Copyright (C) Jeremy Allison 2001-2002
Copyright (C) Simo Sorce 2001
Copyright (C) Jim McDonough <jmcd@us.ibm.com> 2003
-
+ Copyright (C) James Peach 2006
+
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 of the License, or
enum protocol_types Protocol = PROTOCOL_COREPLUS;
/* a default finfo structure to ensure all fields are sensible */
-file_info def_finfo = {-1,0,0,0,0,0,0,"",""};
+file_info def_finfo;
/* this is used by the chaining code */
int chain_size = 0;
return True;
}
+/***********************************************************************
+ Free memory allocated to global objects
+***********************************************************************/
+
+void gfree_names(void)
+{
+ SAFE_FREE( smb_myname );
+ SAFE_FREE( smb_myworkgroup );
+ SAFE_FREE( smb_scope );
+ free_netbios_names_array();
+}
+
+void gfree_all( void )
+{
+ gfree_names();
+ gfree_loadparm();
+ gfree_case_tables();
+ gfree_debugsyms();
+ gfree_charcnv();
+ gfree_messages();
+
+ /* release the talloc null_context memory last */
+ talloc_disable_null_tracking();
+}
+
const char *my_netbios_names(int i)
{
return smb_my_netbios_names[i];
return;
}
- if (mem_ctx != NULL)
+ if (mem_ctx != NULL) {
*gids = TALLOC_REALLOC_ARRAY(mem_ctx, *gids, gid_t, *num_gids+1);
- else
+ } else {
*gids = SMB_REALLOC_ARRAY(*gids, gid_t, *num_gids+1);
+ }
- if (*gids == NULL)
+ if (*gids == NULL) {
return;
+ }
(*gids)[*num_gids] = gid;
*num_gids += 1;
(*num ) = NULL;
while ((p = Atoic(p, &val, ":,")) != NULL && (*p) != ':') {
- uint32 *tn;
-
- tn = SMB_REALLOC_ARRAY((*num), uint32, (*count)+1);
- if (tn == NULL) {
- SAFE_FREE(*num);
+ *num = SMB_REALLOC_ARRAY((*num), uint32, (*count)+1);
+ if (!(*num)) {
return NULL;
- } else
- (*num) = tn;
+ }
(*num)[(*count)] = val;
(*count)++;
p++;
int set_message(char *buf,int num_words,int num_bytes,BOOL zero)
{
- if (zero)
+ if (zero && (num_words || num_bytes)) {
memset(buf + smb_size,'\0',num_words*2 + num_bytes);
+ }
SCVAL(buf,smb_wct,num_words);
SSVAL(buf,smb_vwv + num_words*SIZEOFWORD,num_bytes);
smb_setlen(buf,smb_size + num_words*2 + num_bytes - 4);
size_t num_to_read_thistime;
size_t num_written = 0;
- if ((buf = SMB_MALLOC(TRANSFER_BUF_SIZE)) == NULL)
+ if ((buf = SMB_MALLOC_ARRAY(char, TRANSFER_BUF_SIZE)) == NULL)
return -1;
while (total < n) {
Become a daemon, discarding the controlling terminal.
****************************************************************************/
-void become_daemon(BOOL Fork)
+void become_daemon(BOOL Fork, BOOL no_process_group)
{
if (Fork) {
if (sys_fork()) {
/* detach from the terminal */
#ifdef HAVE_SETSID
- setsid();
+ if (!no_process_group) setsid();
#elif defined(TIOCNOTTY)
- {
+ if (!no_process_group) {
int i = sys_open("/dev/tty", O_RDWR, 0);
if (i != -1) {
ioctl(i, (int) TIOCNOTTY, (char *)0);
/****************************************************************************
Expand a pointer to be a particular size.
+ Note that this version of Realloc has an extra parameter that decides
+ whether to free the passed in storage on allocation failure or if the
+ new size is zero.
+
+ This is designed for use in the typical idiom of :
+
+ p = SMB_REALLOC(p, size)
+ if (!p) {
+ return error;
+ }
+
+ and not to have to keep track of the old 'p' contents to free later, nor
+ to worry if the size parameter was zero. In the case where NULL is returned
+ we guarentee that p has been freed.
+
+ If free later semantics are desired, then pass 'free_old_on_error' as False which
+ guarentees that the old contents are not freed on error, even if size == 0. To use
+ this idiom use :
+
+ tmp = SMB_REALLOC_KEEP_OLD_ON_ERROR(p, size);
+ if (!tmp) {
+ SAFE_FREE(p);
+ return error;
+ } else {
+ p = tmp;
+ }
+
+ Changes were instigated by Coverity error checking. JRA.
****************************************************************************/
-void *Realloc(void *p,size_t size)
+void *Realloc(void *p, size_t size, BOOL free_old_on_error)
{
void *ret=NULL;
if (size == 0) {
- SAFE_FREE(p);
- DEBUG(5,("Realloc asked for 0 bytes\n"));
+ if (free_old_on_error) {
+ SAFE_FREE(p);
+ }
+ DEBUG(2,("Realloc asked for 0 bytes\n"));
return NULL;
}
#if defined(PARANOID_MALLOC_CHECKER)
- if (!p)
+ if (!p) {
ret = (void *)malloc_(size);
- else
+ } else {
ret = (void *)realloc_(p,size);
+ }
#else
- if (!p)
+ if (!p) {
ret = (void *)malloc(size);
- else
+ } else {
ret = (void *)realloc(p,size);
+ }
#endif
- if (!ret)
+ if (!ret) {
+ if (free_old_on_error && p) {
+ SAFE_FREE(p);
+ }
DEBUG(0,("Memory allocation error: failed to expand to %d bytes\n",(int)size));
+ }
return(ret);
}
Type-safe realloc.
****************************************************************************/
-void *realloc_array(void *p,size_t el_size, unsigned int count)
+void *realloc_array(void *p, size_t el_size, unsigned int count, BOOL free_old_on_error)
{
if (count >= MAX_ALLOC_SIZE/el_size) {
+ if (free_old_on_error) {
+ SAFE_FREE(p);
+ }
return NULL;
}
- return Realloc(p,el_size*count);
+ return Realloc(p, el_size*count, free_old_on_error);
}
/****************************************************************************
- (Hopefully) efficient array append
+ (Hopefully) efficient array append.
****************************************************************************/
+
void add_to_large_array(TALLOC_CTX *mem_ctx, size_t element_size,
- void *element, void **array, uint32 *num_elements,
+ void *element, void *_array, uint32 *num_elements,
ssize_t *array_size)
{
- if (*array_size < 0)
+ void **array = (void **)_array;
+
+ if (*array_size < 0) {
return;
+ }
if (*array == NULL) {
if (*array_size == 0) {
goto error;
}
- if (mem_ctx != NULL)
+ if (mem_ctx != NULL) {
*array = TALLOC(mem_ctx, element_size * (*array_size));
- else
+ } else {
*array = SMB_MALLOC(element_size * (*array_size));
+ }
- if (*array == NULL)
+ if (*array == NULL) {
goto error;
+ }
}
if (*num_elements == *array_size) {
goto error;
}
- if (mem_ctx != NULL)
+ if (mem_ctx != NULL) {
*array = TALLOC_REALLOC(mem_ctx, *array,
element_size * (*array_size));
- else
+ } else {
*array = SMB_REALLOC(*array,
element_size * (*array_size));
+ }
- if (*array == NULL)
+ if (*array == NULL) {
goto error;
+ }
}
memcpy((char *)(*array) + element_size*(*num_elements),
BOOL process_exists(const struct process_id pid)
{
+ if (procid_is_me(&pid)) {
+ return True;
+ }
+
if (!procid_is_local(&pid)) {
/* This *SEVERELY* needs fixing. */
return True;
return (gid_t)-1;
}
-/*******************************************************************
- legacy wrapper for smb_panic2()
-********************************************************************/
-void smb_panic( const char *why )
-{
- smb_panic2( why, True );
-}
-
/*******************************************************************
Something really nasty happened - panic !
********************************************************************/
-#ifdef HAVE_LIBEXC_H
-#include <libexc.h>
-#endif
-
-void smb_panic2(const char *why, BOOL decrement_pid_count )
+void smb_panic(const char *const why)
{
char *cmd;
int result;
-#ifdef HAVE_BACKTRACE_SYMBOLS
- void *backtrace_stack[BACKTRACE_STACK_SIZE];
- size_t backtrace_size;
- char **backtrace_strings;
-#endif
#ifdef DEVELOPER
{
}
#endif
- /* only smbd needs to decrement the smbd counter in connections.tdb */
- if ( decrement_pid_count )
- decrement_smbd_process_count();
+ DEBUG(0,("PANIC (pid %llu): %s\n",
+ (unsigned long long)sys_getpid(), why));
+ log_stack_trace();
cmd = lp_panic_action();
if (cmd && *cmd) {
DEBUG(0, ("smb_panic(): action returned status %d\n",
WEXITSTATUS(result)));
}
- DEBUG(0,("PANIC: %s\n", why));
-#ifdef HAVE_BACKTRACE_SYMBOLS
+ dump_core();
+}
+
+/*******************************************************************
+ Print a backtrace of the stack to the debug log. This function
+ DELIBERATELY LEAKS MEMORY. The expectation is that you should
+ exit shortly after calling it.
+********************************************************************/
+
+#ifdef HAVE_LIBUNWIND_H
+#include <libunwind.h>
+#endif
+
+#ifdef HAVE_EXECINFO_H
+#include <execinfo.h>
+#endif
+
+#ifdef HAVE_LIBEXC_H
+#include <libexc.h>
+#endif
+
+void log_stack_trace(void)
+{
+#ifdef HAVE_LIBUNWIND
+ /* Try to use libunwind before any other technique since on ia64
+ * libunwind correctly walks the stack in more circumstances than
+ * backtrace.
+ */
+ unw_cursor_t cursor;
+ unw_context_t uc;
+ unsigned i = 0;
+
+ char procname[256];
+ unw_word_t ip, sp, off;
+
+ procname[sizeof(procname) - 1] = '\0';
+
+ if (unw_getcontext(&uc) != 0) {
+ goto libunwind_failed;
+ }
+
+ if (unw_init_local(&cursor, &uc) != 0) {
+ goto libunwind_failed;
+ }
+
+ DEBUG(0, ("BACKTRACE:\n"));
+
+ do {
+ ip = sp = 0;
+ unw_get_reg(&cursor, UNW_REG_IP, &ip);
+ unw_get_reg(&cursor, UNW_REG_SP, &sp);
+
+ switch (unw_get_proc_name(&cursor,
+ procname, sizeof(procname) - 1, &off) ) {
+ case 0:
+ /* Name found. */
+ case -UNW_ENOMEM:
+ /* Name truncated. */
+ DEBUGADD(0, (" #%u %s + %#llx [ip=%#llx] [sp=%#llx]\n",
+ i, procname, (long long)off,
+ (long long)ip, (long long) sp));
+ break;
+ default:
+ /* case -UNW_ENOINFO: */
+ /* case -UNW_EUNSPEC: */
+ /* No symbol name found. */
+ DEBUGADD(0, (" #%u %s [ip=%#llx] [sp=%#llx]\n",
+ i, "<unknown symbol>",
+ (long long)ip, (long long) sp));
+ }
+ ++i;
+ } while (unw_step(&cursor) > 0);
+
+ return;
+
+libunwind_failed:
+ DEBUG(0, ("unable to produce a stack trace with libunwind\n"));
+
+#elif HAVE_BACKTRACE_SYMBOLS
+ void *backtrace_stack[BACKTRACE_STACK_SIZE];
+ size_t backtrace_size;
+ char **backtrace_strings;
+
/* get the backtrace (stack frames) */
backtrace_size = backtrace(backtrace_stack,BACKTRACE_STACK_SIZE);
backtrace_strings = backtrace_symbols(backtrace_stack, backtrace_size);
#elif HAVE_LIBEXC
-#define NAMESIZE 32 /* Arbitrary */
-
/* The IRIX libexc library provides an API for unwinding the stack. See
* libexc(3) for details. Apparantly trace_back_stack leaks memory, but
* since we are about to abort anyway, it hardly matters.
- *
- * Note that if we paniced due to a SIGSEGV or SIGBUS (or similar) this
- * will fail with a nasty message upon failing to open the /proc entry.
*/
- {
- __uint64_t addrs[BACKTRACE_STACK_SIZE];
- char * names[BACKTRACE_STACK_SIZE];
- char namebuf[BACKTRACE_STACK_SIZE * NAMESIZE];
- int i;
- int levels;
+#define NAMESIZE 32 /* Arbitrary */
- ZERO_ARRAY(addrs);
- ZERO_ARRAY(names);
- ZERO_ARRAY(namebuf);
+ __uint64_t addrs[BACKTRACE_STACK_SIZE];
+ char * names[BACKTRACE_STACK_SIZE];
+ char namebuf[BACKTRACE_STACK_SIZE * NAMESIZE];
- /* We need to be root so we can open our /proc entry to walk
- * our stack. It also helps when we want to dump core.
- */
- become_root();
+ int i;
+ int levels;
- for (i = 0; i < BACKTRACE_STACK_SIZE; i++) {
- names[i] = namebuf + (i * NAMESIZE);
- }
+ ZERO_ARRAY(addrs);
+ ZERO_ARRAY(names);
+ ZERO_ARRAY(namebuf);
- levels = trace_back_stack(0, addrs, names,
- BACKTRACE_STACK_SIZE, NAMESIZE - 1);
+ /* We need to be root so we can open our /proc entry to walk
+ * our stack. It also helps when we want to dump core.
+ */
+ become_root();
- DEBUG(0, ("BACKTRACE: %d stack frames:\n", levels));
- for (i = 0; i < levels; i++) {
- DEBUGADD(0, (" #%d 0x%llx %s\n", i, addrs[i], names[i]));
- }
- }
+ for (i = 0; i < BACKTRACE_STACK_SIZE; i++) {
+ names[i] = namebuf + (i * NAMESIZE);
+ }
+
+ levels = trace_back_stack(0, addrs, names,
+ BACKTRACE_STACK_SIZE, NAMESIZE - 1);
+
+ DEBUG(0, ("BACKTRACE: %d stack frames:\n", levels));
+ for (i = 0; i < levels; i++) {
+ DEBUGADD(0, (" #%d 0x%llx %s\n", i, addrs[i], names[i]));
+ }
#undef NAMESIZE
-#endif
- dbgflush();
-#ifdef SIGABRT
- CatchSignal(SIGABRT,SIGNAL_CAST SIG_DFL);
+#else
+ DEBUG(0, ("unable to produce a stack trace on this platform\n"));
#endif
- abort();
}
/*******************************************************************
return(NULL);
#endif
-#ifdef HAVE_BROKEN_READDIR
+#ifdef HAVE_BROKEN_READDIR_NAME
/* using /usr/ucb/cc is BAD */
dname = dname - 2;
#endif
/****************************************************************************
Simple routine to do POSIX file locking. Cruft in NFS and 64->32 bit mapping
is dealt with in posix.c
+ Returns True if the lock was granted, False otherwise.
****************************************************************************/
BOOL fcntl_lock(int fd, int op, SMB_OFF_T offset, SMB_OFF_T count, int type)
SMB_STRUCT_FLOCK lock;
int ret;
- DEBUG(8,("fcntl_lock %d %d %.0f %.0f %d\n",fd,op,(double)offset,(double)count,type));
+ DEBUG(8,("fcntl_lock fd=%d op=%d offset=%.0f count=%.0f type=%d\n",
+ fd,op,(double)offset,(double)count,type));
lock.l_type = type;
lock.l_whence = SEEK_SET;
ret = sys_fcntl_ptr(fd,op,&lock);
- if (ret == -1 && errno != 0)
- DEBUG(3,("fcntl_lock: fcntl lock gave errno %d (%s)\n",errno,strerror(errno)));
-
- /* a lock query */
- if (op == SMB_F_GETLK) {
- if ((ret != -1) &&
- (lock.l_type != F_UNLCK) &&
- (lock.l_pid != 0) &&
- (lock.l_pid != sys_getpid())) {
- DEBUG(3,("fcntl_lock: fd %d is locked by pid %d\n",fd,(int)lock.l_pid));
- return(True);
- }
-
- /* it must be not locked or locked by me */
- return(False);
- }
-
- /* a lock set or unset */
if (ret == -1) {
+ int sav = errno;
DEBUG(3,("fcntl_lock: lock failed at offset %.0f count %.0f op %d type %d (%s)\n",
(double)offset,(double)count,op,type,strerror(errno)));
- return(False);
+ errno = sav;
+ return False;
}
/* everything went OK */
DEBUG(8,("fcntl_lock: Lock call successful\n"));
- return(True);
+ return True;
+}
+
+/****************************************************************************
+ Simple routine to query existing file locks. Cruft in NFS and 64->32 bit mapping
+ is dealt with in posix.c
+ Returns True if we have information regarding this lock region (and returns
+ F_UNLCK in *ptype if the region is unlocked). False if the call failed.
+****************************************************************************/
+
+BOOL fcntl_getlock(int fd, SMB_OFF_T *poffset, SMB_OFF_T *pcount, int *ptype, pid_t *ppid)
+{
+ SMB_STRUCT_FLOCK lock;
+ int ret;
+
+ DEBUG(8,("fcntl_getlock fd=%d offset=%.0f count=%.0f type=%d\n",
+ fd,(double)*poffset,(double)*pcount,*ptype));
+
+ lock.l_type = *ptype;
+ lock.l_whence = SEEK_SET;
+ lock.l_start = *poffset;
+ lock.l_len = *pcount;
+ lock.l_pid = 0;
+
+ ret = sys_fcntl_ptr(fd,SMB_F_GETLK,&lock);
+
+ if (ret == -1) {
+ int sav = errno;
+ DEBUG(3,("fcntl_getlock: lock request failed at offset %.0f count %.0f type %d (%s)\n",
+ (double)*poffset,(double)*pcount,*ptype,strerror(errno)));
+ errno = sav;
+ return False;
+ }
+
+ *ptype = lock.l_type;
+ *poffset = lock.l_start;
+ *pcount = lock.l_len;
+ *ppid = lock.l_pid;
+
+ DEBUG(3,("fcntl_getlock: fd %d is returned info %d pid %u\n",
+ fd, (int)lock.l_type, (unsigned int)lock.l_pid));
+ return True;
}
#undef DBGC_CLASS
void name_to_fqdn(fstring fqdn, const char *name)
{
struct hostent *hp = sys_gethostbyname(name);
+
if ( hp && hp->h_name && *hp->h_name ) {
- DEBUG(10,("name_to_fqdn: lookup for %s -> %s.\n", name, hp->h_name));
- fstrcpy(fqdn,hp->h_name);
+ char *full = NULL;
+
+ /* find out if the fqdn is returned as an alias
+ * to cope with /etc/hosts files where the first
+ * name is not the fqdn but the short name */
+ if (hp->h_aliases && (! strchr_m(hp->h_name, '.'))) {
+ int i;
+ for (i = 0; hp->h_aliases[i]; i++) {
+ if (strchr_m(hp->h_aliases[i], '.')) {
+ full = hp->h_aliases[i];
+ break;
+ }
+ }
+ }
+ if (full && (StrCaseCmp(full, "localhost.localdomain") == 0)) {
+ DEBUG(1, ("WARNING: your /etc/hosts file may be broken!\n"));
+ DEBUGADD(1, (" Specifing the machine hostname for address 127.0.0.1 may lead\n"));
+ DEBUGADD(1, (" to Kerberos authentication probelms as localhost.localdomain\n"));
+ DEBUGADD(1, (" may end up being used instead of the real machine FQDN.\n"));
+ full = hp->h_name;
+ }
+
+ if (!full) {
+ full = hp->h_name;
+ }
+
+ DEBUG(10,("name_to_fqdn: lookup for %s -> %s.\n", name, full));
+ fstrcpy(fqdn, full);
} else {
DEBUG(10,("name_to_fqdn: lookup for %s failed.\n", name));
fstrcpy(fqdn, name);
{
void *result;
- if (ptr == NULL)
- return NULL;
-
result = talloc_check_name(ptr, name);
if (result != NULL)
return result;
{
return True;
}
+
+int this_is_smp(void)
+{
+#if defined(HAVE_SYSCONF)
+
+#if defined(SYSCONF_SC_NPROC_ONLN)
+ return (sysconf(_SC_NPROC_ONLN) > 1) ? 1 : 0;
+#elif defined(SYSCONF_SC_NPROCESSORS_ONLN)
+ return (sysconf(_SC_NPROCESSORS_ONLN) > 1) ? 1 : 0;
+#else
+ return 0;
+#endif
+
+#else
+ return 0;
+#endif
+}
git.samba.org / metze / old / v3-2-winbind-ndr.git / blobdiff