* directory or online at http://www.openafs.org/dl/license10.html
*/
+/*
+ * A general description of the supergroup changes
+ * to ptserver:
+ *
+ * In AFS users can be members of groups. When you add
+ * a user, u1, to a group, g1, you add the user id for u1
+ * to an entries list in g1. This is a list of all the
+ * members of g1.
+ * You also add the id for g1 to an entries list in u1.
+ * This is a list of all the groups this user is a
+ * member of.
+ * The list has room for 10 entries. If more are required,
+ * a continuation record is created.
+ *
+ * With UMICH mods, u1 can be a group. When u1 is a group
+ * a new field is required to list the groups this group
+ * is a member of (since the entries field is used to
+ * list it's members). This new field is supergroups and
+ * has two entries. If more are required, a continuation
+ * record is formed.
+ * There are two additional fields required, nextsg is
+ * an address of the next continuation record for this
+ * group, and countsg is the count for the number of
+ * groups this group is a member of.
+ *
+ *
+ *
+ * 09/18/95 jjm Add mdw's changes to afs-3.3a Changes:
+ * (1) Add the parameter -groupdepth or -depth to
+ * define the maximum search depth for supergroups.
+ * Define the variable depthsg to be the value of
+ * the parameter. The default value is set to 5.
+ *
+ * (3) Make sure the sizes of the struct prentry and
+ * struct prentryg are equal. If they aren't equal
+ * the pt database will be corrupted.
+ * The error is reported with an fprintf statement,
+ * but this doesn't print when ptserver is started by
+ * bos, so all you see is an error message in the
+ * /usr/afs/logs/BosLog file. If you start the
+ * ptserver without bos the fprintf will print
+ * on stdout.
+ * The program will terminate with a PT_EXIT(1).
+ *
+ *
+ * Transarc does not currently use opcodes past 520, but
+ * they *could* decide at any time to use more opcodes.
+ * If they did, then one part of our local mods,
+ * ListSupergroups, would break. I've therefore
+ * renumbered it to 530, and put logic in to enable the
+ * old opcode to work (for now).
+ *
+ * 2/1/98 jjm Add mdw's changes for bit mapping for supergroups
+ * Overview:
+ * Before fetching a supergroup, this version of ptserver
+ * checks to see if it was marked "found" and "not a
+ * member". If and only if so, it doesn't fetch the group.
+ * Since this should be the case with most groups, this
+ * should save a significant amount of CPU in redundant
+ * fetches of the same group. After fetching the group,
+ * it sets "found", and either sets or clears "not a
+ * member", depending on if the group was a member of
+ * other groups. When it writes group entries to the
+ * database, it clears the "found" flag.
+ */
+
+#if defined(SUPERGROUPS)
+/*
+ * A general description of the supergroup changes
+ * to ptserver:
+ *
+ * In AFS users can be members of groups. When you add a user, u1,
+ * to a group, g1, you add the user id for u1 to an entries list
+ * in g1. This is a list of all the members of g1. You also add
+ * the id for g1 to an entries list in u1. This is a list of all
+ * the groups this user is a member of. The list has room for 10
+ * entries. If more are required, a continuation record is created.
+ *
+ * With UMICH mods, u1 can be a group. When u1 is a group a new
+ * field is required to list the groups this group is a member of
+ * (since the entries field is used to list it's members). This
+ * new field is supergroups and has two entries. If more are
+ * required, a continuation record is formed.
+ *
+ * There are two additional fields required, nextsg is an address
+ * of the next continuation record for this group, and countsg is
+ * the count for the number of groups this group is a member of.
+ *
+ * Bit mapping support for supergroups:
+ *
+ * Before fetching a supergroup, this version of ptserver checks to
+ * see if it was marked "found" and "not a member". If and only if
+ * so, it doesn't fetch the group. Since this should be the case
+ * with most groups, this should save a significant amount of CPU in
+ * redundant fetches of the same group. After fetching the group, it
+ * sets "found", and either sets or clears "not a member", depending
+ * on if the group was a member of other groups. When it writes
+ * group entries to the database, it clears the "found" flag.
+ */
+#endif
+
#include <afsconfig.h>
#include <afs/param.h>
-RCSID("$Header$");
+RCSID
+ ("$Header$");
#include <afs/stds.h>
#ifdef AFS_AIX32_ENV
#endif
#include <sys/types.h>
#include <stdio.h>
-#ifdef AFS_NT40_ENV
+#ifdef AFS_NT40_ENV
#include <winsock2.h>
#include <WINNT/afsevent.h>
#else
#include <netdb.h>
#include <netinet/in.h>
#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#else
+#ifdef HAVE_STRINGS_H
+#include <strings.h>
+#endif
+#endif
#include <rx/xdr.h>
#include <rx/rx.h>
#include <rx/rx_globals.h>
struct ubik_dbase *dbase;
struct afsconf_dir *prdir;
-extern afs_int32 ubik_lastYesTime;
-extern afs_int32 ubik_nBuffers;
+#if defined(SUPERGROUPS)
+extern afs_int32 depthsg;
+#endif
extern int afsconf_ServerAuth();
extern int afsconf_CheckAuth();
-int pr_realmNameLen;
+int pr_realmNameLen;
char *pr_realmName;
+static struct afsconf_cell info;
+
+extern int prp_group_default;
+extern int prp_user_default;
+
#include "AFS_component_version_number.c"
+int
+prp_access_mask(s)
+ char *s;
+{
+ int r;
+ if (*s >= '0' && *s <= '9') {
+ return strtol(s, NULL, 0);
+ }
+ r = 0;
+ while (*s) switch(*s++)
+ {
+ case 'S': r |= PRP_STATUS_ANY; break;
+ case 's': r |= PRP_STATUS_MEM; break;
+ case 'O': r |= PRP_OWNED_ANY; break;
+ case 'M': r |= PRP_MEMBER_ANY; break;
+ case 'm': r |= PRP_MEMBER_MEM; break;
+ case 'A': r |= PRP_ADD_ANY; break;
+ case 'a': r |= PRP_ADD_MEM; break;
+ case 'r': r |= PRP_REMOVE_MEM; break;
+ }
+ return r;
+}
+
/* check whether caller is authorized to manage RX statistics */
-int pr_rxstat_userok(call)
- struct rx_call *call;
+int
+pr_rxstat_userok(call)
+ struct rx_call *call;
{
- return afsconf_SuperUser(prdir, call, (char *)0);
+ return afsconf_SuperUser(prdir, call, NULL);
}
-void main (argc, argv)
- int argc;
- char **argv;
+int
+main(int argc, char **argv)
{
register afs_int32 code;
afs_int32 myHost;
char hostname[64];
struct rx_service *tservice;
struct rx_securityClass *sc[3];
- extern struct rx_securityClass *rxnull_NewServerSecurityObject();
- extern struct rx_securityClass *rxkad_NewServerSecurityObject();
extern int RXSTATS_ExecuteRequest();
extern int PR_ExecuteRequest();
#if 0
struct ktc_encryptionKey tkey;
#endif
- struct afsconf_cell info;
- int kerberosKeys; /* set if found some keys */
- afs_int32 i,j;
+ int kerberosKeys; /* set if found some keys */
int lwps = 3;
char clones[MAXHOSTSPERCELL];
const char *pr_dbaseName;
char *whoami = "ptserver";
- int a;
- char arg[100];
-
+ int a;
+ char arg[100];
+
#ifdef AFS_AIX32_ENV
/*
* The following signal action for AIX is necessary so that in case of a
* generated which, in many cases, isn't too useful.
*/
struct sigaction nsa;
-
+
sigemptyset(&nsa.sa_mask);
nsa.sa_handler = SIG_DFL;
nsa.sa_flags = SA_FULLDUMP;
sigaction(SIGABRT, &nsa, NULL);
sigaction(SIGSEGV, &nsa, NULL);
#endif
- osi_audit (PTS_StartEvent, 0, AUD_END);
+ osi_audit(PTS_StartEvent, 0, AUD_END);
/* Initialize dirpaths */
if (!(initAFSDirPath() & AFSDIR_SERVER_PATHS_OK)) {
#ifdef AFS_NT40_ENV
- ReportErrorEventAlt(AFSEVT_SVR_NO_INSTALL_DIR, 0, argv[0],0);
+ ReportErrorEventAlt(AFSEVT_SVR_NO_INSTALL_DIR, 0, argv[0], 0);
#endif
- fprintf(stderr,"%s: Unable to obtain AFS server directory.\n", argv[0]);
+ fprintf(stderr, "%s: Unable to obtain AFS server directory.\n",
+ argv[0]);
exit(2);
}
pr_dbaseName = AFSDIR_SERVER_PRDB_FILEPATH;
- for (a=1; a<argc; a++) {
+#if defined(SUPERGROUPS)
+ /* make sure the structures for database records are the same size */
+ if ((sizeof(struct prentry) != ENTRYSIZE)
+ || (sizeof(struct prentryg) != ENTRYSIZE)) {
+ fprintf(stderr,
+ "The structures for the database records are different"
+ " sizes\n" "struct prentry = %d\n" "struct prentryg = %d\n"
+ "ENTRYSIZE = %d\n", sizeof(struct prentry),
+ sizeof(struct prentryg), ENTRYSIZE);
+ PT_EXIT(1);
+ }
+#endif
+
+ for (a = 1; a < argc; a++) {
int alen;
- lcstring (arg, argv[a], sizeof(arg));
- alen = strlen (arg);
- if ((strncmp (arg, "-database", alen) == 0) ||
- (strncmp (arg, "-db", alen) == 0)) {
+ lcstring(arg, argv[a], sizeof(arg));
+ alen = strlen(arg);
+ if ((strncmp(arg, "-database", alen) == 0)
+ || (strncmp(arg, "-db", alen) == 0)) {
pr_dbaseName = argv[++a]; /* specify a database */
+ } else if (strncmp(arg, "-p", alen) == 0) {
+ lwps = atoi(argv[++a]);
+ if (lwps > 16) { /* maximum of 16 */
+ printf("Warning: '-p %d' is too big; using %d instead\n",
+ lwps, 16);
+ lwps = 16;
+ } else if (lwps < 3) { /* minimum of 3 */
+ printf("Warning: '-p %d' is too small; using %d instead\n",
+ lwps, 3);
+ lwps = 3;
+ }
+ } else if (strncmp(arg, "-rebuild", alen) == 0) /* rebuildDB++ */
+ ;
+#if defined(SUPERGROUPS)
+ else if ((strncmp(arg, "-groupdepth", alen) == 0)
+ || (strncmp(arg, "-depth", alen) == 0)) {
+ depthsg = atoi(argv[++a]); /* Max search depth for supergroups */
}
- else if (strncmp(arg, "-p", alen) == 0) {
- lwps = atoi(argv[++a]);
- if (lwps > 16) { /* maximum of 16 */
- printf("Warning: '-p %d' is too big; using %d instead\n",
- lwps, 16);
- lwps = 16;
- } else if (lwps < 3) { /* minimum of 3 */
- printf("Warning: '-p %d' is too small; using %d instead\n",
- lwps, 3);
- lwps = 3;
- }
+#endif
+ else if (strncmp(arg, "-default_access", alen) == 0) {
+ prp_user_default = prp_access_mask(argv[++a]);
+ prp_group_default = prp_access_mask(argv[++a]);
}
- else if (strncmp (arg, "-rebuild", alen) == 0) /* rebuildDB++ */ ;
- else if (strncmp (arg, "-enable_peer_stats", alen) == 0) {
+ else if (strncmp(arg, "-enable_peer_stats", alen) == 0) {
rx_enablePeerRPCStats();
- }
- else if (strncmp (arg, "-enable_process_stats", alen) == 0) {
+ } else if (strncmp(arg, "-enable_process_stats", alen) == 0) {
rx_enableProcessRPCStats();
}
#ifndef AFS_NT40_ENV
- else if (strncmp(arg, "-syslog", alen)==0) {
+ else if (strncmp(arg, "-syslog", alen) == 0) {
/* set syslog logging flag */
serverLogSyslog = 1;
- }
- else if (strncmp(arg, "-syslog=", MIN(8,alen))==0) {
+ } else if (strncmp(arg, "-syslog=", MIN(8, alen)) == 0) {
serverLogSyslog = 1;
- serverLogSyslogFacility = atoi(arg+8);
+ serverLogSyslogFacility = atoi(arg + 8);
}
#endif
else if (*arg == '-') {
- usage:
-
- /* hack in help flag support */
+ /* hack in help flag support */
+#if defined(SUPERGROUPS)
#ifndef AFS_NT40_ENV
- printf ("Usage: ptserver [-database <db path>] "
- "[-syslog[=FACILITY]] "
- "[-p <number of processes>] [-rebuild] "
- "[-enable_peer_stats] [-enable_process_stats] "
- "[-help]\n");
+ printf("Usage: ptserver [-database <db path>] "
+ "[-syslog[=FACILITY]] "
+ "[-p <number of processes>] [-rebuild] "
+ "[-groupdepth <depth>] "
+ "[-enable_peer_stats] [-enable_process_stats] "
+ "[-default_access default_user_access default_group_access] "
+ "[-help]\n");
+#else /* AFS_NT40_ENV */
+ printf("Usage: ptserver [-database <db path>] "
+ "[-p <number of processes>] [-rebuild] "
+ "[-default_access default_user_access default_group_access] "
+ "[-groupdepth <depth>] " "[-help]\n");
+#endif
#else
- printf ("Usage: ptserver [-database <db path>] "
- "[-p <number of processes>] [-rebuild] "
- "[-help]\n");
+#ifndef AFS_NT40_ENV
+ printf("Usage: ptserver [-database <db path>] "
+ "[-syslog[=FACILITY]] "
+ "[-p <number of processes>] [-rebuild] "
+ "[-enable_peer_stats] [-enable_process_stats] "
+ "[-default_access default_user_access default_group_access] "
+ "[-help]\n");
+#else /* AFS_NT40_ENV */
+ printf("Usage: ptserver [-database <db path>] "
+ "[-default_access default_user_access default_group_access] "
+ "[-p <number of processes>] [-rebuild] " "[-help]\n");
+#endif
#endif
- fflush(stdout);
+ fflush(stdout);
PT_EXIT(1);
}
+#if defined(SUPERGROUPS)
+ else {
+ fprintf(stderr, "Unrecognized arg: '%s' ignored!\n", arg);
+ }
+#endif
}
- OpenLog(AFSDIR_SERVER_PTLOG_FILEPATH); /* set up logging */
+#ifndef AFS_NT40_ENV
+ serverLogSyslogTag = "ptserver";
+#endif
+ OpenLog(AFSDIR_SERVER_PTLOG_FILEPATH); /* set up logging */
SetupLogSignals();
-
+
prdir = afsconf_Open(AFSDIR_SERVER_ETC_DIRPATH);
if (!prdir) {
- fprintf (stderr, "ptserver: can't open configuration directory.\n");
+ fprintf(stderr, "ptserver: can't open configuration directory.\n");
PT_EXIT(1);
}
if (afsconf_GetNoAuthFlag(prdir))
- printf ("ptserver: running unauthenticated\n");
+ printf("ptserver: running unauthenticated\n");
-#ifdef AFS_NT40_ENV
+#ifdef AFS_NT40_ENV
/* initialize winsock */
- if (afs_winsockInit()<0) {
- ReportErrorEventAlt(AFSEVT_SVR_WINSOCK_INIT_FAILED, 0,
- argv[0],0);
-
- fprintf(stderr, "ptserver: couldn't initialize winsock. \n");
- PT_EXIT(1);
+ if (afs_winsockInit() < 0) {
+ ReportErrorEventAlt(AFSEVT_SVR_WINSOCK_INIT_FAILED, 0, argv[0], 0);
+
+ fprintf(stderr, "ptserver: couldn't initialize winsock. \n");
+ PT_EXIT(1);
}
#endif
/* get this host */
- gethostname(hostname,sizeof(hostname));
+ gethostname(hostname, sizeof(hostname));
th = gethostbyname(hostname);
if (!th) {
- fprintf (stderr, "ptserver: couldn't get address of this host.\n");
+ fprintf(stderr, "ptserver: couldn't get address of this host.\n");
PT_EXIT(1);
}
memcpy(&myHost, th->h_addr, sizeof(afs_int32));
-
+
/* get list of servers */
- code = afsconf_GetExtendedCellInfo(prdir,(char *)0,"afsprot",
- &info, &clones);
+ code =
+ afsconf_GetExtendedCellInfo(prdir, NULL, "afsprot", &info, &clones);
if (code) {
- com_err (whoami, code, "Couldn't get server list");
+ com_err(whoami, code, "Couldn't get server list");
PT_EXIT(2);
}
pr_realmName = info.name;
- pr_realmNameLen = strlen (pr_realmName);
+ pr_realmNameLen = strlen(pr_realmName);
#if 0
/* get keys */
- code = afsconf_GetKey(prdir,999,&tkey);
+ code = afsconf_GetKey(prdir, 999, &tkey);
if (code) {
- com_err (whoami, code, "couldn't get bcrypt keys from key file, ignoring.");
+ com_err(whoami, code,
+ "couldn't get bcrypt keys from key file, ignoring.");
}
#endif
- { afs_int32 kvno; /* see if there is a KeyFile here */
+ {
+ afs_int32 kvno; /* see if there is a KeyFile here */
struct ktc_encryptionKey key;
- code = afsconf_GetLatestKey (prdir, &kvno, &key);
+ code = afsconf_GetLatestKey(prdir, &kvno, &key);
kerberosKeys = (code == 0);
if (!kerberosKeys)
- printf ("ptserver: can't find any Kerberos keys, code = %d, ignoring\n", code);
+ printf
+ ("ptserver: can't find any Kerberos keys, code = %d, ignoring\n",
+ code);
}
if (kerberosKeys) {
/* initialize ubik */
* CoEntry this adds up to as much as 1+1+39*3 = 119. If all these entries
* and the header are in separate Ubik buffers then 120 buffers may be
* required. */
- ubik_nBuffers = 120 + /*fudge*/40;
- code = ubik_ServerInitByInfo(myHost, htons(AFSCONF_PROTPORT), &info,
- &clones, pr_dbaseName, &dbase);
+ ubik_nBuffers = 120 + /*fudge */ 40;
+ code =
+ ubik_ServerInitByInfo(myHost, htons(AFSCONF_PROTPORT), &info, &clones,
+ pr_dbaseName, &dbase);
if (code) {
- com_err (whoami, code, "Ubik init failed");
+ com_err(whoami, code, "Ubik init failed");
PT_EXIT(2);
}
+#if defined(SUPERGROUPS)
+ pt_hook_write();
+#endif
+
sc[0] = rxnull_NewServerSecurityObject();
sc[1] = 0;
if (kerberosKeys) {
- sc[2] = rxkad_NewServerSecurityObject
- (0, prdir, afsconf_GetKey, (char *)0);
- }
- else sc[2] = sc[0];
+ sc[2] = rxkad_NewServerSecurityObject(0, prdir, afsconf_GetKey, NULL);
+ } else
+ sc[2] = sc[0];
- /* These two lines disallow jumbograms */
- rx_maxReceiveSize = OLD_MAX_PACKET_SIZE;
- rxi_nSendFrags = rxi_nRecvFrags = 1;
+ /* Disable jumbograms */
+ rx_SetNoJumbo();
- tservice = rx_NewService(0,PRSRV,"Protection Server",sc,3,PR_ExecuteRequest);
+ tservice =
+ rx_NewService(0, PRSRV, "Protection Server", sc, 3,
+ PR_ExecuteRequest);
if (tservice == (struct rx_service *)0) {
- fprintf (stderr, "ptserver: Could not create new rx service.\n");
+ fprintf(stderr, "ptserver: Could not create new rx service.\n");
PT_EXIT(3);
}
- rx_SetMinProcs(tservice,2);
- rx_SetMaxProcs(tservice,lwps);
+ rx_SetMinProcs(tservice, 2);
+ rx_SetMaxProcs(tservice, lwps);
- tservice = rx_NewService(0,RX_STATS_SERVICE_ID,"rpcstats",sc,3,RXSTATS_ExecuteRequest);
+ tservice =
+ rx_NewService(0, RX_STATS_SERVICE_ID, "rpcstats", sc, 3,
+ RXSTATS_ExecuteRequest);
if (tservice == (struct rx_service *)0) {
- fprintf (stderr, "ptserver: Could not create new rx service.\n");
+ fprintf(stderr, "ptserver: Could not create new rx service.\n");
PT_EXIT(3);
}
- rx_SetMinProcs(tservice,2);
- rx_SetMaxProcs(tservice,4);
+ rx_SetMinProcs(tservice, 2);
+ rx_SetMaxProcs(tservice, 4);
/* allow super users to manage RX statistics */
rx_SetRxStatUserOk(pr_rxstat_userok);
rx_StartServer(1);
- osi_audit (PTS_FinishEvent, -1, AUD_END);
+ osi_audit(PTS_FinishEvent, -1, AUD_END);
+ exit(0);
}