Menu
▾
▴
[linux-iscsi-devel] [ linux-iscsi-Bugs-939904 ] n/w card stops functioning after few I/Os on iSCSI disks
|
From: SourceForge.net <no...@so...> - 2004-04-28 01:47:07
|
Bugs item #939904, was opened at 2004-04-22 13:32 Message generated for change (Settings changed) made by krishmnc You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=387023&aid=939904&group_id=26396 Category: iscsi-drvr Group: 4.0.1 >Status: Closed >Resolution: Fixed Priority: 5 Submitted By: Krishna Murthy N.C (krishmnc) Assigned to: Krishna Murthy N.C (krishmnc) Summary: n/w card stops functioning after few I/Os on iSCSI disks Initial Comment: Hi, I was testing our iSCSI initiator driver on linux-2.6.4. The host machine had Ethernet Pro 100 ethernet cards using which iSCSI sessions were established. Few minutes after starting I/O on iSCSI disks the following messages started appearing on the console: Apr 12 12:04:10 linux-2 kernel: iscsi-tx: page allocation failure. order:0, mode:0x20 Apr 12 12:04:10 linux-2 kernel: Call Trace: Apr 12 12:04:10 linux-2 kernel: [<c014b25b>] __alloc_pages+0x33b/0x3d0 Apr 12 12:04:11 linux-2 kernel: [<c01212b5>] scheduler_tick+0x105/0x680 Apr 12 12:04:11 linux-2 kernel: [<c014b317>] __get_free_pages+0x27/0x40 Apr 12 12:04:11 linux-2 kernel: [<c014f46c>] inet_sendmsg+0x4b/0x60 Apr 12 12:04:11 linux-2 kernel: [<c032710b>] cache_grow+0xdc/0x3b0 Apr 12 12:04:11 linux-2 kernel: [<c014f85d>] cache_alloc_refill+0x11d/0x530 Apr 12 12:04:11 linux-2 kernel: [<c010bff2>] apic_timer_interrupt+0x1a/0x20 Apr 12 12:04:11 linux-2 kernel: [<c0150415>] kmem_cache_alloc+0x205/0x230 Apr 12 12:04:11 linux-2 kernel: [<c032b343>] sock_sendmsg+0xcb/0xd0 Apr 12 12:04:11 linux-2 kernel: [<c011b6dd>] smp_apic_timer_interrupt+0xcd/0x140 Apr 12 12:04:11 linux-2 kernel: [<c010bff2>] apic_timer_interrupt+0x1a/0x20 Apr 12 12:04:11 linux-2 kernel: [<e0900430>] iscsi_sendmsg+0x60/0x70 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<e08fd825>] iscsi_xmit_task+0x1f5/0xa10 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<c011b6dd>] smp_apic_timer_interrupt+0xcd/0x140 Apr 12 12:04:11 linux-2 kernel: [<c010bff2>] apic_timer_interrupt+0x1a/0x20 Apr 12 12:04:11 linux-2 kernel: [<e0900430>] iscsi_sendmsg+0x60/0x70 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<e08e59fb>] kunmap_sg+0x1b/0x20 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<e09054a0>] iscsi_xmit_data+0x520/0x940 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<c010df30>] do_IRQ+0x120/0x230 Apr 12 12:04:11 linux-2 kernel: [<e08fabfa>] alloc_task+0x5a/0x100 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<e08fabfa>] alloc_task+0x5a/0x100 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<e08faea9>] add_session_task+0x39/0x60 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<e0904b61>] iscsi_xmit_queued_cmnds+0x2d1/0x6f0 [iscsi_sfnet] Apr 12 12:04:11 linux-2 kernel: [<e08e6400>] iscsi_tx_thread+0x850/0xbc0 [iscsi_sfnet] Apr 12 12:04:12 linux-2 kernel: [<c010913d>] kernel_thread_helper+0x5/0x18 Apr 12 12:04:12 linux-2 kernel: Apr 12 12:04:12 linux-2 kernel: Apr 12 12:04:12 linux-2 kernel: speedo_rx+0x265/0x380 Apr 12 12:04:12 linux-2 kernel: [<c026466e>] speedo_interrupt+0x2ee/0x310 Apr 12 12:04:12 linux-2 kernel: [<c014f262>] cache_init_objs+0xe2/0x1e0 Apr 12 12:04:12 linux-2 kernel: [<c010da9b>] handle_IRQ_event+0x3b/0x70 Apr 12 12:04:12 linux-2 kernel: [<c010def2>] do_IRQ+0xe2/0x230 Apr 12 12:04:12 linux-2 kernel: [<c010bf70>] common_interrupt+0x18/0x20 Apr 12 12:04:12 linux-2 kernel: [<c011007b>] handle_vm86_fault+0x6ab/0xa70 Apr 12 12:04:12 linux-2 kernel: [<c015026b>] kmem_cache_alloc+0x5b/0x230 Apr 12 12:04:12 linux-2 kernel: [<c032b343>] alloc_skb+0x23/0xf0 Apr 12 12:04:12 linux-2 kernel: [<c0356a49>] tcp_sendmsg+0x1149/0x13f0 Apr 12 12:04:12 linux-2 kernel: [<c034bca6>] ip_rcv+0x3b6/0x550 Apr 12 12:04:12 linux-2 kernel: [<c037ab9b>] inet_sendmsg+0x4b/0x60 Apr 12 12:04:12 linux-2 kernel: [<c032710b>] sock_sendmsg+0xcb/0xd0 Apr 12 12:04:12 linux-2 kernel: [<c011b6dd>] smp_apic_timer_interrupt+0xcd/0x140 Apr 12 12:04:12 linux-2 kernel: [<c010bf70>] common_interrupt+0x18/0x20 Apr 12 12:04:12 linux-2 kernel: [<e0900430>] iscsi_sendmsg+0x60/0x70 [iscsi_sfnet] Apr 12 12:04:12 linux-2 kernel: [<e0905419>] iscsi_xmit_data+0x499/0x940 [iscsi_sfnet] Apr 12 12:04:12 linux-2 kernel: [<c0330e80>] net_rx_action+0x80/0x120 Apr 12 12:04:13 linux-2 kernel: [<c010df85>] do_IRQ+0x175/0x230 Apr 12 12:04:13 linux-2 kernel: [<e09059ff>] iscsi_xmit_r2t_data+0x13f/0x310 [iscsi_sfnet] Apr 12 12:04:13 linux-2 kernel: [<e08e63f6>] iscsi_tx_thread+0x846/0xbc0 [iscsi_sfnet] Apr 12 12:04:13 linux-2 kernel: [<c0121fc0>] default_wake_function+0x0/0x20 Apr 12 12:04:13 linux-2 kernel: [<c0121fc0>] default_wake_function+0x0/0x20 Apr 12 12:04:13 linux-2 kernel: [<e08e5bb0>] iscsi_tx_thread+0x0/0xbc0 [iscsi_sfnet] Apr 12 12:04:13 linux-2 kernel: [<c010913d>] kernel_thread_helper+0x5/0x18 Apr 12 12:04:13 linux-2 kernel: Apr 12 12:04:13 linux-2 kernel: eth1: can't fill rx buffer (force 0)! Apr 12 12:04:13 linux-2 kernel: eth1: can't fill rx buffer (force 1)! Apr 12 12:04:13 linux-2 kernel: eth1: can't fill rx buffer (force 0)! Apr 12 12:04:13 linux-2 kernel: eth1: can't fill rx buffer (force 0)! Apr 12 12:04:13 linux-2 kernel: eth1: can't fill rx buffer (force 1)! Apr 12 12:04:13 linux-2 kernel: eth1: can't fill rx buffer (force 0)! Apr 12 12:04:13 linux-2 kernel: eth1: can't fill rx buffer (force 0)! Apr 12 12:04:14 linux-2 kernel: eth1: can't fill rx buffer (force 0)! Apr 12 12:04:17 linux-2 kernel: eth1: RxAbort command stalled -------------------------------------------------------------------------------- After this eth1 ceased to function until the system was rebooted. Even ping failed. I tried the same test on another machine which had a different n/w card. Though I saw messages about page allocation failure, the card continued to function normally. ---------------------------------------------------------------------- Comment By: Krishna Murthy N.C (krishmnc) Date: 2004-04-27 18:47 Message: Logged In: YES user_id=824504 Intel ethernet controller Ethernet Pro 100 has two drivers eepro and e100. When the former was used, the problem was seen. No problem was seen when e100 was used. The above behaviour has been documented in README. The next release will have updated README. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=387023&aid=939904&group_id=26396 |