From: lhh@sourceware.org <lhh@sourceware.org>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] cluster/rgmanager ChangeLog make/defines.mk.in ...
Date: 27 Mar 2007 19:33:24 -0000 [thread overview]
Message-ID: <20070327193324.22260.qmail@sourceware.org> (raw)
CVSROOT: /cvs/cluster
Module name: cluster
Changes by: lhh at sourceware.org 2007-03-27 20:33:21
Modified files:
rgmanager : ChangeLog
rgmanager/make : defines.mk.input
rgmanager/src/clulib: Makefile clulog.c vft.c
rgmanager/src/daemons: groups.c main.c nodeevent.c resrules.c
rg_locks.c rg_thread.c
rgmanager/src/utils: clulog.c
Added files:
rgmanager/src/clulib: wrap_lock.c
Log message:
Merge patch from Crosswalk team
Team: Leonard Maiorani, Scott Cannata, Henry Harris
* Always check malloc() return codes
* Fix errant clu_unlock() calls in vft.c in cases where clu_lock() failed
* Add ability to wrap pthread_mutex / pthread_rwlock calls for better
stability
* Fix improper pthread_mutex_destroy() semantics
Patches:
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/ChangeLog.diff?cvsroot=cluster&r1=1.31&r2=1.32
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/make/defines.mk.input.diff?cvsroot=cluster&r1=1.6&r2=1.7
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/clulib/wrap_lock.c.diff?cvsroot=cluster&r1=NONE&r2=1.1
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/clulib/Makefile.diff?cvsroot=cluster&r1=1.11&r2=1.12
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/clulib/clulog.c.diff?cvsroot=cluster&r1=1.5&r2=1.6
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/clulib/vft.c.diff?cvsroot=cluster&r1=1.17&r2=1.18
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/daemons/groups.c.diff?cvsroot=cluster&r1=1.29&r2=1.30
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/daemons/main.c.diff?cvsroot=cluster&r1=1.35&r2=1.36
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/daemons/nodeevent.c.diff?cvsroot=cluster&r1=1.5&r2=1.6
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/daemons/resrules.c.diff?cvsroot=cluster&r1=1.19&r2=1.20
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/daemons/rg_locks.c.diff?cvsroot=cluster&r1=1.8&r2=1.9
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/daemons/rg_thread.c.diff?cvsroot=cluster&r1=1.18&r2=1.19
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/rgmanager/src/utils/clulog.c.diff?cvsroot=cluster&r1=1.3&r2=1.4
--- cluster/rgmanager/ChangeLog 2006/10/25 15:39:41 1.31
+++ cluster/rgmanager/ChangeLog 2007/03/27 19:33:19 1.32
@@ -1,3 +1,11 @@
+2007-03-27 Lon Hohberger <lhh@redhat.com>
+ * Merge patch from Crosswalk development team:
+ * Scott Cannata
+ * Henry Harris
+ * Leonard Maiorani
+
+=====
+
2006-10-25 Lon Hohberger <lhh@redhat.com>
* src/resources/clusterfs.sh: Fix unmounting problem (#212074)
--- cluster/rgmanager/make/defines.mk.input 2007/03/22 23:24:58 1.6
+++ cluster/rgmanager/make/defines.mk.input 2007/03/27 19:33:19 1.7
@@ -32,3 +32,11 @@
# Default CFLAGS
CFLAGS += -DSHAREDIR=\"@SHAREDIR@\" -Wall -Werror -Wstrict-prototypes -Wshadow ${INCLUDE}
+
+#
+# Wrappers around pthread_mutex / pthread_rwlock calls for deadlock
+# detection (and other things)
+#
+#CFLAGS += -DSHAREDIR=\"@SHAREDIR@\" -Wall -Werror -Wstrict-prototypes -Wshadow ${INCLUDE} -DWRAP_LOCKS
+#LDFLAGS += -Wl,-wrap,pthread_mutex_lock,-wrap,pthread_mutex_unlock,-wrap,pthread_rwlock_rdlock,-wrap,pthread_rwlock_wrlock,-wrap,pthread_rwlock_unlock
+
/cvs/cluster/cluster/rgmanager/src/clulib/wrap_lock.c,v --> standard output
revision 1.1
--- cluster/rgmanager/src/clulib/wrap_lock.c
+++ - 2007-03-27 20:33:21.897905000 +0100
@@ -0,0 +1,224 @@
+/*
+ Copyright Red Hat, Inc. 2007
+ Copyright Crosswalk 2006-2007
+
+ This program is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2, or (at your option) any
+ later version.
+
+ This program is distributed in the hope that it will be useful, but
+ WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; see the file COPYING. If not, write to the
+ Free Software Foundation, Inc., 675 Mass Ave, Cambridge,
+ MA 02139, USA.
+*/
+#ifdef WRAP_LOCKS
+#include <stdio.h>
+#include <sys/types.h>
+#include <gettid.h>
+#include <pthread.h>
+#include <string.h>
+#include <errno.h>
+#include <string.h>
+#include <signal.h>
+
+int __real_pthread_mutex_lock(pthread_mutex_t *lock);
+int
+__wrap_pthread_mutex_lock(pthread_mutex_t *lock)
+{
+ int status;
+ struct timespec delay;
+
+ while (1) {
+ status = __real_pthread_mutex_lock(lock);
+
+ switch(status) {
+ case EDEADLK:
+ /* Already own it: Note the error, but continue */
+ fprintf(stderr, "[%d] %s(%p): %s; continuing\n",
+ gettid(),
+ __FUNCTION__, lock, strerror(status));
+ /* deliberate fallthrough */
+ case 0:
+ return 0;
+ case EBUSY:
+ /* Try again */
+ break;
+ default:
+ /* Other return codes */
+ fprintf(stderr, "[%d] %s(%p): %s\n", gettid(),
+ __FUNCTION__, lock, strerror(status));
+ raise(SIGSEGV);
+ /* EINVAL? */
+ return 0;
+ }
+
+ delay.tv_sec = 0;
+ delay.tv_nsec = 100000;
+ nanosleep(&delay, NULL);
+ }
+
+ /* Not reached */
+ return 0;
+}
+
+
+int __real_pthread_mutex_unlock(pthread_mutex_t *lock);
+int
+__wrap_pthread_mutex_unlock(pthread_mutex_t *lock)
+{
+ int status;
+ struct timespec delay;
+
+ while (1) {
+ status = __real_pthread_mutex_unlock(lock);
+
+ switch(status) {
+ case EPERM:
+ /* Don't own it: Note the error, but continue */
+ fprintf(stderr, "[%d] %s(%p): %s; continuing\n",
+ gettid(),
+ __FUNCTION__, lock, strerror(status));
+ /* deliberate fallthrough */
+ case 0:
+ return 0;
+ default:
+ fprintf(stderr, "[%d] %s(%p): %s\n", gettid(),
+ __FUNCTION__, lock, strerror(status));
+ raise(SIGSEGV);
+ return 0;
+ }
+
+ delay.tv_sec = 0;
+ delay.tv_nsec = 100000;
+ nanosleep(&delay, NULL);
+ }
+
+ /* Not reached */
+ return 0;
+}
+
+
+int __real_pthread_rwlock_rdlock(pthread_rwlock_t *lock);
+int
+__wrap_pthread_rwlock_rdlock(pthread_rwlock_t *lock)
+{
+ int status;
+ struct timespec delay;
+
+ while (1) {
+ status = __real_pthread_rwlock_rdlock(lock);
+
+ switch(status) {
+ case EDEADLK:
+ /* Already own it: Note the error, but continue */
+ fprintf(stderr, "[%d] %s(%p): %s; continuing\n",
+ gettid(),
+ __FUNCTION__, lock, strerror(status));
+ /* deliberate fallthrough */
+ case 0:
+ return 0;
+ case EBUSY:
+ /* Try again */
+ break;
+ default:
+ /* Other return codes */
+ fprintf(stderr, "[%d] %s(%p): %s\n", gettid(),
+ __FUNCTION__, lock, strerror(status));
+ raise(SIGSEGV);
+ /* EINVAL? */
+ return 0;
+ }
+
+ delay.tv_sec = 0;
+ delay.tv_nsec = 100000;
+ nanosleep(&delay, NULL);
+ }
+
+ /* Not reached */
+ return 0;
+}
+
+
+int __real_pthread_rwlock_wrlock(pthread_rwlock_t *lock);
+int
+__wrap_pthread_rwlock_wrlock(pthread_rwlock_t *lock)
+{
+ int status;
+ struct timespec delay;
+
+ while (1) {
+ status = __real_pthread_rwlock_wrlock(lock);
+
+ switch(status) {
+ case EDEADLK:
+ /* Already own it: Note the error, but continue */
+ fprintf(stderr, "[%d] %s(%p): %s; continuing\n",
+ gettid(),
+ __FUNCTION__, lock, strerror(status));
+ /* deliberate fallthrough */
+ case 0:
+ return 0;
+ case EBUSY:
+ /* Try again */
+ break;
+ default:
+ /* Other return codes */
+ fprintf(stderr, "[%d] %s(%p): %s\n", gettid(),
+ __FUNCTION__, lock, strerror(status));
+ raise(SIGSEGV);
+ /* EINVAL? */
+ return 0;
+ }
+
+ delay.tv_sec = 0;
+ delay.tv_nsec = 100000;
+ nanosleep(&delay, NULL);
+ }
+
+ /* Not reached */
+ return 0;
+}
+
+
+int __real_pthread_rwlock_unlock(pthread_rwlock_t *lock);
+int
+__wrap_pthread_rwlock_unlock(pthread_rwlock_t *lock)
+{
+ int status;
+ struct timespec delay;
+
+ while (1) {
+ status = __real_pthread_rwlock_unlock(lock);
+
+ switch(status) {
+ case EPERM:
+ /* Don't own it: Note the error, but continue */
+ fprintf(stderr, "[%d] %s(%p): %s; continuing\n",
+ gettid(),
+ __FUNCTION__, lock, strerror(status));
+ /* deliberate fallthrough */
+ case 0:
+ return 0;
+ default:
+ fprintf(stderr, "[%d] %s(%p): %s\n", gettid(),
+ __FUNCTION__, lock, strerror(status));
+ raise(SIGSEGV);
+ return 0;
+ }
+
+ delay.tv_sec = 0;
+ delay.tv_nsec = 100000;
+ nanosleep(&delay, NULL);
+ }
+
+ /* Not reached */
+ return 0;
+}
+#endif
+
--- cluster/rgmanager/src/clulib/Makefile 2007/03/20 17:09:56 1.11
+++ cluster/rgmanager/src/clulib/Makefile 2007/03/27 19:33:20 1.12
@@ -33,7 +33,8 @@
libclulib.a: clulog.o daemon_init.o signals.o msgsimple.o \
gettid.o rg_strings.o message.o members.o fdops.o \
- lock.o cman.o vft.o msg_cluster.o msg_socket.o
+ lock.o cman.o vft.o msg_cluster.o msg_socket.o \
+ wrap_lock.o
${AR} cru $@ $^
ranlib $@
--- cluster/rgmanager/src/clulib/clulog.c 2006/06/02 17:37:10 1.5
+++ cluster/rgmanager/src/clulib/clulog.c 2007/03/27 19:33:20 1.6
@@ -20,7 +20,7 @@
/** @file
* Library routines for communicating with the logging daemon.
*
- * $Id: clulog.c,v 1.5 2006/06/02 17:37:10 lhh Exp $
+ * $Id: clulog.c,v 1.6 2007/03/27 19:33:20 lhh Exp $
*
* Author: Jeff Moyer <moyer@missioncriticallinux.com>
*/
@@ -50,7 +50,7 @@
#include <string.h>
-static const char *version __attribute__ ((unused)) = "$Revision: 1.5 $";
+static const char *version __attribute__ ((unused)) = "$Revision: 1.6 $";
#ifdef DEBUG
#include <assert.h>
@@ -70,7 +70,12 @@
static int syslog_facility = LOG_DAEMON;
static char *daemon_name = NULL;
static pid_t daemon_pid = -1;
+
+#ifdef WRAP_LOCKS
+static pthread_mutex_t log_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+#else
static pthread_mutex_t log_mutex = PTHREAD_MUTEX_INITIALIZER;
+#endif
CODE logger_prioritynames[] =
{ {"emerg", LOG_EMERG},
--- cluster/rgmanager/src/clulib/vft.c 2006/10/23 22:47:00 1.17
+++ cluster/rgmanager/src/clulib/vft.c 2007/03/27 19:33:20 1.18
@@ -50,8 +50,13 @@
* TODO: We could make it thread safe, but this might be unnecessary work
* Solution: Super-coarse-grained-bad-code-locking!
*/
+#ifdef WRAP_LOCKS
+static pthread_mutex_t key_list_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+static pthread_mutex_t vf_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+#else
static pthread_mutex_t key_list_mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_mutex_t vf_mutex = PTHREAD_MUTEX_INITIALIZER;
+#endif /* WRAP_LOCKS */
static pthread_t vf_thread = (pthread_t)-1;
static int vf_thread_ready = 0;
static vf_vote_cb_t default_vote_cb = NULL;
@@ -788,6 +793,15 @@
free(key_node->kn_data);
key_node->kn_datalen = vnp->vn_datalen;
key_node->kn_data = malloc(vnp->vn_datalen);
+
+ /*
+ * Need to check return of malloc always
+ */
+ if (key_node->kn_data == NULL) {
+ fprintf (stderr, "malloc fail err=%d\n", errno);
+ return -1;
+ }
+
memcpy(key_node->kn_data, vnp->vn_data, vnp->vn_datalen);
free(vnp);
@@ -1013,6 +1027,13 @@
}
newnode = malloc(sizeof(*newnode));
+
+ if (newnode == NULL) {
+ fprintf(stderr, "malloc fail3 err=%d\n", errno);
+ pthread_mutex_unlock(&key_list_mutex);
+ return -1;
+ }
+
newnode->kn_data = NULL;
memset(newnode,0,sizeof(*newnode));
newnode->kn_keyid = strdup(keyid);
@@ -1136,7 +1157,6 @@
snprintf(lock_name, sizeof(lock_name), "usrm::vf");
l = clu_lock(LKM_EXMODE, &lockp, 0, lock_name);
if (l < 0) {
- clu_unlock(&lockp);
pthread_mutex_unlock(&vf_mutex);
return l;
}
@@ -1153,7 +1173,7 @@
}
#ifdef DEBUG
- printf("aight, need responses from %d guys\n", remain);
+ printf("Allright, need responses from %d members\n", remain);
#endif
pthread_mutex_lock(&key_list_mutex);
@@ -1195,6 +1215,8 @@
*/
if (msg_open(MSG_CLUSTER, 0, _port, &everyone, 0) < 0) {
printf("msg_open: fail: %s\n", strerror(errno));
+ clu_unlock(&lockp);
+ pthread_mutex_unlock(&vf_mutex);
return -1;
}
@@ -1432,9 +1454,7 @@
snprintf(lock_name, sizeof(lock_name), "usrm::vf");
l = clu_lock(LKM_EXMODE, &lockp, 0, lock_name);
if (l < 0) {
- clu_unlock(&lockp);
pthread_mutex_unlock(&vf_mutex);
- printf("Couldn't lock %s\n", keyid);
return l;
}
--- cluster/rgmanager/src/daemons/groups.c 2007/03/20 17:09:57 1.29
+++ cluster/rgmanager/src/daemons/groups.c 2007/03/27 19:33:20 1.30
@@ -42,8 +42,13 @@
static resource_node_t *_tree = NULL;
static fod_t *_domains = NULL;
+#ifdef WRAP_LOCKS
+pthread_mutex_t config_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+pthread_mutex_t status_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+#else
pthread_mutex_t config_mutex = PTHREAD_MUTEX_INITIALIZER;
pthread_mutex_t status_mutex = PTHREAD_MUTEX_INITIALIZER;
+#endif
pthread_rwlock_t resource_lock = PTHREAD_RWLOCK_INITIALIZER;
void res_build_name(char *, size_t, resource_t *);
--- cluster/rgmanager/src/daemons/main.c 2007/03/20 17:09:57 1.35
+++ cluster/rgmanager/src/daemons/main.c 2007/03/27 19:33:20 1.36
@@ -403,7 +403,8 @@
/* Peek-a-boo */
sz = msg_receive(ctx, msg_hdr, sizeof(msgbuf), 1);
if (sz < sizeof (generic_msg_hdr)) {
- clulog(LOG_ERR, "#37: Error receiving message header (%d)\n", sz);
+ clulog(LOG_ERR,
+ "#37: Error receiving message header (%d)\n", sz);
goto out;
}
--- cluster/rgmanager/src/daemons/nodeevent.c 2007/03/20 17:09:57 1.5
+++ cluster/rgmanager/src/daemons/nodeevent.c 2007/03/27 19:33:20 1.6
@@ -35,8 +35,12 @@
/**
* Node event queue.
*/
-static nevent_t *event_queue = NULL;
+#ifdef WRAP_LOCKS
+static pthread_mutex_t ne_queue_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+#else
static pthread_mutex_t ne_queue_mutex = PTHREAD_MUTEX_INITIALIZER;
+#endif
+static nevent_t *event_queue = NULL;
static pthread_t ne_thread = 0;
int ne_queue_request(int local, int nodeid, int state);
--- cluster/rgmanager/src/daemons/resrules.c 2007/03/22 23:46:58 1.19
+++ cluster/rgmanager/src/daemons/resrules.c 2007/03/27 19:33:20 1.20
@@ -815,6 +815,10 @@
n = read(fd, buf, sizeof(buf));
if (n < 0) {
+
+ if (errno == EINTR)
+ continue;
+
if (*file)
free(*file);
return -1;
--- cluster/rgmanager/src/daemons/rg_locks.c 2006/12/18 21:55:27 1.8
+++ cluster/rgmanager/src/daemons/rg_locks.c 2007/03/27 19:33:20 1.9
@@ -36,11 +36,17 @@
static int _rg_statuscnt = 0;
static int _rg_statusmax = 5; /* XXX */
-static pthread_mutex_t locks_mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_cond_t unlock_cond = PTHREAD_COND_INITIALIZER;
static pthread_cond_t zero_cond = PTHREAD_COND_INITIALIZER;
static pthread_cond_t init_cond = PTHREAD_COND_INITIALIZER;
+
+#ifdef WRAP_LOCKS
+static pthread_mutex_t locks_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+static pthread_mutex_t _ccs_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+#else
+static pthread_mutex_t locks_mutex = PTHREAD_MUTEX_INITIALIZER;
static pthread_mutex_t _ccs_mutex = PTHREAD_MUTEX_INITIALIZER;
+#endif
#ifdef NO_CCS
static xmlDocPtr ccs_doc = NULL;
@@ -317,4 +323,3 @@
pthread_mutex_unlock(&locks_mutex);
return 0;
}
-
--- cluster/rgmanager/src/daemons/rg_thread.c 2007/03/10 00:20:54 1.18
+++ cluster/rgmanager/src/daemons/rg_thread.c 2007/03/27 19:33:20 1.19
@@ -42,8 +42,12 @@
* Resource thread queue head.
*/
static resthread_t *resthread_list = NULL;
-static pthread_mutex_t reslist_mutex = PTHREAD_MUTEX_INITIALIZER;
+#ifdef WRAP_LOCKS
+static pthread_mutex_t reslist_mutex = PTHREAD_ERRORCHECK_MUTEX_INITIALIZER_NP;
+#else
+static pthread_mutex_t reslist_mutex = PTHREAD_MUTEX_INITIALIZER;
+#endif
static resthread_t *find_resthread_byname(const char *resgroupname);
static int spawn_if_needed(const char *resgroupname);
@@ -163,7 +167,7 @@
char myname[256];
resthread_t *myself;
request_t *req;
- int ret = RG_EFAIL, error = 0;
+ int ret = RG_EFAIL, error = 0, mystatus;
rg_inc_threads();
@@ -453,7 +457,22 @@
raise(SIGSEGV);
}
- pthread_mutex_destroy(&my_queue_mutex);
+ mystatus = pthread_mutex_destroy(&my_queue_mutex);
+ if (mystatus != 0)
+ {
+ if (mystatus == EBUSY) {
+ pthread_mutex_unlock(&my_queue_mutex);
+ }
+
+ mystatus = pthread_mutex_destroy(&my_queue_mutex);
+ if (mystatus != 0) {
+ fprintf (stderr, "mutex_destroy=%d err=%d %p\n",
+ mystatus, errno, &my_queue_mutex);
+
+ fflush (stderr);
+ }
+ }
+
list_remove(&resthread_list, myself);
free(myself);
--- cluster/rgmanager/src/utils/clulog.c 2006/08/18 20:33:24 1.3
+++ cluster/rgmanager/src/utils/clulog.c 2007/03/27 19:33:20 1.4
@@ -122,6 +122,12 @@
/* Add two bytes for linefeed and NULL terminator */
len = strlen(argv[argc-1]) + 2;
logmsg = (char*)malloc(strlen(argv[argc-1])+2);
+ if (logmsg == NULL) {
+ fprintf(stderr,
+ "clulog: malloc fail err=%d\n", errno);
+ exit(0);
+ }
+
snprintf(logmsg, len, "%s\n", argv[argc-1]);
if (!cmdline_loglevel) {
reply other threads:[~2007-03-27 19:33 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20070327193324.22260.qmail@sourceware.org \
--to=lhh@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.