diff options
39 files changed, 1464 insertions, 367 deletions
diff --git a/Documentation/filesystems/nfs/exporting.rst b/Documentation/filesystems/nfs/exporting.rst index 4b30daee399af7..198d805d611c67 100644 --- a/Documentation/filesystems/nfs/exporting.rst +++ b/Documentation/filesystems/nfs/exporting.rst @@ -241,3 +241,10 @@ following flags are defined: all of an inode's dirty data on last close. Exports that behave this way should set EXPORT_OP_FLUSH_ON_CLOSE so that NFSD knows to skip waiting for writeback when closing such files. + + EXPORT_OP_ASYNC_LOCK - Indicates a capable filesystem to do async lock + requests from lockd. Only set EXPORT_OP_ASYNC_LOCK if the filesystem has + it's own ->lock() functionality as core posix_lock_file() implementation + has no async lock request handling yet. For more information about how to + indicate an async lock request from a ->lock() file_operations struct, see + fs/locks.c and comment for the function vfs_lock_file(). diff --git a/Documentation/netlink/specs/nfsd.yaml b/Documentation/netlink/specs/nfsd.yaml new file mode 100644 index 00000000000000..403d3e3a04f37a --- /dev/null +++ b/Documentation/netlink/specs/nfsd.yaml @@ -0,0 +1,74 @@ +# SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) + +name: nfsd +protocol: genetlink +uapi-header: linux/nfsd_netlink.h + +doc: NFSD configuration over generic netlink. + +attribute-sets: + - + name: rpc-status + attributes: + - + name: xid + type: u32 + byte-order: big-endian + - + name: flags + type: u32 + - + name: prog + type: u32 + - + name: version + type: u8 + - + name: proc + type: u32 + - + name: service_time + type: s64 + - + name: pad + type: pad + - + name: saddr4 + type: u32 + byte-order: big-endian + display-hint: ipv4 + - + name: daddr4 + type: u32 + byte-order: big-endian + display-hint: ipv4 + - + name: saddr6 + type: binary + display-hint: ipv6 + - + name: daddr6 + type: binary + display-hint: ipv6 + - + name: sport + type: u16 + byte-order: big-endian + - + name: dport + type: u16 + byte-order: big-endian + - + name: compound-ops + type: u32 + multi-attr: true + +operations: + list: + - + name: rpc-status-get + doc: dump pending nfsd rpc + attribute-set: rpc-status + dump: + pre: nfsd-nl-rpc-status-get-start + post: nfsd-nl-rpc-status-get-done diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c index 6579948070a482..b441c706c2b82f 100644 --- a/fs/lockd/svc.c +++ b/fs/lockd/svc.c @@ -24,7 +24,6 @@ #include <linux/uio.h> #include <linux/smp.h> #include <linux/mutex.h> -#include <linux/kthread.h> #include <linux/freezer.h> #include <linux/inetdevice.h> @@ -135,11 +134,11 @@ lockd(void *vrqstp) * The main request loop. We don't terminate until the last * NFS mount or NFS daemon has gone away. */ - while (!kthread_should_stop()) { + while (!svc_thread_should_stop(rqstp)) { /* update sv_maxconn if it has changed */ rqstp->rq_server->sv_maxconn = nlm_max_connections; - nlmsvc_retry_blocked(); + nlmsvc_retry_blocked(rqstp); svc_recv(rqstp); } if (nlmsvc_ops) diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c index 43aeba9de55cbb..2dc10900ad1c33 100644 --- a/fs/lockd/svclock.c +++ b/fs/lockd/svclock.c @@ -30,7 +30,6 @@ #include <linux/sunrpc/svc_xprt.h> #include <linux/lockd/nlm.h> #include <linux/lockd/lockd.h> -#include <linux/kthread.h> #include <linux/exportfs.h> #define NLMDBG_FACILITY NLMDBG_SVCLOCK @@ -481,9 +480,7 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file, struct nlm_host *host, struct nlm_lock *lock, int wait, struct nlm_cookie *cookie, int reclaim) { -#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) struct inode *inode = nlmsvc_file_inode(file); -#endif struct nlm_block *block = NULL; int error; int mode; @@ -497,7 +494,7 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file, (long long)lock->fl.fl_end, wait); - if (nlmsvc_file_file(file)->f_op->lock) { + if (!exportfs_lock_op_is_async(inode->i_sb->s_export_op)) { async_block = wait; wait = 0; } @@ -543,6 +540,25 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file, goto out; } + spin_lock(&nlm_blocked_lock); + /* + * If this is a lock request for an already pending + * lock request we return nlm_lck_blocked without calling + * vfs_lock_file() again. Otherwise we have two pending + * requests on the underlaying ->lock() implementation but + * only one nlm_block to being granted by lm_grant(). + */ + if (exportfs_lock_op_is_async(inode->i_sb->s_export_op) && + !list_empty(&block->b_list)) { + spin_unlock(&nlm_blocked_lock); + ret = nlm_lck_blocked; + goto out; + } + + /* Append to list of blocked */ + nlmsvc_insert_block_locked(block, NLM_NEVER); + spin_unlock(&nlm_blocked_lock); + if (!wait) lock->fl.fl_flags &= ~FL_SLEEP; mode = lock_to_openmode(&lock->fl); @@ -552,16 +568,12 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file, dprintk("lockd: vfs_lock_file returned %d\n", error); switch (error) { case 0: + nlmsvc_remove_block(block); ret = nlm_granted; goto out; case -EAGAIN: - /* - * If this is a blocking request for an - * already pending lock request then we need - * to put it back on lockd's block list - */ - if (wait) - break; + if (!wait) + nlmsvc_remove_block(block); ret = async_block ? nlm_lck_blocked : nlm_lck_denied; goto out; case FILE_LOCK_DEFERRED: @@ -572,17 +584,16 @@ nlmsvc_lock(struct svc_rqst *rqstp, struct nlm_file *file, ret = nlmsvc_defer_lock_rqst(rqstp, block); goto out; case -EDEADLK: + nlmsvc_remove_block(block); ret = nlm_deadlock; goto out; default: /* includes ENOLCK */ + nlmsvc_remove_block(block); ret = nlm_lck_denied_nolocks; goto out; } ret = nlm_lck_blocked; - - /* Append to list of blocked */ - nlmsvc_insert_block(block, NLM_NEVER); out: mutex_unlock(&file->f_mutex); nlmsvc_release_block(block); @@ -1020,13 +1031,13 @@ retry_deferred_block(struct nlm_block *block) * be retransmitted. */ void -nlmsvc_retry_blocked(void) +nlmsvc_retry_blocked(struct svc_rqst *rqstp) { unsigned long timeout = MAX_SCHEDULE_TIMEOUT; struct nlm_block *block; spin_lock(&nlm_blocked_lock); - while (!list_empty(&nlm_blocked) && !kthread_should_stop()) { + while (!list_empty(&nlm_blocked) && !svc_thread_should_stop(rqstp)) { block = list_entry(nlm_blocked.next, struct nlm_block, b_list); if (block->b_when == NLM_NEVER) diff --git a/fs/locks.c b/fs/locks.c index 76ad05f8070ad9..d4e49a990a8daa 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -2264,11 +2264,13 @@ out: * To avoid blocking kernel daemons, such as lockd, that need to acquire POSIX * locks, the ->lock() interface may return asynchronously, before the lock has * been granted or denied by the underlying filesystem, if (and only if) - * lm_grant is set. Callers expecting ->lock() to return asynchronously - * will only use F_SETLK, not F_SETLKW; they will set FL_SLEEP if (and only if) - * the request is for a blocking lock. When ->lock() does return asynchronously, - * it must return FILE_LOCK_DEFERRED, and call ->lm_grant() when the lock - * request completes. + * lm_grant is set. Additionally EXPORT_OP_ASYNC_LOCK in export_operations + * flags need to be set. + * + * Callers expecting ->lock() to return asynchronously will only use F_SETLK, + * not F_SETLKW; they will set FL_SLEEP if (and only if) the request is for a + * blocking lock. When ->lock() does return asynchronously, it must return + * FILE_LOCK_DEFERRED, and call ->lm_grant() when the lock request completes. * If the request is for non-blocking lock the file system should return * FILE_LOCK_DEFERRED then try to get the lock and call the callback routine * with the result. If the request timed out the callback routine will return a diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index 466ebf1d41b2b7..4ffa1f469e90a9 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -78,7 +78,7 @@ nfs4_callback_svc(void *vrqstp) set_freezable(); - while (!kthread_freezable_should_stop(NULL)) + while (!svc_thread_should_stop(rqstp)) svc_recv(rqstp); svc_exit_thread(rqstp); @@ -86,45 +86,6 @@ nfs4_callback_svc(void *vrqstp) } #if defined(CONFIG_NFS_V4_1) -/* - * The callback service for NFSv4.1 callbacks - */ -static int -nfs41_callback_svc(void *vrqstp) -{ - struct svc_rqst *rqstp = vrqstp; - struct svc_serv *serv = rqstp->rq_server; - struct rpc_rqst *req; - int error; - DEFINE_WAIT(wq); - - set_freezable(); - - while (!kthread_freezable_should_stop(NULL)) { - prepare_to_wait(&serv->sv_cb_waitq, &wq, TASK_IDLE); - spin_lock_bh(&serv->sv_cb_lock); - if (!list_empty(&serv->sv_cb_list)) { - req = list_first_entry(&serv->sv_cb_list, - struct rpc_rqst, rq_bc_list); - list_del(&req->rq_bc_list); - spin_unlock_bh(&serv->sv_cb_lock); - finish_wait(&serv->sv_cb_waitq, &wq); - dprintk("Invoking bc_svc_process()\n"); - error = bc_svc_process(serv, req, rqstp); - dprintk("bc_svc_process() returned w/ error code= %d\n", - error); - } else { - spin_unlock_bh(&serv->sv_cb_lock); - if (!kthread_should_stop()) - schedule(); - finish_wait(&serv->sv_cb_waitq, &wq); - } - } - - svc_exit_thread(rqstp); - return 0; -} - static inline void nfs_callback_bc_serv(u32 minorversion, struct rpc_xprt *xprt, struct svc_serv *serv) { @@ -237,10 +198,7 @@ static struct svc_serv *nfs_callback_create_svc(int minorversion) cb_info->users); threadfn = nfs4_callback_svc; -#if defined(CONFIG_NFS_V4_1) - if (minorversion) - threadfn = nfs41_callback_svc; -#else +#if !defined(CONFIG_NFS_V4_1) if (minorversion) return ERR_PTR(-ENOTSUPP); #endif diff --git a/fs/nfsd/Makefile b/fs/nfsd/Makefile index 6fffc8f03f7408..b8736a82e57c47 100644 --- a/fs/nfsd/Makefile +++ b/fs/nfsd/Makefile @@ -12,7 +12,8 @@ nfsd-y += trace.o nfsd-y += nfssvc.o nfsctl.o nfsfh.o vfs.o \ export.o auth.o lockd.o nfscache.o \ - stats.o filecache.o nfs3proc.o nfs3xdr.o + stats.o filecache.o nfs3proc.o nfs3xdr.o \ + netlink.o nfsd-$(CONFIG_NFSD_V2) += nfsproc.o nfsxdr.o nfsd-$(CONFIG_NFSD_V2_ACL) += nfs2acl.o nfsd-$(CONFIG_NFSD_V3_ACL) += nfs3acl.o diff --git a/fs/nfsd/filecache.c b/fs/nfsd/filecache.c index ee9c923192e08a..07bf219f9ae482 100644 --- a/fs/nfsd/filecache.c +++ b/fs/nfsd/filecache.c @@ -989,22 +989,21 @@ nfsd_file_do_acquire(struct svc_rqst *rqstp, struct svc_fh *fhp, unsigned char need = may_flags & NFSD_FILE_MAY_MASK; struct net *net = SVC_NET(rqstp); struct nfsd_file *new, *nf; - const struct cred *cred; + bool stale_retry = true; bool open_retry = true; struct inode *inode; __be32 status; int ret; +retry: status = fh_verify(rqstp, fhp, S_IFREG, may_flags|NFSD_MAY_OWNER_OVERRIDE); if (status != nfs_ok) return status; inode = d_inode(fhp->fh_dentry); - cred = get_current_cred(); -retry: rcu_read_lock(); - nf = nfsd_file_lookup_locked(net, cred, inode, need, want_gc); + nf = nfsd_file_lookup_locked(net, current_cred(), inode, need, want_gc); rcu_read_unlock(); if (nf) { @@ -1026,7 +1025,7 @@ retry: rcu_read_lock(); spin_lock(&inode->i_lock); - nf = nfsd_file_lookup_locked(net, cred, inode, need, want_gc); + nf = nfsd_file_lookup_locked(net, current_cred(), inode, need, want_gc); if (unlikely(nf)) { spin_unlock(&inode->i_lock); rcu_read_unlock(); @@ -1058,6 +1057,7 @@ wait_for_construction: goto construction_err; } open_retry = false; + fh_put(fhp); goto retry; } this_cpu_inc(nfsd_file_cache_hits); @@ -1074,7 +1074,6 @@ out: nfsd_file_check_write_error(nf); *pnf = nf; } - put_cred(cred); trace_nfsd_file_acquire(rqstp, inode, may_flags, nf, status); return status; @@ -1088,8 +1087,20 @@ open_file: status = nfs_ok; trace_nfsd_file_opened(nf, status); } else { - status = nfsd_open_verified(rqstp, fhp, may_flags, - &nf->nf_file); + ret = nfsd_open_verified(rqstp, fhp, may_flags, + &nf->nf_file); + if (ret == -EOPENSTALE && stale_retry) { + stale_retry = false; + nfsd_file_unhash(nf); + clear_and_wake_up_bit(NFSD_FILE_PENDING, + &nf->nf_flags); + if (refcount_dec_and_test(&nf->nf_ref)) + nfsd_file_free(nf); + nf = NULL; + fh_put(fhp); + goto retry; + } + status = nfserrno(ret); trace_nfsd_file_open(nf, status); } } else diff --git a/fs/nfsd/netlink.c b/fs/nfsd/netlink.c new file mode 100644 index 00000000000000..0e1d635ec5f912 --- /dev/null +++ b/fs/nfsd/netlink.c @@ -0,0 +1,32 @@ +// SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) +/* Do not edit directly, auto-generated from: */ +/* Documentation/netlink/specs/nfsd.yaml */ +/* YNL-GEN kernel source */ + +#include <net/netlink.h> +#include <net/genetlink.h> + +#include "netlink.h" + +#include <uapi/linux/nfsd_netlink.h> + +/* Ops table for nfsd */ +static const struct genl_split_ops nfsd_nl_ops[] = { + { + .cmd = NFSD_CMD_RPC_STATUS_GET, + .start = nfsd_nl_rpc_status_get_start, + .dumpit = nfsd_nl_rpc_status_get_dumpit, + .done = nfsd_nl_rpc_status_get_done, + .flags = GENL_CMD_CAP_DUMP, + }, +}; + +struct genl_family nfsd_nl_family __ro_after_init = { + .name = NFSD_FAMILY_NAME, + .version = NFSD_FAMILY_VERSION, + .netnsok = true, + .parallel_ops = true, + .module = THIS_MODULE, + .split_ops = nfsd_nl_ops, + .n_split_ops = ARRAY_SIZE(nfsd_nl_ops), +}; diff --git a/fs/nfsd/netlink.h b/fs/nfsd/netlink.h new file mode 100644 index 00000000000000..d83dd6bdee927e --- /dev/null +++ b/fs/nfsd/netlink.h @@ -0,0 +1,22 @@ +/* SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) */ +/* Do not edit directly, auto-generated from: */ +/* Documentation/netlink/specs/nfsd.yaml */ +/* YNL-GEN kernel header */ + +#ifndef _LINUX_NFSD_GEN_H +#define _LINUX_NFSD_GEN_H + +#include <net/netlink.h> +#include <net/genetlink.h> + +#include <uapi/linux/nfsd_netlink.h> + +int nfsd_nl_rpc_status_get_start(struct netlink_callback *cb); +int nfsd_nl_rpc_status_get_done(struct netlink_callback *cb); + +int nfsd_nl_rpc_status_get_dumpit(struct sk_buff *skb, + struct netlink_callback *cb); + +extern struct genl_family nfsd_nl_family; + +#endif /* _LINUX_NFSD_GEN_H */ diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index 4039ffcf90ba58..92bc109dabe638 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -84,7 +84,21 @@ static void encode_uint32(struct xdr_stream *xdr, u32 n) static void encode_bitmap4(struct xdr_stream *xdr, const __u32 *bitmap, size_t len) { - WARN_ON_ONCE(xdr_stream_encode_uint32_array(xdr, bitmap, len) < 0); + xdr_stream_encode_uint32_array(xdr, bitmap, len); +} + +static int decode_cb_fattr4(struct xdr_stream *xdr, uint32_t *bitmap, + struct nfs4_cb_fattr *fattr) +{ + fattr->ncf_cb_change = 0; + fattr->ncf_cb_fsize = 0; + if (bitmap[0] & FATTR4_WORD0_CHANGE) + if (xdr_stream_decode_u64(xdr, &fattr->ncf_cb_change) < 0) + return -NFSERR_BAD_XDR; + if (bitmap[0] & FATTR4_WORD0_SIZE) + if (xdr_stream_decode_u64(xdr, &fattr->ncf_cb_fsize) < 0) + return -NFSERR_BAD_XDR; + return 0; } /* @@ -358,6 +372,30 @@ encode_cb_recallany4args(struct xdr_stream *xdr, } /* + * CB_GETATTR4args + * struct CB_GETATTR4args { + * nfs_fh4 fh; + * bitmap4 attr_request; + * }; + * + * The size and change attributes are the only one + * guaranteed to be serviced by the client. + */ +static void +encode_cb_getattr4args(struct xdr_stream *xdr, struct nfs4_cb_compound_hdr *hdr, + struct nfs4_cb_fattr *fattr) +{ + struct nfs4_delegation *dp = + container_of(fattr, struct nfs4_delegation, dl_cb_fattr); + struct knfsd_fh *fh = &dp->dl_stid.sc_file->fi_fhandle; + + encode_nfs_cb_opnum4(xdr, OP_CB_GETATTR); + encode_nfs_fh4(xdr, fh); + encode_bitmap4(xdr, fattr->ncf_cb_bmap, ARRAY_SIZE(fattr->ncf_cb_bmap)); + hdr->nops++; +} + +/* * CB_SEQUENCE4args * * struct CB_SEQUENCE4args { @@ -493,6 +531,26 @@ static void nfs4_xdr_enc_cb_null(struct rpc_rqst *req, struct xdr_stream *xdr, } /* + * 20.1. Operation 3: CB_GETATTR - Get Attributes + */ +static void nfs4_xdr_enc_cb_getattr(struct rpc_rqst *req, + struct xdr_stream *xdr, const void *data) +{ + const struct nfsd4_callback *cb = data; + struct nfs4_cb_fattr *ncf = + container_of(cb, struct nfs4_cb_fattr, ncf_getattr); + struct nfs4_cb_compound_hdr hdr = { + .ident = cb->cb_clp->cl_cb_ident, + .minorversion = cb->cb_clp->cl_minorversion, + }; + + encode_cb_compound4args(xdr, &hdr); + encode_cb_sequence4args(xdr, cb, &hdr); + encode_cb_getattr4args(xdr, &hdr, ncf); + encode_cb_nops(&hdr); +} + +/* * 20.2. Operation 4: CB_RECALL - Recall a Delegation */ static void nfs4_xdr_enc_cb_recall(struct rpc_rqst *req, struct xdr_stream *xdr, @@ -548,6 +606,42 @@ static int nfs4_xdr_dec_cb_null(struct rpc_rqst *req, struct xdr_stream *xdr, } /* + * 20.1. Operation 3: CB_GETATTR - Get Attributes + */ +static int nfs4_xdr_dec_cb_getattr(struct rpc_rqst *rqstp, + struct xdr_stream *xdr, + void *data) +{ + struct nfsd4_callback *cb = data; + struct nfs4_cb_compound_hdr hdr; + int status; + u32 bitmap[3] = {0}; + u32 attrlen; + struct nfs4_cb_fattr *ncf = + container_of(cb, struct nfs4_cb_fattr, ncf_getattr); + + status = decode_cb_compound4res(xdr, &hdr); + if (unlikely(status)) + return status; + + status = decode_cb_sequence4res(xdr, cb); + if (unlikely(status || cb->cb_seq_status)) + return status; + + status = decode_cb_op_status(xdr, OP_CB_GETATTR, &cb->cb_status); + if (status) + return status; + if (xdr_stream_decode_uint32_array(xdr, bitmap, 3) < 0) + return -NFSERR_BAD_XDR; + if (xdr_stream_decode_u32(xdr, &attrlen) < 0) + return -NFSERR_BAD_XDR; + if (attrlen > (sizeof(ncf->ncf_cb_change) + sizeof(ncf->ncf_cb_fsize))) + return -NFSERR_BAD_XDR; + status = decode_cb_fattr4(xdr, bitmap, ncf); + return status; +} + +/* * 20.2. Operation 4: CB_RECALL - Recall a Delegation */ static int nfs4_xdr_dec_cb_recall(struct rpc_rqst *rqstp, @@ -855,6 +949,7 @@ static const struct rpc_procinfo nfs4_cb_procedures[] = { PROC(CB_NOTIFY_LOCK, COMPOUND, cb_notify_lock, cb_notify_lock), PROC(CB_OFFLOAD, COMPOUND, cb_offload, cb_offload), PROC(CB_RECALL_ANY, COMPOUND, cb_recall_any, cb_recall_any), + PROC(CB_GETATTR, COMPOUND, cb_getattr, cb_getattr), }; static unsigned int nfs4_cb_counts[ARRAY_SIZE(nfs4_cb_procedures)]; diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index 4199ede0583c7d..d7e88c7beba32e 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -1329,7 +1329,8 @@ extern void nfs_sb_deactive(struct super_block *sb); * setup a work entry in the ssc delayed unmount list. */ static __be32 nfsd4_ssc_setup_dul(struct nfsd_net *nn, char *ipaddr, - struct nfsd4_ssc_umount_item **nsui) + struct nfsd4_ssc_umount_item **nsui, + struct svc_rqst *rqstp) { struct nfsd4_ssc_umount_item *ni = NULL; struct nfsd4_ssc_umount_item *work = NULL; @@ -1351,7 +1352,7 @@ try_again: spin_unlock(&nn->nfsd_ssc_lock); /* allow 20secs for mount/unmount for now - revisit */ - if (kthread_should_stop() || + if (svc_thread_should_stop(rqstp) || (schedule_timeout(20*HZ) == 0)) { finish_wait(&nn->nfsd_ssc_waitq, &wait); kfree(work); @@ -1467,7 +1468,7 @@ nfsd4_interssc_connect(struct nl4_server *nss, struct svc_rqst *rqstp, goto out_free_rawdata; snprintf(dev_name, len + 5, "%s%s%s:/", startsep, ipaddr, endsep); - status = nfsd4_ssc_setup_dul(nn, ipaddr, nsui); + status = nfsd4_ssc_setup_dul(nn, ipaddr, nsui, rqstp); if (status) goto out_free_devname; if ((*nsui)->nsui_vfsmount) @@ -1642,6 +1643,7 @@ static ssize_t _nfsd_copy_file_range(struct nfsd4_copy *copy, if (bytes_total == 0) bytes_total = ULLONG_MAX; do { + /* Only async copies can be stopped here */ if (kthread_should_stop()) break; bytes_copied = nfsd_copy_file_range(src, src_pos, dst, dst_pos, @@ -1760,6 +1762,7 @@ static int nfsd4_do_async_copy(void *data) struct nfsd4_copy *copy = (struct nfsd4_copy *)data; __be32 nfserr; + trace_nfsd_copy_do_async(copy); if (nfsd4_ssc_is_inter(copy)) { struct file *filp; @@ -1798,21 +1801,27 @@ nfsd4_copy(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, __be32 status; struct nfsd4_copy *async_copy = NULL; + copy->cp_clp = cstate->clp; if (nfsd4_ssc_is_inter(copy)) { + trace_nfsd_copy_inter(copy); if (!inter_copy_offload_enable || nfsd4_copy_is_sync(copy)) { status = nfserr_notsupp; goto out; } status = nfsd4_setup_inter_ssc(rqstp, cstate, copy); - if (status) + if (status) { + trace_nfsd_copy_done(copy, status); return nfserr_offload_denied; + } } else { + trace_nfsd_copy_intra(copy); status = nfsd4_setup_intra_ssc(rqstp, cstate, copy); - if (status) + if (status) { + trace_nfsd_copy_done(copy, status); return status; + } } - copy->cp_clp = cstate->clp; memcpy(©->fh, &cstate->current_fh.fh_handle, sizeof(struct knfsd_fh)); if (nfsd4_copy_is_async(copy)) { @@ -1847,6 +1856,7 @@ nfsd4_copy(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, copy->nf_dst->nf_file, true); } out: + trace_nfsd_copy_done(copy, status); release_copy_files(copy); return status; out_err: diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 8534693eb6a497..a39b62ef87d542 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -127,6 +127,7 @@ static void free_session(struct nfsd4_session *); static const struct nfsd4_callback_ops nfsd4_cb_recall_ops; static const struct nfsd4_callback_ops nfsd4_cb_notify_lock_ops; +static const struct nfsd4_callback_ops nfsd4_cb_getattr_ops; static struct workqueue_struct *laundry_wq; @@ -1187,6 +1188,10 @@ alloc_init_deleg(struct nfs4_client *clp, struct nfs4_file *fp, dp->dl_recalled = false; nfsd4_init_cb(&dp->dl_recall, dp->dl_stid.sc_client, &nfsd4_cb_recall_ops, NFSPROC4_CLNT_CB_RECALL); + nfsd4_init_cb(&dp->dl_cb_fattr.ncf_getattr, dp->dl_stid.sc_client, + &nfsd4_cb_getattr_ops, NFSPROC4_CLNT_CB_GETATTR); + dp->dl_cb_fattr.ncf_file_modified = false; + dp->dl_cb_fattr.ncf_cb_bmap[0] = FATTR4_WORD0_CHANGE | FATTR4_WORD0_SIZE; get_nfs4_file(fp); dp->dl_stid.sc_file = fp; return dp; @@ -2894,11 +2899,56 @@ nfsd4_cb_recall_any_release(struct nfsd4_callback *cb) spin_unlock(&nn->client_lock); } +static int +nfsd4_cb_getattr_done(struct nfsd4_callback *cb, struct rpc_task *task) +{ + struct nfs4_cb_fattr *ncf = + container_of(cb, struct nfs4_cb_fattr, ncf_getattr); + + ncf->ncf_cb_status = task->tk_status; + switch (task->tk_status) { + case -NFS4ERR_DELAY: + rpc_delay(task, 2 * HZ); + return 0; + default: + return 1; + } +} + +static void +nfsd4_cb_getattr_release(struct nfsd4_callback *cb) +{ + struct nfs4_cb_fattr *ncf = + container_of(cb, struct nfs4_cb_fattr, ncf_getattr); + struct nfs4_delegation *dp = + container_of(ncf, struct nfs4_delegation, dl_cb_fattr); + + nfs4_put_stid(&dp->dl_stid); + clear_bit(CB_GETATTR_BUSY, &ncf->ncf_cb_flags); + wake_up_bit(&ncf->ncf_cb_flags, CB_GETATTR_BUSY); +} + static const struct nfsd4_callback_ops nfsd4_cb_recall_any_ops = { .done = nfsd4_cb_recall_any_done, .release = nfsd4_cb_recall_any_release, }; +static const struct nfsd4_callback_ops nfsd4_cb_getattr_ops = { + .done = nfsd4_cb_getattr_done, + .release = nfsd4_cb_getattr_release, +}; + +void nfs4_cb_getattr(struct nfs4_cb_fattr *ncf) +{ + struct nfs4_delegation *dp = + container_of(ncf, struct nfs4_delegation, dl_cb_fattr); + + if (test_and_set_bit(CB_GETATTR_BUSY, &ncf->ncf_cb_flags)) + return; + refcount_inc(&dp->dl_stid.sc_count); + nfsd4_run_cb(&ncf->ncf_getattr); +} + static struct nfs4_client *create_client(struct xdr_netobj name, struct svc_rqst *rqstp, nfs4_verifier *verf) { @@ -5634,6 +5684,8 @@ nfs4_open_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp, struct svc_fh *parent = NULL; int cb_up; int status = 0; + struct kstat stat; + struct path path; cb_up = nfsd4_cb_channel_good(oo->oo_owner.so_client); open->op_recall = 0; @@ -5671,6 +5723,18 @@ nfs4_open_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp, if (open->op_share_access & NFS4_SHARE_ACCESS_WRITE) { open->op_delegate_type = NFS4_OPEN_DELEGATE_WRITE; trace_nfsd_deleg_write(&dp->dl_stid.sc_stateid); + path.mnt = currentfh->fh_export->ex_path.mnt; + path.dentry = currentfh->fh_dentry; + if (vfs_getattr(&path, &stat, + (STATX_SIZE | STATX_CTIME | STATX_CHANGE_COOKIE), + AT_STATX_SYNC_AS_STAT)) { + nfs4_put_stid(&dp->dl_stid); + destroy_delegation(dp); + goto out_no_deleg; + } + dp->dl_cb_fattr.ncf_cur_fsize = stat.size; + dp->dl_cb_fattr.ncf_initial_cinfo = + nfsd4_change_attribute(&stat, d_inode(currentfh->fh_dentry)); } else { open->op_delegate_type = NFS4_OPEN_DELEGATE_READ; trace_nfsd_deleg_read(&dp->dl_stid.sc_stateid); @@ -7487,6 +7551,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, struct nfsd4_blocked_lock *nbl = NULL; struct file_lock *file_lock = NULL; struct file_lock *conflock = NULL; + struct super_block *sb; __be32 status = 0; int lkflg; int err; @@ -7508,6 +7573,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, dprintk("NFSD: nfsd4_lock: permission denied!\n"); return status; } + sb = cstate->current_fh.fh_dentry->d_sb; if (lock->lk_is_new) { if (nfsd4_has_session(cstate)) @@ -7559,7 +7625,8 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, fp = lock_stp->st_stid.sc_file; switch (lock->lk_type) { case NFS4_READW_LT: - if (nfsd4_has_session(cstate)) + if (nfsd4_has_session(cstate) || + exportfs_lock_op_is_async(sb->s_export_op)) fl_flags |= FL_SLEEP; fallthrough; case NFS4_READ_LT: @@ -7571,7 +7638,8 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, fl_type = F_RDLCK; break; case NFS4_WRITEW_LT: - if (nfsd4_has_session(cstate)) + if (nfsd4_has_session(cstate) || + exportfs_lock_op_is_async(sb->s_export_op)) fl_flags |= FL_SLEEP; fallthrough; case NFS4_WRITE_LT: @@ -7599,7 +7667,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, * for file locks), so don't attempt blocking lock notifications * on those filesystems: */ - if (nf->nf_file->f_op->lock) + if (!exportfs_lock_op_is_async(sb->s_export_op)) fl_flags &= ~FL_SLEEP; nbl = find_or_allocate_block(lock_sop, &fp->fi_fhandle, nn); @@ -8403,6 +8471,8 @@ nfsd4_get_writestateid(struct nfsd4_compound_state *cstate, * nfsd4_deleg_getattr_conflict - Recall if GETATTR causes conflict * @rqstp: RPC transaction context * @inode: file to be checked for a conflict + * @modified: return true if file was modified + * @size: new size of file if modified is true * * This function is called when there is a conflict between a write * delegation and a change/size GETATTR from another client. The server @@ -8411,21 +8481,21 @@ nfsd4_get_writestateid(struct nfsd4_compound_state *cstate, * delegation before replying to the GETATTR. See RFC 8881 section * 18.7.4. * - * The current implementation does not support CB_GETATTR yet. However - * this can avoid recalling the delegation could be added in follow up - * work. - * * Returns 0 if there is no conflict; otherwise an nfs_stat * code is returned. */ __be32 -nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp, struct inode *inode) +nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp, struct inode *inode, + bool *modified, u64 *size) { __be32 status; struct file_lock_context *ctx; struct file_lock *fl; struct nfs4_delegation *dp; + struct iattr attrs; + struct nfs4_cb_fattr *ncf; + *modified = false; ctx = locks_inode_context(inode); if (!ctx) return 0; @@ -8452,10 +8522,34 @@ nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp, struct inode *inode) break_lease: spin_unlock(&ctx->flc_lock); nfsd_stats_wdeleg_getattr_inc(); - status = nfserrno(nfsd_open_break_lease(inode, NFSD_MAY_READ)); - if (status != nfserr_jukebox || - !nfsd_wait_for_delegreturn(rqstp, inode)) - return status; + + dp = fl->fl_owner; + ncf = &dp->dl_cb_fattr; + nfs4_cb_getattr(&dp->dl_cb_fattr); + wait_on_bit(&ncf->ncf_cb_flags, CB_GETATTR_BUSY, TASK_INTERRUPTIBLE); + if (ncf->ncf_cb_status) { + status = nfserrno(nfsd_open_break_lease(inode, NFSD_MAY_READ)); + if (status != nfserr_jukebox || + !nfsd_wait_for_delegreturn(rqstp, inode)) + return status; + } + if (!ncf->ncf_file_modified && + (ncf->ncf_initial_cinfo != ncf->ncf_cb_change || + ncf->ncf_cur_fsize != ncf->ncf_cb_fsize)) + ncf->ncf_file_modified = true; + if (ncf->ncf_file_modified) { + /* + * The server would not update the file's metadata + * with the client's modified size. + */ + attrs.ia_mtime = attrs.ia_ctime = current_time(inode); + attrs.ia_valid = ATTR_MTIME | ATTR_CTIME; + setattr_copy(&nop_mnt_idmap, inode, &attrs); + mark_inode_dirty(inode); + ncf->ncf_cur_fsize = ncf->ncf_cb_fsize; + *size = ncf->ncf_cur_fsize; + *modified = true; + } return 0; } break; diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 2e40c74d2f7278..69504da6c4d4e7 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -2975,6 +2975,8 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp, .dentry = dentry, }; struct nfsd_net *nn = net_generic(SVC_NET(rqstp), nfsd_net_id); + bool file_modified; + u64 size = 0; BUG_ON(bmval1 & NFSD_WRITEONLY_ATTRS_WORD1); BUG_ON(!nfsd_attrs_supported(minorversion, bmval)); @@ -2985,7 +2987,8 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp, goto out; } if (bmval0 & (FATTR4_WORD0_CHANGE | FATTR4_WORD0_SIZE)) { - status = nfsd4_deleg_getattr_conflict(rqstp, d_inode(dentry)); + status = nfsd4_deleg_getattr_conflict(rqstp, d_inode(dentry), + &file_modified, &size); if (status) goto out; } @@ -3112,7 +3115,10 @@ nfsd4_encode_fattr(struct xdr_stream *xdr, struct svc_fh *fhp, p = xdr_reserve_space(xdr, 8); if (!p) goto out_resource; - p = xdr_encode_hyper(p, stat.size); + if (file_modified) + p = xdr_encode_hyper(p, size); + else + p = xdr_encode_hyper(p, stat.size); } if (bmval0 & FATTR4_WORD0_LINK_SUPPORT) { p = xdr_reserve_space(xdr, 4); diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 7ed02fb88a362c..b71744e355a86d 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -26,6 +26,7 @@ #include "pnfs.h" #include "filecache.h" #include "trace.h" +#include "netlink.h" /* * We have a single directory with several nodes in it. @@ -1496,6 +1497,204 @@ static int create_proc_exports_entry(void) unsigned int nfsd_net_id; /** + * nfsd_nl_rpc_status_get_start - Prepare rpc_status_get dumpit + * @cb: netlink metadata and command arguments + * + * Return values: + * %0: The rpc_status_get command may proceed + * %-ENODEV: There is no NFSD running in this namespace + */ +int nfsd_nl_rpc_status_get_start(struct netlink_callback *cb) +{ + struct nfsd_net *nn = net_generic(sock_net(cb->skb->sk), nfsd_net_id); + int ret = -ENODEV; + + mutex_lock(&nfsd_mutex); + if (nn->nfsd_serv) { + svc_get(nn->nfsd_serv); + ret = 0; + } + mutex_unlock(&nfsd_mutex); + + return ret; +} + +static int nfsd_genl_rpc_status_compose_msg(struct sk_buff *skb, + struct netlink_callback *cb, + struct nfsd_genl_rqstp *rqstp) +{ + void *hdr; + u32 i; + + hdr = genlmsg_put(skb, NETLINK_CB(cb->skb).portid, cb->nlh->nlmsg_seq, + &nfsd_nl_family, NLM_F_MULTI, + NFSD_CMD_RPC_STATUS_GET); + if (!hdr) + return -ENOBUFS; + + if (nla_put_be32(skb, NFSD_A_RPC_STATUS_XID, rqstp->rq_xid) || + nla_put_u32(skb, NFSD_A_RPC_STATUS_FLAGS, rqstp->rq_flags) || + nla_put_u32(skb, NFSD_A_RPC_STATUS_PROG, rqstp->rq_prog) || + nla_put_u32(skb, NFSD_A_RPC_STATUS_PROC, rqstp->rq_proc) || + nla_put_u8(skb, NFSD_A_RPC_STATUS_VERSION, rqstp->rq_vers) || + nla_put_s64(skb, NFSD_A_RPC_STATUS_SERVICE_TIME, + ktime_to_us(rqstp->rq_stime), + NFSD_A_RPC_STATUS_PAD)) + return -ENOBUFS; + + switch (rqstp->rq_saddr.sa_family) { + case AF_INET: { + const struct sockaddr_in *s_in, *d_in; + + s_in = (const struct sockaddr_in *)&rqstp->rq_saddr; + d_in = (const struct sockaddr_in *)&rqstp->rq_daddr; + if (nla_put_in_addr(skb, NFSD_A_RPC_STATUS_SADDR4, + s_in->sin_addr.s_addr) || + nla_put_in_addr(skb, NFSD_A_RPC_STATUS_DADDR4, + d_in->sin_addr.s_addr) || + nla_put_be16(skb, NFSD_A_RPC_STATUS_SPORT, + s_in->sin_port) || + nla_put_be16(skb, NFSD_A_RPC_STATUS_DPORT, + d_in->sin_port)) + return -ENOBUFS; + break; + } + case AF_INET6: { + const struct sockaddr_in6 *s_in, *d_in; + + s_in = (const struct sockaddr_in6 *)&rqstp->rq_saddr; + d_in = (const struct sockaddr_in6 *)&rqstp->rq_daddr; + if (nla_put_in6_addr(skb, NFSD_A_RPC_STATUS_SADDR6, + &s_in->sin6_addr) || + nla_put_in6_addr(skb, NFSD_A_RPC_STATUS_DADDR6, + &d_in->sin6_addr) || + nla_put_be16(skb, NFSD_A_RPC_STATUS_SPORT, + s_in->sin6_port) || + nla_put_be16(skb, NFSD_A_RPC_STATUS_DPORT, + d_in->sin6_port)) + return -ENOBUFS; + break; + } + } + + for (i = 0; i < rqstp->rq_opcnt; i++) + if (nla_put_u32(skb, NFSD_A_RPC_STATUS_COMPOUND_OPS, + rqstp->rq_opnum[i])) + return -ENOBUFS; + + genlmsg_end(skb, hdr); + return 0; +} + +/** + * nfsd_nl_rpc_status_get_dumpit - Handle rpc_status_get dumpit + * @skb: reply buffer + * @cb: netlink metadata and command arguments + * + * Returns the size of the reply or a negative errno. + */ +int nfsd_nl_rpc_status_get_dumpit(struct sk_buff *skb, + struct netlink_callback *cb) +{ + struct nfsd_net *nn = net_generic(sock_net(skb->sk), nfsd_net_id); + int i, ret, rqstp_index = 0; + + rcu_read_lock(); + + for (i = 0; i < nn->nfsd_serv->sv_nrpools; i++) { + struct svc_rqst *rqstp; + + if (i < cb->args[0]) /* already consumed */ + continue; + + rqstp_index = 0; + list_for_each_entry_rcu(rqstp, + &nn->nfsd_serv->sv_pools[i].sp_all_threads, + rq_all) { + struct nfsd_genl_rqstp genl_rqstp; + unsigned int status_counter; + + if (rqstp_index++ < cb->args[1]) /* already consumed */ + continue; + /* + * Acquire rq_status_counter before parsing the rqst + * fields. rq_status_counter is set to an odd value in + * order to notify the consumers the rqstp fields are + * meaningful. + */ + status_counter = + smp_load_acquire(&rqstp->rq_status_counter); + if (!(status_counter & 1)) + continue; + + genl_rqstp.rq_xid = rqstp->rq_xid; + genl_rqstp.rq_flags = rqstp->rq_flags; + genl_rqstp.rq_vers = rqstp->rq_vers; + genl_rqstp.rq_prog = rqstp->rq_prog; + genl_rqstp.rq_proc = rqstp->rq_proc; + genl_rqstp.rq_stime = rqstp->rq_stime; + genl_rqstp.rq_opcnt = 0; + memcpy(&genl_rqstp.rq_daddr, svc_daddr(rqstp), + sizeof(struct sockaddr)); + memcpy(&genl_rqstp.rq_saddr, svc_addr(rqstp), + sizeof(struct sockaddr)); + +#ifdef CONFIG_NFSD_V4 + if (rqstp->rq_vers == NFS4_VERSION && + rqstp->rq_proc == NFSPROC4_COMPOUND) { + /* NFSv4 compound */ + struct nfsd4_compoundargs *args; + int j; + + args = rqstp->rq_argp; + genl_rqstp.rq_opcnt = args->opcnt; + for (j = 0; j < genl_rqstp.rq_opcnt; j++) + genl_rqstp.rq_opnum[j] = + args->ops[j].opnum; + } +#endif /* CONFIG_NFSD_V4 */ + + /* + * Acquire rq_status_counter before reporting the rqst + * fields to the user. + */ + if (smp_load_acquire(&rqstp->rq_status_counter) != + status_counter) + continue; + + ret = nfsd_genl_rpc_status_compose_msg(skb, cb, + &genl_rqstp); + if (ret) + goto out; + } + } + + cb->args[0] = i; + cb->args[1] = rqstp_index; + ret = skb->len; +out: + rcu_read_unlock(); + + return ret; +} + +/** + * nfsd_nl_rpc_status_get_done - rpc_status_get dumpit post-processing + * @cb: netlink metadata and command arguments + * + * Return values: + * %0: Success + */ +int nfsd_nl_rpc_status_get_done(struct netlink_callback *cb) +{ + mutex_lock(&nfsd_mutex); + nfsd_put(sock_net(cb->skb->sk)); + mutex_unlock(&nfsd_mutex); + + return 0; +} + +/** * nfsd_net_init - Prepare the nfsd_net portion of a new net namespace * @net: a freshly-created network namespace * @@ -1589,6 +1788,10 @@ static int __init init_nfsd(void) retval = register_filesystem(&nfsd_fs_type); if (retval) goto out_free_all; + retval = genl_register_family(&nfsd_nl_family); + if (retval) + goto out_free_all; + return 0; out_free_all: nfsd4_destroy_laundry_wq(); @@ -1613,6 +1816,7 @@ out_free_slabs: static void __exit exit_nfsd(void) { + genl_unregister_family(&nfsd_nl_family); unregister_filesystem(&nfsd_fs_type); nfsd4_destroy_laundry_wq(); unregister_cld_notifier(); diff --git a/fs/nfsd/nfsd.h b/fs/nfsd/nfsd.h index 11c14faa6c67be..f5ff42f41ee7fe 100644 --- a/fs/nfsd/nfsd.h +++ b/fs/nfsd/nfsd.h @@ -62,6 +62,23 @@ struct readdir_cd { __be32 err; /* 0, nfserr, or nfserr_eof */ }; +/* Maximum number of operations per session compound */ +#define NFSD_MAX_OPS_PER_COMPOUND 50 + +struct nfsd_genl_rqstp { + struct sockaddr rq_daddr; + struct sockaddr rq_saddr; + unsigned long rq_flags; + ktime_t rq_stime; + __be32 rq_xid; + u32 rq_vers; + u32 rq_prog; + u32 rq_proc; + + /* NFSv4 compound */ + u32 rq_opcnt; + u32 rq_opnum[NFSD_MAX_OPS_PER_COMPOUND]; +}; extern struct svc_program nfsd_program; extern const struct svc_version nfsd_version2, nfsd_version3, nfsd_version4; diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c index c7af1095f6b549..c5890cdfe97bd0 100644 --- a/fs/nfsd/nfssvc.c +++ b/fs/nfsd/nfssvc.c @@ -713,14 +713,13 @@ int nfsd_nrpools(struct net *net) int nfsd_get_nrthreads(int n, int *nthreads, struct net *net) { - int i = 0; struct nfsd_net *nn = net_generic(net, nfsd_net_id); + struct svc_serv *serv = nn->nfsd_serv; + int i; - if (nn->nfsd_serv != NULL) { - for (i = 0; i < nn->nfsd_serv->sv_nrpools && i < n; i++) - nthreads[i] = nn->nfsd_serv->sv_pools[i].sp_nrthreads; - } - + if (serv) + for (i = 0; i < serv->sv_nrpools && i < n; i++) + nthreads[i] = atomic_read(&serv->sv_pools[i].sp_nrthreads); return 0; } @@ -957,7 +956,7 @@ nfsd(void *vrqstp) /* * The main request loop */ - while (!kthread_should_stop()) { + while (!svc_thread_should_stop(rqstp)) { /* Update sv_maxconn if it has changed */ rqstp->rq_server->sv_maxconn = nn->max_connections; @@ -998,6 +997,15 @@ int nfsd_dispatch(struct svc_rqst *rqstp) if (!proc->pc_decode(rqstp, &rqstp->rq_arg_stream)) goto out_decode_err; + /* + * Release rq_status_counter setting it to an odd value after the rpc + * request has been properly parsed. rq_status_counter is used to + * notify the consumers if the rqstp fields are stable + * (rq_status_counter is odd) or not meaningful (rq_status_counter + * is even). + */ + smp_store_release(&rqstp->rq_status_counter, rqstp->rq_status_counter | 1); + rp = NULL; switch (nfsd_cache_lookup(rqstp, &rp)) { case RC_DOIT: @@ -1015,6 +1023,12 @@ int nfsd_dispatch(struct svc_rqst *rqstp) if (!proc->pc_encode(rqstp, &rqstp->rq_res_stream)) goto out_encode_err; + /* + * Release rq_status_counter setting it to an even value after the rpc + * request has been properly processed. + */ + smp_store_release(&rqstp->rq_status_counter, rqstp->rq_status_counter + 1); + nfsd_cache_update(rqstp, rp, rqstp->rq_cachetype, statp + 1); out_cached_reply: return 1; diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h index cbddcf484dbac7..f96eaa8e9413da 100644 --- a/fs/nfsd/state.h +++ b/fs/nfsd/state.h @@ -117,6 +117,24 @@ struct nfs4_cpntf_state { time64_t cpntf_time; /* last time stateid used */ }; +struct nfs4_cb_fattr { + struct nfsd4_callback ncf_getattr; + u32 ncf_cb_status; + u32 ncf_cb_bmap[1]; + + /* from CB_GETATTR reply */ + u64 ncf_cb_change; + u64 ncf_cb_fsize; + + unsigned long ncf_cb_flags; + bool ncf_file_modified; + u64 ncf_initial_cinfo; + u64 ncf_cur_fsize; +}; + +/* bits for ncf_cb_flags */ +#define CB_GETATTR_BUSY 0 + /* * Represents a delegation stateid. The nfs4_client holds references to these * and they are put when it is being destroyed or when the delegation is @@ -150,6 +168,9 @@ struct nfs4_delegation { int dl_retries; struct nfsd4_callback dl_recall; bool dl_recalled; + + /* for CB_GETATTR */ + struct nfs4_cb_fattr dl_cb_fattr; }; #define cb_to_delegation(cb) \ @@ -174,8 +195,6 @@ static inline struct nfs4_delegation *delegstateid(struct nfs4_stid *s) /* Maximum number of slots per session. 160 is useful for long haul TCP */ #define NFSD_MAX_SLOTS_PER_SESSION 160 -/* Maximum number of operations per session compound */ -#define NFSD_MAX_OPS_PER_COMPOUND 50 /* Maximum session per slot cache size */ #define NFSD_SLOT_CACHE_SIZE 2048 /* Maximum number of NFSD_SLOT_CACHE_SIZE slots per session */ @@ -642,6 +661,7 @@ enum nfsd4_cb_op { NFSPROC4_CLNT_CB_SEQUENCE, NFSPROC4_CLNT_CB_NOTIFY_LOCK, NFSPROC4_CLNT_CB_RECALL_ANY, + NFSPROC4_CLNT_CB_GETATTR, }; /* Returns true iff a is later than b: */ @@ -734,5 +754,6 @@ static inline bool try_to_expire_client(struct nfs4_client *clp) } extern __be32 nfsd4_deleg_getattr_conflict(struct svc_rqst *rqstp, - struct inode *inode); + struct inode *inode, bool *file_modified, u64 *size); +extern void nfs4_cb_getattr(struct nfs4_cb_fattr *ncf); #endif /* NFSD4_STATE_H */ diff --git a/fs/nfsd/trace.h b/fs/nfsd/trace.h index 80390434887183..fbc0ccb404241f 100644 --- a/fs/nfsd/trace.h +++ b/fs/nfsd/trace.h @@ -1863,6 +1863,93 @@ TRACE_EVENT(nfsd_end_grace, ) ); +DECLARE_EVENT_CLASS(nfsd_copy_class, + TP_PROTO( + const struct nfsd4_copy *copy + ), + TP_ARGS(copy), + TP_STRUCT__entry( + __field(bool, intra) + __field(bool, async) + __field(u32, src_cl_boot) + __field(u32, src_cl_id) + __field(u32, src_so_id) + __field(u32, src_si_generation) + __field(u32, dst_cl_boot) + __field(u32, dst_cl_id) + __field(u32, dst_so_id) + __field(u32, dst_si_generation) + __field(u64, src_cp_pos) + __field(u64, dst_cp_pos) + __field(u64, cp_count) + __sockaddr(addr, sizeof(struct sockaddr_in6)) + ), + TP_fast_assign( + const stateid_t *src_stp = ©->cp_src_stateid; + const stateid_t *dst_stp = ©->cp_dst_stateid; + + __entry->intra = test_bit(NFSD4_COPY_F_INTRA, ©->cp_flags); + __entry->async = !test_bit(NFSD4_COPY_F_SYNCHRONOUS, ©->cp_flags); + __entry->src_cl_boot = src_stp->si_opaque.so_clid.cl_boot; + __entry->src_cl_id = src_stp->si_opaque.so_clid.cl_id; + __entry->src_so_id = src_stp->si_opaque.so_id; + __entry->src_si_generation = src_stp->si_generation; + __entry->dst_cl_boot = dst_stp->si_opaque.so_clid.cl_boot; + __entry->dst_cl_id = dst_stp->si_opaque.so_clid.cl_id; + __entry->dst_so_id = dst_stp->si_opaque.so_id; + __entry->dst_si_generation = dst_stp->si_generation; + __entry->src_cp_pos = copy->cp_src_pos; + __entry->dst_cp_pos = copy->cp_dst_pos; + __entry->cp_count = copy->cp_count; + __assign_sockaddr(addr, ©->cp_clp->cl_addr, + sizeof(struct sockaddr_in6)); + ), + TP_printk("client=%pISpc intra=%d async=%d " + "src_stateid[si_generation:0x%x cl_boot:0x%x cl_id:0x%x so_id:0x%x] " + "dst_stateid[si_generation:0x%x cl_boot:0x%x cl_id:0x%x so_id:0x%x] " + "cp_src_pos=%llu cp_dst_pos=%llu cp_count=%llu", + __get_sockaddr(addr), __entry->intra, __entry->async, + __entry->src_si_generation, __entry->src_cl_boot, + __entry->src_cl_id, __entry->src_so_id, + __entry->dst_si_generation, __entry->dst_cl_boot, + __entry->dst_cl_id, __entry->dst_so_id, + __entry->src_cp_pos, __entry->dst_cp_pos, __entry->cp_count + ) +); + +#define DEFINE_COPY_EVENT(name) \ +DEFINE_EVENT(nfsd_copy_class, nfsd_copy_##name, \ + TP_PROTO(const struct nfsd4_copy *copy), \ + TP_ARGS(copy)) + +DEFINE_COPY_EVENT(inter); +DEFINE_COPY_EVENT(intra); +DEFINE_COPY_EVENT(do_async); + +TRACE_EVENT(nfsd_copy_done, + TP_PROTO( + const struct nfsd4_copy *copy, + __be32 status + ), + TP_ARGS(copy, status), + TP_STRUCT__entry( + __field(int, status) + __field(bool, intra) + __field(bool, async) + __sockaddr(addr, sizeof(struct sockaddr_in6)) + ), + TP_fast_assign( + __entry->status = be32_to_cpu(status); + __entry->intra = test_bit(NFSD4_COPY_F_INTRA, ©->cp_flags); + __entry->async = !test_bit(NFSD4_COPY_F_SYNCHRONOUS, ©->cp_flags); + __assign_sockaddr(addr, ©->cp_clp->cl_addr, + sizeof(struct sockaddr_in6)); + ), + TP_printk("addr=%pISpc status=%d intra=%d async=%d ", + __get_sockaddr(addr), __entry->status, __entry->intra, __entry->async + ) +); + #endif /* _NFSD_TRACE_H */ #undef TRACE_INCLUDE_PATH diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 48260cf68fde8b..5bf3cffde83125 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -337,6 +337,24 @@ out: return err; } +static void +commit_reset_write_verifier(struct nfsd_net *nn, struct svc_rqst *rqstp, + int err) +{ + switch (err) { + case -EAGAIN: + case -ESTALE: + /* + * Neither of these are the result of a problem with + * durable storage, so avoid a write verifier reset. + */ + break; + default: + nfsd_reset_write_verifier(nn); + trace_nfsd_writeverf_reset(nn, rqstp, err); + } +} + /* * Commit metadata changes to stable storage. */ @@ -647,8 +665,7 @@ __be32 nfsd4_clone_file_range(struct svc_rqst *rqstp, &nfsd4_get_cstate(rqstp)->current_fh, dst_pos, count, status); - nfsd_reset_write_verifier(nn); - trace_nfsd_writeverf_reset(nn, rqstp, status); + commit_reset_write_verifier(nn, rqstp, status); ret = nfserrno(status); } } @@ -823,7 +840,7 @@ int nfsd_open_break_lease(struct inode *inode, int access) * and additional flags. * N.B. After this call fhp needs an fh_put */ -static __be32 +static int __nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, int may_flags, struct file **filp) { @@ -831,14 +848,12 @@ __nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, struct inode *inode; struct file *file; int flags = O_RDONLY|O_LARGEFILE; - __be32 err; - int host_err = 0; + int host_err = -EPERM; path.mnt = fhp->fh_export->ex_path.mnt; path.dentry = fhp->fh_dentry; inode = d_inode(path.dentry); - err = nfserr_perm; if (IS_APPEND(inode) && (may_flags & NFSD_MAY_WRITE)) goto out; @@ -847,7 +862,7 @@ __nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, host_err = nfsd_open_break_lease(inode, may_flags); if (host_err) /* NOMEM or WOULDBLOCK */ - goto out_nfserr; + goto out; if (may_flags & NFSD_MAY_WRITE) { if (may_flags & NFSD_MAY_READ) @@ -859,13 +874,13 @@ __nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, file = dentry_open(&path, flags, current_cred()); if (IS_ERR(file)) { host_err = PTR_ERR(file); - goto out_nfserr; + goto out; } host_err = ima_file_check(file, may_flags); if (host_err) { fput(file); - goto out_nfserr; + goto out; } if (may_flags & NFSD_MAY_64BIT_COOKIE) @@ -874,10 +889,8 @@ __nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, file->f_mode |= FMODE_32BITHASH; *filp = file; -out_nfserr: - err = nfserrno(host_err); out: - return err; + return host_err; } __be32 @@ -885,6 +898,7 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, int may_flags, struct file **filp) { __be32 err; + int host_err; bool retried = false; validate_process_creds(); @@ -904,12 +918,13 @@ nfsd_open(struct svc_rqst *rqstp, struct svc_fh *fhp, umode_t type, retry: err = fh_verify(rqstp, fhp, type, may_flags); if (!err) { - err = __nfsd_open(rqstp, fhp, type, may_flags, filp); - if (err == nfserr_stale && !retried) { + host_err = __nfsd_open(rqstp, fhp, type, may_flags, filp); + if (host_err == -EOPENSTALE && !retried) { retried = true; fh_put(fhp); goto retry; } + err = nfserrno(host_err); } validate_process_creds(); return err; @@ -922,13 +937,13 @@ retry: * @may_flags: internal permission flags * @filp: OUT: open "struct file *" * - * Returns an nfsstat value in network byte order. + * Returns zero on success, or a negative errno value. */ -__be32 +int nfsd_open_verified(struct svc_rqst *rqstp, struct svc_fh *fhp, int may_flags, struct file **filp) { - __be32 err; + int err; validate_process_creds(); err = __nfsd_open(rqstp, fhp, S_IFREG, may_flags, filp); @@ -1172,8 +1187,7 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct nfsd_file *nf, host_err = vfs_iter_write(file, &iter, &pos, flags); file_end_write(file); if (host_err < 0) { - nfsd_reset_write_verifier(nn); - trace_nfsd_writeverf_reset(nn, rqstp, host_err); + commit_reset_write_verifier(nn, rqstp, host_err); goto out_nfserr; } *cnt = host_err; @@ -1185,10 +1199,8 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct nfsd_file *nf, if (stable && use_wgather) { host_err = wait_for_concurrent_writes(file); - if (host_err < 0) { - nfsd_reset_write_verifier(nn); - trace_nfsd_writeverf_reset(nn, rqstp, host_err); - } + if (host_err < 0) + commit_reset_write_verifier(nn, rqstp, host_err); } out_nfserr: @@ -1331,8 +1343,7 @@ nfsd_commit(struct svc_rqst *rqstp, struct svc_fh *fhp, struct nfsd_file *nf, err = nfserr_notsupp; break; default: - nfsd_reset_write_verifier(nn); - trace_nfsd_writeverf_reset(nn, rqstp, err2); + commit_reset_write_verifier(nn, rqstp, err2); err = nfserrno(err2); } } else diff --git a/fs/nfsd/vfs.h b/fs/nfsd/vfs.h index a6890ea7b765b6..e3c29596f4df12 100644 --- a/fs/nfsd/vfs.h +++ b/fs/nfsd/vfs.h @@ -104,8 +104,8 @@ __be32 nfsd_setxattr(struct svc_rqst *rqstp, struct svc_fh *fhp, int nfsd_open_break_lease(struct inode *, int); __be32 nfsd_open(struct svc_rqst *, struct svc_fh *, umode_t, int, struct file **); -__be32 nfsd_open_verified(struct svc_rqst *, struct svc_fh *, - int, struct file **); +int nfsd_open_verified(struct svc_rqst *rqstp, struct svc_fh *fhp, + int may_flags, struct file **filp); __be32 nfsd_splice_read(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, loff_t offset, unsigned long *count, diff --git a/fs/nfsd/xdr4cb.h b/fs/nfsd/xdr4cb.h index 0d39af1b00a0f3..e8b00309c449fe 100644 --- a/fs/nfsd/xdr4cb.h +++ b/fs/nfsd/xdr4cb.h @@ -54,3 +54,21 @@ #define NFS4_dec_cb_recall_any_sz (cb_compound_dec_hdr_sz + \ cb_sequence_dec_sz + \ op_dec_sz) + +/* + * 1: CB_GETATTR opcode (32-bit) + * N: file_handle + * 1: number of entry in attribute array (32-bit) + * 1: entry 0 in attribute array (32-bit) + */ +#define NFS4_enc_cb_getattr_sz (cb_compound_enc_hdr_sz + \ + cb_sequence_enc_sz + \ + 1 + enc_nfs4_fh_sz + 1 + 1) +/* + * 4: fattr_bitmap_maxsz + * 1: attribute array len + * 2: change attr (64-bit) + * 2: size (64-bit) + */ +#define NFS4_dec_cb_getattr_sz (cb_compound_dec_hdr_sz + \ + cb_sequence_dec_sz + 4 + 1 + 2 + 2 + op_dec_sz) diff --git a/include/linux/exportfs.h b/include/linux/exportfs.h index 11fbd0ee137080..6dd993240fccc3 100644 --- a/include/linux/exportfs.h +++ b/include/linux/exportfs.h @@ -224,9 +224,23 @@ struct export_operations { atomic attribute updates */ #define EXPORT_OP_FLUSH_ON_CLOSE (0x20) /* fs flushes file data on close */ +#define EXPORT_OP_ASYNC_LOCK (0x40) /* fs can do async lock request */ unsigned long flags; }; +/** + * exportfs_lock_op_is_async() - export op supports async lock operation + * @export_ops: the nfs export operations to check + * + * Returns true if the nfs export_operations structure has + * EXPORT_OP_ASYNC_LOCK in their flags set + */ +static inline bool +exportfs_lock_op_is_async(const struct export_operations *export_ops) +{ + return export_ops->flags & EXPORT_OP_ASYNC_LOCK; +} + extern int exportfs_encode_inode_fh(struct inode *inode, struct fid *fid, int *max_len, struct inode *parent, int flags); diff --git a/include/linux/llist.h b/include/linux/llist.h index 85bda2d02d65be..2c982ff7475ab0 100644 --- a/include/linux/llist.h +++ b/include/linux/llist.h @@ -74,6 +74,33 @@ static inline void init_llist_head(struct llist_head *list) } /** + * init_llist_node - initialize lock-less list node + * @node: the node to be initialised + * + * In cases where there is a need to test if a node is on + * a list or not, this initialises the node to clearly + * not be on any list. + */ +static inline void init_llist_node(struct llist_node *node) +{ + node->next = node; +} + +/** + * llist_on_list - test if a lock-list list node is on a list + * @node: the node to test + * + * When a node is on a list the ->next pointer will be NULL or + * some other node. It can never point to itself. We use that + * in init_llist_node() to record that a node is not on any list, + * and here to test whether it is on any list. + */ +static inline bool llist_on_list(const struct llist_node *node) +{ + return node->next != node; +} + +/** * llist_entry - get the struct of this entry * @ptr: the &struct llist_node pointer. * @type: the type of the struct this is embedded in. @@ -249,6 +276,25 @@ static inline struct llist_node *__llist_del_all(struct llist_head *head) extern struct llist_node *llist_del_first(struct llist_head *head); +/** + * llist_del_first_init - delete first entry from lock-list and mark is as being off-list + * @head: the head of lock-less list to delete from. + * + * This behave the same as llist_del_first() except that llist_init_node() is called + * on the returned node so that llist_on_list() will report false for the node. + */ +static inline struct llist_node *llist_del_first_init(struct llist_head *head) +{ + struct llist_node *n = llist_del_first(head); + + if (n) + init_llist_node(n); + return n; +} + +extern bool llist_del_first_this(struct llist_head *head, + struct llist_node *this); + struct llist_node *llist_reverse_order(struct llist_node *head); #endif /* LLIST_H */ diff --git a/include/linux/lockd/lockd.h b/include/linux/lockd/lockd.h index 0f016d69c996e5..9f565416d18671 100644 --- a/include/linux/lockd/lockd.h +++ b/include/linux/lockd/lockd.h @@ -282,7 +282,7 @@ __be32 nlmsvc_testlock(struct svc_rqst *, struct nlm_file *, struct nlm_host *, struct nlm_lock *, struct nlm_lock *, struct nlm_cookie *); __be32 nlmsvc_cancel_blocked(struct net *net, struct nlm_file *, struct nlm_lock *); -void nlmsvc_retry_blocked(void); +void nlmsvc_retry_blocked(struct svc_rqst *rqstp); void nlmsvc_traverse_blocks(struct nlm_host *, struct nlm_file *, nlm_host_match_fn_t match); void nlmsvc_grant_reply(struct nlm_cookie *, __be32); diff --git a/include/linux/lwq.h b/include/linux/lwq.h new file mode 100644 index 00000000000000..d081d5cf8e33fa --- /dev/null +++ b/include/linux/lwq.h @@ -0,0 +1,124 @@ +/* SPDX-License-Identifier: GPL-2.0-only */ + +#ifndef LWQ_H +#define LWQ_H +/* + * Light-weight single-linked queue built from llist + * + * Entries can be enqueued from any context with no locking. + * Entries can be dequeued from process context with integrated locking. + * + * This is particularly suitable when work items are queued in + * BH or IRQ context, and where work items are handled one at a time + * by dedicated threads. + */ +#include <linux/container_of.h> +#include <linux/spinlock.h> +#include <linux/llist.h> + +struct lwq_node { + struct llist_node node; +}; + +struct lwq { + spinlock_t lock; + struct llist_node *ready; /* entries to be dequeued */ + struct llist_head new; /* entries being enqueued */ +}; + +/** + * lwq_init - initialise a lwq + * @q: the lwq object + */ +static inline void lwq_init(struct lwq *q) +{ + spin_lock_init(&q->lock); + q->ready = NULL; + init_llist_head(&q->new); +} + +/** + * lwq_empty - test if lwq contains any entry + * @q: the lwq object + * + * This empty test contains an acquire barrier so that if a wakeup + * is sent when lwq_dequeue returns true, it is safe to go to sleep after + * a test on lwq_empty(). + */ +static inline bool lwq_empty(struct lwq *q) +{ + /* acquire ensures ordering wrt lwq_enqueue() */ + return smp_load_acquire(&q->ready) == NULL && llist_empty(&q->new); +} + +struct llist_node *__lwq_dequeue(struct lwq *q); +/** + * lwq_dequeue - dequeue first (oldest) entry from lwq + * @q: the queue to dequeue from + * @type: the type of object to return + * @member: them member in returned object which is an lwq_node. + * + * Remove a single object from the lwq and return it. This will take + * a spinlock and so must always be called in the same context, typcially + * process contet. + */ +#define lwq_dequeue(q, type, member) \ + ({ struct llist_node *_n = __lwq_dequeue(q); \ + _n ? container_of(_n, type, member.node) : NULL; }) + +struct llist_node *lwq_dequeue_all(struct lwq *q); + +/** + * lwq_for_each_safe - iterate over detached queue allowing deletion + * @_n: iterator variable + * @_t1: temporary struct llist_node ** + * @_t2: temporary struct llist_node * + * @_l: address of llist_node pointer from lwq_dequeue_all() + * @_member: member in _n where lwq_node is found. + * + * Iterate over members in a dequeued list. If the iterator variable + * is set to NULL, the iterator removes that entry from the queue. + */ +#define lwq_for_each_safe(_n, _t1, _t2, _l, _member) \ + for (_t1 = (_l); \ + *(_t1) ? (_n = container_of(*(_t1), typeof(*(_n)), _member.node),\ + _t2 = ((*_t1)->next), \ + true) \ + : false; \ + (_n) ? (_t1 = &(_n)->_member.node.next, 0) \ + : ((*(_t1) = (_t2)), 0)) + +/** + * lwq_enqueue - add a new item to the end of the queue + * @n - the lwq_node embedded in the item to be added + * @q - the lwq to append to. + * + * No locking is needed to append to the queue so this can + * be called from any context. + * Return %true is the list may have previously been empty. + */ +static inline bool lwq_enqueue(struct lwq_node *n, struct lwq *q) +{ + /* acquire enqures ordering wrt lwq_dequeue */ + return llist_add(&n->node, &q->new) && + smp_load_acquire(&q->ready) == NULL; +} + +/** + * lwq_enqueue_batch - add a list of new items to the end of the queue + * @n - the lwq_node embedded in the first item to be added + * @q - the lwq to append to. + * + * No locking is needed to append to the queue so this can + * be called from any context. + * Return %true is the list may have previously been empty. + */ +static inline bool lwq_enqueue_batch(struct llist_node *n, struct lwq *q) +{ + struct llist_node *e = n; + + /* acquire enqures ordering wrt lwq_dequeue */ + return llist_add_batch(llist_reverse_order(n), e, &q->new) && + smp_load_acquire(&q->ready) == NULL; +} +#endif /* LWQ_H */ diff --git a/include/linux/sunrpc/svc.h b/include/linux/sunrpc/svc.h index dbf5b21feafe48..b10f987509cc80 100644 --- a/include/linux/sunrpc/svc.h +++ b/include/linux/sunrpc/svc.h @@ -17,6 +17,7 @@ #include <linux/sunrpc/xdr.h> #include <linux/sunrpc/auth.h> #include <linux/sunrpc/svcauth.h> +#include <linux/lwq.h> #include <linux/wait.h> #include <linux/mm.h> #include <linux/pagevec.h> @@ -33,10 +34,10 @@ */ struct svc_pool { unsigned int sp_id; /* pool id; also node id on NUMA */ - spinlock_t sp_lock; /* protects all fields */ - struct list_head sp_sockets; /* pending sockets */ - unsigned int sp_nrthreads; /* # of threads in pool */ + struct lwq sp_xprts; /* pending transports */ + atomic_t sp_nrthreads; /* # of threads in pool */ struct list_head sp_all_threads; /* all server threads */ + struct llist_head sp_idle_threads; /* idle server threads */ /* statistics on pool operation */ struct percpu_counter sp_messages_arrived; @@ -49,7 +50,8 @@ struct svc_pool { /* bits for sp_flags */ enum { SP_TASK_PENDING, /* still work to do even if no xprt is queued */ - SP_CONGESTED, /* all threads are busy, none idle */ + SP_NEED_VICTIM, /* One thread needs to agree to exit */ + SP_VICTIM_REMAINS, /* One thread needs to actually exit */ }; @@ -88,12 +90,9 @@ struct svc_serv { int (*sv_threadfn)(void *data); #if defined(CONFIG_SUNRPC_BACKCHANNEL) - struct list_head sv_cb_list; /* queue for callback requests + struct lwq sv_cb_list; /* queue for callback requests * that arrive over the same * connection */ - spinlock_t sv_cb_lock; /* protects the svc_cb_list */ - wait_queue_head_t sv_cb_waitq; /* sleep here if there are no - * entries in the svc_cb_list */ bool sv_bc_enabled; /* service uses backchannel */ #endif /* CONFIG_SUNRPC_BACKCHANNEL */ }; @@ -186,6 +185,7 @@ extern u32 svc_max_payload(const struct svc_rqst *rqstp); */ struct svc_rqst { struct list_head rq_all; /* all threads list */ + struct llist_node rq_idle; /* On the idle list */ struct rcu_head rq_rcu_head; /* for RCU deferred kfree */ struct svc_xprt * rq_xprt; /* transport ptr */ @@ -251,6 +251,7 @@ struct svc_rqst { * net namespace */ void ** rq_lease_breaker; /* The v4 client breaking a lease */ + unsigned int rq_status_counter; /* RPC processing counter */ }; /* bits for rq_flags */ @@ -261,8 +262,7 @@ enum { RQ_DROPME, /* drop current reply */ RQ_SPLICE_OK, /* turned off in gss privacy to prevent * encrypting page cache pages */ - RQ_VICTIM, /* about to be shut down */ - RQ_BUSY, /* request is busy */ + RQ_VICTIM, /* Have agreed to shut down */ RQ_DATA, /* request has data */ }; @@ -301,6 +301,28 @@ static inline struct sockaddr *svc_daddr(const struct svc_rqst *rqst) return (struct sockaddr *) &rqst->rq_daddr; } +/** + * svc_thread_should_stop - check if this thread should stop + * @rqstp: the thread that might need to stop + * + * To stop an svc thread, the pool flags SP_NEED_VICTIM and SP_VICTIM_REMAINS + * are set. The first thread which sees SP_NEED_VICTIM clears it, becoming + * the victim using this function. It should then promptly call + * svc_exit_thread() to complete the process, clearing SP_VICTIM_REMAINS + * so the task waiting for a thread to exit can wake and continue. + * + * Return values: + * %true: caller should invoke svc_exit_thread() + * %false: caller should do nothing + */ +static inline bool svc_thread_should_stop(struct svc_rqst *rqstp) +{ + if (test_and_clear_bit(SP_NEED_VICTIM, &rqstp->rq_pool->sp_flags)) + set_bit(RQ_VICTIM, &rqstp->rq_flags); + + return test_bit(RQ_VICTIM, &rqstp->rq_flags); +} + struct svc_deferred_req { u32 prot; /* protocol (UDP or TCP) */ struct svc_xprt *xprt; @@ -413,8 +435,7 @@ struct svc_serv * svc_create_pooled(struct svc_program *, unsigned int, int svc_set_num_threads(struct svc_serv *, struct svc_pool *, int); int svc_pool_stats_open(struct svc_serv *serv, struct file *file); void svc_process(struct svc_rqst *rqstp); -int bc_svc_process(struct svc_serv *, struct rpc_rqst *, - struct svc_rqst *); +void svc_process_bc(struct rpc_rqst *req, struct svc_rqst *rqstp); int svc_register(const struct svc_serv *, struct net *, const int, const unsigned short, const unsigned short); diff --git a/include/linux/sunrpc/svc_xprt.h b/include/linux/sunrpc/svc_xprt.h index fa55d12dc76512..8e20cd60e2e712 100644 --- a/include/linux/sunrpc/svc_xprt.h +++ b/include/linux/sunrpc/svc_xprt.h @@ -54,7 +54,7 @@ struct svc_xprt { const struct svc_xprt_ops *xpt_ops; struct kref xpt_ref; struct list_head xpt_list; - struct list_head xpt_ready; + struct lwq_node xpt_ready; unsigned long xpt_flags; struct svc_serv *xpt_server; /* service for transport */ diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h index 4ecc89301eb74d..f85d3a0daca26f 100644 --- a/include/linux/sunrpc/xprt.h +++ b/include/linux/sunrpc/xprt.h @@ -57,6 +57,7 @@ struct xprt_class; struct seq_file; struct svc_serv; struct net; +#include <linux/lwq.h> /* * This describes a complete RPC request @@ -121,7 +122,7 @@ struct rpc_rqst { int rq_ntrans; #if defined(CONFIG_SUNRPC_BACKCHANNEL) - struct list_head rq_bc_list; /* Callback service list */ + struct lwq_node rq_bc_list; /* Callback service list */ unsigned long rq_bc_pa_state; /* Backchannel prealloc state */ struct list_head rq_bc_pa_list; /* Backchannel prealloc list */ #endif /* CONFIG_SUNRPC_BACKCHANEL */ diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h index 6beb38c1dcb5eb..337c90787fb1a9 100644 --- a/include/trace/events/sunrpc.h +++ b/include/trace/events/sunrpc.h @@ -1677,7 +1677,6 @@ DEFINE_SVCXDRBUF_EVENT(sendto); svc_rqst_flag(DROPME) \ svc_rqst_flag(SPLICE_OK) \ svc_rqst_flag(VICTIM) \ - svc_rqst_flag(BUSY) \ svc_rqst_flag_end(DATA) #undef svc_rqst_flag diff --git a/include/uapi/linux/nfsd_netlink.h b/include/uapi/linux/nfsd_netlink.h new file mode 100644 index 00000000000000..c8ae72466ee644 --- /dev/null +++ b/include/uapi/linux/nfsd_netlink.h @@ -0,0 +1,39 @@ +/* SPDX-License-Identifier: ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) */ +/* Do not edit directly, auto-generated from: */ +/* Documentation/netlink/specs/nfsd.yaml */ +/* YNL-GEN uapi header */ + +#ifndef _UAPI_LINUX_NFSD_H +#define _UAPI_LINUX_NFSD_H + +#define NFSD_FAMILY_NAME "nfsd" +#define NFSD_FAMILY_VERSION 1 + +enum { + NFSD_A_RPC_STATUS_XID = 1, + NFSD_A_RPC_STATUS_FLAGS, + NFSD_A_RPC_STATUS_PROG, + NFSD_A_RPC_STATUS_VERSION, + NFSD_A_RPC_STATUS_PROC, + NFSD_A_RPC_STATUS_SERVICE_TIME, + NFSD_A_RPC_STATUS_PAD, + NFSD_A_RPC_STATUS_SADDR4, + NFSD_A_RPC_STATUS_DADDR4, + NFSD_A_RPC_STATUS_SADDR6, + NFSD_A_RPC_STATUS_DADDR6, + NFSD_A_RPC_STATUS_SPORT, + NFSD_A_RPC_STATUS_DPORT, + NFSD_A_RPC_STATUS_COMPOUND_OPS, + + __NFSD_A_RPC_STATUS_MAX, + NFSD_A_RPC_STATUS_MAX = (__NFSD_A_RPC_STATUS_MAX - 1) +}; + +enum { + NFSD_CMD_RPC_STATUS_GET = 1, + + __NFSD_CMD_MAX, + NFSD_CMD_MAX = (__NFSD_CMD_MAX - 1) +}; + +#endif /* _UAPI_LINUX_NFSD_H */ diff --git a/lib/Kconfig b/lib/Kconfig index c686f4adc1246a..76fe64f933fc7b 100644 --- a/lib/Kconfig +++ b/lib/Kconfig @@ -729,6 +729,11 @@ config PARMAN config OBJAGG tristate "objagg" if COMPILE_TEST +config LWQ_TEST + bool "Boot-time test for lwq queuing" + help + Run boot-time test of light-weight queuing. + endmenu config GENERIC_IOREMAP diff --git a/lib/Makefile b/lib/Makefile index 740109b6e2c89f..d0c116b706e6ba 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -45,7 +45,7 @@ obj-y += lockref.o obj-y += bcd.o sort.o parser.o debug_locks.o random32.o \ bust_spinlocks.o kasprintf.o bitmap.o scatterlist.o \ list_sort.o uuid.o iov_iter.o clz_ctz.o \ - bsearch.o find_bit.o llist.o memweight.o kfifo.o \ + bsearch.o find_bit.o llist.o lwq.o memweight.o kfifo.o \ percpu-refcount.o rhashtable.o base64.o \ once.o refcount.o rcuref.o usercopy.o errseq.o bucket_locks.o \ generic-radix-tree.o diff --git a/lib/llist.c b/lib/llist.c index 6e668fa5a2c66f..f21d0cfbbaaabc 100644 --- a/lib/llist.c +++ b/lib/llist.c @@ -66,6 +66,34 @@ struct llist_node *llist_del_first(struct llist_head *head) EXPORT_SYMBOL_GPL(llist_del_first); /** + * llist_del_first_this - delete given entry of lock-less list if it is first + * @head: the head for your lock-less list + * @this: a list entry. + * + * If head of the list is given entry, delete and return %true else + * return %false. + * + * Multiple callers can safely call this concurrently with multiple + * llist_add() callers, providing all the callers offer a different @this. + */ +bool llist_del_first_this(struct llist_head *head, + struct llist_node *this) +{ + struct llist_node *entry, *next; + + /* acquire ensures orderig wrt try_cmpxchg() is llist_del_first() */ + entry = smp_load_acquire(&head->first); + do { + if (entry != this) + return false; + next = READ_ONCE(entry->next); + } while (!try_cmpxchg(&head->first, &entry, next)); + + return true; +} +EXPORT_SYMBOL_GPL(llist_del_first_this); + +/** * llist_reverse_order - reverse order of a llist chain * @head: first item of the list to be reversed * diff --git a/lib/lwq.c b/lib/lwq.c new file mode 100644 index 00000000000000..8a723b29b39e50 --- /dev/null +++ b/lib/lwq.c @@ -0,0 +1,155 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Light-weight single-linked queue. + * + * Entries are enqueued to the head of an llist, with no blocking. + * This can happen in any context. + * + * Entries are dequeued using a spinlock to protect against multiple + * access. The llist is staged in reverse order, and refreshed + * from the llist when it exhausts. + * + * This is particularly suitable when work items are queued in BH or + * IRQ context, and where work items are handled one at a time by + * dedicated threads. + */ +#include <linux/rcupdate.h> +#include <linux/lwq.h> + +struct llist_node *__lwq_dequeue(struct lwq *q) +{ + struct llist_node *this; + + if (lwq_empty(q)) + return NULL; + spin_lock(&q->lock); + this = q->ready; + if (!this && !llist_empty(&q->new)) { + /* ensure queue doesn't appear transiently lwq_empty */ + smp_store_release(&q->ready, (void *)1); + this = llist_reverse_order(llist_del_all(&q->new)); + if (!this) + q->ready = NULL; + } + if (this) + q->ready = llist_next(this); + spin_unlock(&q->lock); + return this; +} +EXPORT_SYMBOL_GPL(__lwq_dequeue); + +/** + * lwq_dequeue_all - dequeue all currently enqueued objects + * @q: the queue to dequeue from + * + * Remove and return a linked list of llist_nodes of all the objects that were + * in the queue. The first on the list will be the object that was least + * recently enqueued. + */ +struct llist_node *lwq_dequeue_all(struct lwq *q) +{ + struct llist_node *r, *t, **ep; + + if (lwq_empty(q)) + return NULL; + + spin_lock(&q->lock); + r = q->ready; + q->ready = NULL; + t = llist_del_all(&q->new); + spin_unlock(&q->lock); + ep = &r; + while (*ep) + ep = &(*ep)->next; + *ep = llist_reverse_order(t); + return r; +} +EXPORT_SYMBOL_GPL(lwq_dequeue_all); + +#if IS_ENABLED(CONFIG_LWQ_TEST) + +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/wait_bit.h> +#include <linux/kthread.h> +#include <linux/delay.h> +struct tnode { + struct lwq_node n; + int i; + int c; +}; + +static int lwq_exercise(void *qv) +{ + struct lwq *q = qv; + int cnt; + struct tnode *t; + + for (cnt = 0; cnt < 10000; cnt++) { + wait_var_event(q, (t = lwq_dequeue(q, struct tnode, n)) != NULL); + t->c++; + if (lwq_enqueue(&t->n, q)) + wake_up_var(q); + } + while (!kthread_should_stop()) + schedule_timeout_idle(1); + return 0; +} + +static int lwq_test(void) +{ + int i; + struct lwq q; + struct llist_node *l, **t1, *t2; + struct tnode *t; + struct task_struct *threads[8]; + + printk(KERN_INFO "testing lwq....\n"); + lwq_init(&q); + printk(KERN_INFO " lwq: run some threads\n"); + for (i = 0; i < ARRAY_SIZE(threads); i++) + threads[i] = kthread_run(lwq_exercise, &q, "lwq-test-%d", i); + for (i = 0; i < 100; i++) { + t = kmalloc(sizeof(*t), GFP_KERNEL); + t->i = i; + t->c = 0; + if (lwq_enqueue(&t->n, &q)) + wake_up_var(&q); + } + /* wait for threads to exit */ + for (i = 0; i < ARRAY_SIZE(threads); i++) + if (!IS_ERR_OR_NULL(threads[i])) + kthread_stop(threads[i]); + printk(KERN_INFO " lwq: dequeue first 50:"); + for (i = 0; i < 50 ; i++) { + if (i && (i % 10) == 0) { + printk(KERN_CONT "\n"); + printk(KERN_INFO " lwq: ... "); + } + t = lwq_dequeue(&q, struct tnode, n); + printk(KERN_CONT " %d(%d)", t->i, t->c); + kfree(t); + } + printk(KERN_CONT "\n"); + l = lwq_dequeue_all(&q); + printk(KERN_INFO " lwq: delete the multiples of 3 (test lwq_for_each_safe())\n"); + lwq_for_each_safe(t, t1, t2, &l, n) { + if ((t->i % 3) == 0) { + t->i = -1; + kfree(t); + t = NULL; + } + } + if (l) + lwq_enqueue_batch(l, &q); + printk(KERN_INFO " lwq: dequeue remaining:"); + while ((t = lwq_dequeue(&q, struct tnode, n)) != NULL) { + printk(KERN_CONT " %d", t->i); + kfree(t); + } + printk(KERN_CONT "\n"); + return 0; +} + +module_init(lwq_test); +#endif /* CONFIG_LWQ_TEST*/ diff --git a/net/sunrpc/backchannel_rqst.c b/net/sunrpc/backchannel_rqst.c index 65a6c6429a53ed..caa94cf57123f4 100644 --- a/net/sunrpc/backchannel_rqst.c +++ b/net/sunrpc/backchannel_rqst.c @@ -83,7 +83,6 @@ static struct rpc_rqst *xprt_alloc_bc_req(struct rpc_xprt *xprt) return NULL; req->rq_xprt = xprt; - INIT_LIST_HEAD(&req->rq_bc_list); /* Preallocate one XDR receive buffer */ if (xprt_alloc_xdr_buf(&req->rq_rcv_buf, gfp_flags) < 0) { @@ -349,10 +348,8 @@ found: } /* - * Add callback request to callback list. The callback - * service sleeps on the sv_cb_waitq waiting for new - * requests. Wake it up after adding enqueing the - * request. + * Add callback request to callback list. Wake a thread + * on the first pool (usually the only pool) to handle it. */ void xprt_complete_bc_request(struct rpc_rqst *req, uint32_t copied) { @@ -369,8 +366,6 @@ void xprt_complete_bc_request(struct rpc_rqst *req, uint32_t copied) dprintk("RPC: add callback request to list\n"); xprt_get(xprt); - spin_lock(&bc_serv->sv_cb_lock); - list_add(&req->rq_bc_list, &bc_serv->sv_cb_list); - wake_up(&bc_serv->sv_cb_waitq); - spin_unlock(&bc_serv->sv_cb_lock); + lwq_enqueue(&req->rq_bc_list, &bc_serv->sv_cb_list); + svc_pool_wake_idle_thread(&bc_serv->sv_pools[0]); } diff --git a/net/sunrpc/svc.c b/net/sunrpc/svc.c index 812fda9d45dd63..7d4c10e609dcd4 100644 --- a/net/sunrpc/svc.c +++ b/net/sunrpc/svc.c @@ -438,9 +438,7 @@ EXPORT_SYMBOL_GPL(svc_bind); static void __svc_init_bc(struct svc_serv *serv) { - INIT_LIST_HEAD(&serv->sv_cb_list); - spin_lock_init(&serv->sv_cb_lock); - init_waitqueue_head(&serv->sv_cb_waitq); + lwq_init(&serv->sv_cb_list); } #else static void @@ -509,9 +507,9 @@ __svc_create(struct svc_program *prog, unsigned int bufsize, int npools, i, serv->sv_name); pool->sp_id = i; - INIT_LIST_HEAD(&pool->sp_sockets); + lwq_init(&pool->sp_xprts); INIT_LIST_HEAD(&pool->sp_all_threads); - spin_lock_init(&pool->sp_lock); + init_llist_head(&pool->sp_idle_threads); percpu_counter_init(&pool->sp_messages_arrived, 0, GFP_KERNEL); percpu_counter_init(&pool->sp_sockets_queued, 0, GFP_KERNEL); @@ -642,7 +640,6 @@ svc_rqst_alloc(struct svc_serv *serv, struct svc_pool *pool, int node) folio_batch_init(&rqstp->rq_fbatch); - __set_bit(RQ_BUSY, &rqstp->rq_flags); rqstp->rq_server = serv; rqstp->rq_pool = pool; @@ -682,10 +679,13 @@ svc_prepare_thread(struct svc_serv *serv, struct svc_pool *pool, int node) serv->sv_nrthreads += 1; spin_unlock_bh(&serv->sv_lock); - spin_lock_bh(&pool->sp_lock); - pool->sp_nrthreads++; + atomic_inc(&pool->sp_nrthreads); + + /* Protected by whatever lock the service uses when calling + * svc_set_num_threads() + */ list_add_rcu(&rqstp->rq_all, &pool->sp_all_threads); - spin_unlock_bh(&pool->sp_lock); + return rqstp; } @@ -701,23 +701,25 @@ svc_prepare_thread(struct svc_serv *serv, struct svc_pool *pool, int node) void svc_pool_wake_idle_thread(struct svc_pool *pool) { struct svc_rqst *rqstp; + struct llist_node *ln; rcu_read_lock(); - list_for_each_entry_rcu(rqstp, &pool->sp_all_threads, rq_all) { - if (test_and_set_bit(RQ_BUSY, &rqstp->rq_flags)) - continue; - + ln = READ_ONCE(pool->sp_idle_threads.first); + if (ln) { + rqstp = llist_entry(ln, struct svc_rqst, rq_idle); WRITE_ONCE(rqstp->rq_qtime, ktime_get()); - wake_up_process(rqstp->rq_task); + if (!task_is_running(rqstp->rq_task)) { + wake_up_process(rqstp->rq_task); + trace_svc_wake_up(rqstp->rq_task->pid); + percpu_counter_inc(&pool->sp_threads_woken); + } rcu_read_unlock(); - percpu_counter_inc(&pool->sp_threads_woken); - trace_svc_wake_up(rqstp->rq_task->pid); return; } rcu_read_unlock(); - set_bit(SP_CONGESTED, &pool->sp_flags); } +EXPORT_SYMBOL_GPL(svc_pool_wake_idle_thread); static struct svc_pool * svc_pool_next(struct svc_serv *serv, struct svc_pool *pool, unsigned int *state) @@ -725,36 +727,38 @@ svc_pool_next(struct svc_serv *serv, struct svc_pool *pool, unsigned int *state) return pool ? pool : &serv->sv_pools[(*state)++ % serv->sv_nrpools]; } -static struct task_struct * -svc_pool_victim(struct svc_serv *serv, struct svc_pool *pool, unsigned int *state) +static struct svc_pool * +svc_pool_victim(struct svc_serv *serv, struct svc_pool *target_pool, + unsigned int *state) { + struct svc_pool *pool; unsigned int i; - struct task_struct *task = NULL; + +retry: + pool = target_pool; if (pool != NULL) { - spin_lock_bh(&pool->sp_lock); + if (atomic_inc_not_zero(&pool->sp_nrthreads)) + goto found_pool; + return NULL; } else { for (i = 0; i < serv->sv_nrpools; i++) { pool = &serv->sv_pools[--(*state) % serv->sv_nrpools]; - spin_lock_bh(&pool->sp_lock); - if (!list_empty(&pool->sp_all_threads)) + if (atomic_inc_not_zero(&pool->sp_nrthreads)) goto found_pool; - spin_unlock_bh(&pool->sp_lock); } return NULL; } found_pool: - if (!list_empty(&pool->sp_all_threads)) { - struct svc_rqst *rqstp; - - rqstp = list_entry(pool->sp_all_threads.next, struct svc_rqst, rq_all); - set_bit(RQ_VICTIM, &rqstp->rq_flags); - list_del_rcu(&rqstp->rq_all); - task = rqstp->rq_task; - } - spin_unlock_bh(&pool->sp_lock); - return task; + set_bit(SP_VICTIM_REMAINS, &pool->sp_flags); + set_bit(SP_NEED_VICTIM, &pool->sp_flags); + if (!atomic_dec_and_test(&pool->sp_nrthreads)) + return pool; + /* Nothing left in this pool any more */ + clear_bit(SP_NEED_VICTIM, &pool->sp_flags); + clear_bit(SP_VICTIM_REMAINS, &pool->sp_flags); + goto retry; } static int @@ -795,18 +799,16 @@ svc_start_kthreads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) static int svc_stop_kthreads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) { - struct svc_rqst *rqstp; - struct task_struct *task; unsigned int state = serv->sv_nrthreads-1; + struct svc_pool *victim; do { - task = svc_pool_victim(serv, pool, &state); - if (task == NULL) + victim = svc_pool_victim(serv, pool, &state); + if (!victim) break; - rqstp = kthread_data(task); - /* Did we lose a race to svo_function threadfn? */ - if (kthread_stop(task) == -EINTR) - svc_exit_thread(rqstp); + svc_pool_wake_idle_thread(victim); + wait_on_bit(&victim->sp_flags, SP_VICTIM_REMAINS, + TASK_IDLE); nrservs++; } while (nrservs < 0); return 0; @@ -832,13 +834,10 @@ svc_stop_kthreads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) int svc_set_num_threads(struct svc_serv *serv, struct svc_pool *pool, int nrservs) { - if (pool == NULL) { + if (!pool) nrservs -= serv->sv_nrthreads; - } else { - spin_lock_bh(&pool->sp_lock); - nrservs -= pool->sp_nrthreads; - spin_unlock_bh(&pool->sp_lock); - } + else + nrservs -= atomic_read(&pool->sp_nrthreads); if (nrservs > 0) return svc_start_kthreads(serv, pool, nrservs); @@ -924,11 +923,9 @@ svc_exit_thread(struct svc_rqst *rqstp) struct svc_serv *serv = rqstp->rq_server; struct svc_pool *pool = rqstp->rq_pool; - spin_lock_bh(&pool->sp_lock); - pool->sp_nrthreads--; - if (!test_and_set_bit(RQ_VICTIM, &rqstp->rq_flags)) - list_del_rcu(&rqstp->rq_all); - spin_unlock_bh(&pool->sp_lock); + list_del_rcu(&rqstp->rq_all); + + atomic_dec(&pool->sp_nrthreads); spin_lock_bh(&serv->sv_lock); serv->sv_nrthreads -= 1; @@ -938,6 +935,11 @@ svc_exit_thread(struct svc_rqst *rqstp) svc_rqst_free(rqstp); svc_put(serv); + /* That svc_put() cannot be the last, because the thread + * waiting for SP_VICTIM_REMAINS to clear must hold + * a reference. So it is still safe to access pool. + */ + clear_and_wake_up_bit(SP_VICTIM_REMAINS, &pool->sp_flags); } EXPORT_SYMBOL_GPL(svc_exit_thread); @@ -1544,24 +1546,20 @@ out_drop: } #if defined(CONFIG_SUNRPC_BACKCHANNEL) -/* - * Process a backchannel RPC request that arrived over an existing - * outbound connection +/** + * svc_process_bc - process a reverse-direction RPC request + * @req: RPC request to be used for client-side processing + * @rqstp: server-side execution context + * */ -int -bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req, - struct svc_rqst *rqstp) +void svc_process_bc(struct rpc_rqst *req, struct svc_rqst *rqstp) { struct rpc_task *task; int proc_error; - int error; - - dprintk("svc: %s(%p)\n", __func__, req); /* Build the svc_rqst used by the common processing routine */ rqstp->rq_xid = req->rq_xid; rqstp->rq_prot = req->rq_xprt->prot; - rqstp->rq_server = serv; rqstp->rq_bc_net = req->rq_xprt->xprt_net; rqstp->rq_addrlen = sizeof(req->rq_xprt->addr); @@ -1590,10 +1588,8 @@ bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req, * been processed by the caller. */ svcxdr_init_decode(rqstp); - if (!xdr_inline_decode(&rqstp->rq_arg_stream, XDR_UNIT * 2)) { - error = -EINVAL; - goto out; - } + if (!xdr_inline_decode(&rqstp->rq_arg_stream, XDR_UNIT * 2)) + return; /* Parse and execute the bc call */ proc_error = svc_process_common(rqstp); @@ -1602,26 +1598,18 @@ bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req, if (!proc_error) { /* Processing error: drop the request */ xprt_free_bc_request(req); - error = -EINVAL; - goto out; + return; } /* Finally, send the reply synchronously */ memcpy(&req->rq_snd_buf, &rqstp->rq_res, sizeof(req->rq_snd_buf)); task = rpc_run_bc_task(req); - if (IS_ERR(task)) { - error = PTR_ERR(task); - goto out; - } + if (IS_ERR(task)) + return; WARN_ON_ONCE(atomic_read(&task->tk_count) != 1); - error = task->tk_status; rpc_put_task(task); - -out: - dprintk("svc: %s(), error=%d\n", __func__, error); - return error; } -EXPORT_SYMBOL_GPL(bc_svc_process); +EXPORT_SYMBOL_GPL(svc_process_bc); #endif /* CONFIG_SUNRPC_BACKCHANNEL */ /** diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index 4cfe9640df4814..fee83d1024bc8d 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -9,7 +9,6 @@ #include <linux/sched/mm.h> #include <linux/errno.h> #include <linux/freezer.h> -#include <linux/kthread.h> #include <linux/slab.h> #include <net/sock.h> #include <linux/sunrpc/addr.h> @@ -17,6 +16,7 @@ #include <linux/sunrpc/svc_xprt.h> #include <linux/sunrpc/svcsock.h> #include <linux/sunrpc/xprt.h> +#include <linux/sunrpc/bc_xprt.h> #include <linux/module.h> #include <linux/netdevice.h> #include <trace/events/sunrpc.h> @@ -201,7 +201,6 @@ void svc_xprt_init(struct net *net, struct svc_xprt_class *xcl, kref_init(&xprt->xpt_ref); xprt->xpt_server = serv; INIT_LIST_HEAD(&xprt->xpt_list); - INIT_LIST_HEAD(&xprt->xpt_ready); INIT_LIST_HEAD(&xprt->xpt_deferred); INIT_LIST_HEAD(&xprt->xpt_users); mutex_init(&xprt->xpt_mutex); @@ -472,9 +471,7 @@ void svc_xprt_enqueue(struct svc_xprt *xprt) pool = svc_pool_for_cpu(xprt->xpt_server); percpu_counter_inc(&pool->sp_sockets_queued); - spin_lock_bh(&pool->sp_lock); - list_add_tail(&xprt->xpt_ready, &pool->sp_sockets); - spin_unlock_bh(&pool->sp_lock); + lwq_enqueue(&xprt->xpt_ready, &pool->sp_xprts); svc_pool_wake_idle_thread(pool); } @@ -487,18 +484,9 @@ static struct svc_xprt *svc_xprt_dequeue(struct svc_pool *pool) { struct svc_xprt *xprt = NULL; - if (list_empty(&pool->sp_sockets)) - goto out; - - spin_lock_bh(&pool->sp_lock); - if (likely(!list_empty(&pool->sp_sockets))) { - xprt = list_first_entry(&pool->sp_sockets, - struct svc_xprt, xpt_ready); - list_del_init(&xprt->xpt_ready); + xprt = lwq_dequeue(&pool->sp_xprts, struct svc_xprt, xpt_ready); + if (xprt) svc_xprt_get(xprt); - } - spin_unlock_bh(&pool->sp_lock); -out: return xprt; } @@ -674,7 +662,7 @@ static bool svc_alloc_arg(struct svc_rqst *rqstp) continue; set_current_state(TASK_IDLE); - if (kthread_should_stop()) { + if (svc_thread_should_stop(rqstp)) { set_current_state(TASK_RUNNING); return false; } @@ -699,7 +687,7 @@ static bool svc_alloc_arg(struct svc_rqst *rqstp) } static bool -rqst_should_sleep(struct svc_rqst *rqstp) +svc_thread_should_sleep(struct svc_rqst *rqstp) { struct svc_pool *pool = rqstp->rq_pool; @@ -708,65 +696,51 @@ rqst_should_sleep(struct svc_rqst *rqstp) return false; /* was a socket queued? */ - if (!list_empty(&pool->sp_sockets)) + if (!lwq_empty(&pool->sp_xprts)) return false; /* are we shutting down? */ - if (kthread_should_stop()) + if (svc_thread_should_stop(rqstp)) return false; - /* are we freezing? */ - if (freezing(current)) - return false; +#if defined(CONFIG_SUNRPC_BACKCHANNEL) + if (svc_is_backchannel(rqstp)) { + if (!lwq_empty(&rqstp->rq_server->sv_cb_list)) + return false; + } +#endif return true; } -static struct svc_xprt *svc_get_next_xprt(struct svc_rqst *rqstp) +static void svc_thread_wait_for_work(struct svc_rqst *rqstp) { - struct svc_pool *pool = rqstp->rq_pool; - - /* rq_xprt should be clear on entry */ - WARN_ON_ONCE(rqstp->rq_xprt); - - rqstp->rq_xprt = svc_xprt_dequeue(pool); - if (rqstp->rq_xprt) - goto out_found; - - set_current_state(TASK_IDLE); - smp_mb__before_atomic(); - clear_bit(SP_CONGESTED, &pool->sp_flags); - clear_bit(RQ_BUSY, &rqstp->rq_flags); - smp_mb__after_atomic(); - - if (likely(rqst_should_sleep(rqstp))) - schedule(); - else + struct svc_pool *pool = rqstp->rq_pool; + + if (svc_thread_should_sleep(rqstp)) { + set_current_state(TASK_IDLE | TASK_FREEZABLE); + llist_add(&rqstp->rq_idle, &pool->sp_idle_threads); + if (likely(svc_thread_should_sleep(rqstp))) + schedule(); + + while (!llist_del_first_this(&pool->sp_idle_threads, + &rqstp->rq_idle)) { + /* Work just became available. This thread can only + * handle it after removing rqstp from the idle + * list. If that attempt failed, some other thread + * must have queued itself after finding no + * work to do, so that thread has taken responsibly + * for this new work. This thread can safely sleep + * until woken again. + */ + schedule(); + set_current_state(TASK_IDLE | TASK_FREEZABLE); + } __set_current_state(TASK_RUNNING); - + } else { + cond_resched(); + } try_to_freeze(); - - set_bit(RQ_BUSY, &rqstp->rq_flags); - smp_mb__after_atomic(); - clear_bit(SP_TASK_PENDING, &pool->sp_flags); - rqstp->rq_xprt = svc_xprt_dequeue(pool); - if (rqstp->rq_xprt) - goto out_found; - - if (kthread_should_stop()) - return NULL; - return NULL; -out_found: - clear_bit(SP_TASK_PENDING, &pool->sp_flags); - /* Normally we will wait up to 5 seconds for any required - * cache information to be provided. - */ - if (!test_bit(SP_CONGESTED, &pool->sp_flags)) - rqstp->rq_chandle.thread_wait = 5*HZ; - else - rqstp->rq_chandle.thread_wait = 1*HZ; - trace_svc_xprt_dequeue(rqstp); - return rqstp->rq_xprt; } static void svc_add_new_temp_xprt(struct svc_serv *serv, struct svc_xprt *newxpt) @@ -785,7 +759,7 @@ static void svc_add_new_temp_xprt(struct svc_serv *serv, struct svc_xprt *newxpt svc_xprt_received(newxpt); } -static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt) +static void svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt) { struct svc_serv *serv = rqstp->rq_server; int len = 0; @@ -826,11 +800,35 @@ static int svc_handle_xprt(struct svc_rqst *rqstp, struct svc_xprt *xprt) len = xprt->xpt_ops->xpo_recvfrom(rqstp); rqstp->rq_reserved = serv->sv_max_mesg; atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved); + if (len <= 0) + goto out; + + trace_svc_xdr_recvfrom(&rqstp->rq_arg); + + clear_bit(XPT_OLD, &xprt->xpt_flags); + + rqstp->rq_chandle.defer = svc_defer; + + if (serv->sv_stats) + serv->sv_stats->netcnt++; + percpu_counter_inc(&rqstp->rq_pool->sp_messages_arrived); + rqstp->rq_stime = ktime_get(); + svc_process(rqstp); } else svc_xprt_received(xprt); out: - return len; + rqstp->rq_res.len = 0; + svc_xprt_release(rqstp); +} + +static void svc_thread_wake_next(struct svc_rqst *rqstp) +{ + if (!svc_thread_should_sleep(rqstp)) + /* More work pending after I dequeued some, + * wake another worker + */ + svc_pool_wake_idle_thread(rqstp->rq_pool); } /** @@ -843,44 +841,51 @@ out: */ void svc_recv(struct svc_rqst *rqstp) { - struct svc_xprt *xprt = NULL; - struct svc_serv *serv = rqstp->rq_server; - int len; + struct svc_pool *pool = rqstp->rq_pool; if (!svc_alloc_arg(rqstp)) - goto out; + return; - try_to_freeze(); - cond_resched(); - if (kthread_should_stop()) - goto out; + svc_thread_wait_for_work(rqstp); - xprt = svc_get_next_xprt(rqstp); - if (!xprt) - goto out; + clear_bit(SP_TASK_PENDING, &pool->sp_flags); - len = svc_handle_xprt(rqstp, xprt); + if (svc_thread_should_stop(rqstp)) { + svc_thread_wake_next(rqstp); + return; + } - /* No data, incomplete (TCP) read, or accept() */ - if (len <= 0) - goto out_release; + rqstp->rq_xprt = svc_xprt_dequeue(pool); + if (rqstp->rq_xprt) { + struct svc_xprt *xprt = rqstp->rq_xprt; - trace_svc_xdr_recvfrom(&rqstp->rq_arg); + svc_thread_wake_next(rqstp); + /* Normally we will wait up to 5 seconds for any required + * cache information to be provided. When there are no + * idle threads, we reduce the wait time. + */ + if (pool->sp_idle_threads.first) + rqstp->rq_chandle.thread_wait = 5 * HZ; + else + rqstp->rq_chandle.thread_wait = 1 * HZ; - clear_bit(XPT_OLD, &xprt->xpt_flags); + trace_svc_xprt_dequeue(rqstp); + svc_handle_xprt(rqstp, xprt); + } - rqstp->rq_chandle.defer = svc_defer; +#if defined(CONFIG_SUNRPC_BACKCHANNEL) + if (svc_is_backchannel(rqstp)) { + struct svc_serv *serv = rqstp->rq_server; + struct rpc_rqst *req; - if (serv->sv_stats) - serv->sv_stats->netcnt++; - percpu_counter_inc(&rqstp->rq_pool->sp_messages_arrived); - rqstp->rq_stime = ktime_get(); - svc_process(rqstp); -out: - return; -out_release: - rqstp->rq_res.len = 0; - svc_xprt_release(rqstp); + req = lwq_dequeue(&serv->sv_cb_list, + struct rpc_rqst, rq_bc_list); + if (req) { + svc_thread_wake_next(rqstp); + svc_process_bc(req, rqstp); + } + } +#endif } EXPORT_SYMBOL_GPL(svc_recv); @@ -890,7 +895,6 @@ EXPORT_SYMBOL_GPL(svc_recv); void svc_drop(struct svc_rqst *rqstp) { trace_svc_drop(rqstp); - svc_xprt_release(rqstp); } EXPORT_SYMBOL_GPL(svc_drop); @@ -906,8 +910,6 @@ void svc_send(struct svc_rqst *rqstp) int status; xprt = rqstp->rq_xprt; - if (!xprt) - return; /* calculate over-all length */ xb = &rqstp->rq_res; @@ -920,7 +922,6 @@ void svc_send(struct svc_rqst *rqstp) status = xprt->xpt_ops->xpo_sendto(rqstp); trace_svc_send(rqstp, status); - svc_xprt_release(rqstp); } /* @@ -1031,7 +1032,6 @@ static void svc_delete_xprt(struct svc_xprt *xprt) spin_lock_bh(&serv->sv_lock); list_del_init(&xprt->xpt_list); - WARN_ON_ONCE(!list_empty(&xprt->xpt_ready)); if (test_bit(XPT_TEMP, &xprt->xpt_flags)) serv->sv_tmpcnt--; spin_unlock_bh(&serv->sv_lock); @@ -1082,36 +1082,26 @@ static int svc_close_list(struct svc_serv *serv, struct list_head *xprt_list, st return ret; } -static struct svc_xprt *svc_dequeue_net(struct svc_serv *serv, struct net *net) +static void svc_clean_up_xprts(struct svc_serv *serv, struct net *net) { - struct svc_pool *pool; struct svc_xprt *xprt; - struct svc_xprt *tmp; int i; for (i = 0; i < serv->sv_nrpools; i++) { - pool = &serv->sv_pools[i]; - - spin_lock_bh(&pool->sp_lock); - list_for_each_entry_safe(xprt, tmp, &pool->sp_sockets, xpt_ready) { - if (xprt->xpt_net != net) - continue; - list_del_init(&xprt->xpt_ready); - spin_unlock_bh(&pool->sp_lock); - return xprt; + struct svc_pool *pool = &serv->sv_pools[i]; + struct llist_node *q, **t1, *t2; + + q = lwq_dequeue_all(&pool->sp_xprts); + lwq_for_each_safe(xprt, t1, t2, &q, xpt_ready) { + if (xprt->xpt_net == net) { + set_bit(XPT_CLOSE, &xprt->xpt_flags); + svc_delete_xprt(xprt); + xprt = NULL; + } } - spin_unlock_bh(&pool->sp_lock); - } - return NULL; -} - -static void svc_clean_up_xprts(struct svc_serv *serv, struct net *net) -{ - struct svc_xprt *xprt; - while ((xprt = svc_dequeue_net(serv, net))) { - set_bit(XPT_CLOSE, &xprt->xpt_flags); - svc_delete_xprt(xprt); + if (q) + lwq_enqueue_batch(q, &pool->sp_xprts); } } diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c index e4d84a13c566e7..8c817e755262df 100644 --- a/net/sunrpc/xprtrdma/backchannel.c +++ b/net/sunrpc/xprtrdma/backchannel.c @@ -263,11 +263,9 @@ void rpcrdma_bc_receive_call(struct rpcrdma_xprt *r_xprt, /* Queue rqst for ULP's callback service */ bc_serv = xprt->bc_serv; xprt_get(xprt); - spin_lock(&bc_serv->sv_cb_lock); - list_add(&rqst->rq_bc_list, &bc_serv->sv_cb_list); - spin_unlock(&bc_serv->sv_cb_lock); + lwq_enqueue(&rqst->rq_bc_list, &bc_serv->sv_cb_list); - wake_up(&bc_serv->sv_cb_waitq); + svc_pool_wake_idle_thread(&bc_serv->sv_pools[0]); r_xprt->rx_stats.bcall_count++; return; |