Direct binding and libumem

Yann Rouillard yann at pleiades.fr.eu.org
Wed Mar 5 00:00:08 CET 2014


Hi Dago,

Yes, that's a known special case that happens in some rare cases.
In your case, the binary uses only malloc, free, realloc and that set of
symbols can't be directly bound to, wether it is in libumem or libc.
I suppose this is done on purpose to allow to change the memory allocator.

I already excluded some native libraries and I will add libumem to the list.

Even better would be to check if the linked symbols can't be directly bound
to but that would make the test more complex and I don't think it's worth
it for now.


Yann


2014-03-04 9:34 GMT+01:00 Dagobert Michelsen <dam at opencsw.org>:

> Hi folks,
>
> I just encountered an interesting case: when compiling the latest xjobs
> (which does
> not have funky linker magic at all) I got the following checkpkg result:
>
> > CHECKPKG_OVERRIDES_CSWxjobs +=
> no-direct-binding|/opt/csw/bin/xjobs|is|not|directly|bound|to|soname|libumem.so.1
>
>
> Inspection of the build process showed that -B direct is correctly passed.
> Using the excellent documentation at the checkpkg error tag wiki at
>   http://wiki.opencsw.org/checkpkg-error-tags#toc31
> showed the following issue:
>
> > dam at unstable10s [unstable10s]:/home/dam/mgar/pkg/xjobs/trunk > $(mgar
> show-buildsys | cut -f1)/bin/check_db_symbols
> work/solaris10-sparc/pkgroot/opt/csw/bin/xjobs
> > Use of uninitialized value $library_path in string eq at
> /home/dam/mgar/pkg/.buildsys/v2/bin/check_db_symbols line 64, <$ldd> line 4.
> > Use of uninitialized value $soname in hash element at
> /home/dam/mgar/pkg/.buildsys/v2/bin/check_db_symbols line 67, <$ldd> line 4.
> >
> > Library                 Directly bound    Not directly bound    Not
> directly bindable
> > libumem.so.1                    0                  0
>  3
> > libc.so.1                      74                  0
>  2
> > libm.so.2                       2                  0
>  0
>
>
> It seems that some libraries contain only not directly bindable symbols,
> essentially these should be excluded from the test.
> @Yann: Do you think this is possible?
>
>
> Best regards
>
>   -- Dago
>
> --
> "You don't become great by trying to be great, you become great by wanting
> to do something,
> and then doing it so hard that you become great in the process." - xkcd
> #896
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opencsw.org/pipermail/maintainers/attachments/20140305/a77f86f3/attachment.html>


More information about the maintainers mailing list