[mirror_ubuntu-bionic-kernel.git] / fs / nfs / flexfilelayout / flexfilelayout.h

/* SPDX-License-Identifier: GPL-2.0 */
/*
 * NFSv4 flexfile layout driver data structures.
 *
 * Copyright (c) 2014, Primary Data, Inc. All rights reserved.
 *
 * Tao Peng <bergwolf@primarydata.com>
 */

#ifndef FS_NFS_NFS4FLEXFILELAYOUT_H
#define FS_NFS_NFS4FLEXFILELAYOUT_H

#define FF_FLAGS_NO_LAYOUTCOMMIT 1
#define FF_FLAGS_NO_IO_THRU_MDS  2
#define FF_FLAGS_NO_READ_IO      4

#include "../pnfs.h"

/* XXX: Let's filter out insanely large mirror count for now to avoid oom
 * due to network error etc. */
#define NFS4_FLEXFILE_LAYOUT_MAX_MIRROR_CNT 4096

/* LAYOUTSTATS report interval in ms */
#define FF_LAYOUTSTATS_REPORT_INTERVAL (60000L)
#define FF_LAYOUTSTATS_MAXDEV 4

struct nfs4_ff_ds_version {
	u32				version;
	u32				minor_version;
	u32				rsize;
	u32				wsize;
	bool				tightly_coupled;
};

/* chained in global deviceid hlist */
struct nfs4_ff_layout_ds {
	struct nfs4_deviceid_node	id_node;
	u32				ds_versions_cnt;
	struct nfs4_ff_ds_version	*ds_versions;
	struct nfs4_pnfs_ds		*ds;
};

struct nfs4_ff_layout_ds_err {
	struct list_head		list; /* linked in mirror error_list */
	u64				offset;
	u64				length;
	int				status;
	enum nfs_opnum4			opnum;
	nfs4_stateid			stateid;
	struct nfs4_deviceid		deviceid;
};

struct nfs4_ff_io_stat {
	__u64				ops_requested;
	__u64				bytes_requested;
	__u64				ops_completed;
	__u64				bytes_completed;
	__u64				bytes_not_delivered;
	ktime_t				total_busy_time;
	ktime_t				aggregate_completion_time;
};

struct nfs4_ff_busy_timer {
	ktime_t start_time;
	atomic_t n_ops;
};

struct nfs4_ff_layoutstat {
	struct nfs4_ff_io_stat io_stat;
	struct nfs4_ff_busy_timer busy_timer;
};

struct nfs4_ff_layout_mirror {
	struct pnfs_layout_hdr		*layout;
	struct list_head		mirrors;
	u32				ds_count;
	u32				efficiency;
	struct nfs4_deviceid		devid;
	struct nfs4_ff_layout_ds	*mirror_ds;
	u32				fh_versions_cnt;
	struct nfs_fh			*fh_versions;
	nfs4_stateid			stateid;
	struct rpc_cred	__rcu		*ro_cred;
	struct rpc_cred	__rcu		*rw_cred;
	atomic_t			ref;
	spinlock_t			lock;
	unsigned long			flags;
	struct nfs4_ff_layoutstat	read_stat;
	struct nfs4_ff_layoutstat	write_stat;
	ktime_t				start_time;
	u32				report_interval;
};

#define NFS4_FF_MIRROR_STAT_AVAIL	(0)

struct nfs4_ff_layout_segment {
	struct pnfs_layout_segment	generic_hdr;
	u64				stripe_unit;
	u32				flags;
	u32				mirror_array_cnt;
	struct nfs4_ff_layout_mirror	**mirror_array;
};

struct nfs4_flexfile_layout {
	struct pnfs_layout_hdr generic_hdr;
	struct pnfs_ds_commit_info commit_info;
	struct list_head	mirrors;
	struct list_head	error_list; /* nfs4_ff_layout_ds_err */
	ktime_t			last_report_time; /* Layoutstat report times */
};

struct nfs4_flexfile_layoutreturn_args {
	struct list_head errors;
	struct nfs42_layoutstat_devinfo devinfo[FF_LAYOUTSTATS_MAXDEV];
	unsigned int num_errors;
	unsigned int num_dev;
	struct page *pages[1];
};

static inline struct nfs4_flexfile_layout *
FF_LAYOUT_FROM_HDR(struct pnfs_layout_hdr *lo)
{
	return container_of(lo, struct nfs4_flexfile_layout, generic_hdr);
}

static inline struct nfs4_ff_layout_segment *
FF_LAYOUT_LSEG(struct pnfs_layout_segment *lseg)
{
	return container_of(lseg,
			    struct nfs4_ff_layout_segment,
			    generic_hdr);
}

static inline struct nfs4_deviceid_node *
FF_LAYOUT_DEVID_NODE(struct pnfs_layout_segment *lseg, u32 idx)
{
	if (idx >= FF_LAYOUT_LSEG(lseg)->mirror_array_cnt ||
	    FF_LAYOUT_LSEG(lseg)->mirror_array[idx] == NULL ||
	    FF_LAYOUT_LSEG(lseg)->mirror_array[idx]->mirror_ds == NULL)
		return NULL;
	return &FF_LAYOUT_LSEG(lseg)->mirror_array[idx]->mirror_ds->id_node;
}

static inline struct nfs4_ff_layout_ds *
FF_LAYOUT_MIRROR_DS(struct nfs4_deviceid_node *node)
{
	return container_of(node, struct nfs4_ff_layout_ds, id_node);
}

static inline struct nfs4_ff_layout_mirror *
FF_LAYOUT_COMP(struct pnfs_layout_segment *lseg, u32 idx)
{
	if (idx >= FF_LAYOUT_LSEG(lseg)->mirror_array_cnt)
		return NULL;
	return FF_LAYOUT_LSEG(lseg)->mirror_array[idx];
}

static inline u32
FF_LAYOUT_MIRROR_COUNT(struct pnfs_layout_segment *lseg)
{
	return FF_LAYOUT_LSEG(lseg)->mirror_array_cnt;
}

static inline bool
ff_layout_no_fallback_to_mds(struct pnfs_layout_segment *lseg)
{
	return FF_LAYOUT_LSEG(lseg)->flags & FF_FLAGS_NO_IO_THRU_MDS;
}

static inline bool
ff_layout_no_read_on_rw(struct pnfs_layout_segment *lseg)
{
	return FF_LAYOUT_LSEG(lseg)->flags & FF_FLAGS_NO_READ_IO;
}

static inline bool
ff_layout_test_devid_unavailable(struct nfs4_deviceid_node *node)
{
	/*
	 * Flexfiles should never mark a DS unavailable, but if it does
	 * print a (ratelimited) warning as this can affect performance.
	 */
	if (nfs4_test_deviceid_unavailable(node)) {
		u32 *p = (u32 *)node->deviceid.data;

		pr_warn_ratelimited("NFS: flexfiles layout referencing an "
				"unavailable device [%x%x%x%x]\n",
				p[0], p[1], p[2], p[3]);
		return true;
	}
	return false;
}

static inline int
nfs4_ff_layout_ds_version(struct pnfs_layout_segment *lseg, u32 ds_idx)
{
	return FF_LAYOUT_COMP(lseg, ds_idx)->mirror_ds->ds_versions[0].version;
}

struct nfs4_ff_layout_ds *
nfs4_ff_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
			    gfp_t gfp_flags);
void nfs4_ff_layout_put_deviceid(struct nfs4_ff_layout_ds *mirror_ds);
void nfs4_ff_layout_free_deviceid(struct nfs4_ff_layout_ds *mirror_ds);
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);
int ff_layout_encode_ds_ioerr(struct xdr_stream *xdr, const struct list_head *head);
void ff_layout_free_ds_ioerr(struct list_head *head);
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);
struct nfs_fh *
nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx);

struct nfs4_pnfs_ds *
nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
			  bool fail_return);

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 rpc_cred *ff_layout_get_ds_cred(struct pnfs_layout_segment *lseg,
				       u32 ds_idx, struct rpc_cred *mdscred);
bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg);
bool ff_layout_avoid_read_on_rw(struct pnfs_layout_segment *lseg);

#endif /* FS_NFS_NFS4FLEXFILELAYOUT_H */
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
d67ae825 TH	2	/*
	3	* NFSv4 flexfile layout driver data structures.
	4	*
	5	* Copyright (c) 2014, Primary Data, Inc. All rights reserved.
	6	*
	7	* Tao Peng <bergwolf@primarydata.com>
	8	*/
	9
	10	#ifndef FS_NFS_NFS4FLEXFILELAYOUT_H
	11	#define FS_NFS_NFS4FLEXFILELAYOUT_H
	12
c0f5f505	13	#define FF_FLAGS_NO_LAYOUTCOMMIT 1
fb1084e3 TH	14	#define FF_FLAGS_NO_IO_THRU_MDS 2
fb1084e3 TH	15	#define FF_FLAGS_NO_READ_IO 4
c0f5f505	16
d67ae825 TH	17	#include "../pnfs.h"
	18
	19	/* XXX: Let's filter out insanely large mirror count for now to avoid oom
	20	* due to network error etc. */
	21	#define NFS4_FLEXFILE_LAYOUT_MAX_MIRROR_CNT 4096
	22
97ba375b PT	23	/* LAYOUTSTATS report interval in ms */
97ba375b PT	24	#define FF_LAYOUTSTATS_REPORT_INTERVAL (60000L)
230bc962	25	#define FF_LAYOUTSTATS_MAXDEV 4
97ba375b	26
d67ae825 TH	27	struct nfs4_ff_ds_version {
	28	u32 version;
	29	u32 minor_version;
	30	u32 rsize;
	31	u32 wsize;
	32	bool tightly_coupled;
	33	};
	34
	35	/* chained in global deviceid hlist */
	36	struct nfs4_ff_layout_ds {
	37	struct nfs4_deviceid_node id_node;
	38	u32 ds_versions_cnt;
	39	struct nfs4_ff_ds_version *ds_versions;
	40	struct nfs4_pnfs_ds *ds;
	41	};
	42
	43	struct nfs4_ff_layout_ds_err {
	44	struct list_head list; /* linked in mirror error_list */
	45	u64 offset;
	46	u64 length;
	47	int status;
	48	enum nfs_opnum4 opnum;
	49	nfs4_stateid stateid;
	50	struct nfs4_deviceid deviceid;
	51	};
	52
abcb7bfc TM	53	struct nfs4_ff_io_stat {
	54	__u64 ops_requested;
	55	__u64 bytes_requested;
	56	__u64 ops_completed;
	57	__u64 bytes_completed;
	58	__u64 bytes_not_delivered;
	59	ktime_t total_busy_time;
	60	ktime_t aggregate_completion_time;
	61	};
	62
	63	struct nfs4_ff_busy_timer {
	64	ktime_t start_time;
	65	atomic_t n_ops;
	66	};
	67
	68	struct nfs4_ff_layoutstat {
	69	struct nfs4_ff_io_stat io_stat;
	70	struct nfs4_ff_busy_timer busy_timer;
	71	};
	72
d67ae825	73	struct nfs4_ff_layout_mirror {
266d12d4 TM	74	struct pnfs_layout_hdr *layout;
266d12d4 TM	75	struct list_head mirrors;
d67ae825 TH	76	u32 ds_count;
d67ae825 TH	77	u32 efficiency;
65990d1a	78	struct nfs4_deviceid devid;
d67ae825 TH	79	struct nfs4_ff_layout_ds *mirror_ds;
	80	u32 fh_versions_cnt;
	81	struct nfs_fh *fh_versions;
	82	nfs4_stateid stateid;
3064b686 JL	83	struct rpc_cred __rcu *ro_cred;
3064b686 JL	84	struct rpc_cred __rcu *rw_cred;
28a0d72c	85	atomic_t ref;
d67ae825	86	spinlock_t lock;
06946c6a	87	unsigned long flags;
abcb7bfc TM	88	struct nfs4_ff_layoutstat read_stat;
abcb7bfc TM	89	struct nfs4_ff_layoutstat write_stat;
d983803d	90	ktime_t start_time;
d0379a5d	91	u32 report_interval;
d67ae825 TH	92	};
d67ae825 TH	93
06946c6a TM	94	#define NFS4_FF_MIRROR_STAT_AVAIL (0)
06946c6a TM	95
d67ae825 TH	96	struct nfs4_ff_layout_segment {
	97	struct pnfs_layout_segment generic_hdr;
	98	u64 stripe_unit;
c0f5f505	99	u32 flags;
d67ae825 TH	100	u32 mirror_array_cnt;
	101	struct nfs4_ff_layout_mirror **mirror_array;
	102	};
	103
	104	struct nfs4_flexfile_layout {
	105	struct pnfs_layout_hdr generic_hdr;
	106	struct pnfs_ds_commit_info commit_info;
266d12d4	107	struct list_head mirrors;
d67ae825	108	struct list_head error_list; /* nfs4_ff_layout_ds_err */
1c8d477a	109	ktime_t last_report_time; /* Layoutstat report times */
d67ae825 TH	110	};
d67ae825 TH	111
5b9b3c85 TM	112	struct nfs4_flexfile_layoutreturn_args {
5b9b3c85 TM	113	struct list_head errors;
230bc962	114	struct nfs42_layoutstat_devinfo devinfo[FF_LAYOUTSTATS_MAXDEV];
5b9b3c85	115	unsigned int num_errors;
230bc962	116	unsigned int num_dev;
d9152114	117	struct page *pages[1];
5b9b3c85 TM	118	};
5b9b3c85 TM	119
d67ae825 TH	120	static inline struct nfs4_flexfile_layout *
	121	FF_LAYOUT_FROM_HDR(struct pnfs_layout_hdr *lo)
	122	{
	123	return container_of(lo, struct nfs4_flexfile_layout, generic_hdr);
	124	}
	125
	126	static inline struct nfs4_ff_layout_segment *
	127	FF_LAYOUT_LSEG(struct pnfs_layout_segment *lseg)
	128	{
	129	return container_of(lseg,
	130	struct nfs4_ff_layout_segment,
	131	generic_hdr);
	132	}
	133
	134	static inline struct nfs4_deviceid_node *
	135	FF_LAYOUT_DEVID_NODE(struct pnfs_layout_segment *lseg, u32 idx)
	136	{
	137	if (idx >= FF_LAYOUT_LSEG(lseg)->mirror_array_cnt \|\|
	138	FF_LAYOUT_LSEG(lseg)->mirror_array[idx] == NULL \|\|
	139	FF_LAYOUT_LSEG(lseg)->mirror_array[idx]->mirror_ds == NULL)
	140	return NULL;
	141	return &FF_LAYOUT_LSEG(lseg)->mirror_array[idx]->mirror_ds->id_node;
	142	}
	143
	144	static inline struct nfs4_ff_layout_ds *
	145	FF_LAYOUT_MIRROR_DS(struct nfs4_deviceid_node *node)
	146	{
	147	return container_of(node, struct nfs4_ff_layout_ds, id_node);
	148	}
	149
	150	static inline struct nfs4_ff_layout_mirror *
	151	FF_LAYOUT_COMP(struct pnfs_layout_segment *lseg, u32 idx)
	152	{
	153	if (idx >= FF_LAYOUT_LSEG(lseg)->mirror_array_cnt)
	154	return NULL;
	155	return FF_LAYOUT_LSEG(lseg)->mirror_array[idx];
	156	}
	157
	158	static inline u32
	159	FF_LAYOUT_MIRROR_COUNT(struct pnfs_layout_segment *lseg)
	160	{
	161	return FF_LAYOUT_LSEG(lseg)->mirror_array_cnt;
	162	}
	163
260074cd TM	164	static inline bool
	165	ff_layout_no_fallback_to_mds(struct pnfs_layout_segment *lseg)
	166	{
	167	return FF_LAYOUT_LSEG(lseg)->flags & FF_FLAGS_NO_IO_THRU_MDS;
	168	}
	169
fb1084e3 TH	170	static inline bool
	171	ff_layout_no_read_on_rw(struct pnfs_layout_segment *lseg)
	172	{
	173	return FF_LAYOUT_LSEG(lseg)->flags & FF_FLAGS_NO_READ_IO;
	174	}
	175
d67ae825 TH	176	static inline bool
	177	ff_layout_test_devid_unavailable(struct nfs4_deviceid_node *node)
	178	{
da066f3f WAA	179	/*
	180	* Flexfiles should never mark a DS unavailable, but if it does
	181	* print a (ratelimited) warning as this can affect performance.
	182	*/
	183	if (nfs4_test_deviceid_unavailable(node)) {
	184	u32 p = (u32 )node->deviceid.data;
	185
	186	pr_warn_ratelimited("NFS: flexfiles layout referencing an "
	187	"unavailable device [%x%x%x%x]\n",
	188	p[0], p[1], p[2], p[3]);
	189	return true;
	190	}
	191	return false;
d67ae825 TH	192	}
	193
	194	static inline int
	195	nfs4_ff_layout_ds_version(struct pnfs_layout_segment *lseg, u32 ds_idx)
	196	{
	197	return FF_LAYOUT_COMP(lseg, ds_idx)->mirror_ds->ds_versions[0].version;
	198	}
	199
	200	struct nfs4_ff_layout_ds *
	201	nfs4_ff_alloc_deviceid_node(struct nfs_server server, struct pnfs_device pdev,
	202	gfp_t gfp_flags);
	203	void nfs4_ff_layout_put_deviceid(struct nfs4_ff_layout_ds *mirror_ds);
	204	void nfs4_ff_layout_free_deviceid(struct nfs4_ff_layout_ds *mirror_ds);
	205	int ff_layout_track_ds_error(struct nfs4_flexfile_layout *flo,
	206	struct nfs4_ff_layout_mirror *mirror, u64 offset,
	207	u64 length, int status, enum nfs_opnum4 opnum,
	208	gfp_t gfp_flags);
5b9b3c85 TM	209	int ff_layout_encode_ds_ioerr(struct xdr_stream xdr, const struct list_head head);
	210	void ff_layout_free_ds_ioerr(struct list_head *head);
	211	unsigned int ff_layout_fetch_ds_ioerr(struct pnfs_layout_hdr *lo,
	212	const struct pnfs_layout_range *range,
	213	struct list_head *head,
	214	unsigned int maxnum);
d67ae825 TH	215	struct nfs_fh *
	216	nfs4_ff_layout_select_ds_fh(struct pnfs_layout_segment *lseg, u32 mirror_idx);
	217
	218	struct nfs4_pnfs_ds *
	219	nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
	220	bool fail_return);
	221
	222	struct rpc_clnt *
	223	nfs4_ff_find_or_create_ds_client(struct pnfs_layout_segment *lseg,
	224	u32 ds_idx,
	225	struct nfs_client *ds_clp,
	226	struct inode *inode);
	227	struct rpc_cred ff_layout_get_ds_cred(struct pnfs_layout_segment lseg,
	228	u32 ds_idx, struct rpc_cred *mdscred);
3b13b4b3	229	bool ff_layout_avoid_mds_available_ds(struct pnfs_layout_segment *lseg);
fb1084e3 TH	230	bool ff_layout_avoid_read_on_rw(struct pnfs_layout_segment *lseg);
fb1084e3 TH	231
d67ae825	232	#endif /* FS_NFS_NFS4FLEXFILELAYOUT_H */