isci: fix frame received locking
Updates to the frame_rcvd before need to be atomic with respect to when
they are evaluated by libsas.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
diff --git a/drivers/scsi/isci/phy.c b/drivers/scsi/isci/phy.c
index 93a401d..c01d762 100644
--- a/drivers/scsi/isci/phy.c
+++ b/drivers/scsi/isci/phy.c
@@ -893,6 +893,7 @@
enum scic_sds_phy_states state = sci_phy->sm.current_state_id;
struct scic_sds_controller *scic = sci_phy->owning_port->owning_controller;
enum sci_status result;
+ unsigned long flags;
switch (state) {
case SCI_PHY_SUB_AWAIT_IAF_UF: {
@@ -911,7 +912,9 @@
if (iaf.frame_type == 0) {
u32 state;
+ spin_lock_irqsave(&iphy->sas_phy.frame_rcvd_lock, flags);
memcpy(&iphy->frame_rcvd.iaf, &iaf, sizeof(iaf));
+ spin_unlock_irqrestore(&iphy->sas_phy.frame_rcvd_lock, flags);
if (iaf.smp_tport) {
/* We got the IAF for an expander PHY go to the final
* state since there are no power requirements for
@@ -954,9 +957,11 @@
frame_index,
(void **)&fis_frame_data);
+ spin_lock_irqsave(&iphy->sas_phy.frame_rcvd_lock, flags);
scic_sds_controller_copy_sata_response(&iphy->frame_rcvd.fis,
frame_header,
fis_frame_data);
+ spin_unlock_irqrestore(&iphy->sas_phy.frame_rcvd_lock, flags);
/* got IAF we can now go to the await spinup semaphore state */
sci_change_state(&sci_phy->sm, SCI_PHY_SUB_FINAL);
