We are working on rebuilding the target to get rid of the (OF) flags. The example given on the open-fcoe.org website uses RHEL 6.1 which is based on 2.6.32 kernel. I would appreciate your recommendation of a stable Linux distribution release of 3.x kernel for FCoE vn2vn testing. Our configuration assumes a raw drive could be accessed by multiple initiators. Would that configuration work? Is multiple initiator supported by tcm_fc? Thanks, Jun On Mon, Apr 28, 2014 at 3:25 PM, Nicholas A. Bellinger <nab@xxxxxxxxxxxxxxx> wrote: > On Mon, 2014-04-28 at 14:36 -0700, Jun Wu wrote: >> Hi Nicholas, >> >> We upgraded Ubuntu 13.10 to the latest 14.04 which has a >> 3.13.0-24-generic kernel. I reproduced the bug with v3.13. V3.14 >> doesn't compile on the kernel. The tfc_io.c is the same in v3.14. >> >> The initiator I am using is open-fcoe-3.11. >> >> root@poc2:~# modinfo fcoe >> filename: /lib/modules/3.13.0-24-generic/kernel/drivers/scsi/fcoe/fcoe.ko >> license: GPL v2 >> description: FCoE >> author: Open-FCoE.org >> srcversion: 6DA44562FC66B71637941E8 >> depends: libfcoe,libfc,scsi_transport_fc >> intree: Y >> vermagic: 3.13.0-24-generic SMP mod_unload modversions >> signer: Magrathea: Glacier signing key >> sig_key: 00:A5:A6:57:59:DE:47:4B:C5:C4:31:20:88:0C:1B:94:A5:39:F4:31 >> sig_hashalgo: sha512 >> parm: ddp_min:Minimum I/O size in bytes for Direct Data >> Placement (DDP). (uint) >> parm: debug_logging:a bit mask of logging levels (int) >> >> root@poc2:~# fcoeadm -v >> 1.0.29 >> >> By issuing "echo eth2 > /sys/module/libfcoe/parameters/create_vn2vn", >> the initiator can see the target drives. >> >> I always run fio with 4KB sequential read on all the drives to >> reproduce the bug. Here is the crash dump with your debug information: >> >> [ 2883.272203] TARGET_CORE[fc]: Unsupported SCSI Opcode 0x85, sending >> CHECK_CONDITION. >> [ 3473.092860] CDB: 0x28 data_length: 4096 t_data_sg: (null) >> t_data_nents: 0se_cmd_flags: 0x00000109 >> [ 3473.092887] ------------[ cut here ]------------ >> [ 3473.092958] kernel BUG at /home/zb/target3.13/target/tcm_fc/tfc_io.c:100! >> [ 3473.093056] invalid opcode: 0000 [#1] SMP >> [ 3473.093123] Modules linked in: ib_srpt tcm_qla2xxx qla2xxx >> tcm_loop(OF) tcm_fc(OF) iscsi_target_mod(OF) target_core_pscsi(OF) >> target_core_file(OF) target_core_iblock(OF) target_core_mod(OF) > > Mmmm, the flags listed here for the target modules (OF) indicate that > the modules are being built both '(O)ut of tree' and are '(F)orcibly > loaded'... > > Also, the source path '/home/zb/target3.13/target/tcm_fc/tfc_io.c' > indicates some broken / strange out-of-tree build process for the target > code..? > > Why are you building the target this way, and why aren't you using the > upstream target sources..? > > --nab > -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
