From: Nicholas Bellinger <nab@xxxxxxxxxxxxxxx>
Hi folks,
This patch series contains two bugfixes to address possible deadlocks in
exception path handling where qla_hw_data->hardware_lock may already be
held.
The first one is possible during TMR LUN_RESET operation while cmd->aborted
status triggers a call to qla2*xx_send_term_exchange with ->hardware_lock
already held. The second occurs during tcm_qla2xxx_handle_cmd() ->
transport_lookup_cmd_lun() lookup failure where tcm_qla2xxx_release_cmd()
needs to be called directly w/ ->hardware_lock held because of the missing
se_device association. The second is fallout from the recent conversion
of the active I/O shutdown code to use ->hardware_lock directly to enforce
sychronization.
Please have a look and let us know if you have any concerns.
Thank you,
--nab
Nicholas Bellinger (2):
qla2xxx: Prevent deadlock for term_exchange w/ hardware_lock held
tcm_qla2xxx: Fix deadlock for tcm_qla2xxx_release_cmd w/
hardware_lock held
drivers/scsi/qla2xxx/qla_target.c | 8 ++-
drivers/target/tcm_qla2xxx/tcm_qla2xxx_fabric.c | 59 ++++++++++++++++-------
2 files changed, 47 insertions(+), 20 deletions(-)
--
1.7.2.5
--
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
