Description of problem: Running the Frysk testsuite triggered the following verbatim kernel messages. Mote that this is *not* truncated output. There really were no additional messages. It looks suspiciously like a corrupt kernel oops, especially since the call trace is empty. invalid opcode: 0000 [#1] SMP last sysfs file: /devices/system/cpu/cpu0/cpufreq/scaling_setspeed Modules linked in: autofs4 hidp rfcomm l2cap bluetooth sunrpc ipv6 acpi_cpufreq freq_table dm_mirror dm_multipath dm_mod video sbs i2c_ec i2c_core dock button b attery ac lp floppy joydev sg snd_hda_intel snd_hda_codec snd_seq_dummy snd_seq_ oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd pcspkr soundcore ide_cd snd_page_alloc parport_pc iTCO_wdt tg3 iTC O_vendor_support parport serio_raw cdrom ata_piix libata sd_mod scsi_mod ext3 jb d ehci_hcd ohci_hcd uhci_hcd CPU: 0 EIP: 0060:[<f7e79381>] Not tainted VLI EFLAGS: 00010207 (2.6.20-1.2925.fc6 #1) EIP is at 0xf7e79381 eax: 71552593 ebx: f755a080 ecx: f7e79280 edx: f7561d40 esi: c2402380 edi: 00000005 ebp: 00000000 esp: c0787b40 ds: 007b es: 007b ss: 0068 Process ptrace_after_ex (pid: 9679, ti=c0787000 task=c2528c70 task.ti=eafbb000) Stack: 00000005 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 Call Trace: ======================= Code: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 40 <8e> ff f7 40 13 e6 f7 00 00 00 00 00 00 00 00 00 00 00 00 00 00 EIP: [<f7e79381>] 0xf7e79381 SS:ESP 0068:c0787b40 Version-Release number of selected component (if applicable): kernel 2.6.20-1.2925.fc6 How reproducible: Somewhat intermittent (mostly a case of usually the testsuite run triggering some kernel bugs or hang - just not always this one) Steps to Reproduce: 1. Obtain frysk source tree (CVS HEAD) 2. mkdir build 3. cd build 4. ../frysk/autogen.sh 5. make 6. make check Actual results: Kernel messages shown above. Expected results: No kernel messages. Additional info: Executing the Frysk testsuite on this kernel on i386 has been rather erratic, with this being one of the possible outcomes. Sometimes the 'make check' completes without any kernel errors getting triggered, but I'd say that I have experienced about 40-50% probability for some kernel problem getting signaled as result of the testsuite execution.
Please try 2.6.20-1.2933.fc6 from Fedora updates-testing and tell me if you still see problems. That includes fixes for several problems vs 1.2925.fc6; I do have some further fixes to be included in a forthcoming update beyond what's in 1.2933, but it is less likely that you will hit those failure modes.
Running with 2.6.20-1.2948.fc6 we no longer see this failure take place.