crypto: qat - fix device reset flow

When the device needs a reset, e.g. when an uncorrectable PCIe AER event
occurs, various services/data structures need to be cleaned up, the
hardware reset and the services/data structures initialized and started.
The code to perform the cleanup and initialization was not performed when
a device reset was done.

This patch moves some of the initialization code out of the .probe entry-
point into a separate function that is now called during probe as well as
after the hardware has been reset.  Similarly, a new function is added for
first cleaning up these services/data structures prior to resetting.  The
new functions are adf_dev_init() and adf_dev_shutdown(), respectively, for
which there are already prototypes but no actual functions just yet and are
now called when the device is reset and during probe/cleanup of the driver.
The down and up flows via ioctl calls has similarly been updated.

In addition, there are two other bugs in the reset flow - one in the logic
for determining whether to schedule a device reset upon receiving an
uncorrectable AER event which prevents the reset flow from being initiated,
and another with clearing the status bit indicating a device is configured
(when resetting the device the configuration remains across the reset so
the bit should not be cleared, otherwise, the necessary services will not
be re-started in adf_dev_start() after the reset - clear the bit only when
actually deleting the configuration).

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

6 files changed, 98 insertions, 16 deletions

diff --git a/drivers/crypto/qat/qat_common/adf_accel_devices.h b/drivers/crypto/qat/qat_common/adf_accel_devices.h
index 2ed425664a16..ed226cee6e29 100644
--- a/drivers/crypto/qat/qat_common/adf_accel_devices.h
+++ b/drivers/crypto/qat/qat_common/adf_accel_devices.h
@@ -148,6 +148,11 @@ struct adf_hw_device_data {
 	int (*alloc_irq)(struct adf_accel_dev *accel_dev);
 	void (*free_irq)(struct adf_accel_dev *accel_dev);
 	void (*enable_error_correction)(struct adf_accel_dev *accel_dev);
+	int (*init_admin_comms)(struct adf_accel_dev *accel_dev);
+	void (*exit_admin_comms)(struct adf_accel_dev *accel_dev);
+	int (*init_arb)(struct adf_accel_dev *accel_dev);
+	void (*exit_arb)(struct adf_accel_dev *accel_dev);
+	void (*enable_ints)(struct adf_accel_dev *accel_dev);
 	const char *fw_name;
 	uint32_t pci_dev_id;
 	uint32_t fuses;
diff --git a/drivers/crypto/qat/qat_common/adf_aer.c b/drivers/crypto/qat/qat_common/adf_aer.c
index 10ce4a2854ab..740dc9e1c70c 100644
--- a/drivers/crypto/qat/qat_common/adf_aer.c
+++ b/drivers/crypto/qat/qat_common/adf_aer.c
@@ -125,8 +125,9 @@ static void adf_device_reset_worker(struct work_struct *work)
 
 	adf_dev_restarting_notify(accel_dev);
 	adf_dev_stop(accel_dev);
+	adf_dev_shutdown(accel_dev);
 	adf_dev_restore(accel_dev);
-	if (adf_dev_start(accel_dev)) {
+	if (adf_dev_init(accel_dev) || adf_dev_start(accel_dev)) {
 		/* The device hanged and we can't restart it so stop here */
 		dev_err(&GET_DEV(accel_dev), "Restart device failed\n");
 		kfree(reset_data);
@@ -148,8 +149,8 @@ static int adf_dev_aer_schedule_reset(struct adf_accel_dev *accel_dev,
 {
 	struct adf_reset_dev_data *reset_data;
 
-	if (adf_dev_started(accel_dev) &&
-	    !test_bit(ADF_STATUS_RESTARTING, &accel_dev->status))
+	if (!adf_dev_started(accel_dev) ||
+	    test_bit(ADF_STATUS_RESTARTING, &accel_dev->status))
 		return 0;
 
 	set_bit(ADF_STATUS_RESTARTING, &accel_dev->status);
diff --git a/drivers/crypto/qat/qat_common/adf_cfg.c b/drivers/crypto/qat/qat_common/adf_cfg.c
index aba7f1d043fb..de16da9070a5 100644
--- a/drivers/crypto/qat/qat_common/adf_cfg.c
+++ b/drivers/crypto/qat/qat_common/adf_cfg.c
@@ -50,6 +50,7 @@
 #include <linux/seq_file.h>
 #include "adf_accel_devices.h"
 #include "adf_cfg.h"
+#include "adf_common_drv.h"
 
 static DEFINE_MUTEX(qat_cfg_read_lock);
 
@@ -159,6 +160,7 @@ void adf_cfg_del_all(struct adf_accel_dev *accel_dev)
 	down_write(&dev_cfg_data->lock);
 	adf_cfg_section_del_all(&dev_cfg_data->sec_list);
 	up_write(&dev_cfg_data->lock);
+	clear_bit(ADF_STATUS_CONFIGURED, &accel_dev->status);
 }
 
 /**
diff --git a/drivers/crypto/qat/qat_common/adf_common_drv.h b/drivers/crypto/qat/qat_common/adf_common_drv.h
index 5e8f9d431e5d..a62e485c8786 100644
--- a/drivers/crypto/qat/qat_common/adf_common_drv.h
+++ b/drivers/crypto/qat/qat_common/adf_common_drv.h
@@ -93,7 +93,7 @@ int adf_service_unregister(struct service_hndl *service);
 int adf_dev_init(struct adf_accel_dev *accel_dev);
 int adf_dev_start(struct adf_accel_dev *accel_dev);
 int adf_dev_stop(struct adf_accel_dev *accel_dev);
-int adf_dev_shutdown(struct adf_accel_dev *accel_dev);
+void adf_dev_shutdown(struct adf_accel_dev *accel_dev);
 
 int adf_ctl_dev_register(void);
 void adf_ctl_dev_unregister(void);
diff --git a/drivers/crypto/qat/qat_common/adf_ctl_drv.c b/drivers/crypto/qat/qat_common/adf_ctl_drv.c
index 7ee93f881db6..74207a6f0516 100644
--- a/drivers/crypto/qat/qat_common/adf_ctl_drv.c
+++ b/drivers/crypto/qat/qat_common/adf_ctl_drv.c
@@ -282,6 +282,8 @@ static int adf_ctl_stop_devices(uint32_t id)
 			if (adf_dev_stop(accel_dev)) {
 				pr_err("QAT: Failed to stop qat_dev%d\n", id);
 				ret = -EFAULT;
+			} else {
+				adf_dev_shutdown(accel_dev);
 			}
 		}
 	}
@@ -343,7 +345,9 @@ static int adf_ctl_ioctl_dev_start(struct file *fp, unsigned int cmd,
 	if (!adf_dev_started(accel_dev)) {
 		pr_info("QAT: Starting acceleration device qat_dev%d.\n",
 			ctl_data->device_id);
-		ret = adf_dev_start(accel_dev);
+		ret = adf_dev_init(accel_dev);
+		if (!ret)
+			ret = adf_dev_start(accel_dev);
 	} else {
 		pr_info("QAT: Acceleration device qat_dev%d already started.\n",
 			ctl_data->device_id);
@@ -351,6 +355,7 @@ static int adf_ctl_ioctl_dev_start(struct file *fp, unsigned int cmd,
 	if (ret) {
 		pr_err("QAT: Failed to start qat_dev%d\n", ctl_data->device_id);
 		adf_dev_stop(accel_dev);
+		adf_dev_shutdown(accel_dev);
 	}
 out:
 	kfree(ctl_data);
diff --git a/drivers/crypto/qat/qat_common/adf_init.c b/drivers/crypto/qat/qat_common/adf_init.c
index 5c0e47a00a87..a3afa0ff9ec6 100644
--- a/drivers/crypto/qat/qat_common/adf_init.c
+++ b/drivers/crypto/qat/qat_common/adf_init.c
@@ -108,26 +108,47 @@ int adf_service_unregister(struct service_hndl *service)
 EXPORT_SYMBOL_GPL(adf_service_unregister);
 
 /**
- * adf_dev_start() - Start acceleration service for the given accel device
- * @accel_dev:    Pointer to acceleration device.
+ * adf_dev_init() - Init data structures and services for the given accel device
+ * @accel_dev: Pointer to acceleration device.
  *
- * Function notifies all the registered services that the acceleration device
- * is ready to be used.
- * To be used by QAT device specific drivers.
+ * Initialize the ring data structures and the admin comms and arbitration
+ * services.
  *
  * Return: 0 on success, error code othewise.
  */
-int adf_dev_start(struct adf_accel_dev *accel_dev)
+int adf_dev_init(struct adf_accel_dev *accel_dev)
 {
 	struct service_hndl *service;
 	struct list_head *list_itr;
 	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
 
+	if (!hw_data) {
+		dev_err(&GET_DEV(accel_dev),
+			"QAT: Failed to init device - hw_data not set\n");
+		return -EFAULT;
+	}
+
 	if (!test_bit(ADF_STATUS_CONFIGURED, &accel_dev->status)) {
 		pr_info("QAT: Device not configured\n");
 		return -EFAULT;
 	}
-	set_bit(ADF_STATUS_STARTING, &accel_dev->status);
+
+	if (adf_init_etr_data(accel_dev)) {
+		dev_err(&GET_DEV(accel_dev), "Failed initialize etr\n");
+		return -EFAULT;
+	}
+
+	if (hw_data->init_admin_comms && hw_data->init_admin_comms(accel_dev)) {
+		dev_err(&GET_DEV(accel_dev), "Failed initialize admin comms\n");
+		return -EFAULT;
+	}
+
+	if (hw_data->init_arb && hw_data->init_arb(accel_dev)) {
+		dev_err(&GET_DEV(accel_dev), "Failed initialize hw arbiter\n");
+		return -EFAULT;
+	}
+
+	hw_data->enable_ints(accel_dev);
 
 	if (adf_ae_init(accel_dev)) {
 		pr_err("QAT: Failed to initialise Acceleration Engine\n");
@@ -178,6 +199,27 @@ int adf_dev_start(struct adf_accel_dev *accel_dev)
 
 	hw_data->enable_error_correction(accel_dev);
 
+	return 0;
+}
+EXPORT_SYMBOL_GPL(adf_dev_init);
+
+/**
+ * adf_dev_start() - Start acceleration service for the given accel device
+ * @accel_dev:    Pointer to acceleration device.
+ *
+ * Function notifies all the registered services that the acceleration device
+ * is ready to be used.
+ * To be used by QAT device specific drivers.
+ *
+ * Return: 0 on success, error code othewise.
+ */
+int adf_dev_start(struct adf_accel_dev *accel_dev)
+{
+	struct service_hndl *service;
+	struct list_head *list_itr;
+
+	set_bit(ADF_STATUS_STARTING, &accel_dev->status);
+
 	if (adf_ae_start(accel_dev)) {
 		pr_err("QAT: AE Start Failed\n");
 		return -EFAULT;
@@ -232,7 +274,6 @@ EXPORT_SYMBOL_GPL(adf_dev_start);
  */
 int adf_dev_stop(struct adf_accel_dev *accel_dev)
 {
-	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
 	struct service_hndl *service;
 	struct list_head *list_itr;
 	int ret, wait = 0;
@@ -241,7 +282,6 @@ int adf_dev_stop(struct adf_accel_dev *accel_dev)
 	    !test_bit(ADF_STATUS_STARTING, &accel_dev->status)) {
 		return 0;
 	}
-	clear_bit(ADF_STATUS_CONFIGURED, &accel_dev->status);
 	clear_bit(ADF_STATUS_STARTING, &accel_dev->status);
 	clear_bit(ADF_STATUS_STARTED, &accel_dev->status);
 
@@ -285,6 +325,29 @@ int adf_dev_stop(struct adf_accel_dev *accel_dev)
 			clear_bit(ADF_STATUS_AE_STARTED, &accel_dev->status);
 	}
 
+	return 0;
+}
+EXPORT_SYMBOL_GPL(adf_dev_stop);
+
+/**
+ * adf_dev_shutdown() - shutdown acceleration services and data strucutures
+ * @accel_dev: Pointer to acceleration device
+ *
+ * Cleanup the ring data structures and the admin comms and arbitration
+ * services.
+ */
+void adf_dev_shutdown(struct adf_accel_dev *accel_dev)
+{
+	struct adf_hw_device_data *hw_data = accel_dev->hw_device;
+	struct service_hndl *service;
+	struct list_head *list_itr;
+
+	if (!hw_data) {
+		dev_err(&GET_DEV(accel_dev),
+			"QAT: Failed to shutdown device - hw_data not set\n");
+		return;
+	}
+
 	if (test_bit(ADF_STATUS_AE_UCODE_LOADED, &accel_dev->status)) {
 		if (adf_ae_fw_release(accel_dev))
 			pr_err("QAT: Failed to release the ucode\n");
@@ -335,9 +398,15 @@ int adf_dev_stop(struct adf_accel_dev *accel_dev)
 	if (!test_bit(ADF_STATUS_RESTARTING, &accel_dev->status))
 		adf_cfg_del_all(accel_dev);
 
-	return 0;
+	if (hw_data->exit_arb)
+		hw_data->exit_arb(accel_dev);
+
+	if (hw_data->exit_admin_comms)
+		hw_data->exit_admin_comms(accel_dev);
+
+	adf_cleanup_etr_data(accel_dev);
 }
-EXPORT_SYMBOL_GPL(adf_dev_stop);
+EXPORT_SYMBOL_GPL(adf_dev_shutdown);
 
 int adf_dev_restarting_notify(struct adf_accel_dev *accel_dev)
 {