[deliverable/linux.git] / drivers / staging / lustre / lustre / mdc / mdc_lib.c

/*
 * GPL HEADER START
 *
 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 only,
 * as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License version 2 for more details (a copy is included
 * in the LICENSE file that accompanied this code).
 *
 * You should have received a copy of the GNU General Public License
 * version 2 along with this program; If not, see
 * http://www.gnu.org/licenses/gpl-2.0.html
 *
 * GPL HEADER END
 */
/*
 * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
 * Use is subject to license terms.
 *
 * Copyright (c) 2011, 2015, Intel Corporation.
 */
/*
 * This file is part of Lustre, http://www.lustre.org/
 * Lustre is a trademark of Sun Microsystems, Inc.
 */

#define DEBUG_SUBSYSTEM S_MDC
#include "../include/lustre_net.h"
#include "../include/lustre/lustre_idl.h"
#include "mdc_internal.h"

static void __mdc_pack_body(struct mdt_body *b, __u32 suppgid)
{
	b->mbo_suppgid = suppgid;
	b->mbo_uid = from_kuid(&init_user_ns, current_uid());
	b->mbo_gid = from_kgid(&init_user_ns, current_gid());
	b->mbo_fsuid = from_kuid(&init_user_ns, current_fsuid());
	b->mbo_fsgid = from_kgid(&init_user_ns, current_fsgid());
	b->mbo_capability = cfs_curproc_cap_pack();
}

void mdc_swap_layouts_pack(struct ptlrpc_request *req,
			   struct md_op_data *op_data)
{
	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
						    &RMF_MDT_BODY);

	__mdc_pack_body(b, op_data->op_suppgids[0]);
	b->mbo_fid1 = op_data->op_fid1;
	b->mbo_fid2 = op_data->op_fid2;
	b->mbo_valid |= OBD_MD_FLID;
}

void mdc_pack_body(struct ptlrpc_request *req, const struct lu_fid *fid,
		   __u64 valid, int ea_size, __u32 suppgid, int flags)
{
	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
						    &RMF_MDT_BODY);
	b->mbo_valid = valid;
	b->mbo_eadatasize = ea_size;
	b->mbo_flags = flags;
	__mdc_pack_body(b, suppgid);
	if (fid) {
		b->mbo_fid1 = *fid;
		b->mbo_valid |= OBD_MD_FLID;
	}
}

/**
 * Pack a name (path component) into a request
 *
 * \param[in] req	request
 * \param[in] field	request field (usually RMF_NAME)
 * \param[in] name	path component
 * \param[in] name_len	length of path component
 *
 * \a field must be present in \a req and of size \a name_len + 1.
 *
 * \a name must be '\0' terminated of length \a name_len and represent
 * a single path component (not contain '/').
 */
static void mdc_pack_name(struct ptlrpc_request *req,
			  const struct req_msg_field *field,
			  const char *name, size_t name_len)
{
	size_t buf_size;
	size_t cpy_len;
	char *buf;

	buf = req_capsule_client_get(&req->rq_pill, field);
	buf_size = req_capsule_get_size(&req->rq_pill, field, RCL_CLIENT);

	LASSERT(name && name_len && buf && buf_size == name_len + 1);

	cpy_len = strlcpy(buf, name, buf_size);

	LASSERT(cpy_len == name_len && lu_name_is_valid_2(buf, cpy_len));
}

void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff,
		      __u32 size, const struct lu_fid *fid)
{
	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
						    &RMF_MDT_BODY);
	b->mbo_fid1 = *fid;
	b->mbo_valid |= OBD_MD_FLID;
	b->mbo_size = pgoff;		       /* !! */
	b->mbo_nlink = size;			/* !! */
	__mdc_pack_body(b, -1);
	b->mbo_mode = LUDA_FID | LUDA_TYPE;
}

/* packing of MDS records */
void mdc_create_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
		     const void *data, int datalen, __u32 mode,
		     __u32 uid, __u32 gid, cfs_cap_t cap_effective, __u64 rdev)
{
	struct mdt_rec_create	*rec;
	char			*tmp;
	__u64			 flags;

	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_create));
	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);

	rec->cr_opcode   = REINT_CREATE;
	rec->cr_fsuid    = uid;
	rec->cr_fsgid    = gid;
	rec->cr_cap      = cap_effective;
	rec->cr_fid1     = op_data->op_fid1;
	rec->cr_fid2     = op_data->op_fid2;
	rec->cr_mode     = mode;
	rec->cr_rdev     = rdev;
	rec->cr_time     = op_data->op_mod_time;
	rec->cr_suppgid1 = op_data->op_suppgids[0];
	rec->cr_suppgid2 = op_data->op_suppgids[1];
	flags = op_data->op_flags & MF_SOM_LOCAL_FLAGS;
	if (op_data->op_bias & MDS_CREATE_VOLATILE)
		flags |= MDS_OPEN_VOLATILE;
	set_mrc_cr_flags(rec, flags);
	rec->cr_bias     = op_data->op_bias;
	rec->cr_umask    = current_umask();

	mdc_pack_name(req, &RMF_NAME, op_data->op_name, op_data->op_namelen);
	if (data) {
		tmp = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
		memcpy(tmp, data, datalen);
	}
}

static __u64 mds_pack_open_flags(__u64 flags, __u32 mode)
{
	__u64 cr_flags = (flags & (FMODE_READ | FMODE_WRITE |
				   MDS_OPEN_FL_INTERNAL));
	if (flags & O_CREAT)
		cr_flags |= MDS_OPEN_CREAT;
	if (flags & O_EXCL)
		cr_flags |= MDS_OPEN_EXCL;
	if (flags & O_TRUNC)
		cr_flags |= MDS_OPEN_TRUNC;
	if (flags & O_APPEND)
		cr_flags |= MDS_OPEN_APPEND;
	if (flags & O_SYNC)
		cr_flags |= MDS_OPEN_SYNC;
	if (flags & O_DIRECTORY)
		cr_flags |= MDS_OPEN_DIRECTORY;
	if (flags & __FMODE_EXEC)
		cr_flags |= MDS_FMODE_EXEC;
	if (cl_is_lov_delay_create(flags))
		cr_flags |= MDS_OPEN_DELAY_CREATE;

	if (flags & O_NONBLOCK)
		cr_flags |= MDS_OPEN_NORESTORE;

	return cr_flags;
}

/* packing of MDS records */
void mdc_open_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
		   __u32 mode, __u64 rdev, __u64 flags, const void *lmm,
		   int lmmlen)
{
	struct mdt_rec_create *rec;
	char *tmp;
	__u64 cr_flags;

	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_create));
	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);

	/* XXX do something about time, uid, gid */
	rec->cr_opcode   = REINT_OPEN;
	rec->cr_fsuid    = from_kuid(&init_user_ns, current_fsuid());
	rec->cr_fsgid    = from_kgid(&init_user_ns, current_fsgid());
	rec->cr_cap      = cfs_curproc_cap_pack();
	rec->cr_fid1 = op_data->op_fid1;
	rec->cr_fid2 = op_data->op_fid2;

	rec->cr_mode     = mode;
	cr_flags = mds_pack_open_flags(flags, mode);
	rec->cr_rdev     = rdev;
	rec->cr_time     = op_data->op_mod_time;
	rec->cr_suppgid1 = op_data->op_suppgids[0];
	rec->cr_suppgid2 = op_data->op_suppgids[1];
	rec->cr_bias     = op_data->op_bias;
	rec->cr_umask    = current_umask();
	rec->cr_old_handle = op_data->op_handle;

	if (op_data->op_name) {
		mdc_pack_name(req, &RMF_NAME, op_data->op_name,
			      op_data->op_namelen);

		if (op_data->op_bias & MDS_CREATE_VOLATILE)
			cr_flags |= MDS_OPEN_VOLATILE;
	}

	if (lmm) {
		cr_flags |= MDS_OPEN_HAS_EA;
		tmp = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
		memcpy(tmp, lmm, lmmlen);
	}
	set_mrc_cr_flags(rec, cr_flags);
}

static inline __u64 attr_pack(unsigned int ia_valid)
{
	__u64 sa_valid = 0;

	if (ia_valid & ATTR_MODE)
		sa_valid |= MDS_ATTR_MODE;
	if (ia_valid & ATTR_UID)
		sa_valid |= MDS_ATTR_UID;
	if (ia_valid & ATTR_GID)
		sa_valid |= MDS_ATTR_GID;
	if (ia_valid & ATTR_SIZE)
		sa_valid |= MDS_ATTR_SIZE;
	if (ia_valid & ATTR_ATIME)
		sa_valid |= MDS_ATTR_ATIME;
	if (ia_valid & ATTR_MTIME)
		sa_valid |= MDS_ATTR_MTIME;
	if (ia_valid & ATTR_CTIME)
		sa_valid |= MDS_ATTR_CTIME;
	if (ia_valid & ATTR_ATIME_SET)
		sa_valid |= MDS_ATTR_ATIME_SET;
	if (ia_valid & ATTR_MTIME_SET)
		sa_valid |= MDS_ATTR_MTIME_SET;
	if (ia_valid & ATTR_FORCE)
		sa_valid |= MDS_ATTR_FORCE;
	if (ia_valid & ATTR_ATTR_FLAG)
		sa_valid |= MDS_ATTR_ATTR_FLAG;
	if (ia_valid & ATTR_KILL_SUID)
		sa_valid |=  MDS_ATTR_KILL_SUID;
	if (ia_valid & ATTR_KILL_SGID)
		sa_valid |= MDS_ATTR_KILL_SGID;
	if (ia_valid & ATTR_CTIME_SET)
		sa_valid |= MDS_ATTR_CTIME_SET;
	if (ia_valid & ATTR_OPEN)
		sa_valid |= MDS_ATTR_FROM_OPEN;
	if (ia_valid & ATTR_BLOCKS)
		sa_valid |= MDS_ATTR_BLOCKS;
	if (ia_valid & MDS_OPEN_OWNEROVERRIDE)
		/* NFSD hack (see bug 5781) */
		sa_valid |= MDS_OPEN_OWNEROVERRIDE;
	return sa_valid;
}

static void mdc_setattr_pack_rec(struct mdt_rec_setattr *rec,
				 struct md_op_data *op_data)
{
	rec->sa_opcode  = REINT_SETATTR;
	rec->sa_fsuid   = from_kuid(&init_user_ns, current_fsuid());
	rec->sa_fsgid   = from_kgid(&init_user_ns, current_fsgid());
	rec->sa_cap     = cfs_curproc_cap_pack();
	rec->sa_suppgid = -1;

	rec->sa_fid    = op_data->op_fid1;
	rec->sa_valid  = attr_pack(op_data->op_attr.ia_valid);
	rec->sa_mode   = op_data->op_attr.ia_mode;
	rec->sa_uid    = from_kuid(&init_user_ns, op_data->op_attr.ia_uid);
	rec->sa_gid    = from_kgid(&init_user_ns, op_data->op_attr.ia_gid);
	rec->sa_size   = op_data->op_attr.ia_size;
	rec->sa_blocks = op_data->op_attr_blocks;
	rec->sa_atime  = LTIME_S(op_data->op_attr.ia_atime);
	rec->sa_mtime  = LTIME_S(op_data->op_attr.ia_mtime);
	rec->sa_ctime  = LTIME_S(op_data->op_attr.ia_ctime);
	rec->sa_attr_flags = op_data->op_attr_flags;
	if ((op_data->op_attr.ia_valid & ATTR_GID) &&
	    in_group_p(op_data->op_attr.ia_gid))
		rec->sa_suppgid =
			from_kgid(&init_user_ns, op_data->op_attr.ia_gid);
	else
		rec->sa_suppgid = op_data->op_suppgids[0];

	rec->sa_bias = op_data->op_bias;
}

static void mdc_ioepoch_pack(struct mdt_ioepoch *epoch,
			     struct md_op_data *op_data)
{
	memcpy(&epoch->handle, &op_data->op_handle, sizeof(epoch->handle));
	epoch->ioepoch = op_data->op_ioepoch;
	epoch->flags = op_data->op_flags & MF_SOM_LOCAL_FLAGS;
}

void mdc_setattr_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
		      void *ea, int ealen, void *ea2, int ea2len)
{
	struct mdt_rec_setattr *rec;
	struct mdt_ioepoch *epoch;
	struct lov_user_md *lum = NULL;

	CLASSERT(sizeof(struct mdt_rec_reint) ==
					sizeof(struct mdt_rec_setattr));
	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
	mdc_setattr_pack_rec(rec, op_data);

	if (op_data->op_flags & (MF_SOM_CHANGE | MF_EPOCH_OPEN)) {
		epoch = req_capsule_client_get(&req->rq_pill, &RMF_MDT_EPOCH);
		mdc_ioepoch_pack(epoch, op_data);
	}

	if (ealen == 0)
		return;

	lum = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
	if (!ea) { /* Remove LOV EA */
		lum->lmm_magic = LOV_USER_MAGIC_V1;
		lum->lmm_stripe_size = 0;
		lum->lmm_stripe_count = 0;
		lum->lmm_stripe_offset = (typeof(lum->lmm_stripe_offset))(-1);
	} else {
		memcpy(lum, ea, ealen);
	}

	if (ea2len == 0)
		return;

	memcpy(req_capsule_client_get(&req->rq_pill, &RMF_LOGCOOKIES), ea2,
	       ea2len);
}

void mdc_unlink_pack(struct ptlrpc_request *req, struct md_op_data *op_data)
{
	struct mdt_rec_unlink *rec;

	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_unlink));
	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);

	rec->ul_opcode   = op_data->op_cli_flags & CLI_RM_ENTRY ?
					REINT_RMENTRY : REINT_UNLINK;
	rec->ul_fsuid    = op_data->op_fsuid;
	rec->ul_fsgid    = op_data->op_fsgid;
	rec->ul_cap      = op_data->op_cap;
	rec->ul_mode     = op_data->op_mode;
	rec->ul_suppgid1 = op_data->op_suppgids[0];
	rec->ul_suppgid2 = -1;
	rec->ul_fid1     = op_data->op_fid1;
	rec->ul_fid2     = op_data->op_fid2;
	rec->ul_time     = op_data->op_mod_time;
	rec->ul_bias     = op_data->op_bias;

	mdc_pack_name(req, &RMF_NAME, op_data->op_name, op_data->op_namelen);
}

void mdc_link_pack(struct ptlrpc_request *req, struct md_op_data *op_data)
{
	struct mdt_rec_link *rec;

	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_link));
	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);

	rec->lk_opcode   = REINT_LINK;
	rec->lk_fsuid    = op_data->op_fsuid; /* current->fsuid; */
	rec->lk_fsgid    = op_data->op_fsgid; /* current->fsgid; */
	rec->lk_cap      = op_data->op_cap;   /* current->cap_effective; */
	rec->lk_suppgid1 = op_data->op_suppgids[0];
	rec->lk_suppgid2 = op_data->op_suppgids[1];
	rec->lk_fid1     = op_data->op_fid1;
	rec->lk_fid2     = op_data->op_fid2;
	rec->lk_time     = op_data->op_mod_time;
	rec->lk_bias     = op_data->op_bias;

	mdc_pack_name(req, &RMF_NAME, op_data->op_name, op_data->op_namelen);
}

void mdc_rename_pack(struct ptlrpc_request *req, struct md_op_data *op_data,
		     const char *old, int oldlen, const char *new, int newlen)
{
	struct mdt_rec_rename *rec;

	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_rename));
	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);

	/* XXX do something about time, uid, gid */
	rec->rn_opcode	 = op_data->op_cli_flags & CLI_MIGRATE ?
				REINT_MIGRATE : REINT_RENAME;
	rec->rn_opcode   = REINT_RENAME;
	rec->rn_fsuid    = op_data->op_fsuid;
	rec->rn_fsgid    = op_data->op_fsgid;
	rec->rn_cap      = op_data->op_cap;
	rec->rn_suppgid1 = op_data->op_suppgids[0];
	rec->rn_suppgid2 = op_data->op_suppgids[1];
	rec->rn_fid1     = op_data->op_fid1;
	rec->rn_fid2     = op_data->op_fid2;
	rec->rn_time     = op_data->op_mod_time;
	rec->rn_mode     = op_data->op_mode;
	rec->rn_bias     = op_data->op_bias;

	mdc_pack_name(req, &RMF_NAME, old, oldlen);

	if (new)
		mdc_pack_name(req, &RMF_SYMTGT, new, newlen);
}

void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, int flags,
		      struct md_op_data *op_data, int ea_size)
{
	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
						    &RMF_MDT_BODY);

	b->mbo_valid = valid;
	if (op_data->op_bias & MDS_CHECK_SPLIT)
		b->mbo_valid |= OBD_MD_FLCKSPLIT;
	if (op_data->op_bias & MDS_CROSS_REF)
		b->mbo_valid |= OBD_MD_FLCROSSREF;
	b->mbo_eadatasize = ea_size;
	b->mbo_flags = flags;
	__mdc_pack_body(b, op_data->op_suppgids[0]);

	b->mbo_fid1 = op_data->op_fid1;
	b->mbo_fid2 = op_data->op_fid2;
	b->mbo_valid |= OBD_MD_FLID;

	if (op_data->op_name)
		mdc_pack_name(req, &RMF_NAME, op_data->op_name,
			      op_data->op_namelen);
}

static void mdc_hsm_release_pack(struct ptlrpc_request *req,
				 struct md_op_data *op_data)
{
	if (op_data->op_bias & MDS_HSM_RELEASE) {
		struct close_data *data;
		struct ldlm_lock *lock;

		data = req_capsule_client_get(&req->rq_pill, &RMF_CLOSE_DATA);

		lock = ldlm_handle2lock(&op_data->op_lease_handle);
		if (lock) {
			data->cd_handle = lock->l_remote_handle;
			LDLM_LOCK_PUT(lock);
		}
		ldlm_cli_cancel(&op_data->op_lease_handle, LCF_LOCAL);

		data->cd_data_version = op_data->op_data_version;
		data->cd_fid = op_data->op_fid2;
	}
}

void mdc_close_pack(struct ptlrpc_request *req, struct md_op_data *op_data)
{
	struct mdt_ioepoch *epoch;
	struct mdt_rec_setattr *rec;

	epoch = req_capsule_client_get(&req->rq_pill, &RMF_MDT_EPOCH);
	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);

	mdc_setattr_pack_rec(rec, op_data);
	/*
	 * The client will zero out local timestamps when losing the IBITS lock
	 * so any new RPC timestamps will update the client inode's timestamps.
	 * There was a defect on the server side which allowed the atime to be
	 * overwritten by a zeroed-out atime packed into the close RPC.
	 *
	 * Proactively clear the MDS_ATTR_ATIME flag in the RPC in this case
	 * to avoid zeroing the atime on old unpatched servers.  See LU-8041.
	 */
	if (rec->sa_atime == 0)
		rec->sa_valid &= ~MDS_ATTR_ATIME;

	mdc_ioepoch_pack(epoch, op_data);
	mdc_hsm_release_pack(req, op_data);
}
Commit	Line	Data
d7e09d03 PT	1	/*
	2	* GPL HEADER START
	3	*
	4	* DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
	5	*
	6	* This program is free software; you can redistribute it and/or modify
	7	* it under the terms of the GNU General Public License version 2 only,
	8	* as published by the Free Software Foundation.
	9	*
	10	* This program is distributed in the hope that it will be useful, but
	11	* WITHOUT ANY WARRANTY; without even the implied warranty of
	12	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	13	* General Public License version 2 for more details (a copy is included
	14	* in the LICENSE file that accompanied this code).
	15	*
	16	* You should have received a copy of the GNU General Public License
	17	* version 2 along with this program; If not, see
6a5b99a4	18	* http://www.gnu.org/licenses/gpl-2.0.html
d7e09d03	19	*
d7e09d03 PT	20	* GPL HEADER END
	21	*/
	22	/*
	23	* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
	24	* Use is subject to license terms.
	25	*
1dc563a6	26	* Copyright (c) 2011, 2015, Intel Corporation.
d7e09d03 PT	27	*/
	28	/*
	29	* This file is part of Lustre, http://www.lustre.org/
	30	* Lustre is a trademark of Sun Microsystems, Inc.
	31	*/
	32
	33	#define DEBUG_SUBSYSTEM S_MDC
05932307 GKH	34	#include "../include/lustre_net.h"
05932307 GKH	35	#include "../include/lustre/lustre_idl.h"
d7e09d03 PT	36	#include "mdc_internal.h"
d7e09d03 PT	37
d7e09d03 PT	38	static void __mdc_pack_body(struct mdt_body *b, __u32 suppgid)
d7e09d03 PT	39	{
2e1b5b8b JH	40	b->mbo_suppgid = suppgid;
	41	b->mbo_uid = from_kuid(&init_user_ns, current_uid());
	42	b->mbo_gid = from_kgid(&init_user_ns, current_gid());
	43	b->mbo_fsuid = from_kuid(&init_user_ns, current_fsuid());
	44	b->mbo_fsgid = from_kgid(&init_user_ns, current_fsgid());
	45	b->mbo_capability = cfs_curproc_cap_pack();
d7e09d03 PT	46	}
d7e09d03 PT	47
d7e09d03 PT	48	void mdc_swap_layouts_pack(struct ptlrpc_request *req,
	49	struct md_op_data *op_data)
	50	{
	51	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
	52	&RMF_MDT_BODY);
	53
	54	__mdc_pack_body(b, op_data->op_suppgids[0]);
2e1b5b8b JH	55	b->mbo_fid1 = op_data->op_fid1;
	56	b->mbo_fid2 = op_data->op_fid2;
	57	b->mbo_valid \|= OBD_MD_FLID;
d7e09d03 PT	58	}
d7e09d03 PT	59
ef2e0f55	60	void mdc_pack_body(struct ptlrpc_request req, const struct lu_fid fid,
d7e09d03 PT	61	__u64 valid, int ea_size, __u32 suppgid, int flags)
	62	{
	63	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
	64	&RMF_MDT_BODY);
2e1b5b8b JH	65	b->mbo_valid = valid;
	66	b->mbo_eadatasize = ea_size;
	67	b->mbo_flags = flags;
d7e09d03 PT	68	__mdc_pack_body(b, suppgid);
d7e09d03 PT	69	if (fid) {
2e1b5b8b JH	70	b->mbo_fid1 = *fid;
2e1b5b8b JH	71	b->mbo_valid \|= OBD_MD_FLID;
d7e09d03 PT	72	}
	73	}
	74
d097d67b JH	75	/**
	76	* Pack a name (path component) into a request
	77	*
	78	* \param[in] req request
	79	* \param[in] field request field (usually RMF_NAME)
	80	* \param[in] name path component
	81	* \param[in] name_len length of path component
	82	*
	83	* \a field must be present in \a req and of size \a name_len + 1.
	84	*
	85	* \a name must be '\0' terminated of length \a name_len and represent
	86	* a single path component (not contain '/').
	87	*/
	88	static void mdc_pack_name(struct ptlrpc_request *req,
	89	const struct req_msg_field *field,
	90	const char *name, size_t name_len)
	91	{
	92	size_t buf_size;
	93	size_t cpy_len;
	94	char *buf;
	95
	96	buf = req_capsule_client_get(&req->rq_pill, field);
	97	buf_size = req_capsule_get_size(&req->rq_pill, field, RCL_CLIENT);
	98
	99	LASSERT(name && name_len && buf && buf_size == name_len + 1);
	100
	101	cpy_len = strlcpy(buf, name, buf_size);
	102
	103	LASSERT(cpy_len == name_len && lu_name_is_valid_2(buf, cpy_len));
	104	}
	105
d7e09d03	106	void mdc_readdir_pack(struct ptlrpc_request *req, __u64 pgoff,
ef2e0f55	107	__u32 size, const struct lu_fid *fid)
d7e09d03 PT	108	{
	109	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
	110	&RMF_MDT_BODY);
2e1b5b8b JH	111	b->mbo_fid1 = *fid;
	112	b->mbo_valid \|= OBD_MD_FLID;
	113	b->mbo_size = pgoff; /* !! */
	114	b->mbo_nlink = size; /* !! */
d7e09d03	115	__mdc_pack_body(b, -1);
2e1b5b8b	116	b->mbo_mode = LUDA_FID \| LUDA_TYPE;
d7e09d03 PT	117	}
	118
	119	/* packing of MDS records */
	120	void mdc_create_pack(struct ptlrpc_request req, struct md_op_data op_data,
	121	const void *data, int datalen, __u32 mode,
	122	__u32 uid, __u32 gid, cfs_cap_t cap_effective, __u64 rdev)
	123	{
	124	struct mdt_rec_create *rec;
	125	char *tmp;
	126	__u64 flags;
	127
	128	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_create));
	129	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
	130
d7e09d03 PT	131	rec->cr_opcode = REINT_CREATE;
	132	rec->cr_fsuid = uid;
	133	rec->cr_fsgid = gid;
	134	rec->cr_cap = cap_effective;
	135	rec->cr_fid1 = op_data->op_fid1;
	136	rec->cr_fid2 = op_data->op_fid2;
	137	rec->cr_mode = mode;
	138	rec->cr_rdev = rdev;
	139	rec->cr_time = op_data->op_mod_time;
	140	rec->cr_suppgid1 = op_data->op_suppgids[0];
	141	rec->cr_suppgid2 = op_data->op_suppgids[1];
	142	flags = op_data->op_flags & MF_SOM_LOCAL_FLAGS;
	143	if (op_data->op_bias & MDS_CREATE_VOLATILE)
	144	flags \|= MDS_OPEN_VOLATILE;
	145	set_mrc_cr_flags(rec, flags);
	146	rec->cr_bias = op_data->op_bias;
	147	rec->cr_umask = current_umask();
	148
d097d67b	149	mdc_pack_name(req, &RMF_NAME, op_data->op_name, op_data->op_namelen);
d7e09d03 PT	150	if (data) {
	151	tmp = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
	152	memcpy(tmp, data, datalen);
	153	}
	154	}
	155
d3a8a4e2	156	static __u64 mds_pack_open_flags(__u64 flags, __u32 mode)
d7e09d03 PT	157	{
d7e09d03 PT	158	__u64 cr_flags = (flags & (FMODE_READ \| FMODE_WRITE \|
c1b66fcc	159	MDS_OPEN_FL_INTERNAL));
d7e09d03 PT	160	if (flags & O_CREAT)
	161	cr_flags \|= MDS_OPEN_CREAT;
	162	if (flags & O_EXCL)
	163	cr_flags \|= MDS_OPEN_EXCL;
	164	if (flags & O_TRUNC)
	165	cr_flags \|= MDS_OPEN_TRUNC;
	166	if (flags & O_APPEND)
	167	cr_flags \|= MDS_OPEN_APPEND;
	168	if (flags & O_SYNC)
	169	cr_flags \|= MDS_OPEN_SYNC;
	170	if (flags & O_DIRECTORY)
	171	cr_flags \|= MDS_OPEN_DIRECTORY;
962dbfd6	172	if (flags & __FMODE_EXEC)
d7e09d03	173	cr_flags \|= MDS_FMODE_EXEC;
38585ccc	174	if (cl_is_lov_delay_create(flags))
d7e09d03 PT	175	cr_flags \|= MDS_OPEN_DELAY_CREATE;
d7e09d03 PT	176
7d5ed06b	177	if (flags & O_NONBLOCK)
d7e09d03 PT	178	cr_flags \|= MDS_OPEN_NORESTORE;
	179
	180	return cr_flags;
	181	}
	182
	183	/* packing of MDS records */
	184	void mdc_open_pack(struct ptlrpc_request req, struct md_op_data op_data,
d3a8a4e2	185	__u32 mode, __u64 rdev, __u64 flags, const void *lmm,
d7e09d03 PT	186	int lmmlen)
	187	{
	188	struct mdt_rec_create *rec;
	189	char *tmp;
	190	__u64 cr_flags;
	191
	192	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_create));
	193	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
	194
	195	/* XXX do something about time, uid, gid */
	196	rec->cr_opcode = REINT_OPEN;
4b1a25f0 PT	197	rec->cr_fsuid = from_kuid(&init_user_ns, current_fsuid());
4b1a25f0 PT	198	rec->cr_fsgid = from_kgid(&init_user_ns, current_fsgid());
d7e09d03	199	rec->cr_cap = cfs_curproc_cap_pack();
cc1cf0e0 RS	200	rec->cr_fid1 = op_data->op_fid1;
	201	rec->cr_fid2 = op_data->op_fid2;
	202
d7e09d03 PT	203	rec->cr_mode = mode;
	204	cr_flags = mds_pack_open_flags(flags, mode);
	205	rec->cr_rdev = rdev;
	206	rec->cr_time = op_data->op_mod_time;
	207	rec->cr_suppgid1 = op_data->op_suppgids[0];
	208	rec->cr_suppgid2 = op_data->op_suppgids[1];
	209	rec->cr_bias = op_data->op_bias;
	210	rec->cr_umask = current_umask();
d3a8a4e2	211	rec->cr_old_handle = op_data->op_handle;
d7e09d03	212
d7e09d03	213	if (op_data->op_name) {
d097d67b JH	214	mdc_pack_name(req, &RMF_NAME, op_data->op_name,
	215	op_data->op_namelen);
	216
d7e09d03 PT	217	if (op_data->op_bias & MDS_CREATE_VOLATILE)
	218	cr_flags \|= MDS_OPEN_VOLATILE;
	219	}
	220
	221	if (lmm) {
	222	cr_flags \|= MDS_OPEN_HAS_EA;
	223	tmp = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
	224	memcpy(tmp, lmm, lmmlen);
	225	}
	226	set_mrc_cr_flags(rec, cr_flags);
	227	}
	228
982ec91f SM	229	static inline __u64 attr_pack(unsigned int ia_valid)
982ec91f SM	230	{
d7e09d03 PT	231	__u64 sa_valid = 0;
	232
	233	if (ia_valid & ATTR_MODE)
	234	sa_valid \|= MDS_ATTR_MODE;
	235	if (ia_valid & ATTR_UID)
	236	sa_valid \|= MDS_ATTR_UID;
	237	if (ia_valid & ATTR_GID)
	238	sa_valid \|= MDS_ATTR_GID;
	239	if (ia_valid & ATTR_SIZE)
	240	sa_valid \|= MDS_ATTR_SIZE;
	241	if (ia_valid & ATTR_ATIME)
	242	sa_valid \|= MDS_ATTR_ATIME;
	243	if (ia_valid & ATTR_MTIME)
	244	sa_valid \|= MDS_ATTR_MTIME;
	245	if (ia_valid & ATTR_CTIME)
	246	sa_valid \|= MDS_ATTR_CTIME;
	247	if (ia_valid & ATTR_ATIME_SET)
	248	sa_valid \|= MDS_ATTR_ATIME_SET;
	249	if (ia_valid & ATTR_MTIME_SET)
	250	sa_valid \|= MDS_ATTR_MTIME_SET;
	251	if (ia_valid & ATTR_FORCE)
	252	sa_valid \|= MDS_ATTR_FORCE;
	253	if (ia_valid & ATTR_ATTR_FLAG)
	254	sa_valid \|= MDS_ATTR_ATTR_FLAG;
	255	if (ia_valid & ATTR_KILL_SUID)
	256	sa_valid \|= MDS_ATTR_KILL_SUID;
	257	if (ia_valid & ATTR_KILL_SGID)
	258	sa_valid \|= MDS_ATTR_KILL_SGID;
	259	if (ia_valid & ATTR_CTIME_SET)
	260	sa_valid \|= MDS_ATTR_CTIME_SET;
e56e17ea	261	if (ia_valid & ATTR_OPEN)
d7e09d03 PT	262	sa_valid \|= MDS_ATTR_FROM_OPEN;
	263	if (ia_valid & ATTR_BLOCKS)
	264	sa_valid \|= MDS_ATTR_BLOCKS;
	265	if (ia_valid & MDS_OPEN_OWNEROVERRIDE)
	266	/* NFSD hack (see bug 5781) */
	267	sa_valid \|= MDS_OPEN_OWNEROVERRIDE;
	268	return sa_valid;
	269	}
	270
	271	static void mdc_setattr_pack_rec(struct mdt_rec_setattr *rec,
	272	struct md_op_data *op_data)
	273	{
	274	rec->sa_opcode = REINT_SETATTR;
4b1a25f0 PT	275	rec->sa_fsuid = from_kuid(&init_user_ns, current_fsuid());
4b1a25f0 PT	276	rec->sa_fsgid = from_kgid(&init_user_ns, current_fsgid());
d7e09d03 PT	277	rec->sa_cap = cfs_curproc_cap_pack();
	278	rec->sa_suppgid = -1;
	279
	280	rec->sa_fid = op_data->op_fid1;
	281	rec->sa_valid = attr_pack(op_data->op_attr.ia_valid);
	282	rec->sa_mode = op_data->op_attr.ia_mode;
4b1a25f0 PT	283	rec->sa_uid = from_kuid(&init_user_ns, op_data->op_attr.ia_uid);
4b1a25f0 PT	284	rec->sa_gid = from_kgid(&init_user_ns, op_data->op_attr.ia_gid);
d7e09d03 PT	285	rec->sa_size = op_data->op_attr.ia_size;
	286	rec->sa_blocks = op_data->op_attr_blocks;
	287	rec->sa_atime = LTIME_S(op_data->op_attr.ia_atime);
	288	rec->sa_mtime = LTIME_S(op_data->op_attr.ia_mtime);
	289	rec->sa_ctime = LTIME_S(op_data->op_attr.ia_ctime);
bb41292b	290	rec->sa_attr_flags = op_data->op_attr_flags;
d7e09d03	291	if ((op_data->op_attr.ia_valid & ATTR_GID) &&
4b1a25f0 PT	292	in_group_p(op_data->op_attr.ia_gid))
	293	rec->sa_suppgid =
	294	from_kgid(&init_user_ns, op_data->op_attr.ia_gid);
d7e09d03 PT	295	else
	296	rec->sa_suppgid = op_data->op_suppgids[0];
	297
	298	rec->sa_bias = op_data->op_bias;
	299	}
	300
	301	static void mdc_ioepoch_pack(struct mdt_ioepoch *epoch,
	302	struct md_op_data *op_data)
	303	{
	304	memcpy(&epoch->handle, &op_data->op_handle, sizeof(epoch->handle));
	305	epoch->ioepoch = op_data->op_ioepoch;
	306	epoch->flags = op_data->op_flags & MF_SOM_LOCAL_FLAGS;
	307	}
	308
	309	void mdc_setattr_pack(struct ptlrpc_request req, struct md_op_data op_data,
	310	void ea, int ealen, void ea2, int ea2len)
	311	{
	312	struct mdt_rec_setattr *rec;
	313	struct mdt_ioepoch *epoch;
	314	struct lov_user_md *lum = NULL;
	315
301af906 SM	316	CLASSERT(sizeof(struct mdt_rec_reint) ==
301af906 SM	317	sizeof(struct mdt_rec_setattr));
d7e09d03 PT	318	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
	319	mdc_setattr_pack_rec(rec, op_data);
	320
d7e09d03 PT	321	if (op_data->op_flags & (MF_SOM_CHANGE \| MF_EPOCH_OPEN)) {
	322	epoch = req_capsule_client_get(&req->rq_pill, &RMF_MDT_EPOCH);
	323	mdc_ioepoch_pack(epoch, op_data);
	324	}
	325
	326	if (ealen == 0)
	327	return;
	328
	329	lum = req_capsule_client_get(&req->rq_pill, &RMF_EADATA);
34e3ff96	330	if (!ea) { /* Remove LOV EA */
d7e09d03 PT	331	lum->lmm_magic = LOV_USER_MAGIC_V1;
	332	lum->lmm_stripe_size = 0;
	333	lum->lmm_stripe_count = 0;
	334	lum->lmm_stripe_offset = (typeof(lum->lmm_stripe_offset))(-1);
	335	} else {
	336	memcpy(lum, ea, ealen);
	337	}
	338
	339	if (ea2len == 0)
	340	return;
	341
	342	memcpy(req_capsule_client_get(&req->rq_pill, &RMF_LOGCOOKIES), ea2,
	343	ea2len);
	344	}
	345
	346	void mdc_unlink_pack(struct ptlrpc_request req, struct md_op_data op_data)
	347	{
	348	struct mdt_rec_unlink *rec;
d7e09d03 PT	349
	350	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_unlink));
	351	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
d7e09d03	352
301af906	353	rec->ul_opcode = op_data->op_cli_flags & CLI_RM_ENTRY ?
d7e09d03	354	REINT_RMENTRY : REINT_UNLINK;
301af906 SM	355	rec->ul_fsuid = op_data->op_fsuid;
	356	rec->ul_fsgid = op_data->op_fsgid;
	357	rec->ul_cap = op_data->op_cap;
	358	rec->ul_mode = op_data->op_mode;
	359	rec->ul_suppgid1 = op_data->op_suppgids[0];
	360	rec->ul_suppgid2 = -1;
	361	rec->ul_fid1 = op_data->op_fid1;
	362	rec->ul_fid2 = op_data->op_fid2;
	363	rec->ul_time = op_data->op_mod_time;
	364	rec->ul_bias = op_data->op_bias;
d7e09d03	365
d097d67b	366	mdc_pack_name(req, &RMF_NAME, op_data->op_name, op_data->op_namelen);
d7e09d03 PT	367	}
	368
	369	void mdc_link_pack(struct ptlrpc_request req, struct md_op_data op_data)
	370	{
	371	struct mdt_rec_link *rec;
d7e09d03 PT	372
	373	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_link));
	374	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
d7e09d03 PT	375
d7e09d03 PT	376	rec->lk_opcode = REINT_LINK;
125ffec0 SM	377	rec->lk_fsuid = op_data->op_fsuid; /* current->fsuid; */
	378	rec->lk_fsgid = op_data->op_fsgid; /* current->fsgid; */
	379	rec->lk_cap = op_data->op_cap; /* current->cap_effective; */
d7e09d03 PT	380	rec->lk_suppgid1 = op_data->op_suppgids[0];
	381	rec->lk_suppgid2 = op_data->op_suppgids[1];
	382	rec->lk_fid1 = op_data->op_fid1;
	383	rec->lk_fid2 = op_data->op_fid2;
	384	rec->lk_time = op_data->op_mod_time;
	385	rec->lk_bias = op_data->op_bias;
	386
d097d67b	387	mdc_pack_name(req, &RMF_NAME, op_data->op_name, op_data->op_namelen);
d7e09d03 PT	388	}
	389
	390	void mdc_rename_pack(struct ptlrpc_request req, struct md_op_data op_data,
	391	const char old, int oldlen, const char new, int newlen)
	392	{
	393	struct mdt_rec_rename *rec;
d7e09d03 PT	394
	395	CLASSERT(sizeof(struct mdt_rec_reint) == sizeof(struct mdt_rec_rename));
	396	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
	397
	398	/* XXX do something about time, uid, gid */
79496845	399	rec->rn_opcode = op_data->op_cli_flags & CLI_MIGRATE ?
79496845	400	REINT_MIGRATE : REINT_RENAME;
d7e09d03 PT	401	rec->rn_opcode = REINT_RENAME;
	402	rec->rn_fsuid = op_data->op_fsuid;
	403	rec->rn_fsgid = op_data->op_fsgid;
	404	rec->rn_cap = op_data->op_cap;
	405	rec->rn_suppgid1 = op_data->op_suppgids[0];
	406	rec->rn_suppgid2 = op_data->op_suppgids[1];
	407	rec->rn_fid1 = op_data->op_fid1;
	408	rec->rn_fid2 = op_data->op_fid2;
	409	rec->rn_time = op_data->op_mod_time;
	410	rec->rn_mode = op_data->op_mode;
	411	rec->rn_bias = op_data->op_bias;
	412
d097d67b	413	mdc_pack_name(req, &RMF_NAME, old, oldlen);
d7e09d03	414
d097d67b JH	415	if (new)
d097d67b JH	416	mdc_pack_name(req, &RMF_SYMTGT, new, newlen);
d7e09d03 PT	417	}
	418
	419	void mdc_getattr_pack(struct ptlrpc_request *req, __u64 valid, int flags,
	420	struct md_op_data *op_data, int ea_size)
	421	{
	422	struct mdt_body *b = req_capsule_client_get(&req->rq_pill,
	423	&RMF_MDT_BODY);
	424
2e1b5b8b	425	b->mbo_valid = valid;
d7e09d03	426	if (op_data->op_bias & MDS_CHECK_SPLIT)
2e1b5b8b	427	b->mbo_valid \|= OBD_MD_FLCKSPLIT;
d7e09d03	428	if (op_data->op_bias & MDS_CROSS_REF)
2e1b5b8b JH	429	b->mbo_valid \|= OBD_MD_FLCROSSREF;
	430	b->mbo_eadatasize = ea_size;
	431	b->mbo_flags = flags;
d7e09d03 PT	432	__mdc_pack_body(b, op_data->op_suppgids[0]);
d7e09d03 PT	433
2e1b5b8b JH	434	b->mbo_fid1 = op_data->op_fid1;
	435	b->mbo_fid2 = op_data->op_fid2;
	436	b->mbo_valid \|= OBD_MD_FLID;
d7e09d03	437
d097d67b JH	438	if (op_data->op_name)
	439	mdc_pack_name(req, &RMF_NAME, op_data->op_name,
	440	op_data->op_namelen);
d7e09d03 PT	441	}
d7e09d03 PT	442
48d23e61 JX	443	static void mdc_hsm_release_pack(struct ptlrpc_request *req,
	444	struct md_op_data *op_data)
	445	{
	446	if (op_data->op_bias & MDS_HSM_RELEASE) {
	447	struct close_data *data;
	448	struct ldlm_lock *lock;
	449
	450	data = req_capsule_client_get(&req->rq_pill, &RMF_CLOSE_DATA);
48d23e61 JX	451
48d23e61 JX	452	lock = ldlm_handle2lock(&op_data->op_lease_handle);
34e3ff96	453	if (lock) {
48d23e61	454	data->cd_handle = lock->l_remote_handle;
ead02808	455	LDLM_LOCK_PUT(lock);
48d23e61 JX	456	}
	457	ldlm_cli_cancel(&op_data->op_lease_handle, LCF_LOCAL);
	458
	459	data->cd_data_version = op_data->op_data_version;
	460	data->cd_fid = op_data->op_fid2;
	461	}
	462	}
	463
d7e09d03 PT	464	void mdc_close_pack(struct ptlrpc_request req, struct md_op_data op_data)
	465	{
	466	struct mdt_ioepoch *epoch;
	467	struct mdt_rec_setattr *rec;
	468
	469	epoch = req_capsule_client_get(&req->rq_pill, &RMF_MDT_EPOCH);
	470	rec = req_capsule_client_get(&req->rq_pill, &RMF_REC_REINT);
	471
	472	mdc_setattr_pack_rec(rec, op_data);
aea7ccd9 NY	473	/*
	474	* The client will zero out local timestamps when losing the IBITS lock
	475	* so any new RPC timestamps will update the client inode's timestamps.
	476	* There was a defect on the server side which allowed the atime to be
	477	* overwritten by a zeroed-out atime packed into the close RPC.
	478	*
	479	* Proactively clear the MDS_ATTR_ATIME flag in the RPC in this case
	480	* to avoid zeroing the atime on old unpatched servers. See LU-8041.
	481	*/
	482	if (rec->sa_atime == 0)
	483	rec->sa_valid &= ~MDS_ATTR_ATIME;
	484
d7e09d03	485	mdc_ioepoch_pack(epoch, op_data);
48d23e61	486	mdc_hsm_release_pack(req, op_data);
d7e09d03	487	}