Merge branch 'linus' into perf/core

Conflicts:
	kernel/hw_breakpoint.c

Merge reason: resolve the conflict.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

arch/Kconfig

@@ -32,8 +32,9 @@ config HAVE_OPROFILE
 
 config KPROBES
 	bool "Kprobes"
-	depends on KALLSYMS && MODULES
+	depends on MODULES
 	depends on HAVE_KPROBES
+	select KALLSYMS
 	help
 	  Kprobes allows you to trap at almost any kernel address and
 	  execute a callback function.  register_kprobe() establishes
@@ -45,7 +46,6 @@ config OPTPROBES
 	def_bool y
 	depends on KPROBES && HAVE_OPTPROBES
 	depends on !PREEMPT
-	select KALLSYMS_ALL
 
 config HAVE_EFFICIENT_UNALIGNED_ACCESS
 	bool

arch/x86/include/asm/hw_breakpoint.h

@@ -20,7 +20,7 @@ struct arch_hw_breakpoint {
 #include <linux/list.h>
 
 /* Available HW breakpoint length encodings */
-#define X86_BREAKPOINT_LEN_X		0x00
+#define X86_BREAKPOINT_LEN_X		0x40
 #define X86_BREAKPOINT_LEN_1		0x40
 #define X86_BREAKPOINT_LEN_2		0x44
 #define X86_BREAKPOINT_LEN_4		0x4c

arch/x86/kernel/hw_breakpoint.c

@@ -206,11 +206,27 @@ int arch_check_bp_in_kernelspace(struct perf_event *bp)
 int arch_bp_generic_fields(int x86_len, int x86_type,
 			   int *gen_len, int *gen_type)
 {
-	/* Len */
-	switch (x86_len) {
-	case X86_BREAKPOINT_LEN_X:
+	/* Type */
+	switch (x86_type) {
+	case X86_BREAKPOINT_EXECUTE:
+		if (x86_len != X86_BREAKPOINT_LEN_X)
+			return -EINVAL;
+
+		*gen_type = HW_BREAKPOINT_X;
 		*gen_len = sizeof(long);
+		return 0;
+	case X86_BREAKPOINT_WRITE:
+		*gen_type = HW_BREAKPOINT_W;
 		break;
+	case X86_BREAKPOINT_RW:
+		*gen_type = HW_BREAKPOINT_W | HW_BREAKPOINT_R;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	/* Len */
+	switch (x86_len) {
 	case X86_BREAKPOINT_LEN_1:
 		*gen_len = HW_BREAKPOINT_LEN_1;
 		break;
@@ -229,21 +245,6 @@ int arch_bp_generic_fields(int x86_len, int x86_type,
 		return -EINVAL;
 	}
 
-	/* Type */
-	switch (x86_type) {
-	case X86_BREAKPOINT_EXECUTE:
-		*gen_type = HW_BREAKPOINT_X;
-		break;
-	case X86_BREAKPOINT_WRITE:
-		*gen_type = HW_BREAKPOINT_W;
-		break;
-	case X86_BREAKPOINT_RW:
-		*gen_type = HW_BREAKPOINT_W | HW_BREAKPOINT_R;
-		break;
-	default:
-		return -EINVAL;
-	}
-
 	return 0;
 }
 
@@ -316,9 +317,6 @@ int arch_validate_hwbkpt_settings(struct perf_event *bp)
 	ret = -EINVAL;
 
 	switch (info->len) {
-	case X86_BREAKPOINT_LEN_X:
-		align = sizeof(long) -1;
-		break;
 	case X86_BREAKPOINT_LEN_1:
 		align = 0;
 		break;

arch/x86/lguest/boot.c

@@ -324,9 +324,8 @@ static void lguest_load_gdt(const struct desc_ptr *desc)
 }
 
 /*
- * For a single GDT entry which changes, we do the lazy thing: alter our GDT,
- * then tell the Host to reload the entire thing.  This operation is so rare
- * that this naive implementation is reasonable.
+ * For a single GDT entry which changes, we simply change our copy and
+ * then tell the host about it.
  */
 static void lguest_write_gdt_entry(struct desc_struct *dt, int entrynum,
 				   const void *desc, int type)
@@ -338,9 +337,13 @@ static void lguest_write_gdt_entry(struct desc_struct *dt, int entrynum,
 }
 
 /*
- * OK, I lied.  There are three "thread local storage" GDT entries which change
+ * There are three "thread local storage" GDT entries which change
  * on every context switch (these three entries are how glibc implements
- * __thread variables).  So we have a hypercall specifically for this case.
+ * __thread variables).  As an optimization, we have a hypercall
+ * specifically for this case.
+ *
+ * Wouldn't it be nicer to have a general LOAD_GDT_ENTRIES hypercall
+ * which took a range of entries?
  */
 static void lguest_load_tls(struct thread_struct *t, unsigned int cpu)
 {

arch/x86/oprofile/nmi_int.c

@@ -671,7 +671,9 @@ static int __init ppro_init(char **cpu_type)
 	case 14:
 		*cpu_type = "i386/core";
 		break;
-	case 15: case 23:
+	case 0x0f:
+	case 0x16:
+	case 0x17:
 		*cpu_type = "i386/core_2";
 		break;
 	case 0x1a:

drivers/char/agp/intel-agp.c

@@ -806,6 +806,8 @@ static const struct intel_driver_description {
 	    "G45/G43", NULL, &intel_i965_driver },
 	{ PCI_DEVICE_ID_INTEL_B43_HB, PCI_DEVICE_ID_INTEL_B43_IG,
 	    "B43", NULL, &intel_i965_driver },
+	{ PCI_DEVICE_ID_INTEL_B43_1_HB, PCI_DEVICE_ID_INTEL_B43_1_IG,
+	    "B43", NULL, &intel_i965_driver },
 	{ PCI_DEVICE_ID_INTEL_G41_HB, PCI_DEVICE_ID_INTEL_G41_IG,
 	    "G41", NULL, &intel_i965_driver },
 	{ PCI_DEVICE_ID_INTEL_IRONLAKE_D_HB, PCI_DEVICE_ID_INTEL_IRONLAKE_D_IG,

drivers/char/agp/intel-agp.h

@@ -186,6 +186,8 @@
 #define PCI_DEVICE_ID_INTEL_Q33_IG          0x29D2
 #define PCI_DEVICE_ID_INTEL_B43_HB          0x2E40
 #define PCI_DEVICE_ID_INTEL_B43_IG          0x2E42
+#define PCI_DEVICE_ID_INTEL_B43_1_HB        0x2E90
+#define PCI_DEVICE_ID_INTEL_B43_1_IG        0x2E92
 #define PCI_DEVICE_ID_INTEL_GM45_HB         0x2A40
 #define PCI_DEVICE_ID_INTEL_GM45_IG         0x2A42
 #define PCI_DEVICE_ID_INTEL_EAGLELAKE_HB        0x2E00

drivers/char/virtio_console.c

@@ -596,6 +596,10 @@ static ssize_t port_fops_write(struct file *filp, const char __user *ubuf,
 	ssize_t ret;
 	bool nonblock;
 
+	/* Userspace could be out to fool us */
+	if (!count)
+		return 0;
+
 	port = filp->private_data;
 
 	nonblock = filp->f_flags & O_NONBLOCK;
@@ -642,7 +646,7 @@ static unsigned int port_fops_poll(struct file *filp, poll_table *wait)
 	poll_wait(filp, &port->waitqueue, wait);
 
 	ret = 0;
-	if (port->inbuf)
+	if (!will_read_block(port))
 		ret |= POLLIN | POLLRDNORM;
 	if (!will_write_block(port))
 		ret |= POLLOUT;

drivers/gpu/drm/i915/i915_drv.c

@@ -170,6 +170,7 @@ static const struct pci_device_id pciidlist[] = {		/* aka */
 	INTEL_VGA_DEVICE(0x2e22, &intel_g45_info),		/* G45_G */
 	INTEL_VGA_DEVICE(0x2e32, &intel_g45_info),		/* G41_G */
 	INTEL_VGA_DEVICE(0x2e42, &intel_g45_info),		/* B43_G */
+	INTEL_VGA_DEVICE(0x2e92, &intel_g45_info),		/* B43_G.1 */
 	INTEL_VGA_DEVICE(0xa001, &intel_pineview_info),
 	INTEL_VGA_DEVICE(0xa011, &intel_pineview_info),
 	INTEL_VGA_DEVICE(0x0042, &intel_ironlake_d_info),

drivers/gpu/drm/i915/i915_gem.c

@@ -2351,14 +2351,21 @@ i915_gem_object_get_fence_reg(struct drm_gem_object *obj)
 
 	reg->obj = obj;
 
-	if (IS_GEN6(dev))
+	switch (INTEL_INFO(dev)->gen) {
+	case 6:
 		sandybridge_write_fence_reg(reg);
-	else if (IS_I965G(dev))
+		break;
+	case 5:
+	case 4:
 		i965_write_fence_reg(reg);
-	else if (IS_I9XX(dev))
+		break;
+	case 3:
 		i915_write_fence_reg(reg);
-	else
+		break;
+	case 2:
 		i830_write_fence_reg(reg);
+		break;
+	}
 
 	trace_i915_gem_object_get_fence(obj, obj_priv->fence_reg,
 			obj_priv->tiling_mode);
@@ -2381,22 +2388,26 @@ i915_gem_clear_fence_reg(struct drm_gem_object *obj)
 	struct drm_i915_gem_object *obj_priv = to_intel_bo(obj);
 	struct drm_i915_fence_reg *reg =
 		&dev_priv->fence_regs[obj_priv->fence_reg];
+	uint32_t fence_reg;
 
-	if (IS_GEN6(dev)) {
+	switch (INTEL_INFO(dev)->gen) {
+	case 6:
 		I915_WRITE64(FENCE_REG_SANDYBRIDGE_0 +
 			     (obj_priv->fence_reg * 8), 0);
-	} else if (IS_I965G(dev)) {
+		break;
+	case 5:
+	case 4:
 		I915_WRITE64(FENCE_REG_965_0 + (obj_priv->fence_reg * 8), 0);
-	} else {
-		uint32_t fence_reg;
-
-		if (obj_priv->fence_reg < 8)
-			fence_reg = FENCE_REG_830_0 + obj_priv->fence_reg * 4;
+		break;
+	case 3:
+		if (obj_priv->fence_reg > 8)
+			fence_reg = FENCE_REG_945_8 + (obj_priv->fence_reg - 8) * 4;
 		else
-			fence_reg = FENCE_REG_945_8 + (obj_priv->fence_reg -
-						       8) * 4;
+	case 2:
+			fence_reg = FENCE_REG_830_0 + obj_priv->fence_reg * 4;
 
 		I915_WRITE(fence_reg, 0);
+		break;
 	}
 
 	reg->obj = NULL;

drivers/gpu/drm/i915/i915_gem_evict.c

@@ -79,6 +79,7 @@ mark_free(struct drm_i915_gem_object *obj_priv,
 	   struct list_head *unwind)
 {
 	list_add(&obj_priv->evict_list, unwind);
+	drm_gem_object_reference(&obj_priv->base);
 	return drm_mm_scan_add_block(obj_priv->gtt_space);
 }
 
@@ -165,6 +166,7 @@ i915_gem_evict_something(struct drm_device *dev, int min_size, unsigned alignmen
 	list_for_each_entry(obj_priv, &unwind_list, evict_list) {
 		ret = drm_mm_scan_remove_block(obj_priv->gtt_space);
 		BUG_ON(ret);
+		drm_gem_object_unreference(&obj_priv->base);
 	}
 
 	/* We expect the caller to unpin, evict all and try again, or give up.
@@ -181,18 +183,21 @@ found:
 			 * scanning, therefore store to be evicted objects on a
 			 * temporary list. */
 			list_move(&obj_priv->evict_list, &eviction_list);
-		}
+		} else
+			drm_gem_object_unreference(&obj_priv->base);
 	}
 
 	/* Unbinding will emit any required flushes */
 	list_for_each_entry_safe(obj_priv, tmp_obj_priv,
 				 &eviction_list, evict_list) {
 #if WATCH_LRU
-		DRM_INFO("%s: evicting %p\n", __func__, obj);
+		DRM_INFO("%s: evicting %p\n", __func__, &obj_priv->base);
 #endif
 		ret = i915_gem_object_unbind(&obj_priv->base);
 		if (ret)
 			return ret;
+
+		drm_gem_object_unreference(&obj_priv->base);
 	}
 
 	/* The just created free hole should be on the top of the free stack

drivers/gpu/drm/i915/i915_suspend.c

@@ -789,16 +789,25 @@ int i915_save_state(struct drm_device *dev)
 		dev_priv->saveSWF2[i] = I915_READ(SWF30 + (i << 2));
 
 	/* Fences */
-	if (IS_I965G(dev)) {
+	switch (INTEL_INFO(dev)->gen) {
+	case 6:
+		for (i = 0; i < 16; i++)
+			dev_priv->saveFENCE[i] = I915_READ64(FENCE_REG_SANDYBRIDGE_0 + (i * 8));
+		break;
+	case 5:
+	case 4:
 		for (i = 0; i < 16; i++)
 			dev_priv->saveFENCE[i] = I915_READ64(FENCE_REG_965_0 + (i * 8));
-	} else {
-		for (i = 0; i < 8; i++)
-			dev_priv->saveFENCE[i] = I915_READ(FENCE_REG_830_0 + (i * 4));
-
+		break;
+	case 3:
 		if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
 			for (i = 0; i < 8; i++)
 				dev_priv->saveFENCE[i+8] = I915_READ(FENCE_REG_945_8 + (i * 4));
+	case 2:
+		for (i = 0; i < 8; i++)
+			dev_priv->saveFENCE[i] = I915_READ(FENCE_REG_830_0 + (i * 4));
+		break;
+
 	}
 
 	return 0;
@@ -815,15 +824,24 @@ int i915_restore_state(struct drm_device *dev)
 	I915_WRITE(HWS_PGA, dev_priv->saveHWS);
 
 	/* Fences */
-	if (IS_I965G(dev)) {
+	switch (INTEL_INFO(dev)->gen) {
+	case 6:
+		for (i = 0; i < 16; i++)
+			I915_WRITE64(FENCE_REG_SANDYBRIDGE_0 + (i * 8), dev_priv->saveFENCE[i]);
+		break;
+	case 5:
+	case 4:
 		for (i = 0; i < 16; i++)
 			I915_WRITE64(FENCE_REG_965_0 + (i * 8), dev_priv->saveFENCE[i]);
-	} else {
-		for (i = 0; i < 8; i++)
-			I915_WRITE(FENCE_REG_830_0 + (i * 4), dev_priv->saveFENCE[i]);
+		break;
+	case 3:
+	case 2:
 		if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))
 			for (i = 0; i < 8; i++)
 				I915_WRITE(FENCE_REG_945_8 + (i * 4), dev_priv->saveFENCE[i+8]);
+		for (i = 0; i < 8; i++)
+			I915_WRITE(FENCE_REG_830_0 + (i * 4), dev_priv->saveFENCE[i]);
+		break;
 	}
 
 	i915_restore_display(dev);

drivers/gpu/drm/i915/intel_crt.c

@@ -188,7 +188,7 @@ static bool intel_ironlake_crt_detect_hotplug(struct drm_connector *connector)
 
 	if (wait_for((I915_READ(PCH_ADPA) & ADPA_CRT_HOTPLUG_FORCE_TRIGGER) == 0,
 		     1000, 1))
-		DRM_ERROR("timed out waiting for FORCE_TRIGGER");
+		DRM_DEBUG_KMS("timed out waiting for FORCE_TRIGGER");
 
 	if (turn_off_dac) {
 		I915_WRITE(PCH_ADPA, temp);
@@ -245,7 +245,7 @@ static bool intel_crt_detect_hotplug(struct drm_connector *connector)
 		if (wait_for((I915_READ(PORT_HOTPLUG_EN) &
 			      CRT_HOTPLUG_FORCE_DETECT) == 0,
 			     1000, 1))
-			DRM_ERROR("timed out waiting for FORCE_DETECT to go off");
+			DRM_DEBUG_KMS("timed out waiting for FORCE_DETECT to go off");
 	}
 
 	stat = I915_READ(PORT_HOTPLUG_STAT);

drivers/gpu/drm/i915/intel_display.c

@@ -2463,11 +2463,19 @@ static bool intel_crtc_mode_fixup(struct drm_crtc *crtc,
 				  struct drm_display_mode *adjusted_mode)
 {
 	struct drm_device *dev = crtc->dev;
+
 	if (HAS_PCH_SPLIT(dev)) {
 		/* FDI link clock is fixed at 2.7G */
 		if (mode->clock * 3 > IRONLAKE_FDI_FREQ * 4)
 			return false;
 	}
+
+	/* XXX some encoders set the crtcinfo, others don't.
+	 * Obviously we need some form of conflict resolution here...
+	 */
+	if (adjusted_mode->crtc_htotal == 0)
+		drm_mode_set_crtcinfo(adjusted_mode, 0);
+
 	return true;
 }
 

fs/ceph/Kconfig

@@ -3,6 +3,7 @@ config CEPH_FS
 	depends on INET && EXPERIMENTAL
 	select LIBCRC32C
 	select CRYPTO_AES
+	select CRYPTO
 	help
 	  Choose Y or M here to include support for mounting the
 	  experimental Ceph distributed file system.  Ceph is an extremely

fs/ceph/addr.c

@@ -411,8 +411,8 @@ static int writepage_nounlock(struct page *page, struct writeback_control *wbc)
 	if (i_size < page_off + len)
 		len = i_size - page_off;
 
-	dout("writepage %p page %p index %lu on %llu~%u\n",
-	     inode, page, page->index, page_off, len);
+	dout("writepage %p page %p index %lu on %llu~%u snapc %p\n",
+	     inode, page, page->index, page_off, len, snapc);
 
 	writeback_stat = atomic_long_inc_return(&client->writeback_count);
 	if (writeback_stat >
@@ -766,7 +766,8 @@ get_more_pages:
 			/* ok */
 			if (locked_pages == 0) {
 				/* prepare async write request */
-				offset = page->index << PAGE_CACHE_SHIFT;
+				offset = (unsigned long long)page->index
+					<< PAGE_CACHE_SHIFT;
 				len = wsize;
 				req = ceph_osdc_new_request(&client->osdc,
 					    &ci->i_layout,

fs/ceph/caps.c

@@ -814,7 +814,7 @@ int __ceph_caps_used(struct ceph_inode_info *ci)
 		used |= CEPH_CAP_PIN;
 	if (ci->i_rd_ref)
 		used |= CEPH_CAP_FILE_RD;
-	if (ci->i_rdcache_ref || ci->i_rdcache_gen)
+	if (ci->i_rdcache_ref || ci->vfs_inode.i_data.nrpages)
 		used |= CEPH_CAP_FILE_CACHE;
 	if (ci->i_wr_ref)
 		used |= CEPH_CAP_FILE_WR;
@@ -1195,10 +1195,14 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
  * asynchronously back to the MDS once sync writes complete and dirty
  * data is written out.
  *
+ * Unless @again is true, skip cap_snaps that were already sent to
+ * the MDS (i.e., during this session).
+ *
  * Called under i_lock.  Takes s_mutex as needed.
  */
 void __ceph_flush_snaps(struct ceph_inode_info *ci,
-			struct ceph_mds_session **psession)
+			struct ceph_mds_session **psession,
+			int again)
 		__releases(ci->vfs_inode->i_lock)
 		__acquires(ci->vfs_inode->i_lock)
 {
@@ -1227,7 +1231,7 @@ retry:
 		 * pages to be written out.
 		 */
 		if (capsnap->dirty_pages || capsnap->writing)
-			continue;
+			break;
 
 		/*
 		 * if cap writeback already occurred, we should have dropped
@@ -1240,6 +1244,13 @@ retry:
 			dout("no auth cap (migrating?), doing nothing\n");
 			goto out;
 		}
+
+		/* only flush each capsnap once */
+		if (!again && !list_empty(&capsnap->flushing_item)) {
+			dout("already flushed %p, skipping\n", capsnap);
+			continue;
+		}
+
 		mds = ci->i_auth_cap->session->s_mds;
 		mseq = ci->i_auth_cap->mseq;
 
@@ -1276,8 +1287,8 @@ retry:
 			      &session->s_cap_snaps_flushing);
 		spin_unlock(&inode->i_lock);
 
-		dout("flush_snaps %p cap_snap %p follows %lld size %llu\n",
-		     inode, capsnap, next_follows, capsnap->size);
+		dout("flush_snaps %p cap_snap %p follows %lld tid %llu\n",
+		     inode, capsnap, capsnap->follows, capsnap->flush_tid);
 		send_cap_msg(session, ceph_vino(inode).ino, 0,
 			     CEPH_CAP_OP_FLUSHSNAP, capsnap->issued, 0,
 			     capsnap->dirty, 0, capsnap->flush_tid, 0, mseq,
@@ -1314,7 +1325,7 @@ static void ceph_flush_snaps(struct ceph_inode_info *ci)
 	struct inode *inode = &ci->vfs_inode;
 
 	spin_lock(&inode->i_lock);
-	__ceph_flush_snaps(ci, NULL);
+	__ceph_flush_snaps(ci, NULL, 0);
 	spin_unlock(&inode->i_lock);
 }
 
@@ -1477,7 +1488,7 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
 
 	/* flush snaps first time around only */
 	if (!list_empty(&ci->i_cap_snaps))
-		__ceph_flush_snaps(ci, &session);
+		__ceph_flush_snaps(ci, &session, 0);
 	goto retry_locked;
 retry:
 	spin_lock(&inode->i_lock);
@@ -1894,7 +1905,7 @@ static void kick_flushing_capsnaps(struct ceph_mds_client *mdsc,
 		if (cap && cap->session == session) {
 			dout("kick_flushing_caps %p cap %p capsnap %p\n", inode,
 			     cap, capsnap);
-			__ceph_flush_snaps(ci, &session);
+			__ceph_flush_snaps(ci, &session, 1);
 		} else {
 			pr_err("%p auth cap %p not mds%d ???\n", inode,
 			       cap, session->s_mds);

fs/ceph/dir.c

@@ -1021,11 +1021,15 @@ out_touch:
 static void ceph_dentry_release(struct dentry *dentry)
 {
 	struct ceph_dentry_info *di = ceph_dentry(dentry);
-	struct inode *parent_inode = dentry->d_parent->d_inode;
-	u64 snapid = ceph_snap(parent_inode);
+	struct inode *parent_inode = NULL;
+	u64 snapid = CEPH_NOSNAP;
 
+	if (!IS_ROOT(dentry)) {
+		parent_inode = dentry->d_parent->d_inode;
+		if (parent_inode)
+			snapid = ceph_snap(parent_inode);
+	}
 	dout("dentry_release %p parent %p\n", dentry, parent_inode);
-
 	if (parent_inode && snapid != CEPH_SNAPDIR) {
 		struct ceph_inode_info *ci = ceph_inode(parent_inode);
 

fs/ceph/inode.c

@@ -845,7 +845,7 @@ static void ceph_set_dentry_offset(struct dentry *dn)
  * the caller) if we fail.
  */
 static struct dentry *splice_dentry(struct dentry *dn, struct inode *in,
-				    bool *prehash)
+				    bool *prehash, bool set_offset)
 {
 	struct dentry *realdn;
 
@@ -877,7 +877,8 @@ static struct dentry *splice_dentry(struct dentry *dn, struct inode *in,
 	}
 	if ((!prehash || *prehash) && d_unhashed(dn))
 		d_rehash(dn);
-	ceph_set_dentry_offset(dn);
+	if (set_offset)
+		ceph_set_dentry_offset(dn);
 out:
 	return dn;
 }
@@ -1062,7 +1063,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 				d_delete(dn);
 				goto done;
 			}
-			dn = splice_dentry(dn, in, &have_lease);
+			dn = splice_dentry(dn, in, &have_lease, true);
 			if (IS_ERR(dn)) {
 				err = PTR_ERR(dn);
 				goto done;
@@ -1105,7 +1106,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 			goto done;
 		}
 		dout(" linking snapped dir %p to dn %p\n", in, dn);
-		dn = splice_dentry(dn, in, NULL);
+		dn = splice_dentry(dn, in, NULL, true);
 		if (IS_ERR(dn)) {
 			err = PTR_ERR(dn);
 			goto done;
@@ -1237,7 +1238,7 @@ retry_lookup:
 				err = PTR_ERR(in);
 				goto out;
 			}
-			dn = splice_dentry(dn, in, NULL);
+			dn = splice_dentry(dn, in, NULL, false);
 			if (IS_ERR(dn))
 				dn = NULL;
 		}

fs/ceph/mds_client.c

@@ -2374,6 +2374,8 @@ static int encode_caps_cb(struct inode *inode, struct ceph_cap *cap,
 						num_fcntl_locks,
 						num_flock_locks);
 		unlock_kernel();
+	} else {
+		err = ceph_pagelist_append(pagelist, &rec, reclen);
 	}
 
 out_free:

fs/ceph/pagelist.c

@@ -5,10 +5,18 @@
 
 #include "pagelist.h"
 
+static void ceph_pagelist_unmap_tail(struct ceph_pagelist *pl)
+{
+	struct page *page = list_entry(pl->head.prev, struct page,
+				       lru);
+	kunmap(page);
+}
+
 int ceph_pagelist_release(struct ceph_pagelist *pl)
 {
 	if (pl->mapped_tail)
-		kunmap(pl->mapped_tail);
+		ceph_pagelist_unmap_tail(pl);
+
 	while (!list_empty(&pl->head)) {
 		struct page *page = list_first_entry(&pl->head, struct page,
 						     lru);
@@ -26,7 +34,7 @@ static int ceph_pagelist_addpage(struct ceph_pagelist *pl)
 	pl->room += PAGE_SIZE;
 	list_add_tail(&page->lru, &pl->head);
 	if (pl->mapped_tail)
-		kunmap(pl->mapped_tail);
+		ceph_pagelist_unmap_tail(pl);
 	pl->mapped_tail = kmap(page);
 	return 0;
 }

fs/ceph/snap.c

@@ -119,6 +119,7 @@ static struct ceph_snap_realm *ceph_create_snap_realm(
 	INIT_LIST_HEAD(&realm->children);
 	INIT_LIST_HEAD(&realm->child_item);
 	INIT_LIST_HEAD(&realm->empty_item);
+	INIT_LIST_HEAD(&realm->dirty_item);
 	INIT_LIST_HEAD(&realm->inodes_with_caps);
 	spin_lock_init(&realm->inodes_with_caps_lock);
 	__insert_snap_realm(&mdsc->snap_realms, realm);
@@ -467,7 +468,7 @@ void ceph_queue_cap_snap(struct ceph_inode_info *ci)
 		INIT_LIST_HEAD(&capsnap->ci_item);
 		INIT_LIST_HEAD(&capsnap->flushing_item);
 
-		capsnap->follows = snapc->seq - 1;
+		capsnap->follows = snapc->seq;
 		capsnap->issued = __ceph_caps_issued(ci, NULL);
 		capsnap->dirty = dirty;
 
@@ -604,6 +605,7 @@ int ceph_update_snap_trace(struct ceph_mds_client *mdsc,
 	struct ceph_snap_realm *realm;
 	int invalidate = 0;
 	int err = -ENOMEM;
+	LIST_HEAD(dirty_realms);
 
 	dout("update_snap_trace deletion=%d\n", deletion);
 more:
@@ -626,24 +628,6 @@ more:
 		}
 	}
 
-	if (le64_to_cpu(ri->seq) > realm->seq) {
-		dout("update_snap_trace updating %llx %p %lld -> %lld\n",
-		     realm->ino, realm, realm->seq, le64_to_cpu(ri->seq));
-		/*
-		 * if the realm seq has changed, queue a cap_snap for every
-		 * inode with open caps.  we do this _before_ we update
-		 * the realm info so that we prepare for writeback under the
-		 * _previous_ snap context.
-		 *
-		 * ...unless it's a snap deletion!
-		 */
-		if (!deletion)
-			queue_realm_cap_snaps(realm);
-	} else {
-		dout("update_snap_trace %llx %p seq %lld unchanged\n",
-		     realm->ino, realm, realm->seq);
-	}
-
 	/* ensure the parent is correct */
 	err = adjust_snap_realm_parent(mdsc, realm, le64_to_cpu(ri->parent));
 	if (err < 0)
@@ -651,6 +635,8 @@ more:
 	invalidate += err;
 
 	if (le64_to_cpu(ri->seq) > realm->seq) {
+		dout("update_snap_trace updating %llx %p %lld -> %lld\n",
+		     realm->ino, realm, realm->seq, le64_to_cpu(ri->seq));
 		/* update realm parameters, snap lists */
 		realm->seq = le64_to_cpu(ri->seq);
 		realm->created = le64_to_cpu(ri->created);
@@ -668,9 +654,17 @@ more:
 		if (err < 0)
 			goto fail;
 
+		/* queue realm for cap_snap creation */
+		list_add(&realm->dirty_item, &dirty_realms);
+
 		invalidate = 1;
 	} else if (!realm->cached_context) {
+		dout("update_snap_trace %llx %p seq %lld new\n",
+		     realm->ino, realm, realm->seq);
 		invalidate = 1;
+	} else {
+		dout("update_snap_trace %llx %p seq %lld unchanged\n",
+		     realm->ino, realm, realm->seq);
 	}
 
 	dout("done with %llx %p, invalidated=%d, %p %p\n", realm->ino,
@@ -683,6 +677,14 @@ more:
 	if (invalidate)
 		rebuild_snap_realms(realm);
 
+	/*
+	 * queue cap snaps _after_ we've built the new snap contexts,
+	 * so that i_head_snapc can be set appropriately.
+	 */
+	list_for_each_entry(realm, &dirty_realms, dirty_item) {
+		queue_realm_cap_snaps(realm);
+	}
+
 	__cleanup_empty_realms(mdsc);
 	return 0;
 
@@ -715,7 +717,7 @@ static void flush_snaps(struct ceph_mds_client *mdsc)
 		igrab(inode);
 		spin_unlock(&mdsc->snap_flush_lock);
 		spin_lock(&inode->i_lock);
-		__ceph_flush_snaps(ci, &session);
+		__ceph_flush_snaps(ci, &session, 0);
 		spin_unlock(&inode->i_lock);
 		iput(inode);
 		spin_lock(&mdsc->snap_flush_lock);
@@ -816,6 +818,7 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc,
 			};
 			struct inode *inode = ceph_find_inode(sb, vino);
 			struct ceph_inode_info *ci;
+			struct ceph_snap_realm *oldrealm;
 
 			if (!inode)
 				continue;
@@ -841,18 +844,19 @@ void ceph_handle_snap(struct ceph_mds_client *mdsc,
 			dout(" will move %p to split realm %llx %p\n",
 			     inode, realm->ino, realm);
 			/*
-			 * Remove the inode from the realm's inode
-			 * list, but don't add it to the new realm
-			 * yet.  We don't want the cap_snap to be
-			 * queued (again) by ceph_update_snap_trace()
-			 * below.  Queue it _now_, under the old context.
+			 * Move the inode to the new realm
 			 */
 			spin_lock(&realm->inodes_with_caps_lock);
 			list_del_init(&ci->i_snap_realm_item);
+			list_add(&ci->i_snap_realm_item,
+				 &realm->inodes_with_caps);
+			oldrealm = ci->i_snap_realm;
+			ci->i_snap_realm = realm;
 			spin_unlock(&realm->inodes_with_caps_lock);
 			spin_unlock(&inode->i_lock);
 
-			ceph_queue_cap_snap(ci);
+			ceph_get_snap_realm(mdsc, realm);
+			ceph_put_snap_realm(mdsc, oldrealm);
 
 			iput(inode);
 			continue;
@@ -880,43 +884,9 @@ skip_inode:
 	ceph_update_snap_trace(mdsc, p, e,
 			       op == CEPH_SNAP_OP_DESTROY);
 
-	if (op == CEPH_SNAP_OP_SPLIT) {
-		/*
-		 * ok, _now_ add the inodes into the new realm.
-		 */
-		for (i = 0; i < num_split_inos; i++) {
-			struct ceph_vino vino = {
-				.ino = le64_to_cpu(split_inos[i]),
-				.snap = CEPH_NOSNAP,
-			};
-			struct inode *inode = ceph_find_inode(sb, vino);
-			struct ceph_inode_info *ci;
-
-			if (!inode)
-				continue;
-			ci = ceph_inode(inode);
-			spin_lock(&inode->i_lock);
-			if (list_empty(&ci->i_snap_realm_item)) {
-				struct ceph_snap_realm *oldrealm =
-					ci->i_snap_realm;
-
-				dout(" moving %p to split realm %llx %p\n",
-				     inode, realm->ino, realm);
-				spin_lock(&realm->inodes_with_caps_lock);
-				list_add(&ci->i_snap_realm_item,
-					 &realm->inodes_with_caps);
-				ci->i_snap_realm = realm;
-				spin_unlock(&realm->inodes_with_caps_lock);
-				ceph_get_snap_realm(mdsc, realm);
-				ceph_put_snap_realm(mdsc, oldrealm);
-			}
-			spin_unlock(&inode->i_lock);
-			iput(inode);
-		}
-
+	if (op == CEPH_SNAP_OP_SPLIT)
 		/* we took a reference when we created the realm, above */
 		ceph_put_snap_realm(mdsc, realm);
-	}
 
 	__cleanup_empty_realms(mdsc);
 

fs/ceph/super.h

@@ -690,6 +690,8 @@ struct ceph_snap_realm {
 
 	struct list_head empty_item;     /* if i have ref==0 */
 
+	struct list_head dirty_item;     /* if realm needs new context */
+
 	/* the current set of snaps for this realm */
 	struct ceph_snap_context *cached_context;
 
@@ -826,7 +828,8 @@ extern void ceph_put_cap_refs(struct ceph_inode_info *ci, int had);
 extern void ceph_put_wrbuffer_cap_refs(struct ceph_inode_info *ci, int nr,
 				       struct ceph_snap_context *snapc);
 extern void __ceph_flush_snaps(struct ceph_inode_info *ci,
-			       struct ceph_mds_session **psession);
+			       struct ceph_mds_session **psession,
+			       int again);
 extern void ceph_check_caps(struct ceph_inode_info *ci, int flags,
 			    struct ceph_mds_session *session);
 extern void ceph_check_delayed_caps(struct ceph_mds_client *mdsc);

include/linux/fs.h

@@ -1093,6 +1093,10 @@ struct file_lock {
 
 #include <linux/fcntl.h>
 
+/* temporary stubs for BKL removal */
+#define lock_flocks() lock_kernel()
+#define unlock_flocks() unlock_kernel()
+
 extern void send_sigio(struct fown_struct *fown, int fd, int band);
 
 #ifdef CONFIG_FILE_LOCKING

kernel/sched.c

@@ -3513,9 +3513,9 @@ void task_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
 	rtime = nsecs_to_cputime(p->se.sum_exec_runtime);
 
 	if (total) {
-		u64 temp;
+		u64 temp = rtime;
 
-		temp = (u64)(rtime * utime);
+		temp *= utime;
 		do_div(temp, total);
 		utime = (cputime_t)temp;
 	} else
@@ -3546,9 +3546,9 @@ void thread_group_times(struct task_struct *p, cputime_t *ut, cputime_t *st)
 	rtime = nsecs_to_cputime(cputime.sum_exec_runtime);
 
 	if (total) {
-		u64 temp;
+		u64 temp = rtime;
 
-		temp = (u64)(rtime * cputime.utime);
+		temp *= cputime.utime;
 		do_div(temp, total);
 		utime = (cputime_t)temp;
 	} else

kernel/sched_fair.c

@@ -3630,7 +3630,7 @@ static inline int nohz_kick_needed(struct rq *rq, int cpu)
 	if (time_before(now, nohz.next_balance))
 		return 0;
 
-	if (!rq->nr_running)
+	if (rq->idle_at_tick)
 		return 0;
 
 	first_pick_cpu = atomic_read(&nohz.first_pick_cpu);