Skip site navigation (1)Skip section navigation (2) 
Date:      Wed, 30 Jun 2021 22:19:30 GMT
From:      Rick Macklem <rmacklem@FreeBSD.org>
To:        src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-main@FreeBSD.org
Subject:   git: c5f4772c66d2 - main - nfscl: Improve "Consider increasing kern.ipc.maxsockbuf" message
Message-ID:  <202106302219.15UMJU5j080956@gitrepo.freebsd.org>
next in thread | raw e-mail | index | archive | help 
The branch main has been updated by rmacklem:

URL: https://cgit.FreeBSD.org/src/commit/?id=c5f4772c66d2eb31b84a84a89c8a284043f03452

commit c5f4772c66d2eb31b84a84a89c8a284043f03452
Author:     Rick Macklem <rmacklem@FreeBSD.org>
AuthorDate: 2021-06-30 22:15:41 +0000
Commit:     Rick Macklem <rmacklem@FreeBSD.org>
CommitDate: 2021-06-30 22:15:41 +0000

    nfscl: Improve "Consider increasing kern.ipc.maxsockbuf" message
    
    When the setting of kern.ipc.maxsockbuf is less than what is
    desired for I/O based on vfs.maxbcachebuf and vfs.nfs.bufpackets,
    a console message of "Consider increasing kern.ipc.maxsockbuf".
    is printed.
    
    This patch modifies the message to provide a suggested value
    for kern.ipc.maxsockbuf.
    Note that the setting is only needed when the NFS rsize/wsize
    is set to vfs.maxbcachebuf.
    
    While here, make nfs_bufpackets global, so that it can be used
    by a future patch that adds a sysctl to set the NFS server's
    maximum I/O size.  Also, remove "sizeof(u_int32_t)" from the maximum
    packet length, since NFS_MAXXDR is already an "overestimate"
    of the actual length.
    
    MFC after:      2 weeks
---
 sys/fs/nfs/nfs_commonkrpc.c | 28 +++++++++++++++++++++-------
 1 file changed, 21 insertions(+), 7 deletions(-)

diff --git a/sys/fs/nfs/nfs_commonkrpc.c b/sys/fs/nfs/nfs_commonkrpc.c
index 04ef04955ce0..63ff02331a0e 100644
--- a/sys/fs/nfs/nfs_commonkrpc.c
+++ b/sys/fs/nfs/nfs_commonkrpc.c
@@ -101,8 +101,8 @@ extern int nfscl_debuglevel;
 extern int nfsrv_lease;
 
 SVCPOOL		*nfscbd_pool;
+int		nfs_bufpackets = 4;
 static int	nfsrv_gsscallbackson = 0;
-static int	nfs_bufpackets = 4;
 static int	nfs_reconnects;
 static int	nfs3_jukebox_delay = 10;
 static int	nfs_skip_wcc_data_onerr = 1;
@@ -180,6 +180,7 @@ newnfs_connect(struct nfsmount *nmp, struct nfssockreq *nrp,
 	struct thread *td = curthread;
 	SVCXPRT *xprt;
 	struct timeval timo;
+	uint64_t tval;
 
 	/*
 	 * We need to establish the socket using the credentials of
@@ -238,8 +239,21 @@ newnfs_connect(struct nfsmount *nmp, struct nfssockreq *nrp,
 	do {
 	    if (error != 0 && pktscale > 2) {
 		if (nmp != NULL && nrp->nr_sotype == SOCK_STREAM &&
-		    pktscale == pktscalesav)
-		    printf("Consider increasing kern.ipc.maxsockbuf\n");
+		    pktscale == pktscalesav) {
+		    /*
+		     * Suggest vfs.nfs.bufpackets * maximum RPC message,
+		     * adjusted for the sb_max->sb_max_adj conversion of
+		     * MCLBYTES / (MSIZE + MCLBYTES) as the minimum setting
+		     * for kern.ipc.maxsockbuf.
+		     */
+		    tval = (NFS_MAXBSIZE + NFS_MAXXDR) * nfs_bufpackets;
+		    tval *= MSIZE + MCLBYTES;
+		    tval += MCLBYTES - 1; /* Round up divide by MCLBYTES. */
+		    tval /= MCLBYTES;
+		    printf("Consider increasing kern.ipc.maxsockbuf to a "
+			"minimum of %ju to support %ubyte NFS I/O\n",
+			(uintmax_t)tval, NFS_MAXBSIZE);
+		}
 		pktscale--;
 	    }
 	    if (nrp->nr_sotype == SOCK_DGRAM) {
@@ -255,10 +269,10 @@ newnfs_connect(struct nfsmount *nmp, struct nfssockreq *nrp,
 		if (nrp->nr_sotype != SOCK_STREAM)
 			panic("nfscon sotype");
 		if (nmp != NULL) {
-			sndreserve = (NFS_MAXBSIZE + NFS_MAXXDR +
-			    sizeof (u_int32_t)) * pktscale;
-			rcvreserve = (NFS_MAXBSIZE + NFS_MAXXDR +
-			    sizeof (u_int32_t)) * pktscale;
+			sndreserve = (NFS_MAXBSIZE + NFS_MAXXDR) *
+			    pktscale;
+			rcvreserve = (NFS_MAXBSIZE + NFS_MAXXDR) *
+			    pktscale;
 		} else {
 			sndreserve = rcvreserve = 1024 * pktscale;
 		}

Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?202106302219.15UMJU5j080956>