RTC to 1.7: fix for hs16 + zero

[Andrew John Hughes]

2010/1/15 Edward Nevill <ed at camswl.com>:
> Hi folks,
>
> Attached is a fix for building hs16 on zero. As far as I can tell it is broken on all zero platforms, whether or not they build with Shark.
>
> The problem is simply that the function is_allocatable in os_linux_zero.cpp is unimplemeted. Now hs14 did not call this but hs16 does seem to call it.
>
> The 'fix' I have implemented is simply to copy the implementation from os_linux_x86.cpp to os_linux_zero.cpp.
>
> I have tested this with the test_gamma but have not done any further testing as I will have to do a complete rebuild to test it fully.
>
> (it cannot be worse than the current implemetation which just called ShouldNotCallThis()).
>
> Ok to commit?
>
> Regards,
> Ed.
>
> --- CUT HERE ---
> diff -ruNE old/icedtea6-1.7/ChangeLog new/icedtea6-1.7/ChangeLog
> --- old/icedtea6-1.7/ChangeLog  2010-01-15 16:57:45.000000000 +0000
> +++ new/icedtea6-1.7/ChangeLog  2010-01-15 17:02:23.000000000 +0000
> @@ -1,3 +1,8 @@
> +2010-01-15  Edard Nevill <ed at camswl.com>
> +
> +        * ports/hotspot/src/os_cpu/linux_zero/vm/os_linux_zero.cpp
> +       Implement is_allocatable to make build word with hs16
> +
>  2010-01-14  Matthias Klose  <doko at ubuntu.com>
>
>        * Makefile.am (stamps/icedtea-npplugin): New target, add file
> diff -ruNE old/icedtea6-1.7/ports/hotspot/src/os_cpu/linux_zero/vm/os_linux_zero.cpp new/icedtea6-1.7/ports/hotspot/src/os_cpu/linux_zero/vm/os_linux_zero.cpp
> --- old/icedtea6-1.7/ports/hotspot/src/os_cpu/linux_zero/vm/os_linux_zero.cpp   2010-01-15 16:57:47.000000000 +0000
> +++ new/icedtea6-1.7/ports/hotspot/src/os_cpu/linux_zero/vm/os_linux_zero.cpp   2010-01-15 16:59:56.000000000 +0000
> @@ -249,7 +249,23 @@
>  }
>
>  bool os::is_allocatable(size_t bytes) {
> -  ShouldNotCallThis();
> +#ifdef AMD64
> +  // unused on amd64?
> +  return true;
> +#else
> +
> +  if (bytes < 2 * G) {
> +    return true;
> +  }
> +
> +  char* addr = reserve_memory(bytes, NULL);
> +
> +  if (addr != NULL) {
> +    release_memory(addr, bytes);
> +  }
> +
> +  return addr != NULL;
> +#endif // AMD64
>  }
>
>  ///////////////////////////////////////////////////////////////////////////////
>
>
>
>
>
>

Looks good to me, please push to HEAD and 1.7.

Thanks,
-- 
Andrew :-)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net

PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint: F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8