Opened 13 years ago
Last modified 10 years ago
#65 reopened defect
Java traps System (000e)
Reported by: | Batchheizer | Owned by: | |
---|---|---|---|
Priority: | Feedback Pending | Milestone: | |
Component: | general | Version: | 1.6.0-b19 Beta 2 |
Severity: | low | Keywords: | trap 000e ring 0 crash |
Cc: | lgrosenthal@… |
Description
If I used big apps like FF, OOo or TB before I start a Java-App Java traps (sometimes) the system. Only Power off/on is possible.
I hope the attached foto helps. If not tell me, what I should do for fixing the problem.
Attachments (5)
Change History (19)
Changed 13 years ago by
Attachment: | TrapJava000e.jpg added |
---|
comment:1 Changed 13 years ago by
Next trap after reboot:
Opened FF (3.6.13)
Opened Commonist
Did some screenshots
Created ticket here
Clicked on "view ticket"
Trap!
comment:2 Changed 13 years ago by
Trap 000E is PAGE FAULT (the page being referenced is not present in memory or the procedure referencing the page does not have enough privilege to access the page. Which may be a software error or sometimes a hardware error, e.g. damaged memory chips).
The value of CS is 0168. Which (according to this nice guide) means the trap happened in the ring 0 code, i.e. in some driver. This way, I doubt that your problem is directly related to Java -- most likely, it just creates a condition for it to pop up. It may be your network card driver, for instance. Or some other system component.
Which OS/2 version do you use?
P.S. I've been seeing TRAP 000e in Firefox very often when I run eCS 2.0 in VBox on Linux (i.e. no Java involved at all).
comment:3 Changed 13 years ago by
Thanks for the fast response.
I use eCS SilverRelease?. The problem only appears in conection with the new JAVA.
comment:4 Changed 13 years ago by
Are you sure it's the new Java? Can you use Innotek Java 1.4.2 in the same conditions and see what you get then?
On my real OS/2 machine, I use the new Java a lot (and sometimes I do use FF, though not very often) and I see no such traps at all, not a single one.
What is your ACPI.PSD line in CONFIG.SYS?
comment:5 Changed 13 years ago by
No ACPI. I keep you up to date if it appears again. But the problem only appeared with the new Java. I use small Java-Apps every day.
comment:6 Changed 13 years ago by
Priority: | major → Feedback Pending |
---|---|
Severity: | → low |
Please try RC2 and report the results.
comment:7 Changed 13 years ago by
JFYI. When trying to force a dump of a hung Java/Odin? process (see http://svn.netlabs.org/odin32/ticket/28#comment:3 for details), I get just the same trap 000E with the same CS contents and the same location value. Don't know so far how useful this information is for either of the problems.
comment:8 Changed 13 years ago by
With the latest builds of Java I did not hava one trap in about 3 months - so ticket could be closed.
comment:9 Changed 13 years ago by
Resolution: | → fixed |
---|---|
Status: | new → closed |
Thank you for the feedback.
Changed 11 years ago by
Attachment: | TRAP_e_starting_Moneydance_while_SeaMonkey_running_2013-06-03.jpg added |
---|
Trap screen when opening Moneydance after SeaMonkey? has been up for some time.
comment:10 Changed 11 years ago by
Cc: | lgrosenthal@… added |
---|---|
Keywords: | trap 000e ring 0 crash added |
Resolution: | fixed |
Status: | closed → reopened |
I am seeing this same issue (ring 0) when opening Moneydance 2011 after SeaMonkey? 2.7.2 has been up and running for some time. If I remember to close SeaMonkey?, the trap never occurs.
Note photo of trap screen. I believe this behavior started with GA4 (currently running GA5), and I saw this on my T43 (no ACPI) and now I see it on my T43p (no ACPI) (different RAM, CPU, NIC, WLAN, but same HDD). It could be a driver (GenMAC), but I can't be certain.
I can set up for a dump, I suppose. Any particular options I should use? I'm guessing you'd want a paddr(all).
comment:11 Changed 11 years ago by
This trap looks familiar. Seems to have something to do with shared memory I suppose. This needs to be properly investigated one day. And the dump is useful of course. Steven Levine is best when it comes to kernel traps, I will ask him to look at that.
Can you check more precisely what OpenJDK release did it start happening with? If you can prove that it doesn't happen with e.g. GA3 that could definitely help us.
comment:12 Changed 11 years ago by
Thanks for the quick follow-up, Dmitriy.
I've done the following to prepare for testing this today (see also, my following up in ticket # 216):
Verified Odin version:
[c:\]yum info libodin Installed Packages Name : libodin Arch : i386 Version : 0.8.9 Release : 1 Size : 3.0 M Repo : installed Summary : Odin runtime URL : http://svn.netlabs.org/odin32/ License : Project Odin Software License Description: Odin is a set of libraries that provide a Win32-compatible runtime : environment. This environment is necesasry to run Win32 : applications on OS/2 as well as native OS/2 applications built from : Win32 sources using Odin SDK. : : This package contains core Odin runtime libraries.
Verified current OpenJDK:
[c:\java160\bin]java -version openjdk version "1.6.0_27" OpenJDK Runtime Environment (build 1.6.0_27-b01-GA5) OpenJDK Client VM (build 20.0-b12, mixed mode)
Renamed c:\Java160 openjdk6_b27_jre_os2_ga5-20130217
Unzipped openjdk6_b25_jre_os2_ga3-20120721.zip
Renamed c:\openjdk6_b25_jre_os2_ga3-20120721.zip Java160
[c:\java160\bin]java -version openjdk version "1.6.0_25" OpenJDK Runtime Environment (build 1.6.0_25-b01-GA3) OpenJDK Client VM (build 20.0-b12, mixed mode)
If I get another trap today, I'll configure for a dump (this system isn't quite set up for it - I can do process dumps, but I don't have a dump partition, so I'll need to swap in another drive for that; luckily, the machine is docked, so I can mount a spare unit in the dock).
I'll follow up when I have more.
Changed 11 years ago by
Attachment: | 2013-06-04_SeaMonkey-memory-use-after-5.5-hours-uptime-normal-browsing-up-to-20-tabs-open-6-currently.pdf added |
---|
Print-to-file from SeaMonkey? after 5.5 hours uptime, up to 20 tabs open, 6 curently, prior to attempting to start Moneydance. Note corruption in pdf.
Changed 11 years ago by
Attachment: | TRAP_e_starting_Moneydance_while_SeaMonkey_running_2013-06-04.jpg added |
---|
Trap screen when opening Moneydance after SeaMonkey? has been up for 5.5 hours (GA3 on Odin 0.8.9).
Changed 11 years ago by
Attachment: | 2013-06-04_SeaMonkey-memory-use-startup-6-tabs-restored.pdf added |
---|
Print-to-file from SeaMonkey? after restart of system, 6 tabs restored. Note improved pdf output.
comment:13 Changed 11 years ago by
One more round of this before I configure for a trap dump. I should have saved the about:memory screens to html and not printed to pdf; I hadn't expected the corruption in the first one. However, it does point to something being wrong before the JVM starts. Note that IcedTea? has been completely removed from the configuration. Of the running tabs, no Flash content was evident, though some Flash had been playing several hours earlier.
comment:14 Changed 10 years ago by
Commenting here as a reminder to myself that I need to provide a dump file. In general I have not seen an actual trap in some time, though the system will become non-responsive as previously described, and upon shutting down, I am presented with a message from Xshutdown that the INI files could not be written. It is sometimes possible to just C-A-D, depending upon how much data had previously been written to the INI files prior to shutdown. Most of the time, however, it is necessary to restore an archive.
Large changes to the desktop (add a half dozen icons and a folder for a new application, create shadows of them, and drag the folder to the Progams folder) will accelerate the problem, and the system usually needs to be restarted immediately following (again, this is after some period of uptime prior to a crash, and with SeaMonkey? 2.7.2 ESR12 open with a number of tabs (Iced Tea is not loaded) and a JavaRDP session going for some time.
Foto of screen