nop... its my fault, yet again, not to use the standard am.config file. so it woz variable "nfs_proto" which is not deducted and null by default.
so, maybe the following patch in amd/srvr_nfs.c will make it more robusts to
crashes:
@@ -956,11 +956,11 @@ no_dns:
* chosen here (right now it always uses datagram sockets).
*/
ITER(fs, fserver, &nfs_srvr_list) {
if (STREQ(host, fs->fs_host) &&
nfs_version == fs->fs_version &&
- STREQ(nfs_proto, fs->fs_proto)) {
+ (!nfs_proto || STREQ(nfs_proto, fs->fs_proto))) {
/*
* fill in the IP address -- this is only needed
* if there is a chance an IP address will change
* between mounts.
* Mike Mitchell, m...@unx.sas.com, 09/08/93
still this global nfs_srvr_list queue is not protected,
although there are portmaper callback pinging function at work....
this am-utils package makes you shiver !
thx for reading this, and long live user-space services.
PTZ.
signature.asc
Description: PGP signature
