Nabble - Linux Test Project

is there anyway to get lcov to generate a function summary report

2008-09-06T11:51:48Z

I'd like to get a web page from lcov that showed the list of unexecuted functions.

Scott Weitzenkamp
SQA and Release Manager
Server Access Virtualization Business Unit
Cisco Systems

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-coverage mailing list
Ltp-coverage@...
https://lists.sourceforge.net/lists/listinfo/ltp-coverage

From forum: ltp-coverage

Re: [TESTING] SMP-nice for group scheduling

2008-09-05T14:36:40Z

Subrata Modak wrote:

> It would be great if you can do this. I need this to be compiled fine
> with gcc alone.

The attached version will compile with:

gcc --std=c99 fairtest.c -o fairtest -lm

>>It would certainly be possible, but the whole config parsing routine
>>would need to be rewritten to allow for commandline-based config
>>information. I'm curious...what's the issue with reading config files
>>to control testcases?
>
>
> this we can manage with the config file. I will find out some way for
> this.

If this turns out to be very difficult, I could rewrite the parsing
code. I'd rather not do it if I don't have to though.

Chris

/*
Author: Nortel Networks
Chris Friesen cfriesen@...
Approval token "Nortel-02-July-2008-01"

Copyright 2008 Nortel Networks

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 (at your option) any later
version.

THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES,
REPRESENTATIONS, AND CONDITIONS INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. See the GNU General Public License for more details.

IN NO EVENT SHALL THE COPYRIGHT OWNER BE LIABLE FOR ANY DIRECT, INDIRECT,
INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE
OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

You should have received a copy of the GNU General Public License along with
this program; if not, contact the Free Software Foundation, Inc., 675 Mass Ave,
Cambridge, MA 02139, USA www.fsf.org/ for a copy.

*/

#define _GNU_SOURCE
#include <stdio.h>
#include <string.h>
#include <sys/types.h>
#include <unistd.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <stdlib.h>
#include <sched.h>
#include <ctype.h>
#include <sys/syscall.h>
#include <sys/mman.h>
#include <time.h>
#include <math.h>
#include <errno.h>
#include <signal.h>
#include <sys/time.h>
#include <sys/resource.h>

#define GROUP_NAMESIZE 16
#define MAX_CPUS 4
#define MAX_GROUPS 10

#define for_all_kids(i,info) for(int i=0;i<info->numkidgroups;i++)
#define kid(i,info) info->kidgroups[i]
#define for_all_hogs(i,info) for(int i=0;i<info->numhogs;i++)
#define hog(i,info) info->hogs[i]

struct acct_data_struct
{
pid_t pid;

//the following are obtained from /proc/<pid>/sched
float cputime;
float wait_count;
float wait_max;
float wait_sum;
float wait_avg;
float ctx;
};

struct groupinfo
{
struct groupinfo *parent;
struct groupinfo *kidgroups[MAX_GROUPS];
int numkidgroups;
int share;
float totalpercent; //expected cpu usage of this group and all child groups
float percent; //expected cpu usage of this group
float net; //net cpu usage of this group
float gross; //gross cpu usage of this group
float kidweight; //sum of shares of child groups with cpu hogs
char name; //group name

int numhogs;
int numkidhogs;
pid_t hogs[MAX_CPUS];
struct acct_data_struct hogdata[MAX_CPUS];
double totalactual; //sum of actual cpu usage of all hogs for this group (and child groups)
};

int delay=-1;
int duration=-1;
int do_yield=0;
int do_sleep=0;
int verbose=0;
int extraverbose=0;
int doaffine=0;
int setfifo=0;

int numcpus;
struct groupinfo rootgroup;

void printresults(struct groupinfo *info);

char groupname[GROUP_NAMESIZE];
char *buildname(struct groupinfo *info)
{
if (info == &rootgroup) {
strcpy(groupname, "root");
return groupname;
}

char *ptr = groupname+GROUP_NAMESIZE-1;
*ptr = 0;
ptr--;
*ptr = info->name;

while((info=info->parent) != &rootgroup) {
ptr--;
*ptr = '/';
ptr--;
*ptr = info->name;
}
return ptr;
}

//can't have group percent be more than numhogs*100
void validatepercents(struct groupinfo *info)
{
if (info->percent/100.0*numcpus > info->numhogs) {
printf("invalid testcase configuration\n");
printf("group \"%s\" cannot use up specified time with given number of hogs\n", buildname(info));
exit(-1);
}

for_all_kids(i,info)
validatepercents(kid(i,info));
}

//determine the number of "descendent" hogs
void calckidhogs(struct groupinfo *info)
{
for_all_kids(i,info)
calckidhogs(kid(i,info));

info->kidweight=0;
for_all_kids(i,info) {
info->numkidhogs += kid(i,info)->numhogs + kid(i,info)->numkidhogs;
if ((kid(i,info)->numhogs > 0) || (kid(i,info)->kidweight))
info->kidweight += kid(i,info)->share;
}
}

//calculate the percent of the system that this group should get
void calcpercents(struct groupinfo *info)
{
float percent;
struct groupinfo *tmp;

//don't do calculation for any group without hogs
if (info == &rootgroup || info->numhogs==0)
goto next;

//this group has cpu hogs. Walk up the tree to the root group,
//calculating our expected cpu allocation. Our share is divided by
//the sum of the weights of our sibling groups with hogs in their
//descendents, times the share of the parent group.
//This is then repeated at the next level up,
//all the way up to the root group.

percent = info->share;

tmp = info;
while(tmp->parent) {
percent *= tmp->parent->share / tmp->parent->kidweight;
tmp = tmp->parent;
}

info->net = percent;
next:
for_all_kids(i,info)
calcpercents(kid(i,info));
}

void calc_expected(struct groupinfo *info)
{
if (info->numhogs)
info->percent = info->net;

for_all_kids(i,info)
calc_expected(kid(i,info));
}

int dumpgroup(struct groupinfo *info)
{
if (!info->name)
return 0;

if (info == &rootgroup)
goto next;

printf("%-s, %5d, %3d, %f\n",
buildname(info), info->share, info->numhogs, info->percent);
next:
for_all_kids(i,info)
dumpgroup(kid(i,info));
return 0;
}

void dumpconfig()
{
printf("using settling delay of %d sec, runtime of %d sec\n",
delay, duration);
printf("group hierarchy (name, weight, hogs, expected usage):\n");
dumpgroup(&rootgroup);
}

struct groupinfo * findgroup(struct groupinfo *info, char *name)
{
if (*name == '*')
return &rootgroup;

//scan kidgroups, try to find group with matching name
for(int i=0;i<MAX_GROUPS;i++) {
struct groupinfo *tmp = kid(i,info);
if (!tmp) {
//haven't found it yet, so doesnt' exist...make it
struct groupinfo *newinfo = (struct groupinfo *)malloc(sizeof(struct groupinfo));
memset(newinfo,0, sizeof(*newinfo));
kid(i,info) = newinfo;
newinfo->name = *name;
newinfo->parent = info;
info->numkidgroups++;
return newinfo;
}

//we found a kid, check the name
if (tmp->name == *name) {
//two possibilities, duplicate name or else a child group
//for now, ignore dupe
return findgroup(tmp, name+2);
}
}
return 0;
}

void validateleafhogs(struct groupinfo *info)
{
if (info->numhogs && info->numkidgroups) {
printf("error in group %s, only leaf groups can have cpu hogs\n",
buildname(info));
exit(-1);
}

for_all_kids(i,info)
validateleafhogs(kid(i,info));
}

int parsegroup(char *line)
{
char *nextptr , *ptr = line;
struct groupinfo *info;
int num;

nextptr = strchr(ptr, ',');
*nextptr = 0;

info = findgroup(&rootgroup, ptr);
if (!info) {
printf("unable to parse group, exiting\n");
exit(-1);
}

//root group is already set up, so ignore it
if (info == &rootgroup)
return 0;

//parse share
ptr = nextptr+1;
nextptr = strchr(ptr, ',');
*nextptr = 0;
int rc = sscanf(ptr, "%d", &num);
if (rc != 1){
printf("unable to parse numhogs, exiting\n");
printf("next ptr: %s\n", ptr);
exit(-1);
}
info->share=num;

//parse numhogs
//it's either 0,1,or 'n'
ptr = nextptr+1;
if (*ptr == '0')
info->numhogs = 0;
else if (*ptr == '1')
info->numhogs = 1;
else if (*ptr == 'n')
info->numhogs = numcpus;
else {
printf("unable to parse numhogs, exiting\n");
printf("next ptr: %s\n", ptr);
exit(-1);
}

return 0;
}

int parseconfig(char *path)
{
FILE * fp;
char * line = NULL;
size_t len = 0;
ssize_t read;
fp = fopen(path, "r");
if (!fp) {
perror("fopen");
return -1;
}

rootgroup.name='*';
rootgroup.share=100;
rootgroup.numhogs=0;
rootgroup.parent=NULL;
rootgroup.net = 0;
rootgroup.gross = 100;

while ((read = getline(&line, &len, fp)) != -1) {
if (!len)
continue;
if (line[0] == '#')
continue;
if (isspace(line[0]))
continue;

if (delay==-1)
sscanf(line, "%d", &delay);
else if (duration==-1)
sscanf(line, "%d", &duration);
else
parsegroup(line);
}
if (line)
free(line);
fclose(fp);
fp=NULL;

validateleafhogs(&rootgroup);
calckidhogs(&rootgroup);
calcpercents(&rootgroup);
calc_expected(&rootgroup);
dumpconfig();

validatepercents(&rootgroup);
return 0;
}

int setshare(int share, char *path)
{
char filename[128];
sprintf(filename, "%s/cpu.shares", path);

int fd = open(filename, O_RDWR);
char buf[1000];
int len=0;
len += sprintf(buf+len, "%d", share);

int rc = write(fd, buf, strlen(buf)+1);
close(fd);

if (rc < 0)
printf("%s: errror writing: %s\n", path, buf);

return 0;

}

int setgroup(pid_t pid, char *path)
{
char members[128];
sprintf(members, "%s/tasks", path);
int fd = open(members, O_RDWR);
char buf[100];
sprintf(buf, "%d", pid);
int rc = write(fd, buf, strlen(buf));
if (rc == -1) {
perror("problem setting task group:");
}
close(fd);
return 0;

}

void makegroups(struct groupinfo *info)
{
if (info != &rootgroup) {
char *name = buildname(info);
mkdir(name, 0777);
setshare(info->share, name);
}

for_all_kids(i,info)
makegroups(kid(i,info));
}

int setupsched()
{
system("mount -t cgroup -o cpu cgroup /dev/cgroup/ &> /dev/null");
chdir("/dev/cgroup");
system("find /dev/cgroup -type d -mindepth 1 "
"| sort -r | xargs rmdir >& /dev/null");
system("mkdir -p sys;cd sys;echo 2 > cpu.shares;for i in `cat ../tasks`; do echo $i > tasks; done");

makegroups(&rootgroup);

return 0;
}

void forkkidgroups(struct groupinfo *info)
{
for_all_hogs(i,info) {
int rc = fork();
if (!rc) {
//printf("pid %d in group %s, cpu %d\n", getpid(), buildname(info), i);
//fflush(stdout);
if (doaffine) {
cpu_set_t cpumask;

CPU_ZERO(&cpumask);
CPU_SET(i, &cpumask);

int rc = syscall(__NR_sched_setaffinity, getpid(), sizeof(cpumask), &cpumask);
if (rc < 0) {
printf("unable to set affinity for pid %d, cpu %d, group %s: %m",
getpid(), i, buildname(info));
}
}
struct sched_param p;
if (setfifo) {
p.sched_priority = 1;
sched_setscheduler(getpid(), SCHED_FIFO, &p);
} else {
p.sched_priority = 0;
sched_setscheduler(getpid(), SCHED_OTHER, &p);
setpriority(PRIO_PROCESS, 0, -10);
}

//infinite loop in child process

/* put ourselves to sleep until woken by parent */
raise(SIGSTOP);

int count = 0;
while(1) {
if (count++ > 1000000){
count=0;
if (do_yield)
sched_yield();
if (do_sleep) {
struct timespec ts = {0,1};
nanosleep(&ts,0);
}
}

}
} else if (rc>0) {
/* need to put kid into non-rt policy so we can set its group */
struct sched_param p;
p.sched_priority = 0;
sched_setscheduler(rc, SCHED_OTHER, &p);

hog(i,info)=rc;
setgroup(rc, buildname(info));
} else {
printf("error, unable to fork child hog %d, group %s: %m\n", i, buildname(info));
exit(-1);
}
}

for_all_kids(i,info)
forkkidgroups(kid(i,info));
}

void killkidgroups(struct groupinfo *info, int sig)
{
for_all_hogs(i,info)
kill(hog(i,info), sig);
for_all_kids(i,info)
killkidgroups(kid(i,info), sig);
}

void resetstats(struct groupinfo *info)
{
for_all_hogs(i,info) {
char cmd[1024];
snprintf(cmd, sizeof(cmd), "echo 0 > /proc/%d/sched", hog(i,info));
system(cmd);
}

for_all_kids(i,info)
resetstats(kid(i,info));
}

void parseline(char *line, char *key, float *val)
{
float tmp;
int conv;

char *s = strstr(line, key);
if (!s)
return;
s = strstr(s, ":");
if (!s)
return;
s++;
conv = sscanf(s, " %f", &tmp);
if (conv==1)
*val = tmp;
else {
printf("error parsing line %s for key %s", line, key);
exit(-1);
}
}

/* exit if we can't read data for this pid */
void getpiddata(pid_t pid, struct acct_data_struct *data)
{
FILE * fp;
char * line = NULL;
size_t len = 0;
ssize_t read;
char name[1024];
float wait_max, wait_sum, wait_count, sum_exec_runtime,ctx;

if (extraverbose) {
snprintf(name, sizeof(name), "cat /proc/%d/sched", pid);
system(name);
}

snprintf(name, sizeof(name), "/proc/%d/sched", pid);

fp = fopen(name, "r");
if (!fp) {
printf("error opening sched stats file for pid %d: %m", pid);
exit(-1);
}
wait_max = wait_sum = wait_count = sum_exec_runtime = 0;

while ((read = getline(&line, &len, fp)) != -1) {
if (!len) {
printf("error reading stats file for pid %d\n", pid);
exit(-1);
}

parseline(line, "se.wait_max", &wait_max);
parseline(line, "se.wait_sum", &wait_sum);
parseline(line, "se.wait_count", &wait_count);
parseline(line, "se.sum_exec_runtime", &sum_exec_runtime);
parseline(line, "nr_switches", &ctx);

}

if (line)
free(line);
fclose(fp);

if (!wait_max || !wait_sum || !wait_count || !sum_exec_runtime || !ctx){
printf("error parsing stats for pid %d\n", pid);
exit(-1);
}

data->cputime = sum_exec_runtime;
data->wait_max = wait_max;
data->wait_sum = wait_sum;
data->wait_count = wait_count;
data->wait_avg = wait_sum/wait_count;
data->ctx = ctx;
}

int updatetime(struct groupinfo *info)
{
for_all_hogs(i,info) {
getpiddata(hog(i,info), &info->hogdata[i]);
info->totalactual+=info->hogdata[i].cputime;
}

for_all_kids(i,info) {
int rc = updatetime(kid(i,info));
if (rc == 1)
return 1;
info->totalactual+=kid(i,info)->totalactual;
}
return 0;
}

int checktime(struct groupinfo *info)
{
for_all_hogs(i,info) {
if (hog(i,info) != info->hogdata[i].pid) {
printf("error, no timing data for pid %d, results unreliable\n", hog(i,info));
return 1;
}
}

for_all_kids(i,info) {
int rc = checktime(kid(i,info));
if (rc == 1)
return 1;
}
return 0;
}

float allowed_variance(struct groupinfo *info, int cpu)
{
//allow 3% of specified weight as error
float allowed_error = info->percent * 0.03;

//with smallest tolerance of 0.25%
if (allowed_error < 0.25)
allowed_error = 0.25;

return allowed_error;
}

const char * meets_share_spec(struct groupinfo *info, int hog)
{
/* note....this is the old calculation, need to update it with the new */
float actual;
if (hog==-1)
actual = info->totalactual*100.0/rootgroup.totalactual;
else
actual = (float)info->hogdata[hog].cputime*100.0*numcpus/rootgroup.totalactual;
float expected = info->percent;
float allowed_error = allowed_variance(info, 0);
if ((fabs(actual - expected) < allowed_error) || (actual > expected))
return "y";
else
return "n";
}

float allowed_lat(struct groupinfo *info, int cpu)
{
float tmp = 1000 / (info->percent * 10.0);
if (tmp < 20)
tmp = 20;
return tmp;
}

const char * meets_lat_spec(struct groupinfo *info, int cpu)
{
float worst = info->hogdata[0].wait_max;
return (allowed_lat(info, cpu) > worst) ? "y" : "n";
}

void printresults_mult(struct groupinfo *info)
{
char tmpbuf[1024];
int pos;

if ((info==&rootgroup) || (info->numhogs == 0))
goto printkidgroups;

info->totalactual=0;
for_all_hogs(i,info)
info->totalactual+=info->hogdata[i].cputime;

//name
printf("%7s", buildname(info));

if (verbose) {
pos=0;
pos += sprintf(tmpbuf + pos, " %s", meets_share_spec(info, -1));
pos += sprintf(tmpbuf + pos, "(%s", meets_share_spec(info, 0));
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%s", meets_share_spec(info, i));
}
printf("%8s)", tmpbuf);
}

if (verbose) {
pos=0;
pos += sprintf(tmpbuf + pos, " %s", meets_lat_spec(info, 0));
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%s", meets_lat_spec(info, i));
}
printf("%12s", tmpbuf);
}

//actual percent
pos=0;
pos += sprintf(tmpbuf + pos, " %5.2f", (float)info->totalactual*100.0/rootgroup.totalactual);
if (verbose) {
pos += sprintf(tmpbuf + pos, "(%5.2f", (float)info->hogdata[0].cputime*100.0*numcpus/rootgroup.totalactual);
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%.2f", (float)info->hogdata[i].cputime*100.0*numcpus/rootgroup.totalactual);
}
printf("%21s)", tmpbuf);
} else
printf("%13s", tmpbuf);

//expected percent
pos=0;
pos += sprintf(tmpbuf + pos, " %5.2f", info->percent);
printf("%14s", tmpbuf);

if (verbose) {
//error
pos=0;
pos += sprintf(tmpbuf + pos, " %5.2f", (float)info->hogdata[0].cputime*100.0*numcpus/rootgroup.totalactual - info->percent);
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%.2f", (float)info->hogdata[i].cputime*100.0*numcpus/rootgroup.totalactual - info->percent);
}
printf("%14s", tmpbuf);

//allowed error?
pos=0;
pos += sprintf(tmpbuf + pos, " %5.2f", allowed_variance(info, 0));
printf("%10s", tmpbuf);

//context switches
pos=0;
pos += sprintf(tmpbuf + pos, " %.0f", info->hogdata[0].ctx);
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%.0f", info->hogdata[i].ctx);
}
printf("%15s", tmpbuf);
}

//avg latency
pos=0;
pos += sprintf(tmpbuf + pos, " %.0f", info->hogdata[0].wait_avg);
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%.0f",info->hogdata[i].wait_avg);
}
printf("%18s", tmpbuf);

//worst latency
pos=0;
pos += sprintf(tmpbuf + pos, " %.0f", info->hogdata[0].wait_max);
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%.0f", info->hogdata[i].wait_max);
}
printf("%18s", tmpbuf);

if (verbose) {
//allowed latency
pos=0;
pos += sprintf(tmpbuf + pos, " %.2f", allowed_lat(info, 0));
for (int i=1;i<info->numhogs;i++) {
pos += sprintf(tmpbuf + pos, "/%.2f", allowed_lat(info, i));
}
printf("%23s", tmpbuf);
}

printf("\n");

printkidgroups:
for_all_kids(i,info)
printresults(kid(i,info));
}

//results for single hog
void printresults_single(struct groupinfo *info)
{
float percent, expected;

if ((info==&rootgroup) || (info->numhogs == 0))
goto printkidgroups;

info->totalactual=0;
for_all_hogs(i,info) {
info->totalactual+=info->hogdata[i].cputime;
}

percent = info->totalactual * 100.0 / rootgroup.totalactual;
expected = info->percent;

printf("%7s", buildname(info));
if (verbose)
printf("%8s%12s ", meets_share_spec(info, -1), meets_lat_spec(info, 0));

//printf("%15.2f", time_ms);

printf("%13.2f%14.2f", percent, expected);
if (verbose)
printf("%14.2f%14.2f%14.0f",
percent - expected, allowed_variance(info, 0), info->hogdata[0].ctx);

printf("%18.0f%18.0f", info->hogdata[0].wait_avg, info->hogdata[0].wait_max);
if (verbose)
printf("%20.2f", allowed_lat(info,0));

printf("\n");

printkidgroups:
//could print time/scheduled breakdown per pid within group

for_all_kids(i,info)
printresults(kid(i,info));
}

void printresults(struct groupinfo *info)
{
if (info->numhogs > 1)
printresults_mult(info);
else
printresults_single(info);
}

void printout(struct groupinfo *info)
{
double expected=duration*numcpus*1000.0;
double err = fabs(rootgroup.totalactual - expected);

//complain if more than 2% discrepancy between actual and expected duration
if (err/expected > 0.02)
printf("Warning, actual cpu time different than expected. actual: %f, expected: %f\n",
rootgroup.totalactual, expected);

if (verbose) {
printf(" meets meets\n");
printf(" group share spec lat spec actual(%%) expected(%%) error(%%) allowed error(%%) ctx switches avg latency(ms) max_latency(ms) allowed_latency(ms)\n");
} else {
printf(" group actual(%%) expected(%%) avg latency(ms) max_latency(ms)\n");
}
printresults(info);
}

int runtest()
{
//configure the scheduler for the test
setupsched();

//create the cpu hogs for each group. They will sleep
//waiting to be sent SIGCONT.
forkkidgroups(&rootgroup);

//tell all hogs to continue
killkidgroups(&rootgroup, SIGCONT);

//let hogs run for initial delay time to get them settled in
sleep(delay);

//pause hogs for a sec and reset all the stats, then start them
killkidgroups(&rootgroup, SIGSTOP);
resetstats(&rootgroup);
killkidgroups(&rootgroup, SIGCONT);

//let hogs run for duration of test
sleep(duration);

//pause all hogs
killkidgroups(&rootgroup, SIGSTOP);

//get the runtime data
updatetime(&rootgroup);

//terminate all hogs
killkidgroups(&rootgroup, SIGTERM);

//dump the data
printout(&rootgroup);

//erase all groups
system("find /dev/cgroup -type d -mindepth 1 "
"| sort -r | xargs rmdir >& /dev/null");

return 0;
}

int main(int argc, char **argv)
{
while(1) {
int c = getopt(argc, argv, "syvVaf");
switch(c) {
case 'y':
printf("doing yield\n");
do_yield = 1;
break;
case 's':
printf("doing sleep\n");
do_sleep = 1;
break;
case 'v':
verbose = 1;
break;
case 'V':
verbose = 1;
extraverbose = 1;
break;
case 'a':
doaffine = 1;
break;
case 'f':
setfifo=1;
break;
case -1:
default:
goto done_options;
}
}
done_options:

signal(SIGCHLD, SIG_IGN);

// bump up the priority of the main task so we can kill the others
struct sched_param p;
p.sched_priority = 2;
sched_setscheduler(getpid(), SCHED_RR, &p);

numcpus = sysconf(_SC_NPROCESSORS_ONLN);

//parse the config file specified in the arg list
int rc = parseconfig(argv[optind]);
if (rc < 0) {
printf("unable to parse config file\n");
return -1;
}

runtest();
}

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-list mailing list
Ltp-list@...
https://lists.sourceforge.net/lists/listinfo/ltp-list

From forum: ltp-list

[PATCH] fix getcpu test on parisc

2008-09-05T13:46:46Z

This patch fixes the getcpu patch for parisc (and probably other platforms as well).

The check for "__GLIBC__ >= 2 && __GLIBC_MINOR__ >= 6" is wrong.
It breaks as soon as a glibc 3.0 would be published.
Replace it with __GLIBC_PREREQ(2,6).

Tested on parisc and i386.

Signed-off-by: Helge Deller <deller@...>

diff -u -p -r1.1 getcpu01.c
--- testcases/kernel/syscalls/getcpu/getcpu01.c 22 Aug 2008 21:09:59 -0000 1.1
+++ testcases/kernel/syscalls/getcpu/getcpu01.c 5 Sep 2008 20:37:25 -0000
@@ -61,8 +61,7 @@
#include <dirent.h>

#if defined(__i386__) || defined(__x86_64__)
- #if defined(__GLIBC__) && defined(__GLIBC_MINOR__) \
- && __GLIBC__ >= 2 && __GLIBC_MINOR__ >= 6
+ #if __GLIBC_PREREQ(2,6)
#if defined(__x86_64__)
#include <utmpx.h>
#endif
@@ -165,12 +164,8 @@ static inline int getcpu(unsigned *cpu_i
{
#if defined(__i386__)
return syscall(318, cpu_id,node_id,cache_struct);
- #elif defined(__x86_64__)
- #if defined(__GLIBC__) && defined(__GLIBC_MINOR__) \
- && __GLIBC__ >= 2 && __GLIBC_MINOR__ >= 6
- return *cpu_id = sched_getcpu();
- #endif
- return 0;
+ #elif __GLIBC_PREREQ(2,6)
+ return *cpu_id = sched_getcpu();
#endif
return 0;
}

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-list mailing list
Ltp-list@...
https://lists.sourceforge.net/lists/listinfo/ltp-list

From forum: ltp-list

Re: [PATCH] Fix utimensat build failure

2008-09-05T05:42:59Z

On Fri, Sep 5, 2008 at 8:16 AM, Subrata Modak
<subrata@...> wrote:

> On Thu, 2008-09-04 at 20:39 +0530, Subrata Modak wrote:
>> Hi Michael,
>>
>> I have seen that in some of my systems, utimensat fails to build because
>> of some definitions missing in system headers. Do, you think we can
>> apply the following patch:
>>
>> Signed-Off-By: Subrata Modak <subrata@...>
>
> I thought this to be prudent until all headers are properly updated.
> Hence i merged this.

Looking at the kernel headers, these two constants are defined in
headers common to all architectures, so I think this should be okay.

Cheers,

Michael

>>
>> ---
>> diff -uprN
>> ltp-full-20080831.orig/testcases/kernel/syscalls/utimensat/utimensat01.c
>> ltp-full-20080831/testcases/kernel/syscalls/utimensat/utimensat01.c
>> ---
>> ltp-full-20080831.orig/testcases/kernel/syscalls/utimensat/utimensat01.c
>> 2008-09-04 20:33:24.000000000 +0530
>> +++ ltp-full-20080831/testcases/kernel/syscalls/utimensat/utimensat01.c
>> 2008-09-04 20:36:39.000000000 +0530
>> @@ -50,6 +50,13 @@
>>
>> #ifdef __NR_utimensat
>>
>> +#ifndef AT_FDCWD
>> + #define AT_FDCWD -100
>> +#endif
>> +#ifndef AT_SYMLINK_NOFOLLOW
>> + #define AT_SYMLINK_NOFOLLOW 0x100
>> +#endif
>> +
>> #define EXIT_bad_usage 3
>> #define EXIT_failed_syscall 3
>>
>> Regards--
>> Subrata
>>
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Ltp-list mailing list
>> Ltp-list@...
>> https://lists.sourceforge.net/lists/listinfo/ltp-list
>
>

--
Michael Kerrisk
Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/
man-pages online: http://www.kernel.org/doc/man-pages/online_pages.html
Found a bug? http://www.kernel.org/doc/man-pages/reporting_bugs.html

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-list mailing list
Ltp-list@...
https://lists.sourceforge.net/lists/listinfo/ltp-list

From forum: ltp-list

[RFC] [OLS 2008 FALLOUT] Issue # 3 (LTP-Robust)

2008-09-05T04:33:53Z

Hi,

With a brief stop over in our fallout discussion, i thought of resuming
discussing the other issues i came up.

As we have proposed in our OLS 2008 paper:
http://ltp.sourceforge.net/documentation/technical_papers/ltp-ols-2008-paper.pdf,

about LTP-robust, which will be a sub-project under LTP. As we are aware
that we have huge limitation in improving code coverage in testing from
the user space alone, as , we do not have sufficient test cases to test
all the features that gets added in the kernel exponentially. The reason
for that is well known, as, we do not have enough volunteers to write
test cases and keep up testing the growing kernel.

Hence, we envisaged the concept of creation of LTP-Robust, which
basically will be a mechanism built in LTP to inject faults in the
kernel. Mechanism of LTP-Robust should be able to do the following:

1) Inject Random faults inside the kernel, when individual test cases
are being run by PAN,
2) PAN should be able to detect that test failures are because of
dynamic faults injected, rather than actual faults in the kernel,
3) Once it is able to decipher this information, it should be able to
restart that test once again with fault-less kernel.

With the above testing methods, we believe we can improve the code
coverage by also touching the error paths inside the kernel, which in
normal circumstance will not be touched by the test code. In order to
achieve the above mechanism, we need to change the PAN (ltp/pan/pan.c)
code.

I would like to know if any of you would be interested to work on this.
Before that let us discuss and sort out things which you think needs
attention before embarking.

Regards--
Subrata

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-list mailing list
Ltp-list@...
https://lists.sourceforge.net/lists/listinfo/ltp-list

From forum: ltp-list

Re: Network and Network stress tests setup!!!

2008-09-05T04:27:51Z

Hi Rishi,

Even I had the same doubt. As per your suggestions I have done the manual setup according to the HOWTO.txt file for LTP.
But still I am getting the same old results.

I have changed the .conf files as per the instructions. I restart the xinetd by "/etc/init.d/xinetd restart" and then do "chkconfig [service-name] on"
Is this method right? Or is there any other way to restart the xinetd?

Regards,
Suyog

On Fri, Sep 5, 2008 at 2:50 PM, Rishikesh K. Rajak <risrajak@...> wrote:

Yes i doubt so. So here the conclusion is, all the testcase related to r is failing/broken. so you have to manually configure rsh, rlogin, ftp etc whichever services you want to enable between client and server.

Please enable and configure all the network services before running LTP network test.

-Rishi
On Fri, 2008-09-05 at 14:40 +0530, Suyog Upadhye wrote:
Hi Rishi,

I am attaching the result log, which LTP creates in /tmp folder.
netpan.log and netstress.log are the files created after running networktests.sh and networkstress.sh respectively.
networkResults.console.log is basically the console output after running networktests.sh.

Kindly note that I have run networkstress.sh script with '-i' (icmp) and -t (tcp) option only.

As per my understanding, I have to export RHOST and PASSWD in networktests.sh which I am doing.
And in networkstress.sh I need to export RHOST, RHOST_HWADDR (mac address), HTTP_DOWNLOAD_DIR, FTP_DOWNLOAD_DIR, FTP_UPLOAD_DIR, FTP_UPLOAD_URLDIR, IPV4_NETWORK, LHOST_IPV4_HOST, RHOST_IPV4_HOST, IPV4_NETWORK_REVERSE. In this case I am only setting RHOST and RHOST_HWADDR. Rest all is not very clear to me.

Regards,
Suyog

On Fri, Sep 5, 2008 at 1:53 PM, Rishikesh K. Rajak <risrajak@...> wrote:

Hi,

Before running ltp-network test you have to configure many things e.g: rsh, ftp, dhcp server etc...

Can you show us your failure so that i can tell you what other configuration you require to run successfully on your machine ?

I hope you are running the latest ltp.

-Rishi

On Fri, 2008-09-05 at 12:28 +0530, Suyog Upadhye wrote:

Hi,

I am trying to run networktests.sh and networkstress.sh on a setup of two machines.
I have same version of LTP on both the machines. I have added RHOST/PASSWD entries in networktests.sh and in networkstress.sh.

I am seeing only 10 test cases out of 33 network test cases are passed and no test case is passing in network stress test cases.

I doubt this is mainly the issue with the setup.
Can you kindly send me a consolidated list of the exports need to be done or if any services need to be restarted?

I am doing the following.
1. Installed all the servers on build machine. (e.g. telnetd, ftp, rlogin, finger etc.)
2. /etc/init.d/xinetd restart
3. service nfs restart
(Both machines can ping to each other normally, but all icmp tests in networkstress.sh are failing.)

Kindly let me know if I am on right track. And also let me know any pointers or a list of software setup instructions.

Regards,
Suyog
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________ Ltp-list mailing list Ltp-list@... https://lists.sourceforge.net/lists/listinfo/ltp-list
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________ Ltp-list mailing list Ltp-list@... https://lists.sourceforge.net/lists/listinfo/ltp-list

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-list mailing list
Ltp-list@...
https://lists.sourceforge.net/lists/listinfo/ltp-list

From forum: ltp-list

Re: [TESTING] x86: implement prctl PR_GET_TSC and PR_SET_TSC in 2.6.26

2008-09-05T04:09:37Z

Hi Erik,

On Thu, 2008-09-04 at 18:44 +0530, Subrata Modak wrote:
> Hi Erik,
>
> On Mon, 2008-09-01 at 19:36 +0200, Erik Bosman wrote:
> > Hi Subrata,
> >
> > Sorry for my late reaction. I'm busy working on my Master thesis right
> > now.

I would disturb you for a brief period once again.

> >
> > On Tue, 26 Aug 2008, Subrata Modak wrote:
> >
> > > Hi Erik,
> > >
> > > Would you kindly let us know how you tested prctl PR_GET_TSC and
> > > PR_SET_TSC, which is added in 2.6.26. Did you write/develop an automated
> > > test case(s) of your own. Let us know if you would like the same to
> > > share with LTP under GPLv2.
> > >
> >
> > One of my patches added three tests to the Documentation/prctl, so you can
> > find them there from 2.6.26 on. They are not automated but it is not hard
> > to change that. I would recommend using disable-tsc-on-off-stress-test.c.
>
> Thank you very much. I would look into them surely, and see how i can
> integrate them inside LTP.

I have run those tests on i686, x86_64 & ppc64 running 2.6.26 kernel.
And here are my observations:

1) Built and ran perfectly on X86_64:

# uname -a
Linux 2.6.26-autokern1 #1 SMP PREEMPT Fri Sep 5 09:40:31 UTC 2008 x86_64
GNU/Linux

# ./disable-tsc-test
rdtsc() == 9766811077875
prctl(PR_GET_TSC, &tsc_val); tsc_val == PR_TSC_ENABLE
rdtsc() == 9766811184102
prctl(PR_SET_TSC, PR_TSC_ENABLE)
rdtsc() == 9766811200743
prctl(PR_SET_TSC, PR_TSC_SIGSEGV)
rdtsc() == [ SIG_SEGV ]
prctl(PR_GET_TSC, &tsc_val); tsc_val == PR_TSC_SIGSEGV
prctl(PR_SET_TSC, PR_TSC_ENABLE)
rdtsc() == 9766811246535
# echo $?
0

# ./disable-tsc-on-off-stress-test
[No further output means we're allright]
# echo $?
0

# ./disable-tsc-ctxt-sw-stress-test
[No further output means we're allright]
# echo $?
0

2) Built fine on i686, but disable-tsc-on-off-stress-test failed:

# uname -a
Linux 2.6.26-autokern1 #1 SMP Fri Sep 5 09:49:37 UTC 2008 i686 GNU/Linux

# ./disable-tsc-test
rdtsc() == 4098191459558
prctl(PR_GET_TSC, &tsc_val); tsc_val == PR_TSC_ENABLE
rdtsc() == 4098192032718
prctl(PR_SET_TSC, PR_TSC_ENABLE)
rdtsc() == 4098192209502
prctl(PR_SET_TSC, PR_TSC_SIGSEGV)
rdtsc() == [ SIG_SEGV ]
prctl(PR_GET_TSC, &tsc_val); tsc_val == PR_TSC_SIGSEGV
prctl(PR_SET_TSC, PR_TSC_ENABLE)
rdtsc() == 4098192880534
# echo $?
0

# ./disable-tsc-on-off-stress-test
Floating point exception
# echo $?
136

# ./disable-tsc-ctxt-sw-stress-test
[No further output means we're allright]
# echo $?
0

3) Build failure on ppc64 with a POWER6 cpu:

# uname -a
Linux 2.6.26-autokern1 #1 SMP Fri Sep 5 05:41:12 EDT 2008 ppc64 ppc64
ppc64 GNU/Linux

# make
cc disable-tsc-ctxt-sw-stress-test.c -o
disable-tsc-ctxt-sw-stress-test
disable-tsc-ctxt-sw-stress-test.c: In function ‘rdtsc’:
disable-tsc-ctxt-sw-stress-test.c:33: error: impossible constraint in
‘asm’
make: *** [disable-tsc-ctxt-sw-stress-test] Error 1

# gcc --version
gcc (GCC) 4.1.2 20071124 (Red Hat 4.1.2-42)
Copyright (C) 2006 Free Software Foundation, Inc.

Can you kindly help us fix the:
1) Floating point exception for disable-tsc-on-off-stress-test, running
2.6.26 on i686,
2) build failure issue(s) with ppc64.

I also see the following issue(s) when i run it on i386 machine running
kernel less than 2.6.26:

$ uname -a
Linux 2.6.18-92.el5 #1 SMP Tue Apr 29 13:16:12 EDT 2008 i686 i686 i386
GNU/Linux

$ ./disable-tsc-test
rdtsc() == 34367216958620
prctl(PR_GET_TSC, &tsc_val); prctl: Invalid argument
tsc_val == [not set]
rdtsc() == 34367217591252
prctl(PR_SET_TSC, PR_TSC_ENABLE)
prctl: Invalid argument
rdtsc() == 34367217645449
prctl(PR_SET_TSC, PR_TSC_SIGSEGV)
prctl: Invalid argument
rdtsc() == 34367217698450
$ echo $?
0

$ ./disable-tsc-ctxt-sw-stress-test
[No further output means we're allright]
prctl: Invalid argument
prctl: Invalid argument
..............
..............
prctl: Invalid argument
$ echo $?
0

$ ./disable-tsc-ctxt-sw-stress-test
[No further output means we're allright]
prctl: Invalid argument
prctl: Invalid argument
..............
..............
prctl: Invalid argument
$ echo $?
0

Should it not fail on any kernel(s) below 2.6.26 ? It shows pass for all
the tests. And also lots of lines like:

prctl: Invalid argument

is printed, which in ideal situation should not be printed at all. What
do you think ?

Regards--
Subrata

>
> >
> > Is it a problem if the test causes a very high load?
>
> No, there is no such problem. But we will run it separately rather than
> with default run. Normally, we run the stress tests separately from the
> normal functional tests.
>
> Regards--
> Subrata
>
> >
> > > We already have the prctl() test cases in LTP and probably with this
> > > addition, we need to add some more tests to the existing scenario. Would
> > > you review the existing ones too:
> > >
> > > http://ltp.cvs.sourceforge.net/ltp/ltp/testcases/kernel/syscalls/prctl/,
> >
> > Prctl is by definition a very diverse syscall, I am not an expert on
> > all prctl calls. I don't think I can help you with many of them (as I
> > don't have the hardware, ) and as I said, I'm rather busy finishing my
> > master thesis.
> >
> > Regards,
> > Erik
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Ltp-list mailing list
> Ltp-list@...
> https://lists.sourceforge.net/lists/listinfo/ltp-list

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Ltp-list mailing list
Ltp-list@...
https://lists.sourceforge.net/lists/listinfo/ltp-list

From forum: ltp-list