);
+		}
+		spin_unlock(&card->ctrl[i].queue_lock);
+	}
+}
+
+int rsxx_eeh_remap_dmas(struct rsxx_cardinfo *card)
+{
+	struct rsxx_dma *dma;
+	int i;
+
+	for (i = 0; i < card->n_targets; i++) {
+		spin_lock(&card->ctrl[i].queue_lock);
+		list_for_each_entry(dma, &card->ctrl[i].queue, list) {
+			dma->dma_addr = pci_map_page(card->dev, dma->page,
+					dma->pg_off, get_dma_size(dma),
+					dma->cmd == HW_CMD_BLK_WRITE ?
+					PCI_DMA_TODEVICE :
+					PCI_DMA_FROMDEVICE);
+			if (!dma->dma_addr) {
+				spin_unlock(&card->ctrl[i].queue_lock);
+				kmem_cache_free(rsxx_dma_pool, dma);
+				return -ENOMEM;
+			}
+		}
+		spin_unlock(&card->ctrl[i].queue_lock);
+	}
+
+	return 0;
+}
 
 int rsxx_dma_init(void)
 {

drivers/block/rsxx/rsxx.h

@@ -27,15 +27,17 @@
 
 /*----------------- IOCTL Definitions -------------------*/
 
+#define RSXX_MAX_DATA 8
+
 struct rsxx_reg_access {
 	__u32 addr;
 	__u32 cnt;
 	__u32 stat;
 	__u32 stream;
-	__u32 data[8];
+	__u32 data[RSXX_MAX_DATA];
 };
 
-#define RSXX_MAX_REG_CNT	(8 * (sizeof(__u32)))
+#define RSXX_MAX_REG_CNT	(RSXX_MAX_DATA * (sizeof(__u32)))
 
 #define RSXX_IOC_MAGIC 'r'
 

drivers/block/rsxx/rsxx_cfg.h

@@ -58,7 +58,7 @@ struct rsxx_card_cfg {
 };
 
 /* Vendor ID Values */
-#define RSXX_VENDOR_ID_TMS_IBM		0
+#define RSXX_VENDOR_ID_IBM		0
 #define RSXX_VENDOR_ID_DSI		1
 #define RSXX_VENDOR_COUNT		2
 

drivers/block/rsxx/rsxx_priv.h

@@ -45,16 +45,13 @@
 
 struct proc_cmd;
 
-#define PCI_VENDOR_ID_TMS_IBM		0x15B6
-#define PCI_DEVICE_ID_RS70_FLASH	0x0019
-#define PCI_DEVICE_ID_RS70D_FLASH	0x001A
-#define PCI_DEVICE_ID_RS80_FLASH	0x001C
-#define PCI_DEVICE_ID_RS81_FLASH	0x001E
+#define PCI_DEVICE_ID_FS70_FLASH	0x04A9
+#define PCI_DEVICE_ID_FS80_FLASH	0x04AA
 
 #define RS70_PCI_REV_SUPPORTED	4
 
 #define DRIVER_NAME "rsxx"
-#define DRIVER_VERSION "3.7"
+#define DRIVER_VERSION "4.0"
 
 /* Block size is 4096 */
 #define RSXX_HW_BLK_SHIFT		12
@@ -67,6 +64,9 @@ struct proc_cmd;
 #define RSXX_MAX_OUTSTANDING_CMDS	255
 #define RSXX_CS_IDX_MASK		0xff
 
+#define STATUS_BUFFER_SIZE8     4096
+#define COMMAND_BUFFER_SIZE8    4096
+
 #define RSXX_MAX_TARGETS	8
 
 struct dma_tracker_list;
@@ -91,6 +91,9 @@ struct rsxx_dma_stats {
 	u32 discards_failed;
 	u32 done_rescheduled;
 	u32 issue_rescheduled;
+	u32 dma_sw_err;
+	u32 dma_hw_fault;
+	u32 dma_cancelled;
 	u32 sw_q_depth;		/* Number of DMAs on the SW queue. */
 	atomic_t hw_q_depth;	/* Number of DMAs queued to HW. */
 };
@@ -116,6 +119,7 @@ struct rsxx_dma_ctrl {
 struct rsxx_cardinfo {
 	struct pci_dev		*dev;
 	unsigned int		halt;
+	unsigned int		eeh_state;
 
 	void			__iomem *regmap;
 	spinlock_t		irq_lock;
@@ -224,6 +228,7 @@ enum rsxx_pci_regmap {
 	PERF_RD512_HI	= 0xac,
 	PERF_WR512_LO	= 0xb0,
 	PERF_WR512_HI	= 0xb4,
+	PCI_RECONFIG	= 0xb8,
 };
 
 enum rsxx_intr {
@@ -237,6 +242,8 @@ enum rsxx_intr {
 	CR_INTR_DMA5	= 0x00000080,
 	CR_INTR_DMA6	= 0x00000100,
 	CR_INTR_DMA7	= 0x00000200,
+	CR_INTR_ALL_C	= 0x0000003f,
+	CR_INTR_ALL_G	= 0x000003ff,
 	CR_INTR_DMA_ALL = 0x000003f5,
 	CR_INTR_ALL	= 0xffffffff,
 };
@@ -253,8 +260,14 @@ enum rsxx_pci_reset {
 	DMA_QUEUE_RESET		= 0x00000001,
 };
 
+enum rsxx_hw_fifo_flush {
+	RSXX_FLUSH_BUSY		= 0x00000002,
+	RSXX_FLUSH_TIMEOUT	= 0x00000004,
+};
+
 enum rsxx_pci_revision {
 	RSXX_DISCARD_SUPPORT = 2,
+	RSXX_EEH_SUPPORT     = 3,
 };
 
 enum rsxx_creg_cmd {
@@ -360,11 +373,17 @@ int rsxx_dma_setup(struct rsxx_cardinfo *card);
 void rsxx_dma_destroy(struct rsxx_cardinfo *card);
 int rsxx_dma_init(void);
 void rsxx_dma_cleanup(void);
+void rsxx_dma_queue_reset(struct rsxx_cardinfo *card);
+int rsxx_dma_configure(struct rsxx_cardinfo *card);
 int rsxx_dma_queue_bio(struct rsxx_cardinfo *card,
 			   struct bio *bio,
 			   atomic_t *n_dmas,
 			   rsxx_dma_cb cb,
 			   void *cb_data);
+int rsxx_hw_buffers_init(struct pci_dev *dev, struct rsxx_dma_ctrl *ctrl);
+int rsxx_eeh_save_issued_dmas(struct rsxx_cardinfo *card);
+void rsxx_eeh_cancel_dmas(struct rsxx_cardinfo *card);
+int rsxx_eeh_remap_dmas(struct rsxx_cardinfo *card);
 
 /***** cregs.c *****/
 int rsxx_creg_write(struct rsxx_cardinfo *card, u32 addr,
@@ -389,10 +408,11 @@ int rsxx_creg_setup(struct rsxx_cardinfo *card);
 void rsxx_creg_destroy(struct rsxx_cardinfo *card);
 int rsxx_creg_init(void);
 void rsxx_creg_cleanup(void);
-
 int rsxx_reg_access(struct rsxx_cardinfo *card,
 			struct rsxx_reg_access __user *ucmd,
 			int read);
+void rsxx_eeh_save_issued_creg(struct rsxx_cardinfo *card);
+void rsxx_kick_creg_queue(struct rsxx_cardinfo *card);
 
 
 

drivers/block/xen-blkback/blkback.c

@@ -164,7 +164,7 @@ static void make_response(struct xen_blkif *blkif, u64 id,
 
 #define foreach_grant_safe(pos, n, rbtree, node) \
 	for ((pos) = container_of(rb_first((rbtree)), typeof(*(pos)), node), \
-	     (n) = rb_next(&(pos)->node); \
+	     (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL; \
 	     &(pos)->node != NULL; \
 	     (pos) = container_of(n, typeof(*(pos)), node), \
 	     (n) = (&(pos)->node != NULL) ? rb_next(&(pos)->node) : NULL)
@@ -381,8 +381,8 @@ irqreturn_t xen_blkif_be_int(int irq, void *dev_id)
 
 static void print_stats(struct xen_blkif *blkif)
 {
-	pr_info("xen-blkback (%s): oo %3d  |  rd %4d  |  wr %4d  |  f %4d"
-		 "  |  ds %4d\n",
+	pr_info("xen-blkback (%s): oo %3llu  |  rd %4llu  |  wr %4llu  |  f %4llu"
+		 "  |  ds %4llu\n",
 		 current->comm, blkif->st_oo_req,
 		 blkif->st_rd_req, blkif->st_wr_req,
 		 blkif->st_f_req, blkif->st_ds_req);
@@ -442,7 +442,7 @@ int xen_blkif_schedule(void *arg)
 }
 
 struct seg_buf {
-	unsigned long buf;
+	unsigned int offset;
 	unsigned int nsec;
 };
 /*
@@ -621,30 +621,21 @@ static int xen_blkbk_map(struct blkif_request *req,
 				 * If this is a new persistent grant
 				 * save the handler
 				 */
-				persistent_gnts[i]->handle = map[j].handle;
-				persistent_gnts[i]->dev_bus_addr =
-					map[j++].dev_bus_addr;
+				persistent_gnts[i]->handle = map[j++].handle;
 			}
 			pending_handle(pending_req, i) =
 				persistent_gnts[i]->handle;
 
 			if (ret)
 				continue;
-
-			seg[i].buf = persistent_gnts[i]->dev_bus_addr |
-				(req->u.rw.seg[i].first_sect << 9);
 		} else {
-			pending_handle(pending_req, i) = map[j].handle;
+			pending_handle(pending_req, i) = map[j++].handle;
 			bitmap_set(pending_req->unmap_seg, i, 1);
 
-			if (ret) {
-				j++;
+			if (ret)
 				continue;
-			}
-
-			seg[i].buf = map[j++].dev_bus_addr |
-				(req->u.rw.seg[i].first_sect << 9);
 		}
+		seg[i].offset = (req->u.rw.seg[i].first_sect << 9);
 	}
 	return ret;
 }
@@ -679,6 +670,16 @@ static int dispatch_discard_io(struct xen_blkif *blkif,
 	return err;
 }
 
+static int dispatch_other_io(struct xen_blkif *blkif,
+			     struct blkif_request *req,
+			     struct pending_req *pending_req)
+{
+	free_req(pending_req);
+	make_response(blkif, req->u.other.id, req->operation,
+		      BLKIF_RSP_EOPNOTSUPP);
+	return -EIO;
+}
+
 static void xen_blk_drain_io(struct xen_blkif *blkif)
 {
 	atomic_set(&blkif->drain, 1);
@@ -800,17 +801,30 @@ __do_block_io_op(struct xen_blkif *blkif)
 
 		/* Apply all sanity checks to /private copy/ of request. */
 		barrier();
-		if (unlikely(req.operation == BLKIF_OP_DISCARD)) {
+
+		switch (req.operation) {
+		case BLKIF_OP_READ:
+		case BLKIF_OP_WRITE:
+		case BLKIF_OP_WRITE_BARRIER:
+		case BLKIF_OP_FLUSH_DISKCACHE:
+			if (dispatch_rw_block_io(blkif, &req, pending_req))
+				goto done;
+			break;
+		case BLKIF_OP_DISCARD:
 			free_req(pending_req);
 			if (dispatch_discard_io(blkif, &req))
-				break;
-		} else if (dispatch_rw_block_io(blkif, &req, pending_req))
+				goto done;
 			break;
+		default:
+			if (dispatch_other_io(blkif, &req, pending_req))
+				goto done;
+			break;
+		}
 
 		/* Yield point for this unbounded loop. */
 		cond_resched();
 	}
-
+done:
 	return more_to_do;
 }
 
@@ -904,7 +918,8 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 		pr_debug(DRV_PFX "access denied: %s of [%llu,%llu] on dev=%04x\n",
 			 operation == READ ? "read" : "write",
 			 preq.sector_number,
-			 preq.sector_number + preq.nr_sects, preq.dev);
+			 preq.sector_number + preq.nr_sects,
+			 blkif->vbd.pdevice);
 		goto fail_response;
 	}
 
@@ -947,7 +962,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 		       (bio_add_page(bio,
 				     pages[i],
 				     seg[i].nsec << 9,
-				     seg[i].buf & ~PAGE_MASK) == 0)) {
+				     seg[i].offset) == 0)) {
 
 			bio = bio_alloc(GFP_KERNEL, nseg-i);
 			if (unlikely(bio == NULL))
@@ -977,13 +992,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 		bio->bi_end_io  = end_block_io_op;
 	}
 
-	/*
-	 * We set it one so that the last submit_bio does not have to call
-	 * atomic_inc.
-	 */
 	atomic_set(&pending_req->pendcnt, nbio);
-
-	/* Get a reference count for the disk queue and start sending I/O */
 	blk_start_plug(&plug);
 
 	for (i = 0; i < nbio; i++)
@@ -1011,6 +1020,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
  fail_put_bio:
 	for (i = 0; i < nbio; i++)
 		bio_put(biolist[i]);
+	atomic_set(&pending_req->pendcnt, 1);
 	__end_block_io_op(pending_req, -EINVAL);
 	msleep(1); /* back off a bit */
 	return -EIO;

drivers/block/xen-blkback/common.h

@@ -77,11 +77,18 @@ struct blkif_x86_32_request_discard {
 	uint64_t       nr_sectors;
 } __attribute__((__packed__));
 
+struct blkif_x86_32_request_other {
+	uint8_t        _pad1;
+	blkif_vdev_t   _pad2;
+	uint64_t       id;           /* private guest value, echoed in resp  */
+} __attribute__((__packed__));
+
 struct blkif_x86_32_request {
 	uint8_t        operation;    /* BLKIF_OP_???                         */
 	union {
 		struct blkif_x86_32_request_rw rw;
 		struct blkif_x86_32_request_discard discard;
+		struct blkif_x86_32_request_other other;
 	} u;
 } __attribute__((__packed__));
 
@@ -113,11 +120,19 @@ struct blkif_x86_64_request_discard {
 	uint64_t       nr_sectors;
 } __attribute__((__packed__));
 
+struct blkif_x86_64_request_other {
+	uint8_t        _pad1;
+	blkif_vdev_t   _pad2;
+	uint32_t       _pad3;        /* offsetof(blkif_..,u.discard.id)==8   */
+	uint64_t       id;           /* private guest value, echoed in resp  */
+} __attribute__((__packed__));
+
 struct blkif_x86_64_request {
 	uint8_t        operation;    /* BLKIF_OP_???                         */
 	union {
 		struct blkif_x86_64_request_rw rw;
 		struct blkif_x86_64_request_discard discard;
+		struct blkif_x86_64_request_other other;
 	} u;
 } __attribute__((__packed__));
 
@@ -172,7 +187,6 @@ struct persistent_gnt {
 	struct page *page;
 	grant_ref_t gnt;
 	grant_handle_t handle;
-	uint64_t dev_bus_addr;
 	struct rb_node node;
 };
 
@@ -208,13 +222,13 @@ struct xen_blkif {
 
 	/* statistics */
 	unsigned long		st_print;
-	int			st_rd_req;
-	int			st_wr_req;
-	int			st_oo_req;
-	int			st_f_req;
-	int			st_ds_req;
-	int			st_rd_sect;
-	int			st_wr_sect;
+	unsigned long long			st_rd_req;
+	unsigned long long			st_wr_req;
+	unsigned long long			st_oo_req;
+	unsigned long long			st_f_req;
+	unsigned long long			st_ds_req;
+	unsigned long long			st_rd_sect;
+	unsigned long long			st_wr_sect;
 
 	wait_queue_head_t	waiting_to_free;
 };
@@ -278,6 +292,11 @@ static inline void blkif_get_x86_32_req(struct blkif_request *dst,
 		dst->u.discard.nr_sectors = src->u.discard.nr_sectors;
 		break;
 	default:
+		/*
+		 * Don't know how to translate this op. Only get the
+		 * ID so failure can be reported to the frontend.
+		 */
+		dst->u.other.id = src->u.other.id;
 		break;
 	}
 }
@@ -309,6 +328,11 @@ static inline void blkif_get_x86_64_req(struct blkif_request *dst,
 		dst->u.discard.nr_sectors = src->u.discard.nr_sectors;
 		break;
 	default:
+		/*
+		 * Don't know how to translate this op. Only get the
+		 * ID so failure can be reported to the frontend.
+		 */
+		dst->u.other.id = src->u.other.id;
 		break;
 	}
 }

drivers/block/xen-blkback/xenbus.c

@@ -230,13 +230,13 @@ int __init xen_blkif_interface_init(void)
 	}								\
 	static DEVICE_ATTR(name, S_IRUGO, show_##name, NULL)
 
-VBD_SHOW(oo_req,  "%d\n", be->blkif->st_oo_req);
-VBD_SHOW(rd_req,  "%d\n", be->blkif->st_rd_req);
-VBD_SHOW(wr_req,  "%d\n", be->blkif->st_wr_req);
-VBD_SHOW(f_req,  "%d\n", be->blkif->st_f_req);
-VBD_SHOW(ds_req,  "%d\n", be->blkif->st_ds_req);
-VBD_SHOW(rd_sect, "%d\n", be->blkif->st_rd_sect);
-VBD_SHOW(wr_sect, "%d\n", be->blkif->st_wr_sect);
+VBD_SHOW(oo_req,  "%llu\n", be->blkif->st_oo_req);
+VBD_SHOW(rd_req,  "%llu\n", be->blkif->st_rd_req);
+VBD_SHOW(wr_req,  "%llu\n", be->blkif->st_wr_req);
+VBD_SHOW(f_req,  "%llu\n", be->blkif->st_f_req);
+VBD_SHOW(ds_req,  "%llu\n", be->blkif->st_ds_req);
+VBD_SHOW(rd_sect, "%llu\n", be->blkif->st_rd_sect);
+VBD_SHOW(wr_sect, "%llu\n", be->blkif->st_wr_sect);
 
 static struct attribute *xen_vbdstat_attrs[] = {
 	&dev_attr_oo_req.attr,

drivers/block/xen-blkfront.c

@@ -44,7 +44,7 @@
 #include <linux/mutex.h>
 #include <linux/scatterlist.h>
 #include <linux/bitmap.h>
-#include <linux/llist.h>
+#include <linux/list.h>
 
 #include <xen/xen.h>
 #include <xen/xenbus.h>
@@ -68,13 +68,12 @@ enum blkif_state {
 struct grant {
 	grant_ref_t gref;
 	unsigned long pfn;
-	struct llist_node node;
+	struct list_head node;
 };
 
 struct blk_shadow {
 	struct blkif_request req;
 	struct request *request;
-	unsigned long frame[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 	struct grant *grants_used[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 };
 
@@ -105,7 +104,7 @@ struct blkfront_info
 	struct work_struct work;
 	struct gnttab_free_callback callback;
 	struct blk_shadow shadow[BLK_RING_SIZE];
-	struct llist_head persistent_gnts;
+	struct list_head persistent_gnts;
 	unsigned int persistent_gnts_c;
 	unsigned long shadow_free;
 	unsigned int feature_flush;
@@ -165,6 +164,69 @@ static int add_id_to_freelist(struct blkfront_info *info,
 	return 0;
 }
 
+static int fill_grant_buffer(struct blkfront_info *info, int num)
+{
+	struct page *granted_page;
+	struct grant *gnt_list_entry, *n;
+	int i = 0;
+
+	while(i < num) {
+		gnt_list_entry = kzalloc(sizeof(struct grant), GFP_NOIO);
+		if (!gnt_list_entry)
+			goto out_of_memory;
+
+		granted_page = alloc_page(GFP_NOIO);
+		if (!granted_page) {
+			kfree(gnt_list_entry);
+			goto out_of_memory;
+		}
+
+		gnt_list_entry->pfn = page_to_pfn(granted_page);
+		gnt_list_entry->gref = GRANT_INVALID_REF;
+		list_add(&gnt_list_entry->node, &info->persistent_gnts);
+		i++;
+	}
+
+	return 0;
+
+out_of_memory:
+	list_for_each_entry_safe(gnt_list_entry, n,
+	                         &info->persistent_gnts, node) {
+		list_del(&gnt_list_entry->node);
+		__free_page(pfn_to_page(gnt_list_entry->pfn));
+		kfree(gnt_list_entry);
+		i--;
+	}
+	BUG_ON(i != 0);
+	return -ENOMEM;
+}
+
+static struct grant *get_grant(grant_ref_t *gref_head,
+                               struct blkfront_info *info)
+{
+	struct grant *gnt_list_entry;
+	unsigned long buffer_mfn;
+
+	BUG_ON(list_empty(&info->persistent_gnts));
+	gnt_list_entry = list_first_entry(&info->persistent_gnts, struct grant,
+	                                  node);
+	list_del(&gnt_list_entry->node);
+
+	if (gnt_list_entry->gref != GRANT_INVALID_REF) {
+		info->persistent_gnts_c--;
+		return gnt_list_entry;
+	}
+
+	/* Assign a gref to this page */
+	gnt_list_entry->gref = gnttab_claim_grant_reference(gref_head);
+	BUG_ON(gnt_list_entry->gref == -ENOSPC);
+	buffer_mfn = pfn_to_mfn(gnt_list_entry->pfn);
+	gnttab_grant_foreign_access_ref(gnt_list_entry->gref,
+	                                info->xbdev->otherend_id,
+	                                buffer_mfn, 0);
+	return gnt_list_entry;
+}
+
 static const char *op_name(int op)
 {
 	static const char *const names[] = {
@@ -293,7 +355,6 @@ static int blkif_ioctl(struct block_device *bdev, fmode_t mode,
 static int blkif_queue_request(struct request *req)
 {
 	struct blkfront_info *info = req->rq_disk->private_data;
-	unsigned long buffer_mfn;
 	struct blkif_request *ring_req;
 	unsigned long id;
 	unsigned int fsect, lsect;
@@ -306,7 +367,6 @@ static int blkif_queue_request(struct request *req)
 	 */
 	bool new_persistent_gnts;
 	grant_ref_t gref_head;
-	struct page *granted_page;
 	struct grant *gnt_list_entry = NULL;
 	struct scatterlist *sg;
 
@@ -370,41 +430,8 @@ static int blkif_queue_request(struct request *req)
 			fsect = sg->offset >> 9;
 			lsect = fsect + (sg->length >> 9) - 1;
 
-			if (info->persistent_gnts_c) {
-				BUG_ON(llist_empty(&info->persistent_gnts));
-				gnt_list_entry = llist_entry(
-					llist_del_first(&info->persistent_gnts),
-					struct grant, node);
-
-				ref = gnt_list_entry->gref;
-				buffer_mfn = pfn_to_mfn(gnt_list_entry->pfn);
-				info->persistent_gnts_c--;
-			} else {
-				ref = gnttab_claim_grant_reference(&gref_head);
-				BUG_ON(ref == -ENOSPC);
-
-				gnt_list_entry =
-					kmalloc(sizeof(struct grant),
-							 GFP_ATOMIC);
-				if (!gnt_list_entry)
-					return -ENOMEM;
-
-				granted_page = alloc_page(GFP_ATOMIC);
-				if (!granted_page) {
-					kfree(gnt_list_entry);
-					return -ENOMEM;
-				}
-
-				gnt_list_entry->pfn =
-					page_to_pfn(granted_page);
-				gnt_list_entry->gref = ref;
-
-				buffer_mfn = pfn_to_mfn(page_to_pfn(
-								granted_page));
-				gnttab_grant_foreign_access_ref(ref,
-					info->xbdev->otherend_id,
-					buffer_mfn, 0);
-			}
+			gnt_list_entry = get_grant(&gref_head, info);
+			ref = gnt_list_entry->gref;
 
 			info->shadow[id].grants_used[i] = gnt_list_entry;
 
@@ -435,7 +462,6 @@ static int blkif_queue_request(struct request *req)
 				kunmap_atomic(shared_data);
 			}
 
-			info->shadow[id].frame[i] = mfn_to_pfn(buffer_mfn);
 			ring_req->u.rw.seg[i] =
 					(struct blkif_request_segment) {
 						.gref       = ref,
@@ -790,9 +816,8 @@ static void blkif_restart_queue(struct work_struct *work)
 
 static void blkif_free(struct blkfront_info *info, int suspend)
 {
-	struct llist_node *all_gnts;
-	struct grant *persistent_gnt, *tmp;
-	struct llist_node *n;
+	struct grant *persistent_gnt;
+	struct grant *n;
 
 	/* Prevent new requests being issued until we fix things up. */
 	spin_lock_irq(&info->io_lock);
@@ -803,22 +828,20 @@ static void blkif_free(struct blkfront_info *info, int suspend)
 		blk_stop_queue(info->rq);
 
 	/* Remove all persistent grants */
-	if (info->persistent_gnts_c) {
-		all_gnts = llist_del_all(&info->persistent_gnts);
-		persistent_gnt = llist_entry(all_gnts, typeof(*(persistent_gnt)), node);
-		while (persistent_gnt) {
-			gnttab_end_foreign_access(persistent_gnt->gref, 0, 0UL);
+	if (!list_empty(&info->persistent_gnts)) {
+		list_for_each_entry_safe(persistent_gnt, n,
+		                         &info->persistent_gnts, node) {
+			list_del(&persistent_gnt->node);
+			if (persistent_gnt->gref != GRANT_INVALID_REF) {
+				gnttab_end_foreign_access(persistent_gnt->gref,
+				                          0, 0UL);
+				info->persistent_gnts_c--;
+			}
 			__free_page(pfn_to_page(persistent_gnt->pfn));
-			tmp = persistent_gnt;
-			n = persistent_gnt->node.next;
-			if (n)
-				persistent_gnt = llist_entry(n, typeof(*(persistent_gnt)), node);
-			else
-				persistent_gnt = NULL;
-			kfree(tmp);
+			kfree(persistent_gnt);
 		}
-		info->persistent_gnts_c = 0;
 	}
+	BUG_ON(info->persistent_gnts_c != 0);
 
 	/* No more gnttab callback work. */
 	gnttab_cancel_free_callback(&info->callback);
@@ -875,7 +898,7 @@ static void blkif_completion(struct blk_shadow *s, struct blkfront_info *info,
 	}
 	/* Add the persistent grant into the list of free grants */
 	for (i = 0; i < s->req.u.rw.nr_segments; i++) {
-		llist_add(&s->grants_used[i]->node, &info->persistent_gnts);
+		list_add(&s->grants_used[i]->node, &info->persistent_gnts);
 		info->persistent_gnts_c++;
 	}
 }
@@ -1013,6 +1036,12 @@ static int setup_blkring(struct xenbus_device *dev,
 
 	sg_init_table(info->sg, BLKIF_MAX_SEGMENTS_PER_REQUEST);
 
+	/* Allocate memory for grants */
+	err = fill_grant_buffer(info, BLK_RING_SIZE *
+	                              BLKIF_MAX_SEGMENTS_PER_REQUEST);
+	if (err)
+		goto fail;
+
 	err = xenbus_grant_ring(dev, virt_to_mfn(info->ring.sring));
 	if (err < 0) {
 		free_page((unsigned long)sring);
@@ -1171,7 +1200,7 @@ static int blkfront_probe(struct xenbus_device *dev,
 	spin_lock_init(&info->io_lock);
 	info->xbdev = dev;
 	info->vdevice = vdevice;
-	init_llist_head(&info->persistent_gnts);
+	INIT_LIST_HEAD(&info->persistent_gnts);
 	info->persistent_gnts_c = 0;
 	info->connected = BLKIF_STATE_DISCONNECTED;
 	INIT_WORK(&info->work, blkif_restart_queue);
@@ -1203,11 +1232,10 @@ static int blkif_recover(struct blkfront_info *info)
 	int j;
 
 	/* Stage 1: Make a safe copy of the shadow state. */
-	copy = kmalloc(sizeof(info->shadow),
+	copy = kmemdup(info->shadow, sizeof(info->shadow),
 		       GFP_NOIO | __GFP_REPEAT | __GFP_HIGH);
 	if (!copy)
 		return -ENOMEM;
-	memcpy(copy, info->shadow, sizeof(info->shadow));
 
 	/* Stage 2: Set up free list. */
 	memset(&info->shadow, 0, sizeof(info->shadow));
@@ -1236,7 +1264,7 @@ static int blkif_recover(struct blkfront_info *info)
 				gnttab_grant_foreign_access_ref(
 					req->u.rw.seg[j].gref,
 					info->xbdev->otherend_id,
-					pfn_to_mfn(info->shadow[req->u.rw.id].frame[j]),
+					pfn_to_mfn(copy[i].grants_used[j]->pfn),
 					0);
 		}
 		info->shadow[req->u.rw.id].req = *req;

include/xen/interface/io/blkif.h

@@ -138,11 +138,21 @@ struct blkif_request_discard {
 	uint8_t        _pad3;
 } __attribute__((__packed__));
 
+struct blkif_request_other {
+	uint8_t      _pad1;
+	blkif_vdev_t _pad2;        /* only for read/write requests         */
+#ifdef CONFIG_X86_64
+	uint32_t     _pad3;        /* offsetof(blkif_req..,u.other.id)==8*/
+#endif
+	uint64_t     id;           /* private guest value, echoed in resp  */
+} __attribute__((__packed__));
+
 struct blkif_request {
 	uint8_t        operation;    /* BLKIF_OP_???                         */
 	union {
 		struct blkif_request_rw rw;
 		struct blkif_request_discard discard;
+		struct blkif_request_other other;
 	} u;
 } __attribute__((__packed__));