| Summary: | JOGL crash with ATI mobility Radeon HD 5470 | ||
|---|---|---|---|
| Product: | [JogAmp] Jogl | Reporter: | guillaume.lurenbaum |
| Component: | windows | Assignee: | Sven Gothel <sgothel> |
| Status: | RESOLVED DUPLICATE | ||
| Severity: | normal | CC: | gouessej, wwalker3 |
| Priority: | P3 | ||
| Version: | 2 | ||
| Hardware: | pc_x86_64 | ||
| OS: | windows | ||
| Type: | --- | SCM Refs: | |
| Workaround: | --- | ||
| Attachments: |
JOGL 1.1.1a test log
Zip of JOGL 2 test log (this log is abnormally long -- check reason) HotSpot error log for JOGL 2 test |
||
|
Description
guillaume.lurenbaum
2011-10-14 09:57:02 CEST
I'll take a look once I'm done with bug 513 (which looks similar). I'll need to email some test cases to you to run on the failing machine (because they're too big to post in Bugzilla) -- is that OK? Yes sure. I'll be happy to run the tests on the failing machine. (In reply to comment #1) > I'll take a look once I'm done with bug 513 (which looks similar). I'll need to > email some test cases to you to run on the failing machine (because they're too > big to post in Bugzilla) -- is that OK? Have you tried this with 64-bit Java and 64-bit JOGL 2? Your bug log looks like you're running 32-bit Java, and I can't tell about the JOGL 2 version. This bug was originally reported at http://forum.jogamp.org/JOGL-crash-with-ATI-mobility-Radeon-HD-5470-td3402682.html. Sorry for the delay, the client was out of his office these last days. The client is running a 64-bits Windows 7 with a 32-bit browser (Firefox) and a 32-bits Java: Java Plug-in 1.6.0_29 Using JRE version 1.6.0_29-b11 Java HotSpot(TM) Client VM os.name = windows 7 os.arch = x86 I've emailed two tests to Guillaume, one to confirm his assertion that JOGL 1 works correctly on the user's machine, and one to try to duplicate the failure seen with JOGL 2. Once he attaches the log and error files for these tests to this bug report, we should have more data to work on. Resent test cases with attachments renamed to try to get around GMail's attachment filter. Created attachment 295 [details]
JOGL 1.1.1a test log
Created attachment 296 [details]
Zip of JOGL 2 test log (this log is abnormally long -- check reason)
Created attachment 297 [details]
HotSpot error log for JOGL 2 test
Attached Guillaume's error logs. The JOGL 2 error log has some very suspicious values coming back from wglGetPixelFormatAttribivARB (a driver call which JOGL 1 doesn't use) so there could be a driver bug that affects JOGL 2 but not JOGL 1. I'll analyze the log output and make sure the suspicious values are from the driver, not from the logging code or JOGL itself. I can also send a C test to Guillaume to see if wglGetPixelFormatAttribivARB acts strangely outside of Java. If so, it's definitely a driver bug. Thank you Wade. Your comment 10 makes very much sense, I concur. Since we only use these ARB extensions in case 'WGL_ARB_pixel_format' is supported by the driver, claimed via the extensions list - it is a driver bug. Shall we close this bug with this summary ? For your information, JOGL1 and LWJGL work fine on this computer. Isn't there a way to catch this error and adapt the JOGL2 behavior? Guillaume. (In reply to comment #11) > Thank you Wade. Your comment 10 makes very much sense, I concur. > Since we only use these ARB extensions in case 'WGL_ARB_pixel_format' is > supported by the driver, > claimed via the extensions list - it is a driver bug. > > Shall we close this bug with this summary ? It looks like the "suspicious values" I saw in the log are due to a bug in my logging code, not a bug in the driver :) I'll send Guillaume a C test that hits the same driver functions as JOGL 2 and we'll see if that works. Thx Wade! It seems that the native call wglCreateContextAttribsARB(..) crashes with the 'certain attributes', where previous calls w/ other attributes don't crash and some even succeed. Is this bug still applicable ? Maybe an AMD driver update fixed this bug ? If the bug still occurs, I guess we need access to such a machine to nail it down. As mentioned in another situation, an old ATI R300 mobile Thinkpad of mine w/ WinXP works fine. Is this still happening w/ latest aggregated build ? <http://jogamp.org/deployment/archive/master/gluegen_598-joal_364-jogl_824-jocl_686-signed/> I reduce it from critical -> normal, since I haven't heard anything about this anymore. Pls re-validate, thank you. Here is my customer's console when trying with v2.0-rc10. The applet stops at this point. Do you have a specific test link I could give him? ------------------------------------------------------------------------------ Java Plug-in 10.7.2.10 Using JRE version 1.7.0_07-b10 Java HotSpot(TM) Client VM User home directory = C:\Users\Darius ---------------------------------------------------- c: clear console window f: finalize objects on finalization queue g: garbage collect h: display this help message l: dump classloader list m: print memory usage o: trigger logging q: hide console r: reload policy configuration s: dump system and deployment properties t: dump thread list v: dump thread stack x: clear classloader cache 0-5: set trace level to <n> ---------------------------------------------------- Match: beginTraversal Match: digest selected JREDesc: JREDesc[version 1.4+, heap=-1--1, args=null, href=http://java.sun.com/products/autodl/j2se, sel=false, null, null], JREInfo: JREInfo for index 0: platform is: 1.7 product is: 1.7.0_07 location is: http://java.sun.com/products/autodl/j2se path is: C:\Program Files (x86)\Java\jre7\bin\javaw.exe args is: native platform is: Windows, x86 [ x86, 32bit ] JavaFX runtime is: JavaFX 2.2.0 found at C:\Program Files (x86)\Java\jre7\ enabled is: true registered is: true system is: true Match: ignoring maxHeap: -1 Match: ignoring InitHeap: -1 Match: digesting vmargs: null Match: digested vmargs: [JVMParameters: isSecure: true, args: ] Match: JVM args after accumulation: [JVMParameters: isSecure: true, args: ] Match: digest LaunchDesc: http://www.altadynqa.com/3d-virtualevents/applet/JOGL2/AltadynSignedLastRelease.jnlp Match: digest properties: [-Dsun.java2d.noddraw=true] Match: JVM args: [JVMParameters: isSecure: true, args: -Dsun.java2d.noddraw=true] Match: digest LaunchDesc: http://www.altadynqa.com/3d-virtualevents/applet/JOGL2/jogl-all-awt.jnlp Match: digest properties: [] Match: JVM args: [JVMParameters: isSecure: true, args: -Dsun.java2d.noddraw=true] Match: digest LaunchDesc: http://www.altadynqa.com/3d-virtualevents/applet/JOGL2/gluegen-rt.jnlp Match: digest properties: [] Match: JVM args: [JVMParameters: isSecure: true, args: -Dsun.java2d.noddraw=true] Match: endTraversal .. Match: JVM args final: -Dsun.java2d.noddraw=true Match: Running JREInfo Version match: 1.7.0.07 == 1.7.0.07 Match: Running JVM args mismatch: have:<-Xmx256m> !satisfy want:<-Dsun.java2d.noddraw=true> Guillaume, is this bug still reproducible with the RC11? It looks like the bug 706. We need an efficient workaround for this bug. we assume it's duplicate of 706 (or vice versa). 706 is more verbose. *** This bug has been marked as a duplicate of bug 706 *** |