[mirror_ubuntu-jammy-kernel.git] / fs / nfs / flexfilelayout / flexfilelayoutdev.c

// SPDX-License-Identifier: GPL-2.0
/*
 * Device operations for the pnfs nfs4 file layout driver.
 *
 * Copyright (c) 2014, Primary Data, Inc. All rights reserved.
 *
 * Tao Peng <bergwolf@primarydata.com>
 */

#include <linux/nfs_fs.h>
#include <linux/vmalloc.h>
#include <linux/module.h>
#include <linux/sunrpc/addr.h>

#include "../internal.h"
#include "../nfs4session.h"
#include "flexfilelayout.h"

#define NFSDBG_FACILITY		NFSDBG_PNFS_LD

static unsigned int dataserver_timeo = NFS_DEF_TCP_RETRANS;
static unsigned int dataserver_retrans;

static bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg);

void nfs4_ff_layout_put_deviceid(struct nfs4_ff_layout_ds *mirror_ds)
{
	if (!IS_ERR_OR_NULL(mirror_ds))
		nfs4_put_deviceid_node(&mirror_ds->id_node);
}

void nfs4_ff_layout_free_deviceid(struct nfs4_ff_layout_ds *mirror_ds)
{
	nfs4_print_deviceid(&mirror_ds->id_node.deviceid);
	nfs4_pnfs_ds_put(mirror_ds->ds);
	kfree(mirror_ds->ds_versions);
	kfree_rcu(mirror_ds, id_node.rcu);
}

/* Decode opaque device data and construct new_ds using it */
struct nfs4_ff_layout_ds *
nfs4_ff_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
			    gfp_t gfp_flags)
{
	struct xdr_stream stream;
	struct xdr_buf buf;
	struct page *scratch;
	struct list_head dsaddrs;
	struct nfs4_pnfs_ds_addr *da;
	struct nfs4_ff_layout_ds *new_ds = NULL;
	struct nfs4_ff_ds_version *ds_versions = NULL;
	u32 mp_count;
	u32 version_count;
	__be32 *p;
	int i, ret = -ENOMEM;

	/* set up xdr stream */
	scratch = alloc_page(gfp_flags);
	if (!scratch)
		goto out_err;

	new_ds = kzalloc(sizeof(struct nfs4_ff_layout_ds), gfp_flags);
	if (!new_ds)
		goto out_scratch;

	nfs4_init_deviceid_node(&new_ds->id_node,
				server,
				&pdev->dev_id);
	INIT_LIST_HEAD(&dsaddrs);

	xdr_init_decode_pages(&stream, &buf, pdev->pages, pdev->pglen);
	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);

	/* multipath count */
	p = xdr_inline_decode(&stream, 4);
	if (unlikely(!p))
		goto out_err_drain_dsaddrs;
	mp_count = be32_to_cpup(p);
	dprintk("%s: multipath ds count %d\n", __func__, mp_count);

	for (i = 0; i < mp_count; i++) {
		/* multipath ds */
		da = nfs4_decode_mp_ds_addr(server->nfs_client->cl_net,
					    &stream, gfp_flags);
		if (da)
			list_add_tail(&da->da_node, &dsaddrs);
	}
	if (list_empty(&dsaddrs)) {
		dprintk("%s: no suitable DS addresses found\n",
			__func__);
		ret = -ENOMEDIUM;
		goto out_err_drain_dsaddrs;
	}

	/* version count */
	p = xdr_inline_decode(&stream, 4);
	if (unlikely(!p))
		goto out_err_drain_dsaddrs;
	version_count = be32_to_cpup(p);
	dprintk("%s: version count %d\n", __func__, version_count);

	ds_versions = kzalloc(version_count * sizeof(struct nfs4_ff_ds_version),
			      gfp_flags);
	if (!ds_versions)
		goto out_scratch;

	for (i = 0; i < version_count; i++) {
		/* 20 = version(4) + minor_version(4) + rsize(4) + wsize(4) +
		 * tightly_coupled(4) */
		p = xdr_inline_decode(&stream, 20);
		if (unlikely(!p))
			goto out_err_drain_dsaddrs;
		ds_versions[i].version = be32_to_cpup(p++);
		ds_versions[i].minor_version = be32_to_cpup(p++);
		ds_versions[i].rsize = nfs_block_size(be32_to_cpup(p++), NULL);
		ds_versions[i].wsize = nfs_block_size(be32_to_cpup(p++), NULL);
		ds_versions[i].tightly_coupled = be32_to_cpup(p);

		if (ds_versions[i].rsize > NFS_MAX_FILE_IO_SIZE)
			ds_versions[i].rsize = NFS_MAX_FILE_IO_SIZE;
		if (ds_versions[i].wsize > NFS_MAX_FILE_IO_SIZE)
			ds_versions[i].wsize = NFS_MAX_FILE_IO_SIZE;

		/*
		 * check for valid major/minor combination.
		 * currently we support dataserver which talk:
		 *   v3, v4.0, v4.1, v4.2
		 */
		if (!((ds_versions[i].version == 3 && ds_versions[i].minor_version == 0) ||
			(ds_versions[i].version == 4 && ds_versions[i].minor_version < 3))) {
			dprintk("%s: [%d] unsupported ds version %d-%d\n", __func__,
				i, ds_versions[i].version,
				ds_versions[i].minor_version);
			ret = -EPROTONOSUPPORT;
			goto out_err_drain_dsaddrs;
		}

		dprintk("%s: [%d] vers %u minor_ver %u rsize %u wsize %u coupled %d\n",
			__func__, i, ds_versions[i].version,
			ds_versions[i].minor_version,
			ds_versions[i].rsize,
			ds_versions[i].wsize,
			ds_versions[i].tightly_coupled);
	}

	new_ds->ds_versions = ds_versions;
	new_ds->ds_versions_cnt = version_count;

	new_ds->ds = nfs4_pnfs_ds_add(&dsaddrs, gfp_flags);
	if (!new_ds->ds)
		goto out_err_drain_dsaddrs;

	/* If DS was already in cache, free ds addrs */
	while (!list_empty(&dsaddrs)) {
		da = list_first_entry(&dsaddrs,
				      struct nfs4_pnfs_ds_addr,
				      da_node);
		list_del_init(&da->da_node);
		kfree(da->da_remotestr);
		kfree(da);
	}

	__free_page(scratch);
	return new_ds;

out_err_drain_dsaddrs:
	while (!list_empty(&dsaddrs)) {
		da = list_first_entry(&dsaddrs, struct nfs4_pnfs_ds_addr,
				      da_node);
		list_del_init(&da->da_node);
		kfree(da->da_remotestr);
		kfree(da);
	}

	kfree(ds_versions);
out_scratch:
	__free_page(scratch);
out_err:
	kfree(new_ds);

	dprintk("%s ERROR: returning %d\n", __func__, ret);
	return NULL;
}

static void ff_layout_mark_devid_invalid(struct pnfs_layout_segment *lseg,
		struct nfs4_deviceid_node *devid)
{
	nfs4_delete_deviceid(devid->ld, devid->nfs_client, &devid->deviceid);
	if (!ff_layout_has_available_ds(lseg))
		pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode,
				lseg);
}

static bool ff_layout_mirror_valid(struct pnfs_layout_segment *lseg,
				   struct nfs4_ff_layout_mirror *mirror,
				   bool create)
{
	if (mirror == NULL || IS_ERR(mirror->mirror_ds))
		goto outerr;
	if (mirror->mirror_ds == NULL) {
		if (create) {
			struct nfs4_deviceid_node *node;
			struct pnfs_layout_hdr *lh = lseg->pls_layout;
			struct nfs4_ff_layout_ds *mirror_ds = ERR_PTR(-ENODEV);

			node = nfs4_find_get_deviceid(NFS_SERVER(lh->plh_inode),
					&mirror->devid, lh->plh_lc_cred,
					GFP_KERNEL);
			if (node)
				mirror_ds = FF_LAYOUT_MIRROR_DS(node);

			/* check for race with another call to this function */
			if (cmpxchg(&mirror->mirror_ds, NULL, mirror_ds) &&
			    mirror_ds != ERR_PTR(-ENODEV))
				nfs4_put_deviceid_node(node);
		} else
			goto outerr;
	}

	if (IS_ERR(mirror->mirror_ds))
		goto outerr;

	if (mirror->mirror_ds->ds == NULL) {
		struct nfs4_deviceid_node *devid;
		devid = &mirror->mirror_ds->id_node;
		ff_layout_mark_devid_invalid(lseg, devid);
		return false;
	}
	return true;
outerr:
	pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode, lseg);
	return false;
}

static void extend_ds_error(struct nfs4_ff_layout_ds_err *err,
			    u64 offset, u64 length)
{
	u64 end;

	end = max_t(u64, pnfs_end_offset(err->offset, err->length),
		    pnfs_end_offset(offset, length));
	err->offset = min_t(u64, err->offset, offset);
	err->length = end - err->offset;
}

static int
ff_ds_error_match(const struct nfs4_ff_layout_ds_err *e1,
		const struct nfs4_ff_layout_ds_err *e2)
{
	int ret;

	if (e1->opnum != e2->opnum)
		return e1->opnum < e2->opnum ? -1 : 1;
	if (e1->status != e2->status)
		return e1->status < e2->status ? -1 : 1;
	ret = memcmp(e1->stateid.data, e2->stateid.data,
			sizeof(e1->stateid.data));
	if (ret != 0)
		return ret;
	ret = memcmp(&e1->deviceid, &e2->deviceid, sizeof(e1->deviceid));
	if (ret != 0)
		return ret;
	if (pnfs_end_offset(e1->offset, e1->length) < e2->offset)
		return -1;
	if (e1->offset > pnfs_end_offset(e2->offset, e2->length))
		return 1;
	/* If ranges overlap or are contiguous, they are the same */
	return 0;
}

static void
ff_layout_add_ds_error_locked(struct nfs4_flexfile_layout *flo,
			      struct nfs4_ff_layout_ds_err *dserr)
{
	struct nfs4_ff_layout_ds_err *err, *tmp;
	struct list_head *head = &flo->error_list;
	int match;

	/* Do insertion sort w/ merges */
	list_for_each_entry_safe(err, tmp, &flo->error_list, list) {
		match = ff_ds_error_match(err, dserr);
		if (match < 0)
			continue;
		if (match > 0) {
			/* Add entry "dserr" _before_ entry "err" */
			head = &err->list;
			break;
		}
		/* Entries match, so merge "err" into "dserr" */
		extend_ds_error(dserr, err->offset, err->length);
		list_replace(&err->list, &dserr->list);
		kfree(err);
		return;
	}

	list_add_tail(&dserr->list, head);
}

int ff_layout_track_ds_error(struct nfs4_flexfile_layout *flo,
			     struct nfs4_ff_layout_mirror *mirror, u64 offset,
			     u64 length, int status, enum nfs_opnum4 opnum,
			     gfp_t gfp_flags)
{
	struct nfs4_ff_layout_ds_err *dserr;

	if (status == 0)
		return 0;

	if (mirror->mirror_ds == NULL)
		return -EINVAL;

	dserr = kmalloc(sizeof(*dserr), gfp_flags);
	if (!dserr)
		return -ENOMEM;

	INIT_LIST_HEAD(&dserr->list);
	dserr->offset = offset;
	dserr->length = length;
	dserr->status = status;
	dserr->opnum = opnum;
	nfs4_stateid_copy(&dserr->stateid, &mirror->stateid);
	memcpy(&dserr->deviceid, &mirror->mirror_ds->id_node.deviceid,
	       NFS4_DEVICEID4_SIZE);

	spin_lock(&flo->generic_hdr.plh_inode->i_lock);
	ff_layout_add_ds_error_locked(flo, dserr);
	spin_unlock(&flo->generic_hdr.plh_inode->i_lock);

	return 0;
}

static struct rpc_cred *
ff_layout_get_mirror_cred(struct nfs4_ff_layout_mirror *mirror, u32 iomode)
{
	struct rpc_cred *cred, __rcu **pcred;

	if (iomode == IOMODE_READ)
		pcred = &mirror->ro_cred;
	else
		pcred = &mirror->rw_cred;

	rcu_read_lock();
	do {
		cred = rcu_dereference(*pcred);
		if (!cred)
			break;

		cred = get_rpccred_rcu(cred);
	} while(!cred);
	rcu_read_unlock();
	return cred;
}

struct nfs_fh *
nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx)
{
	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, mirror_idx);
	struct nfs_fh *fh = NULL;

	if (!ff_layout_mirror_valid(lseg, mirror, false)) {
		pr_err_ratelimited("NFS: %s: No data server for mirror offset index %d\n",
			__func__, mirror_idx);
		goto out;
	}

	/* FIXME: For now assume there is only 1 version available for the DS */
	fh = &mirror->fh_versions[0];
out:
	return fh;
}

/**
 * nfs4_ff_layout_prepare_ds - prepare a DS connection for an RPC call
 * @lseg: the layout segment we're operating on
 * @ds_idx: index of the DS to use
 * @fail_return: return layout on connect failure?
 *
 * Try to prepare a DS connection to accept an RPC call. This involves
 * selecting a mirror to use and connecting the client to it if it's not
 * already connected.
 *
 * Since we only need a single functioning mirror to satisfy a read, we don't
 * want to return the layout if there is one. For writes though, any down
 * mirror should result in a LAYOUTRETURN. @fail_return is how we distinguish
 * between the two cases.
 *
 * Returns a pointer to a connected DS object on success or NULL on failure.
 */
struct nfs4_pnfs_ds *
nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
			  bool fail_return)
{
	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx);
	struct nfs4_pnfs_ds *ds = NULL;
	struct nfs4_deviceid_node *devid;
	struct inode *ino = lseg->pls_layout->plh_inode;
	struct nfs_server *s = NFS_SERVER(ino);
	unsigned int max_payload;
	int status;

	if (!ff_layout_mirror_valid(lseg, mirror, true)) {
		pr_err_ratelimited("NFS: %s: No data server for offset index %d\n",
			__func__, ds_idx);
		goto out;
	}

	devid = &mirror->mirror_ds->id_node;
	if (ff_layout_test_devid_unavailable(devid))
		goto out_fail;

	ds = mirror->mirror_ds->ds;
	/* matching smp_wmb() in _nfs4_pnfs_v3/4_ds_connect */
	smp_rmb();
	if (ds->ds_clp)
		goto out;

	/* FIXME: For now we assume the server sent only one version of NFS
	 * to use for the DS.
	 */
	status = nfs4_pnfs_ds_connect(s, ds, devid, dataserver_timeo,
			     dataserver_retrans,
			     mirror->mirror_ds->ds_versions[0].version,
			     mirror->mirror_ds->ds_versions[0].minor_version);

	/* connect success, check rsize/wsize limit */
	if (!status) {
		max_payload =
			nfs_block_size(rpc_max_payload(ds->ds_clp->cl_rpcclient),
				       NULL);
		if (mirror->mirror_ds->ds_versions[0].rsize > max_payload)
			mirror->mirror_ds->ds_versions[0].rsize = max_payload;
		if (mirror->mirror_ds->ds_versions[0].wsize > max_payload)
			mirror->mirror_ds->ds_versions[0].wsize = max_payload;
		goto out;
	}
out_fail:
	ff_layout_track_ds_error(FF_LAYOUT_FROM_HDR(lseg->pls_layout),
				 mirror, lseg->pls_range.offset,
				 lseg->pls_range.length, NFS4ERR_NXIO,
				 OP_ILLEGAL, GFP_NOIO);
	if (fail_return || !ff_layout_has_available_ds(lseg))
		pnfs_error_mark_layout_for_return(ino, lseg);
	ds = NULL;
out:
	return ds;
}

struct rpc_cred *
ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg, u32 ds_idx,
		      struct rpc_cred *mdscred)
{
	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx);
	struct rpc_cred *cred;

	if (mirror) {
		cred = ff_layout_get_mirror_cred(mirror, lseg->pls_range.iomode);
		if (!cred)
			cred = get_rpccred(mdscred);
	} else {
		cred = get_rpccred(mdscred);
	}
	return cred;
}

/**
* Find or create a DS rpc client with th MDS server rpc client auth flavor
* in the nfs_client cl_ds_clients list.
*/
struct rpc_clnt *
nfs4_ff_find_or_create_ds_client(struct pnfs_layout_segment *lseg, u32 ds_idx,
				 struct nfs_client *ds_clp, struct inode *inode)
{
	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx);

	switch (mirror->mirror_ds->ds_versions[0].version) {
	case 3:
		/* For NFSv3 DS, flavor is set when creating DS connections */
		return ds_clp->cl_rpcclient;
	case 4:
		return nfs4_find_or_create_ds_client(ds_clp, inode);
	default:
		BUG();
	}
}

void ff_layout_free_ds_ioerr(struct list_head *head)
{
	struct nfs4_ff_layout_ds_err *err;

	while (!list_empty(head)) {
		err = list_first_entry(head,
				struct nfs4_ff_layout_ds_err,
				list);
		list_del(&err->list);
		kfree(err);
	}
}

/* called with inode i_lock held */
int ff_layout_encode_ds_ioerr(struct xdr_stream *xdr, const struct list_head *head)
{
	struct nfs4_ff_layout_ds_err *err;
	__be32 *p;

	list_for_each_entry(err, head, list) {
		/* offset(8) + length(8) + stateid(NFS4_STATEID_SIZE)
		 * + array length + deviceid(NFS4_DEVICEID4_SIZE)
		 * + status(4) + opnum(4)
		 */
		p = xdr_reserve_space(xdr,
				28 + NFS4_STATEID_SIZE + NFS4_DEVICEID4_SIZE);
		if (unlikely(!p))
			return -ENOBUFS;
		p = xdr_encode_hyper(p, err->offset);
		p = xdr_encode_hyper(p, err->length);
		p = xdr_encode_opaque_fixed(p, &err->stateid,
					    NFS4_STATEID_SIZE);
		/* Encode 1 error */
		*p++ = cpu_to_be32(1);
		p = xdr_encode_opaque_fixed(p, &err->deviceid,
					    NFS4_DEVICEID4_SIZE);
		*p++ = cpu_to_be32(err->status);
		*p++ = cpu_to_be32(err->opnum);
		dprintk("%s: offset %llu length %llu status %d op %d\n",
			__func__, err->offset, err->length, err->status,
			err->opnum);
	}

	return 0;
}

static
unsigned int do_layout_fetch_ds_ioerr(struct pnfs_layout_hdr *lo,
				      const struct pnfs_layout_range *range,
				      struct list_head *head,
				      unsigned int maxnum)
{
	struct nfs4_flexfile_layout *flo = FF_LAYOUT_FROM_HDR(lo);
	struct inode *inode = lo->plh_inode;
	struct nfs4_ff_layout_ds_err *err, *n;
	unsigned int ret = 0;

	spin_lock(&inode->i_lock);
	list_for_each_entry_safe(err, n, &flo->error_list, list) {
		if (!pnfs_is_range_intersecting(err->offset,
				pnfs_end_offset(err->offset, err->length),
				range->offset,
				pnfs_end_offset(range->offset, range->length)))
			continue;
		if (!maxnum)
			break;
		list_move(&err->list, head);
		maxnum--;
		ret++;
	}
	spin_unlock(&inode->i_lock);
	return ret;
}

unsigned int ff_layout_fetch_ds_ioerr(struct pnfs_layout_hdr *lo,
				      const struct pnfs_layout_range *range,
				      struct list_head *head,
				      unsigned int maxnum)
{
	unsigned int ret;

	ret = do_layout_fetch_ds_ioerr(lo, range, head, maxnum);
	/* If we're over the max, discard all remaining entries */
	if (ret == maxnum) {
		LIST_HEAD(discard);
		do_layout_fetch_ds_ioerr(lo, range, &discard, -1);
		ff_layout_free_ds_ioerr(&discard);
	}
	return ret;
}

static bool ff_read_layout_has_available_ds(struct pnfs_layout_segment *lseg)
{
	struct nfs4_ff_layout_mirror *mirror;
	struct nfs4_deviceid_node *devid;
	u32 idx;

	for (idx = 0; idx < FF_LAYOUT_MIRROR_COUNT(lseg); idx++) {
		mirror = FF_LAYOUT_COMP(lseg, idx);
		if (mirror) {
			if (!mirror->mirror_ds)
				return true;
			if (IS_ERR(mirror->mirror_ds))
				continue;
			devid = &mirror->mirror_ds->id_node;
			if (!ff_layout_test_devid_unavailable(devid))
				return true;
		}
	}

	return false;
}

static bool ff_rw_layout_has_available_ds(struct pnfs_layout_segment *lseg)
{
	struct nfs4_ff_layout_mirror *mirror;
	struct nfs4_deviceid_node *devid;
	u32 idx;

	for (idx = 0; idx < FF_LAYOUT_MIRROR_COUNT(lseg); idx++) {
		mirror = FF_LAYOUT_COMP(lseg, idx);
		if (!mirror || IS_ERR(mirror->mirror_ds))
			return false;
		if (!mirror->mirror_ds)
			continue;
		devid = &mirror->mirror_ds->id_node;
		if (ff_layout_test_devid_unavailable(devid))
			return false;
	}

	return FF_LAYOUT_MIRROR_COUNT(lseg) != 0;
}

static bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg)
{
	if (lseg->pls_range.iomode == IOMODE_READ)
		return  ff_read_layout_has_available_ds(lseg);
	/* Note: RW layout needs all mirrors available */
	return ff_rw_layout_has_available_ds(lseg);
}

bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg)
{
	return ff_layout_no_fallback_to_mds(lseg) ||
	       ff_layout_has_available_ds(lseg);
}

bool ff_layout_avoid_read_on_rw(struct pnfs_layout_segment *lseg)
{
	return lseg->pls_range.iomode == IOMODE_RW &&
	       ff_layout_no_read_on_rw(lseg);
}

module_param(dataserver_retrans, uint, 0644);
MODULE_PARM_DESC(dataserver_retrans, "The  number of times the NFSv4.1 client "
			"retries a request before it attempts further "
			" recovery  action.");
module_param(dataserver_timeo, uint, 0644);
MODULE_PARM_DESC(dataserver_timeo, "The time (in tenths of a second) the "
			"NFSv4.1  client  waits for a response from a "
			" data server before it retries an NFS request.");
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
d67ae825 TH	2	/*
	3	* Device operations for the pnfs nfs4 file layout driver.
	4	*
	5	* Copyright (c) 2014, Primary Data, Inc. All rights reserved.
	6	*
	7	* Tao Peng <bergwolf@primarydata.com>
	8	*/
	9
	10	#include <linux/nfs_fs.h>
	11	#include <linux/vmalloc.h>
	12	#include <linux/module.h>
	13	#include <linux/sunrpc/addr.h>
	14
	15	#include "../internal.h"
	16	#include "../nfs4session.h"
	17	#include "flexfilelayout.h"
	18
	19	#define NFSDBG_FACILITY NFSDBG_PNFS_LD
	20
15d03055 TM	21	static unsigned int dataserver_timeo = NFS_DEF_TCP_RETRANS;
15d03055 TM	22	static unsigned int dataserver_retrans;
d67ae825	23
65990d1a FI	24	static bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg);
65990d1a FI	25
d67ae825 TH	26	void nfs4_ff_layout_put_deviceid(struct nfs4_ff_layout_ds *mirror_ds)
d67ae825 TH	27	{
65990d1a	28	if (!IS_ERR_OR_NULL(mirror_ds))
d67ae825 TH	29	nfs4_put_deviceid_node(&mirror_ds->id_node);
	30	}
	31
	32	void nfs4_ff_layout_free_deviceid(struct nfs4_ff_layout_ds *mirror_ds)
	33	{
	34	nfs4_print_deviceid(&mirror_ds->id_node.deviceid);
	35	nfs4_pnfs_ds_put(mirror_ds->ds);
1feb2616	36	kfree(mirror_ds->ds_versions);
84a80f62	37	kfree_rcu(mirror_ds, id_node.rcu);
d67ae825 TH	38	}
	39
	40	/* Decode opaque device data and construct new_ds using it */
	41	struct nfs4_ff_layout_ds *
	42	nfs4_ff_alloc_deviceid_node(struct nfs_server server, struct pnfs_device pdev,
	43	gfp_t gfp_flags)
	44	{
	45	struct xdr_stream stream;
	46	struct xdr_buf buf;
	47	struct page *scratch;
	48	struct list_head dsaddrs;
	49	struct nfs4_pnfs_ds_addr *da;
	50	struct nfs4_ff_layout_ds *new_ds = NULL;
	51	struct nfs4_ff_ds_version *ds_versions = NULL;
	52	u32 mp_count;
	53	u32 version_count;
	54	__be32 *p;
	55	int i, ret = -ENOMEM;
	56
	57	/* set up xdr stream */
	58	scratch = alloc_page(gfp_flags);
	59	if (!scratch)
	60	goto out_err;
	61
	62	new_ds = kzalloc(sizeof(struct nfs4_ff_layout_ds), gfp_flags);
	63	if (!new_ds)
	64	goto out_scratch;
	65
	66	nfs4_init_deviceid_node(&new_ds->id_node,
	67	server,
	68	&pdev->dev_id);
	69	INIT_LIST_HEAD(&dsaddrs);
	70
	71	xdr_init_decode_pages(&stream, &buf, pdev->pages, pdev->pglen);
	72	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
	73
	74	/* multipath count */
	75	p = xdr_inline_decode(&stream, 4);
	76	if (unlikely(!p))
	77	goto out_err_drain_dsaddrs;
	78	mp_count = be32_to_cpup(p);
	79	dprintk("%s: multipath ds count %d\n", __func__, mp_count);
	80
	81	for (i = 0; i < mp_count; i++) {
	82	/* multipath ds */
	83	da = nfs4_decode_mp_ds_addr(server->nfs_client->cl_net,
	84	&stream, gfp_flags);
	85	if (da)
	86	list_add_tail(&da->da_node, &dsaddrs);
	87	}
	88	if (list_empty(&dsaddrs)) {
	89	dprintk("%s: no suitable DS addresses found\n",
	90	__func__);
	91	ret = -ENOMEDIUM;
	92	goto out_err_drain_dsaddrs;
	93	}
	94
	95	/* version count */
	96	p = xdr_inline_decode(&stream, 4);
	97	if (unlikely(!p))
	98	goto out_err_drain_dsaddrs;
	99	version_count = be32_to_cpup(p);
	100	dprintk("%s: version count %d\n", __func__, version_count);
	101
102	ds_versions = kzalloc(version_count * sizeof(struct nfs4_ff_ds_version),
103	gfp_flags);
104	if (!ds_versions)
105	goto out_scratch;
106
107	for (i = 0; i < version_count; i++) {
108	/* 20 = version(4) + minor_version(4) + rsize(4) + wsize(4) +
109	* tightly_coupled(4) */
110	p = xdr_inline_decode(&stream, 20);
111	if (unlikely(!p))
112	goto out_err_drain_dsaddrs;
113	ds_versions[i].version = be32_to_cpup(p++);
114	ds_versions[i].minor_version = be32_to_cpup(p++);
115	ds_versions[i].rsize = nfs_block_size(be32_to_cpup(p++), NULL);
116	ds_versions[i].wsize = nfs_block_size(be32_to_cpup(p++), NULL);
117	ds_versions[i].tightly_coupled = be32_to_cpup(p);
118
119	if (ds_versions[i].rsize > NFS_MAX_FILE_IO_SIZE)
120	ds_versions[i].rsize = NFS_MAX_FILE_IO_SIZE;
121	if (ds_versions[i].wsize > NFS_MAX_FILE_IO_SIZE)
122	ds_versions[i].wsize = NFS_MAX_FILE_IO_SIZE;
123
a7878ca1 TM	124	/*
	125	* check for valid major/minor combination.
	126	* currently we support dataserver which talk:
	127	* v3, v4.0, v4.1, v4.2
	128	*/
	129	if (!((ds_versions[i].version == 3 && ds_versions[i].minor_version == 0) \|\|
	130	(ds_versions[i].version == 4 && ds_versions[i].minor_version < 3))) {
d67ae825 TH	131	dprintk("%s: [%d] unsupported ds version %d-%d\n", __func__,
	132	i, ds_versions[i].version,
	133	ds_versions[i].minor_version);
	134	ret = -EPROTONOSUPPORT;
	135	goto out_err_drain_dsaddrs;
	136	}
	137
	138	dprintk("%s: [%d] vers %u minor_ver %u rsize %u wsize %u coupled %d\n",
	139	__func__, i, ds_versions[i].version,
	140	ds_versions[i].minor_version,
	141	ds_versions[i].rsize,
	142	ds_versions[i].wsize,
	143	ds_versions[i].tightly_coupled);
	144	}
	145
	146	new_ds->ds_versions = ds_versions;
	147	new_ds->ds_versions_cnt = version_count;
	148
	149	new_ds->ds = nfs4_pnfs_ds_add(&dsaddrs, gfp_flags);
	150	if (!new_ds->ds)
	151	goto out_err_drain_dsaddrs;
	152
	153	/* If DS was already in cache, free ds addrs */
	154	while (!list_empty(&dsaddrs)) {
	155	da = list_first_entry(&dsaddrs,
	156	struct nfs4_pnfs_ds_addr,
	157	da_node);
	158	list_del_init(&da->da_node);
	159	kfree(da->da_remotestr);
	160	kfree(da);
	161	}
	162
	163	__free_page(scratch);
	164	return new_ds;
	165
	166	out_err_drain_dsaddrs:
	167	while (!list_empty(&dsaddrs)) {
	168	da = list_first_entry(&dsaddrs, struct nfs4_pnfs_ds_addr,
	169	da_node);
	170	list_del_init(&da->da_node);
	171	kfree(da->da_remotestr);
	172	kfree(da);
	173	}
	174
	175	kfree(ds_versions);
	176	out_scratch:
	177	__free_page(scratch);
	178	out_err:
	179	kfree(new_ds);
	180
	181	dprintk("%s ERROR: returning %d\n", __func__, ret);
	182	return NULL;
	183	}
	184
889d94d4 TM	185	static void ff_layout_mark_devid_invalid(struct pnfs_layout_segment *lseg,
	186	struct nfs4_deviceid_node *devid)
	187	{
1c48cee8	188	nfs4_delete_deviceid(devid->ld, devid->nfs_client, &devid->deviceid);
889d94d4 TM	189	if (!ff_layout_has_available_ds(lseg))
	190	pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode,
	191	lseg);
	192	}
	193
	194	static bool ff_layout_mirror_valid(struct pnfs_layout_segment *lseg,
65990d1a FI	195	struct nfs4_ff_layout_mirror *mirror,
65990d1a FI	196	bool create)
889d94d4	197	{
65990d1a FI	198	if (mirror == NULL \|\| IS_ERR(mirror->mirror_ds))
	199	goto outerr;
	200	if (mirror->mirror_ds == NULL) {
	201	if (create) {
	202	struct nfs4_deviceid_node *node;
	203	struct pnfs_layout_hdr *lh = lseg->pls_layout;
	204	struct nfs4_ff_layout_ds *mirror_ds = ERR_PTR(-ENODEV);
	205
	206	node = nfs4_find_get_deviceid(NFS_SERVER(lh->plh_inode),
	207	&mirror->devid, lh->plh_lc_cred,
	208	GFP_KERNEL);
	209	if (node)
	210	mirror_ds = FF_LAYOUT_MIRROR_DS(node);
	211
	212	/* check for race with another call to this function */
	213	if (cmpxchg(&mirror->mirror_ds, NULL, mirror_ds) &&
	214	mirror_ds != ERR_PTR(-ENODEV))
	215	nfs4_put_deviceid_node(node);
	216	} else
	217	goto outerr;
889d94d4	218	}
f17f8a14 TM	219
	220	if (IS_ERR(mirror->mirror_ds))
	221	goto outerr;
	222
889d94d4 TM	223	if (mirror->mirror_ds->ds == NULL) {
	224	struct nfs4_deviceid_node *devid;
	225	devid = &mirror->mirror_ds->id_node;
	226	ff_layout_mark_devid_invalid(lseg, devid);
	227	return false;
	228	}
	229	return true;
65990d1a FI	230	outerr:
	231	pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode, lseg);
	232	return false;
889d94d4 TM	233	}
889d94d4 TM	234
d67ae825 TH	235	static void extend_ds_error(struct nfs4_ff_layout_ds_err *err,
	236	u64 offset, u64 length)
	237	{
	238	u64 end;
	239
17822b20 TM	240	end = max_t(u64, pnfs_end_offset(err->offset, err->length),
17822b20 TM	241	pnfs_end_offset(offset, length));
d67ae825 TH	242	err->offset = min_t(u64, err->offset, offset);
	243	err->length = end - err->offset;
	244	}
	245
b819ed4b TM	246	static int
	247	ff_ds_error_match(const struct nfs4_ff_layout_ds_err *e1,
	248	const struct nfs4_ff_layout_ds_err *e2)
d67ae825	249	{
b819ed4b TM	250	int ret;
	251
	252	if (e1->opnum != e2->opnum)
	253	return e1->opnum < e2->opnum ? -1 : 1;
	254	if (e1->status != e2->status)
	255	return e1->status < e2->status ? -1 : 1;
93b717fd TM	256	ret = memcmp(e1->stateid.data, e2->stateid.data,
93b717fd TM	257	sizeof(e1->stateid.data));
b819ed4b TM	258	if (ret != 0)
	259	return ret;
	260	ret = memcmp(&e1->deviceid, &e2->deviceid, sizeof(e1->deviceid));
	261	if (ret != 0)
	262	return ret;
17822b20	263	if (pnfs_end_offset(e1->offset, e1->length) < e2->offset)
b819ed4b	264	return -1;
17822b20	265	if (e1->offset > pnfs_end_offset(e2->offset, e2->length))
b819ed4b TM	266	return 1;
	267	/* If ranges overlap or are contiguous, they are the same */
	268	return 0;
d67ae825 TH	269	}
d67ae825 TH	270
b819ed4b	271	static void
d67ae825 TH	272	ff_layout_add_ds_error_locked(struct nfs4_flexfile_layout *flo,
	273	struct nfs4_ff_layout_ds_err *dserr)
	274	{
b819ed4b TM	275	struct nfs4_ff_layout_ds_err err, tmp;
	276	struct list_head *head = &flo->error_list;
	277	int match;
	278
	279	/* Do insertion sort w/ merges */
	280	list_for_each_entry_safe(err, tmp, &flo->error_list, list) {
	281	match = ff_ds_error_match(err, dserr);
	282	if (match < 0)
	283	continue;
	284	if (match > 0) {
	285	/* Add entry "dserr" _before_ entry "err" */
	286	head = &err->list;
d67ae825 TH	287	break;
d67ae825 TH	288	}
b819ed4b TM	289	/* Entries match, so merge "err" into "dserr" */
b819ed4b TM	290	extend_ds_error(dserr, err->offset, err->length);
cb067935	291	list_replace(&err->list, &dserr->list);
b819ed4b	292	kfree(err);
cb067935	293	return;
d67ae825 TH	294	}
d67ae825 TH	295
b819ed4b	296	list_add_tail(&dserr->list, head);
d67ae825 TH	297	}
	298
	299	int ff_layout_track_ds_error(struct nfs4_flexfile_layout *flo,
	300	struct nfs4_ff_layout_mirror *mirror, u64 offset,
	301	u64 length, int status, enum nfs_opnum4 opnum,
	302	gfp_t gfp_flags)
	303	{
	304	struct nfs4_ff_layout_ds_err *dserr;
d67ae825 TH	305
	306	if (status == 0)
	307	return 0;
	308
	309	if (mirror->mirror_ds == NULL)
	310	return -EINVAL;
	311
d67ae825 TH	312	dserr = kmalloc(sizeof(*dserr), gfp_flags);
	313	if (!dserr)
	314	return -ENOMEM;
	315
	316	INIT_LIST_HEAD(&dserr->list);
	317	dserr->offset = offset;
	318	dserr->length = length;
	319	dserr->status = status;
	320	dserr->opnum = opnum;
	321	nfs4_stateid_copy(&dserr->stateid, &mirror->stateid);
	322	memcpy(&dserr->deviceid, &mirror->mirror_ds->id_node.deviceid,
	323	NFS4_DEVICEID4_SIZE);
	324
	325	spin_lock(&flo->generic_hdr.plh_inode->i_lock);
b819ed4b	326	ff_layout_add_ds_error_locked(flo, dserr);
d67ae825	327	spin_unlock(&flo->generic_hdr.plh_inode->i_lock);
d67ae825 TH	328
	329	return 0;
	330	}
	331
57f3f4c0 JL	332	static struct rpc_cred *
	333	ff_layout_get_mirror_cred(struct nfs4_ff_layout_mirror *mirror, u32 iomode)
	334	{
3064b686	335	struct rpc_cred cred, __rcu *pcred;
57f3f4c0	336
3064b686 JL	337	if (iomode == IOMODE_READ)
	338	pcred = &mirror->ro_cred;
	339	else
	340	pcred = &mirror->rw_cred;
57f3f4c0 JL	341
	342	rcu_read_lock();
	343	do {
	344	cred = rcu_dereference(*pcred);
	345	if (!cred)
	346	break;
	347
	348	cred = get_rpccred_rcu(cred);
	349	} while(!cred);
	350	rcu_read_unlock();
	351	return cred;
	352	}
	353
d67ae825 TH	354	struct nfs_fh *
	355	nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx)
	356	{
	357	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, mirror_idx);
	358	struct nfs_fh *fh = NULL;
d67ae825	359
65990d1a	360	if (!ff_layout_mirror_valid(lseg, mirror, false)) {
889d94d4	361	pr_err_ratelimited("NFS: %s: No data server for mirror offset index %d\n",
d67ae825	362	__func__, mirror_idx);
d67ae825 TH	363	goto out;
	364	}
	365
	366	/* FIXME: For now assume there is only 1 version available for the DS */
	367	fh = &mirror->fh_versions[0];
	368	out:
	369	return fh;
	370	}
	371
95e2b7e9 JL	372	/**
	373	* nfs4_ff_layout_prepare_ds - prepare a DS connection for an RPC call
	374	* @lseg: the layout segment we're operating on
	375	* @ds_idx: index of the DS to use
	376	* @fail_return: return layout on connect failure?
	377	*
	378	* Try to prepare a DS connection to accept an RPC call. This involves
	379	* selecting a mirror to use and connecting the client to it if it's not
	380	* already connected.
	381	*
	382	* Since we only need a single functioning mirror to satisfy a read, we don't
	383	* want to return the layout if there is one. For writes though, any down
	384	* mirror should result in a LAYOUTRETURN. @fail_return is how we distinguish
	385	* between the two cases.
	386	*
	387	* Returns a pointer to a connected DS object on success or NULL on failure.
	388	*/
d67ae825 TH	389	struct nfs4_pnfs_ds *
	390	nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
	391	bool fail_return)
	392	{
	393	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx);
	394	struct nfs4_pnfs_ds *ds = NULL;
	395	struct nfs4_deviceid_node *devid;
	396	struct inode *ino = lseg->pls_layout->plh_inode;
	397	struct nfs_server *s = NFS_SERVER(ino);
	398	unsigned int max_payload;
a33e4b03	399	int status;
d67ae825	400
65990d1a	401	if (!ff_layout_mirror_valid(lseg, mirror, true)) {
889d94d4	402	pr_err_ratelimited("NFS: %s: No data server for offset index %d\n",
d67ae825	403	__func__, ds_idx);
d67ae825 TH	404	goto out;
	405	}
	406
	407	devid = &mirror->mirror_ds->id_node;
	408	if (ff_layout_test_devid_unavailable(devid))
3dc14735	409	goto out_fail;
d67ae825 TH	410
	411	ds = mirror->mirror_ds->ds;
	412	/* matching smp_wmb() in _nfs4_pnfs_v3/4_ds_connect */
	413	smp_rmb();
	414	if (ds->ds_clp)
90a0be00	415	goto out;
d67ae825 TH	416
	417	/* FIXME: For now we assume the server sent only one version of NFS
	418	* to use for the DS.
	419	*/
a33e4b03	420	status = nfs4_pnfs_ds_connect(s, ds, devid, dataserver_timeo,
d67ae825 TH	421	dataserver_retrans,
d67ae825 TH	422	mirror->mirror_ds->ds_versions[0].version,
7d38de3f	423	mirror->mirror_ds->ds_versions[0].minor_version);
d67ae825 TH	424
d67ae825 TH	425	/* connect success, check rsize/wsize limit */
260f32ad	426	if (!status) {
d67ae825 TH	427	max_payload =
	428	nfs_block_size(rpc_max_payload(ds->ds_clp->cl_rpcclient),
	429	NULL);
	430	if (mirror->mirror_ds->ds_versions[0].rsize > max_payload)
	431	mirror->mirror_ds->ds_versions[0].rsize = max_payload;
	432	if (mirror->mirror_ds->ds_versions[0].wsize > max_payload)
	433	mirror->mirror_ds->ds_versions[0].wsize = max_payload;
3dc14735	434	goto out;
d67ae825	435	}
da066f3f	436	out_fail:
3dc14735 TM	437	ff_layout_track_ds_error(FF_LAYOUT_FROM_HDR(lseg->pls_layout),
	438	mirror, lseg->pls_range.offset,
	439	lseg->pls_range.length, NFS4ERR_NXIO,
	440	OP_ILLEGAL, GFP_NOIO);
3dc14735 TM	441	if (fail_return \|\| !ff_layout_has_available_ds(lseg))
	442	pnfs_error_mark_layout_for_return(ino, lseg);
	443	ds = NULL;
d67ae825 TH	444	out:
	445	return ds;
	446	}
	447
	448	struct rpc_cred *
	449	ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg, u32 ds_idx,
	450	struct rpc_cred *mdscred)
	451	{
	452	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx);
547a6376	453	struct rpc_cred *cred;
d67ae825	454
57f3f4c0 JL	455	if (mirror) {
	456	cred = ff_layout_get_mirror_cred(mirror, lseg->pls_range.iomode);
	457	if (!cred)
	458	cred = get_rpccred(mdscred);
	459	} else {
	460	cred = get_rpccred(mdscred);
	461	}
d67ae825 TH	462	return cred;
	463	}
	464
	465	/**
	466	* Find or create a DS rpc client with th MDS server rpc client auth flavor
	467	* in the nfs_client cl_ds_clients list.
	468	*/
	469	struct rpc_clnt *
	470	nfs4_ff_find_or_create_ds_client(struct pnfs_layout_segment *lseg, u32 ds_idx,
	471	struct nfs_client ds_clp, struct inode inode)
	472	{
	473	struct nfs4_ff_layout_mirror *mirror = FF_LAYOUT_COMP(lseg, ds_idx);
	474
	475	switch (mirror->mirror_ds->ds_versions[0].version) {
	476	case 3:
	477	/* For NFSv3 DS, flavor is set when creating DS connections */
	478	return ds_clp->cl_rpcclient;
	479	case 4:
	480	return nfs4_find_or_create_ds_client(ds_clp, inode);
	481	default:
	482	BUG();
	483	}
	484	}
	485
5b9b3c85 TM	486	void ff_layout_free_ds_ioerr(struct list_head *head)
	487	{
	488	struct nfs4_ff_layout_ds_err *err;
	489
	490	while (!list_empty(head)) {
	491	err = list_first_entry(head,
	492	struct nfs4_ff_layout_ds_err,
	493	list);
	494	list_del(&err->list);
	495	kfree(err);
	496	}
	497	}
	498
d67ae825	499	/* called with inode i_lock held */
5b9b3c85	500	int ff_layout_encode_ds_ioerr(struct xdr_stream xdr, const struct list_head head)
d67ae825	501	{
5b9b3c85	502	struct nfs4_ff_layout_ds_err *err;
d67ae825 TH	503	__be32 *p;
d67ae825 TH	504
5b9b3c85	505	list_for_each_entry(err, head, list) {
d67ae825	506	/* offset(8) + length(8) + stateid(NFS4_STATEID_SIZE)
d1354907 TM	507	* + array length + deviceid(NFS4_DEVICEID4_SIZE)
d1354907 TM	508	* + status(4) + opnum(4)
d67ae825 TH	509	*/
d67ae825 TH	510	p = xdr_reserve_space(xdr,
d1354907	511	28 + NFS4_STATEID_SIZE + NFS4_DEVICEID4_SIZE);
d67ae825 TH	512	if (unlikely(!p))
	513	return -ENOBUFS;
	514	p = xdr_encode_hyper(p, err->offset);
	515	p = xdr_encode_hyper(p, err->length);
	516	p = xdr_encode_opaque_fixed(p, &err->stateid,
	517	NFS4_STATEID_SIZE);
d1354907 TM	518	/* Encode 1 error */
d1354907 TM	519	*p++ = cpu_to_be32(1);
d67ae825 TH	520	p = xdr_encode_opaque_fixed(p, &err->deviceid,
	521	NFS4_DEVICEID4_SIZE);
	522	*p++ = cpu_to_be32(err->status);
	523	*p++ = cpu_to_be32(err->opnum);
5b9b3c85	524	dprintk("%s: offset %llu length %llu status %d op %d\n",
d67ae825	525	__func__, err->offset, err->length, err->status,
5b9b3c85	526	err->opnum);
d67ae825 TH	527	}
	528
	529	return 0;
	530	}
	531
5b9b3c85 TM	532	static
	533	unsigned int do_layout_fetch_ds_ioerr(struct pnfs_layout_hdr *lo,
	534	const struct pnfs_layout_range *range,
	535	struct list_head *head,
	536	unsigned int maxnum)
	537	{
	538	struct nfs4_flexfile_layout *flo = FF_LAYOUT_FROM_HDR(lo);
	539	struct inode *inode = lo->plh_inode;
	540	struct nfs4_ff_layout_ds_err err, n;
	541	unsigned int ret = 0;
	542
	543	spin_lock(&inode->i_lock);
	544	list_for_each_entry_safe(err, n, &flo->error_list, list) {
	545	if (!pnfs_is_range_intersecting(err->offset,
	546	pnfs_end_offset(err->offset, err->length),
	547	range->offset,
	548	pnfs_end_offset(range->offset, range->length)))
	549	continue;
	550	if (!maxnum)
	551	break;
	552	list_move(&err->list, head);
	553	maxnum--;
	554	ret++;
	555	}
	556	spin_unlock(&inode->i_lock);
	557	return ret;
	558	}
	559
	560	unsigned int ff_layout_fetch_ds_ioerr(struct pnfs_layout_hdr *lo,
	561	const struct pnfs_layout_range *range,
	562	struct list_head *head,
	563	unsigned int maxnum)
	564	{
	565	unsigned int ret;
	566
	567	ret = do_layout_fetch_ds_ioerr(lo, range, head, maxnum);
	568	/* If we're over the max, discard all remaining entries */
	569	if (ret == maxnum) {
	570	LIST_HEAD(discard);
	571	do_layout_fetch_ds_ioerr(lo, range, &discard, -1);
	572	ff_layout_free_ds_ioerr(&discard);
	573	}
	574	return ret;
	575	}
	576
81d6dc8b	577	static bool ff_read_layout_has_available_ds(struct pnfs_layout_segment *lseg)
d67ae825 TH	578	{
	579	struct nfs4_ff_layout_mirror *mirror;
	580	struct nfs4_deviceid_node *devid;
81d6dc8b	581	u32 idx;
d67ae825 TH	582
	583	for (idx = 0; idx < FF_LAYOUT_MIRROR_COUNT(lseg); idx++) {
	584	mirror = FF_LAYOUT_COMP(lseg, idx);
65990d1a FI	585	if (mirror) {
	586	if (!mirror->mirror_ds)
	587	return true;
	588	if (IS_ERR(mirror->mirror_ds))
	589	continue;
d67ae825 TH	590	devid = &mirror->mirror_ds->id_node;
	591	if (!ff_layout_test_devid_unavailable(devid))
	592	return true;
	593	}
	594	}
	595
	596	return false;
	597	}
	598
81d6dc8b TM	599	static bool ff_rw_layout_has_available_ds(struct pnfs_layout_segment *lseg)
	600	{
	601	struct nfs4_ff_layout_mirror *mirror;
	602	struct nfs4_deviceid_node *devid;
	603	u32 idx;
	604
	605	for (idx = 0; idx < FF_LAYOUT_MIRROR_COUNT(lseg); idx++) {
	606	mirror = FF_LAYOUT_COMP(lseg, idx);
65990d1a	607	if (!mirror \|\| IS_ERR(mirror->mirror_ds))
81d6dc8b	608	return false;
65990d1a FI	609	if (!mirror->mirror_ds)
65990d1a FI	610	continue;
81d6dc8b TM	611	devid = &mirror->mirror_ds->id_node;
	612	if (ff_layout_test_devid_unavailable(devid))
	613	return false;
	614	}
	615
	616	return FF_LAYOUT_MIRROR_COUNT(lseg) != 0;
	617	}
	618
65990d1a	619	static bool ff_layout_has_available_ds(struct pnfs_layout_segment *lseg)
81d6dc8b TM	620	{
	621	if (lseg->pls_range.iomode == IOMODE_READ)
	622	return ff_read_layout_has_available_ds(lseg);
	623	/* Note: RW layout needs all mirrors available */
	624	return ff_rw_layout_has_available_ds(lseg);
	625	}
	626
3b13b4b3 TH	627	bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg)
	628	{
	629	return ff_layout_no_fallback_to_mds(lseg) \|\|
	630	ff_layout_has_available_ds(lseg);
	631	}
	632
fb1084e3 TH	633	bool ff_layout_avoid_read_on_rw(struct pnfs_layout_segment *lseg)
	634	{
	635	return lseg->pls_range.iomode == IOMODE_RW &&
	636	ff_layout_no_read_on_rw(lseg);
	637	}
	638
d67ae825 TH	639	module_param(dataserver_retrans, uint, 0644);
	640	MODULE_PARM_DESC(dataserver_retrans, "The number of times the NFSv4.1 client "
	641	"retries a request before it attempts further "
	642	" recovery action.");
	643	module_param(dataserver_timeo, uint, 0644);
	644	MODULE_PARM_DESC(dataserver_timeo, "The time (in tenths of a second) the "
	645	"NFSv4.1 client waits for a response from a "
	646	" data server before it retries an NFS request.");