Author: trasz
Date: Tue Feb 19 10:28:59 2019
New Revision: 344274
URL: https://svnweb.freebsd.org/changeset/base/344274
Log:
MFC r342114:
Add kern.rpc.gss.client_max, to make it possible to bump it easily.
This can drastically lower the load on gssd(8) on large NFS servers.
Modified:
stable/12/sys/rpc/rpcsec_gss/svc_rpcsec_gss.c
Directory Properties:
stable/12/ (props changed)
Modified: stable/12/sys/rpc/rpcsec_gss/svc_rpcsec_gss.c
==============================================================================
--- stable/12/sys/rpc/rpcsec_gss/svc_rpcsec_gss.c Tue Feb 19 10:17:49
2019 (r344273)
+++ stable/12/sys/rpc/rpcsec_gss/svc_rpcsec_gss.c Tue Feb 19 10:28:59
2019 (r344274)
@@ -171,9 +171,22 @@ struct svc_rpc_gss_cookedcred {
#define CLIENT_HASH_SIZE 256
#define CLIENT_MAX 128
+u_int svc_rpc_gss_client_max = CLIENT_MAX;
+
+SYSCTL_NODE(_kern, OID_AUTO, rpc, CTLFLAG_RW, 0, "RPC");
+SYSCTL_NODE(_kern_rpc, OID_AUTO, gss, CTLFLAG_RW, 0, "GSS");
+
+SYSCTL_UINT(_kern_rpc_gss, OID_AUTO, client_max, CTLFLAG_RW,
+ &svc_rpc_gss_client_max, 0,
+ "Max number of rpc-gss clients");
+
+static u_int svc_rpc_gss_client_count;
+SYSCTL_UINT(_kern_rpc_gss, OID_AUTO, client_count, CTLFLAG_RD,
+ &svc_rpc_gss_client_count, 0,
+ "Number of rpc-gss clients");
+
struct svc_rpc_gss_client_list svc_rpc_gss_client_hash[CLIENT_HASH_SIZE];
struct svc_rpc_gss_client_list svc_rpc_gss_clients;
-static size_t svc_rpc_gss_client_count;
static uint32_t svc_rpc_gss_next_clientid = 1;
static void
@@ -669,7 +682,7 @@ svc_rpc_gss_timeout_clients(void)
*/
sx_xlock(&svc_rpc_gss_lock);
client = TAILQ_LAST(&svc_rpc_gss_clients, svc_rpc_gss_client_list);
- while (svc_rpc_gss_client_count > CLIENT_MAX && client != NULL) {
+ while (svc_rpc_gss_client_count > svc_rpc_gss_client_max && client !=
NULL) {
svc_rpc_gss_forget_client_locked(client);
sx_xunlock(&svc_rpc_gss_lock);
svc_rpc_gss_release_client(client);
_______________________________________________
[email protected] mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "[email protected]"
