Description of problem: Oops while trying to reproduce Bug 206813. Version-Release number of selected component (if applicable): kernel-2.6.17-1.2647.fc6 gdb-6.5-8.fc6 ElectricFence-2.2.2-20.2.2 How reproducible: Did not try. Steps to Reproduce: 1. EF_ALLOW_MALLOC_0=1 LD_PRELOAD=/usr/lib/libefence.so gdb ./bla 2. b inner_main 3. run 4. On (incorrect) "guile> " prompt try to CTRL-C/kill gdb. Actual results: Attached oops, ps(1) commands and others start locking up. Expected results: Non-crashed kernel. Additional info: This kernel has still many crashes on SMP, it runs fine with "maxcpus=1".
Created attachment 136755 [details] Oops log
Hi Jan, I'm having trouble reproducing this (kernel-2_6_18-1_2725_el5 i686 UP). [root@plop ~]# wc -l /proc/`pidof gdb`/maps 9256 /proc/2524/maps [root@plop ~]# wc -l /proc/`pidof bla`/maps 6998 /proc/2527/maps which is a lot but not unexpected given ElectricFence. Would you mind trying if you can reproduce with a recent FC/RHEL kernel?
Unable to reproduce it myself, tried only on kernel-xen-2.6.18-1.2747.fc6.i686, though (the original bugreport was against native non-XEN kernel). Even the common crashes for SMP / maxcpus>1 got fixed even on non-XEN kernels since that time so it could be just a consequence. Thanks for the reproducibility attempt. It could be probablt also CLOSED as RAWHIDE.