[Buildroot] [PATCH] package/dmalloc: Fix for ARC with enabled BR2_SSP_ALL

[Fabrice Fontaine]

Hi Yann,

Le mer. 10 févr. 2021 à 22:45, Yann E. MORIN <yann.morin.1998 at free.fr> a écrit :
>
> Veronika, All,
>
> +Fabrice, see below...
>
> On 2021-02-09 13:47 +0300, Veronika Kremneva spake thusly:
> > When building dmalloc for ARC while having "-fstack-protector-all" option on,
> > we are getting the following error:
> > ------------------------------->8---------------------------
> > >>> dmalloc 5.5.2 Installing to target
> > cp -dpf $HOME/buildroot/output/host/arc-buildroot-linux-gnu/sysroot/usr/lib/libdmalloc*.so
> > $HOME/buildroot/output/target/usr/lib
> > cp -dpf $HOME/buildroot/output/host/arc-buildroot-linux-gnu/sysroot/usr/bin/dmalloc
> > $HOME/buildroot/output/target/usr/bin/dmalloc
> > ERROR: architecture for "/usr/lib/libdmallocthcxx.so" is "ARCompact", should be "ARCv2"
> > ERROR: architecture for "/usr/lib/libdmallocth.so" is "ARCompact", should be "ARCv2"
> > ERROR: architecture for "/usr/lib/libdmalloc.so" is "ARCompact", should be "ARCv2"
> > ERROR: architecture for "/usr/lib/libdmallocxx.so" is "ARCompact", should be "ARCv2"
> > make: *** [$HOME/buildroot/output/build/dmalloc-5.5.2/.stamp_installed] Error 1
> > ------------------------------->8---------------------------
> >
> > You can observe this error in following builds:
> > http://autobuild.buildroot.net/results/57f/57fc51cee4ec9339ea3be23f8c8c9f7ecca1a857/
> > http://autobuild.buildroot.net/results/b4d/b4d457369f35e1942c294b0ea3b3cab4cae0d7a3/
> > http://autobuild.buildroot.net/results/cfc/cfc111076c955dd5c1125c6e34a161121ea11200/
> > http://autobuild.buildroot.net/results/1d0/1d0ffcefbf545d1748e928603fd07c0a8de98eb5/
> >
> > In all of those builds architecture "ARCv2" was explicitly chosen during configuration
> > yet we are getting this error. As we go deeper we find that the object file that linker
> > is trying to process contains no target code thus the default "ARCompact" is being used.
> > But why the object file is empty? Went to check link args:
> > ------------------------------->8---------------------------
> > $HOME/buildroot/output/host/bin/arc-buildroot-linux-gnu-ld: conftest.a(conftest.o):
> > relocation R_ARC_32_ME against `__stack_chk_guard' can not be used when making a shared
> > object; recompile with -fPIC
> > $HOME/buildrootc/output/host/bin/arc-buildroot-linux-gnu-ld: BFD (GNU Binutils) 2.34.50.20200611
> > assertion fail elf32-arc.c:1802
> > $HOME/buildroot/output/host/bin/arc-buildroot-linux-gnu-ld: unrecognized option '-all'
> > $HOME/buildroot/output/host/bin/arc-buildroot-linux-gnu-ld: use the --help option for usage information
> > configure:4475: result: $HOME/buildroot/output/host/bin/arc-buildroot-linux-gnu-ld -G -o $@.t
> > ------------------------------->8---------------------------
> >
> > Looks like we are missing the -fPIC flag in DMALLOC_CFLAGS. As shared libraries should always
> > be compiled with "-fPIC" following patch adds this flag to the dmalloc.mk.
>
> While reviewing that patch, I see that Fabrice already sent a similar
> patch quite some time ago now:
>
>     https://patchwork.ozlabs.org/project/buildroot/patch/20200601092104.830051-1-fontaine.fabrice@gmail.com/
>
> > Signed-off-by: Veronika Kremneva <kremneva at synopsys.com>
> > ---
> >  package/dmalloc/dmalloc.mk | 4 ++++
> >  1 file changed, 4 insertions(+)
> >
> > diff --git a/package/dmalloc/dmalloc.mk b/package/dmalloc/dmalloc.mk
> > index 38b2c02027..638fe0fcda 100644
> > --- a/package/dmalloc/dmalloc.mk
> > +++ b/package/dmalloc/dmalloc.mk
> > @@ -16,6 +16,10 @@ DMALLOC_INSTALL_STAGING = YES
> >  DMALLOC_CONF_OPTS = --enable-shlib
> >  DMALLOC_CFLAGS = $(TARGET_CFLAGS)
> >
> > +ifeq ($(BR2_STATIC_LIBS),)
> > +DMALLOC_CFLAGS += -fPIC
>
> So, it seems that both of you arrived to a similar conclusion, with a
> similar analysis.
>
> I think the change from Fabrice would in the end be better, if it at
> least because it drops a bundled patch.
>
> Except for that, I wonder if we care about not adding -fPIC for static
> libs. Indeed, the current patch unconditionally adds -fPIC to all
> builds, shared and static alike...
>
> Besides, -fPIC is almost innocuous for static builds (maybe a very
> slight overhead to resolve or jump to some symbols)...
>
> What's your thoughts, both Veronika and Fabrice?
It seems that we're not always consistent with -fPIC handling:
 - most of the packages only set -fPIC if BR2_STATIC_LIBS is not set
 - but some packages (e.g. argp-standalone, ficl, jimtcl, lbase64 ...)
set it unconditionally
So I have no strong opinion on both options.
>
> Regards,
> Yann E. MORIN.
>
> > +endif
> > +
> >  ifeq ($(BR2_INSTALL_LIBSTDCPP),y)
> >  DMALLOC_CONF_OPTS += --enable-cxx
> >  else
> > --
> > 2.16.2
> >
> > _______________________________________________
> > buildroot mailing list
> > buildroot at busybox.net
> > http://lists.busybox.net/mailman/listinfo/buildroot
>
> --
> .-----------------.--------------------.------------------.--------------------.
> |  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
> | +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
> | +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
> | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
> '------------------------------^-------^------------------^--------------------'
Best Regards,

Fabrice