RFR (XS/T) 8245703: 32-bit build failures after JDK-8243392

[Thomas Stüfe]

Sorry for breaking 32bit. I am sure I had tested an earlier version, but
apparently not the latest one I pushed.

As for the style, I do not really care. I mostly use #ifdef xxx for
covering multiple ine,s xxx_ONLY() for single lines. I have no strong
emotions here.

Also this is a trivial patch, if you want to push right away it would be
fine with me. Thanks for fixing!

..Thomas



On Mon, May 25, 2020 at 10:00 AM Aleksey Shipilev <shade at redhat.com> wrote:

> On 5/25/20 9:56 AM, David Holmes wrote:
> >> diff -r 438102b06fc0 src/hotspot/share/memory/metaspaceShared.cpp
> >> --- a/src/hotspot/share/memory/metaspaceShared.cpp      Mon May 25
> 11:29:56 2020 +0800
> >> +++ b/src/hotspot/share/memory/metaspaceShared.cpp      Mon May 25
> 08:51:16 2020 +0200
> >> @@ -259,5 +259,7 @@
> >>     assert(DumpSharedSpaces, "should be called for dump time only");
> >>
> >> +#ifdef _LP64
> >>     check_SharedBaseAddress();
> >> +#endif
> >
> > I'd prefer
> >
> > LP64_ONLY(check_SharedBaseAddress();)
> >
> > but either way looks good and trivial.
>
> Yeah. Up to Thomas to decide which style to use here, I guess. I tried to
> match the format of
> subsequent block:
>
>  void MetaspaceShared::initialize_dumptime_shared_and_meta_spaces() {
>    assert(DumpSharedSpaces, "should be called for dump time only");
>
> +#ifdef _LP64
>    check_SharedBaseAddress();
> +#endif
>
>    const size_t reserve_alignment =
> MetaspaceShared::reserved_space_alignment();
>    char* shared_base = (char*)align_up((char*)SharedBaseAddress,
> reserve_alignment);
>
>  #ifdef _LP64
>    assert(CompressedKlassPointers::is_valid_base((address)shared_base),
> "Sanity");
>    // On 64-bit VM we reserve a 4G range and, if
> UseCompressedClassPointers=1,
>    //  will use that to house both the archives and the ccs. See below for
>    //  details.
>    const uint64_t UnscaledClassSpaceMax = (uint64_t(max_juint) + 1);
>
>
> --
> Thanks,
> -Aleksey
>
>