OSDN Git Service
Carl Shapiro [Mon, 1 Nov 2010 23:17:11 +0000 (16:17 -0700)]
Merge "HPROF is here to stay, make WITH_HPROF the default." into dalvik-dev
Carl Shapiro [Mon, 1 Nov 2010 22:43:18 +0000 (15:43 -0700)]
HPROF is here to stay, make WITH_HPROF the default.
Change-Id: Ic9947461daa21641fd9d34d52b6c7998716d8ab2
Dan Bornstein [Mon, 1 Nov 2010 19:58:08 +0000 (12:58 -0700)]
Always dump blocks in label order.
This removes some cases of spurious test failure. Also did some
minor whitespace and commenting cleanup while I was in the territory.
Change-Id: I6b8f6d0c340625c31bde78394e77abf429367af3
Carl Shapiro [Mon, 1 Nov 2010 00:54:04 +0000 (17:54 -0700)]
Merge "Check that the debug registry has been started before visting it." into dalvik-dev
Carl Shapiro [Mon, 1 Nov 2010 00:53:56 +0000 (17:53 -0700)]
Merge "Resume all threads after performing an HPROF dump." into dalvik-dev
Carl Shapiro [Sat, 30 Oct 2010 04:32:52 +0000 (21:32 -0700)]
Check that the debug registry has been started before visting it.
Change-Id: I54605ebcb9a07846bbefc61ea48ba9e148a202fb
Carl Shapiro [Sat, 30 Oct 2010 02:08:57 +0000 (19:08 -0700)]
Resume all threads after performing an HPROF dump.
Change-Id: I83df8359a5d57e6b7cd79d75acc979024e50effb
Dan Bornstein [Sat, 30 Oct 2010 01:35:06 +0000 (18:35 -0700)]
am
79544e58: am
464104bb: am
1fe34326: am
fbe51551: (-s ours) Remove a pointless test, which just tested a test support class.
* commit '
79544e584c5a21c2551051732f7daa0b85afac02':
Remove a pointless test, which just tested a test support class.
Dan Bornstein [Sat, 30 Oct 2010 01:33:04 +0000 (18:33 -0700)]
am
464104bb: am
1fe34326: am
fbe51551: (-s ours) Remove a pointless test, which just tested a test support class.
* commit '
464104bb1bc96949d7638aa79d16431e6ab8da78':
Remove a pointless test, which just tested a test support class.
Dan Bornstein [Sat, 30 Oct 2010 01:30:57 +0000 (18:30 -0700)]
am
1fe34326: am
fbe51551: (-s ours) Remove a pointless test, which just tested a test support class.
* commit '
1fe3432661f96bf57b1e3bd91c3c01bd7564e6ae':
Remove a pointless test, which just tested a test support class.
Dan Bornstein [Sat, 30 Oct 2010 01:29:27 +0000 (18:29 -0700)]
am
fbe51551: (-s ours) Remove a pointless test, which just tested a test support class.
* commit '
fbe515510323ee65ce594b8eda5547c1cbfb8ad9':
Remove a pointless test, which just tested a test support class.
Dan Bornstein [Sat, 30 Oct 2010 01:23:19 +0000 (18:23 -0700)]
Remove a pointless test, which just tested a test support class.
Change-Id: Ia6d47860ad71f013fef5631bcd434c2f6f1d346f
Dan Bornstein [Sat, 30 Oct 2010 00:05:43 +0000 (17:05 -0700)]
am
99bb6ef3: am
32d0157d: am
66b96b42: am
3144eec0: (-s ours) Replace the implementation of this class.
* commit '
99bb6ef3043d4628e6e0da012df9920cd31735ce':
Replace the implementation of this class.
Dan Bornstein [Sat, 30 Oct 2010 00:04:21 +0000 (17:04 -0700)]
am
32d0157d: am
66b96b42: am
3144eec0: (-s ours) Replace the implementation of this class.
* commit '
32d0157dfef33fa48d195d2a467dd511ec7ef347':
Replace the implementation of this class.
Dan Bornstein [Sat, 30 Oct 2010 00:02:39 +0000 (17:02 -0700)]
am
66b96b42: am
3144eec0: (-s ours) Replace the implementation of this class.
* commit '
66b96b42ba1d21027b71ca65821648925c661a78':
Replace the implementation of this class.
Dan Bornstein [Sat, 30 Oct 2010 00:00:32 +0000 (17:00 -0700)]
am
3144eec0: (-s ours) Replace the implementation of this class.
* commit '
3144eec027b6a0de24f024c286e4544b1aca683b':
Replace the implementation of this class.
Carl Shapiro [Fri, 29 Oct 2010 22:35:17 +0000 (15:35 -0700)]
Merge "Reclassify the pin table as a VM internal root. The native stack classification requires an associated thread but are not directly associated with a thread. If we encounter an object in the pin table during an hprof dump hat will fail." into dalvik-dev
Dan Bornstein [Fri, 29 Oct 2010 22:27:54 +0000 (15:27 -0700)]
Replace the implementation of this class.
Use the Apache Harmony version.
Change-Id: Ica3b4c69ccd1cd4c91325c52fad5a71695d58ba4
Carl Shapiro [Fri, 29 Oct 2010 22:14:28 +0000 (15:14 -0700)]
Reclassify the pin table as a VM internal root. The native stack
classification requires an associated thread but are not directly
associated with a thread. If we encounter an object in the pin table
during an hprof dump hat will fail.
Change-Id: I9e49f88cd632584b5818c702cf80f3f96ec5da85
Dan Bornstein [Fri, 29 Oct 2010 20:01:27 +0000 (13:01 -0700)]
A bunch of minor cleanups.
I'm reacquainting myself with this code, and figured I might as well
do some of that nit-picky code convention tweakage while I was there.
The only non-whitespace non-comment changes I made were to change the
scope of a couple of methods to be more appropriate.
Change-Id: I550978ed0f5b9ccec938faef7c00d09532832ee4
Dan Bornstein [Fri, 29 Oct 2010 03:11:53 +0000 (20:11 -0700)]
Fix expected output of dx test 086.
It experienced both innocuous block order shift and changes to
content due to the earlier constant value propagation fix.
Change-Id: If97d9cc82e763e9bbd7c3a8ba834f2d994eb694b
Dan Bornstein [Fri, 29 Oct 2010 03:11:19 +0000 (20:11 -0700)]
Merge "Fix the expected output of dx test 083." into dalvik-dev
Dan Bornstein [Fri, 29 Oct 2010 03:08:00 +0000 (20:08 -0700)]
Fix the expected output of dx test 083.
The recent fix to constant value propagation altered the expected output.
Change-Id: I57a2c1d60f2bc1060e116199d7924944ca1eef43
Carl Shapiro [Wed, 27 Oct 2010 04:07:41 +0000 (21:07 -0700)]
Separate HPROF from the GC.
In the beginning, the only way to traverse the roots and heap was to
piggyback off the garbage collector. As such, HPROF was implemented
by instrumenting the root- and object traversal routines to check a
mode flag and call into HPROF during a GC when the flag was set.
This change moves the HPROF calls out of the GC and into callbacks
invoked through the visitor. Notably, it allows HPROF dumps to be
computed at any point in time without invoking a GC and potentially
destroying evidence relating to the cause of an OOM.
Change-Id: I2b74c4f10f35af3ca33b7c0bbfe470a8b586ff66
Dan Bornstein [Fri, 29 Oct 2010 00:44:06 +0000 (17:44 -0700)]
Fix dx test 087.
This is another case of block rearrangement.
Change-Id: I402d0056f8b6d314cef2efcfea80337777b20a7d
Dan Bornstein [Fri, 29 Oct 2010 00:37:08 +0000 (17:37 -0700)]
Update the expected.txt for dx test 094.
As with a couple previous cases, it looks like the block order
has changed in the output, which isn't significant.
Change-Id: I4aef1da2252b3994875349570aca21043144eed4
Dan Bornstein [Thu, 28 Oct 2010 16:54:57 +0000 (09:54 -0700)]
Up the version number.
Change-Id: Ibf1e0471d1991c0b9ef7df847b427a3aead6fd4d
Dan Bornstein [Wed, 27 Oct 2010 23:30:00 +0000 (16:30 -0700)]
Fix propagation of constants.
In particular, when constructing a rop instruction for a move, if
the source has a known constant value, make the target end up also
having that value.
Change-Id: If172337889134e4a682bb0d8eb4e07ca284ac80e
jeffhao [Fri, 22 Oct 2010 00:44:43 +0000 (17:44 -0700)]
Added multi-threaded support to processing of class files in dx.
Change-Id: Ideb54c414073a9651b21ce0697e6444fa80f146d
Dan Bornstein [Wed, 27 Oct 2010 22:39:49 +0000 (15:39 -0700)]
Clean up dx test 085.
The jasmin source had a bug which dx didn't previously catch, which is
now fixed, and I also removed a superfluous <init> method. With that
cleaned up, the test now fails but in a different way, which I believe
is indicative of a real problem, where it looks like a known constant
value is not getting properly propagated.
Change-Id: I3d31045eecc6d71cff2bca98c948dc58371d35a1
Dan Bornstein [Wed, 27 Oct 2010 22:09:29 +0000 (15:09 -0700)]
Dx test 083's expected output was wrong.
Like many of these tests, the expected output may change from time to
time due to irrelevant (to the test) changes to unrelated parts of dx.
In this case, at some point the rop block dumper seems to have started
sorting its blocks by label.
Change-Id: Ic82cdf8c7f98713bb67e4267c5dc91f0836a859d
Dan Bornstein [Wed, 27 Oct 2010 21:35:07 +0000 (14:35 -0700)]
Merge "The failure of dx test 105 wasn't actually spurious. Fixed the code." into dalvik-dev
Andy McFadden [Wed, 27 Oct 2010 20:32:07 +0000 (13:32 -0700)]
am
cd507233: am
7f91cf6b: am
3cf48016: Set capabilities sooner.
Andy McFadden [Wed, 27 Oct 2010 20:28:22 +0000 (13:28 -0700)]
am
7f91cf6b: am
3cf48016: Set capabilities sooner.
Andy McFadden [Wed, 27 Oct 2010 20:25:34 +0000 (13:25 -0700)]
am
3cf48016: Set capabilities sooner.
Andy McFadden [Wed, 27 Oct 2010 18:25:13 +0000 (11:25 -0700)]
Set capabilities sooner.
Bug
3135433.
Change-Id: I22fde728bd5d65774f8fdf1fa45956fe18358c4c
buzbee [Wed, 27 Oct 2010 19:49:48 +0000 (12:49 -0700)]
Merge "x86: Improve ArrayIndexOutOfBoundsException detail messages" into dalvik-dev
Dan Bornstein [Wed, 27 Oct 2010 18:51:16 +0000 (11:51 -0700)]
The failure of dx test 105 wasn't actually spurious. Fixed the code.
In particular, for array load and store instructions, dx was being too
aggressive about letting the contents of the stack override the
implied type expected by the instructions. This led to cases where dx
would emit code instead of reporting an error.
The failure on the array load side of things implied that there also
needed to be a new regression test on the store side, since the
existing array store test didn't catch this. So, I added that too in
this change.
Finally, I took this opportunity to do minor whitespace-type cleanups
in a couple of related files that I opened during the course of the
investigation.
Change-Id: I72c644f66afb1108ae43a129ac81b010d072155a
buzbee [Wed, 27 Oct 2010 18:33:15 +0000 (11:33 -0700)]
x86: Improve ArrayIndexOutOfBoundsException detail messages
issue
3122084
Changes the x86 interpreter to use the new dvmThrowAIOOBE call
for out of bounds exception w/ more useful info.
Change-Id: Ic196cc99c561bf31953993a6928f574e5c5bf831
Elliott Hughes [Wed, 27 Oct 2010 18:24:59 +0000 (11:24 -0700)]
Merge "Remove nio cruft." into dalvik-dev
Elliott Hughes [Wed, 27 Oct 2010 17:11:12 +0000 (10:11 -0700)]
Remove nio cruft.
Change-Id: Ie0dd478f499e9fa40a4af8f10b81df7144a3e813
Dan Bornstein [Tue, 26 Oct 2010 23:23:07 +0000 (16:23 -0700)]
Fix a test.
The expected output was wrong. (A previous dx change made the output in
the case in question better.)
Change-Id: I94a180efc83968197d7a0a0fad49cdc6e02338cb
Andy McFadden [Tue, 26 Oct 2010 23:00:38 +0000 (16:00 -0700)]
Merge "Properly close <address> tag." into dalvik-dev
Andy McFadden [Tue, 26 Oct 2010 22:59:58 +0000 (15:59 -0700)]
Properly close <address> tag.
Change-Id: I4f86849de25f1d1a35f2356c1f118f3f063674a4
Jeff Hao [Tue, 26 Oct 2010 21:35:54 +0000 (14:35 -0700)]
Merge "Implement the new instruction formats." into dalvik-dev
Andy McFadden [Tue, 26 Oct 2010 19:36:50 +0000 (12:36 -0700)]
Merge "Improve dexdump temp file handling." into dalvik-dev
Dan Bornstein [Tue, 26 Oct 2010 19:00:07 +0000 (12:00 -0700)]
Implement the new instruction formats.
I also refactored some of the old code in the process, to minimize
duplication, and I also did a minor cleanup of the format spec.
Change-Id: Ib07ea4ade52aa6ec4fa69000798e0cfb8f10c3a9
Ben Cheng [Tue, 26 Oct 2010 17:20:43 +0000 (10:20 -0700)]
am
0b33af0c: am
5d992dfa: am
36bd1345: Fix for array lower-bound check for count-down loops.
Ben Cheng [Tue, 26 Oct 2010 17:18:38 +0000 (10:18 -0700)]
am
5d992dfa: am
36bd1345: Fix for array lower-bound check for count-down loops.
Ben Cheng [Tue, 26 Oct 2010 17:16:28 +0000 (10:16 -0700)]
am
36bd1345: Fix for array lower-bound check for count-down loops.
Andy McFadden [Tue, 26 Oct 2010 14:44:13 +0000 (07:44 -0700)]
Merge "Correct verifier merging of primitive arrays" into dalvik-dev
Ben Cheng [Tue, 26 Oct 2010 03:54:31 +0000 (20:54 -0700)]
Fix for array lower-bound check for count-down loops.
If the counted loop is composed as
for (int i = hi; i >= lo; i--) {
.. = array[i];
}
The hoisted lower-bound check should be asserting "lo >= 0".
On the other hand if the counted loop is composed as
for (int i = hi; i > lo; i--) {
.. = array[i];
}
The hoisted lower-bound check should be asserting "lo + 1 >= 0".
Previously these two checks are reversed.
Bug:
3130818
Change-Id: Ibc5a6daa837880e9986e45bbc29d1a5e548be3ae
Andy McFadden [Mon, 25 Oct 2010 23:13:46 +0000 (16:13 -0700)]
Correct verifier merging of primitive arrays
Merging of primitive array types (e.g. float[][] + String[]) was being
handled by the general-purpose type merge, which was determining the
result to be an Object. It should have been an Object[].
This changes the verifier to treat arrays of primitive as arrays, but
with one fewer dimension.
Bug:
3130586
Change-Id: I35a6c27256f69f142fa2a8c22bc5f995096c9ae1
Andy McFadden [Mon, 25 Oct 2010 20:29:43 +0000 (13:29 -0700)]
Improve dexdump temp file handling.
When dexdump is handed a zip or APK, it has to unpack classes.dex to a
temp file. It currently tries /tmp and /sdcard, neither of which are
likely to exist on a Windows system.
We now print a warning when those directories don't exist, and try to
create the temp file in the current directory.
Bug
3125957.
Change-Id: I7724ce48c35f22ed4b5116dfe5783243107df6a1
Jeff Hao [Mon, 25 Oct 2010 18:04:36 +0000 (11:04 -0700)]
Merge "A little instruction format housecleaning and prep work." into dalvik-dev
Dan Bornstein [Mon, 25 Oct 2010 01:30:58 +0000 (18:30 -0700)]
A little instruction format housecleaning and prep work.
Change-Id: Iea9d79d23515090f75cdc32c2123dbfb56e27d3c
Andy McFadden [Mon, 25 Oct 2010 14:56:01 +0000 (07:56 -0700)]
Merge "Allow nested instruction counting." into dalvik-dev
Andy McFadden [Mon, 25 Oct 2010 14:55:05 +0000 (07:55 -0700)]
Merge "Lock thread list, show join target" into dalvik-dev
Andy McFadden [Fri, 22 Oct 2010 22:26:07 +0000 (15:26 -0700)]
Allow nested instruction counting.
It is traditional to store counts in variables with more than one
bit of storage.
Also, fixed some of the decorations.
Bug
2863796.
Change-Id: I167d6803e955159e35e75d1e2924754145ea1361
Jeff Hao [Fri, 22 Oct 2010 21:14:58 +0000 (14:14 -0700)]
Merge "Don't force optimizations during class init in dexopt." into dalvik-dev
Andy McFadden [Fri, 22 Oct 2010 20:32:12 +0000 (13:32 -0700)]
Lock thread list, show join target
There were some paths to "thread list lock must be held" functions that
weren't locking the thread list. We now do trylock/unlock in dumpFrames
to ensure that doesn't happen.
Also, append "tid=N" to the "waiting on" line in stack trace output
when we're waiting on a VMThread object (i.e. we're in Thread.join()).
Bug
2827015 (for the thread list lock).
Change-Id: I05aa27dd440c671b22ef7cfe47e31dfe54b0ed2d
Ben Cheng [Fri, 22 Oct 2010 20:32:07 +0000 (13:32 -0700)]
resolved conflicts for merge of
dfc99234 to dalvik-dev
Change-Id: I0b5b19f9f6cd7aa0b3b51e0f9b87ce21c3175771
Andy McFadden [Fri, 22 Oct 2010 19:41:33 +0000 (12:41 -0700)]
Don't force optimizations during class init in dexopt.
Some "optimizations" are mandatory, notably those having to do with
access to volatile fields. We get into an odd situation in dexopt
where we have to initialize and execute <clinit> in a few classes
(primarily Class and Object), but if optimizations are disabled we
don't want to write optimized instructions into the output DEX.
This modifies the class init code to skip the mandatory optimizations
if we're in dexopt. Either the optimization will happen as a normal
part of dexopt operation, or it won't (in which case it'll be applied
when the class is first loaded in a normal VM).
Bug
3099161.
Change-Id: I11ee5affd5e3ac5a1e8583241acdf3c4de033a96
Ben Cheng [Fri, 22 Oct 2010 17:49:18 +0000 (10:49 -0700)]
Merge "Avoid conditional loads if WORKAROUND_CORTEX_A9_745320 is defined."
Ben Cheng [Fri, 22 Oct 2010 17:15:03 +0000 (10:15 -0700)]
Avoid conditional loads if WORKAROUND_CORTEX_A9_745320 is defined.
No noticeable performance impact by this change.
Bug:
3117632
Change-Id: I31c6adc6cb9999498bb456f1e87f6f04f33e4144
Elliott Hughes [Fri, 22 Oct 2010 00:44:25 +0000 (17:44 -0700)]
Remove junk from platform.S now armv4t is gone.
Change-Id: I30079aacc753c89cfc3a3f64bd900a0cc858d65f
Elliott Hughes [Thu, 21 Oct 2010 20:02:56 +0000 (13:02 -0700)]
Detail messages for ArrayIndexOutOfBoundsExceptions.
This adds ARM fast interpreter and JIT support. x86 is still missing.
Change-Id: Ide46fd9dcd06780193848f594ce7d1491d7f5a96
Elliott Hughes [Thu, 21 Oct 2010 23:07:33 +0000 (16:07 -0700)]
Merge "Remove armv4t support." into dalvik-dev
Elliott Hughes [Thu, 21 Oct 2010 22:44:11 +0000 (15:44 -0700)]
Remove armv4t support.
This was never finished, and hasn't been maintained. And the hardware's
totally obsolete.
Change-Id: Icbf78e507de3b8be8a8f995b8423cc3c678ab0d3
Carl Shapiro [Thu, 21 Oct 2010 22:43:55 +0000 (15:43 -0700)]
Merge "Improve the card table scan." into dalvik-dev
Carl Shapiro [Thu, 21 Oct 2010 22:35:19 +0000 (15:35 -0700)]
Improve the card table scan.
Instead of scanning the table on card at a time, leverage a few
observations to make the scan somewhat more efficient. First, fewer
than 1% of cards are dirtied during the concurrent mark. We can scan
the card table by words instead of inspecting cards individually.
Second, dirtied cards tend to appear in clusters and are often
adjacent to one another. Instead of starting scans at each dirty
card, we scan the objects on consecutive cards together by resuming
the scan at the lowest address visited.
I have not been able to get good profiling data from the VM so I am
not sure how much this helps. This code may need some refining once
we get our hands on applications with larger heaps.
Change-Id: I30bc6bb14bc268fe404bc1629e44247f1ef8a6f2
Dan Bornstein [Thu, 21 Oct 2010 21:55:56 +0000 (14:55 -0700)]
Minor clarifications and cleanup.
Change-Id: I571ca9ac8f7a258a6e5d19208f382f42ee0aab22
Dan Bornstein [Thu, 21 Oct 2010 21:04:43 +0000 (14:04 -0700)]
Merge "One more format and our first set of wide opcodes." into dalvik-dev
Patrick Scott [Thu, 21 Oct 2010 20:30:48 +0000 (13:30 -0700)]
am
28074c85: Merge "Fix debug statements."
Merge commit '
28074c8562f785bc6d7d785299bfb7eb6d111036' into dalvik-dev
* commit '
28074c8562f785bc6d7d785299bfb7eb6d111036':
Fix debug statements.
Dan Bornstein [Thu, 21 Oct 2010 00:32:39 +0000 (17:32 -0700)]
One more format and our first set of wide opcodes.
Change-Id: I16731c1b66b40022c2e10c93640583ae43659701
Patrick Scott [Thu, 21 Oct 2010 19:43:44 +0000 (12:43 -0700)]
Merge "Fix debug statements."
Andy McFadden [Thu, 21 Oct 2010 19:38:34 +0000 (12:38 -0700)]
Merge "Rebuild the x86-atom interpreter." into dalvik-dev
Carl Shapiro [Thu, 21 Oct 2010 19:10:28 +0000 (12:10 -0700)]
am
3a3f5d65: am
493d84f8: am
2e3ec716: Fix class loader interning following a String.intern.
Merge commit '
3a3f5d65f9bbaf5eb5885b4a72b9600dc0dbbe4e' into dalvik-dev
* commit '
3a3f5d65f9bbaf5eb5885b4a72b9600dc0dbbe4e':
Fix class loader interning following a String.intern.
Patrick Scott [Thu, 21 Oct 2010 18:50:15 +0000 (14:50 -0400)]
Fix debug statements.
Shows up without NDEBUG.
Change-Id: I1270266369beb783242283af60ee268118c8fb3d
Andy McFadden [Thu, 21 Oct 2010 18:27:22 +0000 (11:27 -0700)]
Rebuild the x86-atom interpreter.
Change-Id: I784e60325a82fd43420e489ee17e2013684c04d2
Elliott Hughes [Thu, 21 Oct 2010 18:13:17 +0000 (11:13 -0700)]
Merge "Better detail messages in ArrayIndexOutOfBoundExceptions." into dalvik-dev
Carl Shapiro [Thu, 21 Oct 2010 17:51:57 +0000 (10:51 -0700)]
am
493d84f8: am
2e3ec716: Fix class loader interning following a String.intern.
Merge commit '
493d84f888a01747967cbc84603358f525fa7e77'
* commit '
493d84f888a01747967cbc84603358f525fa7e77':
Fix class loader interning following a String.intern.
Elliott Hughes [Thu, 21 Oct 2010 00:14:57 +0000 (17:14 -0700)]
Better detail messages in ArrayIndexOutOfBoundExceptions.
The RI only includes the index. We've traditionally included nothing. This
patch fixes the portable interpreter to include both the index and the array
length.
Later patches will address the ARM- and x86-specific code.
Change-Id: I9d0e6baacced4e1d33e6cd75965017a38571af67
Dan Bornstein [Wed, 20 Oct 2010 23:36:08 +0000 (16:36 -0700)]
Merge "New instruction formats for expansion beyond 256 opcodes." into dalvik-dev
Dan Bornstein [Tue, 19 Oct 2010 23:50:25 +0000 (16:50 -0700)]
New instruction formats for expansion beyond 256 opcodes.
Change-Id: Id57cee56a9b49a2d6434ab2c41657d4a4cb9ed61
Carl Shapiro [Tue, 19 Oct 2010 00:26:46 +0000 (17:26 -0700)]
Improve the error report generated by the card table verifier.
With this change, when a gray object has been found on an unmarked
card all references are sent to the log stream. This will usually
provide enough context to diagnose issues with the write barrier.
Change-Id: Ib6b7d98160e18393f54fc6ebc38c9ca9ca7dad2e
Carl Shapiro [Wed, 20 Oct 2010 04:37:28 +0000 (21:37 -0700)]
am
2e3ec716: Fix class loader interning following a String.intern.
Merge commit '
2e3ec7160b265d520cd755f8c20163410d822373' into gingerbread-plus-aosp
* commit '
2e3ec7160b265d520cd755f8c20163410d822373':
Fix class loader interning following a String.intern.
Carl Shapiro [Wed, 20 Oct 2010 00:27:13 +0000 (17:27 -0700)]
Fix class loader interning following a String.intern.
Strings can be intered by the class loader, in the case of string
literals, or by the user, through String.intern. Literal strings
exist for the life of their referencing classes. User strings are
weak and may be garbage collected when unreferenced. These two
classes of strings are kept in seprate tables for the conveniance of
the garbage collector during root traversal.
When a class loader interns a string that was already interned by the
user the runtime must move the string from the intern table to the
literal table to increase the reference strength. Previously, this
was implemented by inserting the incoming string into the literal
table and removing any matching strings in the intern table. This
transition lost pointer equality. With this change we first insert
the exact string from the intern table into the literal table and
secondly remove its reference from the intern table. By moving the
string between tables pointer equality is preserved.
At this point lookupInternedString should be split into two functions,
possibly by pulled up the relevant bits into the public interface
functions. Since this change will be merged to gingerbread I will
leave the clean up to a separate change.
Bug:
3098960
Carl Shapiro [Wed, 20 Oct 2010 00:27:13 +0000 (17:27 -0700)]
Fix class loader interning following a String.intern.
Strings can be intered by the class loader, in the case of string
literals, or by the user, through String.intern. Literal strings
exist for the life of their referencing classes. User strings are
weak and may be garbage collected when unreferenced. These two
classes of strings are kept in seprate tables for the conveniance of
the garbage collector during root traversal.
When a class loader interns a string that was already interned by the
user the runtime must move the string from the intern table to the
literal table to increase the reference strength. Previously, this
was implemented by inserting the incoming string into the literal
table and removing any matching strings in the intern table. This
transition lost pointer equality. With this change we first insert
the exact string from the intern table into the literal table and
secondly remove its reference from the intern table. By moving the
string between tables pointer equality is preserved.
At this point lookupInternedString should be split into two functions,
possibly by pulled up the relevant bits into the public interface
functions. Since this change will be merged to gingerbread I will
leave the clean up to a separate change.
Bug:
3098960
Change-Id: I11d24d0afa9b7ca2e53fefb1b5e6364795e14d44
Andy McFadden [Tue, 19 Oct 2010 20:45:31 +0000 (13:45 -0700)]
am
d2ba02c2: (-s ours) am
da34e27a: (-s ours) am
3db6149c: Merge "Update heap dump notes. DO NOT MERGE." into gingerbread
Merge commit '
d2ba02c297bc8214611e347e27edee9f41772474' into dalvik-dev
* commit '
d2ba02c297bc8214611e347e27edee9f41772474':
Update heap dump notes. DO NOT MERGE.
Andy McFadden [Tue, 19 Oct 2010 20:37:01 +0000 (13:37 -0700)]
am
da34e27a: (-s ours) am
3db6149c: Merge "Update heap dump notes. DO NOT MERGE." into gingerbread
Merge commit '
da34e27a6de6eec8b4f1036c9ba4922a12a5bf19'
* commit '
da34e27a6de6eec8b4f1036c9ba4922a12a5bf19':
Update heap dump notes. DO NOT MERGE.
Andy McFadden [Tue, 19 Oct 2010 19:02:25 +0000 (12:02 -0700)]
Merge "Update heap dump notes." into dalvik-dev
Carl Shapiro [Tue, 19 Oct 2010 18:34:00 +0000 (11:34 -0700)]
am
6b08a477: am
e8debfcf: am
c6fb7a5e: Merge "Exit early if the we encounter a null reference." into gingerbread
Merge commit '
6b08a4774e13844802cc324320681ba26479899e' into dalvik-dev
* commit '
6b08a4774e13844802cc324320681ba26479899e':
Exit early if the we encounter a null reference.
Andy McFadden [Tue, 19 Oct 2010 18:28:36 +0000 (11:28 -0700)]
am
3db6149c: Merge "Update heap dump notes. DO NOT MERGE." into gingerbread
Merge commit '
3db6149c9891089ed1b357ab69db5c57c9c1fb6c' into gingerbread-plus-aosp
* commit '
3db6149c9891089ed1b357ab69db5c57c9c1fb6c':
Update heap dump notes. DO NOT MERGE.
Andy McFadden [Tue, 19 Oct 2010 18:25:02 +0000 (11:25 -0700)]
Update heap dump notes.
The release order changed slightly.
Bug
3110872 is related.
Change-Id: I2b2982a29910ee58746188e9dbd61ffd9c3b3d64
Andy McFadden [Tue, 19 Oct 2010 18:19:28 +0000 (11:19 -0700)]
Merge "Update heap dump notes. DO NOT MERGE." into gingerbread
Carl Shapiro [Tue, 19 Oct 2010 18:18:23 +0000 (11:18 -0700)]
am
e8debfcf: am
c6fb7a5e: Merge "Exit early if the we encounter a null reference." into gingerbread
Merge commit '
e8debfcf2d16ba18474ca51200e9b7f2d912dee3'
* commit '
e8debfcf2d16ba18474ca51200e9b7f2d912dee3':
Exit early if the we encounter a null reference.
Andy McFadden [Tue, 19 Oct 2010 17:56:38 +0000 (10:56 -0700)]
am
078241d5: am
8ca802ff: am
7b4c1d52: Pop call frame before boxing result
Merge commit '
078241d5d7b44f145b93979665cdc6e5df2b093b' into dalvik-dev
* commit '
078241d5d7b44f145b93979665cdc6e5df2b093b':
Pop call frame before boxing result
buzbee [Tue, 19 Oct 2010 17:56:34 +0000 (10:56 -0700)]
am
1e09b407: am
8cb3d855: am
e09a766f: x86 mterp: fix cherry-pick/merge, issue
3095771
Merge commit '
1e09b40718a6dc9db0ce34a7aacc93a3df45c664' into dalvik-dev
* commit '
1e09b40718a6dc9db0ce34a7aacc93a3df45c664':
x86 mterp: fix cherry-pick/merge, issue
3095771
Andy McFadden [Tue, 19 Oct 2010 17:44:34 +0000 (10:44 -0700)]
am
8ca802ff: am
7b4c1d52: Pop call frame before boxing result
Merge commit '
8ca802ff96060b730119b58379c2e6366afdbd3e'
* commit '
8ca802ff96060b730119b58379c2e6366afdbd3e':
Pop call frame before boxing result