During an exp-run for llvm 13 (see bug 258209), it turned out that
java/openjdk11 through openjdk13 fail to build with clang 13:
=== Output from failing command(s) repeated here ===
* For target jdk__packages_attribute.done:
These crashes are all caused by the markOop/markOopDesc classes, which
are used to keep track of objects, and which are 'marked' using the low
few bits. (See
https://github.com/openjdk/jdk13u/blob/master/src/hotspot/share/oops/markOop.hpp
).
After some laborious bisecting, I found out that these crashes start
occuring after the upstream commit https://github.com
/llvm/llvm-project/commit/16d03818412 (Return "[CGCall] Annotate this
argument with alignment").
What happens afterwards, is that clang considers the "this" pointer to
always be aligned to the alignment of the actual object, and then
masking or adding a few low bits is not working as expected.
The reason openjdk14 and higher work fine with clang 13, and don't crash
similarly, is that the OpenJDK people completely redid the
markOop/markOopDesc classes in
ae5615c614
("8229258: Rework markOop and markOopDesc into a simpler mark word value
carrier"). E.g, the markOopDesc class was renamed to markWord, and
*stores* a pointer-like value instead of *being* a pointer-like value.
This is a much safer way of handling things.
However, this upstream commit is *very* large, as are a few of its
follow-ups, which is probably the reason why it has not been backported
to JDKs <= 13. I tried manually backporting it, but got lost in many
nasty patch conflicts and problems.
As a workaround, build openjdk8 through 13 with clang12 from the
devel/llvm12 port, for the time being.
In addition, allow openjdk14 through 17 to be built with clang 13, by
adding -Wno-unused-but-set-parameter to the compilation flags.
PR: 258954
Approved by: maintainer timeout (2 weeks)
MFH: 2021Q4
Running java or javac executables fails with:
Error: dl failure on line 910
Error: failed /usr/local/openjdk8/jre/lib/ppc64le/server/libjvm.so, because /usr/local/openjdk8/jre/lib/ppc64le/server/libjvm.so: Undefined symbol "_ZN5frameC1EPlPh"
$ make checksum WITH=TEST
===> License GPLv2 accepted by the user
===> openjdk8-8.302.08.1 depends on file: /usr/local/sbin/pkg - found
===> Fetching all distfiles required by openjdk8-8.302.08.1 for building
=> No SHA256 checksum recorded for jtreg4.1-b08.tar.gz.
=> No suitable checksum found for jtreg4.1-b08.tar.gz.
=> SHA256 Checksum OK for battleblow-jdk8u-8.302.08.1-jdk8u302-b08.1_GH0.tar.gz.
*** Error code 1
Reported by: Philipp Ost (via ports@ list)
powerpc64le needs --disable-jfr and USE_PRECOMPILED_HEADER=0 - same as powerpc64 elfv2.
Clang refuses to build little-endian binaries for power7, this needs to be bumped to power8 (this is a bug in openjdk, POWER7 is not LE-compatible).
-minsert-sched-nops=regroup_exact -mno-multiple -mno-string are not supported by clang and need to be removed.
FreeBSD uses sys/endian.h instead of byteswap.h and bswap{16,32,64} instead of bswap_{16,32,64}.
PR: 251247
Approved by: java (maintainer timeout)
Remove files/patch-hotspot_src_cpu_ppc_vm_sharedRuntime__ppc.cpp to fix patching.
I'm not sure why this patch was added. Removing it doesn't break build on powerpc64 (both elfv1 and elfv2). There's no include for alloca.h in sharedRuntime_ppc.cpp.
Chase the devel/libffi update
Bump portrevision of all dependent ports to chace shard library version bump
in libffi.
Update LIB_DEPENDS lines where needed to not require a specific version of
libffi.so.
PR: 247028 (for tracking)
Since the maintainer didn't respond, I'm adding the bootstrap to my own directory on freefall.
PR: 242965
Submitted by: Mikael Urankar (original version)
Approved by: java (maintainer timeout)
* Having /proc mounted is not generally critical, but there is at least one
piece of functionality that does rely on its presence at the moment
(e.g. OperatingSystemMXBean)
* Remove portions of pkg-message that mention needing /proc mounted.
This hasn't been the case for a while afaik (I don't have it mounted
and haven't observed any related problems).
* FreeBSD crashes on infinite recursion rather than throwing a stack
overflow because the stack address of the fault is in the page below
the guard area. Workaround this by rounding down the fault address to
the nearest page boundary. Investigation is still under way into what
may be causing this but this appears to prevent it in simple test cases.
PR: 222146
* Updated the fontconfig.patch file
* Simplified FONTCONFIG to use EXTRA_PATCHES rather than a separate target
PR: 239358
Submitted by: John Hein <jcfyecrayz@liamekaens.com>, fluffy
