Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

"ENOBUF" on socket writes isn't really fatal; we may just be too fast

Browse Source 
for the driver. Don't log the error, just try again. Could try to
be smart and do a backoff, but it's probably not worth the trouble.
This commit is contained in:
fvdl 2000-09-19 22:21:21 +00:00
parent 8d38a9fc8e
commit 874d41c959
1 changed files with 23 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

32
sys/nfs/nfs_socket.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: nfs_socket.c,v 1.59 2000/09/19 17:04:51 bjh21 Exp $ */
/* $NetBSD: nfs_socket.c,v 1.60 2000/09/19 22:21:21 fvdl Exp $ */
/*
* Copyright (c) 1989, 1991, 1993, 1995
@ -441,15 +441,29 @@ nfs_send(so, nam, top, rep)
(struct mbuf *)0, flags);
if (error) {
if (rep) {
log(LOG_INFO, "nfs send error %d for server %s\n",error,
rep->r_nmp->nm_mountp->mnt_stat.f_mntfromname);
/*
* Deal with errors for the client side.
*/
if (rep->r_flags & R_SOFTTERM)
error = EINTR;
else
if (error == ENOBUFS && so->so_type == SOCK_DGRAM) {
/*
* We're too fast for the network/driver,
* and UDP isn't flowcontrolled.
* We need to resend. This is not fatal,
* just try again.
*
* Could be smarter here by doing some sort
* of a backoff, but this is rare.
*/
rep->r_flags |= R_MUSTRESEND;
} else {
log(LOG_INFO, "nfs send error %d for %s\n",
error,
rep->r_nmp->nm_mountp->mnt_stat.f_mntfromname);
/*
* Deal with errors for the client side.
*/
if (rep->r_flags & R_SOFTTERM)
error = EINTR;
else
rep->r_flags |= R_MUSTRESEND;
}
} else
log(LOG_INFO, "nfsd send error %d\n", error);