Linux-libre 4.8.2-gnupck-4.8.2-gnu

4 files changed, 118 insertions, 56 deletions

diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c
index 17a42e4eb..217847679 100644
--- a/fs/nfs/blocklayout/blocklayout.c
+++ b/fs/nfs/blocklayout/blocklayout.c
@@ -102,14 +102,15 @@ static inline void put_parallel(struct parallel_io *p)
 }
 
 static struct bio *
-bl_submit_bio(int rw, struct bio *bio)
+bl_submit_bio(struct bio *bio)
 {
 	if (bio) {
 		get_parallel(bio->bi_private);
 		dprintk("%s submitting %s bio %u@%llu\n", __func__,
-			rw == READ ? "read" : "write", bio->bi_iter.bi_size,
+			bio_op(bio) == READ ? "read" : "write",
+			bio->bi_iter.bi_size,
 			(unsigned long long)bio->bi_iter.bi_sector);
-		submit_bio(rw, bio);
+		submit_bio(bio);
 	}
 	return NULL;
 }
@@ -158,7 +159,7 @@ do_add_page_to_bio(struct bio *bio, int npg, int rw, sector_t isect,
 	if (disk_addr < map->start || disk_addr >= map->start + map->len) {
 		if (!dev->map(dev, disk_addr, map))
 			return ERR_PTR(-EIO);
-		bio = bl_submit_bio(rw, bio);
+		bio = bl_submit_bio(bio);
 	}
 	disk_addr += map->disk_offset;
 	disk_addr -= map->start;
@@ -174,9 +175,10 @@ retry:
 				disk_addr >> SECTOR_SHIFT, end_io, par);
 		if (!bio)
 			return ERR_PTR(-ENOMEM);
+		bio_set_op_attrs(bio, rw, 0);
 	}
 	if (bio_add_page(bio, page, *len, offset) < *len) {
-		bio = bl_submit_bio(rw, bio);
+		bio = bl_submit_bio(bio);
 		goto retry;
 	}
 	return bio;
@@ -252,7 +254,7 @@ bl_read_pagelist(struct nfs_pgio_header *header)
 	for (i = pg_index; i < header->page_array.npages; i++) {
 		if (extent_length <= 0) {
 			/* We've used up the previous extent */
-			bio = bl_submit_bio(READ, bio);
+			bio = bl_submit_bio(bio);
 
 			/* Get the next one */
 			if (!ext_tree_lookup(bl, isect, &be, false)) {
@@ -273,7 +275,7 @@ bl_read_pagelist(struct nfs_pgio_header *header)
 		}
 
 		if (is_hole(&be)) {
-			bio = bl_submit_bio(READ, bio);
+			bio = bl_submit_bio(bio);
 			/* Fill hole w/ zeroes w/o accessing device */
 			dprintk("%s Zeroing page for hole\n", __func__);
 			zero_user_segment(pages[i], pg_offset, pg_len);
@@ -306,7 +308,7 @@ bl_read_pagelist(struct nfs_pgio_header *header)
 		header->res.count = (isect << SECTOR_SHIFT) - header->args.offset;
 	}
 out:
-	bl_submit_bio(READ, bio);
+	bl_submit_bio(bio);
 	blk_finish_plug(&plug);
 	put_parallel(par);
 	return PNFS_ATTEMPTED;
@@ -344,7 +346,7 @@ static void bl_write_cleanup(struct work_struct *work)
 			PAGE_SIZE - 1) & (loff_t)PAGE_MASK;
 
 		ext_tree_mark_written(bl, start >> SECTOR_SHIFT,
-					(end - start) >> SECTOR_SHIFT);
+					(end - start) >> SECTOR_SHIFT, end);
 	}
 
 	pnfs_ld_write_done(hdr);
@@ -398,7 +400,7 @@ bl_write_pagelist(struct nfs_pgio_header *header, int sync)
 	for (i = pg_index; i < header->page_array.npages; i++) {
 		if (extent_length <= 0) {
 			/* We've used up the previous extent */
-			bio = bl_submit_bio(WRITE, bio);
+			bio = bl_submit_bio(bio);
 			/* Get the next one */
 			if (!ext_tree_lookup(bl, isect, &be, true)) {
 				header->pnfs_error = -EINVAL;
@@ -427,7 +429,7 @@ bl_write_pagelist(struct nfs_pgio_header *header, int sync)
 
 	header->res.count = header->args.count;
 out:
-	bl_submit_bio(WRITE, bio);
+	bl_submit_bio(bio);
 	blk_finish_plug(&plug);
 	put_parallel(par);
 	return PNFS_ATTEMPTED;
diff --git a/fs/nfs/blocklayout/blocklayout.h b/fs/nfs/blocklayout/blocklayout.h
index 18e6fd0b9..efc007f00 100644
--- a/fs/nfs/blocklayout/blocklayout.h
+++ b/fs/nfs/blocklayout/blocklayout.h
@@ -141,6 +141,7 @@ struct pnfs_block_layout {
 	struct rb_root		bl_ext_ro;
 	spinlock_t		bl_ext_lock;   /* Protects list manipulation */
 	bool			bl_scsi_layout;
+	u64			bl_lwb;
 };
 
 static inline struct pnfs_block_layout *
@@ -182,7 +183,7 @@ int ext_tree_insert(struct pnfs_block_layout *bl,
 int ext_tree_remove(struct pnfs_block_layout *bl, bool rw, sector_t start,
 		sector_t end);
 int ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start,
-		sector_t len);
+		sector_t len, u64 lwb);
 bool ext_tree_lookup(struct pnfs_block_layout *bl, sector_t isect,
 		struct pnfs_block_extent *ret, bool rw);
 int ext_tree_prepare_commit(struct nfs4_layoutcommit_args *arg);
diff --git a/fs/nfs/blocklayout/dev.c b/fs/nfs/blocklayout/dev.c
index e5b896752..a69ef4e9c 100644
--- a/fs/nfs/blocklayout/dev.c
+++ b/fs/nfs/blocklayout/dev.c
@@ -65,8 +65,8 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b)
 		if (!p)
 			return -EIO;
 		b->simple.nr_sigs = be32_to_cpup(p++);
-		if (!b->simple.nr_sigs) {
-			dprintk("no signature\n");
+		if (!b->simple.nr_sigs || b->simple.nr_sigs > PNFS_BLOCK_MAX_UUIDS) {
+			dprintk("Bad signature count: %d\n", b->simple.nr_sigs);
 			return -EIO;
 		}
 
@@ -89,7 +89,8 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b)
 			memcpy(&b->simple.sigs[i].sig, p,
 				b->simple.sigs[i].sig_len);
 
-			b->simple.len += 8 + 4 + b->simple.sigs[i].sig_len;
+			b->simple.len += 8 + 4 + \
+				(XDR_QUADLEN(b->simple.sigs[i].sig_len) << 2);
 		}
 		break;
 	case PNFS_BLOCK_VOLUME_SLICE:
@@ -104,7 +105,12 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b)
 		p = xdr_inline_decode(xdr, 4);
 		if (!p)
 			return -EIO;
+
 		b->concat.volumes_count = be32_to_cpup(p++);
+		if (b->concat.volumes_count > PNFS_BLOCK_MAX_DEVICES) {
+			dprintk("Too many volumes: %d\n", b->concat.volumes_count);
+			return -EIO;
+		}
 
 		p = xdr_inline_decode(xdr, b->concat.volumes_count * 4);
 		if (!p)
@@ -116,8 +122,13 @@ nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b)
 		p = xdr_inline_decode(xdr, 8 + 4);
 		if (!p)
 			return -EIO;
+
 		p = xdr_decode_hyper(p, &b->stripe.chunk_size);
 		b->stripe.volumes_count = be32_to_cpup(p++);
+		if (b->stripe.volumes_count > PNFS_BLOCK_MAX_DEVICES) {
+			dprintk("Too many volumes: %d\n", b->stripe.volumes_count);
+			return -EIO;
+		}
 
 		p = xdr_inline_decode(xdr, b->stripe.volumes_count * 4);
 		if (!p)
@@ -224,18 +235,20 @@ bl_parse_simple(struct nfs_server *server, struct pnfs_block_dev *d,
 		struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask)
 {
 	struct pnfs_block_volume *v = &volumes[idx];
+	struct block_device *bdev;
 	dev_t dev;
 
 	dev = bl_resolve_deviceid(server, v, gfp_mask);
 	if (!dev)
 		return -EIO;
 
-	d->bdev = blkdev_get_by_dev(dev, FMODE_READ | FMODE_WRITE, NULL);
-	if (IS_ERR(d->bdev)) {
+	bdev = blkdev_get_by_dev(dev, FMODE_READ | FMODE_WRITE, NULL);
+	if (IS_ERR(bdev)) {
 		printk(KERN_WARNING "pNFS: failed to open device %d:%d (%ld)\n",
-			MAJOR(dev), MINOR(dev), PTR_ERR(d->bdev));
-		return PTR_ERR(d->bdev);
+			MAJOR(dev), MINOR(dev), PTR_ERR(bdev));
+		return PTR_ERR(bdev);
 	}
+	d->bdev = bdev;
 
 
 	d->len = i_size_read(d->bdev->bd_inode);
@@ -287,44 +300,71 @@ bl_validate_designator(struct pnfs_block_volume *v)
 	}
 }
 
+/*
+ * Try to open the udev path for the WWN.  At least on Debian the udev
+ * by-id path will always point to the dm-multipath device if one exists.
+ */
+static struct block_device *
+bl_open_udev_path(struct pnfs_block_volume *v)
+{
+	struct block_device *bdev;
+	const char *devname;
+
+	devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%*phN",
+				v->scsi.designator_len, v->scsi.designator);
+	if (!devname)
+		return ERR_PTR(-ENOMEM);
+
+	bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL);
+	if (IS_ERR(bdev)) {
+		pr_warn("pNFS: failed to open device %s (%ld)\n",
+			devname, PTR_ERR(bdev));
+	}
+
+	kfree(devname);
+	return bdev;
+}
+
+/*
+ * Try to open the RH/Fedora specific dm-mpath udev path for this WWN, as the
+ * wwn- links will only point to the first discovered SCSI device there.
+ */
+static struct block_device *
+bl_open_dm_mpath_udev_path(struct pnfs_block_volume *v)
+{
+	struct block_device *bdev;
+	const char *devname;
+
+	devname = kasprintf(GFP_KERNEL,
+			"/dev/disk/by-id/dm-uuid-mpath-%d%*phN",
+			v->scsi.designator_type,
+			v->scsi.designator_len, v->scsi.designator);
+	if (!devname)
+		return ERR_PTR(-ENOMEM);
+
+	bdev = blkdev_get_by_path(devname, FMODE_READ | FMODE_WRITE, NULL);
+	kfree(devname);
+	return bdev;
+}
+
 static int
 bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 		struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask)
 {
 	struct pnfs_block_volume *v = &volumes[idx];
+	struct block_device *bdev;
 	const struct pr_ops *ops;
-	const char *devname;
 	int error;
 
 	if (!bl_validate_designator(v))
 		return -EINVAL;
 
-	switch (v->scsi.designator_len) {
-	case 8:
-		devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%8phN",
-				v->scsi.designator);
-		break;
-	case 12:
-		devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%12phN",
-				v->scsi.designator);
-		break;
-	case 16:
-		devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/wwn-0x%16phN",
-				v->scsi.designator);
-		break;
-	default:
-		return -EINVAL;
-	}
-
-	d->bdev = blkdev_get_by_path(devname, FMODE_READ, NULL);
-	if (IS_ERR(d->bdev)) {
-		pr_warn("pNFS: failed to open device %s (%ld)\n",
-			devname, PTR_ERR(d->bdev));
-		kfree(devname);
-		return PTR_ERR(d->bdev);
-	}
-
-	kfree(devname);
+	bdev = bl_open_dm_mpath_udev_path(v);
+	if (IS_ERR(bdev))
+		bdev = bl_open_udev_path(v);
+	if (IS_ERR(bdev))
+		return PTR_ERR(bdev);
+	d->bdev = bdev;
 
 	d->len = i_size_read(d->bdev->bd_inode);
 	d->map = bl_map_simple;
@@ -352,7 +392,7 @@ bl_parse_scsi(struct nfs_server *server, struct pnfs_block_dev *d,
 	return 0;
 
 out_blkdev_put:
-	blkdev_put(d->bdev, FMODE_READ);
+	blkdev_put(d->bdev, FMODE_READ | FMODE_WRITE);
 	return error;
 }
 
diff --git a/fs/nfs/blocklayout/extent_tree.c b/fs/nfs/blocklayout/extent_tree.c
index 720b3ff55..c85fbfd2d 100644
--- a/fs/nfs/blocklayout/extent_tree.c
+++ b/fs/nfs/blocklayout/extent_tree.c
@@ -121,6 +121,16 @@ ext_try_to_merge_right(struct rb_root *root, struct pnfs_block_extent *be)
 	return be;
 }
 
+static void __ext_put_deviceids(struct list_head *head)
+{
+	struct pnfs_block_extent *be, *tmp;
+
+	list_for_each_entry_safe(be, tmp, head, be_list) {
+		nfs4_put_deviceid_node(be->be_device);
+		kfree(be);
+	}
+}
+
 static void
 __ext_tree_insert(struct rb_root *root,
 		struct pnfs_block_extent *new, bool merge_ok)
@@ -163,7 +173,8 @@ free_new:
 }
 
 static int
-__ext_tree_remove(struct rb_root *root, sector_t start, sector_t end)
+__ext_tree_remove(struct rb_root *root,
+		sector_t start, sector_t end, struct list_head *tmp)
 {
 	struct pnfs_block_extent *be;
 	sector_t len1 = 0, len2 = 0;
@@ -223,8 +234,7 @@ __ext_tree_remove(struct rb_root *root, sector_t start, sector_t end)
 			struct pnfs_block_extent *next = ext_tree_next(be);
 
 			rb_erase(&be->be_node, root);
-			nfs4_put_deviceid_node(be->be_device);
-			kfree(be);
+			list_add_tail(&be->be_list, tmp);
 			be = next;
 		}
 
@@ -350,16 +360,18 @@ int ext_tree_remove(struct pnfs_block_layout *bl, bool rw,
 		sector_t start, sector_t end)
 {
 	int err, err2;
+	LIST_HEAD(tmp);
 
 	spin_lock(&bl->bl_ext_lock);
-	err = __ext_tree_remove(&bl->bl_ext_ro, start, end);
+	err = __ext_tree_remove(&bl->bl_ext_ro, start, end, &tmp);
 	if (rw) {
-		err2 = __ext_tree_remove(&bl->bl_ext_rw, start, end);
+		err2 = __ext_tree_remove(&bl->bl_ext_rw, start, end, &tmp);
 		if (!err)
 			err = err2;
 	}
 	spin_unlock(&bl->bl_ext_lock);
 
+	__ext_put_deviceids(&tmp);
 	return err;
 }
 
@@ -390,18 +402,19 @@ ext_tree_split(struct rb_root *root, struct pnfs_block_extent *be,
 
 int
 ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start,
-		sector_t len)
+		sector_t len, u64 lwb)
 {
 	struct rb_root *root = &bl->bl_ext_rw;
 	sector_t end = start + len;
 	struct pnfs_block_extent *be;
 	int err = 0;
+	LIST_HEAD(tmp);
 
 	spin_lock(&bl->bl_ext_lock);
 	/*
 	 * First remove all COW extents or holes from written to range.
 	 */
-	err = __ext_tree_remove(&bl->bl_ext_ro, start, end);
+	err = __ext_tree_remove(&bl->bl_ext_ro, start, end, &tmp);
 	if (err)
 		goto out;
 
@@ -458,7 +471,11 @@ ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start,
 		}
 	}
 out:
+	if (bl->bl_lwb < lwb)
+		bl->bl_lwb = lwb;
 	spin_unlock(&bl->bl_ext_lock);
+
+	__ext_put_deviceids(&tmp);
 	return err;
 }
 
@@ -503,7 +520,7 @@ static __be32 *encode_scsi_range(struct pnfs_block_extent *be, __be32 *p)
 }
 
 static int ext_tree_encode_commit(struct pnfs_block_layout *bl, __be32 *p,
-		size_t buffer_size, size_t *count)
+		size_t buffer_size, size_t *count, __u64 *lastbyte)
 {
 	struct pnfs_block_extent *be;
 	int ret = 0;
@@ -527,6 +544,8 @@ static int ext_tree_encode_commit(struct pnfs_block_layout *bl, __be32 *p,
 			p = encode_block_extent(be, p);
 		be->be_tag = EXTENT_COMMITTING;
 	}
+	*lastbyte = bl->bl_lwb - 1;
+	bl->bl_lwb = 0;
 	spin_unlock(&bl->bl_ext_lock);
 
 	return ret;
@@ -549,7 +568,7 @@ ext_tree_prepare_commit(struct nfs4_layoutcommit_args *arg)
 	arg->layoutupdate_pages = &arg->layoutupdate_page;
 
 retry:
-	ret = ext_tree_encode_commit(bl, start_p + 1, buffer_size, &count);
+	ret = ext_tree_encode_commit(bl, start_p + 1, buffer_size, &count, &arg->lastbytewritten);
 	if (unlikely(ret)) {
 		ext_tree_free_commitdata(arg, buffer_size);