Re: getProcState may return wrong value on consecutive calls

Hi,

Looking at the source code, it seems sigar.getProcState(pid) ends calling proc_stat_read in linux_sigar.c

At the beginning of this method there is this code:

    linux_proc_stat_t *pstat = &sigar->last_proc_stat;    int status;    time_t timenow = time(NULL);    /*      * short-lived cache read/parse of last /proc/pid/stat     * as this info is spread out across a few functions.     */    if (pstat->pid == pid) {        if ((timenow - pstat->mtime) < SIGAR_LAST_PROC_EXPIRE) {            return SIGAR_OK;        }    }

SIGAR_LAST_PROC_EXPIRE is defined in a header file and represents 2 seconds. So if the last call was for same pid and less than two seconds ago, it returns a cached result (last_proc_stat from sigar structure). But the cached result is updated only if the process is alive:

    pstat->pid = pid;    pstat->mtime = timenow;    status = SIGAR_PROC_FILE2STR(buffer, pid, PROC_PSTAT);    if (status != SIGAR_OK) {        return status;    }

To sum up, once a process has died, if you call getProcState twice on the same Sigar instance in a less than two seconds interval, you will get the last ProcState value known by Sigar (when the process was still alive).

Is that right?

Thanks and regards,

Thomas