I am running a kernel with the latest patches for netpoll/netconsole oops patch. When doing syrq-b I get the log trace below and the node oops. I see that this bug was fixed before by applying the operation via a workqueue. The vanilla code has this implemented differently by adding the emergancy_restart and using it in sysrq-b. 2007 Mar 5 15:30:12 node1 WARNING: kernel: CPU: 0 2007 Mar 5 15:30:12 node1 WARNING: kernel: EIP: 0060:[<f88554ea>] Not tainted VLI 2007 Mar 5 15:30:12 node1 WARNING: kernel: EFLAGS: 00010002 (2.6.9-42dbg) 2007 Mar 5 15:30:12 node1 WARNING: kernel: EIP is at e1000_clean+0x9f/0x10d [e1000] 2007 Mar 5 15:30:12 node1 WARNING: kernel: eax: 00100100 ebx: f6f3c000 ecx: f6f3c100 edx: 00200200 2007 Mar 5 15:30:12 node1 WARNING: kernel: esi: f6f3c240 edi: 00000092 ebp: d26ffdb4 esp: d26ffd94 2007 Mar 5 15:30:12 node1 WARNING: kernel: ds: 007b es: 007b ss: 0068 2007 Mar 5 15:30:12 node1 WARNING: kernel: Process reboot (pid: 16277, threadinfo=d26ff000 task=f6f491b0) 2007 Mar 5 15:30:12 node1 WARNING: kernel: Stack: c0113e13 00006080 00000000 d26ffdbc 00000000 f6f3c000 c5304900 ca844a80 2007 Mar 5 15:30:12 node1 WARNING: kernel: d26ffdc8 c02c4b56 00000010 f6f3c000 c5304900 d26ffdd8 c02c4be6 f6f3c000 2007 Mar 5 15:30:12 node1 WARNING: kernel: c047145c d26ffe00 c02c4f06 f6f3c130 d26ff000 c5304900 f6f3c880 f9a25a60 2007 Mar 5 15:30:12 node1 WARNING: kernel: Call Trace: 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c010700d>] show_stack+0x7a/0x90 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c010718e>] show_registers+0x152/0x1ca 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c0107382>] die+0x100/0x19f 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c01171a5>] do_page_fault+0x400/0x67f 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c0326d37>] error_code+0x2f/0x38 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c02c4b56>] poll_napi+0x7b/0x8f 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c02c4be6>] netpoll_poll_dev+0x5c/0x5e 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c02c4f06>] netpoll_send_skb+0xcc/0x130 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<f9a25145>] write_msg+0x145/0x1a2 [netconsole] 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011ee8f>] __call_console_drivers+0x48/0x57 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011efd8>] call_console_drivers+0xd5/0xf2 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011f2eb>] release_console_sem+0x26/0xba 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011f233>] vprintk+0x13e/0x183 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011f0f3>] printk+0x18/0x1a 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c012b978>] sys_reboot+0x138/0x2e6 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c032623f>] syscall_call+0x7/0xb 2007 Mar 5 15:30:12 node1 WARNING: kernel: Code: 7d 5a f6 86 b4 00 00 00 03 74 07 89 f0 e8 9e e5 ff ff 9c 5f fa 8b 43 24 a8 20 74 73 8d 8b 00 01 00 00 8b 83 00 01 00 00 8b 51 04 <89> 50 04 89 02 c7 41 04 00 02 20 00 c7 83 00 01 00 00 00 01 10 2007 Mar 5 15:30:12 node1 WARNING: kernel: <0>Fatal exception: panic in 5 seconds 2007 Mar 5 15:30:12 node1 MAJOR: kernel: bad: scheduling while atomic! 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c010703a>] dump_stack+0x17/0x19 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c0323c3b>] schedule+0x93b/0xae5 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c03245f3>] schedule_timeout+0x83/0x172 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c01073ef>] die+0x16d/0x19f 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c01171a5>] do_page_fault+0x400/0x67f 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c0326d37>] error_code+0x2f/0x38 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c02c4b56>] poll_napi+0x7b/0x8f 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c02c4be6>] netpoll_poll_dev+0x5c/0x5e 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c02c4f06>] netpoll_send_skb+0xcc/0x130 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<f9a25145>] write_msg+0x145/0x1a2 [netconsole] 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011ee8f>] __call_console_drivers+0x48/0x57 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011efd8>] call_console_drivers+0xd5/0xf2 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011f2eb>] release_console_sem+0x26/0xba 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011f233>] vprintk+0x13e/0x183 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c011f0f3>] printk+0x18/0x1a 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c012b978>] sys_reboot+0x138/0x2e6 2007 Mar 5 15:30:12 node1 WARNING: kernel: [<c032623f>] syscall_call+0x7/0xb
wierd. Appears as though for this stack trace to occur, dev_wrapper would most likely have to return NULL, indicating that the passed in net_device wasnt allocated using alloc_etherdev, or other routine that set IFF_EXTENDED. e1000 does have some vestigual use of garbage like thats, but its all been rendered inoperative since its part of their still-broken multi-rx queue work. I note this is a debug kernel that you've built, does the same thing occur on a stock U4 kernel? Have you made any changes to the e1000 driver in your debug kernel?
We're still waiting for answers to the questions in comment #1.
ping. Any update?
closing, no updates