[Rd] repeatable segfault - on Mac

[luke-tierney at uiowa.edu]

On Thu, 8 Sep 2011, Martin Maechler wrote:

>>>>>> Simon Urbanek <simon.urbanek at r-project.org>
>>>>>>     on Thu, 8 Sep 2011 09:33:23 -0400 writes:
>
>    > On Sep 8, 2011, at 4:01 AM, Martin Maechler wrote:
>
>    >>>>>>> David Winsemius <dwinsemius at comcast.net>
>    >>>>>>> on Tue, 6 Sep 2011 00:38:13 -0400 writes:
>    >>
>    >>> I can reproduce:
>    >>> eigen(crossprod(matrix(1:2000, 50)) + (0+0i), T, T)
>    >>
>    >>> *** caught segfault ***
>    >>> address 0x102d0e028, cause 'memory not mapped'
>    >>
>    >>> Traceback:
>    >>> 1: .Call("La_rs_cmplx", x, only.values, PACKAGE = "base")
>    >>> 2: eigen(crossprod(matrix(1:2000, 50)) + (0 + (0+0i)), T, T)
>    >>> ##---- after restart
>    >>>> sessionInfo()
>    >>> R version 2.13.1 (2011-07-08)
>    >>> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
>    >>
>    >>> locale:
>    >>> [1] en_US.UTF-8/en_US.UTF-8/C/C/en_US.UTF-8/en_US.UTF-8
>    >>
>    >>> attached base packages:
>    >>> [1] stats     graphics  grDevices utils     datasets  methods   base
>    >>
>    >>> With OSX 10.5.8
>    >>
>    >>> Also happens with 32 bit R
>    >>> [R.app GUI 1.41 (5874) i386-apple-darwin9.8.0]
>    >>
>    >> It does not happen on Linux (different versions)
>    >> nor Windows (2.13.1 patched, early August).
>    >>
>    >> As this is calling LAPACK code,
>    >> I guess that this is yet another case where the Mac version
>    >> of optimized BLAS / LAPACK is playing wrongly.
>    >>
>
>    > Nope, this is R's BLAS / LAPACK!
>
> yes, it seems, as Peter is confirming as well.
>
> *BUT* as far as we currently know, the problem was never seen
> outside of MacOS X , right?
>

My experience is that the Mac OS malloc is less forgiving of writing
outside allocated memory than the one usually used on Linux, so some
things show up only on macs or are easier to trigger on macs even
though the problems exist else where as well. Valgrind probably makes
up for that most of the time but maybe not always.

Best,

luke


> Martin Maechler
>
>    >>> David Winsemius
>    >>
>    >>> On Sep 6, 2011, at 12:12 AM, robin hankin wrote:
>    >>
>    >>>> Hi.  macosx 10.6.8
>    >>>>
>    >>>> With R-2.13.1 and also revision 56948 I get the following repeatable
>    >>>> segfault:
>    >>>>
>    >>>>
>    >>>>
>    >>>> wt118:~% R --vanilla --quiet
>    >>>>> R.Version()
>    >>>> $platform
>    >>>> [1] "x86_64-apple-darwin9.8.0"
>    >>>>
>    >>>> $arch
>    >>>> [1] "x86_64"
>    >>>>
>    >>>> $os
>    >>>> [1] "darwin9.8.0"
>    >>>>
>    >>>> $system
>    >>>> [1] "x86_64, darwin9.8.0"
>    >>>>
>    >>>> $status
>    >>>> [1] ""
>    >>>>
>    >>>> $major
>    >>>> [1] "2"
>    >>>>
>    >>>> $minor
>    >>>> [1] "13.1"
>    >>>>
>    >>>> $year
>    >>>> [1] "2011"
>    >>>>
>    >>>> $month
>    >>>> [1] "07"
>    >>>>
>    >>>> $day
>    >>>> [1] "08"
>    >>>>
>    >>>> $`svn rev`
>    >>>> [1] "56322"
>    >>>>
>    >>>> $language
>    >>>> [1] "R"
>    >>>>
>    >>>> $version.string
>    >>>> [1] "R version 2.13.1 (2011-07-08)"
>    >>>>
>    >>>>> eigen(crossprod(matrix(1:2000, 50)) + (0+0i), T, T)
>    >>>>
>    >>>> *** caught segfault ***
>    >>>> address 0x1038000a8, cause 'memory not mapped'
>    >>>>
>    >>>> Traceback:
>    >>>> 1: .Call("La_rs_cmplx", x, only.values, PACKAGE = "base")
>    >>>> 2: eigen(crossprod(matrix(1:2000, 50)) + (0 + (0+0i)), T, T)
>    >>>>
>    >>>> Possible actions:
>    >>>> 1: abort (with core dump, if enabled)
>    >>>> 2: normal R exit
>    >>>> 3: exit R without saving workspace
>    >>>> 4: exit R saving workspace
>    >>>> Selection: 2
>    >>>> wt118:~%
>    >>>>
>    >>>>
>    >>>>
>    >>>>
>    >>>>
>    >>>> --
>    >>>> Robin Hankin
>    >>>> Uncertainty Analyst
>    >>>> hankin.robin at gmail.com
>    >>>>
>    >>>> ______________________________________________
>    >>>> R-devel at r-project.org mailing list
>    >>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>    >>
>    >>> David Winsemius, MD
>    >>> Heritage Laboratories
>    >>> West Hartford, CT
>    >>
>    >>> ______________________________________________
>    >>> R-devel at r-project.org mailing list
>    >>> https://stat.ethz.ch/mailman/listinfo/r-devel
>    >>
>    >> ______________________________________________
>    >> R-devel at r-project.org mailing list
>    >> https://stat.ethz.ch/mailman/listinfo/r-devel
>    >>
>    >>
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

-- 
Luke Tierney
Chair, Statistics and Actuarial Science
Ralph E. Wareham Professor of Mathematical Sciences
University of Iowa                  Phone:             319-335-3386
Department of Statistics and        Fax:               319-335-3017
    Actuarial Science
241 Schaeffer Hall                  email:      luke at stat.uiowa.edu
Iowa City, IA 52242                 WWW:  http://www.stat.uiowa.edu