On Thu, Jan 14, 2016 at 10:36:37AM -0800, Bart Van Assche wrote: > These patches look interesting to me but I would appreciate it if an > additional aspect would be addressed by this patch series. How about > inserting a patch before patch 1/5 that makes the target core set a flag > with the meaning "shutdown_session()" has been called instead of relying on > sess_tearing_down ? The current implementation of the target core is based > on the assumption that the shutdown_session() callback calls > target_sess_cmd_list_set_waiting(). I'd like to get rid of this assumption > such that I can move that call in the ib_srpt and tcm_qla2xxx drivers from > the shutdown_session() callback to srpt_release_channel_work(). These two > drivers are namely the only drivers for which the > target_sess_cmd_list_set_waiting() and target_wait_for_sess_cmds() calls > occur from different functions. The next step would be to merge these two > functions into a single function. In an ideal world we'd do target_sess_cmd_list_set_waiting before calling ->close_session and include waiting for the command as part of the session deregistration. But currently there are only a few drivers that even need this scheme, so this will need a careful audit of all fabrics drivers first. -- 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
