[PATCH 3.19.y-ckt 071/102] fnic: Use the local variable instead of I/O flag to acquire io_req_lock in fnic_queuecommand() to avoid deadloack

From: Kamal Mostafa
Date: Tue Sep 22 2015 - 14:07:38 EST


3.19.8-ckt7 -stable review patch. If anyone has any objections, please let me know.

------------------

From: Hiral Shah <hishah@xxxxxxxxx>

commit db196935d9562abec4510f48d887bc1f1e054fcf upstream.

We added changes in fnic driver patch 1.6.0.16 to acquire
io_req_lock in fnic_queuecommand() before issuing I/O so that io completion
is serialized. But when releasing the lock we check for the I/O flag and
this could be modified if IO abort occurs before I/O completion. In this case
we wont release the lock and causes deadlock in some scenerios. Using the
local variable to check the IO lock status will resolve the problem.

Fixes: 41df7b02db82cf6c14f094757bac3830d10a827f
Signed-off-by: Hiral Shah <hishah@xxxxxxxxx>
Signed-off-by: Sesidhar Baddela <sebaddel@xxxxxxxxx>
Signed-off-by: Anil Chintalapati <achintal@xxxxxxxxx>
Reviewed-by: Martin K. Petersen <martin.petersen@xxxxxxxxxx>
Signed-off-by: James Bottomley <JBottomley@xxxxxxxx>
Signed-off-by: Kamal Mostafa <kamal@xxxxxxxxxxxxx>
---
drivers/scsi/fnic/fnic.h | 2 +-
drivers/scsi/fnic/fnic_scsi.c | 4 +++-
2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/fnic/fnic.h b/drivers/scsi/fnic/fnic.h
index 26270c3..ce129e5 100644
--- a/drivers/scsi/fnic/fnic.h
+++ b/drivers/scsi/fnic/fnic.h
@@ -39,7 +39,7 @@

#define DRV_NAME "fnic"
#define DRV_DESCRIPTION "Cisco FCoE HBA Driver"
-#define DRV_VERSION "1.6.0.17"
+#define DRV_VERSION "1.6.0.17a"
#define PFX DRV_NAME ": "
#define DFX DRV_NAME "%d: "

diff --git a/drivers/scsi/fnic/fnic_scsi.c b/drivers/scsi/fnic/fnic_scsi.c
index 155b286..25436cd 100644
--- a/drivers/scsi/fnic/fnic_scsi.c
+++ b/drivers/scsi/fnic/fnic_scsi.c
@@ -425,6 +425,7 @@ static int fnic_queuecommand_lck(struct scsi_cmnd *sc, void (*done)(struct scsi_
unsigned long ptr;
struct fc_rport_priv *rdata;
spinlock_t *io_lock = NULL;
+ int io_lock_acquired = 0;

if (unlikely(fnic_chk_state_flags_locked(fnic, FNIC_FLAGS_IO_BLOCKED)))
return SCSI_MLQUEUE_HOST_BUSY;
@@ -518,6 +519,7 @@ static int fnic_queuecommand_lck(struct scsi_cmnd *sc, void (*done)(struct scsi_
spin_lock_irqsave(io_lock, flags);

/* initialize rest of io_req */
+ io_lock_acquired = 1;
io_req->port_id = rport->port_id;
io_req->start_time = jiffies;
CMD_STATE(sc) = FNIC_IOREQ_CMD_PENDING;
@@ -571,7 +573,7 @@ out:
(((u64)CMD_FLAGS(sc) >> 32) | CMD_STATE(sc)));

/* if only we issued IO, will we have the io lock */
- if (CMD_FLAGS(sc) & FNIC_IO_INITIALIZED)
+ if (io_lock_acquired)
spin_unlock_irqrestore(io_lock, flags);

atomic_dec(&fnic->in_flight);
--
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/