Community
Participate
Working Groups
This bug is used to track the attach, re-attach problem (as detailed in bug 194081) that is specific to IBM JREs.
Created attachment 141080 [details] Patch to solve reattach crash with PerformanceExample on IBM JVM The test case PerformanceExample will crash with IBM JVM in reattach scenario. This problem does not happen on Sun JVM. It is related with both JVM and TPTP profilers code. In a TPTP class CThreadSet, a private list implementation is used. But when delete element from list, the delete function assert next pointer of list elements must be not NULL but does not check NULL condition in TPTP release version. Different Thread Id allocation mechanism in JVM implementation may cause this NULL pointer access. Dear committers, please help to review the patch code. After this patch is applied in TPTP 4.5.2.1, it should be ported in TPTP 4.6.1 too. Thanks, Chengrui
Created attachment 141081 [details] Profilers binary files on WinXP for testing usuage This is the profilers binary files for testing usuage.
(In reply to comment #2) > Created an attachment (id=141081) [details] > Profilers binary files on WinXP for testing usuage > > This is the profilers binary files for testing usuage. > The patch does not crash the IBM JRE anymore, but there is no data collected for memory analysis while data collection works fine with execution analysis.
(In reply to comment #3) > (In reply to comment #2) > > Created an attachment (id=141081) [details] [details] > > Profilers binary files on WinXP for testing usuage > > > > This is the profilers binary files for testing usuage. > > > > The patch does not crash the IBM JRE anymore, but there is no data collected > for memory analysis while data collection works fine with execution analysis. > Hi, Eugene, The memory analysis problem exists in previous TPTP too. The patch just add NULL pointer conditon to avoid crash problem. I think memory analysis problem should be filed with another new bug. Since crash problem with IBM JRE has been solved, this bug should be fixed. Thanks, Chengrui
(In reply to comment #4) > (In reply to comment #3) > > (In reply to comment #2) > > > Created an attachment (id=141081) [details] [details] [details] > > > Profilers binary files on WinXP for testing usuage > > > > > > This is the profilers binary files for testing usuage. > > > > > > > The patch does not crash the IBM JRE anymore, but there is no data collected > > for memory analysis while data collection works fine with execution analysis. > > > Hi, Eugene, > The memory analysis problem exists in previous TPTP too. The patch just add > NULL pointer conditon to avoid crash problem. I think memory analysis problem > should be filed with another new bug. Since crash problem with IBM JRE has been > solved, this bug should be fixed. > > Thanks, > Chengrui > Ok. I have opened bug 282952 for the memory problem. This patch is good for the original reattach problem. Please submit PMC approval request for inclusion in TPTP 4521.
Requesting approval to include the following defect in TPTP 4.5.2.1 branch: With IBM JRE, after re-attach, no more data is collected for methods already entered. (https://bugs.eclipse.org/bugs/show_bug.cgi?id=282531) 1. Explain why you believe this is a stop-ship defect. How does the defect manifest itself, and how will users of TPTP / consuming products be affected if the defect is not fixed? This reattach problem is related with bug 194081. After shadow stack solution patch applied, it is found IBM JRE will crash when reattach profiling with test case PerformanceExample. The problem is related both with TPTP profilers NULL pointer check code and JRE implementation. 2. Is there a work-around? If so, why do you believe the work-around is insufficient? No. 3. Is this a regression or API breakage? Explain. No. 4. Does this require new API? No. 5. Who performed the code review? Eugene. 6. Is there a test case attached to the bugzilla record? The test case PerformanceExample is used. 7. What is the nature of the fix? What is the scope of the fix? What is the risk associated with this fix? This fix just adds NULL pointer check conditions in class CThreadSet methods. Risk is very low. 8. Is this fix related to any standards that TPTP adheres to? If so, who has validated that the fix continues to adhere to the standard? No.
Requesting PMC approval for TPTP 4.5.2.1 Milestone 3.
Jonathan, Please check in the patch on behalf of Chengrui. Thanks!
Patch checked into 4.5.2.1 w/ PMC approval.
Verified in consuming product.
*** Bug 325539 has been marked as a duplicate of this bug. ***
