Greetings,
I installed a NIC with a 8110/8169 chip in a server to upgrade to
gigabgit bandwidth. I get a lot of these in dmesg:
Badness in __kfree_skb at net/core/skbuff.c:291
[<c03e2ff1>] __kfree_skb+0xb1/0x100
[<c03e658d>] sk_stream_kill_queues+0x4d/0x180
[<c042e14c>] tcp_v4_destroy_sock+0x1c/0x180
[<c041d4d7>] tcp_destroy_sock+0x47/0x120
[<c042561a>] tcp_rcv_state_process+0x51a/0x9c0
[<c042d1b9>] tcp_v4_do_rcv+0x99/0x120
[<c042d7d0>] tcp_v4_rcv+0x590/0x770
[<c0413c7f>] ip_local_deliver+0xbf/0x1c0
[<c0414260>] ip_local_deliver_finish+0x0/0x130
[<c04140d6>] ip_rcv+0x356/0x4e0
[<c0414390>] ip_rcv_finish+0x0/0x270
[<c01b1165>] pathrelse_and_restore+0x45/0x50
[<c03e88c7>] netif_receive_skb+0x127/0x190
[<c036eeac>] rtl8169_rx_vlan_skb+0x15c/0x1a0
[<c037134a>] rtl8169_rx_interrupt+0x18a/0x400
[<c0371a50>] pci_dma_sync_single_for_device+0x0/0x1
[<c03716cb>] rtl8169_interrupt+0x10b/0x150
[<c01385a0>] handle_IRQ_event+0x30/0x70
[<c0138682>] __do_IRQ+0xa2/0xe0
[<c0448056>] tcp_in_window+0x496/0x4a0
[<c0104d79>] do_IRQ+0x19/0x30
[<c0103292>] common_interrupt+0x1a/0x20
[<c0139ad6>] find_get_page+0x26/0x30
[<c0158373>] __find_get_block_slow+0x53/0x170
[<c044a1c1>] ipt_do_table+0x251/0x330
[<c0159190>] __find_get_block+0x80/0xe0
[<c0159227>] __getblk+0x37/0x70
[<c01b15f7>] search_by_key+0x97/0xd80
[<c03edbfc>] neigh_resolve_output+0xec/0x190
[<c0416c6e>] ip_finish_output+0xfe/0x220
[<c041730a>] ip_queue_xmit+0x36a/0x510
[<c0419300>] dst_output+0x0/0x30
[<c019aab9>] make_cpu_key+0x59/0x70
[<c019cfad>] inode2sd+0xfd/0x160
[<c012cc20>] autoremove_wake_function+0x0/0x60
[<c019aab9>] make_cpu_key+0x59/0x70
[<c019d326>] reiserfs_update_sd_size+0xa6/0x220
[<c0139b3f>] find_lock_page+0x1f/0x90
[<c019aab9>] make_cpu_key+0x59/0x70
[<c01a8424>] reiserfs_dirty_inode+0x74/0x90
[<c01a2365>] reiserfs_submit_file_region_for_write+0x215/0x280
[<c01a3342>] reiserfs_file_write+0x4f2/0x690
[<c0414260>] ip_local_deliver_finish+0x0/0x130
[<c04140d6>] ip_rcv+0x356/0x4e0
[<c0414390>] ip_rcv_finish+0x0/0x270
[<c03e88c7>] netif_receive_skb+0x127/0x190
[<c036eeac>] rtl8169_rx_vlan_skb+0x15c/0x1a0
[<c03713a9>] rtl8169_rx_interrupt+0x1e9/0x400
[<c03716b5>] rtl8169_interrupt+0xf5/0x150
[<c0156ad7>] vfs_write+0xa7/0x180
[<c0156c81>] sys_write+0x51/0x80
[<c01030d5>] syscall_call+0x7/0xb
The server crashed about twice a day.
I tried the following kernels:
2.6.13.1
2.6.13.2
2.6.12.6
(in this order)
The only other change on the server was the usage of vlans since the new
NIC. I also get dead loop/fix it messages, but I think those are the
consequence of the above error...
Any ideas?
Bye
Bgs
Please:
- try 2.6.14-rc2
- open a problem report at http://bugzilla.kernel.org
- send a _complete_ dmesg/lspci -vx/lsmod
- send the vlan-related commands
--
Ueimor
Bgs wrote:
> Greetings,
>
> I installed a NIC with a 8110/8169 chip in a server to upgrade to
> gigabgit bandwidth. I get a lot of these in dmesg:
My best guess from the little information available is that this looks
like 4k stacks being used together with reiserfs and a stack overflow
corrupting the stack away.
Your .config would be helpful to confirm this, though.
I would suggest you turn CONFIG_4KSTACKS off or use a different filesystem.
--
Paulo Marques - http://www.grupopie.com
The rule is perfect: in all matters of opinion our
adversaries are insane.
Mark Twain
Paulo Marques <[email protected]> :
[...]
> I would suggest you turn CONFIG_4KSTACKS off or use a different filesystem.
This is not needed. Tommy Christensen noticed that the r8169 driver
called a forbidden function in an irq context.
I'll have to figure why it was not noticed during the last testing round
but that's a different story.
--
Ueimor