On Friday 11 January 2008 05:08:59 Sébastien Dugué wrote:
> On Thu, 10 Jan 2008 12:01:31 -0800 Darren Hart
<[EMAIL PROTECTED]> wrote:
> > On Thursday 10 January 2008 09:47:16 Darren Hart wrote:
> > > On Thursday 10 January 2008 03:12:35 Sebastien Dugue wrote:
> > > > Add a Missed period (calc took too long) check to
> > > > periodic_cpu_load just as is done in periodic_cpu_load_single.
> > > >
> > > > Signed-off-by: Sebastien Dugue <[EMAIL PROTECTED]>
> > >
> > > Acked-by: Darren Hart <[EMAIL PROTECTED]>
> >
> > Eeek, sorry. Nacked! See below:
> > > > Cc: Darren Hart <[EMAIL PROTECTED]>
> > > > Cc: Tim Chavez <[EMAIL PROTECTED]>
> > > > ---
> > > > .../func/periodic_cpu_load/periodic_cpu_load.c | 11
> > > > +++++++++-- 1 files changed, 9 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git
> > > > a/testcases/realtime/func/periodic_cpu_load/periodic_cpu_load.c
> > > > b/testcases/realtime/func/periodic_cpu_load/periodic_cpu_load.c
> > > > index 2bf9c53..3c39ac3 100644
> > > > ---
> > > > a/testcases/realtime/func/periodic_cpu_load/periodic_cpu_load.c
> > > > +++
> > > > b/testcases/realtime/func/periodic_cpu_load/periodic_cpu_load.c
> > > > @@ -127,7 +127,7 @@ void *periodic_thread(void *thread)
> > > > void*(*func)(void*) = parg->func;
> > > >
> > > > int i = 0;
> > > > - nsec_t next;
> > > > + nsec_t next, now;
> > > > nsec_t exe_start, exe_end, exe_time;
> > > >
> > > > next = rt_gettime();
> > > > @@ -146,7 +146,14 @@ void *periodic_thread(void *thread)
> > > > dat[t->id].records[i].y = exe_time/NS_PER_US;
> > > >
> > > > i++;
> > > > - rt_nanosleep(next - rt_gettime());
> > > > +
> > > > + now = rt_gettime();
> > > > + if (now > next) {
> > > > + printf("Missed period, aborting (calc took too
> > > > long)\n");
> > > > + fail[t->id] = 1;
> >
> > fail is defined as an int, not an array. This causes the build to
> > fail. Simply replacing this line with:
> >
> > fail = 1;
> >
> > is an adequate fix.
>
> I redefined fail as: int fail[THREADS_PER_GROUP * NUM_GROUPS]; in
> the previous patch. I can't see how it can fail to build (and it does
> build and run fine here).
>
Apologies. Since I didn't ack #14 in the series, I didn't apply it.
Without it, this patch will of course fail to build. With the fixes
applied to 14, this also get's an Ack from me.
Acked-by: Darren Hart <[EMAIL PROTECTED]>
--
Darren Hart
IBM Linux Technology Center
Real-Time Linux Team
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
