[ARGUS] tv_nsec potentially invalid value

[Ming Fu via Argus-info]

Hi Carter,

Great! I only looked at the header. Didn't notice the 5.0.2 branch is newer.

Thanks
Ming

-----Original Message-----
From: Carter Bullard <carter at qosient.com>
Sent: Monday, November 25, 2024 10:27 PM
To: Ming Fu <Ming.Fu at esentire.com>
Cc: argus-info at lists.andrew.cmu.edu
Subject: Re: [ARGUS] tv_nsec potentially invalid value

Yes !!!  We fixed this in 5.0.2 which was released last week ... should be there ...
Carter


> On Nov 25, 2024, at 9:56 PM, Ming Fu via Argus-info <argus-info at lists.andrew.cmu.edu> wrote:
>
> Hi,
>
> I noticed a few argus 5.0.0 unexpected stops. Very hard to reproduce, but I think the problem is here:
>
> https://github.com/openargus/argus/blob/main/argus/ArgusSource.c#L4856
>
>         tts->tv_sec  = tvp->tv_sec + 0;
>         tts->tv_nsec = (tvp->tv_usec * 1000) + 250000000;
>         if (tts->tv_nsec > 1000000000) {
>            tts->tv_sec++;
>            tts->tv_nsec -= 1000000000;
>         }
>         if ((retn = pthread_cond_timedwait(&stask->cond, &stask->lock, tts))) {
>         ....
>
> The error log is:
> argus-rt[2591]: ArgusSourceProcess: pthread_cond_timedwait() error EINVAL
>
> The compare at line 4856 should be
>         if (tts->tv_nsec >= 1000000000) {
>
> 1000000000 is an invalid nsec number.
>
> Regards,
> Ming