On 02/11/2016 02:02 PM, Bart Van Assche wrote: > Hi Doug, > > This series of patches is what I came up with while testing the most > recent version of my SCSI target patch series (see also > http://thread.gmane.org/gmane.linux.scsi.target.devel/10905). This > includes fixes for ABORT handling and lockups during driver shutdown. > All review comments that have been posted so far have been addressed in > this patch series. > > Changes compared to v3 of this patch series: > - Removed node_acl_size member from the srpt_template structure. > > Changes compared to v2 of this patch series: > - Moved patch "Simplify srpt_handle_tsk_mgmt()" to the front of this > patch series and added a "Cc: stable" tag. > - Removed the UNKNOWN_TMR definition again from > include/target/target_core_base.h. > - Switched to list_empty_careful() in patch "Use a mutex to protect the > channel list" such that it is no longer needed to introduce RCU in > this patch series. > - Dropped patch "Do not complain about initiator names without leading > 0x" because it modifies the same code as a pending patch series from > Nic. > > Changes compared to v1 of this patch series: > - Rebased this patch series on top of kernel v4.5-rc1. > - As proposed by Alex Estrin, modified patch "Simplify > srpt_handle_tsk_mgmt()" such that task management function code > validity is now checked by the target core. > - Added a comment in patch "Simplify channel state management". > - Changed "return true" into "return 1" in patch "Simplify > srpt_shutdown_session()" as suggested by Christoph. > - Fixed patch "Fix srpt_handle_cmd() error paths" based on the feedback > provided by Sagi. > - Split patch "Eliminate srpt_find_channel()" in three patches. > - Left out patch "Fix a rare crash in srpt_close_session()". It is no > longer needed because srpt_disconnect_ch() is now called with the > sdev mutex held. > - Added several new patches. > > The patches in this patch series are: > 0001-IB-srpt-Simplify-srpt_handle_tsk_mgmt.patch > 0002-IB-srpt-Add-parentheses-around-sizeof-argument.patch > 0003-IB-srpt-Remove-struct-srpt_node_acl.patch > 0004-IB-srpt-Inline-srpt_sdev_name.patch > 0005-IB-srpt-Inline-srpt_get_ch_state.patch > 0006-IB-srpt-Introduce-target_reverse_dma_direction.patch > 0007-IB-srpt-Use-scsilun_to_int.patch > 0008-IB-srpt-Simplify-channel-state-management.patch > 0009-IB-srpt-Simplify-srpt_shutdown_session.patch > 0010-IB-srpt-Fix-srpt_close_session.patch > 0011-IB-srpt-Fix-srpt_handle_cmd-error-paths.patch > 0012-IB-srpt-Fix-how-aborted-commands-are-processed.patch > 0013-IB-srpt-Inline-trivial-CM-callback-functions.patch > 0014-IB-srpt-Eliminate-srpt_find_channel.patch > 0015-IB-srpt-Log-private-data-associated-with-REJ.patch > 0016-IB-srpt-Use-a-mutex-to-protect-the-channel-list.patch > 0017-IB-srpt-Detect-session-shutdown-reliably.patch > 0018-IB-srpt-Fix-srpt_write_pending.patch > 0019-IB-srpt-Log-out-all-initiators-if-a-port-is-disabled.patch > 0020-IB-srpt-Introduce-srpt_process_wait_list.patch > 0021-IB-srpt-Fix-wait-list-processing.patch Modulo the one patch I touched up, the rest looked good. I've picked this up for 4.6. -- Doug Ledford <dledford@xxxxxxxxxx> GPG KeyID: 0E572FDD
Attachment:
signature.asc
Description: OpenPGP digital signature
