Java causes crash on dual Athlon MP

[groovemaneuver]

Hi there,

I've got a Dual Athlon MP @ 1.2 GHz w/ 2 GB RAM and a GeForceFX 5200 running Fedora Core 4 (2.6.14-1.1653_FC4smp). Anytime I try to run a Java-based application (NetBeans, Zend Studio, Azureus, etc), as soon as any Swing components appear on the screen, the computer becomes completely unresponsive -- although I can still move the mouse around the screen. About half of the time, I can hit the power button to initiate an ACPI shutdown, but the other half of the time, I have to yank the power cable. This doesn't happen using the open source "nv" driver, but I need 3-D acceleration, so that's not an option.

I've tried various versions of the nVidia driver, ranging from 1.0-7676 to the current 1.0-8178, and the result has been the same. The following is the output from /var/log/messages...

Dec 30 09:03:44 kernel: [<c014d128>] __kmalloc+0x8d/0x8f
Dec 30 09:03:44 kernel: [<f96cb3c4>] nv_alloc_kernel_mapping+0x148/0x24d [nvidia]
Dec 30 09:03:44 kernel: [<f963fc5c>] _nv003788rm+0x20/0x24 [nvidia]
Dec 30 09:03:44 kernel: [<f94a2218>] _nv002326rm+0x78/0xd4 [nvidia]
Dec 30 09:03:44 kernel: [<f95c6201>] _nv005646rm+0x389/0x440 [nvidia]
Dec 30 09:03:44 kernel: [<c011a3d6>] activate_task+0x8f/0x9e
Dec 30 09:03:44 kernel: [<c0133e47>] autoremove_wake_function+0x15/0x37
Dec 30 09:03:44 kernel: [<c011c463>] __wake_up_common+0x39/0x59
Dec 30 09:03:44 kernel: [<f94927cd>] _nv002423rm+0x1d/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f94927cd>] _nv002423rm+0x1d/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f94927d6>] _nv002423rm+0x26/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f95e870b>] _nv005100rm+0x23/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f95e870b>] _nv005100rm+0x23/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f95e8734>] _nv005098rm+0x24/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f95e8734>] _nv005098rm+0x24/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f9682775>] _nv001858rm+0x161/0x504 [nvidia]
Dec 30 09:03:44 kernel: [<f9492603>] _nv002429rm+0x6f/0x84 [nvidia]
Dec 30 09:03:44 kernel: [<f96844e8>] _nv001849rm+0xd4/0x12c [nvidia]
Dec 30 09:03:44 kernel: [<f95e79f1>] _nv005136rm+0x441/0x484 [nvidia]
Dec 30 09:03:44 kernel: [<f95e8734>] _nv005098rm+0x24/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f9682415>] _nv001856rm+0x1d5/0x240 [nvidia]
Dec 30 09:03:44 kernel: [<f96823f2>] _nv001856rm+0x1b2/0x240 [nvidia]
Dec 30 09:03:44 kernel: [<f95c5ab4>] _nv005468rm+0x79c/0xaf0 [nvidia]
Dec 30 09:03:44 kernel: [<f95e7be3>] _nv004906rm+0x8f/0x9c [nvidia]
Dec 30 09:03:44 kernel: [<f94927d6>] _nv002423rm+0x26/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f94927cd>] _nv002423rm+0x1d/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f95e870b>] _nv005100rm+0x23/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f94927d6>] _nv002423rm+0x26/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f94927cd>] _nv002423rm+0x1d/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f94927d6>] _nv002423rm+0x26/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f95e865d>] _nv005103rm+0x19/0x40 [nvidia]
Dec 30 09:03:44 kernel: [<f95e870b>] _nv005100rm+0x23/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f94927d6>] _nv002423rm+0x26/0x2c [nvidia]
Dec 30 09:03:44 kernel: [<f95e8734>] _nv005098rm+0x24/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f95e7f59>] _nv004943rm+0xa9/0xe8 [nvidia]
Dec 30 09:03:44 kernel: [<f9646505>] _nv003507rm+0x251/0x484 [nvidia]
Dec 30 09:03:44 kernel: [<f96464f5>] _nv003507rm+0x241/0x484 [nvidia]
Dec 30 09:03:44 kernel: [<f9584540>] _nv000436rm+0x20/0xa0 [nvidia]
Dec 30 09:03:44 kernel: [<f96405d0>] _nv003514rm+0x90/0xd4 [nvidia]
Dec 30 09:03:44 kernel: [<f9491cfc>] _nv002448rm+0x58/0x100 [nvidia]
Dec 30 09:03:44 kernel: [<f95e870b>] _nv005100rm+0x23/0x28 [nvidia]
Dec 30 09:03:44 kernel: [<f94a51a7>] _nv001646rm+0x87/0xa8 [nvidia]
Dec 30 09:03:44 kernel: [<c02b0bd9>] pci_read+0x29/0x2e
Dec 30 09:03:44 kernel: [<f94a961f>] rm_isr_bh+0x47/0x74 [nvidia]
Dec 30 09:03:44 kernel: [<f94a963a>] rm_isr_bh+0x62/0x74 [nvidia]
Dec 30 09:03:44 kernel: [<c01293aa>] run_timer_softirq+0x13c/0x1b2
Dec 30 09:03:44 kernel: [<f96ca84b>] nv_kern_isr_bh+0x1e/0x23 [nvidia]
Dec 30 09:03:44 kernel: [<c012594b>] tasklet_action+0x52/0xb3
Dec 30 09:03:44 kernel: [<c0125632>] __do_softirq+0x72/0xdc
Dec 30 09:03:44 kernel: [<c0105c03>] do_softirq+0x4b/0x4f
Dec 30 09:03:44 kernel: =======================
Dec 30 09:03:44 kernel: [<c0105ae5>] do_IRQ+0x55/0x86
Dec 30 09:03:44 kernel: [<c011616c>] smp_apic_timer_interrupt+0xc1/0xca
Dec 30 09:03:44 kernel: [<c01043da>] common_interrupt+0x1a/0x20
Dec 30 09:03:44 kernel: [<c021465b>] acpi_processor_idle+0x0/0x291
Dec 30 09:03:44 kernel: [<c021474f>] acpi_processor_idle+0xf4/0x291
Dec 30 09:03:44 kernel: [<c0101a14>] cpu_idle+0x4e/0x63
Dec 30 09:03:44 kernel: [<c03ea885>] start_kernel+0x175/0x1cc
Dec 30 09:03:44 kernel: [<c03ea309>] unknown_bootoption+0x0/0x1b6

Any insight or suggestions would be much appreciated.

Thanks,
Chris

[zander]

Please generate and attach an nvidia-bug-report.log file; does the problem reproduce with the latest update kernel? If so, please attach the complete Oops output, not just the stack trace.

[pzgren]

The fist problem that the computer becomes completely unresponsive, sound like here in the past with a fresh false self installed java runtime environment 1.5.06 with tomcat, mysql and opencms from me, where I made the mistake not to follow the Fedora guidelines. The whole baby runs with a 98% cpu time from gcj all the time...heavy to use the desktop.

http://fedoranews.org/mediawiki/inde...e_Java_for_FC4

Marcus

[groovemaneuver]

Hi there,

Sorry about not attaching the bug report file. I even read the instructions. Whoops. My system is using FC4's latest kernel and the latest nvidia driver. How do I go about getting the complete kernel oops output as you requested? Also, I've tried booting with every combination of "acpi=off|on|force" and with or without "noapic" with no perceivable difference.

In reply to pzgren, I'm using Sun's JDK and I have installed it the way I install any other RPM. This works fine on my Pentium systems and non-SMP Athlon. Only my Dual Athlon has the problem.

Thanks!
Chris

[pzgren]

Originally Posted by groovemaneuver

In reply to pzgren, I'm using Sun's JDK and I have installed it the way I install any other RPM. This works fine on my Pentium systems and non-SMP Athlon.

An the sun' rpm methode is the false one for FC.... like I said.

Marcus

[groovemaneuver]

Greetings,

Although it's not the ideal fix, I've "solved" the problem by switching out my home server (dual PIII @ 800 MHz) with the problematic dual Athlon workstation. I can now launch Java-based programs without any problems, even with the nVidia driver loaded. Since the dual Athlon is now my server, I threw in an old TNT video card which is only there for the occasional console access. The GeForceFX 5200 works like a charm in the dual PIII, and aside from the system being a tad slower, the Dual PIII is rock-solid.

I need this system functional for work tomorrow, so I have no more time to troubleshoot. Thanks for the tips and I'm sorry that I had to abandon this thread.

Chris