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

Pull more s390 updates from Martin Schwidefsky:
 "The big thing in this second merge for s390 is the new eBPF JIT from
  Michael which replaces the old 32-bit backend.

  The remaining commits are bug fixes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/pci: add locking for fmb access
  s390/pci: extract software counters from fmb
  s390/dasd: Fix unresumed device after suspend/resume having no paths
  s390/dasd: fix unresumed device after suspend/resume
  s390/dasd: fix inability to set a DASD device offline
  s390/mm: Fix memory hotplug for unaligned standby memory
  s390/bpf: Add s390x eBPF JIT compiler backend
  s390: Use bool function return values of true/false not 1/0

arch/s390/Kconfig

@@ -115,7 +115,7 @@ config S390
 	select HAVE_ARCH_SECCOMP_FILTER
 	select HAVE_ARCH_TRACEHOOK
 	select HAVE_ARCH_TRANSPARENT_HUGEPAGE
-	select HAVE_BPF_JIT if PACK_STACK
+	select HAVE_BPF_JIT if PACK_STACK && HAVE_MARCH_Z9_109_FEATURES
 	select HAVE_CMPXCHG_DOUBLE
 	select HAVE_CMPXCHG_LOCAL
 	select HAVE_DEBUG_KMEMLEAK

arch/s390/include/asm/dma-mapping.h

@@ -42,7 +42,7 @@ static inline int dma_supported(struct device *dev, u64 mask)
 static inline bool dma_capable(struct device *dev, dma_addr_t addr, size_t size)
 {
 	if (!dev->dma_mask)
-		return 0;
+		return false;
 	return addr + size - 1 <= *dev->dma_mask;
 }
 

arch/s390/include/asm/pci.h

@@ -7,6 +7,7 @@
 #define PCI_BAR_COUNT	6
 
 #include <linux/pci.h>
+#include <linux/mutex.h>
 #include <asm-generic/pci.h>
 #include <asm-generic/pci-dma-compat.h>
 #include <asm/pci_clp.h>
@@ -44,10 +45,6 @@ struct zpci_fmb {
 	u64 rpcit_ops;
 	u64 dma_rbytes;
 	u64 dma_wbytes;
-	/* software counters */
-	atomic64_t allocated_pages;
-	atomic64_t mapped_pages;
-	atomic64_t unmapped_pages;
 } __packed __aligned(16);
 
 enum zpci_state {
@@ -80,6 +77,7 @@ struct zpci_dev {
 	u8		pft;		/* pci function type */
 	u16		domain;
 
+	struct mutex lock;
 	u8 pfip[CLP_PFIP_NR_SEGMENTS];	/* pci function internal path */
 	u32 uid;			/* user defined id */
 	u8 util_str[CLP_UTIL_STR_LEN];	/* utility string */
@@ -111,6 +109,10 @@ struct zpci_dev {
 	/* Function measurement block */
 	struct zpci_fmb *fmb;
 	u16		fmb_update;	/* update interval */
+	/* software counters */
+	atomic64_t allocated_pages;
+	atomic64_t mapped_pages;
+	atomic64_t unmapped_pages;
 
 	enum pci_bus_speed max_bus_speed;
 

arch/s390/net/bpf_jit.S

 /*
  * BPF Jit compiler for s390, help functions.
  *
- * Copyright IBM Corp. 2012
+ * Copyright IBM Corp. 2012,2015
  *
  * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *	      Michael Holzheu <holzheu@linux.vnet.ibm.com>
  */
+
 #include <linux/linkage.h>
+#include "bpf_jit.h"
 
 /*
  * Calling convention:
- * registers %r2, %r6-%r8, %r10-%r11, %r13, %r15 are call saved
- *   %r2: skb pointer
- *   %r3: offset parameter
- *   %r5: BPF A accumulator
- *   %r8: return address
- *   %r9: save register for skb pointer
- *   %r10: skb->data
- *   %r11: skb->len - skb->data_len (headlen)
- *   %r12: BPF X accumulator
+ * registers %r7-%r10, %r11,%r13, and %r15 are call saved
+ *
+ * Input (64 bit):
+ *   %r3 (%b2) = offset into skb data
+ *   %r6 (%b5) = return address
+ *   %r7 (%b6) = skb pointer
+ *   %r12      = skb data pointer
+ *
+ * Output:
+ *   %r14= %b0 = return value (read skb value)
+ *
+ * Work registers: %r2,%r4,%r5,%r14
  *
  * skb_copy_bits takes 4 parameters:
  *   %r2 = skb pointer
  *   %r3 = offset into skb data
  *   %r4 = pointer to temp buffer
  *   %r5 = length to copy
+ *   Return value in %r2: 0 = ok
+ *
+ * bpf_internal_load_pointer_neg_helper takes 3 parameters:
+ *   %r2 = skb pointer
+ *   %r3 = offset into data
+ *   %r4 = length to copy
+ *   Return value in %r2: Pointer to data
  */
-#define SKBDATA	%r8
-
-	/* A = *(u32 *) (skb->data+K+X) */
-ENTRY(sk_load_word_ind)
-	ar	%r3,%r12		# offset += X
-	bmr	%r8			# < 0 -> return with cc
-
-	/* A = *(u32 *) (skb->data+K) */
-ENTRY(sk_load_word)
-	llgfr	%r1,%r3			# extend offset
-	ahi	%r3,4			# offset + 4
-	clr	%r11,%r3		# hlen <= offset + 4 ?
-	jl	sk_load_word_slow
-	l	%r5,0(%r1,%r10)		# get word from skb
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
 
-sk_load_word_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,160(%r15)		# pointer to temp buffer
-	lghi	%r5,4			# 4 bytes
-	brasl	%r14,skb_copy_bits	# get data from skb
-	l	%r5,160(%r15)		# load result from temp buffer
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
+#define SKF_MAX_NEG_OFF	-0x200000	/* SKF_LL_OFF from filter.h */
 
-	/* A = *(u16 *) (skb->data+K+X) */
-ENTRY(sk_load_half_ind)
-	ar	%r3,%r12		# offset += X
-	bmr	%r8			# < 0 -> return with cc
-
-	/* A = *(u16 *) (skb->data+K) */
-ENTRY(sk_load_half)
-	llgfr	%r1,%r3			# extend offset
-	ahi	%r3,2			# offset + 2
-	clr	%r11,%r3		# hlen <= offset + 2 ?
-	jl	sk_load_half_slow
-	llgh	%r5,0(%r1,%r10)		# get half from skb
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
-
-sk_load_half_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,162(%r15)		# pointer to temp buffer
-	lghi	%r5,2			# 2 bytes
-	brasl	%r14,skb_copy_bits	# get data from skb
-	xc	160(2,%r15),160(%r15)
-	l	%r5,160(%r15)		# load result from temp buffer
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
+/*
+ * Load SIZE bytes from SKB
+ */
+#define sk_load_common(NAME, SIZE, LOAD)				\
+ENTRY(sk_load_##NAME);							\
+	ltgr	%r3,%r3;		/* Is offset negative? */	\
+	jl	sk_load_##NAME##_slow_neg;				\
+ENTRY(sk_load_##NAME##_pos);						\
+	aghi	%r3,SIZE;		/* Offset + SIZE */		\
+	clg	%r3,STK_OFF_HLEN(%r15);	/* Offset + SIZE > hlen? */	\
+	jh	sk_load_##NAME##_slow;					\
+	LOAD	%r14,-SIZE(%r3,%r12);	/* Get data from skb */		\
+	b	OFF_OK(%r6);		/* Return */			\
+									\
+sk_load_##NAME##_slow:;							\
+	lgr	%r2,%r7;		/* Arg1 = skb pointer */	\
+	aghi	%r3,-SIZE;		/* Arg2 = offset */		\
+	la	%r4,STK_OFF_TMP(%r15);	/* Arg3 = temp bufffer */	\
+	lghi	%r5,SIZE;		/* Arg4 = size */		\
+	brasl	%r14,skb_copy_bits;	/* Get data from skb */		\
+	LOAD	%r14,STK_OFF_TMP(%r15);	/* Load from temp bufffer */	\
+	ltgr	%r2,%r2;		/* Set cc to (%r2 != 0) */	\
+	br	%r6;			/* Return */
 
-	/* A = *(u8 *) (skb->data+K+X) */
-ENTRY(sk_load_byte_ind)
-	ar	%r3,%r12		# offset += X
-	bmr	%r8			# < 0 -> return with cc
+sk_load_common(word, 4, llgf)	/* r14 = *(u32 *) (skb->data+offset) */
+sk_load_common(half, 2, llgh)	/* r14 = *(u16 *) (skb->data+offset) */
 
-	/* A = *(u8 *) (skb->data+K) */
+/*
+ * Load 1 byte from SKB (optimized version)
+ */
+	/* r14 = *(u8 *) (skb->data+offset) */
 ENTRY(sk_load_byte)
-	llgfr	%r1,%r3			# extend offset
-	clr	%r11,%r3		# hlen < offset ?
-	jle	sk_load_byte_slow
-	lhi	%r5,0
-	ic	%r5,0(%r1,%r10)		# get byte from skb
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
+	ltgr	%r3,%r3			# Is offset negative?
+	jl	sk_load_byte_slow_neg
+ENTRY(sk_load_byte_pos)
+	clg	%r3,STK_OFF_HLEN(%r15)	# Offset >= hlen?
+	jnl	sk_load_byte_slow
+	llgc	%r14,0(%r3,%r12)	# Get byte from skb
+	b	OFF_OK(%r6)		# Return OK
 
 sk_load_byte_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,163(%r15)		# pointer to temp buffer
-	lghi	%r5,1			# 1 byte
-	brasl	%r14,skb_copy_bits	# get data from skb
-	xc	160(3,%r15),160(%r15)
-	l	%r5,160(%r15)		# load result from temp buffer
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
+	lgr	%r2,%r7			# Arg1 = skb pointer
+					# Arg2 = offset
+	la	%r4,STK_OFF_TMP(%r15)	# Arg3 = pointer to temp buffer
+	lghi	%r5,1			# Arg4 = size (1 byte)
+	brasl	%r14,skb_copy_bits	# Get data from skb
+	llgc	%r14,STK_OFF_TMP(%r15)	# Load result from temp buffer
+	ltgr	%r2,%r2			# Set cc to (%r2 != 0)
+	br	%r6			# Return cc
+
+#define sk_negative_common(NAME, SIZE, LOAD)				\
+sk_load_##NAME##_slow_neg:;						\
+	cgfi	%r3,SKF_MAX_NEG_OFF;					\
+	jl	bpf_error;						\
+	lgr	%r2,%r7;		/* Arg1 = skb pointer */	\
+					/* Arg2 = offset */		\
+	lghi	%r4,SIZE;		/* Arg3 = size */		\
+	brasl	%r14,bpf_internal_load_pointer_neg_helper;		\
+	ltgr	%r2,%r2;						\
+	jz	bpf_error;						\
+	LOAD	%r14,0(%r2);		/* Get data from pointer */	\
+	xr	%r3,%r3;		/* Set cc to zero */		\
+	br	%r6;			/* Return cc */
 
-	/* X = (*(u8 *)(skb->data+K) & 0xf) << 2 */
-ENTRY(sk_load_byte_msh)
-	llgfr	%r1,%r3			# extend offset
-	clr	%r11,%r3		# hlen < offset ?
-	jle	sk_load_byte_msh_slow
-	lhi	%r12,0
-	ic	%r12,0(%r1,%r10)	# get byte from skb
-	nill	%r12,0x0f
-	sll	%r12,2
-	xr	%r1,%r1			# set cc to zero
-	br	%r8
+sk_negative_common(word, 4, llgf)
+sk_negative_common(half, 2, llgh)
+sk_negative_common(byte, 1, llgc)
 
-sk_load_byte_msh_slow:
-	lgr	%r9,%r2			# save %r2
-	lgr	%r3,%r1			# offset
-	la	%r4,163(%r15)		# pointer to temp buffer
-	lghi	%r5,1			# 1 byte
-	brasl	%r14,skb_copy_bits	# get data from skb
-	xc	160(3,%r15),160(%r15)
-	l	%r12,160(%r15)		# load result from temp buffer
-	nill	%r12,0x0f
-	sll	%r12,2
-	ltgr	%r2,%r2			# set cc to (%r2 != 0)
-	lgr	%r2,%r9			# restore %r2
-	br	%r8
+bpf_error:
+# force a return 0 from jit handler
+	ltgr	%r15,%r15	# Set condition code
+	br	%r6

arch/s390/net/bpf_jit.h

+/*
+ * BPF Jit compiler defines
+ *
+ * Copyright IBM Corp. 2012,2015
+ *
+ * Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ *	      Michael Holzheu <holzheu@linux.vnet.ibm.com>
+ */
+
+#ifndef __ARCH_S390_NET_BPF_JIT_H
+#define __ARCH_S390_NET_BPF_JIT_H
+
+#ifndef __ASSEMBLY__
+
+#include <linux/filter.h>
+#include <linux/types.h>
+
+extern u8 sk_load_word_pos[], sk_load_half_pos[], sk_load_byte_pos[];
+extern u8 sk_load_word[], sk_load_half[], sk_load_byte[];
+
+#endif /* __ASSEMBLY__ */
+
+/*
+ * Stackframe layout (packed stack):
+ *
+ *				    ^ high
+ *	      +---------------+     |
+ *	      | old backchain |     |
+ *	      +---------------+     |
+ *	      |   r15 - r6    |     |
+ * BFP	   -> +===============+     |
+ *	      |		      |     |
+ *	      |   BPF stack   |     |
+ *	      |		      |     |
+ *	      +---------------+     |
+ *	      | 8 byte hlen   |     |
+ * R15+168 -> +---------------+     |
+ *	      | 4 byte align  |     |
+ *	      +---------------+     |
+ *	      | 4 byte temp   |     |
+ *	      | for bpf_jit.S |     |
+ * R15+160 -> +---------------+     |
+ *	      | new backchain |     |
+ * R15+152 -> +---------------+     |
+ *	      | + 152 byte SA |     |
+ * R15	   -> +---------------+     + low
+ *
+ * We get 160 bytes stack space from calling function, but only use
+ * 11 * 8 byte (old backchain + r15 - r6) for storing registers.
+ */
+#define STK_OFF (MAX_BPF_STACK + 8 + 4 + 4 + (160 - 11 * 8))
+#define STK_OFF_TMP	160	/* Offset of tmp buffer on stack */
+#define STK_OFF_HLEN	168	/* Offset of SKB header length on stack */
+
+/* Offset to skip condition code check */
+#define OFF_OK		4
+
+#endif /* __ARCH_S390_NET_BPF_JIT_H */

arch/s390/pci/pci.c

@@ -190,6 +190,11 @@ int zpci_fmb_enable_device(struct zpci_dev *zdev)
 		return -ENOMEM;
 	WARN_ON((u64) zdev->fmb & 0xf);
 
+	/* reset software counters */
+	atomic64_set(&zdev->allocated_pages, 0);
+	atomic64_set(&zdev->mapped_pages, 0);
+	atomic64_set(&zdev->unmapped_pages, 0);
+
 	args.fmb_addr = virt_to_phys(zdev->fmb);
 	return mod_pci(zdev, ZPCI_MOD_FC_SET_MEASURE, 0, &args);
 }
@@ -822,6 +827,7 @@ int zpci_create_device(struct zpci_dev *zdev)
 	if (rc)
 		goto out;
 
+	mutex_init(&zdev->lock);
 	if (zdev->state == ZPCI_FN_STATE_CONFIGURED) {
 		rc = zpci_enable_device(zdev);
 		if (rc)

arch/s390/pci/pci_debug.c

@@ -31,12 +31,25 @@ static char *pci_perf_names[] = {
 	"Refresh operations",
 	"DMA read bytes",
 	"DMA write bytes",
-	/* software counters */
+};
+
+static char *pci_sw_names[] = {
 	"Allocated pages",
 	"Mapped pages",
 	"Unmapped pages",
 };
 
+static void pci_sw_counter_show(struct seq_file *m)
+{
+	struct zpci_dev *zdev = m->private;
+	atomic64_t *counter = &zdev->allocated_pages;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(pci_sw_names); i++, counter++)
+		seq_printf(m, "%26s:\t%llu\n", pci_sw_names[i],
+			   atomic64_read(counter));
+}
+
 static int pci_perf_show(struct seq_file *m, void *v)
 {
 	struct zpci_dev *zdev = m->private;
@@ -45,7 +58,10 @@ static int pci_perf_show(struct seq_file *m, void *v)
 
 	if (!zdev)
 		return 0;
+
+	mutex_lock(&zdev->lock);
 	if (!zdev->fmb) {
+		mutex_unlock(&zdev->lock);
 		seq_puts(m, "FMB statistics disabled\n");
 		return 0;
 	}
@@ -65,12 +81,9 @@ static int pci_perf_show(struct seq_file *m, void *v)
 		for (i = 4; i < 6; i++)
 			seq_printf(m, "%26s:\t%llu\n",
 				   pci_perf_names[i], *(stat + i));
-	/* software counters */
-	for (i = 6; i < ARRAY_SIZE(pci_perf_names); i++)
-		seq_printf(m, "%26s:\t%llu\n",
-			   pci_perf_names[i],
-			   atomic64_read((atomic64_t *) (stat + i)));
 
+	pci_sw_counter_show(m);
+	mutex_unlock(&zdev->lock);
 	return 0;
 }
 
@@ -88,19 +101,17 @@ static ssize_t pci_perf_seq_write(struct file *file, const char __user *ubuf,
 	if (rc)
 		return rc;
 
+	mutex_lock(&zdev->lock);
 	switch (val) {
 	case 0:
 		rc = zpci_fmb_disable_device(zdev);
-		if (rc)
-			return rc;
 		break;
 	case 1:
 		rc = zpci_fmb_enable_device(zdev);
-		if (rc)
-			return rc;
 		break;
 	}
-	return count;
+	mutex_unlock(&zdev->lock);
+	return rc ? rc : count;
 }
 
 static int pci_perf_seq_open(struct inode *inode, struct file *filp)

arch/s390/pci/pci_dma.c

@@ -300,7 +300,7 @@ static dma_addr_t s390_dma_map_pages(struct device *dev, struct page *page,
 		flags |= ZPCI_TABLE_PROTECTED;
 
 	if (!dma_update_trans(zdev, pa, dma_addr, size, flags)) {
-		atomic64_add(nr_pages, &zdev->fmb->mapped_pages);
+		atomic64_add(nr_pages, &zdev->mapped_pages);
 		return dma_addr + (offset & ~PAGE_MASK);
 	}
 
@@ -328,7 +328,7 @@ static void s390_dma_unmap_pages(struct device *dev, dma_addr_t dma_addr,
 		zpci_err_hex(&dma_addr, sizeof(dma_addr));
 	}
 
-	atomic64_add(npages, &zdev->fmb->unmapped_pages);
+	atomic64_add(npages, &zdev->unmapped_pages);
 	iommu_page_index = (dma_addr - zdev->start_dma) >> PAGE_SHIFT;
 	dma_free_iommu(zdev, iommu_page_index, npages);
 }
@@ -357,7 +357,7 @@ static void *s390_dma_alloc(struct device *dev, size_t size,
 		return NULL;
 	}
 
-	atomic64_add(size / PAGE_SIZE, &zdev->fmb->allocated_pages);
+	atomic64_add(size / PAGE_SIZE, &zdev->allocated_pages);
 	if (dma_handle)
 		*dma_handle = map;
 	return (void *) pa;
@@ -370,7 +370,7 @@ static void s390_dma_free(struct device *dev, size_t size,
 	struct zpci_dev *zdev = get_zdev(to_pci_dev(dev));
 
 	size = PAGE_ALIGN(size);
-	atomic64_sub(size / PAGE_SIZE, &zdev->fmb->allocated_pages);
+	atomic64_sub(size / PAGE_SIZE, &zdev->allocated_pages);
 	s390_dma_unmap_pages(dev, dma_handle, size, DMA_BIDIRECTIONAL, NULL);
 	free_pages((unsigned long) pa, get_order(size));
 }

drivers/s390/block/dasd.c

@@ -579,7 +579,8 @@ void dasd_kick_device(struct dasd_device *device)
 {
 	dasd_get_device(device);
 	/* queue call to dasd_kick_device to the kernel event daemon. */
-	schedule_work(&device->kick_work);
+	if (!schedule_work(&device->kick_work))
+		dasd_put_device(device);
 }
 EXPORT_SYMBOL(dasd_kick_device);
 
@@ -599,7 +600,8 @@ void dasd_reload_device(struct dasd_device *device)
 {
 	dasd_get_device(device);
 	/* queue call to dasd_reload_device to the kernel event daemon. */
-	schedule_work(&device->reload_device);
+	if (!schedule_work(&device->reload_device))
+		dasd_put_device(device);
 }
 EXPORT_SYMBOL(dasd_reload_device);
 
@@ -619,7 +621,8 @@ void dasd_restore_device(struct dasd_device *device)
 {
 	dasd_get_device(device);
 	/* queue call to dasd_restore_device to the kernel event daemon. */
-	schedule_work(&device->restore_device);
+	if (!schedule_work(&device->restore_device))
+		dasd_put_device(device);
 }
 
 /*
@@ -2163,18 +2166,22 @@ static int _dasd_sleep_on(struct dasd_ccw_req *maincqr, int interruptible)
 			cqr->intrc = -ENOLINK;
 			continue;
 		}
-		/* Don't try to start requests if device is stopped */
-		if (interruptible) {
-			rc = wait_event_interruptible(
-				generic_waitq, !(device->stopped));
-			if (rc == -ERESTARTSYS) {
-				cqr->status = DASD_CQR_FAILED;
-				maincqr->intrc = rc;
-				continue;
-			}
-		} else
-			wait_event(generic_waitq, !(device->stopped));
-
+		/*
+		 * Don't try to start requests if device is stopped
+		 * except path verification requests
+		 */
+		if (!test_bit(DASD_CQR_VERIFY_PATH, &cqr->flags)) {
+			if (interruptible) {
+				rc = wait_event_interruptible(
+					generic_waitq, !(device->stopped));
+				if (rc == -ERESTARTSYS) {
+					cqr->status = DASD_CQR_FAILED;
+					maincqr->intrc = rc;
+					continue;
+				}
+			} else
+				wait_event(generic_waitq, !(device->stopped));
+		}
 		if (!cqr->callback)
 			cqr->callback = dasd_wakeup_cb;
 
@@ -2524,6 +2531,11 @@ static void __dasd_process_request_queue(struct dasd_block *block)
 			__blk_end_request_all(req, -EIO);
 		return;
 	}
+
+	/* if device ist stopped do not fetch new requests */
+	if (basedev->stopped)
+		return;
+
 	/* Now we try to fetch requests from the request queue */
 	while ((req = blk_peek_request(queue))) {
 		if (basedev->features & DASD_FEATURE_READONLY &&

drivers/s390/block/dasd_eckd.c

@@ -1628,7 +1628,8 @@ static void dasd_eckd_kick_validate_server(struct dasd_device *device)
 		return;
 	}
 	/* queue call to do_validate_server to the kernel event daemon. */
-	schedule_work(&device->kick_validate);
+	if (!schedule_work(&device->kick_validate))
+		dasd_put_device(device);
 }
 
 static u32 get_fcx_max_data(struct dasd_device *device)

drivers/s390/char/sclp_cmd.c

@@ -315,10 +315,29 @@ static int sclp_mem_change_state(unsigned long start, unsigned long size,
 			rc |= sclp_assign_storage(incr->rn);
 		else
 			sclp_unassign_storage(incr->rn);
+		if (rc == 0)
+			incr->standby = online ? 0 : 1;
 	}
 	return rc ? -EIO : 0;
 }
 
+static bool contains_standby_increment(unsigned long start, unsigned long end)
+{
+	struct memory_increment *incr;
+	unsigned long istart;
+
+	list_for_each_entry(incr, &sclp_mem_list, list) {
+		istart = rn2addr(incr->rn);
+		if (end - 1 < istart)
+			continue;
+		if (start > istart + sclp_rzm - 1)
+			continue;
+		if (incr->standby)
+			return true;
+	}
+	return false;
+}
+
 static int sclp_mem_notifier(struct notifier_block *nb,
 			     unsigned long action, void *data)
 {
@@ -334,8 +353,16 @@ static int sclp_mem_notifier(struct notifier_block *nb,
 	for_each_clear_bit(id, sclp_storage_ids, sclp_max_storage_id + 1)
 		sclp_attach_storage(id);
 	switch (action) {
-	case MEM_ONLINE:
 	case MEM_GOING_OFFLINE:
+		/*
+		 * We do not allow to set memory blocks offline that contain
+		 * standby memory. This is done to simplify the "memory online"
+		 * case.
+		 */
+		if (contains_standby_increment(start, start + size))
+			rc = -EPERM;
+		break;
+	case MEM_ONLINE:
 	case MEM_CANCEL_OFFLINE:
 		break;
 	case MEM_GOING_ONLINE:
@@ -361,6 +388,21 @@ static struct notifier_block sclp_mem_nb = {
 	.notifier_call = sclp_mem_notifier,
 };
 
+static void __init align_to_block_size(unsigned long long *start,
+				       unsigned long long *size)
+{
+	unsigned long long start_align, size_align, alignment;
+
+	alignment = memory_block_size_bytes();
+	start_align = roundup(*start, alignment);
+	size_align = rounddown(*start + *size, alignment) - start_align;
+
+	pr_info("Standby memory at 0x%llx (%lluM of %lluM usable)\n",
+		*start, size_align >> 20, *size >> 20);
+	*start = start_align;
+	*size = size_align;
+}
+
 static void __init add_memory_merged(u16 rn)
 {
 	static u16 first_rn, num;
@@ -382,7 +424,9 @@ static void __init add_memory_merged(u16 rn)
 		goto skip_add;
 	if (memory_end_set && (start + size > memory_end))
 		size = memory_end - start;
-	add_memory(0, start, size);
+	align_to_block_size(&start, &size);
+	if (size)
+		add_memory(0, start, size);
 skip_add:
 	first_rn = rn;
 	num = 1;