New Defects reported by Coverity Scan for RTEMS

[Joel Sherrill]

On Tue, Aug 30, 2022 at 12:05 AM Sebastian Huber <
sebastian.huber at embedded-brains.de> wrote:

> On 30/08/2022 00:56, scan-admin at coverity.com wrote:
> > ** CID 1512552:  High impact quality  (Y2K38_SAFETY)
> > /cpukit/score/src/kern_tc.c: 1804 in _Timecounter_Windup()
> >
> >
> >
> ________________________________________________________________________________________________________
> > *** CID 1512552:  High impact quality  (Y2K38_SAFETY)
> > /cpukit/score/src/kern_tc.c: 1804 in _Timecounter_Windup()
> > 1798          /* Go live with the new struct timehands. */
> > 1799     #ifdef FFCLOCK
> > 1800          switch (sysclock_active) {
> > 1801          case SYSCLOCK_FBCK:
> > 1802     #endif
> > 1803                  time_second = th->th_microtime.tv_sec;
> >>>>      CID 1512552:  High impact quality  (Y2K38_SAFETY)
> >>>>      A "time_t" value is stored in an integer with too few bits to
> accommodate it.  The expression "th->th_offset.sec" is cast to "int32_t".
> > 1804                  time_uptime = th->th_offset.sec;
> > 1805     #ifdef FFCLOCK
> > 1806                  break;
> > 1807          case SYSCLOCK_FFWD:
> > 1808                  time_second = fftimehands->tick_time_lerp.sec;
> > 1809                  time_uptime = fftimehands->tick_time_lerp.sec -
> ffclock_boottime.sec;
> >
> > ** CID 1512551:    (Y2K38_SAFETY)
>
> This seems to be a new Coverity feature. The Newlib time_t definition is:
>
> #if defined(_USE_LONG_TIME_T) || __LONG_MAX__ > 0x7fffffffL
> #define _TIME_T_ long
> #else
> #define _TIME_T_ __int_least64_t
> #endif
> typedef _TIME_T_        __time_t;
>
> Does Coverity use the Newlib header files? The _USE_LONG_TIME_T should
> be undefined for RTEMS.
>

Yes it should. It works by doing something like this:

cov-build waf|make

And looking at the newlib headers, I agree everything looks like it should
be defined to _int_least64_t. And preprocessing a simple file that includes
<sys/time.h> shows that it is typed to that.

But.... something else is going on. time_uptime is defined to
_Timecounter_Time_uptime
which is an int32_t.

$ grep -r _Timecounter_Time_uptime cpukit/
cpukit/include/rtems/score/timecounter.h:extern volatile int32_t
_Timecounter_Time_uptime;
cpukit/include/machine/_kernel_time.h:#define   time_uptime
_Timecounter_Time_uptime
cpukit/include/machine/_timecounter.h:extern volatile int32_t
_Timecounter_Time_uptime;
cpukit/rtems/src/clockgetuptimeseconds.c:  return _Timecounter_Time_uptime
- 1;
cpukit/score/src/kern_tc.c: *  ::_Timecounter, ::_Timecounter_Time_second,
and ::_Timecounter_Time_uptime
cpukit/score/src/kern_tc.c:#define      time_uptime _Timecounter_Time_uptime
cpukit/score/src/kern_ntptime.c:#define time_uptime _Timecounter_Time_uptime

The declaration in kern_tc.c doesn't align with the prototype in
_timecounter.h
once the macro redefinition is taken into account.

I'm surprised that gcc isn't generating a warning of some type for the
mismatch.

--joel


> --
> embedded brains GmbH
> Herr Sebastian HUBER
> Dornierstr. 4
> 82178 Puchheim
> Germany
> email: sebastian.huber at embedded-brains.de
> phone: +49-89-18 94 741 - 16
> fax:   +49-89-18 94 741 - 08
>
> Registergericht: Amtsgericht München
> Registernummer: HRB 157899
> Vertretungsberechtigte Geschäftsführer: Peter Rasmussen, Thomas Dörfler
> Unsere Datenschutzerklärung finden Sie hier:
> https://embedded-brains.de/datenschutzerklaerung/
> _______________________________________________
> devel mailing list
> devel at rtems.org
> http://lists.rtems.org/mailman/listinfo/devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20220830/7ccbe2ce/attachment.htm>