NFSD: Add a mechanism to wait for a DELEGRETURN

Subsequent patches will use this mechanism to wake up an operation
that is waiting for a client to return a delegation.

The new tracepoint records whether the wait timed out or was
properly awoken by the expected DELEGRETURN:

            nfsd-1155  [002] 83799.493199: nfsd_delegret_wakeup: xid=0x14b7d6ef fh_hash=0xf6826792 (timed out)

Suggested-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>

diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
index 561f3556b1d22ec5dffa5a1cae1c88f602fe1bc4..54bc70427ce330b4a9f38f2fb15db3214f965598 100644 (file)
--- a/fs/nfsd/nfs4state.c
+++ b/fs/nfsd/nfs4state.c
@@ -4717,6 +4717,35 @@ nfs4_share_conflict(struct svc_fh *current_fh, unsigned int deny_type)
        return ret;
 }
 
+static bool nfsd4_deleg_present(const struct inode *inode)
+{
+       struct file_lock_context *ctx = smp_load_acquire(&inode->i_flctx);
+
+       return ctx && !list_empty_careful(&ctx->flc_lease);
+}
+
+/**
+ * nfsd_wait_for_delegreturn - wait for delegations to be returned
+ * @rqstp: the RPC transaction being executed
+ * @inode: in-core inode of the file being waited for
+ *
+ * The timeout prevents deadlock if all nfsd threads happen to be
+ * tied up waiting for returning delegations.
+ *
+ * Return values:
+ *   %true: delegation was returned
+ *   %false: timed out waiting for delegreturn
+ */
+bool nfsd_wait_for_delegreturn(struct svc_rqst *rqstp, struct inode *inode)
+{
+       long __maybe_unused timeo;
+
+       timeo = wait_var_event_timeout(inode, !nfsd4_deleg_present(inode),
+                                      NFSD_DELEGRETURN_TIMEOUT);
+       trace_nfsd_delegret_wakeup(rqstp, inode, timeo);
+       return timeo > 0;
+}
+
 static void nfsd4_cb_recall_prepare(struct nfsd4_callback *cb)
 {
        struct nfs4_delegation *dp = cb_to_delegation(cb);
@@ -6779,6 +6808,7 @@ nfsd4_delegreturn(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
        if (status)
                goto put_stateid;
 
+       wake_up_var(d_inode(cstate->current_fh.fh_dentry));
        destroy_delegation(dp);
 put_stateid:
        nfs4_put_stid(&dp->dl_stid);

diff --git a/fs/nfsd/nfsd.h b/fs/nfsd/nfsd.h
index 57a468ed85c358b377af699a562c0b0fa6e21ed6..6ab4ad41ae84e0fb65bf429211970f9f4864a182 100644 (file)
--- a/fs/nfsd/nfsd.h
+++ b/fs/nfsd/nfsd.h
@@ -164,6 +164,7 @@ char * nfs4_recoverydir(void);
 bool nfsd4_spo_must_allow(struct svc_rqst *rqstp);
 int nfsd4_create_laundry_wq(void);
 void nfsd4_destroy_laundry_wq(void);
+bool nfsd_wait_for_delegreturn(struct svc_rqst *rqstp, struct inode *inode);
 #else
 static inline int nfsd4_init_slabs(void) { return 0; }
 static inline void nfsd4_free_slabs(void) { }
@@ -179,6 +180,11 @@ static inline bool nfsd4_spo_must_allow(struct svc_rqst *rqstp)
 }
 static inline int nfsd4_create_laundry_wq(void) { return 0; };
 static inline void nfsd4_destroy_laundry_wq(void) {};
+static inline bool nfsd_wait_for_delegreturn(struct svc_rqst *rqstp,
+                                             struct inode *inode)
+{
+       return false;
+}
 #endif
 
 /*
@@ -343,6 +349,7 @@ void                nfsd_lockd_shutdown(void);
 #define        NFSD_COURTESY_CLIENT_TIMEOUT    (24 * 60 * 60)  /* seconds */
 #define        NFSD_CLIENT_MAX_TRIM_PER_RUN    128
 #define        NFS4_CLIENTS_PER_GB             1024
+#define NFSD_DELEGRETURN_TIMEOUT       (HZ / 34)       /* 30ms */
 
 /*
  * The following attributes are currently not supported by the NFSv4 server:

diff --git a/fs/nfsd/trace.h b/fs/nfsd/trace.h
index ec8e08315779a8753c8adda194aa27dda59f84cd..06a96e955bd00dca4cbf3bfd33cd9ba4b165a531 100644 (file)
--- a/fs/nfsd/trace.h
+++ b/fs/nfsd/trace.h
@@ -538,6 +538,29 @@ DEFINE_NFSD_COPY_ERR_EVENT(clone_file_range_err);
 #include "filecache.h"
 #include "vfs.h"
 
+TRACE_EVENT(nfsd_delegret_wakeup,
+       TP_PROTO(
+               const struct svc_rqst *rqstp,
+               const struct inode *inode,
+               long timeo
+       ),
+       TP_ARGS(rqstp, inode, timeo),
+       TP_STRUCT__entry(
+               __field(u32, xid)
+               __field(const void *, inode)
+               __field(long, timeo)
+       ),
+       TP_fast_assign(
+               __entry->xid = be32_to_cpu(rqstp->rq_xid);
+               __entry->inode = inode;
+               __entry->timeo = timeo;
+       ),
+       TP_printk("xid=0x%08x inode=%p%s",
+                 __entry->xid, __entry->inode,
+                 __entry->timeo == 0 ? " (timed out)" : ""
+       )
+);
+
 DECLARE_EVENT_CLASS(nfsd_stateid_class,
        TP_PROTO(stateid_t *stp),
        TP_ARGS(stp),