From b93f56a246f3ef514c508c63960dd9a1d41d55a2 Mon Sep 17 00:00:00 2001 From: Vladislav Bolkhovitin Date: Fri, 27 May 2016 02:43:57 +0000 Subject: [PATCH] Blocked revisions 6873-6879,6882,6884,6886-6888,6890-6891 via svnmerge ........ r6873 | vlnb | 2016-04-20 19:19:25 -0700 (Wed, 20 Apr 2016) | 3 lines Some more cleanups ........ r6874 | bvassche | 2016-04-25 11:36:00 -0700 (Mon, 25 Apr 2016) | 2 lines scst_vdisk: Fix a regression introduced by r6873 ........ r6875 | bvassche | 2016-04-25 21:13:42 -0700 (Mon, 25 Apr 2016) | 5 lines scst: Move kref_get_unless_zero() definition to This patch is a slightly modified version of a patch supplied by Israel Rukshin . ........ r6876 | bvassche | 2016-04-26 07:12:17 -0700 (Tue, 26 Apr 2016) | 12 lines isert: fix double free of a fake request A nop that is sent from the scst is treated as a fake request. When a fake request is allocated we release it immediately, so it's wrong to release it also on a completion error (isert_pdu_err). This commit fix a NULL dereference bug when receiving completion with error on this nop. Signed-off-by: Israel Rukshin Signed-off-by: Ariel Nahum ........ r6877 | bvassche | 2016-04-26 07:12:47 -0700 (Tue, 26 Apr 2016) | 9 lines isert: wait for RDMA CM event to complete before destroy resources There is a race between RDMACM event handler and isert_conn_free. The event handler use the connection resources and isert_conn_free destroy them. This commit fix multiple NULL dereference bugs. Signed-off-by: Israel Rukshin ........ r6878 | bvassche | 2016-04-26 07:13:13 -0700 (Tue, 26 Apr 2016) | 8 lines isert: clean tmo timer when freeing the connection The tmo timer is set when allocating a new connection. We need to release it in case it didn't execute yet. Else timer execution will lead to a NULL dereference bug on conn. Signed-off-by: Israel Rukshin ........ r6879 | bvassche | 2016-04-26 07:13:33 -0700 (Tue, 26 Apr 2016) | 10 lines isert: fix race between teardown flow and connect flow It is possible to get ESTABLISHED RDMACM event while the connection is already started teardown flow (i.e. addr change). At teardown the refcount is reduced to zero and we start releasing the connection. In established event We need to check conn is not in teardown flow by checking its ref count is not 0. Signed-off-by: Israel Rukshin ........ r6882 | bvassche | 2016-05-01 15:41:20 -0700 (Sun, 01 May 2016) | 4 lines scst_sysfs, LUN replacement: Initialize 'read_only' before using it See also r6834. ........ r6884 | vlnb | 2016-05-03 19:47:33 -0700 (Tue, 03 May 2016) | 5 lines scst_vdisk: extend vend_specific_id to 128 bytes From Adrian Saul ........ r6886 | vlnb | 2016-05-05 18:53:30 -0700 (Thu, 05 May 2016) | 6 lines scst, EXPERIMENTAL: improve commands CPU affinity Now commands on the fast path processed by the same thread throughout all its lifetime. ........ r6887 | vlnb | 2016-05-05 19:20:41 -0700 (Thu, 05 May 2016) | 5 lines cleanup: remove support of old kernels without cancel_delayed_work_sync Otherwise we can not use scst_mutex on works ........ r6888 | vlnb | 2016-05-05 20:51:21 -0700 (Thu, 05 May 2016) | 5 lines scst: follow up to r6886 Change local and more global queue priority ratio to 2:1 ........ r6890 | vlnb | 2016-05-06 17:49:43 -0700 (Fri, 06 May 2016) | 5 lines scst: add possibility of polling + few minor doc cleanups See README diff for more details ........ r6891 | bvassche | 2016-05-06 21:30:46 -0700 (Fri, 06 May 2016) | 1 line scst, iscsi-scst, scstadmin: Spell successful and unsuccessful correctly ........ git-svn-id: http://svn.code.sf.net/p/scst/svn/branches/3.1.x@6895 d57e44dd-8a1f-0410-8b47-8ef2f437770f