LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH 00/11] hisi_sas: some misc patches
@ 2018-05-02 15:56 John Garry
  2018-05-02 15:56 ` [PATCH 01/11] scsi: hisi_sas: optimise the usage of DQ locking John Garry
                   ` (11 more replies)
  0 siblings, 12 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen; +Cc: linux-scsi, linuxarm, linux-kernel, John Garry

This patchset introduces some misc changes for the
driver. These include:
- fixes for potential problems related to SCSI EH
  and device removal races
- fix protection info size for all hw versions
- some SoC bug workarounds
- minor optimisations
- other more minor things

John Garry (2):
  scsi: hisi_sas: stop controller timer for reset
  scsi: hisi_sas: add readl poll timeout helper wrappers

Xiang Chen (7):
  scsi: hisi_sas: optimise the usage of DQ locking
  scsi: hisi_sas: Add some checks to avoid free'ing a sas_task twice
  scsi: hisi_sas: check host frozen before calling "done" function
  scsi: hisi_sas: fix PI memory size
  scsi: hisi_sas: check sas_dev gone earlier in hisi_sas_abort_task()
  scsi: hisi_sas: update PHY linkrate after a controller reset
  scsi: hisi_sas: config ATA de-reset as an constrained command for v3
    hw

Xiaofei Tan (2):
  scsi: hisi_sas: remove redundant handling to event95 for v3
  scsi: hisi_sas: workaround a v3 hw hilink bug

 drivers/scsi/hisi_sas/hisi_sas.h       |   4 +-
 drivers/scsi/hisi_sas/hisi_sas_main.c  |  37 +++++++--
 drivers/scsi/hisi_sas/hisi_sas_v2_hw.c |  70 +++++++++-------
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 142 +++++++++++++++++++++------------
 4 files changed, 166 insertions(+), 87 deletions(-)

-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 01/11] scsi: hisi_sas: optimise the usage of DQ locking
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 02/11] scsi: hisi_sas: Add some checks to avoid free'ing a sas_task twice John Garry
                   ` (10 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiang Chen, John Garry

From: Xiang Chen <chenxiang66@hisilicon.com>

In the DQ tasklet processing it is not necessary to take the DQ
lock, as there is no contention between adding slots to the CQ and
removing slots from the matching DQ.

In addition, since we run each DQ in a separate tasklet context,
there would be no possible contention between DQ processing
running for the same queue in parallel.

It is still necessary to take hisi_hba lock when free'ing slots.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 3 ---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 3 ---
 2 files changed, 6 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
index a5abde8..384e4ef 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
@@ -3151,14 +3151,12 @@ static void cq_tasklet_v2_hw(unsigned long val)
 	struct hisi_sas_complete_v2_hdr *complete_queue;
 	u32 rd_point = cq->rd_point, wr_point, dev_id;
 	int queue = cq->id;
-	struct hisi_sas_dq *dq = &hisi_hba->dq[queue];
 
 	if (unlikely(hisi_hba->reject_stp_links_msk))
 		phys_try_accept_stp_links_v2_hw(hisi_hba);
 
 	complete_queue = hisi_hba->complete_hdr[queue];
 
-	spin_lock(&dq->lock);
 	wr_point = hisi_sas_read32(hisi_hba, COMPL_Q_0_WR_PTR +
 				   (0x14 * queue));
 
@@ -3208,7 +3206,6 @@ static void cq_tasklet_v2_hw(unsigned long val)
 	/* update rd_point */
 	cq->rd_point = rd_point;
 	hisi_sas_write32(hisi_hba, COMPL_Q_0_RD_PTR + (0x14 * queue), rd_point);
-	spin_unlock(&dq->lock);
 }
 
 static irqreturn_t cq_interrupt_v2_hw(int irq_no, void *p)
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 33735a7..afc1242 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -1721,11 +1721,9 @@ static void cq_tasklet_v3_hw(unsigned long val)
 	struct hisi_sas_complete_v3_hdr *complete_queue;
 	u32 rd_point = cq->rd_point, wr_point;
 	int queue = cq->id;
-	struct hisi_sas_dq *dq = &hisi_hba->dq[queue];
 
 	complete_queue = hisi_hba->complete_hdr[queue];
 
-	spin_lock(&dq->lock);
 	wr_point = hisi_sas_read32(hisi_hba, COMPL_Q_0_WR_PTR +
 				   (0x14 * queue));
 
@@ -1752,7 +1750,6 @@ static void cq_tasklet_v3_hw(unsigned long val)
 	/* update rd_point */
 	cq->rd_point = rd_point;
 	hisi_sas_write32(hisi_hba, COMPL_Q_0_RD_PTR + (0x14 * queue), rd_point);
-	spin_unlock(&dq->lock);
 }
 
 static irqreturn_t cq_interrupt_v3_hw(int irq_no, void *p)
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 02/11] scsi: hisi_sas: Add some checks to avoid free'ing a sas_task twice
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
  2018-05-02 15:56 ` [PATCH 01/11] scsi: hisi_sas: optimise the usage of DQ locking John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 03/11] scsi: hisi_sas: check host frozen before calling "done" function John Garry
                   ` (9 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiang Chen, John Garry

From: Xiang Chen <chenxiang66@hisilicon.com>

If the SCSI host enters EH, any pending IO will be processed
by SCSI EH. However it is possible that SCSI EH will try to
abort the IO and also at the same time the IO completes in the
driver. In this situation there is a small changes of freeing the
sas_task twice.

Then if another IO re-uses freed sas_task before the second time
of free'ing sas_task, it is possible that freeing incorrect sas_task.

So to avoid this situation, add some checks to crease reliability.
The sas_task task state flag SAS_TASK_STATE_ABORTED is used to
mutually protect the LLDD and libsas free'ing the task.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_main.c  |  4 ++++
 drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 22 +++++++---------------
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 21 +++++++--------------
 3 files changed, 18 insertions(+), 29 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index d1a61b1..52746e2 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -1174,10 +1174,14 @@ static int hisi_sas_abort_task(struct sas_task *task)
 		return TMF_RESP_FUNC_FAILED;
 	}
 
+	spin_lock_irqsave(&task->task_state_lock, flags);
 	if (task->task_state_flags & SAS_TASK_STATE_DONE) {
+		spin_unlock_irqrestore(&task->task_state_lock, flags);
 		rc = TMF_RESP_FUNC_COMPLETE;
 		goto out;
 	}
+	task->task_state_flags |= SAS_TASK_STATE_ABORTED;
+	spin_unlock_irqrestore(&task->task_state_lock, flags);
 
 	sas_dev->dev_status = HISI_SAS_DEV_EH;
 	if (task->lldd_task && task->task_proto & SAS_PROTOCOL_SSP) {
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
index 384e4ef..8ca0044 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
@@ -2386,7 +2386,6 @@ static void slot_err_v2_hw(struct hisi_hba *hisi_hba,
 	struct hisi_sas_complete_v2_hdr *complete_hdr =
 			&complete_queue[slot->cmplt_queue_slot];
 	unsigned long flags;
-	int aborted;
 
 	if (unlikely(!task || !task->lldd_task || !task->dev))
 		return -EINVAL;
@@ -2396,7 +2395,6 @@ static void slot_err_v2_hw(struct hisi_hba *hisi_hba,
 	sas_dev = device->lldd_dev;
 
 	spin_lock_irqsave(&task->task_state_lock, flags);
-	aborted = task->task_state_flags & SAS_TASK_STATE_ABORTED;
 	task->task_state_flags &=
 		~(SAS_TASK_STATE_PENDING | SAS_TASK_AT_INITIATOR);
 	spin_unlock_irqrestore(&task->task_state_lock, flags);
@@ -2404,15 +2402,6 @@ static void slot_err_v2_hw(struct hisi_hba *hisi_hba,
 	memset(ts, 0, sizeof(*ts));
 	ts->resp = SAS_TASK_COMPLETE;
 
-	if (unlikely(aborted)) {
-		dev_dbg(dev, "slot_complete: task(%p) aborted\n", task);
-		ts->stat = SAS_ABORTED_TASK;
-		spin_lock_irqsave(&hisi_hba->lock, flags);
-		hisi_sas_slot_task_free(hisi_hba, task, slot);
-		spin_unlock_irqrestore(&hisi_hba->lock, flags);
-		return ts->stat;
-	}
-
 	if (unlikely(!sas_dev)) {
 		dev_dbg(dev, "slot complete: port has no device\n");
 		ts->stat = SAS_PHY_DOWN;
@@ -2523,13 +2512,16 @@ static void slot_err_v2_hw(struct hisi_hba *hisi_hba,
 	}
 
 out:
+	hisi_sas_slot_task_free(hisi_hba, task, slot);
+	sts = ts->stat;
 	spin_lock_irqsave(&task->task_state_lock, flags);
+	if (task->task_state_flags & SAS_TASK_STATE_ABORTED) {
+		spin_unlock_irqrestore(&task->task_state_lock, flags);
+		dev_info(dev, "slot complete: task(%p) aborted\n", task);
+		return SAS_ABORTED_TASK;
+	}
 	task->task_state_flags |= SAS_TASK_STATE_DONE;
 	spin_unlock_irqrestore(&task->task_state_lock, flags);
-	spin_lock_irqsave(&hisi_hba->lock, flags);
-	hisi_sas_slot_task_free(hisi_hba, task, slot);
-	spin_unlock_irqrestore(&hisi_hba->lock, flags);
-	sts = ts->stat;
 
 	if (task->task_done)
 		task->task_done(task);
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index afc1242..7346110 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -1576,7 +1576,6 @@ static irqreturn_t fatal_axi_int_v3_hw(int irq_no, void *p)
 			hisi_hba->complete_hdr[slot->cmplt_queue];
 	struct hisi_sas_complete_v3_hdr *complete_hdr =
 			&complete_queue[slot->cmplt_queue_slot];
-	int aborted;
 	unsigned long flags;
 
 	if (unlikely(!task || !task->lldd_task || !task->dev))
@@ -1587,21 +1586,12 @@ static irqreturn_t fatal_axi_int_v3_hw(int irq_no, void *p)
 	sas_dev = device->lldd_dev;
 
 	spin_lock_irqsave(&task->task_state_lock, flags);
-	aborted = task->task_state_flags & SAS_TASK_STATE_ABORTED;
 	task->task_state_flags &=
 		~(SAS_TASK_STATE_PENDING | SAS_TASK_AT_INITIATOR);
 	spin_unlock_irqrestore(&task->task_state_lock, flags);
 
 	memset(ts, 0, sizeof(*ts));
 	ts->resp = SAS_TASK_COMPLETE;
-	if (unlikely(aborted)) {
-		dev_dbg(dev, "slot complete: task(%p) aborted\n", task);
-		ts->stat = SAS_ABORTED_TASK;
-		spin_lock_irqsave(&hisi_hba->lock, flags);
-		hisi_sas_slot_task_free(hisi_hba, task, slot);
-		spin_unlock_irqrestore(&hisi_hba->lock, flags);
-		return ts->stat;
-	}
 
 	if (unlikely(!sas_dev)) {
 		dev_dbg(dev, "slot complete: port has not device\n");
@@ -1699,13 +1689,16 @@ static irqreturn_t fatal_axi_int_v3_hw(int irq_no, void *p)
 	}
 
 out:
+	hisi_sas_slot_task_free(hisi_hba, task, slot);
+	sts = ts->stat;
 	spin_lock_irqsave(&task->task_state_lock, flags);
+	if (task->task_state_flags & SAS_TASK_STATE_ABORTED) {
+		spin_unlock_irqrestore(&task->task_state_lock, flags);
+		dev_info(dev, "slot complete: task(%p) aborted\n", task);
+		return SAS_ABORTED_TASK;
+	}
 	task->task_state_flags |= SAS_TASK_STATE_DONE;
 	spin_unlock_irqrestore(&task->task_state_lock, flags);
-	spin_lock_irqsave(&hisi_hba->lock, flags);
-	hisi_sas_slot_task_free(hisi_hba, task, slot);
-	spin_unlock_irqrestore(&hisi_hba->lock, flags);
-	sts = ts->stat;
 
 	if (task->task_done)
 		task->task_done(task);
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 03/11] scsi: hisi_sas: check host frozen before calling "done" function
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
  2018-05-02 15:56 ` [PATCH 01/11] scsi: hisi_sas: optimise the usage of DQ locking John Garry
  2018-05-02 15:56 ` [PATCH 02/11] scsi: hisi_sas: Add some checks to avoid free'ing a sas_task twice John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 04/11] scsi: hisi_sas: fix PI memory size John Garry
                   ` (8 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiang Chen, John Garry

From: Xiang Chen <chenxiang66@hisilicon.com>

When the host is frozen in SCSI EH state, at any point after
the LLDD sets SAS_TASK_STATE_DONE for the sas_task task state,
libsas may free the task; see sas_scsi_find_task().

This puts the LLDD in a difficult position, in that once it
sets SAS_TASK_STATE_DONE for the task state it should not
reference the sas_task again. But the LLDD needs will check
the sas_task indirectly in calling
task->task_done()->sas_scsi_task_done() or sas_ata_task_done()
(to check if the host is frozen state actually).

And the LLDD cannot set SAS_TASK_STATE_DONE for the task
state after task->task_done() is called (as the sas_task is
free'd at this point).

This situation would seem to be a problem made by libsas.

To work around, check in the LLDD whether the host is in
frozen state to ensure it is ok to call task->task_done()
function. If in the frozen state, we rely on SCSI EH and
libsas to free the sas_task directly.

We do not do this for the following IO types:
- SMP - they are managed in libsas directly, outside SCSI EH
- Any internally originated IO, for similar reason

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas.h       |  1 +
 drivers/scsi/hisi_sas/hisi_sas_main.c  |  3 +++
 drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 14 ++++++++++++++
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 14 ++++++++++++++
 4 files changed, 32 insertions(+)

diff --git a/drivers/scsi/hisi_sas/hisi_sas.h b/drivers/scsi/hisi_sas/hisi_sas.h
index d413d05..147cfaf 100644
--- a/drivers/scsi/hisi_sas/hisi_sas.h
+++ b/drivers/scsi/hisi_sas/hisi_sas.h
@@ -196,6 +196,7 @@ struct hisi_sas_slot {
 	dma_addr_t cmd_hdr_dma;
 	struct work_struct abort_slot;
 	struct timer_list internal_abort_timer;
+	bool is_internal;
 };
 
 struct hisi_sas_tmf_task {
diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index 52746e2..8f8e642 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -382,6 +382,8 @@ static int hisi_sas_task_prep(struct sas_task *task, struct hisi_sas_dq
 	slot->cmd_hdr = &cmd_hdr_base[dlvry_queue_slot];
 	slot->task = task;
 	slot->port = port;
+	if (is_tmf)
+		slot->is_internal = true;
 	task->lldd_task = slot;
 	INIT_WORK(&slot->abort_slot, hisi_sas_slot_abort);
 
@@ -1486,6 +1488,7 @@ static int hisi_sas_query_task(struct sas_task *task)
 	slot->cmd_hdr = &cmd_hdr_base[dlvry_queue_slot];
 	slot->task = task;
 	slot->port = port;
+	slot->is_internal = true;
 	task->lldd_task = slot;
 
 	slot->buf = dma_pool_alloc(hisi_hba->buffer_pool,
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
index 8ca0044..6dda6eb 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
@@ -2380,18 +2380,21 @@ static void slot_err_v2_hw(struct hisi_hba *hisi_hba,
 	struct device *dev = hisi_hba->dev;
 	struct task_status_struct *ts;
 	struct domain_device *device;
+	struct sas_ha_struct *ha;
 	enum exec_status sts;
 	struct hisi_sas_complete_v2_hdr *complete_queue =
 			hisi_hba->complete_hdr[slot->cmplt_queue];
 	struct hisi_sas_complete_v2_hdr *complete_hdr =
 			&complete_queue[slot->cmplt_queue_slot];
 	unsigned long flags;
+	bool is_internal = slot->is_internal;
 
 	if (unlikely(!task || !task->lldd_task || !task->dev))
 		return -EINVAL;
 
 	ts = &task->task_status;
 	device = task->dev;
+	ha = device->port->ha;
 	sas_dev = device->lldd_dev;
 
 	spin_lock_irqsave(&task->task_state_lock, flags);
@@ -2523,6 +2526,17 @@ static void slot_err_v2_hw(struct hisi_hba *hisi_hba,
 	task->task_state_flags |= SAS_TASK_STATE_DONE;
 	spin_unlock_irqrestore(&task->task_state_lock, flags);
 
+	if (!is_internal && (task->task_proto != SAS_PROTOCOL_SMP)) {
+		spin_lock_irqsave(&device->done_lock, flags);
+		if (test_bit(SAS_HA_FROZEN, &ha->state)) {
+			spin_unlock_irqrestore(&device->done_lock, flags);
+			dev_info(dev, "slot complete: task(%p) ignored\n ",
+				 task);
+			return sts;
+		}
+		spin_unlock_irqrestore(&device->done_lock, flags);
+	}
+
 	if (task->task_done)
 		task->task_done(task);
 
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 7346110..5c0d968 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -1571,18 +1571,21 @@ static irqreturn_t fatal_axi_int_v3_hw(int irq_no, void *p)
 	struct device *dev = hisi_hba->dev;
 	struct task_status_struct *ts;
 	struct domain_device *device;
+	struct sas_ha_struct *ha;
 	enum exec_status sts;
 	struct hisi_sas_complete_v3_hdr *complete_queue =
 			hisi_hba->complete_hdr[slot->cmplt_queue];
 	struct hisi_sas_complete_v3_hdr *complete_hdr =
 			&complete_queue[slot->cmplt_queue_slot];
 	unsigned long flags;
+	bool is_internal = slot->is_internal;
 
 	if (unlikely(!task || !task->lldd_task || !task->dev))
 		return -EINVAL;
 
 	ts = &task->task_status;
 	device = task->dev;
+	ha = device->port->ha;
 	sas_dev = device->lldd_dev;
 
 	spin_lock_irqsave(&task->task_state_lock, flags);
@@ -1700,6 +1703,17 @@ static irqreturn_t fatal_axi_int_v3_hw(int irq_no, void *p)
 	task->task_state_flags |= SAS_TASK_STATE_DONE;
 	spin_unlock_irqrestore(&task->task_state_lock, flags);
 
+	if (!is_internal && (task->task_proto != SAS_PROTOCOL_SMP)) {
+		spin_lock_irqsave(&device->done_lock, flags);
+		if (test_bit(SAS_HA_FROZEN, &ha->state)) {
+			spin_unlock_irqrestore(&device->done_lock, flags);
+			dev_info(dev, "slot complete: task(%p) ignored\n ",
+				 task);
+			return sts;
+		}
+		spin_unlock_irqrestore(&device->done_lock, flags);
+	}
+
 	if (task->task_done)
 		task->task_done(task);
 
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 04/11] scsi: hisi_sas: fix PI memory size
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (2 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 03/11] scsi: hisi_sas: check host frozen before calling "done" function John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 05/11] scsi: hisi_sas: check sas_dev gone earlier in hisi_sas_abort_task() John Garry
                   ` (7 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiang Chen, John Garry

From: Xiang Chen <chenxiang66@hisilicon.com>

There are 28 bytes of protection information record of SSP for
v3 hw, 16 bytes for v2 hw, and probably 24 for v1 hw (forgotten
now).

So use a value big enough in hisi_sas_command_table_ssp.prot to
cover all cases.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas.h b/drivers/scsi/hisi_sas/hisi_sas.h
index 147cfaf..04a40c4 100644
--- a/drivers/scsi/hisi_sas/hisi_sas.h
+++ b/drivers/scsi/hisi_sas/hisi_sas.h
@@ -410,7 +410,7 @@ struct hisi_sas_command_table_ssp {
 	union {
 		struct {
 			struct ssp_command_iu task;
-			u32 prot[6];
+			u32 prot[7];
 		};
 		struct ssp_tmf_iu ssp_task;
 		struct xfer_rdy_iu xfer_rdy;
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 05/11] scsi: hisi_sas: check sas_dev gone earlier in hisi_sas_abort_task()
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (3 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 04/11] scsi: hisi_sas: fix PI memory size John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 06/11] scsi: hisi_sas: stop controller timer for reset John Garry
                   ` (6 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiang Chen, John Garry

From: Xiang Chen <chenxiang66@hisilicon.com>

It is possible to dereference a NULL-pointer in hisi_sas_abort_task()
in special scenario when the device has been removed.

If an SMP task times-out, it will call hisi_sas_abort_task() to
recover. And currently there is a check in hisi_sas_abort_task() to
avoid the situation of processing the abort for the removed device.

However we have an ordering problem, in that we may reference
a task for the removed device before checking if the device has
been removed.

Fix this by only referencing the sas_dev after we know it is still
present.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_main.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index 8f8e642..24416bb 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -1166,15 +1166,16 @@ static int hisi_sas_abort_task(struct sas_task *task)
 	struct hisi_sas_tmf_task tmf_task;
 	struct domain_device *device = task->dev;
 	struct hisi_sas_device *sas_dev = device->lldd_dev;
-	struct hisi_hba *hisi_hba = dev_to_hisi_hba(task->dev);
-	struct device *dev = hisi_hba->dev;
+	struct hisi_hba *hisi_hba;
+	struct device *dev;
 	int rc = TMF_RESP_FUNC_FAILED;
 	unsigned long flags;
 
-	if (!sas_dev) {
-		dev_warn(dev, "Device has been removed\n");
+	if (!sas_dev)
 		return TMF_RESP_FUNC_FAILED;
-	}
+
+	hisi_hba = dev_to_hisi_hba(task->dev);
+	dev = hisi_hba->dev;
 
 	spin_lock_irqsave(&task->task_state_lock, flags);
 	if (task->task_state_flags & SAS_TASK_STATE_DONE) {
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 06/11] scsi: hisi_sas: stop controller timer for reset
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (4 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 05/11] scsi: hisi_sas: check sas_dev gone earlier in hisi_sas_abort_task() John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 07/11] scsi: hisi_sas: update PHY linkrate after a controller reset John Garry
                   ` (5 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, John Garry, Xiaofei Tan

We should only have the timer enabled after PHY up after
controller reset, so disable prior to reset.

Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_main.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index 24416bb..1f27f84 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -1130,6 +1130,9 @@ static int hisi_sas_controller_reset(struct hisi_hba *hisi_hba)
 	old_state = hisi_hba->hw->get_phys_state(hisi_hba);
 
 	scsi_block_requests(shost);
+	if (timer_pending(&hisi_hba->timer))
+		del_timer_sync(&hisi_hba->timer);
+
 	set_bit(HISI_SAS_REJECT_CMD_BIT, &hisi_hba->flags);
 	rc = hisi_hba->hw->soft_reset(hisi_hba);
 	if (rc) {
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 07/11] scsi: hisi_sas: update PHY linkrate after a controller reset
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (5 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 06/11] scsi: hisi_sas: stop controller timer for reset John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 08/11] scsi: hisi_sas: config ATA de-reset as an constrained command for v3 hw John Garry
                   ` (4 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiang Chen, John Garry

From: Xiang Chen <chenxiang66@hisilicon.com>

After the controller is reset, we currently may not honour the
PHY max linkrate set via sysfs, in that after a reset we always
revert to max linkrate of 12Gbps, ignoring the value set via
sysfs.

This patch modifies to policy to set the programmed PHY linkrate,
honouring the max linkrate programmed via sysfs.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas.h       |  1 +
 drivers/scsi/hisi_sas/hisi_sas_main.c  | 16 ++++++++++++++++
 drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 31 ++++++++++++++++++-------------
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 31 ++++++++++++++++++-------------
 4 files changed, 53 insertions(+), 26 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas.h b/drivers/scsi/hisi_sas/hisi_sas.h
index 04a40c4..4410538 100644
--- a/drivers/scsi/hisi_sas/hisi_sas.h
+++ b/drivers/scsi/hisi_sas/hisi_sas.h
@@ -463,4 +463,5 @@ extern void hisi_sas_slot_task_free(struct hisi_hba *hisi_hba,
 extern bool hisi_sas_notify_phy_event(struct hisi_sas_phy *phy,
 				enum hisi_sas_phy_event event);
 extern void hisi_sas_release_tasks(struct hisi_hba *hisi_hba);
+extern u8 hisi_sas_get_prog_phy_linkrate_mask(enum sas_linkrate max);
 #endif
diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c
index 1f27f84..ff5b8d7 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_main.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_main.c
@@ -135,6 +135,22 @@ int hisi_sas_get_ncq_tag(struct sas_task *task, u32 *tag)
 }
 EXPORT_SYMBOL_GPL(hisi_sas_get_ncq_tag);
 
+/*
+ * This function assumes linkrate mask fits in 8 bits, which it
+ * does for all HW versions supported.
+ */
+u8 hisi_sas_get_prog_phy_linkrate_mask(enum sas_linkrate max)
+{
+	u16 rate = 0;
+	int i;
+
+	max -= SAS_LINK_RATE_1_5_GBPS;
+	for (i = 0; i <= max; i++)
+		rate |= 1 << (i * 2);
+	return rate;
+}
+EXPORT_SYMBOL_GPL(hisi_sas_get_prog_phy_linkrate_mask);
+
 static struct hisi_hba *dev_to_hisi_hba(struct domain_device *device)
 {
 	return device->port->ha->lldd_ha;
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
index 6dda6eb..9e68731 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
@@ -1216,7 +1216,22 @@ static void init_reg_v2_hw(struct hisi_hba *hisi_hba)
 	}
 
 	for (i = 0; i < hisi_hba->n_phy; i++) {
-		hisi_sas_phy_write32(hisi_hba, i, PROG_PHY_LINK_RATE, 0x855);
+		struct hisi_sas_phy *phy = &hisi_hba->phy[i];
+		struct asd_sas_phy *sas_phy = &phy->sas_phy;
+		u32 prog_phy_link_rate = 0x800;
+
+		if (!sas_phy->phy || (sas_phy->phy->maximum_linkrate <
+				SAS_LINK_RATE_1_5_GBPS)) {
+			prog_phy_link_rate = 0x855;
+		} else {
+			enum sas_linkrate max = sas_phy->phy->maximum_linkrate;
+
+			prog_phy_link_rate =
+				hisi_sas_get_prog_phy_linkrate_mask(max) |
+				0x800;
+		}
+		hisi_sas_phy_write32(hisi_hba, i, PROG_PHY_LINK_RATE,
+			prog_phy_link_rate);
 		hisi_sas_phy_write32(hisi_hba, i, SAS_PHY_CTRL, sas_phy_ctrl);
 		hisi_sas_phy_write32(hisi_hba, i, SL_TOUT_CFG, 0x7d7d7d7d);
 		hisi_sas_phy_write32(hisi_hba, i, SL_CONTROL, 0x0);
@@ -1585,13 +1600,10 @@ static enum sas_linkrate phy_get_max_linkrate_v2_hw(void)
 static void phy_set_linkrate_v2_hw(struct hisi_hba *hisi_hba, int phy_no,
 		struct sas_phy_linkrates *r)
 {
-	u32 prog_phy_link_rate =
-		hisi_sas_phy_read32(hisi_hba, phy_no, PROG_PHY_LINK_RATE);
 	struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
 	struct asd_sas_phy *sas_phy = &phy->sas_phy;
-	int i;
 	enum sas_linkrate min, max;
-	u32 rate_mask = 0;
+	u32 prog_phy_link_rate = 0x800;
 
 	if (r->maximum_linkrate == SAS_LINK_RATE_UNKNOWN) {
 		max = sas_phy->phy->maximum_linkrate;
@@ -1604,14 +1616,7 @@ static void phy_set_linkrate_v2_hw(struct hisi_hba *hisi_hba, int phy_no,
 
 	sas_phy->phy->maximum_linkrate = max;
 	sas_phy->phy->minimum_linkrate = min;
-
-	max -= SAS_LINK_RATE_1_5_GBPS;
-
-	for (i = 0; i <= max; i++)
-		rate_mask |= 1 << (i * 2);
-
-	prog_phy_link_rate &= ~0xff;
-	prog_phy_link_rate |= rate_mask;
+	prog_phy_link_rate |= hisi_sas_get_prog_phy_linkrate_mask(max);
 
 	disable_phy_v2_hw(hisi_hba, phy_no);
 	msleep(100);
diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 5c0d968..ffa3cea 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -429,7 +429,22 @@ static void init_reg_v3_hw(struct hisi_hba *hisi_hba)
 	hisi_sas_write32(hisi_hba, HYPER_STREAM_ID_EN_CFG, 1);
 
 	for (i = 0; i < hisi_hba->n_phy; i++) {
-		hisi_sas_phy_write32(hisi_hba, i, PROG_PHY_LINK_RATE, 0x855);
+		struct hisi_sas_phy *phy = &hisi_hba->phy[i];
+		struct asd_sas_phy *sas_phy = &phy->sas_phy;
+		u32 prog_phy_link_rate = 0x800;
+
+		if (!sas_phy->phy || (sas_phy->phy->maximum_linkrate <
+				SAS_LINK_RATE_1_5_GBPS)) {
+			prog_phy_link_rate = 0x855;
+		} else {
+			enum sas_linkrate max = sas_phy->phy->maximum_linkrate;
+
+			prog_phy_link_rate =
+				hisi_sas_get_prog_phy_linkrate_mask(max) |
+				0x800;
+		}
+		hisi_sas_phy_write32(hisi_hba, i, PROG_PHY_LINK_RATE,
+			prog_phy_link_rate);
 		hisi_sas_phy_write32(hisi_hba, i, SAS_RX_TRAIN_TIMER, 0x13e80);
 		hisi_sas_phy_write32(hisi_hba, i, CHL_INT0, 0xffffffff);
 		hisi_sas_phy_write32(hisi_hba, i, CHL_INT1, 0xffffffff);
@@ -1869,13 +1884,10 @@ static int hisi_sas_v3_init(struct hisi_hba *hisi_hba)
 static void phy_set_linkrate_v3_hw(struct hisi_hba *hisi_hba, int phy_no,
 		struct sas_phy_linkrates *r)
 {
-	u32 prog_phy_link_rate =
-		hisi_sas_phy_read32(hisi_hba, phy_no, PROG_PHY_LINK_RATE);
 	struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
 	struct asd_sas_phy *sas_phy = &phy->sas_phy;
-	int i;
 	enum sas_linkrate min, max;
-	u32 rate_mask = 0;
+	u32 prog_phy_link_rate = 0x800;
 
 	if (r->maximum_linkrate == SAS_LINK_RATE_UNKNOWN) {
 		max = sas_phy->phy->maximum_linkrate;
@@ -1888,14 +1900,7 @@ static void phy_set_linkrate_v3_hw(struct hisi_hba *hisi_hba, int phy_no,
 
 	sas_phy->phy->maximum_linkrate = max;
 	sas_phy->phy->minimum_linkrate = min;
-
-	max -= SAS_LINK_RATE_1_5_GBPS;
-
-	for (i = 0; i <= max; i++)
-		rate_mask |= 1 << (i * 2);
-
-	prog_phy_link_rate &= ~0xff;
-	prog_phy_link_rate |= rate_mask;
+	prog_phy_link_rate |= hisi_sas_get_prog_phy_linkrate_mask(max);
 
 	disable_phy_v3_hw(hisi_hba, phy_no);
 	msleep(100);
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 08/11] scsi: hisi_sas: config ATA de-reset as an constrained command for v3 hw
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (6 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 07/11] scsi: hisi_sas: update PHY linkrate after a controller reset John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 09/11] scsi: hisi_sas: remove redundant handling to event95 for v3 John Garry
                   ` (3 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiang Chen, John Garry

From: Xiang Chen <chenxiang66@hisilicon.com>

As a unconstrained command, a command can be sent to SATA
disk even if SATA disk status is BUSY, ERR or DRQ.

If an ATA reset assert is successful but ATA reset de-assert
fails, then it will retry the  reset de-assert. If reset de-
assert retry is successful, we think it is okay to probe the
device but actually it still has Err status.

Apparently we need to retry the ATA reset assertion and de-
assertion instead for this mentioned scenario.

As such, we config ATA reset assert as a constrained command,
if ATA reset de-assert fails, then ATA reset de-assert
retry will also fail. Then we will retry the proper process
of ATA reset assert and de-assert again.

Signed-off-by: Xiang Chen <chenxiang66@hisilicon.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index ffa3cea..026faee 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -351,10 +351,11 @@ struct hisi_sas_err_record_v3 {
 #define DIR_TO_DEVICE 2
 #define DIR_RESERVED 3
 
-#define CMD_IS_UNCONSTRAINT(cmd) \
-	((cmd == ATA_CMD_READ_LOG_EXT) || \
-	(cmd == ATA_CMD_READ_LOG_DMA_EXT) || \
-	(cmd == ATA_CMD_DEV_RESET))
+#define FIS_CMD_IS_UNCONSTRAINED(fis) \
+	((fis.command == ATA_CMD_READ_LOG_EXT) || \
+	(fis.command == ATA_CMD_READ_LOG_DMA_EXT) || \
+	((fis.command == ATA_CMD_DEV_RESET) && \
+	((fis.control & ATA_SRST) != 0)))
 
 static u32 hisi_sas_read32(struct hisi_hba *hisi_hba, u32 off)
 {
@@ -1075,7 +1076,7 @@ static int prep_ata_v3_hw(struct hisi_hba *hisi_hba,
 		<< CMD_HDR_FRAME_TYPE_OFF;
 	dw1 |= sas_dev->device_id << CMD_HDR_DEV_ID_OFF;
 
-	if (CMD_IS_UNCONSTRAINT(task->ata_task.fis.command))
+	if (FIS_CMD_IS_UNCONSTRAINED(task->ata_task.fis))
 		dw1 |= 1 << CMD_HDR_UNCON_CMD_OFF;
 
 	hdr->dw1 = cpu_to_le32(dw1);
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 09/11] scsi: hisi_sas: remove redundant handling to event95 for v3
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (7 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 08/11] scsi: hisi_sas: config ATA de-reset as an constrained command for v3 hw John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 10/11] scsi: hisi_sas: add readl poll timeout helper wrappers John Garry
                   ` (2 subsequent siblings)
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiaofei Tan, John Garry

From: Xiaofei Tan <tanxiaofei@huawei.com>

Event95 is used for DFX purpose. The relevant bit for this
interrupt in the ENT_INT_SRC_MSK3 register has been disabled,
so remove the processing.

Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 026faee..d6e705f 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -1331,14 +1331,9 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 {
 	struct hisi_hba *hisi_hba = p;
 	struct device *dev = hisi_hba->dev;
-	u32 ent_msk, ent_tmp, irq_msk;
+	u32 irq_msk;
 	int phy_no = 0;
 
-	ent_msk = hisi_sas_read32(hisi_hba, ENT_INT_SRC_MSK3);
-	ent_tmp = ent_msk;
-	ent_msk |= ENT_INT_SRC_MSK3_ENT95_MSK_MSK;
-	hisi_sas_write32(hisi_hba, ENT_INT_SRC_MSK3, ent_msk);
-
 	irq_msk = hisi_sas_read32(hisi_hba, CHNL_INT_STATUS)
 				& 0xeeeeeeee;
 
@@ -1415,8 +1410,6 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 		phy_no++;
 	}
 
-	hisi_sas_write32(hisi_hba, ENT_INT_SRC_MSK3, ent_tmp);
-
 	return IRQ_HANDLED;
 }
 
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 10/11] scsi: hisi_sas: add readl poll timeout helper wrappers
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (8 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 09/11] scsi: hisi_sas: remove redundant handling to event95 for v3 John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-02 15:56 ` [PATCH 11/11] scsi: hisi_sas: workaround a v3 hw hilink bug John Garry
  2018-05-08  5:16 ` [PATCH 00/11] hisi_sas: some misc patches Martin K. Petersen
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen; +Cc: linux-scsi, linuxarm, linux-kernel, John Garry

It is common to use readl poll timeout helpers in the
driver, so create custom wrappers.

Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 28 ++++++++++++++++++++++------
 1 file changed, 22 insertions(+), 6 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index d6e705f..28bb71e 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -394,6 +394,20 @@ static u32 hisi_sas_phy_read32(struct hisi_hba *hisi_hba,
 	return readl(regs);
 }
 
+#define hisi_sas_read32_poll_timeout(off, val, cond, delay_us,		\
+				     timeout_us)			\
+({									\
+	void __iomem *regs = hisi_hba->regs + off;			\
+	readl_poll_timeout(regs, val, cond, delay_us, timeout_us);	\
+})
+
+#define hisi_sas_read32_poll_timeout_atomic(off, val, cond, delay_us,	\
+					    timeout_us)			\
+({									\
+	void __iomem *regs = hisi_hba->regs + off;			\
+	readl_poll_timeout_atomic(regs, val, cond, delay_us, timeout_us);\
+})
+
 static void init_reg_v3_hw(struct hisi_hba *hisi_hba)
 {
 	struct pci_dev *pdev = hisi_hba->pci_dev;
@@ -684,8 +698,8 @@ static int reset_hw_v3_hw(struct hisi_hba *hisi_hba)
 	udelay(50);
 
 	/* Ensure axi bus idle */
-	ret = readl_poll_timeout(hisi_hba->regs + AXI_CFG, val, !val,
-			20000, 1000000);
+	ret = hisi_sas_read32_poll_timeout(AXI_CFG, val, !val,
+					   20000, 1000000);
 	if (ret) {
 		dev_err(dev, "axi bus is not idle, ret = %d!\n", ret);
 		return -EIO;
@@ -1977,8 +1991,9 @@ static int soft_reset_v3_hw(struct hisi_hba *hisi_hba)
 	hisi_sas_write32(hisi_hba, AXI_MASTER_CFG_BASE + AM_CTRL_GLOBAL, 0x1);
 
 	/* wait until bus idle */
-	rc = readl_poll_timeout(hisi_hba->regs + AXI_MASTER_CFG_BASE +
-		AM_CURR_TRANS_RETURN, status, status == 0x3, 10, 100);
+	rc = hisi_sas_read32_poll_timeout(AXI_MASTER_CFG_BASE +
+					  AM_CURR_TRANS_RETURN, status,
+					  status == 0x3, 10, 100);
 	if (rc) {
 		dev_err(dev, "axi bus is not idle, rc = %d\n", rc);
 		return rc;
@@ -2396,8 +2411,9 @@ static int hisi_sas_v3_suspend(struct pci_dev *pdev, pm_message_t state)
 		AM_CTRL_GLOBAL, reg_val);
 
 	/* wait until bus idle */
-	rc = readl_poll_timeout(hisi_hba->regs + AXI_MASTER_CFG_BASE +
-		AM_CURR_TRANS_RETURN, status, status == 0x3, 10, 100);
+	rc = hisi_sas_read32_poll_timeout(AXI_MASTER_CFG_BASE +
+					  AM_CURR_TRANS_RETURN, status,
+					  status == 0x3, 10, 100);
 	if (rc) {
 		dev_err(dev, "axi bus is not idle, rc = %d\n", rc);
 		clear_bit(HISI_SAS_REJECT_CMD_BIT, &hisi_hba->flags);
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* [PATCH 11/11] scsi: hisi_sas: workaround a v3 hw hilink bug
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (9 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 10/11] scsi: hisi_sas: add readl poll timeout helper wrappers John Garry
@ 2018-05-02 15:56 ` John Garry
  2018-05-08  5:16 ` [PATCH 00/11] hisi_sas: some misc patches Martin K. Petersen
  11 siblings, 0 replies; 13+ messages in thread
From: John Garry @ 2018-05-02 15:56 UTC (permalink / raw)
  To: jejb, martin.petersen
  Cc: linux-scsi, linuxarm, linux-kernel, Xiaofei Tan, John Garry

From: Xiaofei Tan <tanxiaofei@huawei.com>

There is an SoC bug of v3 hw development version. When hot-
unplugging a directly attached disk, the PHY down interrupt
may not happen. It is very easy to appear on some boards.

When this issue occurs, the controller will receive many invalid
dword frames, and the "alos" fields of register HILINK_ERR_DFX
can indicate that disk was unplugged.

As an workaround solution, this patch detects this issue in
the channel interrupt, and workaround it by following steps:
- Disable the PHY
- Clear error code and interrupt
- Enable the PHY

Then the HW will reissue PHY down interrupt.

Signed-off-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: John Garry <john.garry@huawei.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 25 ++++++++++++++++++++++++-
 1 file changed, 24 insertions(+), 1 deletion(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 28bb71e..492c3be 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -106,6 +106,7 @@
 #define COMPL_Q_0_RD_PTR		0x4f0
 #define AWQOS_AWCACHE_CFG	0xc84
 #define ARQOS_ARCACHE_CFG	0xc88
+#define HILINK_ERR_DFX		0xe04
 
 /* phy registers requiring init */
 #define PORT_BASE			(0x2000)
@@ -167,6 +168,7 @@
 #define CHL_INT1_DMAC_RX_AXI_RD_ERR_OFF	22
 #define CHL_INT2			(PORT_BASE + 0x1bc)
 #define CHL_INT2_SL_IDAF_TOUT_CONF_OFF	0
+#define CHL_INT2_RX_INVLD_DW_OFF	30
 #define CHL_INT2_STP_LINK_TIMEOUT_OFF	31
 #define CHL_INT0_MSK			(PORT_BASE + 0x1c0)
 #define CHL_INT1_MSK			(PORT_BASE + 0x1c4)
@@ -1345,6 +1347,7 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 {
 	struct hisi_hba *hisi_hba = p;
 	struct device *dev = hisi_hba->dev;
+	struct pci_dev *pci_dev = hisi_hba->pci_dev;
 	u32 irq_msk;
 	int phy_no = 0;
 
@@ -1410,8 +1413,28 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 
 			hisi_sas_phy_write32(hisi_hba, phy_no,
 					     CHL_INT2, irq_value2);
-		}
 
+			if ((irq_value2 & BIT(CHL_INT2_RX_INVLD_DW_OFF)) &&
+			    (pci_dev->revision == 0x20)) {
+				u32 reg_value;
+				int rc;
+
+				rc = hisi_sas_read32_poll_timeout_atomic(
+					HILINK_ERR_DFX, reg_value,
+					!((reg_value >> 8) & BIT(phy_no)),
+					1000, 10000);
+				if (rc) {
+					disable_phy_v3_hw(hisi_hba, phy_no);
+					hisi_sas_phy_write32(hisi_hba, phy_no,
+						CHL_INT2,
+						BIT(CHL_INT2_RX_INVLD_DW_OFF));
+					hisi_sas_phy_read32(hisi_hba, phy_no,
+						ERR_CNT_INVLD_DW);
+					mdelay(1);
+					enable_phy_v3_hw(hisi_hba, phy_no);
+				}
+			}
+		}
 
 		if (irq_msk & (2 << (phy_no * 4)) && irq_value0) {
 			hisi_sas_phy_write32(hisi_hba, phy_no,
-- 
1.9.1

^ permalink raw reply	[flat|nested] 13+ messages in thread

* Re: [PATCH 00/11] hisi_sas: some misc patches
  2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
                   ` (10 preceding siblings ...)
  2018-05-02 15:56 ` [PATCH 11/11] scsi: hisi_sas: workaround a v3 hw hilink bug John Garry
@ 2018-05-08  5:16 ` Martin K. Petersen
  11 siblings, 0 replies; 13+ messages in thread
From: Martin K. Petersen @ 2018-05-08  5:16 UTC (permalink / raw)
  To: John Garry; +Cc: jejb, martin.petersen, linux-scsi, linuxarm, linux-kernel


John,

> This patchset introduces some misc changes for the
> driver. These include:
> - fixes for potential problems related to SCSI EH
>   and device removal races
> - fix protection info size for all hw versions
> - some SoC bug workarounds
> - minor optimisations
> - other more minor things

Applied to 4.18/scsi-queue. Thanks!

-- 
Martin K. Petersen	Oracle Linux Engineering

^ permalink raw reply	[flat|nested] 13+ messages in thread

end of thread, other threads:[~2018-05-08  5:16 UTC | newest]

Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-05-02 15:56 [PATCH 00/11] hisi_sas: some misc patches John Garry
2018-05-02 15:56 ` [PATCH 01/11] scsi: hisi_sas: optimise the usage of DQ locking John Garry
2018-05-02 15:56 ` [PATCH 02/11] scsi: hisi_sas: Add some checks to avoid free'ing a sas_task twice John Garry
2018-05-02 15:56 ` [PATCH 03/11] scsi: hisi_sas: check host frozen before calling "done" function John Garry
2018-05-02 15:56 ` [PATCH 04/11] scsi: hisi_sas: fix PI memory size John Garry
2018-05-02 15:56 ` [PATCH 05/11] scsi: hisi_sas: check sas_dev gone earlier in hisi_sas_abort_task() John Garry
2018-05-02 15:56 ` [PATCH 06/11] scsi: hisi_sas: stop controller timer for reset John Garry
2018-05-02 15:56 ` [PATCH 07/11] scsi: hisi_sas: update PHY linkrate after a controller reset John Garry
2018-05-02 15:56 ` [PATCH 08/11] scsi: hisi_sas: config ATA de-reset as an constrained command for v3 hw John Garry
2018-05-02 15:56 ` [PATCH 09/11] scsi: hisi_sas: remove redundant handling to event95 for v3 John Garry
2018-05-02 15:56 ` [PATCH 10/11] scsi: hisi_sas: add readl poll timeout helper wrappers John Garry
2018-05-02 15:56 ` [PATCH 11/11] scsi: hisi_sas: workaround a v3 hw hilink bug John Garry
2018-05-08  5:16 ` [PATCH 00/11] hisi_sas: some misc patches Martin K. Petersen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).