Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

Pull s390 updates from Martin Schwidefsky:
 "The first part of the s390 updates for 4.14:

   - Add machine type 0x3906 for IBM z14

   - Add IBM z14 TLB flushing improvements for KVM guests

   - Exploit the TOD clock epoch extension to provide a continuous TOD
     clock afer 2042/09/17

   - Add NIAI spinlock hints for IBM z14

   - Rework the vmcp driver and use CMA for the respone buffer of z/VM
     CP commands

   - Drop some s390 specific asm headers and use the generic version

   - Add block discard for DASD-FBA devices under z/VM

   - Add average request times to DASD statistics

   - A few of those constify patches which seem to be in vogue right now

   - Cleanup and bug fixes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (50 commits)
  s390/mm: avoid empty zero pages for KVM guests to avoid postcopy hangs
  s390/dasd: Add discard support for FBA devices
  s390/zcrypt: make CPRBX const
  s390/uaccess: avoid mvcos jump label
  s390/mm: use generic mm_hooks
  s390/facilities: fix typo
  s390/vmcp: simplify vmcp_response_free()
  s390/topology: Remove the unused parent_node() macro
  s390/dasd: Change unsigned long long to unsigned long
  s390/smp: convert cpuhp_setup_state() return code to zero on success
  s390: fix 'novx' early parameter handling
  s390/dasd: add average request times to dasd statistics
  s390/scm: use common completion path
  s390/pci: log changes to uid checking
  s390/vmcp: simplify vmcp_ioctl()
  s390/vmcp: return -ENOTTY for unknown ioctl commands
  s390/vmcp: split vmcp header file and move to uapi
  s390/vmcp: make use of contiguous memory allocator
  s390/cpcmd,vmcp: avoid GFP_DMA allocations
  s390/vmcp: fix uaccess check and avoid undefined behavior
  ...

26 files changed, 398 insertions, 102 deletions

diff --git a/drivers/s390/block/dasd.c b/drivers/s390/block/dasd.c
index 670ac0a4ef49..9c97ad1ee121 100644
--- a/drivers/s390/block/dasd.c
+++ b/drivers/s390/block/dasd.c
@@ -801,11 +801,12 @@ static void dasd_profile_end(struct dasd_block *block,
 			     struct dasd_ccw_req *cqr,
 			     struct request *req)
 {
-	long strtime, irqtime, endtime, tottime;	/* in microseconds */
-	long tottimeps, sectors;
+	unsigned long strtime, irqtime, endtime, tottime;
+	unsigned long tottimeps, sectors;
 	struct dasd_device *device;
 	int sectors_ind, tottime_ind, tottimeps_ind, strtime_ind;
 	int irqtime_ind, irqtimeps_ind, endtime_ind;
+	struct dasd_profile_info *data;
 
 	device = cqr->startdev;
 	if (!(dasd_global_profile_level ||
@@ -835,6 +836,11 @@ static void dasd_profile_end(struct dasd_block *block,
 
 	spin_lock(&dasd_global_profile.lock);
 	if (dasd_global_profile.data) {
+		data = dasd_global_profile.data;
+		data->dasd_sum_times += tottime;
+		data->dasd_sum_time_str += strtime;
+		data->dasd_sum_time_irq += irqtime;
+		data->dasd_sum_time_end += endtime;
 		dasd_profile_end_add_data(dasd_global_profile.data,
 					  cqr->startdev != block->base,
 					  cqr->cpmode == 1,
@@ -847,7 +853,12 @@ static void dasd_profile_end(struct dasd_block *block,
 	spin_unlock(&dasd_global_profile.lock);
 
 	spin_lock(&block->profile.lock);
-	if (block->profile.data)
+	if (block->profile.data) {
+		data = block->profile.data;
+		data->dasd_sum_times += tottime;
+		data->dasd_sum_time_str += strtime;
+		data->dasd_sum_time_irq += irqtime;
+		data->dasd_sum_time_end += endtime;
 		dasd_profile_end_add_data(block->profile.data,
 					  cqr->startdev != block->base,
 					  cqr->cpmode == 1,
@@ -856,10 +867,16 @@ static void dasd_profile_end(struct dasd_block *block,
 					  tottimeps_ind, strtime_ind,
 					  irqtime_ind, irqtimeps_ind,
 					  endtime_ind);
+	}
 	spin_unlock(&block->profile.lock);
 
 	spin_lock(&device->profile.lock);
-	if (device->profile.data)
+	if (device->profile.data) {
+		data = device->profile.data;
+		data->dasd_sum_times += tottime;
+		data->dasd_sum_time_str += strtime;
+		data->dasd_sum_time_irq += irqtime;
+		data->dasd_sum_time_end += endtime;
 		dasd_profile_end_add_data(device->profile.data,
 					  cqr->startdev != block->base,
 					  cqr->cpmode == 1,
@@ -868,6 +885,7 @@ static void dasd_profile_end(struct dasd_block *block,
 					  tottimeps_ind, strtime_ind,
 					  irqtime_ind, irqtimeps_ind,
 					  endtime_ind);
+	}
 	spin_unlock(&device->profile.lock);
 }
 
@@ -989,6 +1007,14 @@ static void dasd_stats_seq_print(struct seq_file *m,
 	seq_printf(m, "total_sectors %u\n", data->dasd_io_sects);
 	seq_printf(m, "total_pav %u\n", data->dasd_io_alias);
 	seq_printf(m, "total_hpf %u\n", data->dasd_io_tpm);
+	seq_printf(m, "avg_total %lu\n", data->dasd_io_reqs ?
+		   data->dasd_sum_times / data->dasd_io_reqs : 0UL);
+	seq_printf(m, "avg_build_to_ssch %lu\n", data->dasd_io_reqs ?
+		   data->dasd_sum_time_str / data->dasd_io_reqs : 0UL);
+	seq_printf(m, "avg_ssch_to_irq %lu\n", data->dasd_io_reqs ?
+		   data->dasd_sum_time_irq / data->dasd_io_reqs : 0UL);
+	seq_printf(m, "avg_irq_to_end %lu\n", data->dasd_io_reqs ?
+		   data->dasd_sum_time_end / data->dasd_io_reqs : 0UL);
 	seq_puts(m, "histogram_sectors ");
 	dasd_stats_array(m, data->dasd_io_secs);
 	seq_puts(m, "histogram_io_times ");
@@ -1639,7 +1665,7 @@ void dasd_int_handler(struct ccw_device *cdev, unsigned long intparm,
 {
 	struct dasd_ccw_req *cqr, *next;
 	struct dasd_device *device;
-	unsigned long long now;
+	unsigned long now;
 	int nrf_suppressed = 0;
 	int fp_suppressed = 0;
 	u8 *sense = NULL;
@@ -3152,7 +3178,9 @@ static int dasd_alloc_queue(struct dasd_block *block)
  */
 static void dasd_setup_queue(struct dasd_block *block)
 {
+	unsigned int logical_block_size = block->bp_block;
 	struct request_queue *q = block->request_queue;
+	unsigned int max_bytes, max_discard_sectors;
 	int max;
 
 	if (block->base->features & DASD_FEATURE_USERAW) {
@@ -3169,7 +3197,7 @@ static void dasd_setup_queue(struct dasd_block *block)
 	}
 	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, q);
 	q->limits.max_dev_sectors = max;
-	blk_queue_logical_block_size(q, block->bp_block);
+	blk_queue_logical_block_size(q, logical_block_size);
 	blk_queue_max_hw_sectors(q, max);
 	blk_queue_max_segments(q, USHRT_MAX);
 	/* with page sized segments we can translate each segement into
@@ -3177,6 +3205,21 @@ static void dasd_setup_queue(struct dasd_block *block)
 	 */
 	blk_queue_max_segment_size(q, PAGE_SIZE);
 	blk_queue_segment_boundary(q, PAGE_SIZE - 1);
+
+	/* Only activate blocklayer discard support for devices that support it */
+	if (block->base->features & DASD_FEATURE_DISCARD) {
+		q->limits.discard_granularity = logical_block_size;
+		q->limits.discard_alignment = PAGE_SIZE;
+
+		/* Calculate max_discard_sectors and make it PAGE aligned */
+		max_bytes = USHRT_MAX * logical_block_size;
+		max_bytes = ALIGN(max_bytes, PAGE_SIZE) - PAGE_SIZE;
+		max_discard_sectors = max_bytes / logical_block_size;
+
+		blk_queue_max_discard_sectors(q, max_discard_sectors);
+		blk_queue_max_write_zeroes_sectors(q, max_discard_sectors);
+		queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, q);
+	}
 }
 
 /*
diff --git a/drivers/s390/block/dasd_3990_erp.c b/drivers/s390/block/dasd_3990_erp.c
index 107cd3361e29..e448a0fc0c09 100644
--- a/drivers/s390/block/dasd_3990_erp.c
+++ b/drivers/s390/block/dasd_3990_erp.c
@@ -2231,7 +2231,7 @@ static void dasd_3990_erp_account_error(struct dasd_ccw_req *erp)
 	struct dasd_device *device = erp->startdev;
 	__u8 lpum = erp->refers->irb.esw.esw1.lpum;
 	int pos = pathmask_to_pos(lpum);
-	unsigned long long clk;
+	unsigned long clk;
 
 	if (!device->path_thrhld)
 		return;
diff --git a/drivers/s390/block/dasd_devmap.c b/drivers/s390/block/dasd_devmap.c
index 779dce069cc5..e38042ce94e6 100644
--- a/drivers/s390/block/dasd_devmap.c
+++ b/drivers/s390/block/dasd_devmap.c
@@ -1634,7 +1634,7 @@ static struct attribute * dasd_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group dasd_attr_group = {
+static const struct attribute_group dasd_attr_group = {
 	.attrs = dasd_attrs,
 };
 
@@ -1676,6 +1676,7 @@ dasd_set_feature(struct ccw_device *cdev, int feature, int flag)
 	spin_unlock(&dasd_devmap_lock);
 	return 0;
 }
+EXPORT_SYMBOL(dasd_set_feature);
 
 
 int
diff --git a/drivers/s390/block/dasd_diag.c b/drivers/s390/block/dasd_diag.c
index 5667146c6a0a..98fb28e49d2c 100644
--- a/drivers/s390/block/dasd_diag.c
+++ b/drivers/s390/block/dasd_diag.c
@@ -235,7 +235,7 @@ static void dasd_ext_handler(struct ext_code ext_code,
 {
 	struct dasd_ccw_req *cqr, *next;
 	struct dasd_device *device;
-	unsigned long long expires;
+	unsigned long expires;
 	unsigned long flags;
 	addr_t ip;
 	int rc;
diff --git a/drivers/s390/block/dasd_eckd.c b/drivers/s390/block/dasd_eckd.c
index c3e5ad641b0b..8eafcd5fa004 100644
--- a/drivers/s390/block/dasd_eckd.c
+++ b/drivers/s390/block/dasd_eckd.c
@@ -3254,11 +3254,7 @@ static struct dasd_ccw_req *dasd_eckd_build_cp_cmd_track(
 	/* 1x prefix + one read/write ccw per track */
 	cplength = 1 + trkcount;
 
-	/* on 31-bit we need space for two 32 bit addresses per page
-	 * on 64-bit one 64 bit address
-	 */
-	datasize = sizeof(struct PFX_eckd_data) +
-		cidaw * sizeof(unsigned long long);
+	datasize = sizeof(struct PFX_eckd_data) + cidaw * sizeof(unsigned long);
 
 	/* Allocate the ccw request. */
 	cqr = dasd_smalloc_request(DASD_ECKD_MAGIC, cplength, datasize,
@@ -3856,7 +3852,7 @@ static struct dasd_ccw_req *dasd_eckd_build_cp_raw(struct dasd_device *startdev,
 	}
 	size = ALIGN(size, 8);
 
-	datasize = size + cidaw * sizeof(unsigned long long);
+	datasize = size + cidaw * sizeof(unsigned long);
 
 	/* Allocate the ccw request. */
 	cqr = dasd_smalloc_request(DASD_ECKD_MAGIC, cplength,
diff --git a/drivers/s390/block/dasd_eckd.h b/drivers/s390/block/dasd_eckd.h
index fb1f537d986a..34e153a6b19c 100644
--- a/drivers/s390/block/dasd_eckd.h
+++ b/drivers/s390/block/dasd_eckd.h
@@ -165,7 +165,7 @@ struct DE_eckd_data {
 	__u8 ga_extended;	/* Global Attributes Extended	*/
 	struct ch_t beg_ext;
 	struct ch_t end_ext;
-	unsigned long long ep_sys_time; /* Ext Parameter - System Time Stamp */
+	unsigned long ep_sys_time; /* Ext Parameter - System Time Stamp */
 	__u8 ep_format;        /* Extended Parameter format byte       */
 	__u8 ep_prio;          /* Extended Parameter priority I/O byte */
 	__u8 ep_reserved1;     /* Extended Parameter Reserved	       */
diff --git a/drivers/s390/block/dasd_erp.c b/drivers/s390/block/dasd_erp.c
index 9e3419124264..6389feb2fb7a 100644
--- a/drivers/s390/block/dasd_erp.c
+++ b/drivers/s390/block/dasd_erp.c
@@ -124,7 +124,7 @@ dasd_default_erp_action(struct dasd_ccw_req *cqr)
 struct dasd_ccw_req *dasd_default_erp_postaction(struct dasd_ccw_req *cqr)
 {
 	int success;
-	unsigned long long startclk, stopclk;
+	unsigned long startclk, stopclk;
 	struct dasd_device *startdev;
 
 	BUG_ON(cqr->refers == NULL || cqr->function == NULL);
diff --git a/drivers/s390/block/dasd_fba.c b/drivers/s390/block/dasd_fba.c
index 462cab5d4302..6168ccdb389c 100644
--- a/drivers/s390/block/dasd_fba.c
+++ b/drivers/s390/block/dasd_fba.c
@@ -174,6 +174,9 @@ dasd_fba_check_characteristics(struct dasd_device *device)
 	if (readonly)
 		set_bit(DASD_FLAG_DEVICE_RO, &device->flags);
 
+	/* FBA supports discard, set the according feature bit */
+	dasd_set_feature(cdev, DASD_FEATURE_DISCARD, 1);
+
 	dev_info(&device->cdev->dev,
 		 "New FBA DASD %04X/%02X (CU %04X/%02X) with %d MB "
 		 "and %d B/blk%s\n",
@@ -247,9 +250,192 @@ static void dasd_fba_check_for_device_change(struct dasd_device *device,
 		dasd_generic_handle_state_change(device);
 };
 
-static struct dasd_ccw_req *dasd_fba_build_cp(struct dasd_device * memdev,
-					      struct dasd_block *block,
-					      struct request *req)
+
+/*
+ * Builds a CCW with no data payload
+ */
+static void ccw_write_no_data(struct ccw1 *ccw)
+{
+	ccw->cmd_code = DASD_FBA_CCW_WRITE;
+	ccw->flags |= CCW_FLAG_SLI;
+	ccw->count = 0;
+}
+
+/*
+ * Builds a CCW that writes only zeroes.
+ */
+static void ccw_write_zero(struct ccw1 *ccw, int count)
+{
+	ccw->cmd_code = DASD_FBA_CCW_WRITE;
+	ccw->flags |= CCW_FLAG_SLI;
+	ccw->count = count;
+	ccw->cda = (__u32) (addr_t) page_to_phys(ZERO_PAGE(0));
+}
+
+/*
+ * Helper function to count the amount of necessary CCWs within a given range
+ * with 4k alignment and command chaining in mind.
+ */
+static int count_ccws(sector_t first_rec, sector_t last_rec,
+		      unsigned int blocks_per_page)
+{
+	sector_t wz_stop = 0, d_stop = 0;
+	int cur_pos = 0;
+	int count = 0;
+
+	if (first_rec % blocks_per_page != 0) {
+		wz_stop = first_rec + blocks_per_page -
+			(first_rec % blocks_per_page) - 1;
+		if (wz_stop > last_rec)
+			wz_stop = last_rec;
+		cur_pos = wz_stop - first_rec + 1;
+		count++;
+	}
+
+	if (last_rec - (first_rec + cur_pos) + 1 >= blocks_per_page) {
+		if ((last_rec - blocks_per_page + 1) % blocks_per_page != 0)
+			d_stop = last_rec - ((last_rec - blocks_per_page + 1) %
+					     blocks_per_page);
+		else
+			d_stop = last_rec;
+
+		cur_pos += d_stop - (first_rec + cur_pos) + 1;
+		count++;
+	}
+
+	if (cur_pos == 0 || first_rec + cur_pos - 1 < last_rec)
+		count++;
+
+	return count;
+}
+
+/*
+ * This function builds a CCW request for block layer discard requests.
+ * Each page in the z/VM hypervisor that represents certain records of an FBA
+ * device will be padded with zeros. This is a special behaviour of the WRITE
+ * command which is triggered when no data payload is added to the CCW.
+ *
+ * Note: Due to issues in some z/VM versions, we can't fully utilise this
+ * special behaviour. We have to keep a 4k (or 8 block) alignment in mind to
+ * work around those issues and write actual zeroes to the unaligned parts in
+ * the request. This workaround might be removed in the future.
+ */
+static struct dasd_ccw_req *dasd_fba_build_cp_discard(
+						struct dasd_device *memdev,
+						struct dasd_block *block,
+						struct request *req)
+{
+	struct LO_fba_data *LO_data;
+	struct dasd_ccw_req *cqr;
+	struct ccw1 *ccw;
+
+	sector_t wz_stop = 0, d_stop = 0;
+	sector_t first_rec, last_rec;
+
+	unsigned int blksize = block->bp_block;
+	unsigned int blocks_per_page;
+	int wz_count = 0;
+	int d_count = 0;
+	int cur_pos = 0; /* Current position within the extent */
+	int count = 0;
+	int cplength;
+	int datasize;
+	int nr_ccws;
+
+	first_rec = blk_rq_pos(req) >> block->s2b_shift;
+	last_rec =
+		(blk_rq_pos(req) + blk_rq_sectors(req) - 1) >> block->s2b_shift;
+	count = last_rec - first_rec + 1;
+
+	blocks_per_page = BLOCKS_PER_PAGE(blksize);
+	nr_ccws = count_ccws(first_rec, last_rec, blocks_per_page);
+
+	/* define extent + nr_ccws * locate record + nr_ccws * single CCW */
+	cplength = 1 + 2 * nr_ccws;
+	datasize = sizeof(struct DE_fba_data) +
+		nr_ccws * (sizeof(struct LO_fba_data) + sizeof(struct ccw1));
+
+	cqr = dasd_smalloc_request(DASD_FBA_MAGIC, cplength, datasize, memdev);
+	if (IS_ERR(cqr))
+		return cqr;
+
+	ccw = cqr->cpaddr;
+
+	define_extent(ccw++, cqr->data, WRITE, blksize, first_rec, count);
+	LO_data = cqr->data + sizeof(struct DE_fba_data);
+
+	/* First part is not aligned. Calculate range to write zeroes. */
+	if (first_rec % blocks_per_page != 0) {
+		wz_stop = first_rec + blocks_per_page -
+			(first_rec % blocks_per_page) - 1;
+		if (wz_stop > last_rec)
+			wz_stop = last_rec;
+		wz_count = wz_stop - first_rec + 1;
+
+		ccw[-1].flags |= CCW_FLAG_CC;
+		locate_record(ccw++, LO_data++, WRITE, cur_pos, wz_count);
+
+		ccw[-1].flags |= CCW_FLAG_CC;
+		ccw_write_zero(ccw++, wz_count * blksize);
+
+		cur_pos = wz_count;
+	}
+
+	/* We can do proper discard when we've got at least blocks_per_page blocks. */
+	if (last_rec - (first_rec + cur_pos) + 1 >= blocks_per_page) {
+		/* is last record at page boundary? */
+		if ((last_rec - blocks_per_page + 1) % blocks_per_page != 0)
+			d_stop = last_rec - ((last_rec - blocks_per_page + 1) %
+					     blocks_per_page);
+		else
+			d_stop = last_rec;
+
+		d_count = d_stop - (first_rec + cur_pos) + 1;
+
+		ccw[-1].flags |= CCW_FLAG_CC;
+		locate_record(ccw++, LO_data++, WRITE, cur_pos, d_count);
+
+		ccw[-1].flags |= CCW_FLAG_CC;
+		ccw_write_no_data(ccw++);
+
+		cur_pos += d_count;
+	}
+
+	/* We might still have some bits left which need to be zeroed. */
+	if (cur_pos == 0 || first_rec + cur_pos - 1 < last_rec) {
+		if (d_stop != 0)
+			wz_count = last_rec - d_stop;
+		else if (wz_stop != 0)
+			wz_count = last_rec - wz_stop;
+		else
+			wz_count = count;
+
+		ccw[-1].flags |= CCW_FLAG_CC;
+		locate_record(ccw++, LO_data++, WRITE, cur_pos, wz_count);
+
+		ccw[-1].flags |= CCW_FLAG_CC;
+		ccw_write_zero(ccw++, wz_count * blksize);
+	}
+
+	if (blk_noretry_request(req) ||
+	    block->base->features & DASD_FEATURE_FAILFAST)
+		set_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags);
+
+	cqr->startdev = memdev;
+	cqr->memdev = memdev;
+	cqr->block = block;
+	cqr->expires = memdev->default_expires * HZ;	/* default 5 minutes */
+	cqr->retries = memdev->default_retries;
+	cqr->buildclk = get_tod_clock();
+	cqr->status = DASD_CQR_FILLED;
+
+	return cqr;
+}
+
+static struct dasd_ccw_req *dasd_fba_build_cp_regular(
+						struct dasd_device *memdev,
+						struct dasd_block *block,
+						struct request *req)
 {
 	struct dasd_fba_private *private = block->base->private;
 	unsigned long *idaws;
@@ -372,6 +558,16 @@ static struct dasd_ccw_req *dasd_fba_build_cp(struct dasd_device * memdev,
 	return cqr;
 }
 
+static struct dasd_ccw_req *dasd_fba_build_cp(struct dasd_device *memdev,
+					      struct dasd_block *block,
+					      struct request *req)
+{
+	if (req_op(req) == REQ_OP_DISCARD || req_op(req) == REQ_OP_WRITE_ZEROES)
+		return dasd_fba_build_cp_discard(memdev, block, req);
+	else
+		return dasd_fba_build_cp_regular(memdev, block, req);
+}
+
 static int
 dasd_fba_free_cp(struct dasd_ccw_req *cqr, struct request *req)
 {
diff --git a/drivers/s390/block/dasd_int.h b/drivers/s390/block/dasd_int.h
index dca7cb1e6f65..f9e25fc03d6b 100644
--- a/drivers/s390/block/dasd_int.h
+++ b/drivers/s390/block/dasd_int.h
@@ -167,6 +167,9 @@ do { \
 	printk(d_loglevel PRINTK_HEADER " " d_string "\n", d_args); \
 } while(0)
 
+/* Macro to calculate number of blocks per page */
+#define BLOCKS_PER_PAGE(blksize) (PAGE_SIZE / blksize)
+
 struct dasd_ccw_req {
 	unsigned int magic;		/* Eye catcher */
 	struct list_head devlist;	/* for dasd_device request queue */
@@ -196,10 +199,10 @@ struct dasd_ccw_req {
 	void *function; 		/* originating ERP action */
 
 	/* these are for statistics only */
-	unsigned long long buildclk;	/* TOD-clock of request generation */
-	unsigned long long startclk;	/* TOD-clock of request start */
-	unsigned long long stopclk;	/* TOD-clock of request interrupt */
-	unsigned long long endclk;	/* TOD-clock of request termination */
+	unsigned long buildclk;		/* TOD-clock of request generation */
+	unsigned long startclk;		/* TOD-clock of request start */
+	unsigned long stopclk;		/* TOD-clock of request interrupt */
+	unsigned long endclk;		/* TOD-clock of request termination */
 
         /* Callback that is called after reaching final status. */
 	void (*callback)(struct dasd_ccw_req *, void *data);
@@ -423,7 +426,7 @@ struct dasd_path {
 	u8 chpid;
 	struct dasd_conf_data *conf_data;
 	atomic_t error_count;
-	unsigned long long errorclk;
+	unsigned long errorclk;
 };
 
 
@@ -454,6 +457,10 @@ struct dasd_profile_info {
 	unsigned int dasd_read_time2[32];  /* hist. of time from start to irq */
 	unsigned int dasd_read_time3[32];  /* hist. of time from irq to end */
 	unsigned int dasd_read_nr_req[32]; /* hist. of # of requests in chanq */
+	unsigned long dasd_sum_times;	   /* sum of request times */
+	unsigned long dasd_sum_time_str;   /* sum of time from build to start */
+	unsigned long dasd_sum_time_irq;   /* sum of time from start to irq */
+	unsigned long dasd_sum_time_end;   /* sum of time from irq to end */
 };
 
 struct dasd_profile {
@@ -535,7 +542,7 @@ struct dasd_block {
 	struct block_device *bdev;
 	atomic_t open_count;
 
-	unsigned long long blocks; /* size of volume in blocks */
+	unsigned long blocks;	   /* size of volume in blocks */
 	unsigned int bp_block;	   /* bytes per block */
 	unsigned int s2b_shift;	   /* log2 (bp_block/512) */
 
diff --git a/drivers/s390/block/dasd_proc.c b/drivers/s390/block/dasd_proc.c
index 70dc2c4cd3f7..7104d6765773 100644
--- a/drivers/s390/block/dasd_proc.c
+++ b/drivers/s390/block/dasd_proc.c
@@ -90,7 +90,7 @@ dasd_devices_show(struct seq_file *m, void *v)
 			seq_printf(m, "n/f	 ");
 		else
 			seq_printf(m,
-				   "at blocksize: %d, %lld blocks, %lld MB",
+				   "at blocksize: %u, %lu blocks, %lu MB",
 				   block->bp_block, block->blocks,
 				   ((block->bp_block >> 9) *
 				    block->blocks) >> 11);
diff --git a/drivers/s390/block/scm_blk.c b/drivers/s390/block/scm_blk.c
index 0071febac9e6..2e7fd966c515 100644
--- a/drivers/s390/block/scm_blk.c
+++ b/drivers/s390/block/scm_blk.c
@@ -249,13 +249,13 @@ static void scm_request_requeue(struct scm_request *scmrq)
 static void scm_request_finish(struct scm_request *scmrq)
 {
 	struct scm_blk_dev *bdev = scmrq->bdev;
+	int *error;
 	int i;
 
 	for (i = 0; i < nr_requests_per_io && scmrq->request[i]; i++) {
-		if (scmrq->error)
-			blk_mq_end_request(scmrq->request[i], scmrq->error);
-		else
-			blk_mq_complete_request(scmrq->request[i]);
+		error = blk_mq_rq_to_pdu(scmrq->request[i]);
+		*error = scmrq->error;
+		blk_mq_complete_request(scmrq->request[i]);
 	}
 
 	atomic_dec(&bdev->queued_reqs);
@@ -415,7 +415,9 @@ void scm_blk_irq(struct scm_device *scmdev, void *data, blk_status_t error)
 
 static void scm_blk_request_done(struct request *req)
 {
-	blk_mq_end_request(req, 0);
+	int *error = blk_mq_rq_to_pdu(req);
+
+	blk_mq_end_request(req, *error);
 }
 
 static const struct block_device_operations scm_blk_devops = {
@@ -448,6 +450,7 @@ int scm_blk_dev_setup(struct scm_blk_dev *bdev, struct scm_device *scmdev)
 	atomic_set(&bdev->queued_reqs, 0);
 
 	bdev->tag_set.ops = &scm_mq_ops;
+	bdev->tag_set.cmd_size = sizeof(int);
 	bdev->tag_set.nr_hw_queues = nr_requests;
 	bdev->tag_set.queue_depth = nr_requests_per_io * nr_requests;
 	bdev->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
diff --git a/drivers/s390/char/Kconfig b/drivers/s390/char/Kconfig
index b3f1c458905f..97c4c9fdd53d 100644
--- a/drivers/s390/char/Kconfig
+++ b/drivers/s390/char/Kconfig
@@ -169,10 +169,21 @@ config VMCP
 	def_bool y
 	prompt "Support for the z/VM CP interface"
 	depends on S390
+	select CMA
 	help
 	  Select this option if you want to be able to interact with the control
 	  program on z/VM
 
+config VMCP_CMA_SIZE
+	int "Memory in MiB reserved for z/VM CP interface"
+	default "4"
+	depends on VMCP
+	help
+	  Specify the default amount of memory in MiB reserved for the z/VM CP
+	  interface. If needed this memory is used for large contiguous memory
+	  allocations. The default can be changed with the kernel command line
+	  parameter "vmcp_cma".
+
 config MONREADER
 	def_tristate m
 	prompt "API for reading z/VM monitor service records"
diff --git a/drivers/s390/char/raw3270.c b/drivers/s390/char/raw3270.c
index 710f2292911d..5d4f053d7c38 100644
--- a/drivers/s390/char/raw3270.c
+++ b/drivers/s390/char/raw3270.c
@@ -1082,7 +1082,7 @@ static struct attribute * raw3270_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group raw3270_attr_group = {
+static const struct attribute_group raw3270_attr_group = {
 	.attrs = raw3270_attrs,
 };
 
diff --git a/drivers/s390/char/sclp_cmd.c b/drivers/s390/char/sclp_cmd.c
index b9c5522b8a68..dff8b94871f0 100644
--- a/drivers/s390/char/sclp_cmd.c
+++ b/drivers/s390/char/sclp_cmd.c
@@ -252,6 +252,7 @@ static int sclp_attach_storage(u8 id)
 	if (!sccb)
 		return -ENOMEM;
 	sccb->header.length = PAGE_SIZE;
+	sccb->header.function_code = 0x40;
 	rc = sclp_sync_request_timeout(0x00080001 | id << 8, sccb,
 				       SCLP_QUEUE_INTERVAL);
 	if (rc)
diff --git a/drivers/s390/char/sclp_config.c b/drivers/s390/char/sclp_config.c
index 1406fb688a26..7003d52c2191 100644
--- a/drivers/s390/char/sclp_config.c
+++ b/drivers/s390/char/sclp_config.c
@@ -135,7 +135,7 @@ static ssize_t sysfs_ofb_data_write(struct file *filp, struct kobject *kobj,
 	return rc ?: count;
 }
 
-static struct bin_attribute ofb_bin_attr = {
+static const struct bin_attribute ofb_bin_attr = {
 	.attr = {
 		.name = "event_data",
 		.mode = S_IWUSR,
diff --git a/drivers/s390/char/sclp_early.c b/drivers/s390/char/sclp_early.c
index efd84d1d178b..bc1fc00910b0 100644
--- a/drivers/s390/char/sclp_early.c
+++ b/drivers/s390/char/sclp_early.c
@@ -39,7 +39,7 @@ struct read_info_sccb {
 	u8	fac84;			/* 84 */
 	u8	fac85;			/* 85 */
 	u8	_pad_86[91 - 86];	/* 86-90 */
-	u8	flags;			/* 91 */
+	u8	fac91;			/* 91 */
 	u8	_pad_92[98 - 92];	/* 92-97 */
 	u8	fac98;			/* 98 */
 	u8	hamaxpow;		/* 99 */
@@ -103,6 +103,8 @@ static void __init sclp_early_facilities_detect(struct read_info_sccb *sccb)
 	sclp.has_kss = !!(sccb->fac98 & 0x01);
 	if (sccb->fac85 & 0x02)
 		S390_lowcore.machine_flags |= MACHINE_FLAG_ESOP;
+	if (sccb->fac91 & 0x40)
+		S390_lowcore.machine_flags |= MACHINE_FLAG_TLB_GUEST;
 	sclp.rnmax = sccb->rnmax ? sccb->rnmax : sccb->rnmax2;
 	sclp.rzm = sccb->rnsize ? sccb->rnsize : sccb->rnsize2;
 	sclp.rzm <<= 20;
@@ -139,7 +141,7 @@ static void __init sclp_early_facilities_detect(struct read_info_sccb *sccb)
 
 	/* Save IPL information */
 	sclp_ipl_info.is_valid = 1;
-	if (sccb->flags & 0x2)
+	if (sccb->fac91 & 0x2)
 		sclp_ipl_info.has_dump = 1;
 	memcpy(&sclp_ipl_info.loadparm, &sccb->loadparm, LOADPARM_LEN);
 
diff --git a/drivers/s390/char/sclp_ocf.c b/drivers/s390/char/sclp_ocf.c
index f59b71776bbd..f9cbb1ab047b 100644
--- a/drivers/s390/char/sclp_ocf.c
+++ b/drivers/s390/char/sclp_ocf.c
@@ -126,7 +126,7 @@ static struct attribute *ocf_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group ocf_attr_group = {
+static const struct attribute_group ocf_attr_group = {
 	.attrs = ocf_attrs,
 };
 
diff --git a/drivers/s390/char/tape_core.c b/drivers/s390/char/tape_core.c
index 3c379da2eef8..9dd4534823b3 100644
--- a/drivers/s390/char/tape_core.c
+++ b/drivers/s390/char/tape_core.c
@@ -175,7 +175,7 @@ static struct attribute *tape_attrs[] = {
 	NULL
 };
 
-static struct attribute_group tape_attr_group = {
+static const struct attribute_group tape_attr_group = {
 	.attrs = tape_attrs,
 };
 
diff --git a/drivers/s390/char/vmcp.c b/drivers/s390/char/vmcp.c
index 98749fa817da..7898bbcc28fc 100644
--- a/drivers/s390/char/vmcp.c
+++ b/drivers/s390/char/vmcp.c
@@ -17,15 +17,85 @@
 #include <linux/kernel.h>
 #include <linux/miscdevice.h>
 #include <linux/slab.h>
+#include <linux/uaccess.h>
 #include <linux/export.h>
+#include <linux/mutex.h>
+#include <linux/cma.h>
+#include <linux/mm.h>
 #include <asm/compat.h>
 #include <asm/cpcmd.h>
 #include <asm/debug.h>
-#include <linux/uaccess.h>
-#include "vmcp.h"
+#include <asm/vmcp.h>
+
+struct vmcp_session {
+	char *response;
+	unsigned int bufsize;
+	unsigned int cma_alloc : 1;
+	int resp_size;
+	int resp_code;
+	struct mutex mutex;
+};
 
 static debug_info_t *vmcp_debug;
 
+static unsigned long vmcp_cma_size __initdata = CONFIG_VMCP_CMA_SIZE * 1024 * 1024;
+static struct cma *vmcp_cma;
+
+static int __init early_parse_vmcp_cma(char *p)
+{
+	vmcp_cma_size = ALIGN(memparse(p, NULL), PAGE_SIZE);
+	return 0;
+}
+early_param("vmcp_cma", early_parse_vmcp_cma);
+
+void __init vmcp_cma_reserve(void)
+{
+	if (!MACHINE_IS_VM)
+		return;
+	cma_declare_contiguous(0, vmcp_cma_size, 0, 0, 0, false, "vmcp", &vmcp_cma);
+}
+
+static void vmcp_response_alloc(struct vmcp_session *session)
+{
+	struct page *page = NULL;
+	int nr_pages, order;
+
+	order = get_order(session->bufsize);
+	nr_pages = ALIGN(session->bufsize, PAGE_SIZE) >> PAGE_SHIFT;
+	/*
+	 * For anything below order 3 allocations rely on the buddy
+	 * allocator. If such low-order allocations can't be handled
+	 * anymore the system won't work anyway.
+	 */
+	if (order > 2)
+		page = cma_alloc(vmcp_cma, nr_pages, 0, GFP_KERNEL);
+	if (page) {
+		session->response = (char *)page_to_phys(page);
+		session->cma_alloc = 1;
+		return;
+	}
+	session->response = (char *)__get_free_pages(GFP_KERNEL | __GFP_RETRY_MAYFAIL, order);
+}
+
+static void vmcp_response_free(struct vmcp_session *session)
+{
+	int nr_pages, order;
+	struct page *page;
+
+	if (!session->response)
+		return;
+	order = get_order(session->bufsize);
+	nr_pages = ALIGN(session->bufsize, PAGE_SIZE) >> PAGE_SHIFT;
+	if (session->cma_alloc) {
+		page = phys_to_page((unsigned long)session->response);
+		cma_release(vmcp_cma, page, nr_pages);
+		session->cma_alloc = 0;
+	} else {
+		free_pages((unsigned long)session->response, order);
+	}
+	session->response = NULL;
+}
+
 static int vmcp_open(struct inode *inode, struct file *file)
 {
 	struct vmcp_session *session;
@@ -51,7 +121,7 @@ static int vmcp_release(struct inode *inode, struct file *file)
 
 	session = file->private_data;
 	file->private_data = NULL;
-	free_pages((unsigned long)session->response, get_order(session->bufsize));
+	vmcp_response_free(session);
 	kfree(session);
 	return 0;
 }
@@ -97,9 +167,7 @@ vmcp_write(struct file *file, const char __user *buff, size_t count,
 		return -ERESTARTSYS;
 	}
 	if (!session->response)
-		session->response = (char *)__get_free_pages(GFP_KERNEL
-						| __GFP_RETRY_MAYFAIL | GFP_DMA,
-						get_order(session->bufsize));
+		vmcp_response_alloc(session);
 	if (!session->response) {
 		mutex_unlock(&session->mutex);
 		kfree(cmd);
@@ -130,8 +198,8 @@ vmcp_write(struct file *file, const char __user *buff, size_t count,
 static long vmcp_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 {
 	struct vmcp_session *session;
+	int ret = -ENOTTY;
 	int __user *argp;
-	int temp;
 
 	session = file->private_data;
 	if (is_compat_task())
@@ -142,28 +210,26 @@ static long vmcp_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
 		return -ERESTARTSYS;
 	switch (cmd) {
 	case VMCP_GETCODE:
-		temp = session->resp_code;
-		mutex_unlock(&session->mutex);
-		return put_user(temp, argp);
+		ret = put_user(session->resp_code, argp);
+		break;
 	case VMCP_SETBUF:
-		free_pages((unsigned long)session->response,
-				get_order(session->bufsize));
-		session->response=NULL;
-		temp = get_user(session->bufsize, argp);
-		if (get_order(session->bufsize) > 8) {
+		vmcp_response_free(session);
+		ret = get_user(session->bufsize, argp);
+		if (ret)
 			session->bufsize = PAGE_SIZE;
-			temp = -EINVAL;
+		if (!session->bufsize || get_order(session->bufsize) > 8) {
+			session->bufsize = PAGE_SIZE;
+			ret = -EINVAL;
 		}
-		mutex_unlock(&session->mutex);
-		return temp;
+		break;
 	case VMCP_GETSIZE:
-		temp = session->resp_size;
-		mutex_unlock(&session->mutex);
-		return put_user(temp, argp);
+		ret = put_user(session->resp_size, argp);
+		break;
 	default:
-		mutex_unlock(&session->mutex);
-		return -ENOIOCTLCMD;
+		break;
 	}
+	mutex_unlock(&session->mutex);
+	return ret;
 }
 
 static const struct file_operations vmcp_fops = {
diff --git a/drivers/s390/char/vmcp.h b/drivers/s390/char/vmcp.h
deleted file mode 100644
index 1e29b0418382..000000000000
--- a/drivers/s390/char/vmcp.h
+++ /dev/null
@@ -1,30 +0,0 @@
-/*
- * Copyright IBM Corp. 2004, 2005
- * Interface implementation for communication with the z/VM control program
- * Version 1.0
- * Author(s): Christian Borntraeger <cborntra@de.ibm.com>
- *
- *
- * z/VMs CP offers the possibility to issue commands via the diagnose code 8
- * this driver implements a character device that issues these commands and
- * returns the answer of CP.
- *
- * The idea of this driver is based on cpint from Neale Ferguson
- */
-
-#include <linux/ioctl.h>
-#include <linux/mutex.h>
-
-#define VMCP_GETCODE _IOR(0x10, 1, int)
-#define VMCP_SETBUF _IOW(0x10, 2, int)
-#define VMCP_GETSIZE _IOR(0x10, 3, int)
-
-struct vmcp_session {
-	unsigned int bufsize;
-	char *response;
-	int resp_size;
-	int resp_code;
-	/* As we use copy_from/to_user, which might     *
-	 * sleep and cannot use a spinlock              */
-	struct mutex mutex;
-};
diff --git a/drivers/s390/cio/chp.c b/drivers/s390/cio/chp.c
index 432fc40990bd..f4166f80c4d4 100644
--- a/drivers/s390/cio/chp.c
+++ b/drivers/s390/cio/chp.c
@@ -143,7 +143,7 @@ static ssize_t chp_measurement_chars_read(struct file *filp,
 				       sizeof(chp->cmg_chars));
 }
 
-static struct bin_attribute chp_measurement_chars_attr = {
+static const struct bin_attribute chp_measurement_chars_attr = {
 	.attr = {
 		.name = "measurement_chars",
 		.mode = S_IRUSR,
@@ -197,7 +197,7 @@ static ssize_t chp_measurement_read(struct file *filp, struct kobject *kobj,
 	return count;
 }
 
-static struct bin_attribute chp_measurement_attr = {
+static const struct bin_attribute chp_measurement_attr = {
 	.attr = {
 		.name = "measurement",
 		.mode = S_IRUSR,
diff --git a/drivers/s390/cio/device.c b/drivers/s390/cio/device.c
index 7be01a58b44f..489b583f263d 100644
--- a/drivers/s390/cio/device.c
+++ b/drivers/s390/cio/device.c
@@ -612,7 +612,7 @@ static struct attribute *io_subchannel_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group io_subchannel_attr_group = {
+static const struct attribute_group io_subchannel_attr_group = {
 	.attrs = io_subchannel_attrs,
 };
 
@@ -626,7 +626,7 @@ static struct attribute * ccwdev_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group ccwdev_attr_group = {
+static const struct attribute_group ccwdev_attr_group = {
 	.attrs = ccwdev_attrs,
 };
 
diff --git a/drivers/s390/crypto/zcrypt_card.c b/drivers/s390/crypto/zcrypt_card.c
index 53436ea52230..f85dacf1c284 100644
--- a/drivers/s390/crypto/zcrypt_card.c
+++ b/drivers/s390/crypto/zcrypt_card.c
@@ -98,7 +98,7 @@ static struct attribute *zcrypt_card_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group zcrypt_card_attr_group = {
+static const struct attribute_group zcrypt_card_attr_group = {
 	.attrs = zcrypt_card_attrs,
 };
 
diff --git a/drivers/s390/crypto/zcrypt_msgtype6.c b/drivers/s390/crypto/zcrypt_msgtype6.c
index 4fddb4319481..afd20cee7ea0 100644
--- a/drivers/s390/crypto/zcrypt_msgtype6.c
+++ b/drivers/s390/crypto/zcrypt_msgtype6.c
@@ -140,7 +140,7 @@ struct function_and_rules_block {
  *   + 0x000A 'MRP     ' (MCL3 'PK' or CEX2C 'PK')
  * - VUD block
  */
-static struct CPRBX static_cprbx = {
+static const struct CPRBX static_cprbx = {
 	.cprb_len	=  0x00DC,
 	.cprb_ver_id	=  0x02,
 	.func_id	= {0x54, 0x32},
diff --git a/drivers/s390/crypto/zcrypt_queue.c b/drivers/s390/crypto/zcrypt_queue.c
index a303f3b2c328..4742be0eec24 100644
--- a/drivers/s390/crypto/zcrypt_queue.c
+++ b/drivers/s390/crypto/zcrypt_queue.c
@@ -89,7 +89,7 @@ static struct attribute *zcrypt_queue_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group zcrypt_queue_attr_group = {
+static const struct attribute_group zcrypt_queue_attr_group = {
 	.attrs = zcrypt_queue_attrs,
 };
 
diff --git a/drivers/s390/net/qeth_l3_sys.c b/drivers/s390/net/qeth_l3_sys.c
index f2f94f59e0fa..1a80ce41425e 100644
--- a/drivers/s390/net/qeth_l3_sys.c
+++ b/drivers/s390/net/qeth_l3_sys.c
@@ -350,7 +350,7 @@ static struct attribute *qeth_l3_device_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group qeth_l3_device_attr_group = {
+static const struct attribute_group qeth_l3_device_attr_group = {
 	.attrs = qeth_l3_device_attrs,
 };
 
@@ -680,7 +680,7 @@ static struct attribute *qeth_ipato_device_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group qeth_device_ipato_group = {
+static const struct attribute_group qeth_device_ipato_group = {
 	.name = "ipa_takeover",
 	.attrs = qeth_ipato_device_attrs,
 };
@@ -843,7 +843,7 @@ static struct attribute *qeth_vipa_device_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group qeth_device_vipa_group = {
+static const struct attribute_group qeth_device_vipa_group = {
 	.name = "vipa",
 	.attrs = qeth_vipa_device_attrs,
 };
@@ -1006,7 +1006,7 @@ static struct attribute *qeth_rxip_device_attrs[] = {
 	NULL,
 };
 
-static struct attribute_group qeth_device_rxip_group = {
+static const struct attribute_group qeth_device_rxip_group = {
 	.name = "rxip",
 	.attrs = qeth_rxip_device_attrs,
 };