RFR(XS): 8232005: [s390] More exception checks missing in interpreter
Reingruber, Richard
richard.reingruber at sap.com
Fri Oct 11 09:35:45 UTC 2019
Hi Martin,
Thanks for analyzing and fixing those.
Have you looked at the following:
cpu/s390/interp_masm_s390.cpp:970: call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::monitorenter),
cpu/s390/interp_masm_s390.cpp-971- monitor, /*check_for_exceptions=*/false);
cpu/s390/interp_masm_s390.cpp:1150: call_VM(noreg, CAST_FROM_FN_PTR(address, InterpreterRuntime::monitorexit),
cpu/s390/interp_masm_s390.cpp-1151- monitor, /*check_for_exceptions=*/false);
cpu/ppc/templateInterpreterGenerator_ppc.cpp:595: __ call_VM(Rexception, CAST_FROM_FN_PTR(address, InterpreterRuntime::create_klass_exception), false);
cpu/ppc/templateInterpreterGenerator_ppc.cpp:598: __ call_VM(Rexception, CAST_FROM_FN_PTR(address, InterpreterRuntime::create_exception), false);
cpu/ppc/templateInterpreterGenerator_ppc.cpp:2108: __ MacroAssembler::call_VM(R4_ARG2, CAST_FROM_FN_PTR(address, InterpreterRuntime::member_name_arg_or_null), R4_ARG2, R19_method, R14_bcp, false);
x86 does the exception check there. I think the check is needed, because async. exceptions can get installed there and if we don't check, they could be dropped.
Thanks, Richard.
(Not yet Reviewer)
-----Original Message-----
From: hotspot-runtime-dev <hotspot-runtime-dev-bounces at openjdk.java.net> On Behalf Of Doerr, Martin
Sent: Dienstag, 8. Oktober 2019 16:42
To: hotspot-runtime-dev at openjdk.java.net; Schmidt, Lutz <lutz.schmidt at sap.com>; Langer, Christoph <christoph.langer at sap.com>
Subject: [CAUTION] RFR(XS): 8232005: [s390] More exception checks missing in interpreter
Hi,
the template interpreter on s390 misses more exception checks when calling VM functions which may install an async exception (like java.lang.ThreadDeath) at a safepoint.
Bug:
https://bugs.openjdk.java.net/browse/JDK-8232005
These are the two places I found:
http://cr.openjdk.java.net/~mdoerr/8232005_s390_check_exceptions/webrev.00/
Please review.
Best regards,
Martin
More information about the hotspot-runtime-dev
mailing list