[Darshan-users] Problem with darshan_log_getmounts and darshan_log_getexe

[Matthieu Dorier]

Hi Phil, 

Thanks for your answer to this and my other messages. 
If the order of function calls matters, then it probably makes sense that I don't get the right data. I'll check that tomorrow. 

Thanks again, 

Matthieu 

----- Mail original -----

> De: "Phil Carns" <carns at mcs.anl.gov>
> À: darshan-users at lists.mcs.anl.gov
> Envoyé: Dimanche 14 Avril 2013 14:25:58
> Objet: Re: [Darshan-users] Problem with darshan_log_getmounts and
> darshan_log_getexe

> On 04/13/2013 07:15 PM, Matthieu Dorier wrote:

> > Hi,
> 

> > I'm using the C API of Darshan and I have some problems with the
> > two
> > functions mentioned in the title of this mail.
> 
> > I'm testing them using 10024744407537054797.bz2 (in
> > http://ftp.mcs.anl.gov/pub/darshan/data/intrepid/2012/9/1/ ).
> 
> > This trace is supposed to show 4 mount points, which I see when
> > using
> > the darshan-parse command line tool, yet darshan_log_getmounts
> > gives
> > me 0 mount points.
> 
> > The "exe" field is supposed to contain "3979759305", but when
> > calling
> > darshan_log_getexe I get the string "2.02".
> 
> > These functions do not return any error code.
> 

> > All other functions are working properly. I'm using version 2.2.5
> > of
> > Darshan.
> 

> > Just to check, here is how I use these functions:
> 

> > int64_t** devs;
> 
> > char*** mnt_pts;
> 
> > char*** fs_types;
> 
> > int count;
> 
> > int err = darshan_log_getmounts(fd,devs,mnt_pts,fs_types,&count);
> 
> > // here c = 0
> 
> > free(devs); free(mnt_pts); free(fs_types);
> 

> > char buffer[CP_EXE_LEN+2];
> 
> > darshan_log_getexe(fd, buffer);
> 

> > Is there a bug somewhere or am I misusing these functions?
> 
> > Thanks!
> 
> Hi Matthieu,

> You'll need to make a slight change to the arguments to getmounts to
> pass in references to the pointers to be filled in, I think:

> int64_t* devs;
> char** mnt_pts;
> char** fs_types;
> int count;
> int err = darshan_log_getmounts(fd,&devs,&mnt_pts,&fs_types,&count);
> If that doesn't work, its also possible that the calls to getexe()
> and getmounts() may need to be in the opposite order, but I'll need
> to double-check on Monday. From glancing at the code it looks like
> either way should be fine.

> One other minor detail is that technically when free'ing mnt_pts and
> fs_types you need to loop from i=0 to i=count-1 and free mnt_pts[i]
> and fs_types[i] as well, though.

> As you can see this API has not been used much beyond our own
> internal tools so far, so isn't as polished as it probably should be
> :)

> thanks,
> -Phil

> > Matthieu Dorier
> 
> > PhD student at ENS Cachan Brittany and IRISA
> 
> > http://people.irisa.fr/Matthieu.Dorier
> 

> > _______________________________________________
> 
> > Darshan-users mailing list Darshan-users at lists.mcs.anl.gov
> > https://lists.mcs.anl.gov/mailman/listinfo/darshan-users
> 
> _______________________________________________
> Darshan-users mailing list
> Darshan-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/darshan-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/darshan-users/attachments/20130414/28dd466b/attachment.html>