When de-hibernating, the ifconfig process segfaults with the following message: /lib/rcscripts/net/ifconfig.sh: line 7: 11536 Segmentation fault LC_ALL=C /sbin/ifconfig It is impossible to kill crashed ifconfig process, it has D in its state column in (h)top, therefore network connection cannot be made. Reproducible: Sometimes
Created attachment 122019 [details] emerge --info
Created attachment 122021 [details] hibernate.conf
So try sys-apps/iproute2 instead... :)
(In reply to comment #3) > So try sys-apps/iproute2 instead... :) I got used to ipconfig. Besides iproute would be workaround, not the solution.
if you cant `kill -9` it, then it's a kernel problem, not userspace
Please compile your kernel with magic sysrq support. Next time you find the hung ifconfig, capture dmesg output and upload it here. Then hit alt+sysrq+t, capture dmesg again, and attach it here.
(In reply to comment #6) dmesg says: BUG: unable to handle kernel paging request at virtual address f98456a0 printing eip: c0192737 *pde = 284dc067 *pte = 00000000 Oops: 0000 [#1] Modules linked in: ac97_bus soundcore nfs lockd sunrpc fglrx(P) agpgart ipt_MASQUERADE iptable_nat nf_nat ipt_LOG ipt_REJECT xt_tcpudp nf_conntrack_ipv4 xt_state nf_conntrack nfnetlink iptable_filter ip_tables x_tables CPU: 0 EIP: 0060:[<c0192737>] Tainted: P VLI EFLAGS: 00210296 (2.6.21-suspend2-r6-bno #2) EIP is at proc_sys_lookup_table_one+0x17/0x70 eax: f98456a0 ebx: f98456a0 ecx: f98456a0 edx: f46491a8 esi: f0c2dc40 edi: f4649190 ebp: ffffffff esp: e5ee5dd8 ds: 007b es: 007b fs: 00d8 gs: 0033 ss: 0068 Process ifconfig (pid: 20179, ti=e5ee4000 task=e83b6030 task.ti=e5ee4000) Stack: e5ee5e74 00200246 f46491a8 00000001 f0c2dc40 f4649190 00000001 c01927d5 e330cd20 f4649190 f46490b0 e5ee5e20 c0192820 f4649098 fffffffe f0c2dc40 e5ee5e74 c0192f05 f4649190 c0415b80 f4649098 e5ee5f44 c0166c58 e5ee5e80 Call Trace: [<c01927d5>] proc_sys_lookup_table+0x45/0x70 [<c0192820>] do_proc_sys_lookup+0x20/0x50 [<c0192f05>] proc_sys_lookup+0x25/0x70 [<c0166c58>] do_lookup+0x148/0x190 [<c0168608>] __link_path_walk+0x718/0xc10 [<c0168b43>] link_path_walk+0x43/0xc0 [<c0112741>] do_page_fault+0x331/0x630 [<c0314378>] sys_socketcall+0x68/0x280 [<c0168d75>] do_path_lookup+0x75/0x190 [<c0167d72>] getname+0xa2/0xd0 [<c016975b>] __user_walk_fd+0x3b/0x60 [<c015f2fb>] sys_faccessat+0x9b/0x160 [<c0112741>] do_page_fault+0x331/0x630 [<c0314378>] sys_socketcall+0x68/0x280 [<c015f3df>] sys_access+0x1f/0x30 [<c0102798>] sysenter_past_esp+0x5d/0x81 ======================= Code: 3f 37 c0 66 c7 40 0e 00 00 31 c0 c3 90 8d b4 26 00 00 00 00 55 bd ff ff ff ff 57 56 53 89 c3 83 ec 0c 89 54 24 08 eb 03 83 c3 2c <8b> 0b 85 c9 74 32 8b 7b 04 85 ff 74 f0 89 e9 31 c0 f2 ae f7 d1 EIP: [<c0192737>] proc_sys_lookup_table_one+0x17/0x70 SS:ESP 0068:e5ee5dd8
Can you reproduce this on an untainted kernel? The fglrx module could be doing nasty things. Please post a new crash trace here if you can.
(In reply to comment #8) Well, as of now I can say that it happens every time when nfs modules are loaded. I've also experienced that nasty crashes and system hangups on a nfs server system where there is no ATI card, but running KdeTV software and saa7134 module tvtuner.
Please reopen when you provide a crash trace from an untainted kernel (see comment #8)