Resguardo inicial

linux-7.2.3/fs/nfs/blocklayout/Makefile

@@ -0,0 +1,7 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Makefile for the pNFS block layout driver kernel module
+#
+obj-$(CONFIG_PNFS_BLOCK) += blocklayoutdriver.o
+
+blocklayoutdriver-y += blocklayout.o dev.o extent_tree.o rpc_pipefs.o

linux-7.2.3/fs/nfs/blocklayout/blocklayout.h

@@ -0,0 +1,204 @@
+/*
+ *  linux/fs/nfs/blocklayout/blocklayout.h
+ *
+ *  Module for the NFSv4.1 pNFS block layout driver.
+ *
+ *  Copyright (c) 2006 The Regents of the University of Michigan.
+ *  All rights reserved.
+ *
+ *  Andy Adamson <andros@citi.umich.edu>
+ *  Fred Isaman <iisaman@umich.edu>
+ *
+ * permission is granted to use, copy, create derivative works and
+ * redistribute this software and such derivative works for any purpose,
+ * so long as the name of the university of michigan is not used in
+ * any advertising or publicity pertaining to the use or distribution
+ * of this software without specific, written prior authorization.  if
+ * the above copyright notice or any other identification of the
+ * university of michigan is included in any copy of any portion of
+ * this software, then the disclaimer below must also be included.
+ *
+ * this software is provided as is, without representation from the
+ * university of michigan as to its fitness for any purpose, and without
+ * warranty by the university of michigan of any kind, either express
+ * or implied, including without limitation the implied warranties of
+ * merchantability and fitness for a particular purpose.  the regents
+ * of the university of michigan shall not be liable for any damages,
+ * including special, indirect, incidental, or consequential damages,
+ * with respect to any claim arising out or in connection with the use
+ * of the software, even if it has been or is hereafter advised of the
+ * possibility of such damages.
+ */
+#ifndef FS_NFS_NFS4BLOCKLAYOUT_H
+#define FS_NFS_NFS4BLOCKLAYOUT_H
+
+#include <linux/device-mapper.h>
+#include <linux/nfs_fs.h>
+#include <linux/sunrpc/rpc_pipe_fs.h>
+
+#include "../nfs4_fs.h"
+#include "../pnfs.h"
+#include "../netns.h"
+
+#define PAGE_CACHE_SECTORS (PAGE_SIZE >> SECTOR_SHIFT)
+#define PAGE_CACHE_SECTOR_SHIFT (PAGE_SHIFT - SECTOR_SHIFT)
+#define SECTOR_SIZE (1 << SECTOR_SHIFT)
+
+struct pnfs_block_dev;
+
+#define PNFS_BLOCK_MAX_UUIDS	4
+#define PNFS_BLOCK_MAX_DEVICES	64
+
+/*
+ * Random upper cap for the uuid length to avoid unbounded allocation.
+ * Not actually limited by the protocol.
+ */
+#define PNFS_BLOCK_UUID_LEN	128
+
+struct pnfs_block_volume {
+	enum pnfs_block_volume_type	type;
+	union {
+		struct {
+			int		len;
+			int		nr_sigs;
+			struct {
+				u64		offset;
+				u32		sig_len;
+				u8		sig[PNFS_BLOCK_UUID_LEN];
+			} sigs[PNFS_BLOCK_MAX_UUIDS];
+		} simple;
+		struct {
+			u64		start;
+			u64		len;
+			u32		volume;
+		} slice;
+		struct {
+			u32		volumes_count;
+			u32		volumes[PNFS_BLOCK_MAX_DEVICES];
+		} concat;
+		struct {
+			u64		chunk_size;
+			u32		volumes_count;
+			u32		volumes[PNFS_BLOCK_MAX_DEVICES];
+		} stripe;
+		struct {
+			enum scsi_code_set		code_set;
+			enum scsi_designator_type	designator_type;
+			int				designator_len;
+			u8				designator[256];
+			u64				pr_key;
+		} scsi;
+	};
+};
+
+struct pnfs_block_dev_map {
+	u64			start;
+	u64			len;
+	u64			disk_offset;
+	struct block_device		*bdev;
+};
+
+struct pnfs_block_dev {
+	struct nfs4_deviceid_node	node;
+
+	u64				start;
+	u64				len;
+
+	enum pnfs_block_volume_type	type;
+	u32				nr_children;
+	struct pnfs_block_dev		*children;
+	u64				chunk_size;
+
+	struct file			*bdev_file;
+	u64				disk_offset;
+	unsigned long			flags;
+
+	u64				pr_key;
+
+	bool (*map)(struct pnfs_block_dev *dev, u64 offset,
+			struct pnfs_block_dev_map *map);
+};
+
+/* pnfs_block_dev flag bits */
+enum {
+	PNFS_BDEV_REGISTERED = 0,
+};
+
+/* sector_t fields are all in 512-byte sectors */
+struct pnfs_block_extent {
+	union {
+		struct rb_node	be_node;
+		struct list_head be_list;
+	};
+	struct nfs4_deviceid_node *be_device;
+	sector_t	be_f_offset;	/* the starting offset in the file */
+	sector_t	be_length;	/* the size of the extent */
+	sector_t	be_v_offset;	/* the starting offset in the volume */
+	enum pnfs_block_extent_state be_state;	/* the state of this extent */
+#define EXTENT_WRITTEN		1
+#define EXTENT_COMMITTING	2
+	unsigned int	be_tag;
+};
+
+struct pnfs_block_layout {
+	struct pnfs_layout_hdr	bl_layout;
+	struct rb_root		bl_ext_rw;
+	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 *
+BLK_LO2EXT(struct pnfs_layout_hdr *lo)
+{
+	return container_of(lo, struct pnfs_block_layout, bl_layout);
+}
+
+static inline struct pnfs_block_layout *
+BLK_LSEG2EXT(struct pnfs_layout_segment *lseg)
+{
+	return BLK_LO2EXT(lseg->pls_layout);
+}
+
+struct bl_pipe_msg {
+	struct rpc_pipe_msg msg;
+	wait_queue_head_t *bl_wq;
+};
+
+struct bl_msg_hdr {
+	u8  type;
+	u16 totallen; /* length of entire message, including hdr itself */
+};
+
+#define BL_DEVICE_UMOUNT               0x0 /* Umount--delete devices */
+#define BL_DEVICE_MOUNT                0x1 /* Mount--create devices*/
+#define BL_DEVICE_REQUEST_INIT         0x0 /* Start request */
+#define BL_DEVICE_REQUEST_PROC         0x1 /* User level process succeeds */
+#define BL_DEVICE_REQUEST_ERR          0x2 /* User level process fails */
+
+/* dev.c */
+bool bl_register_dev(struct pnfs_block_dev *d);
+struct nfs4_deviceid_node *bl_alloc_deviceid_node(struct nfs_server *server,
+		struct pnfs_device *pdev, gfp_t gfp_mask);
+void bl_free_deviceid_node(struct nfs4_deviceid_node *d);
+
+/* extent_tree.c */
+int ext_tree_insert(struct pnfs_block_layout *bl,
+		struct pnfs_block_extent *new);
+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, 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);
+void ext_tree_mark_committed(struct nfs4_layoutcommit_args *arg, int status);
+
+/* rpc_pipefs.c */
+dev_t bl_resolve_deviceid(struct nfs_server *server,
+		struct pnfs_block_volume *b, gfp_t gfp_mask);
+int __init bl_init_pipefs(void);
+void bl_cleanup_pipefs(void);
+
+#endif /* FS_NFS_NFS4BLOCKLAYOUT_H */

linux-7.2.3/fs/nfs/blocklayout/dev.c

@@ -0,0 +1,592 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2014-2016 Christoph Hellwig.
+ */
+#include <linux/sunrpc/svc.h>
+#include <linux/blkdev.h>
+#include <linux/nfs4.h>
+#include <linux/nfs_fs.h>
+#include <linux/nfs_xdr.h>
+#include <linux/pr.h>
+
+#include "blocklayout.h"
+#include "../nfs4trace.h"
+
+#define NFSDBG_FACILITY		NFSDBG_PNFS_LD
+
+static void bl_unregister_scsi(struct pnfs_block_dev *dev)
+{
+	struct block_device *bdev = file_bdev(dev->bdev_file);
+	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
+	int status;
+
+	status = ops->pr_register(bdev, dev->pr_key, 0, false);
+	if (status)
+		trace_bl_pr_key_unreg_err(bdev, dev->pr_key, status);
+	else
+		trace_bl_pr_key_unreg(bdev, dev->pr_key);
+}
+
+static bool bl_register_scsi(struct pnfs_block_dev *dev)
+{
+	struct block_device *bdev = file_bdev(dev->bdev_file);
+	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
+	int status;
+
+	if (test_and_set_bit(PNFS_BDEV_REGISTERED, &dev->flags))
+		return true;
+
+	status = ops->pr_register(bdev, 0, dev->pr_key, true);
+	if (status) {
+		trace_bl_pr_key_reg_err(bdev, dev->pr_key, status);
+		return false;
+	}
+	trace_bl_pr_key_reg(bdev, dev->pr_key);
+	return true;
+}
+
+static void bl_unregister_dev(struct pnfs_block_dev *dev)
+{
+	u32 i;
+
+	if (dev->nr_children) {
+		for (i = 0; i < dev->nr_children; i++)
+			bl_unregister_dev(&dev->children[i]);
+		return;
+	}
+
+	if (dev->type == PNFS_BLOCK_VOLUME_SCSI &&
+		test_and_clear_bit(PNFS_BDEV_REGISTERED, &dev->flags))
+		bl_unregister_scsi(dev);
+}
+
+bool bl_register_dev(struct pnfs_block_dev *dev)
+{
+	u32 i;
+
+	if (dev->nr_children) {
+		for (i = 0; i < dev->nr_children; i++) {
+			if (!bl_register_dev(&dev->children[i])) {
+				while (i > 0)
+					bl_unregister_dev(&dev->children[--i]);
+				return false;
+			}
+		}
+		return true;
+	}
+
+	if (dev->type == PNFS_BLOCK_VOLUME_SCSI)
+		return bl_register_scsi(dev);
+	return true;
+}
+
+static void
+bl_free_device(struct pnfs_block_dev *dev)
+{
+	bl_unregister_dev(dev);
+
+	if (dev->nr_children) {
+		int i;
+
+		for (i = 0; i < dev->nr_children; i++)
+			bl_free_device(&dev->children[i]);
+		kfree(dev->children);
+	} else {
+		if (dev->bdev_file)
+			fput(dev->bdev_file);
+	}
+}
+
+void
+bl_free_deviceid_node(struct nfs4_deviceid_node *d)
+{
+	struct pnfs_block_dev *dev =
+		container_of(d, struct pnfs_block_dev, node);
+
+	bl_free_device(dev);
+	kfree_rcu(dev, node.rcu);
+}
+
+static int
+nfs4_block_decode_volume(struct xdr_stream *xdr, struct pnfs_block_volume *b)
+{
+	__be32 *p;
+	int i;
+
+	p = xdr_inline_decode(xdr, 4);
+	if (!p)
+		return -EIO;
+	b->type = be32_to_cpup(p++);
+
+	switch (b->type) {
+	case PNFS_BLOCK_VOLUME_SIMPLE:
+		p = xdr_inline_decode(xdr, 4);
+		if (!p)
+			return -EIO;
+		b->simple.nr_sigs = be32_to_cpup(p++);
+		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;
+		}
+
+		b->simple.len = 4 + 4;
+		for (i = 0; i < b->simple.nr_sigs; i++) {
+			p = xdr_inline_decode(xdr, 8 + 4);
+			if (!p)
+				return -EIO;
+			p = xdr_decode_hyper(p, &b->simple.sigs[i].offset);
+			b->simple.sigs[i].sig_len = be32_to_cpup(p++);
+			if (b->simple.sigs[i].sig_len > PNFS_BLOCK_UUID_LEN) {
+				pr_info("signature too long: %d\n",
+					b->simple.sigs[i].sig_len);
+				return -EIO;
+			}
+
+			p = xdr_inline_decode(xdr, b->simple.sigs[i].sig_len);
+			if (!p)
+				return -EIO;
+			memcpy(&b->simple.sigs[i].sig, p,
+				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:
+		p = xdr_inline_decode(xdr, 8 + 8 + 4);
+		if (!p)
+			return -EIO;
+		p = xdr_decode_hyper(p, &b->slice.start);
+		p = xdr_decode_hyper(p, &b->slice.len);
+		b->slice.volume = be32_to_cpup(p++);
+		break;
+	case PNFS_BLOCK_VOLUME_CONCAT:
+		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)
+			return -EIO;
+		for (i = 0; i < b->concat.volumes_count; i++)
+			b->concat.volumes[i] = be32_to_cpup(p++);
+		break;
+	case PNFS_BLOCK_VOLUME_STRIPE:
+		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)
+			return -EIO;
+		for (i = 0; i < b->stripe.volumes_count; i++)
+			b->stripe.volumes[i] = be32_to_cpup(p++);
+		break;
+	case PNFS_BLOCK_VOLUME_SCSI:
+		p = xdr_inline_decode(xdr, 4 + 4 + 4);
+		if (!p)
+			return -EIO;
+		b->scsi.code_set = be32_to_cpup(p++);
+		b->scsi.designator_type = be32_to_cpup(p++);
+		b->scsi.designator_len = be32_to_cpup(p++);
+		p = xdr_inline_decode(xdr, b->scsi.designator_len);
+		if (!p)
+			return -EIO;
+		if (b->scsi.designator_len > 256)
+			return -EIO;
+		memcpy(&b->scsi.designator, p, b->scsi.designator_len);
+		p = xdr_inline_decode(xdr, 8);
+		if (!p)
+			return -EIO;
+		p = xdr_decode_hyper(p, &b->scsi.pr_key);
+		break;
+	default:
+		dprintk("unknown volume type!\n");
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static bool bl_map_simple(struct pnfs_block_dev *dev, u64 offset,
+		struct pnfs_block_dev_map *map)
+{
+	map->start = dev->start;
+	map->len = dev->len;
+	map->disk_offset = dev->disk_offset;
+	map->bdev = file_bdev(dev->bdev_file);
+	return true;
+}
+
+static bool bl_map_concat(struct pnfs_block_dev *dev, u64 offset,
+		struct pnfs_block_dev_map *map)
+{
+	int i;
+
+	for (i = 0; i < dev->nr_children; i++) {
+		struct pnfs_block_dev *child = &dev->children[i];
+
+		if (child->start > offset ||
+		    child->start + child->len <= offset)
+			continue;
+
+		child->map(child, offset - child->start, map);
+		return true;
+	}
+
+	dprintk("%s: ran off loop!\n", __func__);
+	return false;
+}
+
+static bool bl_map_stripe(struct pnfs_block_dev *dev, u64 offset,
+		struct pnfs_block_dev_map *map)
+{
+	struct pnfs_block_dev *child;
+	u64 chunk;
+	u32 chunk_idx;
+	u64 disk_chunk;
+	u64 disk_offset;
+
+	chunk = div_u64(offset, dev->chunk_size);
+	disk_chunk = div_u64_rem(chunk, dev->nr_children, &chunk_idx);
+
+	if (chunk_idx >= dev->nr_children) {
+		dprintk("%s: invalid chunk idx %d (%lld/%lld)\n",
+			__func__, chunk_idx, offset, dev->chunk_size);
+		/* error, should not happen */
+		return false;
+	}
+
+	/* truncate offset to the beginning of the stripe */
+	offset = chunk * dev->chunk_size;
+
+	/* disk offset of the stripe */
+	disk_offset = disk_chunk * dev->chunk_size;
+
+	child = &dev->children[chunk_idx];
+	child->map(child, disk_offset, map);
+
+	map->start += offset;
+	map->disk_offset += disk_offset;
+	map->len = dev->chunk_size;
+	return true;
+}
+
+static int
+bl_parse_deviceid(struct nfs_server *server, struct pnfs_block_dev *d,
+		struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask);
+
+
+static int
+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 file *bdev_file;
+	dev_t dev;
+
+	dev = bl_resolve_deviceid(server, v, gfp_mask);
+	if (!dev)
+		return -EIO;
+
+	bdev_file = bdev_file_open_by_dev(dev, BLK_OPEN_READ | BLK_OPEN_WRITE,
+				       NULL, NULL);
+	if (IS_ERR(bdev_file)) {
+		printk(KERN_WARNING "pNFS: failed to open device %d:%d (%ld)\n",
+			MAJOR(dev), MINOR(dev), PTR_ERR(bdev_file));
+		return PTR_ERR(bdev_file);
+	}
+	d->bdev_file = bdev_file;
+	d->len = bdev_nr_bytes(file_bdev(bdev_file));
+	d->map = bl_map_simple;
+
+	printk(KERN_INFO "pNFS: using block device %s\n",
+		file_bdev(bdev_file)->bd_disk->disk_name);
+	return 0;
+}
+
+static bool
+bl_validate_designator(struct pnfs_block_volume *v)
+{
+	switch (v->scsi.designator_type) {
+	case PS_DESIGNATOR_EUI64:
+		if (v->scsi.code_set != PS_CODE_SET_BINARY)
+			return false;
+
+		if (v->scsi.designator_len != 8 &&
+		    v->scsi.designator_len != 10 &&
+		    v->scsi.designator_len != 16)
+			return false;
+
+		return true;
+	case PS_DESIGNATOR_NAA:
+		if (v->scsi.code_set != PS_CODE_SET_BINARY)
+			return false;
+
+		if (v->scsi.designator_len != 8 &&
+		    v->scsi.designator_len != 16)
+			return false;
+
+		return true;
+	case PS_DESIGNATOR_T10:
+	case PS_DESIGNATOR_NAME:
+		pr_err("pNFS: unsupported designator "
+			"(code set %d, type %d, len %d.\n",
+			v->scsi.code_set,
+			v->scsi.designator_type,
+			v->scsi.designator_len);
+		return false;
+	default:
+		pr_err("pNFS: invalid designator "
+			"(code set %d, type %d, len %d.\n",
+			v->scsi.code_set,
+			v->scsi.designator_type,
+			v->scsi.designator_len);
+		return false;
+	}
+}
+
+static struct file *
+bl_open_path(struct pnfs_block_volume *v, const char *prefix)
+{
+	struct file *bdev_file;
+	const char *devname;
+
+	devname = kasprintf(GFP_KERNEL, "/dev/disk/by-id/%s%*phN",
+			prefix, v->scsi.designator_len, v->scsi.designator);
+	if (!devname)
+		return ERR_PTR(-ENOMEM);
+
+	bdev_file = bdev_file_open_by_path(devname,
+			BLK_OPEN_READ | BLK_OPEN_WRITE, NULL, NULL);
+	if (IS_ERR(bdev_file)) {
+		dprintk("failed to open device %s (%ld)\n",
+			devname, PTR_ERR(bdev_file));
+	} else {
+		pr_info("pNFS: using block device %s\n",
+			file_bdev(bdev_file)->bd_disk->disk_name);
+	}
+
+	kfree(devname);
+	return bdev_file;
+}
+
+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;
+	struct file *bdev_file;
+	int error;
+
+	if (!bl_validate_designator(v))
+		return -EINVAL;
+
+	/*
+	 * Try to open the RH/Fedora specific dm-mpath udev path first, as the
+	 * wwn- links will only point to the first discovered SCSI device there.
+	 * On other distributions like Debian, the default SCSI by-id path will
+	 * point to the dm-multipath device if one exists.
+	 */
+	bdev_file = bl_open_path(v, "dm-uuid-mpath-0x");
+	if (IS_ERR(bdev_file))
+		bdev_file = bl_open_path(v, "wwn-0x");
+	if (IS_ERR(bdev_file))
+		bdev_file = bl_open_path(v, "nvme-eui.");
+	if (IS_ERR(bdev_file)) {
+		pr_warn("pNFS: no device found for volume %*phN\n",
+			v->scsi.designator_len, v->scsi.designator);
+		return PTR_ERR(bdev_file);
+	}
+	d->bdev_file = bdev_file;
+	bdev = file_bdev(bdev_file);
+
+	d->len = bdev_nr_bytes(bdev);
+	d->map = bl_map_simple;
+	d->pr_key = v->scsi.pr_key;
+
+	if (d->len == 0) {
+		error = -ENODEV;
+		goto out_blkdev_put;
+	}
+
+	ops = bdev->bd_disk->fops->pr_ops;
+	if (!ops) {
+		pr_err("pNFS: block device %s does not support reservations.",
+				bdev->bd_disk->disk_name);
+		error = -EINVAL;
+		goto out_blkdev_put;
+	}
+
+	return 0;
+
+out_blkdev_put:
+	fput(d->bdev_file);
+	return error;
+}
+
+static int
+bl_parse_slice(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];
+	int ret;
+
+	ret = bl_parse_deviceid(server, d, volumes, v->slice.volume, gfp_mask);
+	if (ret)
+		return ret;
+
+	d->disk_offset = v->slice.start;
+	d->len = v->slice.len;
+	return 0;
+}
+
+static int
+bl_parse_concat(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];
+	u64 len = 0;
+	int ret, i;
+
+	d->children = kzalloc_objs(struct pnfs_block_dev,
+				   v->concat.volumes_count, gfp_mask);
+	if (!d->children)
+		return -ENOMEM;
+
+	for (i = 0; i < v->concat.volumes_count; i++) {
+		ret = bl_parse_deviceid(server, &d->children[i],
+				volumes, v->concat.volumes[i], gfp_mask);
+		if (ret)
+			return ret;
+
+		d->nr_children++;
+		d->children[i].start += len;
+		len += d->children[i].len;
+	}
+
+	d->len = len;
+	d->map = bl_map_concat;
+	return 0;
+}
+
+static int
+bl_parse_stripe(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];
+	u64 len = 0;
+	int ret, i;
+
+	d->children = kzalloc_objs(struct pnfs_block_dev,
+				   v->stripe.volumes_count, gfp_mask);
+	if (!d->children)
+		return -ENOMEM;
+
+	for (i = 0; i < v->stripe.volumes_count; i++) {
+		ret = bl_parse_deviceid(server, &d->children[i],
+				volumes, v->stripe.volumes[i], gfp_mask);
+		if (ret)
+			return ret;
+
+		d->nr_children++;
+		len += d->children[i].len;
+	}
+
+	d->len = len;
+	d->chunk_size = v->stripe.chunk_size;
+	d->map = bl_map_stripe;
+	return 0;
+}
+
+static int
+bl_parse_deviceid(struct nfs_server *server, struct pnfs_block_dev *d,
+		struct pnfs_block_volume *volumes, int idx, gfp_t gfp_mask)
+{
+	d->type = volumes[idx].type;
+
+	switch (d->type) {
+	case PNFS_BLOCK_VOLUME_SIMPLE:
+		return bl_parse_simple(server, d, volumes, idx, gfp_mask);
+	case PNFS_BLOCK_VOLUME_SLICE:
+		return bl_parse_slice(server, d, volumes, idx, gfp_mask);
+	case PNFS_BLOCK_VOLUME_CONCAT:
+		return bl_parse_concat(server, d, volumes, idx, gfp_mask);
+	case PNFS_BLOCK_VOLUME_STRIPE:
+		return bl_parse_stripe(server, d, volumes, idx, gfp_mask);
+	case PNFS_BLOCK_VOLUME_SCSI:
+		return bl_parse_scsi(server, d, volumes, idx, gfp_mask);
+	default:
+		dprintk("unsupported volume type: %d\n", d->type);
+		return -EIO;
+	}
+}
+
+struct nfs4_deviceid_node *
+bl_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
+		gfp_t gfp_mask)
+{
+	struct nfs4_deviceid_node *node = NULL;
+	struct pnfs_block_volume *volumes;
+	struct pnfs_block_dev *top;
+	struct xdr_stream xdr;
+	struct xdr_buf buf;
+	struct folio *scratch;
+	int nr_volumes, ret, i;
+	__be32 *p;
+
+	scratch = folio_alloc(gfp_mask, 0);
+	if (!scratch)
+		goto out;
+
+	xdr_init_decode_pages(&xdr, &buf, pdev->pages, pdev->pglen);
+	xdr_set_scratch_folio(&xdr, scratch);
+
+	p = xdr_inline_decode(&xdr, sizeof(__be32));
+	if (!p)
+		goto out_free_scratch;
+	nr_volumes = be32_to_cpup(p++);
+
+	volumes = kzalloc_objs(struct pnfs_block_volume, nr_volumes, gfp_mask);
+	if (!volumes)
+		goto out_free_scratch;
+
+	for (i = 0; i < nr_volumes; i++) {
+		ret = nfs4_block_decode_volume(&xdr, &volumes[i]);
+		if (ret < 0)
+			goto out_free_volumes;
+	}
+
+	top = kzalloc_obj(*top, gfp_mask);
+	if (!top)
+		goto out_free_volumes;
+
+	ret = bl_parse_deviceid(server, top, volumes, nr_volumes - 1, gfp_mask);
+
+	node = &top->node;
+	nfs4_init_deviceid_node(node, server, &pdev->dev_id);
+	if (ret)
+		nfs4_mark_deviceid_unavailable(node);
+
+out_free_volumes:
+	kfree(volumes);
+out_free_scratch:
+	folio_put(scratch);
+out:
+	return node;
+}

linux-7.2.3/fs/nfs/blocklayout/extent_tree.c

@@ -0,0 +1,726 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2014-2016 Christoph Hellwig.
+ */
+
+#include <linux/vmalloc.h>
+
+#include "blocklayout.h"
+#include "../nfs4trace.h"
+
+#define NFSDBG_FACILITY		NFSDBG_PNFS_LD
+
+static inline struct pnfs_block_extent *
+ext_node(struct rb_node *node)
+{
+	return rb_entry(node, struct pnfs_block_extent, be_node);
+}
+
+static struct pnfs_block_extent *
+ext_tree_first(struct rb_root *root)
+{
+	struct rb_node *node = rb_first(root);
+	return node ? ext_node(node) : NULL;
+}
+
+static struct pnfs_block_extent *
+ext_tree_prev(struct pnfs_block_extent *be)
+{
+	struct rb_node *node = rb_prev(&be->be_node);
+	return node ? ext_node(node) : NULL;
+}
+
+static struct pnfs_block_extent *
+ext_tree_next(struct pnfs_block_extent *be)
+{
+	struct rb_node *node = rb_next(&be->be_node);
+	return node ? ext_node(node) : NULL;
+}
+
+static inline sector_t
+ext_f_end(struct pnfs_block_extent *be)
+{
+	return be->be_f_offset + be->be_length;
+}
+
+static struct pnfs_block_extent *
+__ext_tree_search(struct rb_root *root, sector_t start)
+{
+	struct rb_node *node = root->rb_node;
+	struct pnfs_block_extent *be = NULL;
+
+	while (node) {
+		be = ext_node(node);
+		if (start < be->be_f_offset)
+			node = node->rb_left;
+		else if (start >= ext_f_end(be))
+			node = node->rb_right;
+		else
+			return be;
+	}
+
+	if (be) {
+		if (start < be->be_f_offset)
+			return be;
+
+		if (start >= ext_f_end(be))
+			return ext_tree_next(be);
+	}
+
+	return NULL;
+}
+
+static bool
+ext_can_merge(struct pnfs_block_extent *be1, struct pnfs_block_extent *be2)
+{
+	if (be1->be_state != be2->be_state)
+		return false;
+	if (be1->be_device != be2->be_device)
+		return false;
+
+	if (be1->be_f_offset + be1->be_length != be2->be_f_offset)
+		return false;
+
+	if (be1->be_state != PNFS_BLOCK_NONE_DATA &&
+	    (be1->be_v_offset + be1->be_length != be2->be_v_offset))
+		return false;
+
+	if (be1->be_state == PNFS_BLOCK_INVALID_DATA &&
+	    be1->be_tag != be2->be_tag)
+		return false;
+
+	return true;
+}
+
+static struct pnfs_block_extent *
+ext_try_to_merge_left(struct rb_root *root, struct pnfs_block_extent *be)
+{
+	struct pnfs_block_extent *left = ext_tree_prev(be);
+
+	if (left && ext_can_merge(left, be)) {
+		left->be_length += be->be_length;
+		rb_erase(&be->be_node, root);
+		nfs4_put_deviceid_node(be->be_device);
+		kfree(be);
+		return left;
+	}
+
+	return be;
+}
+
+static struct pnfs_block_extent *
+ext_try_to_merge_right(struct rb_root *root, struct pnfs_block_extent *be)
+{
+	struct pnfs_block_extent *right = ext_tree_next(be);
+
+	if (right && ext_can_merge(be, right)) {
+		be->be_length += right->be_length;
+		rb_erase(&right->be_node, root);
+		nfs4_put_deviceid_node(right->be_device);
+		kfree(right);
+	}
+
+	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)
+{
+	struct rb_node **p = &root->rb_node, *parent = NULL;
+	struct pnfs_block_extent *be;
+
+	while (*p) {
+		parent = *p;
+		be = ext_node(parent);
+
+		if (new->be_f_offset < be->be_f_offset) {
+			if (merge_ok && ext_can_merge(new, be)) {
+				be->be_f_offset = new->be_f_offset;
+				if (be->be_state != PNFS_BLOCK_NONE_DATA)
+					be->be_v_offset = new->be_v_offset;
+				be->be_length += new->be_length;
+				be = ext_try_to_merge_left(root, be);
+				goto free_new;
+			}
+			p = &(*p)->rb_left;
+		} else if (new->be_f_offset >= ext_f_end(be)) {
+			if (merge_ok && ext_can_merge(be, new)) {
+				be->be_length += new->be_length;
+				be = ext_try_to_merge_right(root, be);
+				goto free_new;
+			}
+			p = &(*p)->rb_right;
+		} else {
+			BUG();
+		}
+	}
+
+	rb_link_node(&new->be_node, parent, p);
+	rb_insert_color(&new->be_node, root);
+	return;
+free_new:
+	nfs4_put_deviceid_node(new->be_device);
+	kfree(new);
+}
+
+static int
+__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;
+	sector_t orig_v_offset;
+	sector_t orig_len;
+
+	be = __ext_tree_search(root, start);
+	if (!be)
+		return 0;
+	if (be->be_f_offset >= end)
+		return 0;
+
+	orig_v_offset = be->be_v_offset;
+	orig_len = be->be_length;
+
+	if (start > be->be_f_offset)
+		len1 = start - be->be_f_offset;
+	if (ext_f_end(be) > end)
+		len2 = ext_f_end(be) - end;
+
+	if (len2 > 0) {
+		if (len1 > 0) {
+			struct pnfs_block_extent *new;
+
+			new = kzalloc_obj(*new, GFP_ATOMIC);
+			if (!new)
+				return -ENOMEM;
+
+			be->be_length = len1;
+
+			new->be_f_offset = end;
+			if (be->be_state != PNFS_BLOCK_NONE_DATA) {
+				new->be_v_offset =
+					orig_v_offset + orig_len - len2;
+			}
+			new->be_length = len2;
+			new->be_state = be->be_state;
+			new->be_tag = be->be_tag;
+			new->be_device = nfs4_get_deviceid(be->be_device);
+
+			__ext_tree_insert(root, new, true);
+		} else {
+			be->be_f_offset = end;
+			if (be->be_state != PNFS_BLOCK_NONE_DATA) {
+				be->be_v_offset =
+					orig_v_offset + orig_len - len2;
+			}
+			be->be_length = len2;
+		}
+	} else {
+		if (len1 > 0) {
+			be->be_length = len1;
+			be = ext_tree_next(be);
+		}
+
+		while (be && ext_f_end(be) <= end) {
+			struct pnfs_block_extent *next = ext_tree_next(be);
+
+			rb_erase(&be->be_node, root);
+			list_add_tail(&be->be_list, tmp);
+			be = next;
+		}
+
+		if (be && be->be_f_offset < end) {
+			len1 = ext_f_end(be) - end;
+			be->be_f_offset = end;
+			if (be->be_state != PNFS_BLOCK_NONE_DATA)
+				be->be_v_offset += be->be_length - len1;
+			be->be_length = len1;
+		}
+	}
+
+	return 0;
+}
+
+int
+ext_tree_insert(struct pnfs_block_layout *bl, struct pnfs_block_extent *new)
+{
+	struct pnfs_block_extent *be;
+	struct rb_root *root;
+	int err = 0;
+
+	switch (new->be_state) {
+	case PNFS_BLOCK_READWRITE_DATA:
+	case PNFS_BLOCK_INVALID_DATA:
+		root = &bl->bl_ext_rw;
+		break;
+	case PNFS_BLOCK_READ_DATA:
+	case PNFS_BLOCK_NONE_DATA:
+		root = &bl->bl_ext_ro;
+		break;
+	default:
+		dprintk("invalid extent type\n");
+		return -EINVAL;
+	}
+
+	spin_lock(&bl->bl_ext_lock);
+retry:
+	be = __ext_tree_search(root, new->be_f_offset);
+	if (!be || be->be_f_offset >= ext_f_end(new)) {
+		__ext_tree_insert(root, new, true);
+	} else if (new->be_f_offset >= be->be_f_offset) {
+		if (ext_f_end(new) <= ext_f_end(be)) {
+			nfs4_put_deviceid_node(new->be_device);
+			kfree(new);
+		} else {
+			sector_t new_len = ext_f_end(new) - ext_f_end(be);
+			sector_t diff = new->be_length - new_len;
+
+			new->be_f_offset += diff;
+			new->be_v_offset += diff;
+			new->be_length = new_len;
+			goto retry;
+		}
+	} else if (ext_f_end(new) <= ext_f_end(be)) {
+		new->be_length = be->be_f_offset - new->be_f_offset;
+		__ext_tree_insert(root, new, true);
+	} else {
+		struct pnfs_block_extent *split;
+		sector_t new_len = ext_f_end(new) - ext_f_end(be);
+		sector_t diff = new->be_length - new_len;
+
+		split = kmemdup(new, sizeof(*new), GFP_ATOMIC);
+		if (!split) {
+			err = -EINVAL;
+			goto out;
+		}
+
+		split->be_length = be->be_f_offset - split->be_f_offset;
+		split->be_device = nfs4_get_deviceid(new->be_device);
+		__ext_tree_insert(root, split, true);
+
+		new->be_f_offset += diff;
+		new->be_v_offset += diff;
+		new->be_length = new_len;
+		goto retry;
+	}
+out:
+	spin_unlock(&bl->bl_ext_lock);
+	return err;
+}
+
+static bool
+__ext_tree_lookup(struct rb_root *root, sector_t isect,
+		struct pnfs_block_extent *ret)
+{
+	struct rb_node *node;
+	struct pnfs_block_extent *be;
+
+	node = root->rb_node;
+	while (node) {
+		be = ext_node(node);
+		if (isect < be->be_f_offset)
+			node = node->rb_left;
+		else if (isect >= ext_f_end(be))
+			node = node->rb_right;
+		else {
+			*ret = *be;
+			return true;
+		}
+	}
+
+	return false;
+}
+
+bool
+ext_tree_lookup(struct pnfs_block_layout *bl, sector_t isect,
+	    struct pnfs_block_extent *ret, bool rw)
+{
+	bool found = false;
+
+	spin_lock(&bl->bl_ext_lock);
+	if (!rw)
+		found = __ext_tree_lookup(&bl->bl_ext_ro, isect, ret);
+	if (!found)
+		found = __ext_tree_lookup(&bl->bl_ext_rw, isect, ret);
+	spin_unlock(&bl->bl_ext_lock);
+
+	return found;
+}
+
+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, &tmp);
+	if (rw) {
+		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;
+}
+
+static int
+ext_tree_split(struct rb_root *root, struct pnfs_block_extent *be,
+		sector_t split)
+{
+	struct pnfs_block_extent *new;
+	sector_t orig_len = be->be_length;
+
+	new = kzalloc_obj(*new, GFP_ATOMIC);
+	if (!new)
+		return -ENOMEM;
+
+	be->be_length = split - be->be_f_offset;
+
+	new->be_f_offset = split;
+	if (be->be_state != PNFS_BLOCK_NONE_DATA)
+		new->be_v_offset = be->be_v_offset + be->be_length;
+	new->be_length = orig_len - be->be_length;
+	new->be_state = be->be_state;
+	new->be_tag = be->be_tag;
+	new->be_device = nfs4_get_deviceid(be->be_device);
+
+	__ext_tree_insert(root, new, false);
+	return 0;
+}
+
+int
+ext_tree_mark_written(struct pnfs_block_layout *bl, sector_t start,
+		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, &tmp);
+	if (err)
+		goto out;
+
+	/*
+	 * Then mark all invalid extents in the range as written to.
+	 */
+	for (be = __ext_tree_search(root, start); be; be = ext_tree_next(be)) {
+		if (be->be_f_offset >= end)
+			break;
+
+		if (be->be_state != PNFS_BLOCK_INVALID_DATA || be->be_tag)
+			continue;
+
+		if (be->be_f_offset < start) {
+			struct pnfs_block_extent *left = ext_tree_prev(be);
+
+			if (left && ext_can_merge(left, be)) {
+				sector_t diff = start - be->be_f_offset;
+
+				left->be_length += diff;
+
+				be->be_f_offset += diff;
+				be->be_v_offset += diff;
+				be->be_length -= diff;
+			} else {
+				err = ext_tree_split(root, be, start);
+				if (err)
+					goto out;
+			}
+		}
+
+		if (ext_f_end(be) > end) {
+			struct pnfs_block_extent *right = ext_tree_next(be);
+
+			if (right && ext_can_merge(be, right)) {
+				sector_t diff = end - be->be_f_offset;
+
+				be->be_length -= diff;
+
+				right->be_f_offset -= diff;
+				right->be_v_offset -= diff;
+				right->be_length += diff;
+			} else {
+				err = ext_tree_split(root, be, end);
+				if (err)
+					goto out;
+			}
+		}
+
+		if (be->be_f_offset >= start && ext_f_end(be) <= end) {
+			be->be_tag = EXTENT_WRITTEN;
+			be = ext_try_to_merge_left(root, be);
+			be = ext_try_to_merge_right(root, be);
+		}
+	}
+out:
+	if (bl->bl_lwb < lwb)
+		bl->bl_lwb = lwb;
+	spin_unlock(&bl->bl_ext_lock);
+
+	__ext_put_deviceids(&tmp);
+	return err;
+}
+
+static size_t ext_tree_layoutupdate_size(struct pnfs_block_layout *bl, size_t count)
+{
+	if (bl->bl_scsi_layout)
+		return sizeof(__be32) + PNFS_SCSI_RANGE_SIZE * count;
+	else
+		return sizeof(__be32) + PNFS_BLOCK_EXTENT_SIZE * count;
+}
+
+static void ext_tree_free_commitdata(struct nfs4_layoutcommit_args *arg,
+		size_t buffer_size)
+{
+	if (arg->layoutupdate_pages != &arg->layoutupdate_page) {
+		int nr_pages = DIV_ROUND_UP(buffer_size, PAGE_SIZE), i;
+
+		for (i = 0; i < nr_pages; i++)
+			put_page(arg->layoutupdate_pages[i]);
+		vfree(arg->start_p);
+		kfree(arg->layoutupdate_pages);
+	} else {
+		put_page(arg->layoutupdate_page);
+	}
+}
+
+static __be32 *encode_block_extent(struct pnfs_block_extent *be, __be32 *p)
+{
+	p = xdr_encode_opaque_fixed(p, be->be_device->deviceid.data,
+			NFS4_DEVICEID4_SIZE);
+	p = xdr_encode_hyper(p, be->be_f_offset << SECTOR_SHIFT);
+	p = xdr_encode_hyper(p, be->be_length << SECTOR_SHIFT);
+	p = xdr_encode_hyper(p, 0LL);
+	*p++ = cpu_to_be32(PNFS_BLOCK_READWRITE_DATA);
+	return p;
+}
+
+static __be32 *encode_scsi_range(struct pnfs_block_extent *be, __be32 *p)
+{
+	p = xdr_encode_hyper(p, be->be_f_offset << SECTOR_SHIFT);
+	return xdr_encode_hyper(p, be->be_length << SECTOR_SHIFT);
+}
+
+/**
+ * ext_tree_try_encode_commit - try to encode all extents into the buffer
+ * @bl: pointer to the layout
+ * @p: pointer to the output buffer
+ * @buffer_size: size of the output buffer
+ * @count: output pointer to the number of encoded extents
+ * @lastbyte: output pointer to the last written byte
+ *
+ * Return values:
+ *   %0: Success, all required extents encoded, outputs are valid
+ *   %-ENOSPC: Buffer too small, nothing encoded, outputs are invalid
+ */
+static int
+ext_tree_try_encode_commit(struct pnfs_block_layout *bl, __be32 *p,
+		size_t buffer_size, size_t *count, __u64 *lastbyte)
+{
+	struct pnfs_block_extent *be;
+
+	spin_lock(&bl->bl_ext_lock);
+	for (be = ext_tree_first(&bl->bl_ext_rw); be; be = ext_tree_next(be)) {
+		if (be->be_state != PNFS_BLOCK_INVALID_DATA ||
+		    be->be_tag != EXTENT_WRITTEN)
+			continue;
+
+		(*count)++;
+		if (ext_tree_layoutupdate_size(bl, *count) > buffer_size) {
+			spin_unlock(&bl->bl_ext_lock);
+			return -ENOSPC;
+		}
+	}
+	for (be = ext_tree_first(&bl->bl_ext_rw); be; be = ext_tree_next(be)) {
+		if (be->be_state != PNFS_BLOCK_INVALID_DATA ||
+		    be->be_tag != EXTENT_WRITTEN)
+			continue;
+
+		if (bl->bl_scsi_layout)
+			p = encode_scsi_range(be, p);
+		else
+			p = encode_block_extent(be, p);
+		be->be_tag = EXTENT_COMMITTING;
+	}
+	*lastbyte = (bl->bl_lwb != 0) ? bl->bl_lwb - 1 : U64_MAX;
+	bl->bl_lwb = 0;
+	spin_unlock(&bl->bl_ext_lock);
+
+	return 0;
+}
+
+/**
+ * ext_tree_encode_commit - encode as much as possible extents into the buffer
+ * @bl: pointer to the layout
+ * @p: pointer to the output buffer
+ * @buffer_size: size of the output buffer
+ * @count: output pointer to the number of encoded extents
+ * @lastbyte: output pointer to the last written byte
+ *
+ * Return values:
+ *   %0: Success, all required extents encoded, outputs are valid
+ *   %-ENOSPC: Buffer too small, some extents are encoded, outputs are valid
+ */
+static int
+ext_tree_encode_commit(struct pnfs_block_layout *bl, __be32 *p,
+		size_t buffer_size, size_t *count, __u64 *lastbyte)
+{
+	struct pnfs_block_extent *be, *be_prev;
+	int ret = 0;
+
+	spin_lock(&bl->bl_ext_lock);
+	for (be = ext_tree_first(&bl->bl_ext_rw); be; be = ext_tree_next(be)) {
+		if (be->be_state != PNFS_BLOCK_INVALID_DATA ||
+		    be->be_tag != EXTENT_WRITTEN)
+			continue;
+
+		(*count)++;
+		if (ext_tree_layoutupdate_size(bl, *count) > buffer_size) {
+			(*count)--;
+			ret = -ENOSPC;
+			break;
+		}
+
+		if (bl->bl_scsi_layout)
+			p = encode_scsi_range(be, p);
+		else
+			p = encode_block_extent(be, p);
+		be->be_tag = EXTENT_COMMITTING;
+		be_prev = be;
+	}
+	if (!ret) {
+		*lastbyte = (bl->bl_lwb != 0) ? bl->bl_lwb - 1 : U64_MAX;
+		bl->bl_lwb = 0;
+	} else {
+		*lastbyte = be_prev->be_f_offset + be_prev->be_length;
+		*lastbyte <<= SECTOR_SHIFT;
+		*lastbyte -= 1;
+	}
+	spin_unlock(&bl->bl_ext_lock);
+
+	return ret;
+}
+
+/**
+ * ext_tree_prepare_commit - encode extents that need to be committed
+ * @arg: layout commit data
+ *
+ * Return values:
+ *   %0: Success, all required extents are encoded
+ *   %-ENOSPC: Some extents are encoded, but not all, due to RPC size limit
+ *   %-ENOMEM: Out of memory, extents not encoded
+ */
+int
+ext_tree_prepare_commit(struct nfs4_layoutcommit_args *arg)
+{
+	struct pnfs_block_layout *bl = BLK_LO2EXT(NFS_I(arg->inode)->layout);
+	size_t count = 0, buffer_size = PAGE_SIZE;
+	__be32 *start_p;
+	int ret;
+
+	arg->layoutupdate_page = alloc_page(GFP_NOFS);
+	if (!arg->layoutupdate_page)
+		return -ENOMEM;
+	start_p = page_address(arg->layoutupdate_page);
+	arg->layoutupdate_pages = &arg->layoutupdate_page;
+
+	ret = ext_tree_try_encode_commit(bl, start_p + 1, buffer_size,
+			&count, &arg->lastbytewritten);
+	if (unlikely(ret)) {
+		ext_tree_free_commitdata(arg, buffer_size);
+
+		buffer_size = NFS_SERVER(arg->inode)->wsize;
+		count = 0;
+
+		arg->layoutupdate_pages =
+			kzalloc_objs(struct page *,
+				     DIV_ROUND_UP(buffer_size, PAGE_SIZE),
+				     GFP_NOFS);
+		if (!arg->layoutupdate_pages)
+			return -ENOMEM;
+
+		start_p = __vmalloc(buffer_size, GFP_NOFS);
+		if (!start_p) {
+			kfree(arg->layoutupdate_pages);
+			return -ENOMEM;
+		}
+
+		ret = ext_tree_encode_commit(bl, start_p + 1, buffer_size,
+				&count, &arg->lastbytewritten);
+	}
+
+	*start_p = cpu_to_be32(count);
+	arg->layoutupdate_len = ext_tree_layoutupdate_size(bl, count);
+
+	if (unlikely(arg->layoutupdate_pages != &arg->layoutupdate_page)) {
+		void *p = start_p, *end = p + arg->layoutupdate_len;
+		struct page *page = NULL;
+		int i = 0;
+
+		arg->start_p = start_p;
+		for ( ; p < end; p += PAGE_SIZE) {
+			page = vmalloc_to_page(p);
+			arg->layoutupdate_pages[i++] = page;
+			get_page(page);
+		}
+	}
+
+	trace_bl_ext_tree_prepare_commit(ret, count,
+			arg->lastbytewritten, !!ret);
+	return ret;
+}
+
+void
+ext_tree_mark_committed(struct nfs4_layoutcommit_args *arg, int status)
+{
+	struct pnfs_block_layout *bl = BLK_LO2EXT(NFS_I(arg->inode)->layout);
+	struct rb_root *root = &bl->bl_ext_rw;
+	struct pnfs_block_extent *be;
+
+	dprintk("%s status %d\n", __func__, status);
+
+	ext_tree_free_commitdata(arg, arg->layoutupdate_len);
+
+	spin_lock(&bl->bl_ext_lock);
+	for (be = ext_tree_first(root); be; be = ext_tree_next(be)) {
+		if (be->be_state != PNFS_BLOCK_INVALID_DATA ||
+		    be->be_tag != EXTENT_COMMITTING)
+			continue;
+
+		if (status) {
+			/*
+			 * Mark as written and try again.
+			 *
+			 * XXX: some real error handling here wouldn't hurt..
+			 */
+			be->be_tag = EXTENT_WRITTEN;
+		} else {
+			be->be_state = PNFS_BLOCK_READWRITE_DATA;
+			be->be_tag = 0;
+		}
+
+		be = ext_try_to_merge_left(root, be);
+		be = ext_try_to_merge_right(root, be);
+	}
+	spin_unlock(&bl->bl_ext_lock);
+}

linux-7.2.3/fs/nfs/blocklayout/rpc_pipefs.c

@@ -0,0 +1,270 @@
+/*
+ *  Copyright (c) 2006,2007 The Regents of the University of Michigan.
+ *  All rights reserved.
+ *
+ *  Andy Adamson <andros@citi.umich.edu>
+ *  Fred Isaman <iisaman@umich.edu>
+ *
+ * permission is granted to use, copy, create derivative works and
+ * redistribute this software and such derivative works for any purpose,
+ * so long as the name of the university of michigan is not used in
+ * any advertising or publicity pertaining to the use or distribution
+ * of this software without specific, written prior authorization.  if
+ * the above copyright notice or any other identification of the
+ * university of michigan is included in any copy of any portion of
+ * this software, then the disclaimer below must also be included.
+ *
+ * this software is provided as is, without representation from the
+ * university of michigan as to its fitness for any purpose, and without
+ * warranty by the university of michigan of any kind, either express
+ * or implied, including without limitation the implied warranties of
+ * merchantability and fitness for a particular purpose.  the regents
+ * of the university of michigan shall not be liable for any damages,
+ * including special, indirect, incidental, or consequential damages,
+ * with respect to any claim arising out or in connection with the use
+ * of the software, even if it has been or is hereafter advised of the
+ * possibility of such damages.
+ */
+
+#include <linux/module.h>
+#include <linux/blkdev.h>
+
+#include "blocklayout.h"
+
+#define NFSDBG_FACILITY         NFSDBG_PNFS_LD
+
+static void
+nfs4_encode_simple(__be32 *p, struct pnfs_block_volume *b)
+{
+	int i;
+
+	*p++ = cpu_to_be32(1);
+	*p++ = cpu_to_be32(b->type);
+	*p++ = cpu_to_be32(b->simple.nr_sigs);
+	for (i = 0; i < b->simple.nr_sigs; i++) {
+		p = xdr_encode_hyper(p, b->simple.sigs[i].offset);
+		p = xdr_encode_opaque(p, b->simple.sigs[i].sig,
+					 b->simple.sigs[i].sig_len);
+	}
+}
+
+dev_t
+bl_resolve_deviceid(struct nfs_server *server, struct pnfs_block_volume *b,
+		gfp_t gfp_mask)
+{
+	struct net *net = server->nfs_client->cl_net;
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+	struct bl_dev_msg *reply = &nn->bl_mount_reply;
+	struct bl_pipe_msg bl_pipe_msg;
+	struct rpc_pipe_msg *msg = &bl_pipe_msg.msg;
+	struct bl_msg_hdr *bl_msg;
+	DECLARE_WAITQUEUE(wq, current);
+	dev_t dev = 0;
+	int rc;
+
+	dprintk("%s CREATING PIPEFS MESSAGE\n", __func__);
+
+	mutex_lock(&nn->bl_mutex);
+	bl_pipe_msg.bl_wq = &nn->bl_wq;
+
+	b->simple.len += 4;	/* single volume */
+	if (b->simple.len > PAGE_SIZE)
+		goto out_unlock;
+
+	memset(msg, 0, sizeof(*msg));
+	msg->len = sizeof(*bl_msg) + b->simple.len;
+	msg->data = kzalloc(msg->len, gfp_mask);
+	if (!msg->data)
+		goto out_unlock;
+
+	bl_msg = msg->data;
+	bl_msg->type = BL_DEVICE_MOUNT;
+	bl_msg->totallen = b->simple.len;
+	nfs4_encode_simple(msg->data + sizeof(*bl_msg), b);
+
+	dprintk("%s CALLING USERSPACE DAEMON\n", __func__);
+	add_wait_queue(&nn->bl_wq, &wq);
+	rc = rpc_queue_upcall(nn->bl_device_pipe, msg);
+	if (rc < 0) {
+		remove_wait_queue(&nn->bl_wq, &wq);
+		goto out_free_data;
+	}
+
+	set_current_state(TASK_UNINTERRUPTIBLE);
+	schedule();
+	remove_wait_queue(&nn->bl_wq, &wq);
+
+	if (reply->status != BL_DEVICE_REQUEST_PROC) {
+		printk(KERN_WARNING "%s failed to decode device: %d\n",
+			__func__, reply->status);
+		goto out_free_data;
+	}
+
+	dev = MKDEV(reply->major, reply->minor);
+out_free_data:
+	kfree(msg->data);
+out_unlock:
+	mutex_unlock(&nn->bl_mutex);
+	return dev;
+}
+
+static ssize_t bl_pipe_downcall(struct file *filp, const char __user *src,
+			 size_t mlen)
+{
+	struct nfs_net *nn = net_generic(file_inode(filp)->i_sb->s_fs_info,
+					 nfs_net_id);
+
+	if (mlen != sizeof (struct bl_dev_msg))
+		return -EINVAL;
+
+	if (copy_from_user(&nn->bl_mount_reply, src, mlen) != 0)
+		return -EFAULT;
+
+	wake_up(&nn->bl_wq);
+
+	return mlen;
+}
+
+static void bl_pipe_destroy_msg(struct rpc_pipe_msg *msg)
+{
+	struct bl_pipe_msg *bl_pipe_msg =
+		container_of(msg, struct bl_pipe_msg, msg);
+
+	if (msg->errno >= 0)
+		return;
+	wake_up(bl_pipe_msg->bl_wq);
+}
+
+static const struct rpc_pipe_ops bl_upcall_ops = {
+	.upcall		= rpc_pipe_generic_upcall,
+	.downcall	= bl_pipe_downcall,
+	.destroy_msg	= bl_pipe_destroy_msg,
+};
+
+static int nfs4blocklayout_register_sb(struct super_block *sb,
+					    struct rpc_pipe *pipe)
+{
+	struct dentry *dir;
+	int err;
+
+	dir = rpc_d_lookup_sb(sb, NFS_PIPE_DIRNAME);
+	if (dir == NULL)
+		return -ENOENT;
+	err = rpc_mkpipe_dentry(dir, "blocklayout", NULL, pipe);
+	dput(dir);
+	return err;
+}
+
+static int rpc_pipefs_event(struct notifier_block *nb, unsigned long event,
+			   void *ptr)
+{
+	struct super_block *sb = ptr;
+	struct net *net = sb->s_fs_info;
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+	int ret = 0;
+
+	if (!try_module_get(THIS_MODULE))
+		return 0;
+
+	if (nn->bl_device_pipe == NULL) {
+		module_put(THIS_MODULE);
+		return 0;
+	}
+
+	switch (event) {
+	case RPC_PIPEFS_MOUNT:
+		ret = nfs4blocklayout_register_sb(sb, nn->bl_device_pipe);
+		break;
+	case RPC_PIPEFS_UMOUNT:
+		rpc_unlink(nn->bl_device_pipe);
+		break;
+	default:
+		ret = -ENOTSUPP;
+		break;
+	}
+	module_put(THIS_MODULE);
+	return ret;
+}
+
+static struct notifier_block nfs4blocklayout_block = {
+	.notifier_call = rpc_pipefs_event,
+};
+
+static int nfs4blocklayout_register_net(struct net *net, struct rpc_pipe *pipe)
+{
+	struct super_block *pipefs_sb;
+	int ret;
+
+	pipefs_sb = rpc_get_sb_net(net);
+	if (!pipefs_sb)
+		return 0;
+	ret = nfs4blocklayout_register_sb(pipefs_sb, pipe);
+	rpc_put_sb_net(net);
+	return ret;
+}
+
+static void nfs4blocklayout_unregister_net(struct net *net,
+					   struct rpc_pipe *pipe)
+{
+	struct super_block *pipefs_sb;
+
+	pipefs_sb = rpc_get_sb_net(net);
+	if (pipefs_sb) {
+		rpc_unlink(pipe);
+		rpc_put_sb_net(net);
+	}
+}
+
+static int nfs4blocklayout_net_init(struct net *net)
+{
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+	int err;
+
+	mutex_init(&nn->bl_mutex);
+	init_waitqueue_head(&nn->bl_wq);
+	nn->bl_device_pipe = rpc_mkpipe_data(&bl_upcall_ops, 0);
+	if (IS_ERR(nn->bl_device_pipe))
+		return PTR_ERR(nn->bl_device_pipe);
+	err = nfs4blocklayout_register_net(net, nn->bl_device_pipe);
+	if (unlikely(err))
+		rpc_destroy_pipe_data(nn->bl_device_pipe);
+	return err;
+}
+
+static void nfs4blocklayout_net_exit(struct net *net)
+{
+	struct nfs_net *nn = net_generic(net, nfs_net_id);
+
+	nfs4blocklayout_unregister_net(net, nn->bl_device_pipe);
+	rpc_destroy_pipe_data(nn->bl_device_pipe);
+	nn->bl_device_pipe = NULL;
+}
+
+static struct pernet_operations nfs4blocklayout_net_ops = {
+	.init = nfs4blocklayout_net_init,
+	.exit = nfs4blocklayout_net_exit,
+};
+
+int __init bl_init_pipefs(void)
+{
+	int ret;
+
+	ret = rpc_pipefs_notifier_register(&nfs4blocklayout_block);
+	if (ret)
+		goto out;
+	ret = register_pernet_subsys(&nfs4blocklayout_net_ops);
+	if (ret)
+		goto out_unregister_notifier;
+	return 0;
+
+out_unregister_notifier:
+	rpc_pipefs_notifier_unregister(&nfs4blocklayout_block);
+out:
+	return ret;
+}
+
+void bl_cleanup_pipefs(void)
+{
+	rpc_pipefs_notifier_unregister(&nfs4blocklayout_block);
+	unregister_pernet_subsys(&nfs4blocklayout_net_ops);
+}