Jan Kiszka wrote: > Teresa Noviello wrote: >> ... >> (at this time i do "rtnet stop") >> >> [353109.925669] Unable to handle kernel paging request at virtual address >> 6b6b6b6b >> [353109.926145] printing eip: >> [353109.926336] e0855c31 >> [353109.926338] *pde = 00000000 >> [353109.926531] Oops: 0000 [#1] >> [353109.926846] PREEMPT >> [353109.927138] Modules linked in: tdma rtmac rtcfg rtcap rt_loopback >> rt_eepro100 rtpacket rtipv4 rtnet rtai_rtdm rtai_sem rtai_lxrt rtai_up >> rtai_hal ide_scsi >> [353109.929170] CPU: 0 >> [353109.929171] EIP: 0060:[<e0855c31>] Not tainted VLI >> [353109.929198] EFLAGS: 00010246 (2.6.15-rtai) >> [353109.929807] EIP is at tdma_detach+0x91/0x200 [tdma] >> [353109.930029] eax: 6b6b6b6b ebx: 6b6b6b6b ecx: d445e910 edx: >> c040e180 > > Ok, crash confirmed here (when I switch on CONFIG_DEBUG_SLAB in the > kernel), but with slightly different location. Looks like some race > during cleanup, a buffer is used after release. Digging deeper, patch > should follow soon. >
Fortunately only a minor issue: a missing "_safe" in a "list_for_each_entry" iteration while detaching a TDMA disciple from a RT-NIC. It's fixed in SVN now, please upgrade to latest revision. Jan
signature.asc
Description: OpenPGP digital signature
