#include <fcntl.h>
#include <sys/stat.h>
#include <sys/types.h>
+#include <dirent.h>
#include <linux/errno.h>
-#include "LTTTypes.h"
+#include <ltt/LTTTypes.h>
#include "parser.h"
-#include <ltt/tracefile.h>
+#include <ltt/trace.h>
-#include "default.h" //yxx test
+#define DIR_NAME_SIZE 256
/* set the offset of the fields belonging to the event,
need the information of the archecture */
-void setFieldsOffset(ltt_tracefile * t, ltt_eventtype *evT, void *evD);
+void setFieldsOffset(LttTracefile * t, LttEventType *evT, void *evD);
/* get the size of the field type according to the archtecture's
size and endian type(info of the archecture) */
-int getFieldtypeSize(ltt_tracefile * t, ltt_eventtype * evT, int offsetRoot,
- int offsetParent, ltt_field * fld, void * evD );
+int getFieldtypeSize(LttTracefile * t, LttEventType * evT, int offsetRoot,
+ int offsetParent, LttField * fld, void * evD );
/* read a fixed size or a block information from the file (fd) */
int readFile(int fd, void * buf, size_t size, char * mesg);
-int readBlock(ltt_tracefile * tf, int whichBlock);
+int readBlock(LttTracefile * tf, int whichBlock);
/* calculate cycles per nsec for current block */
-void getCyclePerNsec(ltt_tracefile * t);
-
-/* functions to release the memory occupied by the tables of a tracefile */
-void freeKey (gpointer key, gpointer value, gpointer user_data);
-void freeValue (gpointer key, gpointer value, gpointer user_data);
-void freeFacilityNameTable (gpointer key, gpointer value, gpointer user_data);
+void getCyclePerNsec(LttTracefile * t);
/* reinitialize the info of the block which is already in the buffer */
-void updateTracefile(ltt_tracefile * tf);
+void updateTracefile(LttTracefile * tf);
/* go to the next event */
-int skipEvent(ltt_tracefile * t);
+int skipEvent(LttTracefile * t);
-/* compare two time (ltt_time), 0:t1=t2, -1:t1<t2, 1:t1>t2 */
-int timecmp(ltt_time * t1, ltt_time * t2);
+/* compare two time (LttTime), 0:t1=t2, -1:t1<t2, 1:t1>t2 */
+int timecmp(LttTime * t1, LttTime * t2);
/* get an integer number */
int getIntNumber(int size1, void *evD);
-/* Time operation macros for ltt_time (struct timespec) */
+/* Time operation macros for LttTime (struct timespec) */
/* (T3 = T2 - T1) */
#define TimeSub(T3, T2, T1) \
do \
/*****************************************************************************
*Function name
- * ltt_tracefile_open : open a trace file, construct a ltt_tracefile
+ * ltt_tracefile_open : open a trace file, construct a LttTracefile
*Input params
- * pathname : path name of the trace file
+ * t : the trace containing the tracefile
+ * fileName : path name of the trace file
*Return value
- * ltt_tracefile * : the structure represents the opened trace file
+ * : a pointer to a tracefile
****************************************************************************/
-ltt_tracefile *ltt_tracefile_open(char * fileName)
+LttTracefile* ltt_tracefile_open(LttTrace * t, char * fileName)
{
- ltt_tracefile * tf;
- struct stat lTDFStat; /* Trace data file status */
- trace_header_event *trace_header;
- block_header a_block_header;
+ LttTracefile * tf;
+ struct stat lTDFStat; /* Trace data file status */
+ BlockStart a_block_start;
- tf = g_new(ltt_tracefile, 1);
+ tf = g_new(LttTracefile, 1);
//open the file
+ tf->name = g_strdup(fileName);
+ tf->trace = t;
tf->fd = open(fileName, O_RDONLY, 0);
if(tf->fd < 0){
g_error("Unable to open input data file %s\n", fileName);
}
// Is the file large enough to contain a trace
- if(lTDFStat.st_size < sizeof(block_header) + sizeof(trace_header_event)){
+ if(lTDFStat.st_size < sizeof(BlockStart) + EVENT_HEADER_SIZE){
g_error("The input data file %s does not contain a trace\n", fileName);
}
//store the size of the file
tf->file_size = lTDFStat.st_size;
-
- //read the first block header
- if(readFile(tf->fd,(void*)&a_block_header, sizeof(block_header),
- "Unable to read block header"))
- exit(1);
-
- //read the trace header event
- trace_header = g_new(trace_header_event, 1);
- if(readFile(tf->fd,(void*)trace_header, sizeof(trace_header_event),
- "Unable to read trace header event"))
- exit(1);
-
- tf->block_number = tf->file_size / trace_header->buffer_size;
- tf->trace_header = (trace_header_event *)trace_header;
+ tf->block_size = t->system_description->ltt_block_size;
+ tf->block_number = tf->file_size / tf->block_size;
+ tf->which_block = 0;
//allocate memory to contain the info of a block
- tf->buffer = (void *) g_new(char, trace_header->buffer_size);
+ tf->buffer = (void *) g_new(char, t->system_description->ltt_block_size);
- //read the last block, get the trace end time
- if(readBlock(tf,tf->block_number)) exit(1);
- tf->end_time = tf->a_block_footer->time;
-
- //read the first block, get the trace start time
- if(tf->block_number != 1)
- if(readBlock(tf,1)) exit(1);
- tf->start_time = tf->a_block_header->time;
-
- //get local machine's data type size and endian type
- getDataEndianType(&(tf->my_arch_size), &(tf->my_arch_endian));
-
- //initialize all tables
- tf->eventtype_number = 0;
- tf->facility_number = 0;
- tf->index_facility = g_hash_table_new (g_int_hash, g_int_equal);
- tf->facility_name = g_hash_table_new (g_str_hash, g_str_equal);
- tf->base_id_name = g_hash_table_new (g_str_hash, g_int_equal);
- tf->eventtype_event_id = g_ptr_array_new();
+ //read the first block
+ if(readBlock(tf,1)) exit(1);
return tf;
}
-/*****************************************************************************
- *Function name
- * ltt_tracefile_close: close a trace file, release all facilities if their
- * usage count == 0 when close_facilities is true
- *Input params
- * t : tracefile which will be closed
- * close_facilities : bool to show if the facilities need to be released
- *Return value
- * int : ????
- ****************************************************************************/
-
-int ltt_tracefile_close(ltt_tracefile *t, int close_facilities)
-{
- //free index_facility table
- g_hash_table_foreach (t->index_facility, freeKey, NULL);
- g_hash_table_destroy(t->index_facility);
-
- //free base_id_facility table
- g_hash_table_foreach (t->base_id_name, freeValue, NULL);
- g_hash_table_destroy(t->base_id_name);
-
- //free eventtype_event_id array
- g_ptr_array_free (t->eventtype_event_id, TRUE);
-
- //free facility_name table
- g_hash_table_foreach (t->facility_name, freeFacilityNameTable,
- GINT_TO_POINTER(1));
- g_hash_table_destroy(t->facility_name);
-
- //free tracefile structure
- g_free(t->trace_header);
- g_free(t->buffer);
- g_free(t);
- return 0;
-}
/*****************************************************************************
- * functions to release the memory occupied by the tables of a tracefile
+ *Open control and per cpu tracefiles
****************************************************************************/
-void freeKey (gpointer key, gpointer value, gpointer user_data)
-{
- g_free(key);
-}
-
-void freeValue (gpointer key, gpointer value, gpointer user_data)
+void ltt_tracefile_open_cpu(LttTrace *t, char * tracefile_name)
{
- g_free(value);
+ LttTracefile * tf;
+ tf = ltt_tracefile_open(t,tracefile_name);
+ t->per_cpu_tracefile_number++;
+ g_ptr_array_add(t->per_cpu_tracefiles, tf);
}
-void freeFacilityNameTable (gpointer key, gpointer value, gpointer user_data)
+void ltt_tracefile_open_control(LttTrace *t, char * control_name)
{
- ltt_facility * fac = (ltt_facility*) value;
- fac->usage_count--;
- if(GPOINTER_TO_INT(user_data) != 0)
- ltt_facility_close(fac);
+ LttTracefile * tf;
+ LttEvent * ev;
+ LttFacility * f;
+ uint16_t evId;
+ void * pos;
+ FacilityLoad fLoad;
+ int i;
+
+ tf = ltt_tracefile_open(t,control_name);
+ t->control_tracefile_number++;
+ g_ptr_array_add(t->control_tracefiles,tf);
+
+ //parse facilities tracefile to get base_id
+ if(strcmp(control_name,"facilities") ==0){
+ while(1){
+ evId = *(uint16_t*)tf->cur_event_pos;
+ if(evId == TRACE_FACILITY_LOAD){
+ pos = tf->cur_event_pos + EVENT_HEADER_SIZE;
+ fLoad.name = (char*)pos;
+ fLoad.checksum = *(LttChecksum*)(pos + sizeof(pos));
+ fLoad.base_code = *(uint32_t*)(pos + sizeof(pos) + sizeof(LttChecksum));
+
+ for(i=0;i<t->facility_number;i++){
+ f = (LttFacility*)g_ptr_array_index(t->facilities,i);
+ if(strcmp(f->name,fLoad.name)==0 && fLoad.checksum==f->checksum){
+ f->base_id = fLoad.base_code;
+ break;
+ }
+ }
+ if(i==t->facility_number)
+ g_error("Facility: %s, checksum: %d is not founded\n",
+ fLoad.name,fLoad.checksum);
+
+ ev = ltt_tracefile_read(tf); //get next event
+ if(!ev) break; //end of tracefile
+ }else if(evId == TRACE_BLOCK_END){
+ //can only reach here if it is the first event
+ g_error("Facilities does not contain any facility_load event\n");
+ }else g_error("Not valid facilities trace file\n");
+ }
+ }
}
/*****************************************************************************
*Function name
- * ltt_tracefile_facility_add: increases the facility usage count and also
- * specifies the base of the numeric range
- * assigned to the event types in the facility
- * for this tracefile
+ * ltt_tracefile_close: close a trace file,
*Input params
- * t : tracefile that the facility will be added
- * f : facility that will be attached to the tracefile
- * base_id : the id for the first event of the facility in the
- * trace file
- *Return value
- * int : ????
+ * t : tracefile which will be closed
****************************************************************************/
-int ltt_tracefile_facility_add(ltt_tracefile *t,ltt_facility *f,int base_id)
+void ltt_tracefile_close(LttTracefile *t)
{
- int * id, *index;
- int i, j, k;
- ltt_eventtype * et;
- gpointer tmpPtr;
-
- //increase the facility usage count
- f->usage_count++;
- t->eventtype_number += f->event_number;
- t->facility_number++;
-
- //insert the facility into index_facility table
- id = g_new(int,1);
- *id = t->facility_number;
- g_hash_table_insert(t->index_facility, (gpointer)id,(gpointer)f);
-
- //insert facility name into table: base_id_name
- id = g_new(int,1);
- *id = base_id;
- g_hash_table_insert(t->base_id_name, (gpointer)(f->name), (gpointer)id);
+ g_free(t->name);
+ g_free(t->buffer);
+ g_free(t);
+}
- //insert facility name into table: facility_name
- g_hash_table_insert(t->facility_name, (gpointer)(f->name), (gpointer)f);
-
- //insert eventtype into the array: eventtype_event_id
- j = base_id + f->event_number;
- k = t->eventtype_event_id->len;
- if(j > t->eventtype_event_id->len){
- for(i=0; i < j - k; i++){
- tmpPtr = (gpointer)g_new(ptr_wrap, 1);
- g_ptr_array_add(t->eventtype_event_id,tmpPtr);
- }
- }
- //initialize the unused elements: NULL
- if(j-k > f->event_number){
- for(i=k; i<base_id; i++){
- tmpPtr = g_ptr_array_index(t->eventtype_event_id, i);
- ((ptr_wrap*)tmpPtr)->ptr = NULL;
- }
+/*****************************************************************************
+ *Get system information
+ ****************************************************************************/
+void getSystemInfo(LttSystemDescription* des, char * pathname)
+{
+ FILE * fp;
+ int i;
+ int entry_number = 15;
+ char buf[DIR_NAME_SIZE];
+ char description[4*DIR_NAME_SIZE];
+ char * ptr;
+
+ fp = fopen(pathname,"r");
+ if(!fp){
+ g_error("Can not open file : %s\n", pathname);
}
-
- //initialize the elements with the eventtypes belonging to the facility
- for(i=0; i<f->event_number; i++){
- tmpPtr = g_ptr_array_index(t->eventtype_event_id, base_id + i);
- ((ptr_wrap*)tmpPtr)->ptr = (gpointer)(f->events[i]);
+
+ while(fgets(buf,DIR_NAME_SIZE, fp)!= NULL){
+ ptr = buf;
+ while(isspace(*ptr)) ptr++;
+ if(strlen(ptr) == 0) continue;
+ break;
}
- //update offset
- for(i=0; i<f->event_number; i++){
- et = f->events[i];
- setFieldsOffset(t, et, NULL);
+ if(strlen(ptr) == 0) g_error("Not a valid file: %s\n", pathname);
+ if(strncmp("<system",ptr,7) !=0)g_error("Not a valid file: %s\n", pathname);
+
+ for(i=0;i<entry_number;i++){
+ if(fgets(buf,DIR_NAME_SIZE, fp)== NULL)
+ g_error("Not a valid file: %s\n", pathname);
+ ptr = buf;
+ while(isspace(*ptr)) ptr++;
+ switch(i){
+ case 0:
+ if(strncmp("node_name=",ptr,10)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->node_name = g_strdup(ptr+10);
+ break;
+ case 1:
+ if(strncmp("domainname=",ptr,11)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->domain_name = g_strdup(ptr+11);
+ break;
+ case 2:
+ if(strncmp("cpu=",ptr,4)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->nb_cpu = (unsigned)atoi(ptr+4);
+ break;
+ case 3:
+ if(strncmp("arch_size=",ptr,10)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ if(strcmp(ptr+10,"\"LP32\"") == 0) des->size = LTT_LP32;
+ else if(strcmp(ptr+10,"\"ILP32\"") == 0) des->size = LTT_ILP32;
+ else if(strcmp(ptr+10,"\"LP64\"") == 0) des->size = LTT_LP64;
+ else if(strcmp(ptr+10,"\"ILP64\"") == 0) des->size = LTT_ILP64;
+ else if(strcmp(ptr+10,"\"UNKNOWN\"") == 0) des->size = LTT_UNKNOWN;
+ break;
+ case 4:
+ if(strncmp("endian=",ptr,7)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ if(strcmp(ptr+7,"\"LITTLE_ENDIAN\"") == 0)
+ des->endian = LTT_LITTLE_ENDIAN;
+ else if(strcmp(ptr+7,"\"BIG_ENDIAN\"") == 0)
+ des->endian = LTT_BIG_ENDIAN;
+ break;
+ case 5:
+ if(strncmp("kernel_name=",ptr,12)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->kernel_name = g_strdup(ptr+12);
+ break;
+ case 6:
+ if(strncmp("kernel_release=",ptr,15)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->kernel_release = g_strdup(ptr+15);
+ break;
+ case 7:
+ if(strncmp("kernel_version=",ptr,15)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->kernel_version = g_strdup(ptr+15);
+ break;
+ case 8:
+ if(strncmp("machine=",ptr,8)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->machine = g_strdup(ptr+8);
+ break;
+ case 9:
+ if(strncmp("processor=",ptr,10)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->processor = g_strdup(ptr+10);
+ break;
+ case 10:
+ if(strncmp("hardware_platform=",ptr,18)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->hardware_platform = g_strdup(ptr+18);
+ break;
+ case 11:
+ if(strncmp("operating_system=",ptr,17)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ des->operating_system = g_strdup(ptr+17);
+ break;
+ case 12:
+ if(strncmp("ltt_major_version=",ptr,18)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ ptr += 18;
+ ptr++;//skip begining "
+ ptr[strlen(ptr)-1] = '\0'; //get rid of the ending "
+ des->ltt_major_version = (unsigned)atoi(ptr);
+ break;
+ case 13:
+ if(strncmp("ltt_minor_version=",ptr,18)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ ptr += 18;
+ ptr++;//skip begining "
+ ptr[strlen(ptr)-1] = '\0'; //get rid of the ending "
+ des->ltt_minor_version = (unsigned)atoi(ptr);
+ break;
+ case 14:
+ if(strncmp("ltt_block_size=",ptr,15)!=0)
+ g_error("Not a valid file: %s\n", pathname);
+ ptr += 15;
+ ptr++;//skip begining "
+ ptr[strlen(ptr)-1] = '\0'; //get rid of the ending "
+ des->ltt_block_size = (unsigned)atoi(ptr);
+ break;
+ default:
+ g_error("Not a valid file: %s\n", pathname);
+ }
}
+ //get description
+ description[0] = '\0';
+ if(fgets(buf,DIR_NAME_SIZE, fp)== NULL)
+ g_error("Not a valid file: %s\n", pathname);
+ ptr = buf;
+ while(isspace(*ptr)) ptr++;
+ if(*ptr != '>') g_error("Not a valid file: %s\n", pathname);
+ while((ptr=fgets(buf,DIR_NAME_SIZE, fp))!= NULL){
+ ptr = buf;
+ while(isspace(*ptr)) ptr++;
+ if(strncmp("</system>",ptr,9) == 0 )break;
+ strcat(description, buf);
+ }
+ if(!ptr)g_error("Not a valid file: %s\n", pathname);
+ if(description[0] = '\0')des->description = NULL;
+ des->description = g_strdup(description);
- return 0;
+ fclose(fp);
}
/*****************************************************************************
- * The following functions used to query the info of the machine where the
- * tracefile was generated. A tracefile may be queried for its architecture
- * type(e.g.,"i386", "powerpc", "powerpcle", "s390", "s390x"), its architecture
- * variant(e.g., "att" versus "sun" for m68k), its operating system (e.g.,
- * "linux","bsd"), its generic architecture, and the machine identity (e.g.,
- * system host name). All character strings belong to the associated tracefile
- * and are freed when it is closed
+ *The following functions get facility/tracefile information
****************************************************************************/
-uint32_t ltt_tracefile_arch_type(ltt_tracefile *t)
+void getFacilityInfo(LttTrace *t, char* eventdefs)
{
- return t->trace_header->arch_type;
-}
+ DIR * dir;
+ struct dirent *entry;
+ char * ptr;
+ int i,j;
+ LttFacility * f;
+ LttEventType * et;
+ LttTracefile * tracefile;
+
+ dir = opendir(eventdefs);
+ if(!dir) g_error("Can not open directory: %s\n", eventdefs);
+
+ while((entry = readdir(dir)) != NULL){
+ ptr = &entry->d_name[strlen(entry->d_name)-4];
+ if(strcmp(ptr,".xml") != 0) continue;
+ ltt_facility_open(t,entry->d_name);
+ }
+ closedir(dir);
+
+ tracefile = (LttTracefile*)g_ptr_array_index(t->per_cpu_tracefiles,0);
+ for(j=0;j<t->facility_number;j++){
+ f = (LttFacility*)g_ptr_array_index(t->facilities, i);
+ for(i=0; i<f->event_number; i++){
+ et = f->events[i];
+ setFieldsOffset(tracefile, et, NULL);
+ }
+ }
+}
-uint32_t ltt_tracefile_arch_variant(ltt_tracefile *t)
+void getControlFileInfo(LttTrace *t, char* control)
{
- return t->trace_header->arch_variant;
+ DIR * dir;
+ struct dirent *entry;
+
+ dir = opendir(control);
+ if(!dir) g_error("Can not open directory: %s\n", control);
+
+ while((entry = readdir(dir)) != NULL){
+ if(strcmp(entry->d_name,"facilities") != 0 ||
+ strcmp(entry->d_name,"interrupts") != 0 ||
+ strcmp(entry->d_name,"processes") != 0) continue;
+
+ ltt_tracefile_open_control(t,entry->d_name);
+ }
+ closedir(dir);
}
-ltt_arch_size ltt_tracefile_arch_size(ltt_tracefile *t)
+void getCpuFileInfo(LttTrace *t, char* cpu)
{
- return t->trace_header->arch_size;
+ DIR * dir;
+ struct dirent *entry;
+
+ dir = opendir(cpu);
+ if(!dir) g_error("Can not open directory: %s\n", cpu);
+
+ while((entry = readdir(dir)) != NULL){
+ if(strcmp(entry->d_name,".") != 0 ||
+ strcmp(entry->d_name,"..") != 0 ){
+ ltt_tracefile_open_cpu(t,entry->d_name);
+ }else continue;
+ }
+ closedir(dir);
}
-ltt_arch_endian ltt_tracefile_arch_endian(ltt_tracefile *t)
-{
- return t->trace_header->arch_endian;
-}
+/*****************************************************************************
+ *A trace is specified as a pathname to the directory containing all the
+ *associated data (control tracefiles, per cpu tracefiles, event
+ *descriptions...).
+ *
+ *When a trace is closed, all the associated facilities, types and fields
+ *are released as well.
+ ****************************************************************************/
-uint32_t ltt_tracefile_system_type(ltt_tracefile *t)
+LttTrace *ltt_trace_open(char *pathname)
{
- return t->trace_header->system_type;
+ LttTrace * t;
+ LttSystemDescription * sys_description;
+ char eventdefs[DIR_NAME_SIZE];
+ char info[DIR_NAME_SIZE];
+ char control[DIR_NAME_SIZE];
+ char cpu[DIR_NAME_SIZE];
+ char tmp[DIR_NAME_SIZE];
+ gboolean has_slash = FALSE;
+
+ //establish the pathname to different directories
+ if(pathname[strlen(pathname)-1] == '/')has_slash = TRUE;
+ strcpy(eventdefs,pathname);
+ if(!has_slash)strcat(eventdefs,"/");
+ strcat(eventdefs,"eventdefs/");
+
+ strcpy(info,pathname);
+ if(!has_slash)strcat(info,"/");
+ strcat(info,"info/");
+
+ strcpy(control,pathname);
+ if(!has_slash)strcat(control,"/");
+ strcat(control,"control/");
+
+ strcpy(cpu,pathname);
+ if(!has_slash)strcat(cpu,"/");
+ strcat(cpu,"cpu/");
+
+ //new trace
+ t = g_new(LttTrace, 1);
+ sys_description = g_new(LttSystemDescription, 1);
+ t->pathname = g_strdup(pathname);
+ t->facility_number = 0;
+ t->control_tracefile_number = 0;
+ t->per_cpu_tracefile_number = 0;
+ t->system_description = sys_description;
+ t->control_tracefiles = g_ptr_array_new();
+ t->per_cpu_tracefiles = g_ptr_array_new();
+ t->facilities = g_ptr_array_new();
+ getDataEndianType(&(t->my_arch_size), &(t->my_arch_endian));
+
+ //get system description
+ strcpy(tmp,info);
+ strcat(tmp,"system.xml");
+ getSystemInfo(sys_description, tmp);
+
+ //get control tracefile info
+ getControlFileInfo(t,control);
+
+ //get cpu tracefile info
+ getCpuFileInfo(t,cpu);
+
+ //get facilities info
+ getFacilityInfo(t,eventdefs);
+
+ return t;
}
-char *ltt_tracefile_system_name(ltt_tracefile *t)
+void ltt_trace_close(LttTrace *t)
{
- return t->trace_header->system_name;
+ int i;
+ LttTracefile * tf;
+ LttFacility * f;
+
+ g_free(t->pathname);
+
+ //free system_description
+ g_free(t->system_description->description);
+ g_free(t->system_description->node_name);
+ g_free(t->system_description->domain_name);
+ g_free(t->system_description->kernel_name);
+ g_free(t->system_description->kernel_release);
+ g_free(t->system_description->kernel_version);
+ g_free(t->system_description->machine);
+ g_free(t->system_description->processor);
+ g_free(t->system_description->hardware_platform);
+ g_free(t->system_description->operating_system);
+ g_free(t->system_description);
+
+ //free control_tracefiles
+ for(i=0;i<t->control_tracefile_number;i++){
+ tf = (LttTracefile*)g_ptr_array_index(t->control_tracefiles,i);
+ ltt_tracefile_close(tf);
+ }
+ g_ptr_array_free(t->control_tracefiles, FALSE);
+
+ //free per_cpu_tracefiles
+ for(i=0;i<t->per_cpu_tracefile_number;i++){
+ tf = (LttTracefile*)g_ptr_array_index(t->per_cpu_tracefiles,i);
+ ltt_tracefile_close(tf);
+ }
+ g_ptr_array_free(t->per_cpu_tracefiles, FALSE);
+
+ //free facilities
+ for(i=0;i<t->facility_number;i++){
+ f = (LttFacility*)g_ptr_array_index(t->facilities,i);
+ ltt_facility_close(f);
+ }
+ g_ptr_array_free(t->facilities, FALSE);
+
+ g_free(t);
}
+
/*****************************************************************************
- *Function name
- * ltt_tracefile_cpu_number: get the number of the cpu
- *Input params
- * t : tracefile
- *Return value
- * unsigned : the number of cpu the system has
+ *Get the system description of the trace
****************************************************************************/
-unsigned ltt_tracefile_cpu_number(ltt_tracefile *t)
+LttSystemDescription *ltt_trace_system_description(LttTrace *t)
{
- return (unsigned)(t->trace_header->cpu_number);
+ return t->system_description;
}
/*****************************************************************************
- *Function name
- * ltt_tracefile_cpu_single: does the tracefile contain events only for a
- * single CPU ?
- *Input params
- * t : tracefile
- *Return value
- * int : 1 for YES, 0 for NO
+ * The following functions discover the facilities of the trace
****************************************************************************/
-int ltt_tracefile_cpu_single(ltt_tracefile *t)
+unsigned ltt_trace_facility_number(LttTrace *t)
{
- if(t->trace_header->cpu_number_used == 1) return 1;
- else return 0;
+ return (unsigned)(t->facility_number);
+}
+
+LttFacility *ltt_trace_facility_get(LttTrace *t, unsigned i)
+{
+ return (LttFacility*)g_ptr_array_index(t->facilities, i);
}
/*****************************************************************************
*Function name
- * ltt_tracefile_cpu_id: which CPU is contained in the tracefile?
+ * ltt_trace_facility_find : find facilities in the trace
*Input params
- * t : tracefile
+ * t : the trace
+ * name : facility name
+ *Output params
+ * position : position of the facility in the trace
*Return value
- * unsigned : cpu id
+ * : the number of facilities
****************************************************************************/
-unsigned ltt_tracefile_cpu_id(ltt_tracefile *t)
+unsigned ltt_trace_facility_find(LttTrace *t, char *name, unsigned *position)
{
- return (unsigned)(t->trace_header->cpu_id);
+ int i, count=0;
+ LttFacility * f;
+ for(i=0;i=t->facility_number;i++){
+ f = (LttFacility*)g_ptr_array_index(t->facilities, i);
+ if(strcmp(f->name,name)==0){
+ count++;
+ if(count==1) *position = i;
+ }else{
+ if(count) break;
+ }
+ }
+ return count;
}
/*****************************************************************************
- * The following functions get the times related to the tracefile
+ * Functions to discover all the event types in the trace
****************************************************************************/
-ltt_time ltt_tracefile_time_start(ltt_tracefile *t)
+unsigned ltt_trace_eventtype_number(LttTrace *t)
{
- return t->start_time;
+ int i;
+ unsigned count = 0;
+ LttFacility * f;
+ for(i=0;i=t->facility_number;i++){
+ f = (LttFacility*)g_ptr_array_index(t->facilities, i);
+ count += f->event_number;
+ }
+ return count;
}
-ltt_time ltt_tracefile_time_end(ltt_tracefile *t)
+LttFacility * ltt_trace_facility_by_id(LttTrace * trace, unsigned id)
{
- return t->end_time;
+ LttFacility * facility;
+ int i;
+ for(i=0;i<trace->facility_number;i++){
+ facility = (LttFacility*) g_ptr_array_index(trace->facilities,i);
+ if(id >= facility->base_id &&
+ id < facility->base_id + facility->event_number)
+ break;
+ }
+ if(i==trace->facility_number) return NULL;
+ else return facility;
}
-ltt_time ltt_tracefile_duration(ltt_tracefile *t)
+LttEventType *ltt_trace_eventtype_get(LttTrace *t, unsigned evId)
{
- ltt_time T;
- TimeSub(T, t->end_time, t->start_time);
- return T;
+ LttFacility * f;
+ f = ltt_trace_facility_by_id(t,evId);
+ if(!f) return NULL;
+ return f->events[evId - f->base_id];
}
/*****************************************************************************
- * The following functions discover the facilities added to the tracefile
+ *There is one "per cpu" tracefile for each CPU, numbered from 0 to
+ *the maximum number of CPU in the system. When the number of CPU installed
+ *is less than the maximum, some positions are unused. There are also a
+ *number of "control" tracefiles (facilities, interrupts...).
****************************************************************************/
-
-unsigned ltt_tracefile_facility_number(ltt_tracefile *t)
+unsigned ltt_trace_control_tracefile_number(LttTrace *t)
{
- return (unsigned)(t->facility_number);
+ return t->control_tracefile_number;
}
-ltt_facility *ltt_tracefile_facility_get(ltt_tracefile *t, unsigned i)
+unsigned ltt_trace_per_cpu_tracefile_number(LttTrace *t)
{
- gconstpointer ptr = (gconstpointer)(&i);
- if(i<=0 || i> t->facility_number)return NULL;
- return (ltt_facility*)g_hash_table_lookup(t->index_facility,ptr);
-}
-
-ltt_facility *ltt_tracefile_facility_get_by_name(ltt_tracefile *t,char *name)
-{
- ltt_facility * fac;
- fac=(ltt_facility*)g_hash_table_lookup(t->facility_name,(gconstpointer)name);
- return fac;
+ return t->per_cpu_tracefile_number;
}
/*****************************************************************************
- * The following functions to discover all the event types in the facilities
- * added to the tracefile. The event type integer id, unique for the trace,
- * is used
+ *It is possible to search for the tracefiles by name or by CPU position.
+ *The index within the tracefiles of the same type is returned if found
+ *and a negative value otherwise.
****************************************************************************/
-unsigned ltt_tracefile_eventtype_number(ltt_tracefile *t)
+int ltt_trace_control_tracefile_find(LttTrace *t, char *name)
{
- return t->eventtype_number;
+ LttTracefile * tracefile;
+ int i;
+ for(i=0;i<t->control_tracefile_number;i++){
+ tracefile = (LttTracefile*)g_ptr_array_index(t->control_tracefiles, i);
+ if(strcmp(tracefile->name, name)==0)break;
+ }
+ if(i == t->control_tracefile_number) return -1;
+ return i;
}
-ltt_eventtype *ltt_tracefile_eventtype_get(ltt_tracefile *t, unsigned evId)
+int ltt_trace_per_cpu_tracefile_find(LttTrace *t, unsigned i)
{
- ptr_wrap * ptr;
- ptr = (ptr_wrap *)g_ptr_array_index(t->eventtype_event_id, (gint)evId);
- return (ltt_eventtype *)(ptr->ptr);
+ LttTracefile * tracefile;
+ int j, name;
+ for(j=0;j<t->per_cpu_tracefile_number;j++){
+ tracefile = (LttTracefile*)g_ptr_array_index(t->per_cpu_tracefiles, j);
+ name = atoi(tracefile->name);
+ if(name == (int)i)break;
+ }
+ if(j == t->per_cpu_tracefile_number) return -1;
+ return j;
}
/*****************************************************************************
- *Function name
- * ltt_tracefile_eventtype_id
- * : given an event type, find its unique id within
- * the tracefile
- *Input params
- * t : tracefile
- * et : event type
- *Return value
- * unsigned : id of the event type in the tracefile
+ *Get a specific tracefile
****************************************************************************/
-unsigned ltt_tracefile_eventtype_id(ltt_tracefile *t, ltt_eventtype *et)
+LttTracefile *ltt_trace_control_tracefile_get(LttTrace *t, unsigned i)
{
- int *id;
- char *name = et->facility->name;
-
- id = (int*)g_hash_table_lookup(t->base_id_name, (gconstpointer)name);
- if(!id)return 0;
- return (unsigned)(*id + et->index);
+ return (LttTracefile*)g_ptr_array_index(t->per_cpu_tracefiles, i);
+}
+
+LttTracefile *ltt_trace_per_cpu_tracefile_get(LttTrace *t, unsigned i)
+{
+ return (LttTracefile*)g_ptr_array_index(t->per_cpu_tracefiles, i);
}
/*****************************************************************************
- *Function name
- * ltt_tracefile_eventtype_root_field
- * : get the root field associated with an event type
- * for the tracefile
- *Input params
- * t : tracefile
- * id : event id
- *Return value
- * ltt_field * : root field of the event
+ *Get the name of a tracefile
****************************************************************************/
-ltt_field *ltt_tracefile_eventtype_root_field(ltt_tracefile *t, unsigned id)
+char *ltt_tracefile_name(LttTracefile *tf)
{
- ltt_eventtype * et;
- et = ltt_tracefile_eventtype_get(t, id);
- if(!et) return NULL;
- return et->root_field;
+ return tf->name;
}
/*****************************************************************************
*Input params
* t : tracefile
* time : criteria of the time
- *Return value
- * int : error code
- * ENOENT, end of the file
- * EINVAL, lseek fail
- * EIO, can not read from the file
****************************************************************************/
-int ltt_tracefile_seek_time(ltt_tracefile *t, ltt_time time)
+void ltt_tracefile_seek_time(LttTracefile *t, LttTime time)
{
int err;
- ltt_time lttTime;
- int headTime = timecmp(&(t->a_block_header->time), &time);
- int tailTime = timecmp(&(t->a_block_footer->time), &time);
+ LttTime lttTime;
+ int headTime = timecmp(&(t->a_block_start->time), &time);
+ int tailTime = timecmp(&(t->a_block_end->time), &time);
if(headTime < 0 && tailTime > 0){
lttTime = getEventTime(t);
err = timecmp(<tTime, &time);
- if(err >= 0){
- if( ( (t->which_block != 1 && t->which_event != 0) ||
- (t->which_block == 1 && t->which_event != 1) ) &&
- ((t->prev_event_time.tv_sec==0 && t->prev_event_time.tv_nsec==0) ||
- timecmp(&t->prev_event_time, &time) >= 0 ) ){
+ if(err > 0){
+ if(t->which_event==1 || timecmp(&t->prev_event_time,&time)<0){
+ return;
+ }else{
updateTracefile(t);
return ltt_tracefile_seek_time(t, time);
}
}else if(err < 0){
err = t->which_block;
- if(ltt_tracefile_read(t) == NULL) return ENOENT;
+ if(ltt_tracefile_read(t) == NULL){
+ g_printf("End of file\n");
+ return;
+ }
if(t->which_block == err)
return ltt_tracefile_seek_time(t,time);
- }
+ }else return;
}else if(headTime > 0){
if(t->which_block == 1){
updateTracefile(t);
t->prev_block_end_time.tv_nsec == 0 ) ||
timecmp(&(t->prev_block_end_time),&time) > 0 ){
err=readBlock(t,t->which_block-1);
- if(err) return err;
+ if(err) g_error("Can not read tracefile: %s\n", t->name);
return ltt_tracefile_seek_time(t, time) ;
}else{
updateTracefile(t);
}else if(tailTime <= 0){
if(t->which_block != t->block_number){
err=readBlock(t,t->which_block+1);
- if(err) return err;
- }else return ENOENT;
+ if(err) g_error("Can not read tracefile: %s\n", t->name);
+ }else {
+ g_printf("End of file\n");
+ return;
+ }
if(tailTime < 0) return ltt_tracefile_seek_time(t, time);
}else if(headTime == 0){
updateTracefile(t);
}
- return 0;
}
/*****************************************************************************
*Input params
* t : tracefile
*Return value
- * ltt_event * : an event to be processed
+ * LttEvent * : an event to be processed
****************************************************************************/
-ltt_event *ltt_tracefile_read(ltt_tracefile *t)
+LttEvent *ltt_tracefile_read(LttTracefile *t)
{
- ltt_event * lttEvent = (ltt_event *)g_new(ltt_event, 1);
- ltt_eventtype * evT;
- ltt_facility * fac;
+ LttEvent * lttEvent = (LttEvent *)g_new(LttEvent, 1);
+ int err;
//update the fields of the current event and go to the next event
- if(skipEvent(t)) return NULL;
+ err = skipEvent(t);
+ if(err == ENOENT) return NULL;
+ if(err == ERANGE) g_error("event id is out of range\n");
+ if(err)g_error("Can not read tracefile\n");
+
+ lttEvent->event_id = (int)(*(uint16_t *)(t->cur_event_pos));
+ if(lttEvent->event_id == TRACE_TIME_HEARTBEAT)
+ t->cur_heart_beat_number++;
t->current_event_time = getEventTime(t);
- lttEvent->event_id = (int)(*(uint8_t *)(t->cur_event_pos));
- evT = ltt_tracefile_eventtype_get(t, (unsigned)lttEvent->event_id);
- fac = evT->facility;
- if(evT->index == TRACE_EV_HEARTBEAT && strcmp(fac->name, "default")==0)
- t->cur_heart_beat_number++;
- lttEvent->cycle_count=*(uint32_t*)(t->cur_event_pos + EVENT_ID_SIZE);
+ lttEvent->time_delta = *(uint32_t*)(t->cur_event_pos + EVENT_ID_SIZE);
+ lttEvent->event_time = t->current_event_time;
+
+ lttEvent->event_cycle_count = ((uint64_t)1)<<32 * t->cur_heart_beat_number
+ + lttEvent->time_delta;
+
lttEvent->tracefile = t;
lttEvent->data = t->cur_event_pos + EVENT_HEADER_SIZE;
* EIO : can not read from the file
****************************************************************************/
-int readBlock(ltt_tracefile * tf, int whichBlock)
+int readBlock(LttTracefile * tf, int whichBlock)
{
off_t nbBytes;
uint32_t lostSize;
if(whichBlock - tf->which_block == 1 && tf->which_block != 0){
- tf->prev_block_end_time = tf->a_block_footer->time;
+ tf->prev_block_end_time = tf->a_block_end->time;
}else{
tf->prev_block_end_time.tv_sec = 0;
tf->prev_block_end_time.tv_nsec = 0;
tf->prev_event_time.tv_sec = 0;
tf->prev_event_time.tv_nsec = 0;
- nbBytes=lseek(tf->fd,(off_t)((whichBlock-1)*tf->trace_header->buffer_size),
- SEEK_SET);
+ nbBytes=lseek(tf->fd,(off_t)((whichBlock-1)*tf->block_size), SEEK_SET);
if(nbBytes == -1) return EINVAL;
- if(readFile(tf->fd,tf->buffer,tf->trace_header->buffer_size,
- "Unable to read a block")) return EIO;
+ if(readFile(tf->fd,tf->buffer,tf->block_size,"Unable to read a block"))
+ return EIO;
+ tf->a_block_start=(BlockStart *) (tf->buffer + EVENT_HEADER_SIZE);
+ lostSize = *(uint32_t*)(tf->buffer + tf->block_size - sizeof(uint32_t));
+ tf->a_block_end=(BlockEnd *)(tf->buffer + tf->block_size -
+ lostSize + EVENT_HEADER_SIZE);
- tf->a_block_header=(block_header *) tf->buffer;
- lostSize = *(uint32_t*)(tf->buffer + tf->trace_header->buffer_size
- - sizeof(uint32_t));
- /* skip event ID and time delta to get the address of the block foot */
- tf->a_block_footer=(block_footer *)(tf->buffer+tf->trace_header->buffer_size
- -lostSize+sizeof(uint8_t)+sizeof(uint32_t));
tf->which_block = whichBlock;
- tf->which_event = 0;
- tf->first_event_pos = tf->buffer + sizeof(block_header);
- if(tf->which_block == 1){
- tf->which_event++;
- tf->first_event_pos += sizeof(trace_header_event);
- }
- tf->cur_event_pos = tf->first_event_pos;
- tf->current_event_time = tf->a_block_header->time;
+ tf->which_event = 1;
+ tf->cur_event_pos = tf->a_block_start + sizeof(BlockStart); //first event
tf->cur_heart_beat_number = 0;
+
+ tf->current_event_time = getEventTime(tf);
getCyclePerNsec(tf);
* tf : tracefile
****************************************************************************/
-void updateTracefile(ltt_tracefile * tf)
+void updateTracefile(LttTracefile * tf)
{
- if(tf->which_block == 1)tf->which_event = 1;
- else tf->which_event = 0;
- tf->cur_event_pos = tf->first_event_pos;
- tf->current_event_time = tf->a_block_header->time;
+ tf->which_event = 1;
+ tf->cur_event_pos = tf->a_block_start + sizeof(BlockStart);
+ tf->current_event_time = getEventTime(tf);
tf->cur_heart_beat_number = 0;
tf->prev_event_time.tv_sec = 0;
* ERANGE : event id is out of range
****************************************************************************/
-int skipEvent(ltt_tracefile * t)
+int skipEvent(LttTracefile * t)
{
int evId, err;
void * evData;
- ltt_eventtype * evT;
- ltt_field * rootFld;
+ LttEventType * evT;
+ LttField * rootFld;
- evId = (int)(*(uint8_t *)(t->cur_event_pos));
+ evId = (int)(*(uint16_t *)(t->cur_event_pos));
evData = t->cur_event_pos + EVENT_HEADER_SIZE;
- evT = ltt_tracefile_eventtype_get(t,(unsigned)evId);
+ evT = ltt_trace_eventtype_get(t->trace,(unsigned)evId);
if(evT) rootFld = evT->root_field;
else return ERANGE;
evT->latest_event = t->which_event;
//the next event is in the next block
- if(t->which_event == t->a_block_header->event_count - 1){
+ if(evId == TRACE_BLOCK_END){
if(t->which_block == t->block_number) return ENOENT;
err = readBlock(t, t->which_block + 1);
if(err) return err;
* t : tracefile
****************************************************************************/
-void getCyclePerNsec(ltt_tracefile * t)
+void getCyclePerNsec(LttTracefile * t)
{
- ltt_time lBufTotalTime; /* Total time for this buffer */
- ltt_cycle_count lBufTotalNSec; /* Total time for this buffer in nsecs */
- ltt_cycle_count lBufTotalCycle;/* Total cycles for this buffer */
+ LttTime lBufTotalTime; /* Total time for this buffer */
+ LttCycleCount lBufTotalNSec; /* Total time for this buffer in nsecs */
+ LttCycleCount lBufTotalCycle;/* Total cycles for this buffer */
/* Calculate the total time for this buffer */
- TimeSub(lBufTotalTime,t->a_block_footer->time, t->a_block_header->time);
+ TimeSub(lBufTotalTime,t->a_block_end->time, t->a_block_start->time);
/* Calculate the total cycles for this bufffer */
- lBufTotalCycle = t->a_block_footer->cycle_count
- - t->a_block_header->cycle_count;
+ lBufTotalCycle = t->a_block_end->cycle_count
+ - t->a_block_start->cycle_count;
/* Convert the total time to nsecs */
lBufTotalNSec = lBufTotalTime.tv_sec * 1000000000 + lBufTotalTime.tv_nsec;
*Input params
* tf : tracefile
*Return value
- * ltt_time : the time of the event
+ * LttTime : the time of the event
****************************************************************************/
-ltt_time getEventTime(ltt_tracefile * tf)
+LttTime getEventTime(LttTracefile * tf)
{
- ltt_time time;
- ltt_cycle_count cycle_count; /* cycle count for the current event */
- ltt_cycle_count lEventTotalCycle; /* Total cycles from start for event */
- double lEventNSec; /* Total usecs from start for event */
- ltt_time lTimeOffset; /* Time offset in struct ltt_time */
+ LttTime time;
+ LttCycleCount cycle_count; // cycle count for the current event
+ LttCycleCount lEventTotalCycle; // Total cycles from start for event
+ double lEventNSec; // Total usecs from start for event
+ LttTime lTimeOffset; // Time offset in struct LttTime
- /* Calculate total time in cycles from start of buffer for this event */
- cycle_count = *(uint32_t*)(tf->cur_event_pos + EVENT_ID_SIZE);
+ // Calculate total time in cycles from start of buffer for this event
+ cycle_count = (LttCycleCount)*(uint32_t*)(tf->cur_event_pos + EVENT_ID_SIZE);
if(tf->cur_heart_beat_number)
cycle_count += ((uint64_t)1)<<32 * tf->cur_heart_beat_number;
- lEventTotalCycle=cycle_count-(tf->a_block_header->cycle_count & 0xFFFFFFFF);
+ lEventTotalCycle = cycle_count - tf->a_block_start->cycle_count;
- /* Convert it to nsecs */
+ // Convert it to nsecs
lEventNSec = lEventTotalCycle / tf->cycle_per_nsec;
- /* Determine offset in struct ltt_time */
+ // Determine offset in struct LttTime
lTimeOffset.tv_nsec = (long)lEventNSec % 1000000000;
lTimeOffset.tv_sec = (long)lEventNSec / 1000000000;
- TimeAdd(time, tf->a_block_header->time, lTimeOffset);
+ TimeAdd(time, tf->a_block_start->time, lTimeOffset);
return time;
}
* evD : event data, it may be NULL
****************************************************************************/
-void setFieldsOffset(ltt_tracefile * t, ltt_eventtype * evT, void * evD)
+void setFieldsOffset(LttTracefile * t, LttEventType * evT, void * evD)
{
- ltt_field * rootFld = evT->root_field;
+ LttField * rootFld = evT->root_field;
// rootFld->base_address = evD;
rootFld->field_size = getFieldtypeSize(t, evT, 0,0,rootFld, evD);
* int : size of the field
****************************************************************************/
-int getFieldtypeSize(ltt_tracefile * t, ltt_eventtype * evT, int offsetRoot,
- int offsetParent, ltt_field * fld, void * evD)
+int getFieldtypeSize(LttTracefile * t, LttEventType * evT, int offsetRoot,
+ int offsetParent, LttField * fld, void * evD)
{
int size, size1, element_number, i, offset1, offset2;
- ltt_type * type = fld->field_type;
+ LttType * type = fld->field_type;
- if(evT->latest_block==t->which_block && evT->latest_event==t->which_event){
- return fld->field_size;
- }
+ if(!t){
+ if(evT->latest_block==t->which_block && evT->latest_event==t->which_event){
+ return fld->field_size;
+ }
+ }
if(fld->field_fixed == 1){
if(fld == evT->root_field) return fld->field_size;
if(type->type_class != LTT_STRUCT && type->type_class != LTT_ARRAY &&
type->type_class != LTT_SEQUENCE && type->type_class != LTT_STRING){
if(fld->field_fixed == -1){
- size = (int) ltt_type_size(t, type);
+ size = (int) ltt_type_size(t->trace, type);
fld->field_fixed = 1;
}else size = fld->field_size;
}else size = fld->field_size;
}else if(type->type_class == LTT_SEQUENCE){
- size1 = (int) ltt_type_size(t, type);
+ size1 = (int) ltt_type_size(t->trace, type);
if(fld->field_fixed == -1){
fld->field_fixed = 0;
size = getFieldtypeSize(t, evT, offsetRoot,0,fld->child[0], NULL);
* int : 0: t1 == t2; -1: t1 < t2; 1: t1 > t2
****************************************************************************/
-int timecmp(ltt_time * t1, ltt_time * t2)
+int timecmp(LttTime * t1, LttTime * t2)
{
- ltt_time T;
+ LttTime T;
TimeSub(T, *t1, *t2);
if(T.tv_sec == 0 && T.tv_nsec == 0) return 0;
else if(T.tv_sec > 0 || (T.tv_sec==0 && T.tv_nsec > 0)) return 1;
* endian : endian type, little or big
****************************************************************************/
-void getDataEndianType(ltt_arch_size * size, ltt_arch_endian * endian)
+void getDataEndianType(LttArchSize * size, LttArchEndian * endian)
{
int i = 1;
char c = (char) i;