Re[4]: [Gc] Native Client support?

Ivan Maidanski ivmai at mail.ru
Wed Dec 29 12:24:13 PST 2010


Hi Manuel,

1. The patch is just a diff between the current CVS snap and the previous one, so it is most probably not usable for other GC versions.
2. Android port in BDWGC is not (or should not) be worse than that for mono/libgc.
3. I haven't test it myself on Android (I only compiled it in the past w/o linking)
4. According to the comment in gcconfig.h, __stack_base__ should be defined in newlib/libc/sys/arm/crt0.S.
5. If something is wrong and you know how to fix it properly, please send me a patch.
6. I suggest you also use -Wall and -fno-strict-aliasing.
7. Is -I libatomic_ops-install/include the same as -I libatomic_ops/include?
8. If you are not defining -DGC_THREADS then there is no sense in -Ilibatomic_ops-install/include.

Regards.

Wed, 29 Dec 2010 06:34:02 +0100 Manuel.Serrano at inria.fr:

> Hi Ivan,
> 
> I'm a little bit confused with the current state of the Android port. Is
> the current CVS version fully operational on Android? Does the current
> GC version supports multi-threading *and* shared libraries *and* dynamic
> loading on Android?
> 
> Currently, I still fail to compile the collector with either the patch you
> send or the version from the CVS. It complains about __stack_base__
> 
> os_dep.c:(.text.GC_get_main_stack_base+0x1c): undefined reference to
> `__stack_base__'
> collect2: ld returned 1 exit status
> 
> I compile (on my Linux-2.6.35 + android-gcc box) with the following options:
> 
> -DHAVE_CONFIG_H -I./include -I./include -fexceptions -O3 -DPLATFORM_ANDROID
> -DNO_DEBUGGING -Iinclude -Ilibatomic_ops-install/include
> -DATOMIC_UNCOLLECTABLE -DNO_EXECUTE_PERMISSION -DFINALIZE_ON_DEMAND
> 
> Am I doing something wrong? Thanks in advance for your help.
> 
> -- 
> Manuel
> 
> _______________________________________________
> Gc mailing list
> Gc at linux.hpl.hp.com
> http://www.hpl.hp.com/hosted/linux/mail-archives/gc/



More information about the Gc mailing list