changeset in /hg/icedtea6: 2009-05-08 Gary Benson <gbenson at red...
Gary Benson
gbenson at redhat.com
Fri May 8 07:49:06 PDT 2009
changeset b9c3b68cc5d7 in /hg/icedtea6
details: http://icedtea.classpath.org/hg/icedtea6?cmd=changeset;node=b9c3b68cc5d7
description:
2009-05-08 Gary Benson <gbenson at redhat.com>
* Makefile.am: Always apply Zero and Shark patches.
* patches/icedtea-test-atomic-operations.patch: Only test on Zero.
* patches/icedtea-eclipse-hotspot-6614100-7b24.patch: Removed.
* HACKING: Updated.
diffstat:
5 files changed, 25 insertions(+), 121 deletions(-)
ChangeLog | 7 +
HACKING | 25 ++----
Makefile.am | 31 +-------
patches/icedtea-eclipse-hotspot-6614100-7b24.patch | 77 --------------------
patches/icedtea-test-atomic-operations.patch | 6 -
diffs (232 lines):
diff -r bce6f516052d -r b9c3b68cc5d7 ChangeLog
--- a/ChangeLog Fri May 08 14:47:09 2009 +0100
+++ b/ChangeLog Fri May 08 15:49:54 2009 +0100
@@ -1,3 +1,10 @@ 2009-05-08 Andrew Haley <aph at redhat.co
+2009-05-08 Gary Benson <gbenson at redhat.com>
+
+ * Makefile.am: Always apply Zero and Shark patches.
+ * patches/icedtea-test-atomic-operations.patch: Only test on Zero.
+ * patches/icedtea-eclipse-hotspot-6614100-7b24.patch: Removed.
+ * HACKING: Updated.
+
2009-05-08 Andrew Haley <aph at redhat.com>
* patches/icedtea-lcms-2.patch: Work around lcms breakage. See
diff -r bce6f516052d -r b9c3b68cc5d7 HACKING
--- a/HACKING Fri May 08 14:47:09 2009 +0100
+++ b/HACKING Fri May 08 15:49:54 2009 +0100
@@ -97,6 +97,12 @@ The following patches are currently appl
* icedtea-java2d-stroker-internal-joint.patch: Fix penultimate joint created by GeneralPath.closePath().
* icedtea-java2d-stroker-internal-close-joint.patch: Fix final joint created by GeneralPath.closePath().
* icedtea-cacao.patch: For the 'java' command, create new thread depending on the current VM.
+* icedtea-bytecodeInterpreter.patch: Replace fast opcodes with opc_default.
+* icedtea-ia64-bugfix.patch: Remove workaround for IA64 GCC bug.
+* icedtea-signature-iterator.patch: Add zero-specific signature handling.
+* icedtea-test-atomic-operations.patch: Add check to see if GCC's sync_lock_test works.
+* icedtea-zero.patch: Generalise architecture support, add ARM and #ifdef out non-zero applicable parts.
+* icedtea-shark.patch: Add support for the Shark JIT.
The following patches are only applied to OpenJDK6 in IcedTea6:
@@ -108,6 +114,7 @@ The following patches are only applied t
* icedtea-timezone.patch : Makes java only look for time zone information in /etc/sysconfig/clock if /etc/localtime is not found (fix
for rh-489586)
* icedtea-dnd-filelists.patch: Fix drag and drop behaviour when dragging a file list between JVMs (S5079469). Backported from OpenJDK.
+* icedtea-signed-types-hot6.patch: Make use of unsigned/signed types explicit.
The following patches are only applied to OpenJDK in IcedTea:
@@ -122,23 +129,6 @@ The following patches are only applied t
calls, replace hexadecimal floating point literals with decimal variants in
java.lang.Double and java.lang.Float.
* icedtea-ecj-pr261.patch: Adds a couple of classes that are omitted from rt.jar. (PR261)
-
-The following patches are only applied for IcedTea builds using the zero-assembler:
-
-* icedtea-bytecodeInterpreter.patch: Replace fast opcodes with opc_default.
-* icedtea-eclipse-hotspot-6614100-7b24.patch: Fix Eclipse crash (S6614100). Fixed in OpenJDK7 b29/hs13.
-* icedtea-ia64-bugfix.patch: Remove workaround for IA64 GCC bug.
-* icedtea-signature-iterator.patch: Add zero-specific signature handling.
-* icedtea-test-atomic-operations.patch: Add check to see if GCC's sync_lock_test works.
-* icedtea-zero.patch: Generalise architecture support, add ARM and #ifdef out non-zero applicable parts.
-
-The following patches are only applied for IcedTea builds not using the zero-assembler:
-
-* icedtea-signed-types-hot6.patch: Make use of unsigned/signed types explicit.
-
-The following patches are only applied for IcedTea builds using the Shark JIT:
-
-* icedtea-shark.patch: Add support for the Shark JIT.
The following patches are only applied when building with experimental NetX plugin support:
@@ -173,3 +163,4 @@ changes to the OpenJDK code base and/or
* icedtea-hotspot-citypeflow.patch: Not required with HS14.
* icedtea-core-zero-build.patch: Not required with HS14.
* icedtea-bytecodeInterpreterWithChecks.patch: Same as icedtea-xslfix.patch (OpenJDK6 only, S6707485); not required with HS14.
+* icedtea-eclipse-hotspot-6614100-7b24.patch: Fix Eclipse crash (S6614100). Fixed in OpenJDK7 b29/hs13. Not required with HS14.
diff -r bce6f516052d -r b9c3b68cc5d7 Makefile.am
--- a/Makefile.am Fri May 08 14:47:09 2009 +0100
+++ b/Makefile.am Fri May 08 15:49:54 2009 +0100
@@ -459,21 +459,6 @@ else
patches/hotspot/$(HSBUILD)/icedtea-gcc-4.3.patch
endif
-# Zero patches
-ZERO_PATCHES = \
- patches/hotspot/$(HSBUILD)/icedtea-bytecodeInterpreter.patch \
- patches/icedtea-signature-iterator.patch \
- patches/icedtea-test-atomic-operations.patch \
- patches/icedtea-zero.patch \
- patches/icedtea-ia64-bugfix.patch \
- patches/hotspot/$(HSBUILD)/icedtea-parisc-opt.patch
-
-if ZERO_BUILD
- ZERO_PATCHES_COND = $(ZERO_PATCHES)
-else
- ZERO_PATCHES_COND =
-endif
-
# Patches needed when not using the newer OpenJDK 7 HotSpot
OLD_HOTSPOT_PATCHES = \
patches/icedtea-signed-types-hot6.patch \
@@ -487,17 +472,15 @@ else
OLD_HOTSPOT_PATCHES_COND = $(OLD_HOTSPOT_PATCHES)
endif
-# If Shark is required, apply shark patch
-if SHARK_BUILD
- SHARK_PATCH = patches/hotspot/$(HSBUILD)/icedtea-shark.patch
-else
- SHARK_PATCH =
-endif
-
ICEDTEA_FSG_PATCHES =
ICEDTEA_PATCHES = \
- $(ZERO_PATCHES_COND) \
+ patches/hotspot/$(HSBUILD)/icedtea-bytecodeInterpreter.patch \
+ patches/icedtea-signature-iterator.patch \
+ patches/icedtea-test-atomic-operations.patch \
+ patches/icedtea-zero.patch \
+ patches/icedtea-ia64-bugfix.patch \
+ patches/hotspot/$(HSBUILD)/icedtea-parisc-opt.patch \
$(OLD_HOTSPOT_PATCHES_COND) \
patches/icedtea-version.patch \
patches/hotspot/$(HSBUILD)/icedtea-version.patch \
@@ -557,7 +540,7 @@ ICEDTEA_PATCHES = \
patches/hotspot/$(HSBUILD)/icedtea-use-idx_t.patch \
patches/hotspot/$(HSBUILD)/icedtea-params-cast-size_t.patch \
patches/icedtea-clean-crypto.patch \
- $(SHARK_PATCH) \
+ patches/hotspot/$(HSBUILD)/icedtea-shark.patch \
$(GCC_PATCH) \
patches/icedtea-arch.patch \
patches/icedtea-lc_ctype.patch \
diff -r bce6f516052d -r b9c3b68cc5d7 patches/icedtea-eclipse-hotspot-6614100-7b24.patch
--- a/patches/icedtea-eclipse-hotspot-6614100-7b24.patch Fri May 08 14:47:09 2009 +0100
+++ /dev/null Thu Jan 01 00:00:00 1970 +0000
@@ -1,77 +0,0 @@
-
-# HG changeset patch
-# User never
-# Date 1212727493 25200
-# Node ID 65fe2bd8883932456557957163bf5c4e4af79469
-# Parent d4dbd9f91680840f6c719b0d23942b4d4eb00837
-6614100: EXCEPTION_ACCESS_VIOLATION while running Eclipse with 1.6.0_05-ea
-Reviewed-by: kvn, jrose, rasbold
-
---- openjdk/hotspot/src/share/vm/opto/cfgnode.cpp Thu Jun 05 15:43:18 2008 -0700
-+++ openjdk/hotspot/src/share/vm/opto/cfgnode.cpp Thu Jun 05 21:44:53 2008 -0700
-@@ -1621,64 +1621,6 @@ Node *PhiNode::Ideal(PhaseGVN *phase, bo
- if (opt != NULL) return opt;
- }
-
-- if (in(1) != NULL && in(1)->Opcode() == Op_AddP && can_reshape) {
-- // Try to undo Phi of AddP:
-- // (Phi (AddP base base y) (AddP base2 base2 y))
-- // becomes:
-- // newbase := (Phi base base2)
-- // (AddP newbase newbase y)
-- //
-- // This occurs as a result of unsuccessful split_thru_phi and
-- // interferes with taking advantage of addressing modes. See the
-- // clone_shift_expressions code in matcher.cpp
-- Node* addp = in(1);
-- const Type* type = addp->in(AddPNode::Base)->bottom_type();
-- Node* y = addp->in(AddPNode::Offset);
-- if (y != NULL && addp->in(AddPNode::Base) == addp->in(AddPNode::Address)) {
-- // make sure that all the inputs are similar to the first one,
-- // i.e. AddP with base == address and same offset as first AddP
-- bool doit = true;
-- for (uint i = 2; i < req(); i++) {
-- if (in(i) == NULL ||
-- in(i)->Opcode() != Op_AddP ||
-- in(i)->in(AddPNode::Base) != in(i)->in(AddPNode::Address) ||
-- in(i)->in(AddPNode::Offset) != y) {
-- doit = false;
-- break;
-- }
-- // Accumulate type for resulting Phi
-- type = type->meet(in(i)->in(AddPNode::Base)->bottom_type());
-- }
-- Node* base = NULL;
-- if (doit) {
-- // Check for neighboring AddP nodes in a tree.
-- // If they have a base, use that it.
-- for (DUIterator_Fast kmax, k = this->fast_outs(kmax); k < kmax; k++) {
-- Node* u = this->fast_out(k);
-- if (u->is_AddP()) {
-- Node* base2 = u->in(AddPNode::Base);
-- if (base2 != NULL && !base2->is_top()) {
-- if (base == NULL)
-- base = base2;
-- else if (base != base2)
-- { doit = false; break; }
-- }
-- }
-- }
-- }
-- if (doit) {
-- if (base == NULL) {
-- base = new (phase->C, in(0)->req()) PhiNode(in(0), type, NULL);
-- for (uint i = 1; i < req(); i++) {
-- base->init_req(i, in(i)->in(AddPNode::Base));
-- }
-- phase->is_IterGVN()->register_new_node_with_optimizer(base);
-- }
-- return new (phase->C, 4) AddPNode(base, base, y);
-- }
-- }
-- }
--
- // Split phis through memory merges, so that the memory merges will go away.
- // Piggy-back this transformation on the search for a unique input....
- // It will be as if the merged memory is the unique value of the phi.
-
diff -r bce6f516052d -r b9c3b68cc5d7 patches/icedtea-test-atomic-operations.patch
--- a/patches/icedtea-test-atomic-operations.patch Fri May 08 14:47:09 2009 +0100
+++ b/patches/icedtea-test-atomic-operations.patch Fri May 08 15:49:54 2009 +0100
@@ -5,11 +5,11 @@ diff -r d384f5a5bd0c hotspot/src/share/v
jint result = JNI_ERR;
DT_RETURN_MARK(CreateJavaVM, jint, (const jint&)result);
-+ // We're about to use Atomic::xchg for locking. On zero platforms
++ // We're about to use Atomic::xchg for locking. On Zero platforms
+ // this is implemented using the GCC builtin __sync_lock_test_and_set
+ // which is not guaranteed to do what we're using it for on all
+ // platforms. So we check it works before doing anything else.
-+#ifdef ASSERT
++#if defined(ZERO) && defined(ASSERT)
+ {
+ jint a = 0xCAFEBABE;
+ jint b = Atomic::xchg(0xC0FFEE, &a);
@@ -18,7 +18,7 @@ diff -r d384f5a5bd0c hotspot/src/share/v
+ assert(a == 0xC0FFEE && b == (jint) 0xCAFEBABE, "Atomic::xchg() works");
+ assert(c == &b && d == &a, "Atomic::xchg_ptr() works");
+ }
-+#endif // ASSERT
++#endif // ZERO && ASSERT
+
// At the moment it's only possible to have one Java VM,
// since some of the runtime state is in global variables.
More information about the distro-pkg-dev
mailing list