James Smart wrote: > This patch set is a request to incorporate the new Broadcom > (Emulex) FC target driver, efct, into the kernel source tree. > > The driver source has been Announced a couple of times, the last > version on 12/18/2018. The driver has been hosted on gitlab for > review has had contributions from the community. > gitlab (git@xxxxxxxxxx:jsmart/efct-Emulex_FC_Target.git) > > The driver integrates into the source tree at the (new) > drivers/scsi/elx subdirectory. > > The driver consists of the following components: > - A libefc_sli subdirectory: This subdirectory contains a library that > encapsulates common definitions and routines for an Emulex SLI-4 > adapter. > - A libefc subdirectory: This subdirectory contains a library of > common routines. Of major import is a number of routines that > implement a FC Discovery engine for target mode. > - An efct subdirectory: This subdirectory contains the efct target > mode device driver. The driver utilizes the above librarys and > plugs into the SCSI LIO interfaces. The driver is SCSI only at > this time. > > The patches populate the libraries and device driver and can only > be compiled as a complete set. > > This driver is completely independent from the lpfc device driver > and there is no overlap on PCI ID's. > > The patches have been cut against the 5.6/scsi-queue branch. > > Thank you to those that have contributed to the driver in the past. > > Review comments welcome! > > -- james > > > V2 modifications: > > Contains the following modifications based on prior review comments: > Indentation/Alignment/Spacing changes > Comments: format cleanup; removed obvious or unnecessary comments; > Added comments for clarity. > Headers use #ifndef comparing for prior inclusion > Cleanup structure names (remove _s suffix) > Encapsulate use of macro arguments > Refactor to remove static function declarations for static local > routines Removed unused variables > Fix SLI4_INTF_VALID_MASK for 32bits > Ensure no BIT() use > Use __ffs() in page count macro > Reorg to move field defines out of structure definition > Commonize command building routines to reduce duplication > LIO interface: > Removed scsi initiator includes > Cleaned up interface defines > Removed lio WWN version attribute. > Expanded macros within logging macros > Cleaned up lio state setting macro > Remove __force use > Modularized session debugfs code so can be easily replaced. > Cleaned up abort task handling. Return after initiating. > Modularized where possible to reduce duplication > Convert from kthread to workqueue use > Remove unused macros > Add missing TARGET_CORE build attribute > Fix kbuild test robot warnings > > Comments not addressed: > Use of __packed: not believed necessary > Session debugfs code remains. There is not yet a common lio > mechanism to replace with. There seems to be an issue with this version and also the code from October on my setup. I am running 5.5.0-rc3 but it also happens on earlier kernel versions. While shutting down the target after some testing I execute rmdir /sys/kernel/config/target/efct/10:00:00:90:fa:f0:89:ba/tpgt_0/acls/10:00:00:90:fa:f0:89:bb but this command never returns and the shutdown script hangs. The code from August [1] and a refactored version [2] do not exhibit this problem. [ 245.485090] efct_TPG[0]_LUN[0] - Removed ACL for InitiatorNode: 10:00:00:90:fa:f0:89:bb Mapped LUN: 0 [ 245.497691] efct_TPG[0] - Freeing ACL for efct InitiatorNode: 10:00:00:90:fa:f0:89:bb Mapped LUN: 0 [ 385.687531] sysrq: Show Blocked State [ 385.687547] task PC stack pid father [ 385.687610] efct:0:0 D 0 3241 2 0x80004000 [ 385.687615] Call Trace: [ 385.687628] __schedule+0x28e/0x7a0 [ 385.687635] ? try_to_del_timer_sync+0x45/0x70 [ 385.687639] ? _raw_spin_lock_irqsave+0x14/0x40 [ 385.687643] schedule+0x46/0xb0 [ 385.687646] schedule_timeout+0x118/0x2d0 [ 385.687650] ? __next_timer_interrupt+0xb0/0xb0 [ 385.687653] wait_for_completion_timeout+0x87/0xf0 [ 385.687657] ? wake_up_q+0x90/0x90 [ 385.687682] efct_intr_thread+0x5a/0xa0 [efct] [ 385.687695] ? efct_device_detach+0x110/0x110 [efct] [ 385.687700] kthread+0xdc/0x110 [ 385.687713] ? efct_device_detach+0x110/0x110 [efct] [ 385.687716] ? kthread_park+0xa0/0xa0 [ 385.687720] ret_from_fork+0x2e/0x40 [ 385.687726] rmdir D 0 3368 3365 0x00000000 [ 385.687730] Call Trace: [ 385.687735] __schedule+0x28e/0x7a0 [ 385.687739] schedule+0x46/0xb0 [ 385.687742] schedule_timeout+0x1bd/0x2d0 [ 385.687762] ? efct_lio_close_session+0x3e/0xd0 [efct] [ 385.687780] ? efct_lio_close_session+0x3e/0xd0 [efct] [ 385.687783] ? wait_for_completion+0x2a/0xe0 [ 385.687786] wait_for_completion+0x8f/0xe0 [ 385.687789] ? wake_up_q+0x90/0x90 [ 385.687820] core_tpg_del_initiator_node_acl+0x73/0x100 [target_core_mod] [ 385.687827] ? config_item_put.part.0+0x57/0xe0 [configfs] [ 385.687845] target_fabric_nacl_base_release+0x20/0x30 [target_core_mod] [ 385.687851] config_item_put.part.0+0x78/0xe0 [configfs] [ 385.687856] config_item_put+0x11/0x20 [configfs] [ 385.687861] configfs_rmdir+0x299/0x300 [configfs] [ 385.687866] vfs_rmdir+0x6a/0x150 [ 385.687869] do_rmdir+0x16d/0x1a0 [ 385.687873] sys_rmdir+0x15/0x20 [ 385.687876] do_fast_syscall_32+0x87/0x280 [ 385.687880] entry_SYSENTER_32+0xaa/0x102 [ 385.687884] EIP: 0xb7eebb89 [ 385.687888] Code: ff 00 06 fc ff 30 06 fc ff 60 06 fc ff 90 06 fc ff d0 06 fc ff 00 07 fc ff 30 07 fc ff 70 07 fc ff 15 06 fc ff 35 06 fc ff 55 <06> fc ff 75 06 fc ff 12 06 fc ff 32 06 fc ff 52 06 fc ff 72 06 fc [ 385.687891] EAX: ffffffda EBX: bf9ea909 ECX: 00000000 EDX: bf9ea909 [ 385.687893] ESI: bf9e9114 EDI: 00000002 EBP: bf9e9078 ESP: bf9e901c [ 385.687895] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b EFLAGS: 00000292 [1] https://repo.or.cz/efct-Emulex_FC_Target/sherbszt.git/shortlog/refs/heads/v2-20190804 [2] https://repo.or.cz/efct-Emulex_FC_Target/sherbszt.git/shortlog/refs/heads/v2-20191125 Sebastian
