On Fri, 2009-12-04 at 15:25 -0500, andros@xxxxxxxxxx wrote: > Fix session reset deadlocks Version 4. > > These patches apply to 2.6.32 > > Fix races and bugs as well as implement a new session draining scheme > designed by Trond. > > 0001-nfs41-add-create-session-into-establish_clid.patch > 0002-nfs41-rename-cl_state-session-SETUP-bit-to-RESET.patch > 0003-nfs41-nfs4_get_lease_time-will-never-session-reset.patch > 0004-nfs41-call-free-slot-from-nfs4_restart_rpc.patch > 0005-nfs41-free-the-slot-on-unhandled-read-errors.patch > 0006-nfs41-fix-switch-in-nfs4_handle_exception.patch > 0007-nfs41-fix-switch-in-nfs4_recovery_handle_error.patch > 0008-nfs41-don-t-clear-tk_action-on-success.patch > 0009-nfs41-remove-nfs4_recover_session.patch > 0010-nfs41-nfs41-fix-state-manager-deadlock-in-session.patch > 0011-nfs41-drain-session-cleanup.patch > 0012-nfs41-only-state-manager-sets-NFS4CLNT_SESSION_SETU.patch > > Testing: > > CONFIG_NFS_V4_1 > v41 mount: Connectathon tests passed. PyNFS testclient.py SESSIONRESET tests > > The INJECT_ERROR testclient.py test where NFS4ERR_BADSESSION was returned > every 50th SEQUENCE operation and the session destroyed > durring a Connectathon basic test run. This passed all but the bigfile test > where the check_lease op->renew_lease nfs4_proc_sequence state manager > session reset call could not get a slot due to the async error handler > restart read/write RPC's getting slots prior to any rpc tasks waiting on > queues. This will be fixed in a subsequent patch set. > > v4 mount: Connectathon tests passed. > > no CONFIG_NFS_V4_1 > v4 mount: Connectathon tests passed. Thanks Andy! Those look good to me. I had to fix them up a bit in order to have them apply on top of the nfs-for-next branch, but nothing major. I'll push them out to the git repository on linux-nfs.org some time during the weekend. Hopefully Ricardo and Alexandros will have sent me their patches too by then... Cheers Trond -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
