Re: [PATCH 15/26] vm: set user.dir, user.home, user.name properties

Tomek Grabiec wrote: Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- vm/jato.c |7 +++ 1 files changed, 7 insertions(+), 0 deletions(-) diff --git a/vm/jato.c b/vm/jato.c index 4a82048..ea176ca 100644 --- a/vm/jato.c +++ b/vm/jato.c @@ -206,6 +206,13 @@ static void

Re: [PATCH 25/26] x86: fix 'this' pointer passing to fixup_vtable() for JNI methods.

Tomek Grabiec wrote: Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- arch/x86/emit-code.c | 17 +++-- 1 files changed, 15 insertions(+), 2 deletions(-) diff --git a/arch/x86/emit-code.c b/arch/x86/emit-code.c index 57de052..521c150 100644 --- a/arch/x86/emit-code.c +++

Re: [PATCH 01/26] vm: introduce vm_thread_get_name()

The series has been applied, thanks! -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core

Re: [PATCH 09/26] regression: introduce JNI regression tests

Hi Tomek, I skipped this patch because it's causing too many problems. The regression test didn't compile: make -C regression/jni make[1]: Entering directory `/home/penberg/src/jato/regression/jni' JAVAHjvm_JNITest.h CXX libjnitest.so In file included from jvm_JNITest.cpp:1:

[PATCH] x86: add USE_DST to INSN_ADD_IMM_REG

add imm, reg also uses the reg, so we should put this in the use-def. Someone should probably go through the whole list to verify that the rest is correct too. Signed-off-by: Vegard Nossum vegard.nos...@gmail.com --- arch/x86/use-def.c |2 +- 1 files changed, 1 insertions(+), 1 deletions(-)

[RFC][PATCH] test: fix spill-reload-test

The test was (as a single bb): add r1, r1, r1 add r2, r2, r2 r2 is expected to be reloaded, but because liveness analysis was wrong, it did not expect r1 to be reloaded. Now we change the code to: add r2, r1, r1 add r2, r2, r2 and expect the reload to appear

[PATCH 2/4] x86: remove unnecessary XXX markers

Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- arch/x86/emit-code.c | 10 +- 1 files changed, 5 insertions(+), 5 deletions(-) diff --git a/arch/x86/emit-code.c b/arch/x86/emit-code.c index 521c150..2afd1fb 100644 --- a/arch/x86/emit-code.c +++ b/arch/x86/emit-code.c @@ -395,7

[PATCH 1/4] vm: cleanup brace placement in jni-interface.c

Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- vm/jni-interface.c | 15 ++- 1 files changed, 10 insertions(+), 5 deletions(-) diff --git a/vm/jni-interface.c b/vm/jni-interface.c index 44efde2..dfba479 100644 --- a/vm/jni-interface.c +++ b/vm/jni-interface.c @@ -52,19 +52,23

[PATCH 3/4] vm: use trace_printf() to print the whole backtrace.

show_function() and show_exe_function() were not using the trace buffer which led to inconsistent backtrace output. Reported-by: Pekka Enberg penb...@cs.helsinki.fi Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- arch/x86/backtrace.c | 28 ++--

[PATCH 4/4] vm: do not try to put null values into system properties.

It is an error because Java API says that getProperty() returns null when property is not set. Reported-by: Pekka Enberg penb...@cs.helsinki.fi Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- vm/jato.c |3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/vm/jato.c

[PATCH] jit: add live-in/use/def/live-out set dumping to liveness analysis tracing

This helped me resolve a bug, so I suppose it's useful in general. Signed-off-by: Vegard Nossum vegard.nos...@gmail.com --- jit/trace-jit.c | 55 ++- 1 files changed, 50 insertions(+), 5 deletions(-) diff --git a/jit/trace-jit.c

[PATCH] vm: add some debugging info in case of itable failure

I needed this for debugging, but getting our hands on this information can be a bit subtle and tricky, so why not just leave it there for future reference. Signed-off-by: Vegard Nossum vegard.nos...@gmail.com --- vm/itable.c | 20 ++-- 1 files changed, 18 insertions(+), 2

[PATCH] vm: implement JNI function GetObjectField()

Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- vm/jni-interface.c | 19 ++- 1 files changed, 18 insertions(+), 1 deletions(-) diff --git a/vm/jni-interface.c b/vm/jni-interface.c index dfba479..eb8468e 100644 --- a/vm/jni-interface.c +++ b/vm/jni-interface.c @@ -524,6

[PATCH] jit: add live-in/use/def/live-out set dumping to liveness analysis tracing

This helped me resolve a bug, so I suppose it's useful in general. Signed-off-by: Vegard Nossum vegard.nos...@gmail.com --- jit/trace-jit.c | 53 +++-- 1 files changed, 47 insertions(+), 6 deletions(-) diff --git a/jit/trace-jit.c

[PATCH] jato: add -Xtrace:method for single-method tracing

This lets us trace the compilation of a single method without executing it. Note that the VM still needs to be initialized and some static class initializers will still be compiled (and traced) and executed. Example: jato -Xtrace:jit -Xtrace:method

[PATCH] x86: add OP_CMPG to insn-selector

This is a copy of OP_CMPL with a call to emulate_fcmpg instead of emulate_fcmpl. Signed-off-by: Vegard Nossum vegard.nos...@gmail.com --- arch/x86/insn-selector.brg | 20 1 files changed, 20 insertions(+), 0 deletions(-) diff --git a/arch/x86/insn-selector.brg

[PATCH] regression: add comparison operators to FloatArithmeticTest

Javac produces both fcmpg and fcmpl instructions for this, verified with javap. Signed-off-by: Vegard Nossum vegard.nos...@gmail.com --- regression/jvm/FloatArithmeticTest.java | 21 + 1 files changed, 21 insertions(+), 0 deletions(-) diff --git

[PATCH] vm: do not use \r character in backtrace printing.

It makes the output corrupted. We make show_exe_function() print to a buffer which is printed only when function succeeds. Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- arch/x86/backtrace.c | 21 +++-- include/vm/backtrace.h |4 +++- vm/signal.c|2 ++

[PATCH 2/3] test: fix warning in vm/stack-trace-stub.c

Signed-off-by: Tomek Grabiec tgrab...@gmail.com --- test/vm/stack-trace-stub.c |1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/test/vm/stack-trace-stub.c b/test/vm/stack-trace-stub.c index d6414b4..7278b4d 100644 --- a/test/vm/stack-trace-stub.c +++

[PATCH] vm: fix trace_flush()

We should not use strtok() to split string into lines because it swollows consequtive '\n' characters. We also do not flush the buffer tail not ended by new line character because we would print the thread name marker ([main]) in the middle of the line next time the buffer is flushed. Tracing