zero broken in jdk/jdk? [resend ]

[Severin Gehwolf]

On Wed, 2018-12-12 at 16:13 -0500, Chris Phillips wrote:
> Hi Magnus
> Oopsn^ 2
> 
> On 2018-12-12 3:58 p.m., Chris Phillips wrote:
> > Oops.
> > 
> > On 2018-12-12 3:51 p.m., Chris Phillips wrote:
> > > Hi Magnus
> > > 
> > > On 2018-12-12 1:53 p.m., "Chris Phillips"@T O wrote:
> > > > Hi Magnus,
> > > > 
> > > > On 12/12/18 01:28 PM, Magnus Ihse Bursie wrote:
> > > > > I get the following build error when building zero in jdk/jdk:
> > > > > 
> > > > > workspace/build/linux-x64-zero/hotspot/variant-zero/libjvm/objs/os_linux.o:
> > > > > In function `os::abort(bool, void*, void const*)':
> > > > > workspace/open/src/hotspot/os/linux/os_linux.cpp:1358: undefined
> > > > > reference to `ClassLoader::close_jrt_image()'
> > > > > collect2: error: ld returned 1 exit status
> > > > > lib/CompileJvm.gmk:172: recipe for target
> > > > > 'workspace/build/linux-x64-zero/support/modules_libs/java.base/server/libjvm.so'
> > > > > failed
> > > > > 
> > > > > Is this a known problem?
> > > > > 
> > > > > /Magnus
> > > > > 
> > > > > 
> > > > 
> > > > Not seen this before, will take a look.
> > > > 
> > > > Chris
> > > > 
> > > > 
> > > 
> > > Yes, somebody broke Zero  (Recent CDS change?)
> > > 
> > > The following fixes it please do open a bug or shall I?
> > > 
> 
> 
> This works, not sure why #ifdef INCLUDE_CDS doesn't.
> 
> (the body of close_jrt_image is under an ifdef INCLUDE_CDS, but maybe
>  there is a higher level #ifdef? )
> 
> diff --git a/src/hotspot/os/linux/os_linux.cpp
> b/src/hotspot/os/linux/os_linux.cpp
> --- a/src/hotspot/os/linux/os_linux.cpp
> +++ b/src/hotspot/os/linux/os_linux.cpp
> @@ -1354,9 +1354,11 @@
>  void os::abort(bool dump_core, void* siginfo, const void* context) {
>    os::shutdown();
>    if (dump_core) {
> +#ifndef ZERO
>      if (UseSharedSpaces && DumpPrivateMappingsInCore) {
>        ClassLoader::close_jrt_image();
>      }
> +#endif
>  #ifndef PRODUCT
>      fdStream out(defaultStream::output_fd());
>      out.print_raw("Current thread is ");

Hi Chris,

I'll create a bug and post this for RFR.

Thanks,
Severin