-
Bug
-
Resolution: Fixed
-
P3
-
9
-
b96
While testing Graal with JDK9 we hit a SEGV with latest C2.
#
# A fatal error has been detected by the Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0x00007f5708e8e89e, pid=13481, tid=13508
#
# JRE version: OpenJDK Runtime Environment (9.0) (build 1.9.0-internal-_2015_11_10_13_33-b00)
# Java VM: OpenJDK 64-Bit Server VM (1.9.0-internal-_2015_11_10_13_33-b00, mixed mode, tiered, jvmci, compressed oops, g1 gc, linux-amd64)
# Problematic frame:
# V [libjvm.so+0x49189e] JVMState::same_calls_as(JVMState const*) const+0xe
#
# Core dump will be written. Default location: Core dumps may be processed with "/usr/share/apport/apport %p %s %c %P" (or dumping to /slave/buildlog/CI-graal-compiler_gate_amd64_linux_postgate_jdk9-SNAPSHOT_jdk8_jdk7/build/graal-compiler/core.13481)
#
# If you would like to submit a bug report, please visit:
# http://bugreport.java.com/bugreport/crash.jsp
#
V [libjvm.so+0x49189e] JVMState::same_calls_as(JVMState const*) const+0xe
V [libjvm.so+0x6ffe0b] IfNode::has_only_uncommon_traps(ProjNode*, ProjNode*&, ProjNode*&, PhaseIterGVN*)+0x1ab
V [libjvm.so+0x702726] IfNode::fold_compares(PhaseIterGVN*)+0x236
V [libjvm.so+0x703a1b] IfNode::Ideal(PhaseGVN*, bool)+0xc9b
Inspection of the code showed it was possible for both dom_caller and caller to be null here:
JVMState* dom_caller = dom_unc->jvms()->caller();
JVMState* caller = unc->jvms()->caller();
if (!dom_caller->same_calls_as(caller)) {
return false;
}
It should at least be checking that both are non-NULL though I think that both of them being NULL would actually be ok. I pushed the following to our local repo http://hg.openjdk.java.net/graal/graal-jvmci-9/hotspot/rev/e9d1036f48c1
#
# A fatal error has been detected by the Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0x00007f5708e8e89e, pid=13481, tid=13508
#
# JRE version: OpenJDK Runtime Environment (9.0) (build 1.9.0-internal-_2015_11_10_13_33-b00)
# Java VM: OpenJDK 64-Bit Server VM (1.9.0-internal-_2015_11_10_13_33-b00, mixed mode, tiered, jvmci, compressed oops, g1 gc, linux-amd64)
# Problematic frame:
# V [libjvm.so+0x49189e] JVMState::same_calls_as(JVMState const*) const+0xe
#
# Core dump will be written. Default location: Core dumps may be processed with "/usr/share/apport/apport %p %s %c %P" (or dumping to /slave/buildlog/CI-graal-compiler_gate_amd64_linux_postgate_jdk9-SNAPSHOT_jdk8_jdk7/build/graal-compiler/core.13481)
#
# If you would like to submit a bug report, please visit:
# http://bugreport.java.com/bugreport/crash.jsp
#
V [libjvm.so+0x49189e] JVMState::same_calls_as(JVMState const*) const+0xe
V [libjvm.so+0x6ffe0b] IfNode::has_only_uncommon_traps(ProjNode*, ProjNode*&, ProjNode*&, PhaseIterGVN*)+0x1ab
V [libjvm.so+0x702726] IfNode::fold_compares(PhaseIterGVN*)+0x236
V [libjvm.so+0x703a1b] IfNode::Ideal(PhaseGVN*, bool)+0xc9b
Inspection of the code showed it was possible for both dom_caller and caller to be null here:
JVMState* dom_caller = dom_unc->jvms()->caller();
JVMState* caller = unc->jvms()->caller();
if (!dom_caller->same_calls_as(caller)) {
return false;
}
It should at least be checking that both are non-NULL though I think that both of them being NULL would actually be ok. I pushed the following to our local repo http://hg.openjdk.java.net/graal/graal-jvmci-9/hotspot/rev/e9d1036f48c1
- relates to
-
JDK-8140574 C2 must re-execute checks after deoptimizing from merged uncommon traps
- Closed