RFR: 8326101: [PPC64] Need to bailout cleanly if creation of stubs fails when code cache is out of space [v2]

[Matthias Baesken]

On Sat, 17 Feb 2024 13:56:06 GMT, Martin Doerr <mdoerr at openjdk.org> wrote:

>> https://github.com/openjdk/jdk/commit/e834a481000f16750b9e9eec13ce9c905c218736 had missed some PPC64 parts. Note that the PPC64 implementation is similar to SPARC, so we need almost the same changes as that platform. `LIR_Assembler::emit_static_call_stub()`, `LIR_Assembler::emit_trampoline_stub_for_call` and `CallStubImpl::emit_trampoline_stub` already record failures internally.
>
> Martin Doerr has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Replace NULL by nullptr

Hi Martin, seems we miss the record_failure at some places too on s390

src/hotspot/cpu/s390/s390.ad 


1406   address base = __ start_a_stub(size_exception_handler());
1407   if (base == nullptr) {
1408     return 0;          // CodeBuffer::expand failed
1409   }
1410 


src/hotspot/cpu/s390/compiledIC_s390.cpp 


56    address stub = __ start_a_stub(CompiledDirectCall::to_interp_stub_size());
57    if (stub == nullptr) {
58      return nullptr;  // CodeBuffer::expand failed.
59    }


do you agree ? Should we file another JBS bug ?

-------------

PR Comment: https://git.openjdk.org/jdk/pull/17902#issuecomment-1952130291