RFR: JDK-8144134: Nightly tests fail with SIGSEGV in Ticks::now()
mikael.gerdin at oracle.com
Fri Nov 27 09:40:44 UTC 2015
Since you added the explicit stack alignment parameter as part of the
integration of Clang compiler support:
What is the reason for setting it?
Isn't the stack alignment already "16" on OS X x64?
Some comments talk about i486 support, if setting the stack alignment
explicitly is only needed on 32 bit builds, would it make sense to only
set stack alignment on such builds?
On 2015-11-27 10:19, Mikael Gerdin wrote:
> Hi Erik,
> On 2015-11-27 10:16, Erik Joelsson wrote:
>> Please review this small fix for the nightly test failure in SA. In my
>> makefile conversion I seem to have accidentally deleted a character. In
>> this case it had rather interesting side effects.
> Interesting indeed!
> Looks good to me.
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8144134
>> diff -r b8bc00e338c4 make/lib/Lib-jdk.hotspot.agent.gmk
>> --- a/make/lib/Lib-jdk.hotspot.agent.gmk
>> +++ b/make/lib/Lib-jdk.hotspot.agent.gmk
>> @@ -79,7 +79,7 @@
>> SA_EXCLUDE_FILES := BsdDebuggerLocal.c ps_proc.c salibelf.c
>> SA_CFLAGS := $(CFLAGS_JDKLIB) \
>> -Damd64 -D_GNU_SOURCE -mno-omit-leaf-frame-pointer \
>> - -mstack-alignment=1 -fPIC
>> + -mstack-alignment=16 -fPIC
>> SA_LDFLAGS := $(LDFLAGS_JDKLIB)
>> SA_LIBS := -framework Foundation -framework JavaNativeFoundation \
>> -framework Security -framework CoreFoundation
More information about the hotspot-dev