diff options
author | Nicholas Bellinger <nab@linux-iscsi.org> | 2017-06-02 20:00:17 -0700 |
---|---|---|
committer | Nicholas Bellinger <nab@linux-iscsi.org> | 2017-06-08 22:24:18 -0700 |
commit | 73d4e580ccc5c3e05cea002f18111f66c9c07034 (patch) | |
tree | c2cfd9bfb4fe98f67467c64dd5fb53ba18d4ac98 /drivers/target/target_core_transport.c | |
parent | 5e0cf5e6c43b9e19fc0284f69e5cd2b4a47523b0 (diff) | |
download | linux-stable-73d4e580ccc5c3e05cea002f18111f66c9c07034.tar.gz linux-stable-73d4e580ccc5c3e05cea002f18111f66c9c07034.tar.bz2 linux-stable-73d4e580ccc5c3e05cea002f18111f66c9c07034.zip |
target: Fix kref->refcount underflow in transport_cmd_finish_abort
This patch fixes a se_cmd->cmd_kref underflow during CMD_T_ABORTED
when a fabric driver drops it's second reference from below the
target_core_tmr.c based callers of transport_cmd_finish_abort().
Recently with the conversion of kref to refcount_t, this bug was
manifesting itself as:
[705519.601034] refcount_t: underflow; use-after-free.
[705519.604034] INFO: NMI handler (kgdb_nmi_handler) took too long to run: 20116.512 msecs
[705539.719111] ------------[ cut here ]------------
[705539.719117] WARNING: CPU: 3 PID: 26510 at lib/refcount.c:184 refcount_sub_and_test+0x33/0x51
Since the original kref atomic_t based kref_put() didn't check for
underflow and only invoked the final callback when zero was reached,
this bug did not manifest in practice since all se_cmd memory is
using preallocated tags.
To address this, go ahead and propigate the existing return from
transport_put_cmd() up via transport_cmd_finish_abort(), and
change transport_cmd_finish_abort() + core_tmr_handle_tas_abort()
callers to only do their local target_put_sess_cmd() if necessary.
Reported-by: Bart Van Assche <bart.vanassche@sandisk.com>
Tested-by: Bart Van Assche <bart.vanassche@sandisk.com>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Himanshu Madhani <himanshu.madhani@qlogic.com>
Cc: Sagi Grimberg <sagig@mellanox.com>
Cc: stable@vger.kernel.org # 3.14+
Tested-by: Gary Guo <ghg@datera.io>
Tested-by: Chu Yuan Lin <cyl@datera.io>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Diffstat (limited to 'drivers/target/target_core_transport.c')
-rw-r--r-- | drivers/target/target_core_transport.c | 9 |
1 files changed, 6 insertions, 3 deletions
diff --git a/drivers/target/target_core_transport.c b/drivers/target/target_core_transport.c index 6025935036c9..f1b3a46bdcaf 100644 --- a/drivers/target/target_core_transport.c +++ b/drivers/target/target_core_transport.c @@ -651,9 +651,10 @@ static void transport_lun_remove_cmd(struct se_cmd *cmd) percpu_ref_put(&lun->lun_ref); } -void transport_cmd_finish_abort(struct se_cmd *cmd, int remove) +int transport_cmd_finish_abort(struct se_cmd *cmd, int remove) { bool ack_kref = (cmd->se_cmd_flags & SCF_ACK_KREF); + int ret = 0; if (cmd->se_cmd_flags & SCF_SE_LUN_CMD) transport_lun_remove_cmd(cmd); @@ -665,9 +666,11 @@ void transport_cmd_finish_abort(struct se_cmd *cmd, int remove) cmd->se_tfo->aborted_task(cmd); if (transport_cmd_check_stop_to_fabric(cmd)) - return; + return 1; if (remove && ack_kref) - transport_put_cmd(cmd); + ret = transport_put_cmd(cmd); + + return ret; } static void target_complete_failure_work(struct work_struct *work) |