[PATCH v5 00/15] add fixes, device quirks, error recovery,

[PATCH v5 00/15] add fixes, device quirks, error recovery,

[Yaniv Gardi]

V5:
as per comment, patch 05/14 in V4, was divided into 2 patches in V5

V4:
fixed DOS EOL symbols in 2 new files

V3:
Due to comments in V2
Removed change 02/17 of V2 
Removed change 03/17 of V2
Removed change 17/17 of V2

V2:
This series should be pushed on top of 15 patches series:
"Big fixes, retries, handle a race condition"
fixed and few comments.

V1:
This series should be pushed on top of 15 patches series:
"Big fixes, retries, handle a race condition"


Yaniv Gardi (15):
  scsi: ufs-qcom: add number of lanes per direction
  scsi: ufs: avoid spurious UFS host controller interrupts
  scsi: ufs: implement scsi host timeout handler
  scsi: ufs: verify hba controller hce reg value
  scsi: ufs: add support to read device and string descriptors
  scsi: ufs: separate device and host quirks
  scsi: ufs: disable vccq if it's not needed by UFS device
  scsi: ufs: make error handling bit faster
  scsi: ufs: add error recovery after DL NAC error
  scsi: ufs: add retry for query descriptors
  scsi: ufs: handle non spec compliant bkops behaviour by device
  scsi: ufs: tune UniPro parameters to optimize hibern8 exit time
  scsi: ufs: fix leakage during link off state
  scsi: ufs: add device quirk delay before putting UFS rails in LPM
  scsi: ufs-qcom: set PA_Local_TX_LCC_Enable before link startup

 .../devicetree/bindings/ufs/ufshcd-pltfrm.txt      |   3 +
 drivers/scsi/ufs/Makefile                          |   2 +-
 drivers/scsi/ufs/ufs-qcom.c                        | 155 ++++-
 drivers/scsi/ufs/ufs-qcom.h                        |   9 +
 drivers/scsi/ufs/ufs.h                             |  33 +
 drivers/scsi/ufs/ufs_quirks.c                      | 104 +++
 drivers/scsi/ufs/ufs_quirks.h                      | 133 ++++
 drivers/scsi/ufs/ufshcd-pltfrm.c                   |  19 +
 drivers/scsi/ufs/ufshcd.c                          | 749 ++++++++++++++++++---
 drivers/scsi/ufs/ufshcd.h                          |  40 +-
 drivers/scsi/ufs/ufshci.h                          |   4 +
 drivers/scsi/ufs/unipro.h                          |  22 +
 12 files changed, 1145 insertions(+), 128 deletions(-)
 create mode 100644 drivers/scsi/ufs/ufs_quirks.c
 create mode 100644 drivers/scsi/ufs/ufs_quirks.h

-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 01/15] scsi: ufs-qcom: add number of lanes per direction

[Yaniv Gardi]

Different platform may have different number of lanes
for the UFS link.
Add parameter to device tree specifying how many lanes
should be configured for the UFS link.

Signed-off-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 .../devicetree/bindings/ufs/ufshcd-pltfrm.txt      |  3 ++
 drivers/scsi/ufs/ufs-qcom.c                        | 39 ++++++++++++----------
 drivers/scsi/ufs/ufshcd-pltfrm.c                   | 19 +++++++++++
 drivers/scsi/ufs/ufshcd.c                          |  1 +
 drivers/scsi/ufs/ufshcd.h                          |  2 ++
 5 files changed, 47 insertions(+), 17 deletions(-)

diff --git a/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt b/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt
index 03c0e98..66f6adf 100644
--- a/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt
+++ b/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt
@@ -38,6 +38,9 @@ Optional properties:
 			  defined or a value in the array is "0" then it is assumed
 			  that the frequency is set by the parent clock or a
 			  fixed rate clock source.
+-lanes-per-direction	: number of lanes available per direction - either 1 or 2.
+			  Note that it is assume same number of lanes is used both
+			  directions at once. If not specified, default is 2 lanes per direction.
 
 Note: If above properties are not defined it can be assumed that the supply
 regulators or clocks are always on.
diff --git a/drivers/scsi/ufs/ufs-qcom.c b/drivers/scsi/ufs/ufs-qcom.c
index 4f38d00..ed57729 100644
--- a/drivers/scsi/ufs/ufs-qcom.c
+++ b/drivers/scsi/ufs/ufs-qcom.c
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2013-2015, Linux Foundation. All rights reserved.
+ * Copyright (c) 2013-2016, Linux Foundation. All rights reserved.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 and
@@ -132,21 +132,24 @@ static int ufs_qcom_enable_lane_clks(struct ufs_qcom_host *host)
 	if (err)
 		goto disable_rx_l0;
 
-	err = ufs_qcom_host_clk_enable(dev, "rx_lane1_sync_clk",
-		host->rx_l1_sync_clk);
-	if (err)
-		goto disable_tx_l0;
+	if (host->hba->lanes_per_direction > 1) {
+		err = ufs_qcom_host_clk_enable(dev, "rx_lane1_sync_clk",
+			host->rx_l1_sync_clk);
+		if (err)
+			goto disable_tx_l0;
 
-	err = ufs_qcom_host_clk_enable(dev, "tx_lane1_sync_clk",
-		host->tx_l1_sync_clk);
-	if (err)
-		goto disable_rx_l1;
+		err = ufs_qcom_host_clk_enable(dev, "tx_lane1_sync_clk",
+			host->tx_l1_sync_clk);
+		if (err)
+			goto disable_rx_l1;
+	}
 
 	host->is_lane_clks_enabled = true;
 	goto out;
 
 disable_rx_l1:
-	clk_disable_unprepare(host->rx_l1_sync_clk);
+	if (host->hba->lanes_per_direction > 1)
+		clk_disable_unprepare(host->rx_l1_sync_clk);
 disable_tx_l0:
 	clk_disable_unprepare(host->tx_l0_sync_clk);
 disable_rx_l0:
@@ -170,14 +173,16 @@ static int ufs_qcom_init_lane_clks(struct ufs_qcom_host *host)
 	if (err)
 		goto out;
 
-	err = ufs_qcom_host_clk_get(dev, "rx_lane1_sync_clk",
-		&host->rx_l1_sync_clk);
-	if (err)
-		goto out;
-
-	err = ufs_qcom_host_clk_get(dev, "tx_lane1_sync_clk",
-		&host->tx_l1_sync_clk);
+	/* In case of single lane per direction, don't read lane1 clocks */
+	if (host->hba->lanes_per_direction > 1) {
+		err = ufs_qcom_host_clk_get(dev, "rx_lane1_sync_clk",
+			&host->rx_l1_sync_clk);
+		if (err)
+			goto out;
 
+		err = ufs_qcom_host_clk_get(dev, "tx_lane1_sync_clk",
+			&host->tx_l1_sync_clk);
+	}
 out:
 	return err;
 }
diff --git a/drivers/scsi/ufs/ufshcd-pltfrm.c b/drivers/scsi/ufs/ufshcd-pltfrm.c
index d2a7b12..718f12e 100644
--- a/drivers/scsi/ufs/ufshcd-pltfrm.c
+++ b/drivers/scsi/ufs/ufshcd-pltfrm.c
@@ -40,6 +40,8 @@
 #include "ufshcd.h"
 #include "ufshcd-pltfrm.h"
 
+#define UFSHCD_DEFAULT_LANES_PER_DIRECTION		2
+
 static int ufshcd_parse_clock_info(struct ufs_hba *hba)
 {
 	int ret = 0;
@@ -277,6 +279,21 @@ void ufshcd_pltfrm_shutdown(struct platform_device *pdev)
 }
 EXPORT_SYMBOL_GPL(ufshcd_pltfrm_shutdown);
 
+static void ufshcd_init_lanes_per_dir(struct ufs_hba *hba)
+{
+	struct device *dev = hba->dev;
+	int ret;
+
+	ret = of_property_read_u32(dev->of_node, "lanes-per-direction",
+		&hba->lanes_per_direction);
+	if (ret) {
+		dev_dbg(hba->dev,
+			"%s: failed to read lanes-per-direction, ret=%d\n",
+			__func__, ret);
+		hba->lanes_per_direction = UFSHCD_DEFAULT_LANES_PER_DIRECTION;
+	}
+}
+
 /**
  * ufshcd_pltfrm_init - probe routine of the driver
  * @pdev: pointer to Platform device handle
@@ -331,6 +348,8 @@ int ufshcd_pltfrm_init(struct platform_device *pdev,
 	pm_runtime_set_active(&pdev->dev);
 	pm_runtime_enable(&pdev->dev);
 
+	ufshcd_init_lanes_per_dir(hba);
+
 	err = ufshcd_init(hba, mmio_base, irq);
 	if (err) {
 		dev_err(dev, "Initialization failed\n");
diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 9c1b94b..a8e42df 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -40,6 +40,7 @@
 #include <linux/async.h>
 #include <linux/devfreq.h>
 
+#include <linux/of.h>
 #include "ufshcd.h"
 #include "unipro.h"
 
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index e3931d0..9ae7f85 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -509,6 +509,8 @@ struct ufs_hba {
 
 	bool wlun_dev_clr_ua;
 
+	/* Number of lanes available (1 or 2) for Rx/Tx */
+	u32 lanes_per_direction;
 	struct ufs_pa_layer_attr pwr_info;
 	struct ufs_pwr_mode_info max_pwr_info;
 
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 02/15] scsi: ufs: avoid spurious UFS host controller interrupts

[Yaniv Gardi]

When control reaches to Linux UFS driver during UFS boot mode, UFS host
controller interrupt status/enable registers may have left over
settings.
In order to avoid any spurious interrupts due to these left overs,
it's important to clear these interrupt status/enable registers before
enabling UFS interrupt handling.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index a8e42df..de7280c 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -5837,6 +5837,21 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq)
 	init_waitqueue_head(&hba->dev_cmd.tag_wq);
 
 	ufshcd_init_clk_gating(hba);
+
+	/*
+	 * In order to avoid any spurious interrupt immediately after
+	 * registering UFS controller interrupt handler, clear any pending UFS
+	 * interrupt status and disable all the UFS interrupts.
+	 */
+	ufshcd_writel(hba, ufshcd_readl(hba, REG_INTERRUPT_STATUS),
+		      REG_INTERRUPT_STATUS);
+	ufshcd_writel(hba, 0, REG_INTERRUPT_ENABLE);
+	/*
+	 * Make sure that UFS interrupts are disabled and any pending interrupt
+	 * status is cleared before registering UFS interrupt handler.
+	 */
+	mb();
+
 	/* IRQ registration */
 	err = devm_request_irq(dev, irq, ufshcd_intr, IRQF_SHARED, UFSHCD, hba);
 	if (err) {
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[Yaniv Gardi]

A race condition exists between request requeueing and scsi layer
error handling:
When UFS driver queuecommand returns a busy status for a request,
it will be requeued and its tag will be freed and set to -1.
At the same time it is possible that the request will timeout and
scsi layer will start error handling for it. The scsi layer reuses
the request and its tag to send error related commands to the device,
however its tag is no longer valid.
As this request was never really sent to the device, there is no
point to start error handling with the device.
Implement the scsi error handling timeout callback and bypass SCSI
error handling for request that were not actually sent to the device.
For such requests simply reset the block layer timer. Otherwise, let
SCSI layer perform the usual error handling.

Signed-off-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 36 ++++++++++++++++++++++++++++++++++++
 1 file changed, 36 insertions(+)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index de7280c..3400ceb 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -4568,6 +4568,41 @@ static void ufshcd_async_scan(void *data, async_cookie_t cookie)
 	ufshcd_probe_hba(hba);
 }
 
+static enum blk_eh_timer_return ufshcd_eh_timed_out(struct scsi_cmnd *scmd)
+{
+	unsigned long flags;
+	struct Scsi_Host *host;
+	struct ufs_hba *hba;
+	int index;
+	bool found = false;
+
+	if (!scmd || !scmd->device || !scmd->device->host)
+		return BLK_EH_NOT_HANDLED;
+
+	host = scmd->device->host;
+	hba = shost_priv(host);
+	if (!hba)
+		return BLK_EH_NOT_HANDLED;
+
+	spin_lock_irqsave(host->host_lock, flags);
+
+	for_each_set_bit(index, &hba->outstanding_reqs, hba->nutrs) {
+		if (hba->lrb[index].cmd == scmd) {
+			found = true;
+			break;
+		}
+	}
+
+	spin_unlock_irqrestore(host->host_lock, flags);
+
+	/*
+	 * Bypass SCSI error handling and reset the block layer timer if this
+	 * SCSI command was not actually dispatched to UFS driver, otherwise
+	 * let SCSI layer handle the error as usual.
+	 */
+	return found ? BLK_EH_NOT_HANDLED : BLK_EH_RESET_TIMER;
+}
+
 static struct scsi_host_template ufshcd_driver_template = {
 	.module			= THIS_MODULE,
 	.name			= UFSHCD,
@@ -4580,6 +4615,7 @@ static struct scsi_host_template ufshcd_driver_template = {
 	.eh_abort_handler	= ufshcd_abort,
 	.eh_device_reset_handler = ufshcd_eh_device_reset_handler,
 	.eh_host_reset_handler   = ufshcd_eh_host_reset_handler,
+	.eh_timed_out		= ufshcd_eh_timed_out,
 	.this_id		= -1,
 	.sg_tablesize		= SG_ALL,
 	.cmd_per_lun		= UFSHCD_CMD_PER_LUN,
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 04/15] scsi: ufs: verify hba controller hce reg value

[Yaniv Gardi]

Sometimes due to hw issues it takes some time to the
host controller register to update. In order to verify the register
has updated, a polling is done until its value is set.

In addition the functions ufshcd_hba_stop() and
ufshcd_wait_for_register() was updated with an additional input
parameter, indicating the timeout between reads will
be done by sleeping or spinning the cpu.

Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 53 ++++++++++++++++++++++++++++-------------------
 drivers/scsi/ufs/ufshcd.h | 12 +++--------
 2 files changed, 35 insertions(+), 30 deletions(-)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 3400ceb..80031e6 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -240,11 +240,13 @@ static inline void ufshcd_disable_irq(struct ufs_hba *hba)
  * @val - wait condition
  * @interval_us - polling interval in microsecs
  * @timeout_ms - timeout in millisecs
+ * @can_sleep - perform sleep or just spin
  *
  * Returns -ETIMEDOUT on error, zero on success
  */
-static int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
-		u32 val, unsigned long interval_us, unsigned long timeout_ms)
+int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
+				u32 val, unsigned long interval_us,
+				unsigned long timeout_ms, bool can_sleep)
 {
 	int err = 0;
 	unsigned long timeout = jiffies + msecs_to_jiffies(timeout_ms);
@@ -253,9 +255,10 @@ static int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
 	val = val & mask;
 
 	while ((ufshcd_readl(hba, reg) & mask) != val) {
-		/* wakeup within 50us of expiry */
-		usleep_range(interval_us, interval_us + 50);
-
+		if (can_sleep)
+			usleep_range(interval_us, interval_us + 50);
+		else
+			udelay(interval_us);
 		if (time_after(jiffies, timeout)) {
 			if ((ufshcd_readl(hba, reg) & mask) != val)
 				err = -ETIMEDOUT;
@@ -1459,7 +1462,7 @@ ufshcd_clear_cmd(struct ufs_hba *hba, int tag)
 	 */
 	err = ufshcd_wait_for_register(hba,
 			REG_UTP_TRANSFER_REQ_DOOR_BELL,
-			mask, ~mask, 1000, 1000);
+			mask, ~mask, 1000, 1000, true);
 
 	return err;
 }
@@ -2815,6 +2818,23 @@ out:
 }
 
 /**
+ * ufshcd_hba_stop - Send controller to reset state
+ * @hba: per adapter instance
+ * @can_sleep: perform sleep or just spin
+ */
+static inline void ufshcd_hba_stop(struct ufs_hba *hba, bool can_sleep)
+{
+	int err;
+
+	ufshcd_writel(hba, CONTROLLER_DISABLE,  REG_CONTROLLER_ENABLE);
+	err = ufshcd_wait_for_register(hba, REG_CONTROLLER_ENABLE,
+					CONTROLLER_ENABLE, CONTROLLER_DISABLE,
+					10, 1, can_sleep);
+	if (err)
+		dev_err(hba->dev, "%s: Controller disable failed\n", __func__);
+}
+
+/**
  * ufshcd_hba_enable - initialize the controller
  * @hba: per adapter instance
  *
@@ -2834,18 +2854,9 @@ static int ufshcd_hba_enable(struct ufs_hba *hba)
 	 * development and testing of this driver. msleep can be changed to
 	 * mdelay and retry count can be reduced based on the controller.
 	 */
-	if (!ufshcd_is_hba_active(hba)) {
-
+	if (!ufshcd_is_hba_active(hba))
 		/* change controller state to "reset state" */
-		ufshcd_hba_stop(hba);
-
-		/*
-		 * This delay is based on the testing done with UFS host
-		 * controller FPGA. The delay can be changed based on the
-		 * host controller used.
-		 */
-		msleep(5);
-	}
+		ufshcd_hba_stop(hba, true);
 
 	/* UniPro link is disabled at this point */
 	ufshcd_set_link_off(hba);
@@ -3898,7 +3909,7 @@ static int ufshcd_clear_tm_cmd(struct ufs_hba *hba, int tag)
 	/* poll for max. 1 sec to clear door bell register by h/w */
 	err = ufshcd_wait_for_register(hba,
 			REG_UTP_TASK_REQ_DOOR_BELL,
-			mask, 0, 1000, 1000);
+			mask, 0, 1000, 1000, true);
 out:
 	return err;
 }
@@ -4180,7 +4191,7 @@ static int ufshcd_host_reset_and_restore(struct ufs_hba *hba)
 
 	/* Reset the host controller */
 	spin_lock_irqsave(hba->host->host_lock, flags);
-	ufshcd_hba_stop(hba);
+	ufshcd_hba_stop(hba, false);
 	spin_unlock_irqrestore(hba->host->host_lock, flags);
 
 	err = ufshcd_hba_enable(hba);
@@ -5133,7 +5144,7 @@ static int ufshcd_link_state_transition(struct ufs_hba *hba,
 		 * Change controller state to "reset state" which
 		 * should also put the link in off/reset state
 		 */
-		ufshcd_hba_stop(hba);
+		ufshcd_hba_stop(hba, true);
 		/*
 		 * TODO: Check if we need any delay to make sure that
 		 * controller is reset
@@ -5609,7 +5620,7 @@ void ufshcd_remove(struct ufs_hba *hba)
 	scsi_remove_host(hba->host);
 	/* disable interrupts */
 	ufshcd_disable_intr(hba, hba->intr_mask);
-	ufshcd_hba_stop(hba);
+	ufshcd_hba_stop(hba, true);
 
 	scsi_host_put(hba->host);
 
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index 9ae7f85..a6d3572 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -590,15 +590,9 @@ int ufshcd_alloc_host(struct device *, struct ufs_hba **);
 void ufshcd_dealloc_host(struct ufs_hba *);
 int ufshcd_init(struct ufs_hba * , void __iomem * , unsigned int);
 void ufshcd_remove(struct ufs_hba *);
-
-/**
- * ufshcd_hba_stop - Send controller to reset state
- * @hba: per adapter instance
- */
-static inline void ufshcd_hba_stop(struct ufs_hba *hba)
-{
-	ufshcd_writel(hba, CONTROLLER_DISABLE,  REG_CONTROLLER_ENABLE);
-}
+int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
+				u32 val, unsigned long interval_us,
+				unsigned long timeout_ms, bool can_sleep);
 
 static inline void check_upiu_size(void)
 {
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 05/15] scsi: ufs: add support to read device and string descriptors

[Yaniv Gardi]

This change adds support to read device descriptor and string descriptor
from a UFS device

Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufs.h    |  1 +
 drivers/scsi/ufs/ufshcd.c | 88 ++++++++++++++++++++++++++++++++++++++++++++++-
 drivers/scsi/ufs/ufshcd.h |  7 ++++
 3 files changed, 95 insertions(+), 1 deletion(-)

diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h
index 54a16ce..aacb235 100644
--- a/drivers/scsi/ufs/ufs.h
+++ b/drivers/scsi/ufs/ufs.h
@@ -43,6 +43,7 @@
 #define GENERAL_UPIU_REQUEST_SIZE 32
 #define QUERY_DESC_MAX_SIZE       255
 #define QUERY_DESC_MIN_SIZE       2
+#define QUERY_DESC_HDR_SIZE       2
 #define QUERY_OSF_SIZE            (GENERAL_UPIU_REQUEST_SIZE - \
 					(sizeof(struct utp_upiu_header)))
 
diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 80031e6..e2ed415 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -39,7 +39,7 @@
 
 #include <linux/async.h>
 #include <linux/devfreq.h>
-
+#include <linux/nls.h>
 #include <linux/of.h>
 #include "ufshcd.h"
 #include "unipro.h"
@@ -232,6 +232,16 @@ static inline void ufshcd_disable_irq(struct ufs_hba *hba)
 	}
 }
 
+/* replace non-printable or non-ASCII characters with spaces */
+static inline void ufshcd_remove_non_printable(char *val)
+{
+	if (!val)
+		return;
+
+	if (*val < 0x20 || *val > 0x7e)
+		*val = ' ';
+}
+
 /*
  * ufshcd_wait_for_register - wait for register value to change
  * @hba - per-adapter interface
@@ -2021,6 +2031,82 @@ static inline int ufshcd_read_power_desc(struct ufs_hba *hba,
 	return ufshcd_read_desc(hba, QUERY_DESC_IDN_POWER, 0, buf, size);
 }
 
+int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size)
+{
+	return ufshcd_read_desc(hba, QUERY_DESC_IDN_DEVICE, 0, buf, size);
+}
+EXPORT_SYMBOL(ufshcd_read_device_desc);
+
+/**
+ * ufshcd_read_string_desc - read string descriptor
+ * @hba: pointer to adapter instance
+ * @desc_index: descriptor index
+ * @buf: pointer to buffer where descriptor would be read
+ * @size: size of buf
+ * @ascii: if true convert from unicode to ascii characters
+ *
+ * Return 0 in case of success, non-zero otherwise
+ */
+int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8 *buf,
+				u32 size, bool ascii)
+{
+	int err = 0;
+
+	err = ufshcd_read_desc(hba,
+				QUERY_DESC_IDN_STRING, desc_index, buf, size);
+
+	if (err) {
+		dev_err(hba->dev, "%s: reading String Desc failed after %d retries. err = %d\n",
+			__func__, QUERY_REQ_RETRIES, err);
+		goto out;
+	}
+
+	if (ascii) {
+		int desc_len;
+		int ascii_len;
+		int i;
+		char *buff_ascii;
+
+		desc_len = buf[0];
+		/* remove header and divide by 2 to move from UTF16 to UTF8 */
+		ascii_len = (desc_len - QUERY_DESC_HDR_SIZE) / 2 + 1;
+		if (size < ascii_len + QUERY_DESC_HDR_SIZE) {
+			dev_err(hba->dev, "%s: buffer allocated size is too small\n",
+					__func__);
+			err = -ENOMEM;
+			goto out;
+		}
+
+		buff_ascii = kmalloc(ascii_len, GFP_KERNEL);
+		if (!buff_ascii) {
+			err = -ENOMEM;
+			goto out_free_buff;
+		}
+
+		/*
+		 * the descriptor contains string in UTF16 format
+		 * we need to convert to utf-8 so it can be displayed
+		 */
+		utf16s_to_utf8s((wchar_t *)&buf[QUERY_DESC_HDR_SIZE],
+				desc_len - QUERY_DESC_HDR_SIZE,
+				UTF16_BIG_ENDIAN, buff_ascii, ascii_len);
+
+		/* replace non-printable or non-ASCII characters with spaces */
+		for (i = 0; i < ascii_len; i++)
+			ufshcd_remove_non_printable(&buff_ascii[i]);
+
+		memset(buf + QUERY_DESC_HDR_SIZE, 0,
+				size - QUERY_DESC_HDR_SIZE);
+		memcpy(buf + QUERY_DESC_HDR_SIZE, buff_ascii, ascii_len);
+		buf[QUERY_DESC_LENGTH_OFFSET] = ascii_len + QUERY_DESC_HDR_SIZE;
+out_free_buff:
+		kfree(buff_ascii);
+	}
+out:
+	return err;
+}
+EXPORT_SYMBOL(ufshcd_read_string_desc);
+
 /**
  * ufshcd_read_unit_desc_param - read the specified unit descriptor parameter
  * @hba: Pointer to adapter instance
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index a6d3572..54e13cc 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -678,6 +678,13 @@ static inline int ufshcd_dme_peer_get(struct ufs_hba *hba,
 	return ufshcd_dme_get_attr(hba, attr_sel, mib_val, DME_PEER);
 }
 
+int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size);
+
+#define ASCII_STD true
+
+int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8 *buf,
+				u32 size, bool ascii);
+
 /* Expose Query-Request API */
 int ufshcd_query_flag(struct ufs_hba *hba, enum query_opcode opcode,
 	enum flag_idn idn, bool *flag_res);
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 06/15] scsi: ufs: separate device and host quirks

[Yaniv Gardi]

Currently we use the host quirks mechanism in order to
handle both device and host controller quirks.
In order to support various of UFS devices we should separate
handling the device quirks from the host controller's.

Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/Makefile     |   2 +-
 drivers/scsi/ufs/ufs.h        |  31 +++++++++++
 drivers/scsi/ufs/ufs_quirks.c | 100 ++++++++++++++++++++++++++++++++++
 drivers/scsi/ufs/ufs_quirks.h | 124 ++++++++++++++++++++++++++++++++++++++++++
 drivers/scsi/ufs/ufshcd.c     |   2 +
 drivers/scsi/ufs/ufshcd.h     |   3 +
 6 files changed, 261 insertions(+), 1 deletion(-)
 create mode 100644 drivers/scsi/ufs/ufs_quirks.c
 create mode 100644 drivers/scsi/ufs/ufs_quirks.h

diff --git a/drivers/scsi/ufs/Makefile b/drivers/scsi/ufs/Makefile
index 8303bcc..8570d41 100644
--- a/drivers/scsi/ufs/Makefile
+++ b/drivers/scsi/ufs/Makefile
@@ -1,5 +1,5 @@
 # UFSHCD makefile
 obj-$(CONFIG_SCSI_UFS_QCOM) += ufs-qcom.o
-obj-$(CONFIG_SCSI_UFSHCD) += ufshcd.o
+obj-$(CONFIG_SCSI_UFSHCD) += ufshcd.o ufs_quirks.o
 obj-$(CONFIG_SCSI_UFSHCD_PCI) += ufshcd-pci.o
 obj-$(CONFIG_SCSI_UFSHCD_PLATFORM) += ufshcd-pltfrm.o
diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h
index aacb235..8ec6356 100644
--- a/drivers/scsi/ufs/ufs.h
+++ b/drivers/scsi/ufs/ufs.h
@@ -196,6 +196,37 @@ enum unit_desc_param {
 	UNIT_DESC_PARAM_LARGE_UNIT_SIZE_M1	= 0x22,
 };
 
+/* Device descriptor parameters offsets in bytes*/
+enum device_desc_param {
+	DEVICE_DESC_PARAM_LEN			= 0x0,
+	DEVICE_DESC_PARAM_TYPE			= 0x1,
+	DEVICE_DESC_PARAM_DEVICE_TYPE		= 0x2,
+	DEVICE_DESC_PARAM_DEVICE_CLASS		= 0x3,
+	DEVICE_DESC_PARAM_DEVICE_SUB_CLASS	= 0x4,
+	DEVICE_DESC_PARAM_PRTCL			= 0x5,
+	DEVICE_DESC_PARAM_NUM_LU		= 0x6,
+	DEVICE_DESC_PARAM_NUM_WLU		= 0x7,
+	DEVICE_DESC_PARAM_BOOT_ENBL		= 0x8,
+	DEVICE_DESC_PARAM_DESC_ACCSS_ENBL	= 0x9,
+	DEVICE_DESC_PARAM_INIT_PWR_MODE		= 0xA,
+	DEVICE_DESC_PARAM_HIGH_PR_LUN		= 0xB,
+	DEVICE_DESC_PARAM_SEC_RMV_TYPE		= 0xC,
+	DEVICE_DESC_PARAM_SEC_LU		= 0xD,
+	DEVICE_DESC_PARAM_BKOP_TERM_LT		= 0xE,
+	DEVICE_DESC_PARAM_ACTVE_ICC_LVL		= 0xF,
+	DEVICE_DESC_PARAM_SPEC_VER		= 0x10,
+	DEVICE_DESC_PARAM_MANF_DATE		= 0x12,
+	DEVICE_DESC_PARAM_MANF_NAME		= 0x14,
+	DEVICE_DESC_PARAM_PRDCT_NAME		= 0x15,
+	DEVICE_DESC_PARAM_SN			= 0x16,
+	DEVICE_DESC_PARAM_OEM_ID		= 0x17,
+	DEVICE_DESC_PARAM_MANF_ID		= 0x18,
+	DEVICE_DESC_PARAM_UD_OFFSET		= 0x1A,
+	DEVICE_DESC_PARAM_UD_LEN		= 0x1B,
+	DEVICE_DESC_PARAM_RTT_CAP		= 0x1C,
+	DEVICE_DESC_PARAM_FRQ_RTC		= 0x1D,
+};
+
 /*
  * Logical Unit Write Protect
  * 00h: LU not write protected
diff --git a/drivers/scsi/ufs/ufs_quirks.c b/drivers/scsi/ufs/ufs_quirks.c
new file mode 100644
index 0000000..476ed01
--- /dev/null
+++ b/drivers/scsi/ufs/ufs_quirks.c
@@ -0,0 +1,100 @@
+/*
+ * Copyright (c) 2013-2016, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include "ufshcd.h"
+#include "ufs_quirks.h"
+
+static struct ufs_dev_fix ufs_fixups[] = {
+	/* UFS cards deviations table */
+	UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL, UFS_DEVICE_NO_VCCQ),
+	UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL,
+		UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS),
+	UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL,
+		UFS_DEVICE_NO_FASTAUTO),
+	UFS_FIX(UFS_VENDOR_TOSHIBA, "THGLF2G9C8KBADG",
+		UFS_DEVICE_QUIRK_PA_TACTIVATE),
+	UFS_FIX(UFS_VENDOR_TOSHIBA, "THGLF2G9D8KBADG",
+		UFS_DEVICE_QUIRK_PA_TACTIVATE),
+
+	END_FIX
+};
+
+static int ufs_get_device_info(struct ufs_hba *hba,
+				struct ufs_device_info *card_data)
+{
+	int err;
+	u8 model_index;
+	u8 str_desc_buf[QUERY_DESC_STRING_MAX_SIZE + 1] = {0};
+	u8 desc_buf[QUERY_DESC_DEVICE_MAX_SIZE];
+
+	err = ufshcd_read_device_desc(hba, desc_buf,
+					QUERY_DESC_DEVICE_MAX_SIZE);
+	if (err) {
+		dev_err(hba->dev, "%s: Failed reading Device Desc. err = %d\n",
+			__func__, err);
+		goto out;
+	}
+
+	/*
+	 * getting vendor (manufacturerID) and Bank Index in big endian
+	 * format
+	 */
+	card_data->wmanufacturerid = desc_buf[DEVICE_DESC_PARAM_MANF_ID] << 8 |
+				     desc_buf[DEVICE_DESC_PARAM_MANF_ID + 1];
+
+	model_index = desc_buf[DEVICE_DESC_PARAM_PRDCT_NAME];
+
+	err = ufshcd_read_string_desc(hba, model_index, str_desc_buf,
+					QUERY_DESC_STRING_MAX_SIZE, ASCII_STD);
+	if (err) {
+		dev_err(hba->dev, "%s: Failed reading Product Name. err = %d\n",
+			__func__, err);
+		goto out;
+	}
+
+	str_desc_buf[QUERY_DESC_STRING_MAX_SIZE] = '\0';
+	strlcpy(card_data->model, (str_desc_buf + QUERY_DESC_HDR_SIZE),
+		min_t(u8, str_desc_buf[QUERY_DESC_LENGTH_OFFSET],
+		      MAX_MODEL_LEN));
+
+	/* Null terminate the model string */
+	card_data->model[MAX_MODEL_LEN] = '\0';
+
+out:
+	return err;
+}
+
+void ufs_advertise_fixup_device(struct ufs_hba *hba)
+{
+	int err;
+	struct ufs_dev_fix *f;
+	struct ufs_device_info card_data;
+
+	card_data.wmanufacturerid = 0;
+
+	err = ufs_get_device_info(hba, &card_data);
+	if (err) {
+		dev_err(hba->dev, "%s: Failed getting device info. err = %d\n",
+			__func__, err);
+		return;
+	}
+
+	for (f = ufs_fixups; f->quirk; f++) {
+		if (((f->card.wmanufacturerid == card_data.wmanufacturerid) ||
+		    (f->card.wmanufacturerid == UFS_ANY_VENDOR)) &&
+		    (STR_PRFX_EQUAL(f->card.model, card_data.model) ||
+		     !strcmp(f->card.model, UFS_ANY_MODEL)))
+			hba->dev_quirks |= f->quirk;
+	}
+}
+EXPORT_SYMBOL(ufs_advertise_fixup_device);
diff --git a/drivers/scsi/ufs/ufs_quirks.h b/drivers/scsi/ufs/ufs_quirks.h
new file mode 100644
index 0000000..01ab166
--- /dev/null
+++ b/drivers/scsi/ufs/ufs_quirks.h
@@ -0,0 +1,124 @@
+/*
+ * Copyright (c) 2014-2016, The Linux Foundation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 and
+ * only version 2 as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ */
+
+#ifndef _UFS_QUIRKS_H_
+#define _UFS_QUIRKS_H_
+
+/* return true if s1 is a prefix of s2 */
+#define STR_PRFX_EQUAL(s1, s2) !strncmp(s1, s2, strlen(s1))
+
+#define UFS_ANY_VENDOR -1
+#define UFS_ANY_MODEL  "ANY_MODEL"
+
+#define MAX_MODEL_LEN 16
+
+#define UFS_VENDOR_TOSHIBA     0x198
+#define UFS_VENDOR_SAMSUNG     0x1CE
+
+/**
+ * ufs_device_info - ufs device details
+ * @wmanufacturerid: card details
+ * @model: card model
+ */
+struct ufs_device_info {
+	u16 wmanufacturerid;
+	char model[MAX_MODEL_LEN + 1];
+};
+
+/**
+ * ufs_dev_fix - ufs device quirk info
+ * @card: ufs card details
+ * @quirk: device quirk
+ */
+struct ufs_dev_fix {
+	struct ufs_device_info card;
+	unsigned int quirk;
+};
+
+#define END_FIX { { 0 }, 0 }
+
+/* add specific device quirk */
+#define UFS_FIX(_vendor, _model, _quirk) \
+		{					  \
+			.card.wmanufacturerid = (_vendor),\
+			.card.model = (_model),		  \
+			.quirk = (_quirk),		  \
+		}
+
+/*
+ * If UFS device is having issue in processing LCC (Line Control
+ * Command) coming from UFS host controller then enable this quirk.
+ * When this quirk is enabled, host controller driver should disable
+ * the LCC transmission on UFS host controller (by clearing
+ * TX_LCC_ENABLE attribute of host to 0).
+ */
+#define UFS_DEVICE_QUIRK_BROKEN_LCC (1 << 0)
+
+/*
+ * Some UFS devices don't need VCCQ rail for device operations. Enabling this
+ * quirk for such devices will make sure that VCCQ rail is not voted.
+ */
+#define UFS_DEVICE_NO_VCCQ (1 << 1)
+
+/*
+ * Some vendor's UFS device sends back to back NACs for the DL data frames
+ * causing the host controller to raise the DFES error status. Sometimes
+ * such UFS devices send back to back NAC without waiting for new
+ * retransmitted DL frame from the host and in such cases it might be possible
+ * the Host UniPro goes into bad state without raising the DFES error
+ * interrupt. If this happens then all the pending commands would timeout
+ * only after respective SW command (which is generally too large).
+ *
+ * We can workaround such device behaviour like this:
+ * - As soon as SW sees the DL NAC error, it should schedule the error handler
+ * - Error handler would sleep for 50ms to see if there are any fatal errors
+ *   raised by UFS controller.
+ *    - If there are fatal errors then SW does normal error recovery.
+ *    - If there are no fatal errors then SW sends the NOP command to device
+ *      to check if link is alive.
+ *        - If NOP command times out, SW does normal error recovery
+ *        - If NOP command succeed, skip the error handling.
+ *
+ * If DL NAC error is seen multiple times with some vendor's UFS devices then
+ * enable this quirk to initiate quick error recovery and also silence related
+ * error logs to reduce spamming of kernel logs.
+ */
+#define UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS (1 << 2)
+
+/*
+ * Some UFS devices may not work properly after resume if the link was kept
+ * in off state during suspend. Enabling this quirk will not allow the
+ * link to be kept in off state during suspend.
+ */
+#define UFS_DEVICE_QUIRK_NO_LINK_OFF	(1 << 3)
+
+/*
+ * Few Toshiba UFS device models advertise RX_MIN_ACTIVATETIME_CAPABILITY as
+ * 600us which may not be enough for reliable hibern8 exit hardware sequence
+ * from UFS device.
+ * To workaround this issue, host should set its PA_TACTIVATE time to 1ms even
+ * if device advertises RX_MIN_ACTIVATETIME_CAPABILITY less than 1ms.
+ */
+#define UFS_DEVICE_QUIRK_PA_TACTIVATE	(1 << 4)
+
+/*
+ * Some UFS memory devices may have really low read/write throughput in
+ * FAST AUTO mode, enable this quirk to make sure that FAST AUTO mode is
+ * never enabled for such devices.
+ */
+#define UFS_DEVICE_NO_FASTAUTO		(1 << 5)
+
+struct ufs_hba;
+void ufs_advertise_fixup_device(struct ufs_hba *hba);
+#endif /* UFS_QUIRKS_H_ */
diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index e2ed415..91a341b 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -42,6 +42,7 @@
 #include <linux/nls.h>
 #include <linux/of.h>
 #include "ufshcd.h"
+#include "ufs_quirks.h"
 #include "unipro.h"
 
 #define UFSHCD_ENABLE_INTRS	(UTP_TRANSFER_REQ_COMPL |\
@@ -4591,6 +4592,7 @@ static int ufshcd_probe_hba(struct ufs_hba *hba)
 	if (ret)
 		goto out;
 
+	ufs_advertise_fixup_device(hba);
 	/* UFS device is also active now */
 	ufshcd_set_ufs_dev_active(hba);
 	ufshcd_force_reset_auto_bkops(hba);
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index 54e13cc..77e79c0 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -470,6 +470,9 @@ struct ufs_hba {
 
 	unsigned int quirks;	/* Deviations from standard UFSHCI spec. */
 
+	/* Device deviations from standard UFS device spec. */
+	unsigned int dev_quirks;
+
 	wait_queue_head_t tm_wq;
 	wait_queue_head_t tm_tag_wq;
 	unsigned long tm_condition;
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 07/15] scsi: ufs: disable vccq if it's not needed by UFS device

[Yaniv Gardi]

Some UFS devices don't require VCCQ rail for device operations hence
this change adds support to recognize such devices and remove vote for
the unused VCCQ rail.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufs.h    |  1 +
 drivers/scsi/ufs/ufshcd.c | 60 +++++++++++++++++++++++++++++++++++++++++++----
 2 files changed, 57 insertions(+), 4 deletions(-)

diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h
index 8ec6356..b291fa6 100644
--- a/drivers/scsi/ufs/ufs.h
+++ b/drivers/scsi/ufs/ufs.h
@@ -501,6 +501,7 @@ struct ufs_vreg {
 	struct regulator *reg;
 	const char *name;
 	bool enabled;
+	bool unused;
 	int min_uV;
 	int max_uV;
 	int min_uA;
diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 91a341b..987cf27 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -195,6 +195,7 @@ static int ufshcd_probe_hba(struct ufs_hba *hba);
 static int __ufshcd_setup_clocks(struct ufs_hba *hba, bool on,
 				 bool skip_ref_clk);
 static int ufshcd_setup_clocks(struct ufs_hba *hba, bool on);
+static int ufshcd_set_vccq_rail_unused(struct ufs_hba *hba, bool unused);
 static int ufshcd_uic_hibern8_exit(struct ufs_hba *hba);
 static int ufshcd_uic_hibern8_enter(struct ufs_hba *hba);
 static inline void ufshcd_add_delay_before_dme_cmd(struct ufs_hba *hba);
@@ -4593,6 +4594,12 @@ static int ufshcd_probe_hba(struct ufs_hba *hba)
 		goto out;
 
 	ufs_advertise_fixup_device(hba);
+
+	ret = ufshcd_set_vccq_rail_unused(hba,
+		(hba->dev_quirks & UFS_DEVICE_NO_VCCQ) ? true : false);
+	if (ret)
+		goto out;
+
 	/* UFS device is also active now */
 	ufshcd_set_ufs_dev_active(hba);
 	ufshcd_force_reset_auto_bkops(hba);
@@ -4743,13 +4750,24 @@ static int ufshcd_config_vreg_load(struct device *dev, struct ufs_vreg *vreg,
 static inline int ufshcd_config_vreg_lpm(struct ufs_hba *hba,
 					 struct ufs_vreg *vreg)
 {
-	return ufshcd_config_vreg_load(hba->dev, vreg, UFS_VREG_LPM_LOAD_UA);
+	if (!vreg)
+		return 0;
+	else if (vreg->unused)
+		return 0;
+	else
+		return ufshcd_config_vreg_load(hba->dev, vreg,
+					       UFS_VREG_LPM_LOAD_UA);
 }
 
 static inline int ufshcd_config_vreg_hpm(struct ufs_hba *hba,
 					 struct ufs_vreg *vreg)
 {
-	return ufshcd_config_vreg_load(hba->dev, vreg, vreg->max_uA);
+	if (!vreg)
+		return 0;
+	else if (vreg->unused)
+		return 0;
+	else
+		return ufshcd_config_vreg_load(hba->dev, vreg, vreg->max_uA);
 }
 
 static int ufshcd_config_vreg(struct device *dev,
@@ -4784,7 +4802,9 @@ static int ufshcd_enable_vreg(struct device *dev, struct ufs_vreg *vreg)
 {
 	int ret = 0;
 
-	if (!vreg || vreg->enabled)
+	if (!vreg)
+		goto out;
+	else if (vreg->enabled || vreg->unused)
 		goto out;
 
 	ret = ufshcd_config_vreg(dev, vreg, true);
@@ -4804,7 +4824,9 @@ static int ufshcd_disable_vreg(struct device *dev, struct ufs_vreg *vreg)
 {
 	int ret = 0;
 
-	if (!vreg || !vreg->enabled)
+	if (!vreg)
+		goto out;
+	else if (!vreg->enabled || vreg->unused)
 		goto out;
 
 	ret = regulator_disable(vreg->reg);
@@ -4910,6 +4932,36 @@ static int ufshcd_init_hba_vreg(struct ufs_hba *hba)
 	return 0;
 }
 
+static int ufshcd_set_vccq_rail_unused(struct ufs_hba *hba, bool unused)
+{
+	int ret = 0;
+	struct ufs_vreg_info *info = &hba->vreg_info;
+
+	if (!info)
+		goto out;
+	else if (!info->vccq)
+		goto out;
+
+	if (unused) {
+		/* shut off the rail here */
+		ret = ufshcd_toggle_vreg(hba->dev, info->vccq, false);
+		/*
+		 * Mark this rail as no longer used, so it doesn't get enabled
+		 * later by mistake
+		 */
+		if (!ret)
+			info->vccq->unused = true;
+	} else {
+		/*
+		 * rail should have been already enabled hence just make sure
+		 * that unused flag is cleared.
+		 */
+		info->vccq->unused = false;
+	}
+out:
+	return ret;
+}
+
 static int __ufshcd_setup_clocks(struct ufs_hba *hba, bool on,
 					bool skip_ref_clk)
 {
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 08/15] scsi: ufs: make error handling bit faster

[Yaniv Gardi]

UFS driver's error handler forcefully tries to clear all the pending
requests. For each pending request in the queue, it waits 1 sec for it
to get cleared. If we have multiple requests in the queue then it's
possible that we might end up waiting for those many seconds before
resetting the host. But note that resetting host would any way clear
all the pending requests from the hardware. Hence this change skips
the forceful clear of the pending requests if we are anyway going to
reset the host (for fatal errors).

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 155 +++++++++++++++++++++++++++++++++-------------
 1 file changed, 112 insertions(+), 43 deletions(-)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 987cf27..dc096f1 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -133,9 +133,11 @@ enum {
 /* UFSHCD UIC layer error flags */
 enum {
 	UFSHCD_UIC_DL_PA_INIT_ERROR = (1 << 0), /* Data link layer error */
-	UFSHCD_UIC_NL_ERROR = (1 << 1), /* Network layer error */
-	UFSHCD_UIC_TL_ERROR = (1 << 2), /* Transport Layer error */
-	UFSHCD_UIC_DME_ERROR = (1 << 3), /* DME error */
+	UFSHCD_UIC_DL_NAC_RECEIVED_ERROR = (1 << 1), /* Data link layer error */
+	UFSHCD_UIC_DL_TCx_REPLAY_ERROR = (1 << 2), /* Data link layer error */
+	UFSHCD_UIC_NL_ERROR = (1 << 3), /* Network layer error */
+	UFSHCD_UIC_TL_ERROR = (1 << 4), /* Transport Layer error */
+	UFSHCD_UIC_DME_ERROR = (1 << 5), /* DME error */
 };
 
 /* Interrupt configuration options */
@@ -3465,31 +3467,18 @@ static void ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 intr_status)
 }
 
 /**
- * ufshcd_transfer_req_compl - handle SCSI and query command completion
+ * __ufshcd_transfer_req_compl - handle SCSI and query command completion
  * @hba: per adapter instance
+ * @completed_reqs: requests to complete
  */
-static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
+static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
+					unsigned long completed_reqs)
 {
 	struct ufshcd_lrb *lrbp;
 	struct scsi_cmnd *cmd;
-	unsigned long completed_reqs;
-	u32 tr_doorbell;
 	int result;
 	int index;
 
-	/* Resetting interrupt aggregation counters first and reading the
-	 * DOOR_BELL afterward allows us to handle all the completed requests.
-	 * In order to prevent other interrupts starvation the DB is read once
-	 * after reset. The down side of this solution is the possibility of
-	 * false interrupt if device completes another request after resetting
-	 * aggregation and before reading the DB.
-	 */
-	if (ufshcd_is_intr_aggr_allowed(hba))
-		ufshcd_reset_intr_aggr(hba);
-
-	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
-	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
-
 	for_each_set_bit(index, &completed_reqs, hba->nutrs) {
 		lrbp = &hba->lrb[index];
 		cmd = lrbp->cmd;
@@ -3519,6 +3508,31 @@ static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
 }
 
 /**
+ * ufshcd_transfer_req_compl - handle SCSI and query command completion
+ * @hba: per adapter instance
+ */
+static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
+{
+	unsigned long completed_reqs;
+	u32 tr_doorbell;
+
+	/* Resetting interrupt aggregation counters first and reading the
+	 * DOOR_BELL afterward allows us to handle all the completed requests.
+	 * In order to prevent other interrupts starvation the DB is read once
+	 * after reset. The down side of this solution is the possibility of
+	 * false interrupt if device completes another request after resetting
+	 * aggregation and before reading the DB.
+	 */
+	if (ufshcd_is_intr_aggr_allowed(hba))
+		ufshcd_reset_intr_aggr(hba);
+
+	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
+	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
+
+	__ufshcd_transfer_req_compl(hba, completed_reqs);
+}
+
+/**
  * ufshcd_disable_ee - disable exception event
  * @hba: per-adapter instance
  * @mask: exception event to disable
@@ -3773,6 +3787,13 @@ out:
 	return;
 }
 
+/* Complete requests that have door-bell cleared */
+static void ufshcd_complete_requests(struct ufs_hba *hba)
+{
+	ufshcd_transfer_req_compl(hba);
+	ufshcd_tmc_handler(hba);
+}
+
 /**
  * ufshcd_err_handler - handle UFS errors that require s/w attention
  * @work: pointer to work structure
@@ -3785,6 +3806,7 @@ static void ufshcd_err_handler(struct work_struct *work)
 	u32 err_tm = 0;
 	int err = 0;
 	int tag;
+	bool needs_reset = false;
 
 	hba = container_of(work, struct ufs_hba, eh_work);
 
@@ -3792,40 +3814,75 @@ static void ufshcd_err_handler(struct work_struct *work)
 	ufshcd_hold(hba, false);
 
 	spin_lock_irqsave(hba->host->host_lock, flags);
-	if (hba->ufshcd_state == UFSHCD_STATE_RESET) {
-		spin_unlock_irqrestore(hba->host->host_lock, flags);
+	if (hba->ufshcd_state == UFSHCD_STATE_RESET)
 		goto out;
-	}
 
 	hba->ufshcd_state = UFSHCD_STATE_RESET;
 	ufshcd_set_eh_in_progress(hba);
 
 	/* Complete requests that have door-bell cleared by h/w */
-	ufshcd_transfer_req_compl(hba);
-	ufshcd_tmc_handler(hba);
-	spin_unlock_irqrestore(hba->host->host_lock, flags);
+	ufshcd_complete_requests(hba);
+	if ((hba->saved_err & INT_FATAL_ERRORS) ||
+	    ((hba->saved_err & UIC_ERROR) &&
+	    (hba->saved_uic_err & (UFSHCD_UIC_DL_PA_INIT_ERROR |
+				   UFSHCD_UIC_DL_NAC_RECEIVED_ERROR |
+				   UFSHCD_UIC_DL_TCx_REPLAY_ERROR))))
+		needs_reset = true;
 
+	/*
+	 * if host reset is required then skip clearing the pending
+	 * transfers forcefully because they will automatically get
+	 * cleared after link startup.
+	 */
+	if (needs_reset)
+		goto skip_pending_xfer_clear;
+
+	/* release lock as clear command might sleep */
+	spin_unlock_irqrestore(hba->host->host_lock, flags);
 	/* Clear pending transfer requests */
-	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs)
-		if (ufshcd_clear_cmd(hba, tag))
-			err_xfer |= 1 << tag;
+	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs) {
+		if (ufshcd_clear_cmd(hba, tag)) {
+			err_xfer = true;
+			goto lock_skip_pending_xfer_clear;
+		}
+	}
 
 	/* Clear pending task management requests */
-	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs)
-		if (ufshcd_clear_tm_cmd(hba, tag))
-			err_tm |= 1 << tag;
+	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs) {
+		if (ufshcd_clear_tm_cmd(hba, tag)) {
+			err_tm = true;
+			goto lock_skip_pending_xfer_clear;
+		}
+	}
 
-	/* Complete the requests that are cleared by s/w */
+lock_skip_pending_xfer_clear:
 	spin_lock_irqsave(hba->host->host_lock, flags);
-	ufshcd_transfer_req_compl(hba);
-	ufshcd_tmc_handler(hba);
-	spin_unlock_irqrestore(hba->host->host_lock, flags);
 
+	/* Complete the requests that are cleared by s/w */
+	ufshcd_complete_requests(hba);
+
+	if (err_xfer || err_tm)
+		needs_reset = true;
+
+skip_pending_xfer_clear:
 	/* Fatal errors need reset */
-	if (err_xfer || err_tm || (hba->saved_err & INT_FATAL_ERRORS) ||
-			((hba->saved_err & UIC_ERROR) &&
-			 (hba->saved_uic_err & UFSHCD_UIC_DL_PA_INIT_ERROR))) {
+	if (needs_reset) {
+		unsigned long max_doorbells = (1UL << hba->nutrs) - 1;
+
+		/*
+		 * ufshcd_reset_and_restore() does the link reinitialization
+		 * which will need atleast one empty doorbell slot to send the
+		 * device management commands (NOP and query commands).
+		 * If there is no slot empty at this moment then free up last
+		 * slot forcefully.
+		 */
+		if (hba->outstanding_reqs == max_doorbells)
+			__ufshcd_transfer_req_compl(hba,
+						    (1UL << (hba->nutrs - 1)));
+
+		spin_unlock_irqrestore(hba->host->host_lock, flags);
 		err = ufshcd_reset_and_restore(hba);
+		spin_lock_irqsave(hba->host->host_lock, flags);
 		if (err) {
 			dev_err(hba->dev, "%s: reset and restore failed\n",
 					__func__);
@@ -3839,9 +3896,18 @@ static void ufshcd_err_handler(struct work_struct *work)
 		hba->saved_err = 0;
 		hba->saved_uic_err = 0;
 	}
+
+	if (!needs_reset) {
+		hba->ufshcd_state = UFSHCD_STATE_OPERATIONAL;
+		if (hba->saved_err || hba->saved_uic_err)
+			dev_err_ratelimited(hba->dev, "%s: exit: saved_err 0x%x saved_uic_err 0x%x",
+			    __func__, hba->saved_err, hba->saved_uic_err);
+	}
+
 	ufshcd_clear_eh_in_progress(hba);
 
 out:
+	spin_unlock_irqrestore(hba->host->host_lock, flags);
 	scsi_unblock_requests(hba->host);
 	ufshcd_release(hba);
 	pm_runtime_put_sync(hba->dev);
@@ -3896,15 +3962,18 @@ static void ufshcd_check_errors(struct ufs_hba *hba)
 	}
 
 	if (queue_eh_work) {
+		/*
+		 * update the transfer error masks to sticky bits, let's do this
+		 * irrespective of current ufshcd_state.
+		 */
+		hba->saved_err |= hba->errors;
+		hba->saved_uic_err |= hba->uic_error;
+
 		/* handle fatal errors only when link is functional */
 		if (hba->ufshcd_state == UFSHCD_STATE_OPERATIONAL) {
 			/* block commands from scsi mid-layer */
 			scsi_block_requests(hba->host);
 
-			/* transfer error masks to sticky bits */
-			hba->saved_err |= hba->errors;
-			hba->saved_uic_err |= hba->uic_error;
-
 			hba->ufshcd_state = UFSHCD_STATE_ERROR;
 			schedule_work(&hba->eh_work);
 		}
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 09/15] scsi: ufs: add error recovery after DL NAC error

[Yaniv Gardi]

Some vendor's UFS device sends back to back NACs for the DL data frames
causing the host controller to raise the DFES error status. Sometimes
such UFS devices send back to back NAC without waiting for new
retransmitted DL frame from the host and in such cases it might be
possible the Host UniPro goes into bad state without raising the DFES
error interrupt. If this happens then all the pending commands would
timeout only after respective SW command (which is generally too
large).

This change workarounds such device behaviour like this:
- As soon as SW sees the DL NAC error, it would schedule the error
  handler
- Error handler would sleep for 50ms to see if there any fatal errors
  raised by UFS controller.
   - If there are fatal errors then SW does normal error recovery.
   - If there are no fatal errors then SW sends the NOP command to
     device to check if link is alive.
       - If NOP command times out, SW does normal error recovery
       - If NOP command succeed, skip the error handling.

If DL NAC error is seen multiple times with some vendor's UFS devices
then enable this quirk to initiate quick error recovery and also
silence related error logs to reduce spamming of kernel logs.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 93 +++++++++++++++++++++++++++++++++++++++++++++++
 drivers/scsi/ufs/ufshci.h |  2 +
 2 files changed, 95 insertions(+)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index dc096f1..e9c7c91 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -3795,6 +3795,79 @@ static void ufshcd_complete_requests(struct ufs_hba *hba)
 }
 
 /**
+ * ufshcd_quirk_dl_nac_errors - This function checks if error handling is
+ *				to recover from the DL NAC errors or not.
+ * @hba: per-adapter instance
+ *
+ * Returns true if error handling is required, false otherwise
+ */
+static bool ufshcd_quirk_dl_nac_errors(struct ufs_hba *hba)
+{
+	unsigned long flags;
+	bool err_handling = true;
+
+	spin_lock_irqsave(hba->host->host_lock, flags);
+	/*
+	 * UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS only workaround the
+	 * device fatal error and/or DL NAC & REPLAY timeout errors.
+	 */
+	if (hba->saved_err & (CONTROLLER_FATAL_ERROR | SYSTEM_BUS_FATAL_ERROR))
+		goto out;
+
+	if ((hba->saved_err & DEVICE_FATAL_ERROR) ||
+	    ((hba->saved_err & UIC_ERROR) &&
+	     (hba->saved_uic_err & UFSHCD_UIC_DL_TCx_REPLAY_ERROR)))
+		goto out;
+
+	if ((hba->saved_err & UIC_ERROR) &&
+	    (hba->saved_uic_err & UFSHCD_UIC_DL_NAC_RECEIVED_ERROR)) {
+		int err;
+		/*
+		 * wait for 50ms to see if we can get any other errors or not.
+		 */
+		spin_unlock_irqrestore(hba->host->host_lock, flags);
+		msleep(50);
+		spin_lock_irqsave(hba->host->host_lock, flags);
+
+		/*
+		 * now check if we have got any other severe errors other than
+		 * DL NAC error?
+		 */
+		if ((hba->saved_err & INT_FATAL_ERRORS) ||
+		    ((hba->saved_err & UIC_ERROR) &&
+		    (hba->saved_uic_err & ~UFSHCD_UIC_DL_NAC_RECEIVED_ERROR)))
+			goto out;
+
+		/*
+		 * As DL NAC is the only error received so far, send out NOP
+		 * command to confirm if link is still active or not.
+		 *   - If we don't get any response then do error recovery.
+		 *   - If we get response then clear the DL NAC error bit.
+		 */
+
+		spin_unlock_irqrestore(hba->host->host_lock, flags);
+		err = ufshcd_verify_dev_init(hba);
+		spin_lock_irqsave(hba->host->host_lock, flags);
+
+		if (err)
+			goto out;
+
+		/* Link seems to be alive hence ignore the DL NAC errors */
+		if (hba->saved_uic_err == UFSHCD_UIC_DL_NAC_RECEIVED_ERROR)
+			hba->saved_err &= ~UIC_ERROR;
+		/* clear NAC error */
+		hba->saved_uic_err &= ~UFSHCD_UIC_DL_NAC_RECEIVED_ERROR;
+		if (!hba->saved_uic_err) {
+			err_handling = false;
+			goto out;
+		}
+	}
+out:
+	spin_unlock_irqrestore(hba->host->host_lock, flags);
+	return err_handling;
+}
+
+/**
  * ufshcd_err_handler - handle UFS errors that require s/w attention
  * @work: pointer to work structure
  */
@@ -3822,6 +3895,17 @@ static void ufshcd_err_handler(struct work_struct *work)
 
 	/* Complete requests that have door-bell cleared by h/w */
 	ufshcd_complete_requests(hba);
+
+	if (hba->dev_quirks & UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS) {
+		bool ret;
+
+		spin_unlock_irqrestore(hba->host->host_lock, flags);
+		/* release the lock as ufshcd_quirk_dl_nac_errors() may sleep */
+		ret = ufshcd_quirk_dl_nac_errors(hba);
+		spin_lock_irqsave(hba->host->host_lock, flags);
+		if (!ret)
+			goto skip_err_handling;
+	}
 	if ((hba->saved_err & INT_FATAL_ERRORS) ||
 	    ((hba->saved_err & UIC_ERROR) &&
 	    (hba->saved_uic_err & (UFSHCD_UIC_DL_PA_INIT_ERROR |
@@ -3897,6 +3981,7 @@ skip_pending_xfer_clear:
 		hba->saved_uic_err = 0;
 	}
 
+skip_err_handling:
 	if (!needs_reset) {
 		hba->ufshcd_state = UFSHCD_STATE_OPERATIONAL;
 		if (hba->saved_err || hba->saved_uic_err)
@@ -3925,6 +4010,14 @@ static void ufshcd_update_uic_error(struct ufs_hba *hba)
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_DATA_LINK_LAYER);
 	if (reg & UIC_DATA_LINK_LAYER_ERROR_PA_INIT)
 		hba->uic_error |= UFSHCD_UIC_DL_PA_INIT_ERROR;
+	else if (hba->dev_quirks &
+		   UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS) {
+		if (reg & UIC_DATA_LINK_LAYER_ERROR_NAC_RECEIVED)
+			hba->uic_error |=
+				UFSHCD_UIC_DL_NAC_RECEIVED_ERROR;
+		else if (reg & UIC_DATA_LINK_LAYER_ERROR_TCx_REPLAY_TIMEOUT)
+			hba->uic_error |= UFSHCD_UIC_DL_TCx_REPLAY_ERROR;
+	}
 
 	/* UIC NL/TL/DME errors needs software retry */
 	reg = ufshcd_readl(hba, REG_UIC_ERROR_CODE_NETWORK_LAYER);
diff --git a/drivers/scsi/ufs/ufshci.h b/drivers/scsi/ufs/ufshci.h
index 0ae0967..2b05bfb 100644
--- a/drivers/scsi/ufs/ufshci.h
+++ b/drivers/scsi/ufs/ufshci.h
@@ -170,6 +170,8 @@ enum {
 #define UIC_DATA_LINK_LAYER_ERROR		UFS_BIT(31)
 #define UIC_DATA_LINK_LAYER_ERROR_CODE_MASK	0x7FFF
 #define UIC_DATA_LINK_LAYER_ERROR_PA_INIT	0x2000
+#define UIC_DATA_LINK_LAYER_ERROR_NAC_RECEIVED	0x0001
+#define UIC_DATA_LINK_LAYER_ERROR_TCx_REPLAY_TIMEOUT 0x0002
 
 /* UECN - Host UIC Error Code Network Layer 40h */
 #define UIC_NETWORK_LAYER_ERROR			UFS_BIT(31)
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 10/15] scsi: ufs: add retry for query descriptors

[Yaniv Gardi]

Query commands have 100ms timeout and it may timeout if they are
issued in parallel to ongoing read/write SCSI commands, this change
adds the retry (max: 10) in case command timeouts.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 55 +++++++++++++++++++++++++++++++----------------
 1 file changed, 37 insertions(+), 18 deletions(-)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index e9c7c91..394054e 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -1875,21 +1875,7 @@ static int ufshcd_query_attr_retry(struct ufs_hba *hba,
 	return ret;
 }
 
-/**
- * ufshcd_query_descriptor - API function for sending descriptor requests
- * hba: per-adapter instance
- * opcode: attribute opcode
- * idn: attribute idn to access
- * index: index field
- * selector: selector field
- * desc_buf: the buffer that contains the descriptor
- * buf_len: length parameter passed to the device
- *
- * Returns 0 for success, non-zero in case of failure.
- * The buf_len parameter will contain, on return, the length parameter
- * received on the response.
- */
-static int ufshcd_query_descriptor(struct ufs_hba *hba,
+static int __ufshcd_query_descriptor(struct ufs_hba *hba,
 			enum query_opcode opcode, enum desc_idn idn, u8 index,
 			u8 selector, u8 *desc_buf, int *buf_len)
 {
@@ -1954,6 +1940,39 @@ out:
 }
 
 /**
+ * ufshcd_query_descriptor_retry - API function for sending descriptor
+ * requests
+ * hba: per-adapter instance
+ * opcode: attribute opcode
+ * idn: attribute idn to access
+ * index: index field
+ * selector: selector field
+ * desc_buf: the buffer that contains the descriptor
+ * buf_len: length parameter passed to the device
+ *
+ * Returns 0 for success, non-zero in case of failure.
+ * The buf_len parameter will contain, on return, the length parameter
+ * received on the response.
+ */
+int ufshcd_query_descriptor_retry(struct ufs_hba *hba,
+			enum query_opcode opcode, enum desc_idn idn, u8 index,
+			u8 selector, u8 *desc_buf, int *buf_len)
+{
+	int err;
+	int retries;
+
+	for (retries = QUERY_REQ_RETRIES; retries > 0; retries--) {
+		err = __ufshcd_query_descriptor(hba, opcode, idn, index,
+						selector, desc_buf, buf_len);
+		if (!err || err == -EINVAL)
+			break;
+	}
+
+	return err;
+}
+EXPORT_SYMBOL(ufshcd_query_descriptor_retry);
+
+/**
  * ufshcd_read_desc_param - read the specified descriptor parameter
  * @hba: Pointer to adapter instance
  * @desc_id: descriptor idn value
@@ -1995,9 +2014,9 @@ static int ufshcd_read_desc_param(struct ufs_hba *hba,
 			return -ENOMEM;
 	}
 
-	ret = ufshcd_query_descriptor(hba, UPIU_QUERY_OPCODE_READ_DESC,
-				      desc_id, desc_index, 0, desc_buf,
-				      &buff_len);
+	ret = ufshcd_query_descriptor_retry(hba, UPIU_QUERY_OPCODE_READ_DESC,
+					desc_id, desc_index, 0, desc_buf,
+					&buff_len);
 
 	if (ret || (buff_len < ufs_query_desc_max_size[desc_id]) ||
 	    (desc_buf[QUERY_DESC_LENGTH_OFFSET] !=
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 11/15] scsi: ufs: handle non spec compliant bkops behaviour by device

[Yaniv Gardi]

We are seeing that some devices are raising the urgent bkops exception
events even when BKOPS status doesn't indicate performace impacted or
critical. Handle these device by determining their urgent bkops status
at runtime.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 53 ++++++++++++++++++++++++++++++++++++++++-------
 drivers/scsi/ufs/ufshcd.h |  6 ++++++
 2 files changed, 52 insertions(+), 7 deletions(-)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 394054e..d28c629 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -3763,7 +3763,7 @@ out:
  */
 static int ufshcd_urgent_bkops(struct ufs_hba *hba)
 {
-	return ufshcd_bkops_ctrl(hba, BKOPS_STATUS_PERF_IMPACT);
+	return ufshcd_bkops_ctrl(hba, hba->urgent_bkops_lvl);
 }
 
 static inline int ufshcd_get_ee_status(struct ufs_hba *hba, u32 *status)
@@ -3772,6 +3772,43 @@ static inline int ufshcd_get_ee_status(struct ufs_hba *hba, u32 *status)
 			QUERY_ATTR_IDN_EE_STATUS, 0, 0, status);
 }
 
+static void ufshcd_bkops_exception_event_handler(struct ufs_hba *hba)
+{
+	int err;
+	u32 curr_status = 0;
+
+	if (hba->is_urgent_bkops_lvl_checked)
+		goto enable_auto_bkops;
+
+	err = ufshcd_get_bkops_status(hba, &curr_status);
+	if (err) {
+		dev_err(hba->dev, "%s: failed to get BKOPS status %d\n",
+				__func__, err);
+		goto out;
+	}
+
+	/*
+	 * We are seeing that some devices are raising the urgent bkops
+	 * exception events even when BKOPS status doesn't indicate performace
+	 * impacted or critical. Handle these device by determining their urgent
+	 * bkops status at runtime.
+	 */
+	if (curr_status < BKOPS_STATUS_PERF_IMPACT) {
+		dev_err(hba->dev, "%s: device raised urgent BKOPS exception for bkops status %d\n",
+				__func__, curr_status);
+		/* update the current status as the urgent bkops level */
+		hba->urgent_bkops_lvl = curr_status;
+		hba->is_urgent_bkops_lvl_checked = true;
+	}
+
+enable_auto_bkops:
+	err = ufshcd_enable_auto_bkops(hba);
+out:
+	if (err < 0)
+		dev_err(hba->dev, "%s: failed to handle urgent bkops %d\n",
+				__func__, err);
+}
+
 /**
  * ufshcd_exception_event_handler - handle exceptions raised by device
  * @work: pointer to work data
@@ -3795,12 +3832,10 @@ static void ufshcd_exception_event_handler(struct work_struct *work)
 	}
 
 	status &= hba->ee_ctrl_mask;
-	if (status & MASK_EE_URGENT_BKOPS) {
-		err = ufshcd_urgent_bkops(hba);
-		if (err < 0)
-			dev_err(hba->dev, "%s: failed to handle urgent bkops %d\n",
-					__func__, err);
-	}
+
+	if (status & MASK_EE_URGENT_BKOPS)
+		ufshcd_bkops_exception_event_handler(hba);
+
 out:
 	pm_runtime_put_sync(hba->dev);
 	return;
@@ -4763,6 +4798,10 @@ static int ufshcd_probe_hba(struct ufs_hba *hba)
 
 	ufshcd_init_pwr_info(hba);
 
+	/* set the default level for urgent bkops */
+	hba->urgent_bkops_lvl = BKOPS_STATUS_PERF_IMPACT;
+	hba->is_urgent_bkops_lvl_checked = false;
+
 	/* UniPro link is active now */
 	ufshcd_set_link_active(hba);
 
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index 77e79c0..65f29aa 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -383,6 +383,9 @@ struct ufs_init_prefetch {
  * @clk_list_head: UFS host controller clocks list node head
  * @pwr_info: holds current power mode
  * @max_pwr_info: keeps the device max valid pwm
+ * @urgent_bkops_lvl: keeps track of urgent bkops level for device
+ * @is_urgent_bkops_lvl_checked: keeps track if the urgent bkops level for
+ *  device is known or not.
  */
 struct ufs_hba {
 	void __iomem *mmio_base;
@@ -538,6 +541,9 @@ struct ufs_hba {
 	struct devfreq *devfreq;
 	struct ufs_clk_scaling clk_scaling;
 	bool is_sys_suspended;
+
+	enum bkops_status urgent_bkops_lvl;
+	bool is_urgent_bkops_lvl_checked;
 };
 
 /* Returns true if clocks can be gated. Otherwise false */
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 12/15] scsi: ufs: tune UniPro parameters to optimize hibern8 exit time

[Yaniv Gardi]

Optimal values of local UniPro parameters like PA_Hibern8Time &
PA_TActivate can help reduce the hibern8 exit latency. If both host and
device supports UniPro ver1.6 or later, these parameters will be
automatically tuned during link startup itself. But if either host or
device doesn't support UniPro ver 1.6 or later, we have to manually
tune them. But to keep manual tuning logic simple, we will only do
manual tuning if local unipro version doesn't support ver1.6 or later.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 118 ++++++++++++++++++++++++++++++++++++++++++++++
 drivers/scsi/ufs/ufshcd.h |   1 +
 drivers/scsi/ufs/ufshci.h |   2 +
 drivers/scsi/ufs/unipro.h |  21 +++++++++
 4 files changed, 142 insertions(+)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index d28c629..c9c4d68 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -570,6 +570,34 @@ static inline int ufshcd_is_hba_active(struct ufs_hba *hba)
 	return (ufshcd_readl(hba, REG_CONTROLLER_ENABLE) & 0x1) ? 0 : 1;
 }
 
+u32 ufshcd_get_local_unipro_ver(struct ufs_hba *hba)
+{
+	/* HCI version 1.0 and 1.1 supports UniPro 1.41 */
+	if ((hba->ufs_version == UFSHCI_VERSION_10) ||
+	    (hba->ufs_version == UFSHCI_VERSION_11))
+		return UFS_UNIPRO_VER_1_41;
+	else
+		return UFS_UNIPRO_VER_1_6;
+}
+EXPORT_SYMBOL(ufshcd_get_local_unipro_ver);
+
+static bool ufshcd_is_unipro_pa_params_tuning_req(struct ufs_hba *hba)
+{
+	/*
+	 * If both host and device support UniPro ver1.6 or later, PA layer
+	 * parameters tuning happens during link startup itself.
+	 *
+	 * We can manually tune PA layer parameters if either host or device
+	 * doesn't support UniPro ver 1.6 or later. But to keep manual tuning
+	 * logic simple, we will only do manual tuning if local unipro version
+	 * doesn't support ver1.6 or later.
+	 */
+	if (ufshcd_get_local_unipro_ver(hba) < UFS_UNIPRO_VER_1_6)
+		return true;
+	else
+		return false;
+}
+
 static void ufshcd_ungate_work(struct work_struct *work)
 {
 	int ret;
@@ -4783,6 +4811,95 @@ out:
 }
 
 /**
+ * ufshcd_tune_pa_tactivate - Tunes PA_TActivate of local UniPro
+ * @hba: per-adapter instance
+ *
+ * PA_TActivate parameter can be tuned manually if UniPro version is less than
+ * 1.61. PA_TActivate needs to be greater than or equal to peerM-PHY's
+ * RX_MIN_ACTIVATETIME_CAPABILITY attribute. This optimal value can help reduce
+ * the hibern8 exit latency.
+ *
+ * Returns zero on success, non-zero error value on failure.
+ */
+static int ufshcd_tune_pa_tactivate(struct ufs_hba *hba)
+{
+	int ret = 0;
+	u32 peer_rx_min_activatetime = 0, tuned_pa_tactivate;
+
+	ret = ufshcd_dme_peer_get(hba,
+				  UIC_ARG_MIB_SEL(
+					RX_MIN_ACTIVATETIME_CAPABILITY,
+					UIC_ARG_MPHY_RX_GEN_SEL_INDEX(0)),
+				  &peer_rx_min_activatetime);
+	if (ret)
+		goto out;
+
+	/* make sure proper unit conversion is applied */
+	tuned_pa_tactivate =
+		((peer_rx_min_activatetime * RX_MIN_ACTIVATETIME_UNIT_US)
+		 / PA_TACTIVATE_TIME_UNIT_US);
+	ret = ufshcd_dme_set(hba, UIC_ARG_MIB(PA_TACTIVATE),
+			     tuned_pa_tactivate);
+
+out:
+	return ret;
+}
+
+/**
+ * ufshcd_tune_pa_hibern8time - Tunes PA_Hibern8Time of local UniPro
+ * @hba: per-adapter instance
+ *
+ * PA_Hibern8Time parameter can be tuned manually if UniPro version is less than
+ * 1.61. PA_Hibern8Time needs to be maximum of local M-PHY's
+ * TX_HIBERN8TIME_CAPABILITY & peer M-PHY's RX_HIBERN8TIME_CAPABILITY.
+ * This optimal value can help reduce the hibern8 exit latency.
+ *
+ * Returns zero on success, non-zero error value on failure.
+ */
+static int ufshcd_tune_pa_hibern8time(struct ufs_hba *hba)
+{
+	int ret = 0;
+	u32 local_tx_hibern8_time_cap = 0, peer_rx_hibern8_time_cap = 0;
+	u32 max_hibern8_time, tuned_pa_hibern8time;
+
+	ret = ufshcd_dme_get(hba,
+			     UIC_ARG_MIB_SEL(TX_HIBERN8TIME_CAPABILITY,
+					UIC_ARG_MPHY_TX_GEN_SEL_INDEX(0)),
+				  &local_tx_hibern8_time_cap);
+	if (ret)
+		goto out;
+
+	ret = ufshcd_dme_peer_get(hba,
+				  UIC_ARG_MIB_SEL(RX_HIBERN8TIME_CAPABILITY,
+					UIC_ARG_MPHY_RX_GEN_SEL_INDEX(0)),
+				  &peer_rx_hibern8_time_cap);
+	if (ret)
+		goto out;
+
+	max_hibern8_time = max(local_tx_hibern8_time_cap,
+			       peer_rx_hibern8_time_cap);
+	/* make sure proper unit conversion is applied */
+	tuned_pa_hibern8time = ((max_hibern8_time * HIBERN8TIME_UNIT_US)
+				/ PA_HIBERN8_TIME_UNIT_US);
+	ret = ufshcd_dme_set(hba, UIC_ARG_MIB(PA_HIBERN8TIME),
+			     tuned_pa_hibern8time);
+out:
+	return ret;
+}
+
+static void ufshcd_tune_unipro_params(struct ufs_hba *hba)
+{
+	if (ufshcd_is_unipro_pa_params_tuning_req(hba)) {
+		ufshcd_tune_pa_tactivate(hba);
+		ufshcd_tune_pa_hibern8time(hba);
+	}
+
+	if (hba->dev_quirks & UFS_DEVICE_QUIRK_PA_TACTIVATE)
+		/* set 1ms timeout for PA_TACTIVATE */
+		ufshcd_dme_set(hba, UIC_ARG_MIB(PA_TACTIVATE), 10);
+}
+
+/**
  * ufshcd_probe_hba - probe hba to detect device and initialize
  * @hba: per-adapter instance
  *
@@ -4814,6 +4931,7 @@ static int ufshcd_probe_hba(struct ufs_hba *hba)
 		goto out;
 
 	ufs_advertise_fixup_device(hba);
+	ufshcd_tune_unipro_params(hba);
 
 	ret = ufshcd_set_vccq_rail_unused(hba,
 		(hba->dev_quirks & UFS_DEVICE_NO_VCCQ) ? true : false);
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index 65f29aa..41d9bfd 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -699,6 +699,7 @@ int ufshcd_query_flag(struct ufs_hba *hba, enum query_opcode opcode,
 	enum flag_idn idn, bool *flag_res);
 int ufshcd_hold(struct ufs_hba *hba, bool async);
 void ufshcd_release(struct ufs_hba *hba);
+u32 ufshcd_get_local_unipro_ver(struct ufs_hba *hba);
 
 /* Wrapper functions for safely calling variant operations */
 static inline const char *ufshcd_get_var_name(struct ufs_hba *hba)
diff --git a/drivers/scsi/ufs/ufshci.h b/drivers/scsi/ufs/ufshci.h
index 2b05bfb..4cb1cc6 100644
--- a/drivers/scsi/ufs/ufshci.h
+++ b/drivers/scsi/ufs/ufshci.h
@@ -92,6 +92,7 @@ enum {
 	UFSHCI_VERSION_10 = 0x00010000, /* 1.0 */
 	UFSHCI_VERSION_11 = 0x00010100, /* 1.1 */
 	UFSHCI_VERSION_20 = 0x00000200, /* 2.0 */
+	UFSHCI_VERSION_21 = 0x00000210, /* 2.1 */
 };
 
 /*
@@ -211,6 +212,7 @@ enum {
 
 /* GenSelectorIndex calculation macros for M-PHY attributes */
 #define UIC_ARG_MPHY_TX_GEN_SEL_INDEX(lane) (lane)
+#define UIC_ARG_MPHY_RX_GEN_SEL_INDEX(lane) (PA_MAXDATALANES + (lane))
 
 #define UIC_ARG_MIB_SEL(attr, sel)	((((attr) & 0xFFFF) << 16) |\
 					 ((sel) & 0xFFFF))
diff --git a/drivers/scsi/ufs/unipro.h b/drivers/scsi/ufs/unipro.h
index 816a8a4..8082020 100644
--- a/drivers/scsi/ufs/unipro.h
+++ b/drivers/scsi/ufs/unipro.h
@@ -15,6 +15,7 @@
 /*
  * M-TX Configuration Attributes
  */
+#define TX_HIBERN8TIME_CAPABILITY		0x000F
 #define TX_MODE					0x0021
 #define TX_HSRATE_SERIES			0x0022
 #define TX_HSGEAR				0x0023
@@ -48,8 +49,12 @@
 #define RX_ENTER_HIBERN8			0x00A7
 #define RX_BYPASS_8B10B_ENABLE			0x00A8
 #define RX_TERMINATION_FORCE_ENABLE		0x0089
+#define RX_MIN_ACTIVATETIME_CAPABILITY		0x008F
+#define RX_HIBERN8TIME_CAPABILITY		0x0092
 
 #define is_mphy_tx_attr(attr)			(attr < RX_MODE)
+#define RX_MIN_ACTIVATETIME_UNIT_US		100
+#define HIBERN8TIME_UNIT_US			100
 /*
  * PHY Adpater attributes
  */
@@ -110,6 +115,12 @@
 #define PA_STALLNOCONFIGTIME	0x15A3
 #define PA_SAVECONFIGTIME	0x15A4
 
+#define PA_TACTIVATE_TIME_UNIT_US	10
+#define PA_HIBERN8_TIME_UNIT_US		100
+
+/* PHY Adapter Protocol Constants */
+#define PA_MAXDATALANES	4
+
 /* PA power modes */
 enum {
 	FAST_MODE	= 1,
@@ -143,6 +154,16 @@ enum ufs_hs_gear_tag {
 	UFS_HS_G3,		/* HS Gear 3 */
 };
 
+enum ufs_unipro_ver {
+	UFS_UNIPRO_VER_RESERVED = 0,
+	UFS_UNIPRO_VER_1_40 = 1, /* UniPro version 1.40 */
+	UFS_UNIPRO_VER_1_41 = 2, /* UniPro version 1.41 */
+	UFS_UNIPRO_VER_1_6 = 3,  /* UniPro version 1.6 */
+	UFS_UNIPRO_VER_MAX = 4,  /* UniPro unsupported version */
+	/* UniPro version field mask in PA_LOCALVERINFO */
+	UFS_UNIPRO_VER_MASK = 0xF,
+};
+
 /*
  * Data Link Layer Attributes
  */
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 13/15] scsi: ufs: fix leakage during link off state

[Yaniv Gardi]

Currently when we try to put the link in off/disabled state during
suspend, it seems link is not being kept in low power mode.
This patch fixes the issue by putting the link in hibern8 first
(so device also puts the link in low power mode) and then stop the
host controller.

Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufshcd.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index c9c4d68..f108a0b 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -5619,6 +5619,16 @@ static int ufshcd_link_state_transition(struct ufs_hba *hba,
 		   (!check_for_bkops || (check_for_bkops &&
 		    !hba->auto_bkops_enabled))) {
 		/*
+		 * Let's make sure that link is in low power mode, we are doing
+		 * this currently by putting the link in Hibern8. Otherway to
+		 * put the link in low power mode is to send the DME end point
+		 * to device and then send the DME reset command to local
+		 * unipro. But putting the link in hibern8 is much faster.
+		 */
+		ret = ufshcd_uic_hibern8_enter(hba);
+		if (ret)
+			goto out;
+		/*
 		 * Change controller state to "reset state" which
 		 * should also put the link in off/reset state
 		 */
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 14/15] scsi: ufs: add device quirk delay before putting UFS rails in LPM

[Yaniv Gardi]

We put the UFS device in sleep state & UFS link in hibern8 state during
runtime suspaned. After this we put all the UFS rails in low power
modes immediately but it seems some devices may still draw more than
sleep current from UFS rails (especially from VCCQ rail) atleast for
500us.
To avoid this situation, this change adds 2ms delay before putting
these UFS rails in LPM mode.

Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufs_quirks.c |  4 ++++
 drivers/scsi/ufs/ufs_quirks.h | 11 ++++++++++-
 drivers/scsi/ufs/ufshcd.c     | 10 ++++++++++
 3 files changed, 24 insertions(+), 1 deletion(-)

diff --git a/drivers/scsi/ufs/ufs_quirks.c b/drivers/scsi/ufs/ufs_quirks.c
index 476ed01..af68007 100644
--- a/drivers/scsi/ufs/ufs_quirks.c
+++ b/drivers/scsi/ufs/ufs_quirks.c
@@ -16,11 +16,15 @@
 
 static struct ufs_dev_fix ufs_fixups[] = {
 	/* UFS cards deviations table */
+	UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL,
+		UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM),
 	UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL, UFS_DEVICE_NO_VCCQ),
 	UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL,
 		UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS),
 	UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL,
 		UFS_DEVICE_NO_FASTAUTO),
+	UFS_FIX(UFS_VENDOR_TOSHIBA, UFS_ANY_MODEL,
+		UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM),
 	UFS_FIX(UFS_VENDOR_TOSHIBA, "THGLF2G9C8KBADG",
 		UFS_DEVICE_QUIRK_PA_TACTIVATE),
 	UFS_FIX(UFS_VENDOR_TOSHIBA, "THGLF2G9D8KBADG",
diff --git a/drivers/scsi/ufs/ufs_quirks.h b/drivers/scsi/ufs/ufs_quirks.h
index 01ab166..ffed0da 100644
--- a/drivers/scsi/ufs/ufs_quirks.h
+++ b/drivers/scsi/ufs/ufs_quirks.h
@@ -18,7 +18,7 @@
 /* return true if s1 is a prefix of s2 */
 #define STR_PRFX_EQUAL(s1, s2) !strncmp(s1, s2, strlen(s1))
 
-#define UFS_ANY_VENDOR -1
+#define UFS_ANY_VENDOR 0xFFFF
 #define UFS_ANY_MODEL  "ANY_MODEL"
 
 #define MAX_MODEL_LEN 16
@@ -119,6 +119,15 @@ struct ufs_dev_fix {
  */
 #define UFS_DEVICE_NO_FASTAUTO		(1 << 5)
 
+/*
+ * It seems some UFS devices may keep drawing more than sleep current
+ * (atleast for 500us) from UFS rails (especially from VCCQ rail).
+ * To avoid this situation, add 2ms delay before putting these UFS
+ * rails in LPM mode.
+ */
+#define UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM	(1 << 6)
+
+
 struct ufs_hba;
 void ufs_advertise_fixup_device(struct ufs_hba *hba);
 #endif /* UFS_QUIRKS_H_ */
diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index f108a0b..da882cf 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -5647,6 +5647,16 @@ out:
 static void ufshcd_vreg_set_lpm(struct ufs_hba *hba)
 {
 	/*
+	 * It seems some UFS devices may keep drawing more than sleep current
+	 * (atleast for 500us) from UFS rails (especially from VCCQ rail).
+	 * To avoid this situation, add 2ms delay before putting these UFS
+	 * rails in LPM mode.
+	 */
+	if (!ufshcd_is_link_active(hba) &&
+	    hba->dev_quirks & UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM)
+		usleep_range(2000, 2100);
+
+	/*
 	 * If UFS device is either in UFS_Sleep turn off VCC rail to save some
 	 * power.
 	 *
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

[PATCH v5 15/15] scsi: ufs-qcom: set PA_Local_TX_LCC_Enable before link startup

[Yaniv Gardi]

Some UFS devices (and may be host) have issues if LCC is
enabled. So we are setting PA_Local_TX_LCC_Enable to 0
before link startup which will make sure that both host
and device TX LCC are disabled once link startup is
completed.

This change also:
- enables the device ref clock before changing to HS mode
  and disables it if entered to PWM mode
- adds printouts of testbus debug registers

Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>

---
 drivers/scsi/ufs/ufs-qcom.c | 116 +++++++++++++++++++++++++++++++++++++++++---
 drivers/scsi/ufs/ufs-qcom.h |   9 ++++
 drivers/scsi/ufs/ufshcd.h   |   9 ++++
 drivers/scsi/ufs/unipro.h   |   1 +
 4 files changed, 128 insertions(+), 7 deletions(-)

diff --git a/drivers/scsi/ufs/ufs-qcom.c b/drivers/scsi/ufs/ufs-qcom.c
index ed57729..3aedf73 100644
--- a/drivers/scsi/ufs/ufs-qcom.c
+++ b/drivers/scsi/ufs/ufs-qcom.c
@@ -16,8 +16,8 @@
 #include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/phy/phy.h>
-
 #include <linux/phy/phy-qcom-ufs.h>
+
 #include "ufshcd.h"
 #include "ufshcd-pltfrm.h"
 #include "unipro.h"
@@ -58,6 +58,12 @@ static void ufs_qcom_dump_regs(struct ufs_hba *hba, int offset, int len,
 			len * 4, false);
 }
 
+static void ufs_qcom_dump_regs_wrapper(struct ufs_hba *hba, int offset, int len,
+		char *prefix, void *priv)
+{
+	ufs_qcom_dump_regs(hba, offset, len, prefix);
+}
+
 static int ufs_qcom_get_connected_tx_lanes(struct ufs_hba *hba, u32 *tx_lanes)
 {
 	int err = 0;
@@ -106,9 +112,11 @@ static void ufs_qcom_disable_lane_clks(struct ufs_qcom_host *host)
 	if (!host->is_lane_clks_enabled)
 		return;
 
-	clk_disable_unprepare(host->tx_l1_sync_clk);
+	if (host->hba->lanes_per_direction > 1)
+		clk_disable_unprepare(host->tx_l1_sync_clk);
 	clk_disable_unprepare(host->tx_l0_sync_clk);
-	clk_disable_unprepare(host->rx_l1_sync_clk);
+	if (host->hba->lanes_per_direction > 1)
+		clk_disable_unprepare(host->rx_l1_sync_clk);
 	clk_disable_unprepare(host->rx_l0_sync_clk);
 
 	host->is_lane_clks_enabled = false;
@@ -272,9 +280,8 @@ static int ufs_qcom_power_up_sequence(struct ufs_hba *hba)
 	ret = ufs_qcom_phy_calibrate_phy(phy, is_rate_B);
 
 	if (ret) {
-		dev_err(hba->dev,
-		"%s: ufs_qcom_phy_calibrate_phy()failed, ret = %d\n",
-		__func__, ret);
+		dev_err(hba->dev, "%s: ufs_qcom_phy_calibrate_phy() failed, ret = %d\n",
+			__func__, ret);
 		goto out;
 	}
 
@@ -524,6 +531,18 @@ static int ufs_qcom_link_startup_notify(struct ufs_hba *hba,
 			err = ufs_qcom_set_dme_vs_core_clk_ctrl_clear_div(hba,
 									  150);
 
+		/*
+		 * Some UFS devices (and may be host) have issues if LCC is
+		 * enabled. So we are setting PA_Local_TX_LCC_Enable to 0
+		 * before link startup which will make sure that both host
+		 * and device TX LCC are disabled once link startup is
+		 * completed.
+		 */
+		if (ufshcd_get_local_unipro_ver(hba) != UFS_UNIPRO_VER_1_41)
+			err = ufshcd_dme_set(hba,
+					UIC_ARG_MIB(PA_LOCAL_TX_LCC_ENABLE),
+					0);
+
 		break;
 	case POST_CHANGE:
 		ufs_qcom_link_startup_post_change(hba);
@@ -967,6 +986,10 @@ static int ufs_qcom_pwr_change_notify(struct ufs_hba *hba,
 			goto out;
 		}
 
+		/* enable the device ref clock before changing to HS mode */
+		if (!ufshcd_is_hs_mode(&hba->pwr_info) &&
+			ufshcd_is_hs_mode(dev_req_params))
+			ufs_qcom_dev_ref_clk_ctrl(host, true);
 		break;
 	case POST_CHANGE:
 		if (ufs_qcom_cfg_timers(hba, dev_req_params->gear_rx,
@@ -994,6 +1017,11 @@ static int ufs_qcom_pwr_change_notify(struct ufs_hba *hba,
 		memcpy(&host->dev_req_params,
 				dev_req_params, sizeof(*dev_req_params));
 		ufs_qcom_update_bus_bw_vote(host);
+
+		/* disable the device ref clock if entered PWM mode */
+		if (ufshcd_is_hs_mode(&hba->pwr_info) &&
+			!ufshcd_is_hs_mode(dev_req_params))
+			ufs_qcom_dev_ref_clk_ctrl(host, false);
 		break;
 	default:
 		ret = -EINVAL;
@@ -1095,6 +1123,9 @@ static int ufs_qcom_setup_clocks(struct ufs_hba *hba, bool on)
 			ufs_qcom_phy_disable_iface_clk(host->generic_phy);
 			goto out;
 		}
+		/* enable the device ref clock for HS mode*/
+		if (ufshcd_is_hs_mode(&hba->pwr_info))
+			ufs_qcom_dev_ref_clk_ctrl(host, true);
 		vote = host->bus_vote.saved_vote;
 		if (vote == host->bus_vote.min_bw_vote)
 			ufs_qcom_update_bus_bw_vote(host);
@@ -1372,6 +1403,74 @@ out:
 	return err;
 }
 
+static void ufs_qcom_print_hw_debug_reg_all(struct ufs_hba *hba,
+		void *priv, void (*print_fn)(struct ufs_hba *hba,
+		int offset, int num_regs, char *str, void *priv))
+{
+	u32 reg;
+	struct ufs_qcom_host *host;
+
+	if (unlikely(!hba)) {
+		pr_err("%s: hba is NULL\n", __func__);
+		return;
+	}
+	if (unlikely(!print_fn)) {
+		dev_err(hba->dev, "%s: print_fn is NULL\n", __func__);
+		return;
+	}
+
+	host = ufshcd_get_variant(hba);
+	if (!(host->dbg_print_en & UFS_QCOM_DBG_PRINT_REGS_EN))
+		return;
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_UFS_DBG_RD_REG_OCSC);
+	print_fn(hba, reg, 44, "UFS_UFS_DBG_RD_REG_OCSC ", priv);
+
+	reg = ufshcd_readl(hba, REG_UFS_CFG1);
+	reg |= UFS_BIT(17);
+	ufshcd_writel(hba, reg, REG_UFS_CFG1);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_UFS_DBG_RD_EDTL_RAM);
+	print_fn(hba, reg, 32, "UFS_UFS_DBG_RD_EDTL_RAM ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_UFS_DBG_RD_DESC_RAM);
+	print_fn(hba, reg, 128, "UFS_UFS_DBG_RD_DESC_RAM ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_UFS_DBG_RD_PRDT_RAM);
+	print_fn(hba, reg, 64, "UFS_UFS_DBG_RD_PRDT_RAM ", priv);
+
+	ufshcd_writel(hba, (reg & ~UFS_BIT(17)), REG_UFS_CFG1);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_DBG_RD_REG_UAWM);
+	print_fn(hba, reg, 4, "UFS_DBG_RD_REG_UAWM ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_DBG_RD_REG_UARM);
+	print_fn(hba, reg, 4, "UFS_DBG_RD_REG_UARM ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_DBG_RD_REG_TXUC);
+	print_fn(hba, reg, 48, "UFS_DBG_RD_REG_TXUC ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_DBG_RD_REG_RXUC);
+	print_fn(hba, reg, 27, "UFS_DBG_RD_REG_RXUC ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_DBG_RD_REG_DFC);
+	print_fn(hba, reg, 19, "UFS_DBG_RD_REG_DFC ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_DBG_RD_REG_TRLUT);
+	print_fn(hba, reg, 34, "UFS_DBG_RD_REG_TRLUT ", priv);
+
+	reg = ufs_qcom_get_debug_reg_offset(host, UFS_DBG_RD_REG_TMRLUT);
+	print_fn(hba, reg, 9, "UFS_DBG_RD_REG_TMRLUT ", priv);
+}
+
+static void ufs_qcom_enable_test_bus(struct ufs_qcom_host *host)
+{
+	if (host->dbg_print_en & UFS_QCOM_DBG_PRINT_TEST_BUS_EN)
+		ufshcd_rmwl(host->hba, TEST_BUS_EN, TEST_BUS_EN, REG_UFS_CFG1);
+	else
+		ufshcd_rmwl(host->hba, TEST_BUS_EN, 0, REG_UFS_CFG1);
+}
+
 static void ufs_qcom_get_default_testbus_cfg(struct ufs_qcom_host *host)
 {
 	/* provide a legal default configuration */
@@ -1480,6 +1579,7 @@ int ufs_qcom_testbus_config(struct ufs_qcom_host *host)
 	ufshcd_rmwl(host->hba, mask,
 		    (u32)host->testbus.select_minor << offset,
 		    reg);
+	ufs_qcom_enable_test_bus(host);
 	ufshcd_release(host->hba);
 	pm_runtime_put_sync(host->hba->dev);
 
@@ -1496,8 +1596,10 @@ static void ufs_qcom_dump_dbg_regs(struct ufs_hba *hba)
 	ufs_qcom_dump_regs(hba, REG_UFS_SYS1CLK_1US, 16,
 			"HCI Vendor Specific Registers ");
 
+	ufs_qcom_print_hw_debug_reg_all(hba, NULL, ufs_qcom_dump_regs_wrapper);
 	ufs_qcom_testbus_read(hba);
 }
+
 /**
  * struct ufs_hba_qcom_vops - UFS QCOM specific variant operations
  *
@@ -1542,7 +1644,7 @@ static int ufs_qcom_probe(struct platform_device *pdev)
  * ufs_qcom_remove - set driver_data of the device to NULL
  * @pdev: pointer to platform device handle
  *
- * Always return 0
+ * Always returns 0
  */
 static int ufs_qcom_remove(struct platform_device *pdev)
 {
diff --git a/drivers/scsi/ufs/ufs-qcom.h b/drivers/scsi/ufs/ufs-qcom.h
index 36249b3..a19307a 100644
--- a/drivers/scsi/ufs/ufs-qcom.h
+++ b/drivers/scsi/ufs/ufs-qcom.h
@@ -241,6 +241,15 @@ struct ufs_qcom_host {
 	struct ufs_qcom_testbus testbus;
 };
 
+static inline u32
+ufs_qcom_get_debug_reg_offset(struct ufs_qcom_host *host, u32 reg)
+{
+	if (host->hw_ver.major <= 0x02)
+		return UFS_CNTLR_2_x_x_VEN_REGS_OFFSET(reg);
+
+	return UFS_CNTLR_3_x_x_VEN_REGS_OFFSET(reg);
+};
+
 #define ufs_qcom_is_link_off(hba) ufshcd_is_link_off(hba)
 #define ufs_qcom_is_link_active(hba) ufshcd_is_link_active(hba)
 #define ufs_qcom_is_link_hibern8(hba) ufshcd_is_link_hibern8(hba)
diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
index 41d9bfd..4bb6566 100644
--- a/drivers/scsi/ufs/ufshcd.h
+++ b/drivers/scsi/ufs/ufshcd.h
@@ -54,6 +54,7 @@
 #include <linux/clk.h>
 #include <linux/completion.h>
 #include <linux/regulator/consumer.h>
+#include "unipro.h"
 
 #include <asm/irq.h>
 #include <asm/byteorder.h>
@@ -689,6 +690,14 @@ static inline int ufshcd_dme_peer_get(struct ufs_hba *hba,
 
 int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size);
 
+static inline bool ufshcd_is_hs_mode(struct ufs_pa_layer_attr *pwr_info)
+{
+	return (pwr_info->pwr_rx == FAST_MODE ||
+		pwr_info->pwr_rx == FASTAUTO_MODE) &&
+		(pwr_info->pwr_tx == FAST_MODE ||
+		pwr_info->pwr_tx == FASTAUTO_MODE);
+}
+
 #define ASCII_STD true
 
 int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8 *buf,
diff --git a/drivers/scsi/ufs/unipro.h b/drivers/scsi/ufs/unipro.h
index 8082020..e2854e4 100644
--- a/drivers/scsi/ufs/unipro.h
+++ b/drivers/scsi/ufs/unipro.h
@@ -75,6 +75,7 @@
 #define PA_MAXRXSPEEDFAST	0x1541
 #define PA_MAXRXSPEEDSLOW	0x1542
 #define PA_TXLINKSTARTUPHS	0x1544
+#define PA_LOCAL_TX_LCC_ENABLE	0x155E
 #define PA_TXSPEEDFAST		0x1565
 #define PA_TXSPEEDSLOW		0x1566
 #define PA_REMOTEVERINFO	0x15A0
-- 
1.8.5.2

-- 
QUALCOMM ISRAEL, on behalf of Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation

Re: [PATCH v5 01/15] scsi: ufs-qcom: add number of lanes per direction

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Different platform may have different number of lanes
> for the UFS link.
> Add parameter to device tree specifying how many lanes
> should be configured for the UFS link.
> 
> Signed-off-by: Gilad Broner <gbroner-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org>
> Signed-off-by: Yaniv Gardi <ygardi-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org>
> 
> ---
Reviewed-by: Hannes Reinecke <hare-l3A5Bk7waGM@public.gmane.org>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare-l3A5Bk7waGM@public.gmane.org			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v5 02/15] scsi: ufs: avoid spurious UFS host controller interrupts

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> When control reaches to Linux UFS driver during UFS boot mode, UFS host
> controller interrupt status/enable registers may have left over
> settings.
> In order to avoid any spurious interrupts due to these left overs,
> it's important to clear these interrupt status/enable registers before
> enabling UFS interrupt handling.
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 15 +++++++++++++++
>  1 file changed, 15 insertions(+)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> A race condition exists between request requeueing and scsi layer
> error handling:
> When UFS driver queuecommand returns a busy status for a request,
> it will be requeued and its tag will be freed and set to -1.
> At the same time it is possible that the request will timeout and
> scsi layer will start error handling for it. The scsi layer reuses
> the request and its tag to send error related commands to the device,
> however its tag is no longer valid.
Hmm. How can the host return a 'busy' status for a request?
From my understanding we have three possibilities:

1) queuecommand returns busy; however, that means that the command has
never been send and this issue shouldn't occur
2) The command returns with BUSY status. But in this case it has already
been returned, so there cannot be any timeout coming in.
3) The host receives a command with a tag which is already in-use.
However, that should have been prevented by the block-layer, which
really should ensure that this situation never happens.

So either way I look at it, it really looks like a bug and adding a
timeout handler will just paper over it.
(Not that a timeout handler is a bad idea, in fact I'm convinced that
you need one. Just not for this purpose.)

So can you elaborate how this 'busy' status comes about?
Is the command sent to the device?

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 04/15] scsi: ufs: verify hba controller hce reg value

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Sometimes due to hw issues it takes some time to the
> host controller register to update. In order to verify the register
> has updated, a polling is done until its value is set.
> 
> In addition the functions ufshcd_hba_stop() and
> ufshcd_wait_for_register() was updated with an additional input
> parameter, indicating the timeout between reads will
> be done by sleeping or spinning the cpu.
> 
> Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 53 ++++++++++++++++++++++++++++-------------------
>  drivers/scsi/ufs/ufshcd.h | 12 +++--------
>  2 files changed, 35 insertions(+), 30 deletions(-)
> 
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 3400ceb..80031e6 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -240,11 +240,13 @@ static inline void ufshcd_disable_irq(struct ufs_hba *hba)
>   * @val - wait condition
>   * @interval_us - polling interval in microsecs
>   * @timeout_ms - timeout in millisecs
> + * @can_sleep - perform sleep or just spin
>   *
>   * Returns -ETIMEDOUT on error, zero on success
>   */
> -static int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
> -		u32 val, unsigned long interval_us, unsigned long timeout_ms)
> +int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
> +				u32 val, unsigned long interval_us,
> +				unsigned long timeout_ms, bool can_sleep)
>  {
>  	int err = 0;
>  	unsigned long timeout = jiffies + msecs_to_jiffies(timeout_ms);
> @@ -253,9 +255,10 @@ static int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
>  	val = val & mask;
>  
>  	while ((ufshcd_readl(hba, reg) & mask) != val) {
> -		/* wakeup within 50us of expiry */
> -		usleep_range(interval_us, interval_us + 50);
> -
> +		if (can_sleep)
> +			usleep_range(interval_us, interval_us + 50);
> +		else
> +			udelay(interval_us);
>  		if (time_after(jiffies, timeout)) {
>  			if ((ufshcd_readl(hba, reg) & mask) != val)
>  				err = -ETIMEDOUT;
> @@ -1459,7 +1462,7 @@ ufshcd_clear_cmd(struct ufs_hba *hba, int tag)
>  	 */
>  	err = ufshcd_wait_for_register(hba,
>  			REG_UTP_TRANSFER_REQ_DOOR_BELL,
> -			mask, ~mask, 1000, 1000);
> +			mask, ~mask, 1000, 1000, true);
>  
>  	return err;
>  }
> @@ -2815,6 +2818,23 @@ out:
>  }
>  
>  /**
> + * ufshcd_hba_stop - Send controller to reset state
> + * @hba: per adapter instance
> + * @can_sleep: perform sleep or just spin
> + */
> +static inline void ufshcd_hba_stop(struct ufs_hba *hba, bool can_sleep)
> +{
> +	int err;
> +
> +	ufshcd_writel(hba, CONTROLLER_DISABLE,  REG_CONTROLLER_ENABLE);
> +	err = ufshcd_wait_for_register(hba, REG_CONTROLLER_ENABLE,
> +					CONTROLLER_ENABLE, CONTROLLER_DISABLE,
> +					10, 1, can_sleep);
> +	if (err)
> +		dev_err(hba->dev, "%s: Controller disable failed\n", __func__);
> +}
> +
Shouldn't you return an error here?
If the controller disable failed you probably need a hard reset or
something, otherwise I would assume that every other command from that
point on will not work as expected.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 05/15] scsi: ufs: add support to read device and string descriptors

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> This change adds support to read device descriptor and string descriptor
> from a UFS device
> 
> Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
> Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufs.h    |  1 +
>  drivers/scsi/ufs/ufshcd.c | 88 ++++++++++++++++++++++++++++++++++++++++++++++-
>  drivers/scsi/ufs/ufshcd.h |  7 ++++
>  3 files changed, 95 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h
> index 54a16ce..aacb235 100644
> --- a/drivers/scsi/ufs/ufs.h
> +++ b/drivers/scsi/ufs/ufs.h
> @@ -43,6 +43,7 @@
>  #define GENERAL_UPIU_REQUEST_SIZE 32
>  #define QUERY_DESC_MAX_SIZE       255
>  #define QUERY_DESC_MIN_SIZE       2
> +#define QUERY_DESC_HDR_SIZE       2
>  #define QUERY_OSF_SIZE            (GENERAL_UPIU_REQUEST_SIZE - \
>  					(sizeof(struct utp_upiu_header)))
>  
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 80031e6..e2ed415 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -39,7 +39,7 @@
>  
>  #include <linux/async.h>
>  #include <linux/devfreq.h>
> -
> +#include <linux/nls.h>
>  #include <linux/of.h>
>  #include "ufshcd.h"
>  #include "unipro.h"
> @@ -232,6 +232,16 @@ static inline void ufshcd_disable_irq(struct ufs_hba *hba)
>  	}
>  }
>  
> +/* replace non-printable or non-ASCII characters with spaces */
> +static inline void ufshcd_remove_non_printable(char *val)
> +{
> +	if (!val)
> +		return;
> +
> +	if (*val < 0x20 || *val > 0x7e)
> +		*val = ' ';
> +}
> +
>  /*
>   * ufshcd_wait_for_register - wait for register value to change
>   * @hba - per-adapter interface
> @@ -2021,6 +2031,82 @@ static inline int ufshcd_read_power_desc(struct ufs_hba *hba,
>  	return ufshcd_read_desc(hba, QUERY_DESC_IDN_POWER, 0, buf, size);
>  }
>  
> +int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size)
> +{
> +	return ufshcd_read_desc(hba, QUERY_DESC_IDN_DEVICE, 0, buf, size);
> +}
> +EXPORT_SYMBOL(ufshcd_read_device_desc);
> +
> +/**
> + * ufshcd_read_string_desc - read string descriptor
> + * @hba: pointer to adapter instance
> + * @desc_index: descriptor index
> + * @buf: pointer to buffer where descriptor would be read
> + * @size: size of buf
> + * @ascii: if true convert from unicode to ascii characters
> + *
> + * Return 0 in case of success, non-zero otherwise
> + */
> +int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8 *buf,
> +				u32 size, bool ascii)
> +{
> +	int err = 0;
> +
> +	err = ufshcd_read_desc(hba,
> +				QUERY_DESC_IDN_STRING, desc_index, buf, size);
> +
> +	if (err) {
> +		dev_err(hba->dev, "%s: reading String Desc failed after %d retries. err = %d\n",
> +			__func__, QUERY_REQ_RETRIES, err);
> +		goto out;
> +	}
> +
> +	if (ascii) {
> +		int desc_len;
> +		int ascii_len;
> +		int i;
> +		char *buff_ascii;
> +
> +		desc_len = buf[0];
> +		/* remove header and divide by 2 to move from UTF16 to UTF8 */
> +		ascii_len = (desc_len - QUERY_DESC_HDR_SIZE) / 2 + 1;
> +		if (size < ascii_len + QUERY_DESC_HDR_SIZE) {
> +			dev_err(hba->dev, "%s: buffer allocated size is too small\n",
> +					__func__);
> +			err = -ENOMEM;
> +			goto out;
> +		}
> +
> +		buff_ascii = kmalloc(ascii_len, GFP_KERNEL);
> +		if (!buff_ascii) {
> +			err = -ENOMEM;
> +			goto out_free_buff;
> +		}
> +
> +		/*
> +		 * the descriptor contains string in UTF16 format
> +		 * we need to convert to utf-8 so it can be displayed
> +		 */
> +		utf16s_to_utf8s((wchar_t *)&buf[QUERY_DESC_HDR_SIZE],
> +				desc_len - QUERY_DESC_HDR_SIZE,
> +				UTF16_BIG_ENDIAN, buff_ascii, ascii_len);
> +
> +		/* replace non-printable or non-ASCII characters with spaces */
> +		for (i = 0; i < ascii_len; i++)
> +			ufshcd_remove_non_printable(&buff_ascii[i]);
> +
> +		memset(buf + QUERY_DESC_HDR_SIZE, 0,
> +				size - QUERY_DESC_HDR_SIZE);
> +		memcpy(buf + QUERY_DESC_HDR_SIZE, buff_ascii, ascii_len);
> +		buf[QUERY_DESC_LENGTH_OFFSET] = ascii_len + QUERY_DESC_HDR_SIZE;
> +out_free_buff:
> +		kfree(buff_ascii);
> +	}
> +out:
> +	return err;
> +}
> +EXPORT_SYMBOL(ufshcd_read_string_desc);
> +
>  /**
>   * ufshcd_read_unit_desc_param - read the specified unit descriptor parameter
>   * @hba: Pointer to adapter instance
> diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
> index a6d3572..54e13cc 100644
> --- a/drivers/scsi/ufs/ufshcd.h
> +++ b/drivers/scsi/ufs/ufshcd.h
> @@ -678,6 +678,13 @@ static inline int ufshcd_dme_peer_get(struct ufs_hba *hba,
>  	return ufshcd_dme_get_attr(hba, attr_sel, mib_val, DME_PEER);
>  }
>  
> +int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size);
> +
> +#define ASCII_STD true
> +
> +int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8 *buf,
> +				u32 size, bool ascii);
> +
>  /* Expose Query-Request API */
>  int ufshcd_query_flag(struct ufs_hba *hba, enum query_opcode opcode,
>  	enum flag_idn idn, bool *flag_res);
> 

Probably shows my ignorance, but _where_ is this function actually used?

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 07/15] scsi: ufs: disable vccq if it's not needed by UFS device

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Some UFS devices don't require VCCQ rail for device operations hence
> this change adds support to recognize such devices and remove vote for
> the unused VCCQ rail.
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufs.h    |  1 +
>  drivers/scsi/ufs/ufshcd.c | 60 +++++++++++++++++++++++++++++++++++++++++++----
>  2 files changed, 57 insertions(+), 4 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 06/15] scsi: ufs: separate device and host quirks

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Currently we use the host quirks mechanism in order to
> handle both device and host controller quirks.
> In order to support various of UFS devices we should separate
> handling the device quirks from the host controller's.
> 
> Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
> Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/Makefile     |   2 +-
>  drivers/scsi/ufs/ufs.h        |  31 +++++++++++
>  drivers/scsi/ufs/ufs_quirks.c | 100 ++++++++++++++++++++++++++++++++++
>  drivers/scsi/ufs/ufs_quirks.h | 124 ++++++++++++++++++++++++++++++++++++++++++
>  drivers/scsi/ufs/ufshcd.c     |   2 +
>  drivers/scsi/ufs/ufshcd.h     |   3 +
>  6 files changed, 261 insertions(+), 1 deletion(-)
>  create mode 100644 drivers/scsi/ufs/ufs_quirks.c
>  create mode 100644 drivers/scsi/ufs/ufs_quirks.h
> 
Sigh. _Of course_ there will be device quirks.
Probably too much to hope to have a new implementation where everyone
agrees on the details.

Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 08/15] scsi: ufs: make error handling bit faster

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> UFS driver's error handler forcefully tries to clear all the pending
> requests. For each pending request in the queue, it waits 1 sec for it
> to get cleared. If we have multiple requests in the queue then it's
> possible that we might end up waiting for those many seconds before
> resetting the host. But note that resetting host would any way clear
> all the pending requests from the hardware. Hence this change skips
> the forceful clear of the pending requests if we are anyway going to
> reset the host (for fatal errors).
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 155 +++++++++++++++++++++++++++++++++-------------
>  1 file changed, 112 insertions(+), 43 deletions(-)
> 
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 987cf27..dc096f1 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -133,9 +133,11 @@ enum {
>  /* UFSHCD UIC layer error flags */
>  enum {
>  	UFSHCD_UIC_DL_PA_INIT_ERROR = (1 << 0), /* Data link layer error */
> -	UFSHCD_UIC_NL_ERROR = (1 << 1), /* Network layer error */
> -	UFSHCD_UIC_TL_ERROR = (1 << 2), /* Transport Layer error */
> -	UFSHCD_UIC_DME_ERROR = (1 << 3), /* DME error */
> +	UFSHCD_UIC_DL_NAC_RECEIVED_ERROR = (1 << 1), /* Data link layer error */
> +	UFSHCD_UIC_DL_TCx_REPLAY_ERROR = (1 << 2), /* Data link layer error */
> +	UFSHCD_UIC_NL_ERROR = (1 << 3), /* Network layer error */
> +	UFSHCD_UIC_TL_ERROR = (1 << 4), /* Transport Layer error */
> +	UFSHCD_UIC_DME_ERROR = (1 << 5), /* DME error */
>  };
>  
>  /* Interrupt configuration options */
> @@ -3465,31 +3467,18 @@ static void ufshcd_uic_cmd_compl(struct ufs_hba *hba, u32 intr_status)
>  }
>  
>  /**
> - * ufshcd_transfer_req_compl - handle SCSI and query command completion
> + * __ufshcd_transfer_req_compl - handle SCSI and query command completion
>   * @hba: per adapter instance
> + * @completed_reqs: requests to complete
>   */
> -static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
> +static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
> +					unsigned long completed_reqs)
>  {
>  	struct ufshcd_lrb *lrbp;
>  	struct scsi_cmnd *cmd;
> -	unsigned long completed_reqs;
> -	u32 tr_doorbell;
>  	int result;
>  	int index;
>  
> -	/* Resetting interrupt aggregation counters first and reading the
> -	 * DOOR_BELL afterward allows us to handle all the completed requests.
> -	 * In order to prevent other interrupts starvation the DB is read once
> -	 * after reset. The down side of this solution is the possibility of
> -	 * false interrupt if device completes another request after resetting
> -	 * aggregation and before reading the DB.
> -	 */
> -	if (ufshcd_is_intr_aggr_allowed(hba))
> -		ufshcd_reset_intr_aggr(hba);
> -
> -	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
> -	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
> -
>  	for_each_set_bit(index, &completed_reqs, hba->nutrs) {
>  		lrbp = &hba->lrb[index];
>  		cmd = lrbp->cmd;
> @@ -3519,6 +3508,31 @@ static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
>  }
>  
>  /**
> + * ufshcd_transfer_req_compl - handle SCSI and query command completion
> + * @hba: per adapter instance
> + */
> +static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
> +{
> +	unsigned long completed_reqs;
> +	u32 tr_doorbell;
> +
> +	/* Resetting interrupt aggregation counters first and reading the
> +	 * DOOR_BELL afterward allows us to handle all the completed requests.
> +	 * In order to prevent other interrupts starvation the DB is read once
> +	 * after reset. The down side of this solution is the possibility of
> +	 * false interrupt if device completes another request after resetting
> +	 * aggregation and before reading the DB.
> +	 */
> +	if (ufshcd_is_intr_aggr_allowed(hba))
> +		ufshcd_reset_intr_aggr(hba);
> +
> +	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
> +	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
> +
> +	__ufshcd_transfer_req_compl(hba, completed_reqs);
> +}
> +
> +/**
>   * ufshcd_disable_ee - disable exception event
>   * @hba: per-adapter instance
>   * @mask: exception event to disable
> @@ -3773,6 +3787,13 @@ out:
>  	return;
>  }
>  
> +/* Complete requests that have door-bell cleared */
> +static void ufshcd_complete_requests(struct ufs_hba *hba)
> +{
> +	ufshcd_transfer_req_compl(hba);
> +	ufshcd_tmc_handler(hba);
> +}
> +
>  /**
>   * ufshcd_err_handler - handle UFS errors that require s/w attention
>   * @work: pointer to work structure
> @@ -3785,6 +3806,7 @@ static void ufshcd_err_handler(struct work_struct *work)
>  	u32 err_tm = 0;
>  	int err = 0;
>  	int tag;
> +	bool needs_reset = false;
>  
>  	hba = container_of(work, struct ufs_hba, eh_work);
>  
> @@ -3792,40 +3814,75 @@ static void ufshcd_err_handler(struct work_struct *work)
>  	ufshcd_hold(hba, false);
>  
>  	spin_lock_irqsave(hba->host->host_lock, flags);
> -	if (hba->ufshcd_state == UFSHCD_STATE_RESET) {
> -		spin_unlock_irqrestore(hba->host->host_lock, flags);
> +	if (hba->ufshcd_state == UFSHCD_STATE_RESET)
>  		goto out;
> -	}
>  
>  	hba->ufshcd_state = UFSHCD_STATE_RESET;
>  	ufshcd_set_eh_in_progress(hba);
>  
>  	/* Complete requests that have door-bell cleared by h/w */
> -	ufshcd_transfer_req_compl(hba);
> -	ufshcd_tmc_handler(hba);
> -	spin_unlock_irqrestore(hba->host->host_lock, flags);
> +	ufshcd_complete_requests(hba);
> +	if ((hba->saved_err & INT_FATAL_ERRORS) ||
> +	    ((hba->saved_err & UIC_ERROR) &&
> +	    (hba->saved_uic_err & (UFSHCD_UIC_DL_PA_INIT_ERROR |
> +				   UFSHCD_UIC_DL_NAC_RECEIVED_ERROR |
> +				   UFSHCD_UIC_DL_TCx_REPLAY_ERROR))))
> +		needs_reset = true;
>  
> +	/*
> +	 * if host reset is required then skip clearing the pending
> +	 * transfers forcefully because they will automatically get
> +	 * cleared after link startup.
> +	 */
> +	if (needs_reset)
> +		goto skip_pending_xfer_clear;
> +
> +	/* release lock as clear command might sleep */
> +	spin_unlock_irqrestore(hba->host->host_lock, flags);
>  	/* Clear pending transfer requests */
> -	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs)
> -		if (ufshcd_clear_cmd(hba, tag))
> -			err_xfer |= 1 << tag;
> +	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs) {
> +		if (ufshcd_clear_cmd(hba, tag)) {
> +			err_xfer = true;
> +			goto lock_skip_pending_xfer_clear;
> +		}
> +	}
>  
>  	/* Clear pending task management requests */
> -	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs)
> -		if (ufshcd_clear_tm_cmd(hba, tag))
> -			err_tm |= 1 << tag;
> +	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs) {
> +		if (ufshcd_clear_tm_cmd(hba, tag)) {
> +			err_tm = true;
> +			goto lock_skip_pending_xfer_clear;
> +		}
> +	}
>  
> -	/* Complete the requests that are cleared by s/w */
> +lock_skip_pending_xfer_clear:
>  	spin_lock_irqsave(hba->host->host_lock, flags);
> -	ufshcd_transfer_req_compl(hba);
> -	ufshcd_tmc_handler(hba);
> -	spin_unlock_irqrestore(hba->host->host_lock, flags);
>  
> +	/* Complete the requests that are cleared by s/w */
> +	ufshcd_complete_requests(hba);
> +
> +	if (err_xfer || err_tm)
> +		needs_reset = true;
> +
> +skip_pending_xfer_clear:
>  	/* Fatal errors need reset */
> -	if (err_xfer || err_tm || (hba->saved_err & INT_FATAL_ERRORS) ||
> -			((hba->saved_err & UIC_ERROR) &&
> -			 (hba->saved_uic_err & UFSHCD_UIC_DL_PA_INIT_ERROR))) {
> +	if (needs_reset) {
> +		unsigned long max_doorbells = (1UL << hba->nutrs) - 1;
> +
> +		/*
> +		 * ufshcd_reset_and_restore() does the link reinitialization
> +		 * which will need atleast one empty doorbell slot to send the
> +		 * device management commands (NOP and query commands).
> +		 * If there is no slot empty at this moment then free up last
> +		 * slot forcefully.
> +		 */
> +		if (hba->outstanding_reqs == max_doorbells)
> +			__ufshcd_transfer_req_compl(hba,
> +						    (1UL << (hba->nutrs - 1)));
> +
> +		spin_unlock_irqrestore(hba->host->host_lock, flags);
>  		err = ufshcd_reset_and_restore(hba);
> +		spin_lock_irqsave(hba->host->host_lock, flags);
>  		if (err) {
>  			dev_err(hba->dev, "%s: reset and restore failed\n",
>  					__func__);
Why don't you reserve a command slot for this case (ie reduce the number
of tags by one)?
That way you would always have at least one slot free, wouldn't you?

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 09/15] scsi: ufs: add error recovery after DL NAC error

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Some vendor's UFS device sends back to back NACs for the DL data frames
> causing the host controller to raise the DFES error status. Sometimes
> such UFS devices send back to back NAC without waiting for new
> retransmitted DL frame from the host and in such cases it might be
> possible the Host UniPro goes into bad state without raising the DFES
> error interrupt. If this happens then all the pending commands would
> timeout only after respective SW command (which is generally too
> large).
> 
> This change workarounds such device behaviour like this:
> - As soon as SW sees the DL NAC error, it would schedule the error
>   handler
> - Error handler would sleep for 50ms to see if there any fatal errors
>   raised by UFS controller.
>    - If there are fatal errors then SW does normal error recovery.
>    - If there are no fatal errors then SW sends the NOP command to
>      device to check if link is alive.
>        - If NOP command times out, SW does normal error recovery
>        - If NOP command succeed, skip the error handling.
> 
> If DL NAC error is seen multiple times with some vendor's UFS devices
> then enable this quirk to initiate quick error recovery and also
> silence related error logs to reduce spamming of kernel logs.
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 93 +++++++++++++++++++++++++++++++++++++++++++++++
>  drivers/scsi/ufs/ufshci.h |  2 +
>  2 files changed, 95 insertions(+)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 10/15] scsi: ufs: add retry for query descriptors

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Query commands have 100ms timeout and it may timeout if they are
> issued in parallel to ongoing read/write SCSI commands, this change
> adds the retry (max: 10) in case command timeouts.
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 55 +++++++++++++++++++++++++++++++----------------
>  1 file changed, 37 insertions(+), 18 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v5 11/15] scsi: ufs: handle non spec compliant bkops behaviour by device

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> We are seeing that some devices are raising the urgent bkops exception
> events even when BKOPS status doesn't indicate performace impacted or
> critical. Handle these device by determining their urgent bkops status
> at runtime.
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 53 ++++++++++++++++++++++++++++++++++++++++-------
>  drivers/scsi/ufs/ufshcd.h |  6 ++++++
>  2 files changed, 52 insertions(+), 7 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 12/15] scsi: ufs: tune UniPro parameters to optimize hibern8 exit time

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Optimal values of local UniPro parameters like PA_Hibern8Time &
> PA_TActivate can help reduce the hibern8 exit latency. If both host and
> device supports UniPro ver1.6 or later, these parameters will be
> automatically tuned during link startup itself. But if either host or
> device doesn't support UniPro ver 1.6 or later, we have to manually
> tune them. But to keep manual tuning logic simple, we will only do
> manual tuning if local unipro version doesn't support ver1.6 or later.
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 118 ++++++++++++++++++++++++++++++++++++++++++++++
>  drivers/scsi/ufs/ufshcd.h |   1 +
>  drivers/scsi/ufs/ufshci.h |   2 +
>  drivers/scsi/ufs/unipro.h |  21 +++++++++
>  4 files changed, 142 insertions(+)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 13/15] scsi: ufs: fix leakage during link off state

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Currently when we try to put the link in off/disabled state during
> suspend, it seems link is not being kept in low power mode.
> This patch fixes the issue by putting the link in hibern8 first
> (so device also puts the link in low power mode) and then stop the
> host controller.
> 
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufshcd.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 14/15] scsi: ufs: add device quirk delay before putting UFS rails in LPM

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> We put the UFS device in sleep state & UFS link in hibern8 state during
> runtime suspaned. After this we put all the UFS rails in low power
> modes immediately but it seems some devices may still draw more than
> sleep current from UFS rails (especially from VCCQ rail) atleast for
> 500us.
> To avoid this situation, this change adds 2ms delay before putting
> these UFS rails in LPM mode.
> 
> Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  drivers/scsi/ufs/ufs_quirks.c |  4 ++++
>  drivers/scsi/ufs/ufs_quirks.h | 11 ++++++++++-
>  drivers/scsi/ufs/ufshcd.c     | 10 ++++++++++
>  3 files changed, 24 insertions(+), 1 deletion(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 15/15] scsi: ufs-qcom: set PA_Local_TX_LCC_Enable before link startup

[Hannes Reinecke]

On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
> Some UFS devices (and may be host) have issues if LCC is
> enabled. So we are setting PA_Local_TX_LCC_Enable to 0
> before link startup which will make sure that both host
> and device TX LCC are disabled once link startup is
> completed.
> 
> This change also:
> - enables the device ref clock before changing to HS mode
>   and disables it if entered to PWM mode
> - adds printouts of testbus debug registers
> 
Please make that in two patches; these issues are unrelated and will
just confuse users if one ever need to dig through the git log.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v5 08/15] scsi: ufs: make error handling bit faster

[ygardi]

> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>> UFS driver's error handler forcefully tries to clear all the pending
>> requests. For each pending request in the queue, it waits 1 sec for it
>> to get cleared. If we have multiple requests in the queue then it's
>> possible that we might end up waiting for those many seconds before
>> resetting the host. But note that resetting host would any way clear
>> all the pending requests from the hardware. Hence this change skips
>> the forceful clear of the pending requests if we are anyway going to
>> reset the host (for fatal errors).
>>
>> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
>> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
>>
>> ---
>>  drivers/scsi/ufs/ufshcd.c | 155
>> +++++++++++++++++++++++++++++++++-------------
>>  1 file changed, 112 insertions(+), 43 deletions(-)
>>
>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>> index 987cf27..dc096f1 100644
>> --- a/drivers/scsi/ufs/ufshcd.c
>> +++ b/drivers/scsi/ufs/ufshcd.c
>> @@ -133,9 +133,11 @@ enum {
>>  /* UFSHCD UIC layer error flags */
>>  enum {
>>  	UFSHCD_UIC_DL_PA_INIT_ERROR = (1 << 0), /* Data link layer error */
>> -	UFSHCD_UIC_NL_ERROR = (1 << 1), /* Network layer error */
>> -	UFSHCD_UIC_TL_ERROR = (1 << 2), /* Transport Layer error */
>> -	UFSHCD_UIC_DME_ERROR = (1 << 3), /* DME error */
>> +	UFSHCD_UIC_DL_NAC_RECEIVED_ERROR = (1 << 1), /* Data link layer error
>> */
>> +	UFSHCD_UIC_DL_TCx_REPLAY_ERROR = (1 << 2), /* Data link layer error */
>> +	UFSHCD_UIC_NL_ERROR = (1 << 3), /* Network layer error */
>> +	UFSHCD_UIC_TL_ERROR = (1 << 4), /* Transport Layer error */
>> +	UFSHCD_UIC_DME_ERROR = (1 << 5), /* DME error */
>>  };
>>
>>  /* Interrupt configuration options */
>> @@ -3465,31 +3467,18 @@ static void ufshcd_uic_cmd_compl(struct ufs_hba
>> *hba, u32 intr_status)
>>  }
>>
>>  /**
>> - * ufshcd_transfer_req_compl - handle SCSI and query command completion
>> + * __ufshcd_transfer_req_compl - handle SCSI and query command
>> completion
>>   * @hba: per adapter instance
>> + * @completed_reqs: requests to complete
>>   */
>> -static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
>> +static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
>> +					unsigned long completed_reqs)
>>  {
>>  	struct ufshcd_lrb *lrbp;
>>  	struct scsi_cmnd *cmd;
>> -	unsigned long completed_reqs;
>> -	u32 tr_doorbell;
>>  	int result;
>>  	int index;
>>
>> -	/* Resetting interrupt aggregation counters first and reading the
>> -	 * DOOR_BELL afterward allows us to handle all the completed requests.
>> -	 * In order to prevent other interrupts starvation the DB is read once
>> -	 * after reset. The down side of this solution is the possibility of
>> -	 * false interrupt if device completes another request after resetting
>> -	 * aggregation and before reading the DB.
>> -	 */
>> -	if (ufshcd_is_intr_aggr_allowed(hba))
>> -		ufshcd_reset_intr_aggr(hba);
>> -
>> -	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
>> -	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
>> -
>>  	for_each_set_bit(index, &completed_reqs, hba->nutrs) {
>>  		lrbp = &hba->lrb[index];
>>  		cmd = lrbp->cmd;
>> @@ -3519,6 +3508,31 @@ static void ufshcd_transfer_req_compl(struct
>> ufs_hba *hba)
>>  }
>>
>>  /**
>> + * ufshcd_transfer_req_compl - handle SCSI and query command completion
>> + * @hba: per adapter instance
>> + */
>> +static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
>> +{
>> +	unsigned long completed_reqs;
>> +	u32 tr_doorbell;
>> +
>> +	/* Resetting interrupt aggregation counters first and reading the
>> +	 * DOOR_BELL afterward allows us to handle all the completed requests.
>> +	 * In order to prevent other interrupts starvation the DB is read once
>> +	 * after reset. The down side of this solution is the possibility of
>> +	 * false interrupt if device completes another request after resetting
>> +	 * aggregation and before reading the DB.
>> +	 */
>> +	if (ufshcd_is_intr_aggr_allowed(hba))
>> +		ufshcd_reset_intr_aggr(hba);
>> +
>> +	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
>> +	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
>> +
>> +	__ufshcd_transfer_req_compl(hba, completed_reqs);
>> +}
>> +
>> +/**
>>   * ufshcd_disable_ee - disable exception event
>>   * @hba: per-adapter instance
>>   * @mask: exception event to disable
>> @@ -3773,6 +3787,13 @@ out:
>>  	return;
>>  }
>>
>> +/* Complete requests that have door-bell cleared */
>> +static void ufshcd_complete_requests(struct ufs_hba *hba)
>> +{
>> +	ufshcd_transfer_req_compl(hba);
>> +	ufshcd_tmc_handler(hba);
>> +}
>> +
>>  /**
>>   * ufshcd_err_handler - handle UFS errors that require s/w attention
>>   * @work: pointer to work structure
>> @@ -3785,6 +3806,7 @@ static void ufshcd_err_handler(struct work_struct
>> *work)
>>  	u32 err_tm = 0;
>>  	int err = 0;
>>  	int tag;
>> +	bool needs_reset = false;
>>
>>  	hba = container_of(work, struct ufs_hba, eh_work);
>>
>> @@ -3792,40 +3814,75 @@ static void ufshcd_err_handler(struct
>> work_struct *work)
>>  	ufshcd_hold(hba, false);
>>
>>  	spin_lock_irqsave(hba->host->host_lock, flags);
>> -	if (hba->ufshcd_state == UFSHCD_STATE_RESET) {
>> -		spin_unlock_irqrestore(hba->host->host_lock, flags);
>> +	if (hba->ufshcd_state == UFSHCD_STATE_RESET)
>>  		goto out;
>> -	}
>>
>>  	hba->ufshcd_state = UFSHCD_STATE_RESET;
>>  	ufshcd_set_eh_in_progress(hba);
>>
>>  	/* Complete requests that have door-bell cleared by h/w */
>> -	ufshcd_transfer_req_compl(hba);
>> -	ufshcd_tmc_handler(hba);
>> -	spin_unlock_irqrestore(hba->host->host_lock, flags);
>> +	ufshcd_complete_requests(hba);
>> +	if ((hba->saved_err & INT_FATAL_ERRORS) ||
>> +	    ((hba->saved_err & UIC_ERROR) &&
>> +	    (hba->saved_uic_err & (UFSHCD_UIC_DL_PA_INIT_ERROR |
>> +				   UFSHCD_UIC_DL_NAC_RECEIVED_ERROR |
>> +				   UFSHCD_UIC_DL_TCx_REPLAY_ERROR))))
>> +		needs_reset = true;
>>
>> +	/*
>> +	 * if host reset is required then skip clearing the pending
>> +	 * transfers forcefully because they will automatically get
>> +	 * cleared after link startup.
>> +	 */
>> +	if (needs_reset)
>> +		goto skip_pending_xfer_clear;
>> +
>> +	/* release lock as clear command might sleep */
>> +	spin_unlock_irqrestore(hba->host->host_lock, flags);
>>  	/* Clear pending transfer requests */
>> -	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs)
>> -		if (ufshcd_clear_cmd(hba, tag))
>> -			err_xfer |= 1 << tag;
>> +	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs) {
>> +		if (ufshcd_clear_cmd(hba, tag)) {
>> +			err_xfer = true;
>> +			goto lock_skip_pending_xfer_clear;
>> +		}
>> +	}
>>
>>  	/* Clear pending task management requests */
>> -	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs)
>> -		if (ufshcd_clear_tm_cmd(hba, tag))
>> -			err_tm |= 1 << tag;
>> +	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs) {
>> +		if (ufshcd_clear_tm_cmd(hba, tag)) {
>> +			err_tm = true;
>> +			goto lock_skip_pending_xfer_clear;
>> +		}
>> +	}
>>
>> -	/* Complete the requests that are cleared by s/w */
>> +lock_skip_pending_xfer_clear:
>>  	spin_lock_irqsave(hba->host->host_lock, flags);
>> -	ufshcd_transfer_req_compl(hba);
>> -	ufshcd_tmc_handler(hba);
>> -	spin_unlock_irqrestore(hba->host->host_lock, flags);
>>
>> +	/* Complete the requests that are cleared by s/w */
>> +	ufshcd_complete_requests(hba);
>> +
>> +	if (err_xfer || err_tm)
>> +		needs_reset = true;
>> +
>> +skip_pending_xfer_clear:
>>  	/* Fatal errors need reset */
>> -	if (err_xfer || err_tm || (hba->saved_err & INT_FATAL_ERRORS) ||
>> -			((hba->saved_err & UIC_ERROR) &&
>> -			 (hba->saved_uic_err & UFSHCD_UIC_DL_PA_INIT_ERROR))) {
>> +	if (needs_reset) {
>> +		unsigned long max_doorbells = (1UL << hba->nutrs) - 1;
>> +
>> +		/*
>> +		 * ufshcd_reset_and_restore() does the link reinitialization
>> +		 * which will need atleast one empty doorbell slot to send the
>> +		 * device management commands (NOP and query commands).
>> +		 * If there is no slot empty at this moment then free up last
>> +		 * slot forcefully.
>> +		 */
>> +		if (hba->outstanding_reqs == max_doorbells)
>> +			__ufshcd_transfer_req_compl(hba,
>> +						    (1UL << (hba->nutrs - 1)));
>> +
>> +		spin_unlock_irqrestore(hba->host->host_lock, flags);
>>  		err = ufshcd_reset_and_restore(hba);
>> +		spin_lock_irqsave(hba->host->host_lock, flags);
>>  		if (err) {
>>  			dev_err(hba->dev, "%s: reset and restore failed\n",
>>  					__func__);
> Why don't you reserve a command slot for this case (ie reduce the number
> of tags by one)?
> That way you would always have at least one slot free, wouldn't you?
>

Hello Hannes,

We are discussing here, a very-very rare scenario where 2 conditions must
co-exist:
1. a fatal error that requires the controller to be reset.
2. all slots are taken.

This 2 conditions, very rarely should happen together.
At that point, it would be better to free the last slot, than to save one
slot for this scenario.
Also, we should remember that reducing the queue-depth for the entire
usual operation, might be a performance hit at some point, where, for
example, a LUN has only 8 slots, will now have 7, which is a hit of 12.5%
of the potential parallelism. So, i would recommend to stick with the
above proposal that not only makes more sense when you estimate the
probability of the conditions to co-exist, but also was tested and proven
to be safe with no performance hit.

Regards,
Yaniv




> Cheers,
>
> Hannes
> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@suse.de			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH v5 05/15] scsi: ufs: add support to read device and string descriptors

[ygardi]

> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>> This change adds support to read device descriptor and string descriptor
>> from a UFS device
>>
>> Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
>> Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
>> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
>>
>> ---
>>  drivers/scsi/ufs/ufs.h    |  1 +
>>  drivers/scsi/ufs/ufshcd.c | 88
>> ++++++++++++++++++++++++++++++++++++++++++++++-
>>  drivers/scsi/ufs/ufshcd.h |  7 ++++
>>  3 files changed, 95 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h
>> index 54a16ce..aacb235 100644
>> --- a/drivers/scsi/ufs/ufs.h
>> +++ b/drivers/scsi/ufs/ufs.h
>> @@ -43,6 +43,7 @@
>>  #define GENERAL_UPIU_REQUEST_SIZE 32
>>  #define QUERY_DESC_MAX_SIZE       255
>>  #define QUERY_DESC_MIN_SIZE       2
>> +#define QUERY_DESC_HDR_SIZE       2
>>  #define QUERY_OSF_SIZE            (GENERAL_UPIU_REQUEST_SIZE - \
>>  					(sizeof(struct utp_upiu_header)))
>>
>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>> index 80031e6..e2ed415 100644
>> --- a/drivers/scsi/ufs/ufshcd.c
>> +++ b/drivers/scsi/ufs/ufshcd.c
>> @@ -39,7 +39,7 @@
>>
>>  #include <linux/async.h>
>>  #include <linux/devfreq.h>
>> -
>> +#include <linux/nls.h>
>>  #include <linux/of.h>
>>  #include "ufshcd.h"
>>  #include "unipro.h"
>> @@ -232,6 +232,16 @@ static inline void ufshcd_disable_irq(struct
>> ufs_hba *hba)
>>  	}
>>  }
>>
>> +/* replace non-printable or non-ASCII characters with spaces */
>> +static inline void ufshcd_remove_non_printable(char *val)
>> +{
>> +	if (!val)
>> +		return;
>> +
>> +	if (*val < 0x20 || *val > 0x7e)
>> +		*val = ' ';
>> +}
>> +
>>  /*
>>   * ufshcd_wait_for_register - wait for register value to change
>>   * @hba - per-adapter interface
>> @@ -2021,6 +2031,82 @@ static inline int ufshcd_read_power_desc(struct
>> ufs_hba *hba,
>>  	return ufshcd_read_desc(hba, QUERY_DESC_IDN_POWER, 0, buf, size);
>>  }
>>
>> +int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size)
>> +{
>> +	return ufshcd_read_desc(hba, QUERY_DESC_IDN_DEVICE, 0, buf, size);
>> +}
>> +EXPORT_SYMBOL(ufshcd_read_device_desc);
>> +
>> +/**
>> + * ufshcd_read_string_desc - read string descriptor
>> + * @hba: pointer to adapter instance
>> + * @desc_index: descriptor index
>> + * @buf: pointer to buffer where descriptor would be read
>> + * @size: size of buf
>> + * @ascii: if true convert from unicode to ascii characters
>> + *
>> + * Return 0 in case of success, non-zero otherwise
>> + */
>> +int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8
>> *buf,
>> +				u32 size, bool ascii)
>> +{
>> +	int err = 0;
>> +
>> +	err = ufshcd_read_desc(hba,
>> +				QUERY_DESC_IDN_STRING, desc_index, buf, size);
>> +
>> +	if (err) {
>> +		dev_err(hba->dev, "%s: reading String Desc failed after %d retries.
>> err = %d\n",
>> +			__func__, QUERY_REQ_RETRIES, err);
>> +		goto out;
>> +	}
>> +
>> +	if (ascii) {
>> +		int desc_len;
>> +		int ascii_len;
>> +		int i;
>> +		char *buff_ascii;
>> +
>> +		desc_len = buf[0];
>> +		/* remove header and divide by 2 to move from UTF16 to UTF8 */
>> +		ascii_len = (desc_len - QUERY_DESC_HDR_SIZE) / 2 + 1;
>> +		if (size < ascii_len + QUERY_DESC_HDR_SIZE) {
>> +			dev_err(hba->dev, "%s: buffer allocated size is too small\n",
>> +					__func__);
>> +			err = -ENOMEM;
>> +			goto out;
>> +		}
>> +
>> +		buff_ascii = kmalloc(ascii_len, GFP_KERNEL);
>> +		if (!buff_ascii) {
>> +			err = -ENOMEM;
>> +			goto out_free_buff;
>> +		}
>> +
>> +		/*
>> +		 * the descriptor contains string in UTF16 format
>> +		 * we need to convert to utf-8 so it can be displayed
>> +		 */
>> +		utf16s_to_utf8s((wchar_t *)&buf[QUERY_DESC_HDR_SIZE],
>> +				desc_len - QUERY_DESC_HDR_SIZE,
>> +				UTF16_BIG_ENDIAN, buff_ascii, ascii_len);
>> +
>> +		/* replace non-printable or non-ASCII characters with spaces */
>> +		for (i = 0; i < ascii_len; i++)
>> +			ufshcd_remove_non_printable(&buff_ascii[i]);
>> +
>> +		memset(buf + QUERY_DESC_HDR_SIZE, 0,
>> +				size - QUERY_DESC_HDR_SIZE);
>> +		memcpy(buf + QUERY_DESC_HDR_SIZE, buff_ascii, ascii_len);
>> +		buf[QUERY_DESC_LENGTH_OFFSET] = ascii_len + QUERY_DESC_HDR_SIZE;
>> +out_free_buff:
>> +		kfree(buff_ascii);
>> +	}
>> +out:
>> +	return err;
>> +}
>> +EXPORT_SYMBOL(ufshcd_read_string_desc);
>> +
>>  /**
>>   * ufshcd_read_unit_desc_param - read the specified unit descriptor
>> parameter
>>   * @hba: Pointer to adapter instance
>> diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
>> index a6d3572..54e13cc 100644
>> --- a/drivers/scsi/ufs/ufshcd.h
>> +++ b/drivers/scsi/ufs/ufshcd.h
>> @@ -678,6 +678,13 @@ static inline int ufshcd_dme_peer_get(struct
>> ufs_hba *hba,
>>  	return ufshcd_dme_get_attr(hba, attr_sel, mib_val, DME_PEER);
>>  }
>>
>> +int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size);
>> +
>> +#define ASCII_STD true
>> +
>> +int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8
>> *buf,
>> +				u32 size, bool ascii);
>> +
>>  /* Expose Query-Request API */
>>  int ufshcd_query_flag(struct ufs_hba *hba, enum query_opcode opcode,
>>  	enum flag_idn idn, bool *flag_res);
>>
>
> Probably shows my ignorance, but _where_ is this function actually used?
>
> Cheers,
> Hannes


Hello Hannes,
I was asked by reviewer Tomas Winkler, to seperate a patch (v4) into 2
patches in v5.
so, this is part I of the seperation, and the functions here, are being
called by the next patch v5 06/15 - in the quirk.c file in
ufs_get_device_info() routine.
hope that helps.

regards,
Yaniv


> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@suse.de			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH v5 08/15] scsi: ufs: make error handling bit faster

[Hannes Reinecke]

On 03/01/2016 05:56 PM, ygardi@codeaurora.org wrote:
>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>> UFS driver's error handler forcefully tries to clear all the pending
>>> requests. For each pending request in the queue, it waits 1 sec for it
>>> to get cleared. If we have multiple requests in the queue then it's
>>> possible that we might end up waiting for those many seconds before
>>> resetting the host. But note that resetting host would any way clear
>>> all the pending requests from the hardware. Hence this change skips
>>> the forceful clear of the pending requests if we are anyway going to
>>> reset the host (for fatal errors).
>>>
>>> Signed-off-by: Subhash Jadavani <subhashj@codeaurora.org>
>>> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
>>>
>>> ---
>>>  drivers/scsi/ufs/ufshcd.c | 155
>>> +++++++++++++++++++++++++++++++++-------------
>>>  1 file changed, 112 insertions(+), 43 deletions(-)
>>>
>>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>>> index 987cf27..dc096f1 100644
>>> --- a/drivers/scsi/ufs/ufshcd.c
>>> +++ b/drivers/scsi/ufs/ufshcd.c
>>> @@ -133,9 +133,11 @@ enum {
>>>  /* UFSHCD UIC layer error flags */
>>>  enum {
>>>  	UFSHCD_UIC_DL_PA_INIT_ERROR = (1 << 0), /* Data link layer error */
>>> -	UFSHCD_UIC_NL_ERROR = (1 << 1), /* Network layer error */
>>> -	UFSHCD_UIC_TL_ERROR = (1 << 2), /* Transport Layer error */
>>> -	UFSHCD_UIC_DME_ERROR = (1 << 3), /* DME error */
>>> +	UFSHCD_UIC_DL_NAC_RECEIVED_ERROR = (1 << 1), /* Data link layer error
>>> */
>>> +	UFSHCD_UIC_DL_TCx_REPLAY_ERROR = (1 << 2), /* Data link layer error */
>>> +	UFSHCD_UIC_NL_ERROR = (1 << 3), /* Network layer error */
>>> +	UFSHCD_UIC_TL_ERROR = (1 << 4), /* Transport Layer error */
>>> +	UFSHCD_UIC_DME_ERROR = (1 << 5), /* DME error */
>>>  };
>>>
>>>  /* Interrupt configuration options */
>>> @@ -3465,31 +3467,18 @@ static void ufshcd_uic_cmd_compl(struct ufs_hba
>>> *hba, u32 intr_status)
>>>  }
>>>
>>>  /**
>>> - * ufshcd_transfer_req_compl - handle SCSI and query command completion
>>> + * __ufshcd_transfer_req_compl - handle SCSI and query command
>>> completion
>>>   * @hba: per adapter instance
>>> + * @completed_reqs: requests to complete
>>>   */
>>> -static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
>>> +static void __ufshcd_transfer_req_compl(struct ufs_hba *hba,
>>> +					unsigned long completed_reqs)
>>>  {
>>>  	struct ufshcd_lrb *lrbp;
>>>  	struct scsi_cmnd *cmd;
>>> -	unsigned long completed_reqs;
>>> -	u32 tr_doorbell;
>>>  	int result;
>>>  	int index;
>>>
>>> -	/* Resetting interrupt aggregation counters first and reading the
>>> -	 * DOOR_BELL afterward allows us to handle all the completed requests.
>>> -	 * In order to prevent other interrupts starvation the DB is read once
>>> -	 * after reset. The down side of this solution is the possibility of
>>> -	 * false interrupt if device completes another request after resetting
>>> -	 * aggregation and before reading the DB.
>>> -	 */
>>> -	if (ufshcd_is_intr_aggr_allowed(hba))
>>> -		ufshcd_reset_intr_aggr(hba);
>>> -
>>> -	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
>>> -	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
>>> -
>>>  	for_each_set_bit(index, &completed_reqs, hba->nutrs) {
>>>  		lrbp = &hba->lrb[index];
>>>  		cmd = lrbp->cmd;
>>> @@ -3519,6 +3508,31 @@ static void ufshcd_transfer_req_compl(struct
>>> ufs_hba *hba)
>>>  }
>>>
>>>  /**
>>> + * ufshcd_transfer_req_compl - handle SCSI and query command completion
>>> + * @hba: per adapter instance
>>> + */
>>> +static void ufshcd_transfer_req_compl(struct ufs_hba *hba)
>>> +{
>>> +	unsigned long completed_reqs;
>>> +	u32 tr_doorbell;
>>> +
>>> +	/* Resetting interrupt aggregation counters first and reading the
>>> +	 * DOOR_BELL afterward allows us to handle all the completed requests.
>>> +	 * In order to prevent other interrupts starvation the DB is read once
>>> +	 * after reset. The down side of this solution is the possibility of
>>> +	 * false interrupt if device completes another request after resetting
>>> +	 * aggregation and before reading the DB.
>>> +	 */
>>> +	if (ufshcd_is_intr_aggr_allowed(hba))
>>> +		ufshcd_reset_intr_aggr(hba);
>>> +
>>> +	tr_doorbell = ufshcd_readl(hba, REG_UTP_TRANSFER_REQ_DOOR_BELL);
>>> +	completed_reqs = tr_doorbell ^ hba->outstanding_reqs;
>>> +
>>> +	__ufshcd_transfer_req_compl(hba, completed_reqs);
>>> +}
>>> +
>>> +/**
>>>   * ufshcd_disable_ee - disable exception event
>>>   * @hba: per-adapter instance
>>>   * @mask: exception event to disable
>>> @@ -3773,6 +3787,13 @@ out:
>>>  	return;
>>>  }
>>>
>>> +/* Complete requests that have door-bell cleared */
>>> +static void ufshcd_complete_requests(struct ufs_hba *hba)
>>> +{
>>> +	ufshcd_transfer_req_compl(hba);
>>> +	ufshcd_tmc_handler(hba);
>>> +}
>>> +
>>>  /**
>>>   * ufshcd_err_handler - handle UFS errors that require s/w attention
>>>   * @work: pointer to work structure
>>> @@ -3785,6 +3806,7 @@ static void ufshcd_err_handler(struct work_struct
>>> *work)
>>>  	u32 err_tm = 0;
>>>  	int err = 0;
>>>  	int tag;
>>> +	bool needs_reset = false;
>>>
>>>  	hba = container_of(work, struct ufs_hba, eh_work);
>>>
>>> @@ -3792,40 +3814,75 @@ static void ufshcd_err_handler(struct
>>> work_struct *work)
>>>  	ufshcd_hold(hba, false);
>>>
>>>  	spin_lock_irqsave(hba->host->host_lock, flags);
>>> -	if (hba->ufshcd_state == UFSHCD_STATE_RESET) {
>>> -		spin_unlock_irqrestore(hba->host->host_lock, flags);
>>> +	if (hba->ufshcd_state == UFSHCD_STATE_RESET)
>>>  		goto out;
>>> -	}
>>>
>>>  	hba->ufshcd_state = UFSHCD_STATE_RESET;
>>>  	ufshcd_set_eh_in_progress(hba);
>>>
>>>  	/* Complete requests that have door-bell cleared by h/w */
>>> -	ufshcd_transfer_req_compl(hba);
>>> -	ufshcd_tmc_handler(hba);
>>> -	spin_unlock_irqrestore(hba->host->host_lock, flags);
>>> +	ufshcd_complete_requests(hba);
>>> +	if ((hba->saved_err & INT_FATAL_ERRORS) ||
>>> +	    ((hba->saved_err & UIC_ERROR) &&
>>> +	    (hba->saved_uic_err & (UFSHCD_UIC_DL_PA_INIT_ERROR |
>>> +				   UFSHCD_UIC_DL_NAC_RECEIVED_ERROR |
>>> +				   UFSHCD_UIC_DL_TCx_REPLAY_ERROR))))
>>> +		needs_reset = true;
>>>
>>> +	/*
>>> +	 * if host reset is required then skip clearing the pending
>>> +	 * transfers forcefully because they will automatically get
>>> +	 * cleared after link startup.
>>> +	 */
>>> +	if (needs_reset)
>>> +		goto skip_pending_xfer_clear;
>>> +
>>> +	/* release lock as clear command might sleep */
>>> +	spin_unlock_irqrestore(hba->host->host_lock, flags);
>>>  	/* Clear pending transfer requests */
>>> -	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs)
>>> -		if (ufshcd_clear_cmd(hba, tag))
>>> -			err_xfer |= 1 << tag;
>>> +	for_each_set_bit(tag, &hba->outstanding_reqs, hba->nutrs) {
>>> +		if (ufshcd_clear_cmd(hba, tag)) {
>>> +			err_xfer = true;
>>> +			goto lock_skip_pending_xfer_clear;
>>> +		}
>>> +	}
>>>
>>>  	/* Clear pending task management requests */
>>> -	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs)
>>> -		if (ufshcd_clear_tm_cmd(hba, tag))
>>> -			err_tm |= 1 << tag;
>>> +	for_each_set_bit(tag, &hba->outstanding_tasks, hba->nutmrs) {
>>> +		if (ufshcd_clear_tm_cmd(hba, tag)) {
>>> +			err_tm = true;
>>> +			goto lock_skip_pending_xfer_clear;
>>> +		}
>>> +	}
>>>
>>> -	/* Complete the requests that are cleared by s/w */
>>> +lock_skip_pending_xfer_clear:
>>>  	spin_lock_irqsave(hba->host->host_lock, flags);
>>> -	ufshcd_transfer_req_compl(hba);
>>> -	ufshcd_tmc_handler(hba);
>>> -	spin_unlock_irqrestore(hba->host->host_lock, flags);
>>>
>>> +	/* Complete the requests that are cleared by s/w */
>>> +	ufshcd_complete_requests(hba);
>>> +
>>> +	if (err_xfer || err_tm)
>>> +		needs_reset = true;
>>> +
>>> +skip_pending_xfer_clear:
>>>  	/* Fatal errors need reset */
>>> -	if (err_xfer || err_tm || (hba->saved_err & INT_FATAL_ERRORS) ||
>>> -			((hba->saved_err & UIC_ERROR) &&
>>> -			 (hba->saved_uic_err & UFSHCD_UIC_DL_PA_INIT_ERROR))) {
>>> +	if (needs_reset) {
>>> +		unsigned long max_doorbells = (1UL << hba->nutrs) - 1;
>>> +
>>> +		/*
>>> +		 * ufshcd_reset_and_restore() does the link reinitialization
>>> +		 * which will need atleast one empty doorbell slot to send the
>>> +		 * device management commands (NOP and query commands).
>>> +		 * If there is no slot empty at this moment then free up last
>>> +		 * slot forcefully.
>>> +		 */
>>> +		if (hba->outstanding_reqs == max_doorbells)
>>> +			__ufshcd_transfer_req_compl(hba,
>>> +						    (1UL << (hba->nutrs - 1)));
>>> +
>>> +		spin_unlock_irqrestore(hba->host->host_lock, flags);
>>>  		err = ufshcd_reset_and_restore(hba);
>>> +		spin_lock_irqsave(hba->host->host_lock, flags);
>>>  		if (err) {
>>>  			dev_err(hba->dev, "%s: reset and restore failed\n",
>>>  					__func__);
>> Why don't you reserve a command slot for this case (ie reduce the number
>> of tags by one)?
>> That way you would always have at least one slot free, wouldn't you?
>>
> 
> Hello Hannes,
> 
> We are discussing here, a very-very rare scenario where 2 conditions must
> co-exist:
> 1. a fatal error that requires the controller to be reset.
> 2. all slots are taken.
> 
> This 2 conditions, very rarely should happen together.
> At that point, it would be better to free the last slot, than to save one
> slot for this scenario.
> Also, we should remember that reducing the queue-depth for the entire
> usual operation, might be a performance hit at some point, where, for
> example, a LUN has only 8 slots, will now have 7, which is a hit of 12.5%
> of the potential parallelism. So, i would recommend to stick with the
> above proposal that not only makes more sense when you estimate the
> probability of the conditions to co-exist, but also was tested and proven
> to be safe with no performance hit.
> 
Okay, good point. I didn't know how many slots there are.
But if you only got 8 you surely don't want to waste one.

Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 05/15] scsi: ufs: add support to read device and string descriptors

[Hannes Reinecke]

On 03/01/2016 06:01 PM, ygardi@codeaurora.org wrote:
>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>> This change adds support to read device descriptor and string descriptor
>>> from a UFS device
>>>
>>> Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
>>> Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
>>> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
>>>
>>> ---
>>>  drivers/scsi/ufs/ufs.h    |  1 +
>>>  drivers/scsi/ufs/ufshcd.c | 88
>>> ++++++++++++++++++++++++++++++++++++++++++++++-
>>>  drivers/scsi/ufs/ufshcd.h |  7 ++++
>>>  3 files changed, 95 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/scsi/ufs/ufs.h b/drivers/scsi/ufs/ufs.h
>>> index 54a16ce..aacb235 100644
>>> --- a/drivers/scsi/ufs/ufs.h
>>> +++ b/drivers/scsi/ufs/ufs.h
>>> @@ -43,6 +43,7 @@
>>>  #define GENERAL_UPIU_REQUEST_SIZE 32
>>>  #define QUERY_DESC_MAX_SIZE       255
>>>  #define QUERY_DESC_MIN_SIZE       2
>>> +#define QUERY_DESC_HDR_SIZE       2
>>>  #define QUERY_OSF_SIZE            (GENERAL_UPIU_REQUEST_SIZE - \
>>>  					(sizeof(struct utp_upiu_header)))
>>>
>>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>>> index 80031e6..e2ed415 100644
>>> --- a/drivers/scsi/ufs/ufshcd.c
>>> +++ b/drivers/scsi/ufs/ufshcd.c
>>> @@ -39,7 +39,7 @@
>>>
>>>  #include <linux/async.h>
>>>  #include <linux/devfreq.h>
>>> -
>>> +#include <linux/nls.h>
>>>  #include <linux/of.h>
>>>  #include "ufshcd.h"
>>>  #include "unipro.h"
>>> @@ -232,6 +232,16 @@ static inline void ufshcd_disable_irq(struct
>>> ufs_hba *hba)
>>>  	}
>>>  }
>>>
>>> +/* replace non-printable or non-ASCII characters with spaces */
>>> +static inline void ufshcd_remove_non_printable(char *val)
>>> +{
>>> +	if (!val)
>>> +		return;
>>> +
>>> +	if (*val < 0x20 || *val > 0x7e)
>>> +		*val = ' ';
>>> +}
>>> +
>>>  /*
>>>   * ufshcd_wait_for_register - wait for register value to change
>>>   * @hba - per-adapter interface
>>> @@ -2021,6 +2031,82 @@ static inline int ufshcd_read_power_desc(struct
>>> ufs_hba *hba,
>>>  	return ufshcd_read_desc(hba, QUERY_DESC_IDN_POWER, 0, buf, size);
>>>  }
>>>
>>> +int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size)
>>> +{
>>> +	return ufshcd_read_desc(hba, QUERY_DESC_IDN_DEVICE, 0, buf, size);
>>> +}
>>> +EXPORT_SYMBOL(ufshcd_read_device_desc);
>>> +
>>> +/**
>>> + * ufshcd_read_string_desc - read string descriptor
>>> + * @hba: pointer to adapter instance
>>> + * @desc_index: descriptor index
>>> + * @buf: pointer to buffer where descriptor would be read
>>> + * @size: size of buf
>>> + * @ascii: if true convert from unicode to ascii characters
>>> + *
>>> + * Return 0 in case of success, non-zero otherwise
>>> + */
>>> +int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8
>>> *buf,
>>> +				u32 size, bool ascii)
>>> +{
>>> +	int err = 0;
>>> +
>>> +	err = ufshcd_read_desc(hba,
>>> +				QUERY_DESC_IDN_STRING, desc_index, buf, size);
>>> +
>>> +	if (err) {
>>> +		dev_err(hba->dev, "%s: reading String Desc failed after %d retries.
>>> err = %d\n",
>>> +			__func__, QUERY_REQ_RETRIES, err);
>>> +		goto out;
>>> +	}
>>> +
>>> +	if (ascii) {
>>> +		int desc_len;
>>> +		int ascii_len;
>>> +		int i;
>>> +		char *buff_ascii;
>>> +
>>> +		desc_len = buf[0];
>>> +		/* remove header and divide by 2 to move from UTF16 to UTF8 */
>>> +		ascii_len = (desc_len - QUERY_DESC_HDR_SIZE) / 2 + 1;
>>> +		if (size < ascii_len + QUERY_DESC_HDR_SIZE) {
>>> +			dev_err(hba->dev, "%s: buffer allocated size is too small\n",
>>> +					__func__);
>>> +			err = -ENOMEM;
>>> +			goto out;
>>> +		}
>>> +
>>> +		buff_ascii = kmalloc(ascii_len, GFP_KERNEL);
>>> +		if (!buff_ascii) {
>>> +			err = -ENOMEM;
>>> +			goto out_free_buff;
>>> +		}
>>> +
>>> +		/*
>>> +		 * the descriptor contains string in UTF16 format
>>> +		 * we need to convert to utf-8 so it can be displayed
>>> +		 */
>>> +		utf16s_to_utf8s((wchar_t *)&buf[QUERY_DESC_HDR_SIZE],
>>> +				desc_len - QUERY_DESC_HDR_SIZE,
>>> +				UTF16_BIG_ENDIAN, buff_ascii, ascii_len);
>>> +
>>> +		/* replace non-printable or non-ASCII characters with spaces */
>>> +		for (i = 0; i < ascii_len; i++)
>>> +			ufshcd_remove_non_printable(&buff_ascii[i]);
>>> +
>>> +		memset(buf + QUERY_DESC_HDR_SIZE, 0,
>>> +				size - QUERY_DESC_HDR_SIZE);
>>> +		memcpy(buf + QUERY_DESC_HDR_SIZE, buff_ascii, ascii_len);
>>> +		buf[QUERY_DESC_LENGTH_OFFSET] = ascii_len + QUERY_DESC_HDR_SIZE;
>>> +out_free_buff:
>>> +		kfree(buff_ascii);
>>> +	}
>>> +out:
>>> +	return err;
>>> +}
>>> +EXPORT_SYMBOL(ufshcd_read_string_desc);
>>> +
>>>  /**
>>>   * ufshcd_read_unit_desc_param - read the specified unit descriptor
>>> parameter
>>>   * @hba: Pointer to adapter instance
>>> diff --git a/drivers/scsi/ufs/ufshcd.h b/drivers/scsi/ufs/ufshcd.h
>>> index a6d3572..54e13cc 100644
>>> --- a/drivers/scsi/ufs/ufshcd.h
>>> +++ b/drivers/scsi/ufs/ufshcd.h
>>> @@ -678,6 +678,13 @@ static inline int ufshcd_dme_peer_get(struct
>>> ufs_hba *hba,
>>>  	return ufshcd_dme_get_attr(hba, attr_sel, mib_val, DME_PEER);
>>>  }
>>>
>>> +int ufshcd_read_device_desc(struct ufs_hba *hba, u8 *buf, u32 size);
>>> +
>>> +#define ASCII_STD true
>>> +
>>> +int ufshcd_read_string_desc(struct ufs_hba *hba, int desc_index, u8
>>> *buf,
>>> +				u32 size, bool ascii);
>>> +
>>>  /* Expose Query-Request API */
>>>  int ufshcd_query_flag(struct ufs_hba *hba, enum query_opcode opcode,
>>>  	enum flag_idn idn, bool *flag_res);
>>>
>>
>> Probably shows my ignorance, but _where_ is this function actually used?
>>
>> Cheers,
>> Hannes
> 
> 
> Hello Hannes,
> I was asked by reviewer Tomas Winkler, to seperate a patch (v4) into 2
> patches in v5.
> so, this is part I of the seperation, and the functions here, are being
> called by the next patch v5 06/15 - in the quirk.c file in
> ufs_get_device_info() routine.
> hope that helps.
> 
Okay; but you should've put that into the patch description.

Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[ygardi]

> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>> A race condition exists between request requeueing and scsi layer
>> error handling:
>> When UFS driver queuecommand returns a busy status for a request,
>> it will be requeued and its tag will be freed and set to -1.
>> At the same time it is possible that the request will timeout and
>> scsi layer will start error handling for it. The scsi layer reuses
>> the request and its tag to send error related commands to the device,
>> however its tag is no longer valid.
> Hmm. How can the host return a 'busy' status for a request?
> From my understanding we have three possibilities:
>
> 1) queuecommand returns busy; however, that means that the command has
> never been send and this issue shouldn't occur
> 2) The command returns with BUSY status. But in this case it has already
> been returned, so there cannot be any timeout coming in.
> 3) The host receives a command with a tag which is already in-use.
> However, that should have been prevented by the block-layer, which
> really should ensure that this situation never happens.
>
> So either way I look at it, it really looks like a bug and adding a
> timeout handler will just paper over it.
> (Not that a timeout handler is a bad idea, in fact I'm convinced that
> you need one. Just not for this purpose.)
>
> So can you elaborate how this 'busy' status comes about?
> Is the command sent to the device?
>
> Cheers,
>
> Hannes


Hi Hannes,

it's going to be a bit long :)
I think you are missing the point.
I will describe a race condition happened to us a while ago, that was
quite difficult to understand and fix.
So, this patch is not about the "busy" returning to the scsi dispatch
routine. it's about the abort triggered after 30 seconds.

imagine a request being queued and sent to the scsi, and then to the ufs.
a timer, initialized to 30 seconds start ticking.
but the request is never sent to the ufs device, as queuecommand() returns
with "SCSI_MLQUEUE_HOST_BUSY"
by looking at the code, this could happen, for example:
	err = ufshcd_hold(hba, true);
	if (err) {
		err = SCSI_MLQUEUE_HOST_BUSY;
		goto out;
	}

so, now, the request should be re-queued, and its timer should be reset.
(REMEMBER THIS POINT, let's call it "POINT A")
BUT, a context switch happens before it's actually re-queued, and CPU is
moving to other tasks, doing other things for 30 seconds. yes, sounds
crazy, but it did happen.

NOW, the timeout_handler invoked, and the scsi_abort() routine start
executing, (since 30 seconds passed with no completion).
so far, so good.
but hey, another context switch happens, right at the beginning of
scsi_abort() routine, before anything useful happens. (this is "POINT B")
so, now, context is going back "POINT A", to the blk_requeue_request()
routine, that is calling:
blk_delete_timer(rq); (which does nothing cause the timer already expired)
and then it calls:
blk_queue_end_tag()
which place "-1" in the tag field of the request, marking the request, as
"not tagged yet".

however, a context switch happens again, and we are back in scsi_abort()
routine ("POINT B"), that now needs to abort this very request, but hey,
in the "tag" field, what it sees is tag "-1" which is obviously wrong.

this patch fixes this very rare race condition:
1. upon timeout, blk_rq_timed_out() is called
2. then it calls rq_timed_out_fn() which eventually call
the new callback presented in this patch: "ufshcd_eh_timed_out()"
3. this routine returns with the right flag:
BLK_EH_NOT_HANDLED or BLK_EH_RESET_TIMER.
4. blk_rq_timed_out() checks the returned value:
in case of BLK_EH_HANDLED, it handles normally, meaning, calling scsi_abort()
in case of BLK_EH_RESET_TIMER it starts a new timer, and scsi_abort()
never called.

hope that helps.
regards,
Yaniv

> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@suse.de			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
>

Re: [PATCH v5 04/15] scsi: ufs: verify hba controller hce reg value

[ygardi]

> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>> Sometimes due to hw issues it takes some time to the
>> host controller register to update. In order to verify the register
>> has updated, a polling is done until its value is set.
>>
>> In addition the functions ufshcd_hba_stop() and
>> ufshcd_wait_for_register() was updated with an additional input
>> parameter, indicating the timeout between reads will
>> be done by sleeping or spinning the cpu.
>>
>> Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
>> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
>>
>> ---
>>  drivers/scsi/ufs/ufshcd.c | 53
>> ++++++++++++++++++++++++++++-------------------
>>  drivers/scsi/ufs/ufshcd.h | 12 +++--------
>>  2 files changed, 35 insertions(+), 30 deletions(-)
>>
>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>> index 3400ceb..80031e6 100644
>> --- a/drivers/scsi/ufs/ufshcd.c
>> +++ b/drivers/scsi/ufs/ufshcd.c
>> @@ -240,11 +240,13 @@ static inline void ufshcd_disable_irq(struct
>> ufs_hba *hba)
>>   * @val - wait condition
>>   * @interval_us - polling interval in microsecs
>>   * @timeout_ms - timeout in millisecs
>> + * @can_sleep - perform sleep or just spin
>>   *
>>   * Returns -ETIMEDOUT on error, zero on success
>>   */
>> -static int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32
>> mask,
>> -		u32 val, unsigned long interval_us, unsigned long timeout_ms)
>> +int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
>> +				u32 val, unsigned long interval_us,
>> +				unsigned long timeout_ms, bool can_sleep)
>>  {
>>  	int err = 0;
>>  	unsigned long timeout = jiffies + msecs_to_jiffies(timeout_ms);
>> @@ -253,9 +255,10 @@ static int ufshcd_wait_for_register(struct ufs_hba
>> *hba, u32 reg, u32 mask,
>>  	val = val & mask;
>>
>>  	while ((ufshcd_readl(hba, reg) & mask) != val) {
>> -		/* wakeup within 50us of expiry */
>> -		usleep_range(interval_us, interval_us + 50);
>> -
>> +		if (can_sleep)
>> +			usleep_range(interval_us, interval_us + 50);
>> +		else
>> +			udelay(interval_us);
>>  		if (time_after(jiffies, timeout)) {
>>  			if ((ufshcd_readl(hba, reg) & mask) != val)
>>  				err = -ETIMEDOUT;
>> @@ -1459,7 +1462,7 @@ ufshcd_clear_cmd(struct ufs_hba *hba, int tag)
>>  	 */
>>  	err = ufshcd_wait_for_register(hba,
>>  			REG_UTP_TRANSFER_REQ_DOOR_BELL,
>> -			mask, ~mask, 1000, 1000);
>> +			mask, ~mask, 1000, 1000, true);
>>
>>  	return err;
>>  }
>> @@ -2815,6 +2818,23 @@ out:
>>  }
>>
>>  /**
>> + * ufshcd_hba_stop - Send controller to reset state
>> + * @hba: per adapter instance
>> + * @can_sleep: perform sleep or just spin
>> + */
>> +static inline void ufshcd_hba_stop(struct ufs_hba *hba, bool can_sleep)
>> +{
>> +	int err;
>> +
>> +	ufshcd_writel(hba, CONTROLLER_DISABLE,  REG_CONTROLLER_ENABLE);
>> +	err = ufshcd_wait_for_register(hba, REG_CONTROLLER_ENABLE,
>> +					CONTROLLER_ENABLE, CONTROLLER_DISABLE,
>> +					10, 1, can_sleep);
>> +	if (err)
>> +		dev_err(hba->dev, "%s: Controller disable failed\n", __func__);
>> +}
>> +
> Shouldn't you return an error here?
> If the controller disable failed you probably need a hard reset or
> something, otherwise I would assume that every other command from that
> point on will not work as expected.
>
> Cheers,
>
> Hannes


Hello Hannes,
The original routine signature is:
void ufshcd_hba_stop(struct ufs_hba *hba);

as you can see, no return value, the reason is simple - there is nothing
we can do if writing to the register fails.

all we wanted to do here, was to add a graceful time to change the
register value. also, we decided to add error msg in case the value is not
change within this timeout.
We can not do anything else, not to say, return error, as there is no
error handling in such case.

So, as far as i see it, we only improved the already exists logic, by
adding some graceful time to the register change, and also, by adding an
error message that was absent before.

thanks,
Yaniv




> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@suse.de			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[Hannes Reinecke]

On 03/01/2016 09:25 PM, ygardi@codeaurora.org wrote:
>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>> A race condition exists between request requeueing and scsi layer
>>> error handling:
>>> When UFS driver queuecommand returns a busy status for a request,
>>> it will be requeued and its tag will be freed and set to -1.
>>> At the same time it is possible that the request will timeout and
>>> scsi layer will start error handling for it. The scsi layer reuses
>>> the request and its tag to send error related commands to the device,
>>> however its tag is no longer valid.
>> Hmm. How can the host return a 'busy' status for a request?
>> From my understanding we have three possibilities:
>>
>> 1) queuecommand returns busy; however, that means that the command has
>> never been send and this issue shouldn't occur
>> 2) The command returns with BUSY status. But in this case it has already
>> been returned, so there cannot be any timeout coming in.
>> 3) The host receives a command with a tag which is already in-use.
>> However, that should have been prevented by the block-layer, which
>> really should ensure that this situation never happens.
>>
>> So either way I look at it, it really looks like a bug and adding a
>> timeout handler will just paper over it.
>> (Not that a timeout handler is a bad idea, in fact I'm convinced that
>> you need one. Just not for this purpose.)
>>
>> So can you elaborate how this 'busy' status comes about?
>> Is the command sent to the device?
>>
>> Cheers,
>>
>> Hannes
> 
> 
> Hi Hannes,
> 
> it's going to be a bit long :)
> I think you are missing the point.
> I will describe a race condition happened to us a while ago, that was
> quite difficult to understand and fix.
> So, this patch is not about the "busy" returning to the scsi dispatch
> routine. it's about the abort triggered after 30 seconds.
> 
> imagine a request being queued and sent to the scsi, and then to the ufs.
> a timer, initialized to 30 seconds start ticking.
> but the request is never sent to the ufs device, as queuecommand() returns
> with "SCSI_MLQUEUE_HOST_BUSY"
> by looking at the code, this could happen, for example:
> 	err = ufshcd_hold(hba, true);
> 	if (err) {
> 		err = SCSI_MLQUEUE_HOST_BUSY;
> 		goto out;
> 	}
> 
Uuhhh.
You probably should not have pointed me to that piece of code ...
open-coding loops in ufshcd_hold() ... shudder.
(Did I ever review that one? Must've ...)
_Anyway_: sleeping in queuecommand is always a bad idea, as then
precisely those issues you've just described will happen.

Couldn't you just call
ufshcd_hold(hba, false)
instead of
ufshcd_hold(hba, true)
?
The request will be requeued more-or-less immediately, avoiding the
issue with timeout handler kicking in.
And the queue will remain blocked until the ungate work item returns, at
which point I/O submission will continue.
As the request will be requeued to the head of the queue there won't be
other I/O competing with tags, so it shouldn't have any adverse effects.

Wouldn't that work?

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 04/15] scsi: ufs: verify hba controller hce reg value

[Hannes Reinecke]

On 03/01/2016 09:32 PM, ygardi@codeaurora.org wrote:
>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>> Sometimes due to hw issues it takes some time to the
>>> host controller register to update. In order to verify the register
>>> has updated, a polling is done until its value is set.
>>>
>>> In addition the functions ufshcd_hba_stop() and
>>> ufshcd_wait_for_register() was updated with an additional input
>>> parameter, indicating the timeout between reads will
>>> be done by sleeping or spinning the cpu.
>>>
>>> Signed-off-by: Raviv Shvili <rshvili@codeaurora.org>
>>> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
>>>
>>> ---
>>>  drivers/scsi/ufs/ufshcd.c | 53
>>> ++++++++++++++++++++++++++++-------------------
>>>  drivers/scsi/ufs/ufshcd.h | 12 +++--------
>>>  2 files changed, 35 insertions(+), 30 deletions(-)
>>>
>>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>>> index 3400ceb..80031e6 100644
>>> --- a/drivers/scsi/ufs/ufshcd.c
>>> +++ b/drivers/scsi/ufs/ufshcd.c
>>> @@ -240,11 +240,13 @@ static inline void ufshcd_disable_irq(struct
>>> ufs_hba *hba)
>>>   * @val - wait condition
>>>   * @interval_us - polling interval in microsecs
>>>   * @timeout_ms - timeout in millisecs
>>> + * @can_sleep - perform sleep or just spin
>>>   *
>>>   * Returns -ETIMEDOUT on error, zero on success
>>>   */
>>> -static int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32
>>> mask,
>>> -		u32 val, unsigned long interval_us, unsigned long timeout_ms)
>>> +int ufshcd_wait_for_register(struct ufs_hba *hba, u32 reg, u32 mask,
>>> +				u32 val, unsigned long interval_us,
>>> +				unsigned long timeout_ms, bool can_sleep)
>>>  {
>>>  	int err = 0;
>>>  	unsigned long timeout = jiffies + msecs_to_jiffies(timeout_ms);
>>> @@ -253,9 +255,10 @@ static int ufshcd_wait_for_register(struct ufs_hba
>>> *hba, u32 reg, u32 mask,
>>>  	val = val & mask;
>>>
>>>  	while ((ufshcd_readl(hba, reg) & mask) != val) {
>>> -		/* wakeup within 50us of expiry */
>>> -		usleep_range(interval_us, interval_us + 50);
>>> -
>>> +		if (can_sleep)
>>> +			usleep_range(interval_us, interval_us + 50);
>>> +		else
>>> +			udelay(interval_us);
>>>  		if (time_after(jiffies, timeout)) {
>>>  			if ((ufshcd_readl(hba, reg) & mask) != val)
>>>  				err = -ETIMEDOUT;
>>> @@ -1459,7 +1462,7 @@ ufshcd_clear_cmd(struct ufs_hba *hba, int tag)
>>>  	 */
>>>  	err = ufshcd_wait_for_register(hba,
>>>  			REG_UTP_TRANSFER_REQ_DOOR_BELL,
>>> -			mask, ~mask, 1000, 1000);
>>> +			mask, ~mask, 1000, 1000, true);
>>>
>>>  	return err;
>>>  }
>>> @@ -2815,6 +2818,23 @@ out:
>>>  }
>>>
>>>  /**
>>> + * ufshcd_hba_stop - Send controller to reset state
>>> + * @hba: per adapter instance
>>> + * @can_sleep: perform sleep or just spin
>>> + */
>>> +static inline void ufshcd_hba_stop(struct ufs_hba *hba, bool can_sleep)
>>> +{
>>> +	int err;
>>> +
>>> +	ufshcd_writel(hba, CONTROLLER_DISABLE,  REG_CONTROLLER_ENABLE);
>>> +	err = ufshcd_wait_for_register(hba, REG_CONTROLLER_ENABLE,
>>> +					CONTROLLER_ENABLE, CONTROLLER_DISABLE,
>>> +					10, 1, can_sleep);
>>> +	if (err)
>>> +		dev_err(hba->dev, "%s: Controller disable failed\n", __func__);
>>> +}
>>> +
>> Shouldn't you return an error here?
>> If the controller disable failed you probably need a hard reset or
>> something, otherwise I would assume that every other command from that
>> point on will not work as expected.
>>
>> Cheers,
>>
>> Hannes
> 
> 
> Hello Hannes,
> The original routine signature is:
> void ufshcd_hba_stop(struct ufs_hba *hba);
> 
> as you can see, no return value, the reason is simple - there is nothing
> we can do if writing to the register fails.
> 
> all we wanted to do here, was to add a graceful time to change the
> register value. also, we decided to add error msg in case the value is not
> change within this timeout.
> We can not do anything else, not to say, return error, as there is no
> error handling in such case.
> 
> So, as far as i see it, we only improved the already exists logic, by
> adding some graceful time to the register change, and also, by adding an
> error message that was absent before.
> 
Thanks for the explanation.

Reviewed-by: Hannes Reinecke <hare@suse.com>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[ygardi]

> On 03/01/2016 09:25 PM, ygardi@codeaurora.org wrote:
>>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>>> A race condition exists between request requeueing and scsi layer
>>>> error handling:
>>>> When UFS driver queuecommand returns a busy status for a request,
>>>> it will be requeued and its tag will be freed and set to -1.
>>>> At the same time it is possible that the request will timeout and
>>>> scsi layer will start error handling for it. The scsi layer reuses
>>>> the request and its tag to send error related commands to the device,
>>>> however its tag is no longer valid.
>>> Hmm. How can the host return a 'busy' status for a request?
>>> From my understanding we have three possibilities:
>>>
>>> 1) queuecommand returns busy; however, that means that the command has
>>> never been send and this issue shouldn't occur
>>> 2) The command returns with BUSY status. But in this case it has
>>> already
>>> been returned, so there cannot be any timeout coming in.
>>> 3) The host receives a command with a tag which is already in-use.
>>> However, that should have been prevented by the block-layer, which
>>> really should ensure that this situation never happens.
>>>
>>> So either way I look at it, it really looks like a bug and adding a
>>> timeout handler will just paper over it.
>>> (Not that a timeout handler is a bad idea, in fact I'm convinced that
>>> you need one. Just not for this purpose.)
>>>
>>> So can you elaborate how this 'busy' status comes about?
>>> Is the command sent to the device?
>>>
>>> Cheers,
>>>
>>> Hannes
>>
>>
>> Hi Hannes,
>>
>> it's going to be a bit long :)
>> I think you are missing the point.
>> I will describe a race condition happened to us a while ago, that was
>> quite difficult to understand and fix.
>> So, this patch is not about the "busy" returning to the scsi dispatch
>> routine. it's about the abort triggered after 30 seconds.
>>
>> imagine a request being queued and sent to the scsi, and then to the
>> ufs.
>> a timer, initialized to 30 seconds start ticking.
>> but the request is never sent to the ufs device, as queuecommand()
>> returns
>> with "SCSI_MLQUEUE_HOST_BUSY"
>> by looking at the code, this could happen, for example:
>> 	err = ufshcd_hold(hba, true);
>> 	if (err) {
>> 		err = SCSI_MLQUEUE_HOST_BUSY;
>> 		goto out;
>> 	}
>>
> Uuhhh.
> You probably should not have pointed me to that piece of code ...
> open-coding loops in ufshcd_hold() ... shudder.
> (Did I ever review that one? Must've ...)
> _Anyway_: sleeping in queuecommand is always a bad idea, as then
> precisely those issues you've just described will happen.
>
> Couldn't you just call
> ufshcd_hold(hba, false)
> instead of
> ufshcd_hold(hba, true)
> ?
> The request will be requeued more-or-less immediately, avoiding the
> issue with timeout handler kicking in.
> And the queue will remain blocked until the ungate work item returns, at
> which point I/O submission will continue.
> As the request will be requeued to the head of the queue there won't be
> other I/O competing with tags, so it shouldn't have any adverse effects.
>
> Wouldn't that work?
>
> Cheers,
>
> Hannes

Hi Hannes

This is a bug, and it should be fixed.
if you choose to bypass it, by calling ufshcd_hold(hba, false), not only
the race condition is still there, and can pop-out at any other point in
the future, but also, not sure what are the consequences of
ufshcd_hold(hba, false) unstead of "true".
so, changing the already tested and working code, (not to return BUSY from
queuecommand) is not a fix.
I strongly recommend we upstream this race-condition fix.

thanks,
Yaniv



> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@suse.de			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[Hannes Reinecke]

On 03/03/2016 05:10 PM, ygardi@codeaurora.org wrote:
>> On 03/01/2016 09:25 PM, ygardi@codeaurora.org wrote:
>>>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>>>> A race condition exists between request requeueing and scsi layer
>>>>> error handling:
>>>>> When UFS driver queuecommand returns a busy status for a request,
>>>>> it will be requeued and its tag will be freed and set to -1.
>>>>> At the same time it is possible that the request will timeout and
>>>>> scsi layer will start error handling for it. The scsi layer reuses
>>>>> the request and its tag to send error related commands to the device,
>>>>> however its tag is no longer valid.
>>>> Hmm. How can the host return a 'busy' status for a request?
>>>> From my understanding we have three possibilities:
>>>>
>>>> 1) queuecommand returns busy; however, that means that the command has
>>>> never been send and this issue shouldn't occur
>>>> 2) The command returns with BUSY status. But in this case it has
>>>> already
>>>> been returned, so there cannot be any timeout coming in.
>>>> 3) The host receives a command with a tag which is already in-use.
>>>> However, that should have been prevented by the block-layer, which
>>>> really should ensure that this situation never happens.
>>>>
>>>> So either way I look at it, it really looks like a bug and adding a
>>>> timeout handler will just paper over it.
>>>> (Not that a timeout handler is a bad idea, in fact I'm convinced that
>>>> you need one. Just not for this purpose.)
>>>>
>>>> So can you elaborate how this 'busy' status comes about?
>>>> Is the command sent to the device?
>>>>
>>>> Cheers,
>>>>
>>>> Hannes
>>>
>>>
>>> Hi Hannes,
>>>
>>> it's going to be a bit long :)
>>> I think you are missing the point.
>>> I will describe a race condition happened to us a while ago, that was
>>> quite difficult to understand and fix.
>>> So, this patch is not about the "busy" returning to the scsi dispatch
>>> routine. it's about the abort triggered after 30 seconds.
>>>
>>> imagine a request being queued and sent to the scsi, and then to the
>>> ufs.
>>> a timer, initialized to 30 seconds start ticking.
>>> but the request is never sent to the ufs device, as queuecommand()
>>> returns
>>> with "SCSI_MLQUEUE_HOST_BUSY"
>>> by looking at the code, this could happen, for example:
>>> 	err = ufshcd_hold(hba, true);
>>> 	if (err) {
>>> 		err = SCSI_MLQUEUE_HOST_BUSY;
>>> 		goto out;
>>> 	}
>>>
>> Uuhhh.
>> You probably should not have pointed me to that piece of code ...
>> open-coding loops in ufshcd_hold() ... shudder.
>> (Did I ever review that one? Must've ...)
>> _Anyway_: sleeping in queuecommand is always a bad idea, as then
>> precisely those issues you've just described will happen.
>>
>> Couldn't you just call
>> ufshcd_hold(hba, false)
>> instead of
>> ufshcd_hold(hba, true)
>> ?
>> The request will be requeued more-or-less immediately, avoiding the
>> issue with timeout handler kicking in.
>> And the queue will remain blocked until the ungate work item returns, at
>> which point I/O submission will continue.
>> As the request will be requeued to the head of the queue there won't be
>> other I/O competing with tags, so it shouldn't have any adverse effects.
>>
>> Wouldn't that work?
>>
>> Cheers,
>>
>> Hannes
> 
> Hi Hannes
> 
> This is a bug, and it should be fixed.
Oh, definitely agreed. The question is _where_.


> if you choose to bypass it, by calling ufshcd_hold(hba, false), not only
> the race condition is still there, and can pop-out at any other point in
> the future, but also, not sure what are the consequences of
> ufshcd_hold(hba, false) unstead of "true".
Well ... seeing it's your driver, I would've thought _you_ should know ...

> so, changing the already tested and working code, (not to return BUSY from
> queuecommand) is not a fix.
Hey, I did _not_ suggest not to retury BUSY from queuecommand.

I was suggesting this patch:

diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
index 9c1b94b..b9295ad 100644
--- a/drivers/scsi/ufs/ufshcd.c
+++ b/drivers/scsi/ufs/ufshcd.c
@@ -1388,7 +1388,7 @@ static int ufshcd_queuecommand(struct Scsi_Host
*host, struct scsi_cmnd *cmd)
                goto out;
        }

-       err = ufshcd_hold(hba, true);
+       err = ufshcd_hold(hba, false);
        if (err) {
                err = SCSI_MLQUEUE_HOST_BUSY;
                clear_bit_unlock(tag, &hba->lrb_in_use);

which, by reading the code, should be avoiding this issue.
I was just asking you if you could give this patch a spin and see if it
works. If not (for whatever reason) I'm happy to accept your patch.
But first I would like to have an explanation why the above would _not_
work.

Unfortunately I don't have the hardware otherwise I'd be running the
tests myself.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke		      zSeries & Storage
hare@suse.de			      +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)

Re: [PATCH v5 01/15] scsi: ufs-qcom: add number of lanes per direction

[Rob Herring]

On Sun, Feb 28, 2016 at 03:32:33PM +0200, Yaniv Gardi wrote:
> Different platform may have different number of lanes
> for the UFS link.
> Add parameter to device tree specifying how many lanes
> should be configured for the UFS link.
> 
> Signed-off-by: Gilad Broner <gbroner@codeaurora.org>
> Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
> 
> ---
>  .../devicetree/bindings/ufs/ufshcd-pltfrm.txt      |  3 ++

Acked-by: Rob Herring <robh@kernel.org>

>  drivers/scsi/ufs/ufs-qcom.c                        | 39 ++++++++++++----------
>  drivers/scsi/ufs/ufshcd-pltfrm.c                   | 19 +++++++++++
>  drivers/scsi/ufs/ufshcd.c                          |  1 +
>  drivers/scsi/ufs/ufshcd.h                          |  2 ++
>  5 files changed, 47 insertions(+), 17 deletions(-)

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[ygardi]

> On 03/03/2016 05:10 PM, ygardi@codeaurora.org wrote:
>>> On 03/01/2016 09:25 PM, ygardi@codeaurora.org wrote:
>>>>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>>>>> A race condition exists between request requeueing and scsi layer
>>>>>> error handling:
>>>>>> When UFS driver queuecommand returns a busy status for a request,
>>>>>> it will be requeued and its tag will be freed and set to -1.
>>>>>> At the same time it is possible that the request will timeout and
>>>>>> scsi layer will start error handling for it. The scsi layer reuses
>>>>>> the request and its tag to send error related commands to the
>>>>>> device,
>>>>>> however its tag is no longer valid.
>>>>> Hmm. How can the host return a 'busy' status for a request?
>>>>> From my understanding we have three possibilities:
>>>>>
>>>>> 1) queuecommand returns busy; however, that means that the command
>>>>> has
>>>>> never been send and this issue shouldn't occur
>>>>> 2) The command returns with BUSY status. But in this case it has
>>>>> already
>>>>> been returned, so there cannot be any timeout coming in.
>>>>> 3) The host receives a command with a tag which is already in-use.
>>>>> However, that should have been prevented by the block-layer, which
>>>>> really should ensure that this situation never happens.
>>>>>
>>>>> So either way I look at it, it really looks like a bug and adding a
>>>>> timeout handler will just paper over it.
>>>>> (Not that a timeout handler is a bad idea, in fact I'm convinced that
>>>>> you need one. Just not for this purpose.)
>>>>>
>>>>> So can you elaborate how this 'busy' status comes about?
>>>>> Is the command sent to the device?
>>>>>
>>>>> Cheers,
>>>>>
>>>>> Hannes
>>>>
>>>>
>>>> Hi Hannes,
>>>>
>>>> it's going to be a bit long :)
>>>> I think you are missing the point.
>>>> I will describe a race condition happened to us a while ago, that was
>>>> quite difficult to understand and fix.
>>>> So, this patch is not about the "busy" returning to the scsi dispatch
>>>> routine. it's about the abort triggered after 30 seconds.
>>>>
>>>> imagine a request being queued and sent to the scsi, and then to the
>>>> ufs.
>>>> a timer, initialized to 30 seconds start ticking.
>>>> but the request is never sent to the ufs device, as queuecommand()
>>>> returns
>>>> with "SCSI_MLQUEUE_HOST_BUSY"
>>>> by looking at the code, this could happen, for example:
>>>> 	err = ufshcd_hold(hba, true);
>>>> 	if (err) {
>>>> 		err = SCSI_MLQUEUE_HOST_BUSY;
>>>> 		goto out;
>>>> 	}
>>>>
>>> Uuhhh.
>>> You probably should not have pointed me to that piece of code ...
>>> open-coding loops in ufshcd_hold() ... shudder.
>>> (Did I ever review that one? Must've ...)
>>> _Anyway_: sleeping in queuecommand is always a bad idea, as then
>>> precisely those issues you've just described will happen.
>>>
>>> Couldn't you just call
>>> ufshcd_hold(hba, false)
>>> instead of
>>> ufshcd_hold(hba, true)
>>> ?
>>> The request will be requeued more-or-less immediately, avoiding the
>>> issue with timeout handler kicking in.
>>> And the queue will remain blocked until the ungate work item returns,
>>> at
>>> which point I/O submission will continue.
>>> As the request will be requeued to the head of the queue there won't be
>>> other I/O competing with tags, so it shouldn't have any adverse
>>> effects.
>>>
>>> Wouldn't that work?
>>>
>>> Cheers,
>>>
>>> Hannes
>>
>> Hi Hannes
>>
>> This is a bug, and it should be fixed.
> Oh, definitely agreed. The question is _where_.
>
>
>> if you choose to bypass it, by calling ufshcd_hold(hba, false), not only
>> the race condition is still there, and can pop-out at any other point in
>> the future, but also, not sure what are the consequences of
>> ufshcd_hold(hba, false) unstead of "true".
> Well ... seeing it's your driver, I would've thought _you_ should know ...
>
>> so, changing the already tested and working code, (not to return BUSY
>> from
>> queuecommand) is not a fix.
> Hey, I did _not_ suggest not to retury BUSY from queuecommand.
>
> I was suggesting this patch:
>
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index 9c1b94b..b9295ad 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -1388,7 +1388,7 @@ static int ufshcd_queuecommand(struct Scsi_Host
> *host, struct scsi_cmnd *cmd)
>                 goto out;
>         }
>
> -       err = ufshcd_hold(hba, true);
> +       err = ufshcd_hold(hba, false);
>         if (err) {
>                 err = SCSI_MLQUEUE_HOST_BUSY;
>                 clear_bit_unlock(tag, &hba->lrb_in_use);
>
> which, by reading the code, should be avoiding this issue.


Hannes,
we are not trying to avoid returning BUSY from queuecommand().
On the contrary. By returning BUSY we actually re-queuing the request
which is exactly what we need to do.
your patch doesn't fix the race condition.

thanks,
Yaniv

> I was just asking you if you could give this patch a spin and see if it
> works. If not (for whatever reason) I'm happy to accept your patch.
> But first I would like to have an explanation why the above would _not_
> work.
>
> Unfortunately I don't have the hardware otherwise I'd be running the
> tests myself.
>
> Cheers,
>
> Hannes
> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@suse.de			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>


--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v5 15/15] scsi: ufs-qcom: set PA_Local_TX_LCC_Enable before link startup

[ygardi]

> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>> Some UFS devices (and may be host) have issues if LCC is
>> enabled. So we are setting PA_Local_TX_LCC_Enable to 0
>> before link startup which will make sure that both host
>> and device TX LCC are disabled once link startup is
>> completed.
>>
>> This change also:
>> - enables the device ref clock before changing to HS mode
>>   and disables it if entered to PWM mode
>> - adds printouts of testbus debug registers
>>
> Please make that in two patches; these issues are unrelated and will
> just confuse users if one ever need to dig through the git log.
>

done. in V6 this patch will be divided into 3 patches.
1. LX_LCC
2. device ref clock
3. debug printouts

thanks,
Yaniv

> Cheers,
>
> Hannes
> --
> Dr. Hannes Reinecke		      zSeries & Storage
> hare@suse.de			      +49 911 74053 688
> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>


--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[ygardi]

Hello, Hannes,

Re-sending

thanks,
Yaniv

>> On 03/03/2016 05:10 PM, ygardi@codeaurora.org wrote:
>>>> On 03/01/2016 09:25 PM, ygardi@codeaurora.org wrote:
>>>>>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>>>>>> A race condition exists between request requeueing and scsi layer
>>>>>>> error handling:
>>>>>>> When UFS driver queuecommand returns a busy status for a request,
>>>>>>> it will be requeued and its tag will be freed and set to -1.
>>>>>>> At the same time it is possible that the request will timeout and
>>>>>>> scsi layer will start error handling for it. The scsi layer reuses
>>>>>>> the request and its tag to send error related commands to the
>>>>>>> device,
>>>>>>> however its tag is no longer valid.
>>>>>> Hmm. How can the host return a 'busy' status for a request?
>>>>>> From my understanding we have three possibilities:
>>>>>>
>>>>>> 1) queuecommand returns busy; however, that means that the command
>>>>>> has
>>>>>> never been send and this issue shouldn't occur
>>>>>> 2) The command returns with BUSY status. But in this case it has
>>>>>> already
>>>>>> been returned, so there cannot be any timeout coming in.
>>>>>> 3) The host receives a command with a tag which is already in-use.
>>>>>> However, that should have been prevented by the block-layer, which
>>>>>> really should ensure that this situation never happens.
>>>>>>
>>>>>> So either way I look at it, it really looks like a bug and adding a
>>>>>> timeout handler will just paper over it.
>>>>>> (Not that a timeout handler is a bad idea, in fact I'm convinced
>>>>>> that
>>>>>> you need one. Just not for this purpose.)
>>>>>>
>>>>>> So can you elaborate how this 'busy' status comes about?
>>>>>> Is the command sent to the device?
>>>>>>
>>>>>> Cheers,
>>>>>>
>>>>>> Hannes
>>>>>
>>>>>
>>>>> Hi Hannes,
>>>>>
>>>>> it's going to be a bit long :)
>>>>> I think you are missing the point.
>>>>> I will describe a race condition happened to us a while ago, that was
>>>>> quite difficult to understand and fix.
>>>>> So, this patch is not about the "busy" returning to the scsi dispatch
>>>>> routine. it's about the abort triggered after 30 seconds.
>>>>>
>>>>> imagine a request being queued and sent to the scsi, and then to the
>>>>> ufs.
>>>>> a timer, initialized to 30 seconds start ticking.
>>>>> but the request is never sent to the ufs device, as queuecommand()
>>>>> returns
>>>>> with "SCSI_MLQUEUE_HOST_BUSY"
>>>>> by looking at the code, this could happen, for example:
>>>>> 	err = ufshcd_hold(hba, true);
>>>>> 	if (err) {
>>>>> 		err = SCSI_MLQUEUE_HOST_BUSY;
>>>>> 		goto out;
>>>>> 	}
>>>>>
>>>> Uuhhh.
>>>> You probably should not have pointed me to that piece of code ...
>>>> open-coding loops in ufshcd_hold() ... shudder.
>>>> (Did I ever review that one? Must've ...)
>>>> _Anyway_: sleeping in queuecommand is always a bad idea, as then
>>>> precisely those issues you've just described will happen.
>>>>
>>>> Couldn't you just call
>>>> ufshcd_hold(hba, false)
>>>> instead of
>>>> ufshcd_hold(hba, true)
>>>> ?
>>>> The request will be requeued more-or-less immediately, avoiding the
>>>> issue with timeout handler kicking in.
>>>> And the queue will remain blocked until the ungate work item returns,
>>>> at
>>>> which point I/O submission will continue.
>>>> As the request will be requeued to the head of the queue there won't
>>>> be
>>>> other I/O competing with tags, so it shouldn't have any adverse
>>>> effects.
>>>>
>>>> Wouldn't that work?
>>>>
>>>> Cheers,
>>>>
>>>> Hannes
>>>
>>> Hi Hannes
>>>
>>> This is a bug, and it should be fixed.
>> Oh, definitely agreed. The question is _where_.
>>
>>
>>> if you choose to bypass it, by calling ufshcd_hold(hba, false), not
>>> only
>>> the race condition is still there, and can pop-out at any other point
>>> in
>>> the future, but also, not sure what are the consequences of
>>> ufshcd_hold(hba, false) unstead of "true".
>> Well ... seeing it's your driver, I would've thought _you_ should know
>> ...
>>
>>> so, changing the already tested and working code, (not to return BUSY
>>> from
>>> queuecommand) is not a fix.
>> Hey, I did _not_ suggest not to retury BUSY from queuecommand.
>>
>> I was suggesting this patch:
>>
>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>> index 9c1b94b..b9295ad 100644
>> --- a/drivers/scsi/ufs/ufshcd.c
>> +++ b/drivers/scsi/ufs/ufshcd.c
>> @@ -1388,7 +1388,7 @@ static int ufshcd_queuecommand(struct Scsi_Host
>> *host, struct scsi_cmnd *cmd)
>>                 goto out;
>>         }
>>
>> -       err = ufshcd_hold(hba, true);
>> +       err = ufshcd_hold(hba, false);
>>         if (err) {
>>                 err = SCSI_MLQUEUE_HOST_BUSY;
>>                 clear_bit_unlock(tag, &hba->lrb_in_use);
>>
>> which, by reading the code, should be avoiding this issue.
>
>
> Hannes,
> we are not trying to avoid returning BUSY from queuecommand().
> On the contrary. By returning BUSY we actually re-queuing the request
> which is exactly what we need to do.
> your patch doesn't fix the race condition.
>
> thanks,
> Yaniv
>
>> I was just asking you if you could give this patch a spin and see if it
>> works. If not (for whatever reason) I'm happy to accept your patch.
>> But first I would like to have an explanation why the above would _not_
>> work.
>>
>> Unfortunately I don't have the hardware otherwise I'd be running the
>> tests myself.
>>
>> Cheers,
>>
>> Hannes
>> --
>> Dr. Hannes Reinecke		      zSeries & Storage
>> hare@suse.de			      +49 911 74053 688
>> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
>> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

Re: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[ygardi]

Hello, Hannes,

Re-sending

thanks,
Yaniv

>> On 03/03/2016 05:10 PM, ygardi@codeaurora.org wrote:
>>>> On 03/01/2016 09:25 PM, ygardi@codeaurora.org wrote:
>>>>>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>>>>>> A race condition exists between request requeueing and scsi layer
>>>>>>> error handling:
>>>>>>> When UFS driver queuecommand returns a busy status for a request,
>>>>>>> it will be requeued and its tag will be freed and set to -1.
>>>>>>> At the same time it is possible that the request will timeout and
>>>>>>> scsi layer will start error handling for it. The scsi layer reuses
>>>>>>> the request and its tag to send error related commands to the
>>>>>>> device,
>>>>>>> however its tag is no longer valid.
>>>>>> Hmm. How can the host return a 'busy' status for a request?
>>>>>> From my understanding we have three possibilities:
>>>>>>
>>>>>> 1) queuecommand returns busy; however, that means that the command
>>>>>> has
>>>>>> never been send and this issue shouldn't occur
>>>>>> 2) The command returns with BUSY status. But in this case it has
>>>>>> already
>>>>>> been returned, so there cannot be any timeout coming in.
>>>>>> 3) The host receives a command with a tag which is already in-use.
>>>>>> However, that should have been prevented by the block-layer, which
>>>>>> really should ensure that this situation never happens.
>>>>>>
>>>>>> So either way I look at it, it really looks like a bug and adding a
>>>>>> timeout handler will just paper over it.
>>>>>> (Not that a timeout handler is a bad idea, in fact I'm convinced
>>>>>> that
>>>>>> you need one. Just not for this purpose.)
>>>>>>
>>>>>> So can you elaborate how this 'busy' status comes about?
>>>>>> Is the command sent to the device?
>>>>>>
>>>>>> Cheers,
>>>>>>
>>>>>> Hannes
>>>>>
>>>>>
>>>>> Hi Hannes,
>>>>>
>>>>> it's going to be a bit long :)
>>>>> I think you are missing the point.
>>>>> I will describe a race condition happened to us a while ago, that was
>>>>> quite difficult to understand and fix.
>>>>> So, this patch is not about the "busy" returning to the scsi dispatch
>>>>> routine. it's about the abort triggered after 30 seconds.
>>>>>
>>>>> imagine a request being queued and sent to the scsi, and then to the
>>>>> ufs.
>>>>> a timer, initialized to 30 seconds start ticking.
>>>>> but the request is never sent to the ufs device, as queuecommand()
>>>>> returns
>>>>> with "SCSI_MLQUEUE_HOST_BUSY"
>>>>> by looking at the code, this could happen, for example:
>>>>> 	err = ufshcd_hold(hba, true);
>>>>> 	if (err) {
>>>>> 		err = SCSI_MLQUEUE_HOST_BUSY;
>>>>> 		goto out;
>>>>> 	}
>>>>>
>>>> Uuhhh.
>>>> You probably should not have pointed me to that piece of code ...
>>>> open-coding loops in ufshcd_hold() ... shudder.
>>>> (Did I ever review that one? Must've ...)
>>>> _Anyway_: sleeping in queuecommand is always a bad idea, as then
>>>> precisely those issues you've just described will happen.
>>>>
>>>> Couldn't you just call
>>>> ufshcd_hold(hba, false)
>>>> instead of
>>>> ufshcd_hold(hba, true)
>>>> ?
>>>> The request will be requeued more-or-less immediately, avoiding the
>>>> issue with timeout handler kicking in.
>>>> And the queue will remain blocked until the ungate work item returns,
>>>> at
>>>> which point I/O submission will continue.
>>>> As the request will be requeued to the head of the queue there won't
>>>> be
>>>> other I/O competing with tags, so it shouldn't have any adverse
>>>> effects.
>>>>
>>>> Wouldn't that work?
>>>>
>>>> Cheers,
>>>>
>>>> Hannes
>>>
>>> Hi Hannes
>>>
>>> This is a bug, and it should be fixed.
>> Oh, definitely agreed. The question is _where_.
>>
>>
>>> if you choose to bypass it, by calling ufshcd_hold(hba, false), not
>>> only
>>> the race condition is still there, and can pop-out at any other point
>>> in
>>> the future, but also, not sure what are the consequences of
>>> ufshcd_hold(hba, false) unstead of "true".
>> Well ... seeing it's your driver, I would've thought _you_ should know
>> ...
>>
>>> so, changing the already tested and working code, (not to return BUSY
>>> from
>>> queuecommand) is not a fix.
>> Hey, I did _not_ suggest not to retury BUSY from queuecommand.
>>
>> I was suggesting this patch:
>>
>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
>> index 9c1b94b..b9295ad 100644
>> --- a/drivers/scsi/ufs/ufshcd.c
>> +++ b/drivers/scsi/ufs/ufshcd.c
>> @@ -1388,7 +1388,7 @@ static int ufshcd_queuecommand(struct Scsi_Host
>> *host, struct scsi_cmnd *cmd)
>>                 goto out;
>>         }
>>
>> -       err = ufshcd_hold(hba, true);
>> +       err = ufshcd_hold(hba, false);
>>         if (err) {
>>                 err = SCSI_MLQUEUE_HOST_BUSY;
>>                 clear_bit_unlock(tag, &hba->lrb_in_use);
>>
>> which, by reading the code, should be avoiding this issue.
>
>
> Hannes,
> we are not trying to avoid returning BUSY from queuecommand().
> On the contrary. By returning BUSY we actually re-queuing the request
> which is exactly what we need to do.
> your patch doesn't fix the race condition.
>
> thanks,
> Yaniv
>
>> I was just asking you if you could give this patch a spin and see if it
>> works. If not (for whatever reason) I'm happy to accept your patch.
>> But first I would like to have an explanation why the above would _not_
>> work.
>>
>> Unfortunately I don't have the hardware otherwise I'd be running the
>> tests myself.
>>
>> Cheers,
>>
>> Hannes
>> --
>> Dr. Hannes Reinecke		      zSeries & Storage
>> hare@suse.de			      +49 911 74053 688
>> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
>> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>


--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: [PATCH v5 03/15] scsi: ufs: implement scsi host timeout handler

[Dolev Raviv]

>> On 03/03/2016 05:10 PM, ygardi@codeaurora.org wrote:
>>>> On 03/01/2016 09:25 PM, ygardi@codeaurora.org wrote:
>>>>>> On 02/28/2016 09:32 PM, Yaniv Gardi wrote:
>>>>>>> A race condition exists between request requeueing and scsi 
>>>>>>> layer error handling:
>>>>>>> When UFS driver queuecommand returns a busy status for a 
>>>>>>> request, it will be requeued and its tag will be freed and set to
-1.
>>>>>>> At the same time it is possible that the request will timeout 
>>>>>>> and scsi layer will start error handling for it. The scsi layer 
>>>>>>> reuses the request and its tag to send error related commands to 
>>>>>>> the device, however its tag is no longer valid.
>>>>>> Hmm. How can the host return a 'busy' status for a request?
>>>>>> From my understanding we have three possibilities:
>>>>>>
>>>>>> 1) queuecommand returns busy; however, that means that the 
>>>>>> command has never been send and this issue shouldn't occur
>>>>>> 2) The command returns with BUSY status. But in this case it has 
>>>>>> already been returned, so there cannot be any timeout coming in.
>>>>>> 3) The host receives a command with a tag which is already in-use.
>>>>>> However, that should have been prevented by the block-layer, 
>>>>>> which really should ensure that this situation never happens.
>>>>>>
>>>>>> So either way I look at it, it really looks like a bug and adding 
>>>>>> a timeout handler will just paper over it.
>>>>>> (Not that a timeout handler is a bad idea, in fact I'm convinced 
>>>>>> that you need one. Just not for this purpose.)
>>>>>>
>>>>>> So can you elaborate how this 'busy' status comes about?
>>>>>> Is the command sent to the device?
>>>>>>
>>>>>> Cheers,
>>>>>>
>>>>>> Hannes
>>>>>
>>>>>
>>>>> Hi Hannes,
>>>>>
>>>>> it's going to be a bit long :)
>>>>> I think you are missing the point.
>>>>> I will describe a race condition happened to us a while ago, that 
>>>>> was quite difficult to understand and fix.
>>>>> So, this patch is not about the "busy" returning to the scsi 
>>>>> dispatch routine. it's about the abort triggered after 30 seconds.
>>>>>
>>>>> imagine a request being queued and sent to the scsi, and then to 
>>>>> the ufs.
>>>>> a timer, initialized to 30 seconds start ticking.
>>>>> but the request is never sent to the ufs device, as queuecommand() 
>>>>> returns with "SCSI_MLQUEUE_HOST_BUSY"
>>>>> by looking at the code, this could happen, for example:
>>>>> 	err = ufshcd_hold(hba, true);
>>>>> 	if (err) {
>>>>> 		err = SCSI_MLQUEUE_HOST_BUSY;
>>>>> 		goto out;
>>>>> 	}
>>>>>
>>>> Uuhhh.
>>>> You probably should not have pointed me to that piece of code ...
>>>> open-coding loops in ufshcd_hold() ... shudder.
>>>> (Did I ever review that one? Must've ...)
>>>> _Anyway_: sleeping in queuecommand is always a bad idea, as then 
>>>> precisely those issues you've just described will happen.
>>>>
>>>> Couldn't you just call
>>>> ufshcd_hold(hba, false)
>>>> instead of
>>>> ufshcd_hold(hba, true)
>>>> ?
>>>> The request will be requeued more-or-less immediately, avoiding the 
>>>> issue with timeout handler kicking in.
>>>> And the queue will remain blocked until the ungate work item 
>>>> returns, at which point I/O submission will continue.
>>>> As the request will be requeued to the head of the queue there 
>>>> won't be other I/O competing with tags, so it shouldn't have any 
>>>> adverse effects.
>>>>
>>>> Wouldn't that work?
>>>>
>>>> Cheers,
>>>>
>>>> Hannes
>>>
>>> Hi Hannes
>>>
>>> This is a bug, and it should be fixed.
>> Oh, definitely agreed. The question is _where_.
>>
>>
>>> if you choose to bypass it, by calling ufshcd_hold(hba, false), not 
>>> only the race condition is still there, and can pop-out at any other 
>>> point in the future, but also, not sure what are the consequences of 
>>> ufshcd_hold(hba, false) unstead of "true".
>> Well ... seeing it's your driver, I would've thought _you_ should 
>> know ...
>>
>>> so, changing the already tested and working code, (not to return 
>>> BUSY from
>>> queuecommand) is not a fix.
>> Hey, I did _not_ suggest not to retury BUSY from queuecommand.
>>
>> I was suggesting this patch:
>>
>> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c 
>> index 9c1b94b..b9295ad 100644
>> --- a/drivers/scsi/ufs/ufshcd.c
>> +++ b/drivers/scsi/ufs/ufshcd.c
>> @@ -1388,7 +1388,7 @@ static int ufshcd_queuecommand(struct Scsi_Host 
>> *host, struct scsi_cmnd *cmd)
>>                 goto out;
>>         }
>>
>> -       err = ufshcd_hold(hba, true);
>> +       err = ufshcd_hold(hba, false);
>>         if (err) {
>>                 err = SCSI_MLQUEUE_HOST_BUSY;
>>                 clear_bit_unlock(tag, &hba->lrb_in_use);
>>
>> which, by reading the code, should be avoiding this issue.
>
>
> Hannes,
> we are not trying to avoid returning BUSY from queuecommand().
> On the contrary. By returning BUSY we actually re-queuing the request 
> which is exactly what we need to do.
> your patch doesn't fix the race condition.
>
> thanks,
> Yaniv
>
>> I was just asking you if you could give this patch a spin and see if 
>> it works. If not (for whatever reason) I'm happy to accept your patch.
>> But first I would like to have an explanation why the above would 
>> _not_ work.
>>
>> Unfortunately I don't have the hardware otherwise I'd be running the 
>> tests myself.
>>
>> Cheers,
>>
>> Hannes
>> --
>> Dr. Hannes Reinecke		      zSeries & Storage
>> hare@suse.de			      +49 911 74053 688
>> SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
>> GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-scsi" 
>> in the body of a message to majordomo@vger.kernel.org More majordomo 
>> info at  http://vger.kernel.org/majordomo-info.html
>>
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" 
> in the body of a message to majordomo@vger.kernel.org More majordomo 
> info at  http://vger.kernel.org/majordomo-info.html
>

I reviewed the patch, you can add 

Reviewed-by: Dolev Raviv <draviv@codeaurora.org>

Thanks,
Dolev
-- 
Qualcomm Israel, on behalf of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux
Foundation Collaborative Project

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html