Re: trying to understand what's going on with timers and boot process

[Mark Seger <Mark dot Seger at hp dot com>]

As I said, my problem is with Time::HiRes which is a perl module.  Maybe 
I should have been a little more specific.  The basic problem is you 
call the modules with the time you want to set the alarm for in usecs.  
The problem is in the broken version if HiRes, which I said is currently 
shipped with all distros of RedHat and probably others as well, calls 
setitimer with that value in its usec field.  here's an example of an 
strace when setting a timer of 2 seconds on a broken implementation of 
HiRes:

setitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={0, 2000000}}, 
{it_interval={0, 0}, it_value={0, 0}}) = 0

notice the second field of it_value is incorrectly set to 2000000, but 
on this system which is running glibc 2.3 it works just fine!  now here 
it is again with a correct version of HiRes:

setitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={2, 0}}, NULL) = 0

As I said in my original note I realize the first call in invalid, but 
it still works correctly with glibc 2.3 and most of the time with 2.5 
and I don't want to force everyone who uses my script to upgrade their 
HiRes (I do have a bugzilla into redhat on this) and so am trying to 
understand the cases where it will fail and why, especially why it only 
appears to fail during boot when 2.5 if glibc is present.

Does that help?

-mark

Carlos O'Donell wrote:
> On Tue, Jul 1, 2008 at 9:09 AM, Mark Seger <Mark.Seger@hp.com> wrote:
>   
> > My issue is I have an open source program that is dependent on HiRes and is
> > being distributed on Fedora 10 and I fear I'm going to start getting a lot
> > of support requests over this.  I do intend to release a new version that
> > checks the version of glibc (I recently discovered I could just run
> > /lib/libc.6.so and it tells me its version) and warn people when there is a
> > problem, but I'm still stumped with the fact that I've only seen errors on
> > 2.5 systems during boot.  Can someone shed any light on what I'm seeing?
> >  Should I be worrying about ALL versions of glibc > 2.3?
> >     
>
> The first step is to develop a pure-C test case. Without a test case
> it's going to be very hard to debug your problem.
>
> Do you have a test case?
>
> Cheers,
> Carlos.
>