Merge patch series "mei: connect to card in D3cold"

Alexander Usyskin <alexander.usyskin@intel.com> says:

When discrete graphic card enters D3cold th CSC engine is powered down.

On wakeup from the D3cold full HECI link reset is required.  The driver
should detect that firmware requests link reset and initiate the link
reset flow.

In the usual flow the connect IOCTL will trigger the wake from D3cold
and corresponding link reset.  The MEI driver invalidates all open
handles on link reset including the one that triggered the wake
rendering this connection unusable.  To break this loop make connect
detect that it is interrupted by link reset and retry connect attempt
after reset was completed.

Link: https://lore.kernel.org/r/20250918130435.3327400-1-alexander.usyskin@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

6 files changed, 62 insertions, 77 deletions

diff --git a/drivers/misc/mei/bus.c b/drivers/misc/mei/bus.c
index 6c8ec5c0d359..38735a41f750 100644
--- a/drivers/misc/mei/bus.c
+++ b/drivers/misc/mei/bus.c
@@ -814,7 +814,7 @@ int mei_cldev_enable(struct mei_cl_device *cldev)
 
 	ret = mei_cl_connect(cl, cldev->me_cl, NULL);
 	if (ret < 0) {
-		dev_err(&cldev->dev, "cannot connect\n");
+		dev_dbg(&cldev->dev, "cannot connect\n");
 		mei_cl_bus_vtag_free(cldev);
 	}
 
diff --git a/drivers/misc/mei/hw-me.c b/drivers/misc/mei/hw-me.c
index 346633ebc6e6..d4612c659784 100644
--- a/drivers/misc/mei/hw-me.c
+++ b/drivers/misc/mei/hw-me.c
@@ -494,43 +494,6 @@ static int mei_me_hw_ready_wait(struct mei_device *dev)
 }
 
 /**
- * mei_me_check_fw_reset - check for the firmware reset error and exception conditions
- *
- * @dev: mei device
- */
-static void mei_me_check_fw_reset(struct mei_device *dev)
-{
-	struct mei_fw_status fw_status;
-	char fw_sts_str[MEI_FW_STATUS_STR_SZ] = {0};
-	int ret;
-	u32 fw_pm_event = 0;
-
-	if (!dev->saved_fw_status_flag)
-		goto end;
-
-	if (dev->gsc_reset_to_pxp == MEI_DEV_RESET_TO_PXP_PERFORMED) {
-		ret = mei_fw_status(dev, &fw_status);
-		if (!ret) {
-			fw_pm_event = fw_status.status[1] & PCI_CFG_HFS_2_PM_EVENT_MASK;
-			if (fw_pm_event != PCI_CFG_HFS_2_PM_CMOFF_TO_CMX_ERROR &&
-			    fw_pm_event != PCI_CFG_HFS_2_PM_CM_RESET_ERROR)
-				goto end;
-		} else {
-			dev_err(&dev->dev, "failed to read firmware status: %d\n", ret);
-		}
-	}
-
-	mei_fw_status2str(&dev->saved_fw_status, fw_sts_str, sizeof(fw_sts_str));
-	dev_warn(&dev->dev, "unexpected reset: fw_pm_event = 0x%x, dev_state = %u fw status = %s\n",
-		 fw_pm_event, dev->saved_dev_state, fw_sts_str);
-
-end:
-	if (dev->gsc_reset_to_pxp == MEI_DEV_RESET_TO_PXP_PERFORMED)
-		dev->gsc_reset_to_pxp = MEI_DEV_RESET_TO_PXP_DONE;
-	dev->saved_fw_status_flag = false;
-}
-
-/**
  * mei_me_hw_start - hw start routine
  *
  * @dev: mei device
@@ -540,8 +503,9 @@ static int mei_me_hw_start(struct mei_device *dev)
 {
 	int ret = mei_me_hw_ready_wait(dev);
 
-	if (kind_is_gsc(dev) || kind_is_gscfi(dev))
-		mei_me_check_fw_reset(dev);
+	if ((kind_is_gsc(dev) || kind_is_gscfi(dev)) &&
+	    dev->gsc_reset_to_pxp == MEI_DEV_RESET_TO_PXP_PERFORMED)
+		dev->gsc_reset_to_pxp = MEI_DEV_RESET_TO_PXP_DONE;
 	if (ret)
 		return ret;
 	dev_dbg(&dev->dev, "hw is ready\n");
@@ -1373,9 +1337,20 @@ irqreturn_t mei_me_irq_thread_handler(int irq, void *dev_id)
 	/*  check if we need to start the dev */
 	if (!mei_host_is_ready(dev)) {
 		if (mei_hw_is_ready(dev)) {
-			dev_dbg(&dev->dev, "we need to start the dev.\n");
-			dev->recvd_hw_ready = true;
-			wake_up(&dev->wait_hw_ready);
+			/* synchronized by dev mutex */
+			if (waitqueue_active(&dev->wait_hw_ready)) {
+				dev_dbg(&dev->dev, "we need to start the dev.\n");
+				dev->recvd_hw_ready = true;
+				wake_up(&dev->wait_hw_ready);
+			} else if (dev->dev_state != MEI_DEV_UNINITIALIZED &&
+				   dev->dev_state != MEI_DEV_POWERING_DOWN &&
+				   dev->dev_state != MEI_DEV_POWER_DOWN) {
+				dev_dbg(&dev->dev, "Force link reset.\n");
+				schedule_work(&dev->reset_work);
+			} else {
+				dev_dbg(&dev->dev, "Ignore this interrupt in state = %d\n",
+					dev->dev_state);
+			}
 		} else {
 			dev_dbg(&dev->dev, "Spurious Interrupt\n");
 		}
diff --git a/drivers/misc/mei/hw.h b/drivers/misc/mei/hw.h
index 2e9cf6f4efb6..3771aa09c592 100644
--- a/drivers/misc/mei/hw.h
+++ b/drivers/misc/mei/hw.h
@@ -27,6 +27,8 @@
 #define MKHI_RCV_TIMEOUT 500 /* receive timeout in msec */
 #define MKHI_RCV_TIMEOUT_SLOW 10000 /* receive timeout in msec, slow FW */
 
+#define MEI_LINK_RESET_WAIT_TIMEOUT_MSEC 500  /* Max wait timeout for link reset, in msec */
+
 /*
  * FW page size for DMA allocations
  */
diff --git a/drivers/misc/mei/init.c b/drivers/misc/mei/init.c
index 4d1d5423b612..b789c4d9c709 100644
--- a/drivers/misc/mei/init.c
+++ b/drivers/misc/mei/init.c
@@ -89,22 +89,6 @@ void mei_cancel_work(struct mei_device *dev)
 }
 EXPORT_SYMBOL_GPL(mei_cancel_work);
 
-static void mei_save_fw_status(struct mei_device *dev)
-{
-	struct mei_fw_status fw_status;
-	int ret;
-
-	ret = mei_fw_status(dev, &fw_status);
-	if (ret) {
-		dev_err(&dev->dev, "failed to read firmware status: %d\n", ret);
-		return;
-	}
-
-	dev->saved_dev_state = dev->dev_state;
-	dev->saved_fw_status_flag = true;
-	memcpy(&dev->saved_fw_status, &fw_status, sizeof(fw_status));
-}
-
 /**
  * mei_reset - resets host and fw.
  *
@@ -128,7 +112,6 @@ int mei_reset(struct mei_device *dev)
 		if (kind_is_gsc(dev) || kind_is_gscfi(dev)) {
 			dev_dbg(&dev->dev, "unexpected reset: dev_state = %s fw status = %s\n",
 				mei_dev_state_str(state), fw_sts_str);
-			mei_save_fw_status(dev);
 		} else {
 			dev_warn(&dev->dev, "unexpected reset: dev_state = %s fw status = %s\n",
 				 mei_dev_state_str(state), fw_sts_str);
@@ -399,7 +382,8 @@ void mei_device_init(struct mei_device *dev,
 	init_waitqueue_head(&dev->wait_hw_ready);
 	init_waitqueue_head(&dev->wait_pg);
 	init_waitqueue_head(&dev->wait_hbm_start);
-	dev->dev_state = MEI_DEV_INITIALIZING;
+	dev->dev_state = MEI_DEV_UNINITIALIZED;
+	init_waitqueue_head(&dev->wait_dev_state);
 	dev->reset_count = 0;
 
 	INIT_LIST_HEAD(&dev->write_list);
@@ -442,5 +426,6 @@ void mei_device_init(struct mei_device *dev,
 		dev->timeouts.hbm = mei_secs_to_jiffies(MEI_HBM_TIMEOUT);
 		dev->timeouts.mkhi_recv = msecs_to_jiffies(MKHI_RCV_TIMEOUT);
 	}
+	dev->timeouts.link_reset_wait = msecs_to_jiffies(MEI_LINK_RESET_WAIT_TIMEOUT_MSEC);
 }
 EXPORT_SYMBOL_GPL(mei_device_init);
diff --git a/drivers/misc/mei/main.c b/drivers/misc/mei/main.c
index 93c518bb5e14..86a73684a373 100644
--- a/drivers/misc/mei/main.c
+++ b/drivers/misc/mei/main.c
@@ -423,6 +423,7 @@ static int mei_ioctl_connect_client(struct file *file,
 	    cl->state != MEI_FILE_DISCONNECTED)
 		return  -EBUSY;
 
+retry:
 	/* find ME client we're trying to connect to */
 	me_cl = mei_me_cl_by_uuid(dev, in_client_uuid);
 	if (!me_cl) {
@@ -454,6 +455,28 @@ static int mei_ioctl_connect_client(struct file *file,
 
 	rets = mei_cl_connect(cl, me_cl, file);
 
+	if (rets && cl->status == -EFAULT &&
+	    (dev->dev_state == MEI_DEV_RESETTING ||
+	     dev->dev_state == MEI_DEV_INIT_CLIENTS)) {
+		/* in link reset, wait for it completion */
+		mutex_unlock(&dev->device_lock);
+		rets = wait_event_interruptible_timeout(dev->wait_dev_state,
+							dev->dev_state == MEI_DEV_ENABLED,
+							dev->timeouts.link_reset_wait);
+		mutex_lock(&dev->device_lock);
+		if (rets < 0) {
+			if (signal_pending(current))
+				rets = -EINTR;
+			goto end;
+		}
+		if (dev->dev_state != MEI_DEV_ENABLED) {
+			rets = -ETIME;
+			goto end;
+		}
+		mei_me_cl_put(me_cl);
+		goto retry;
+	}
+
 end:
 	mei_me_cl_put(me_cl);
 	return rets;
@@ -646,7 +669,7 @@ static long mei_ioctl(struct file *file, unsigned int cmd, unsigned long data)
 	struct mei_cl *cl = file->private_data;
 	struct mei_connect_client_data conn;
 	struct mei_connect_client_data_vtag conn_vtag;
-	const uuid_le *cl_uuid;
+	uuid_le cl_uuid;
 	struct mei_client *props;
 	u8 vtag;
 	u32 notify_get, notify_req;
@@ -674,18 +697,18 @@ static long mei_ioctl(struct file *file, unsigned int cmd, unsigned long data)
 			rets = -EFAULT;
 			goto out;
 		}
-		cl_uuid = &conn.in_client_uuid;
+		cl_uuid = conn.in_client_uuid;
 		props = &conn.out_client_properties;
 		vtag = 0;
 
-		rets = mei_vt_support_check(dev, cl_uuid);
+		rets = mei_vt_support_check(dev, &cl_uuid);
 		if (rets == -ENOTTY)
 			goto out;
 		if (!rets)
-			rets = mei_ioctl_connect_vtag(file, cl_uuid, props,
+			rets = mei_ioctl_connect_vtag(file, &cl_uuid, props,
 						      vtag);
 		else
-			rets = mei_ioctl_connect_client(file, cl_uuid, props);
+			rets = mei_ioctl_connect_client(file, &cl_uuid, props);
 		if (rets)
 			goto out;
 
@@ -707,14 +730,14 @@ static long mei_ioctl(struct file *file, unsigned int cmd, unsigned long data)
 			goto out;
 		}
 
-		cl_uuid = &conn_vtag.connect.in_client_uuid;
+		cl_uuid = conn_vtag.connect.in_client_uuid;
 		props = &conn_vtag.out_client_properties;
 		vtag = conn_vtag.connect.vtag;
 
-		rets = mei_vt_support_check(dev, cl_uuid);
+		rets = mei_vt_support_check(dev, &cl_uuid);
 		if (rets == -EOPNOTSUPP)
 			cl_dbg(dev, cl, "FW Client %pUl does not support vtags\n",
-				cl_uuid);
+				&cl_uuid);
 		if (rets)
 			goto out;
 
@@ -724,7 +747,7 @@ static long mei_ioctl(struct file *file, unsigned int cmd, unsigned long data)
 			goto out;
 		}
 
-		rets = mei_ioctl_connect_vtag(file, cl_uuid, props, vtag);
+		rets = mei_ioctl_connect_vtag(file, &cl_uuid, props, vtag);
 		if (rets)
 			goto out;
 
@@ -1120,6 +1143,8 @@ void mei_set_devstate(struct mei_device *dev, enum mei_dev_state state)
 
 	dev->dev_state = state;
 
+	wake_up_interruptible_all(&dev->wait_dev_state);
+
 	if (!dev->cdev)
 		return;
 
diff --git a/drivers/misc/mei/mei_dev.h b/drivers/misc/mei/mei_dev.h
index 4ab2b9100fd4..0bf8d552c3ea 100644
--- a/drivers/misc/mei/mei_dev.h
+++ b/drivers/misc/mei/mei_dev.h
@@ -57,7 +57,8 @@ enum file_state {
 
 /* MEI device states */
 enum mei_dev_state {
-	MEI_DEV_INITIALIZING = 0,
+	MEI_DEV_UNINITIALIZED = 0,
+	MEI_DEV_INITIALIZING,
 	MEI_DEV_INIT_CLIENTS,
 	MEI_DEV_ENABLED,
 	MEI_DEV_RESETTING,
@@ -465,6 +466,7 @@ struct mei_dev_timeouts {
 	unsigned int d0i3; /* D0i3 set/unset max response time, in jiffies */
 	unsigned long hbm; /* HBM operation timeout, in jiffies */
 	unsigned long mkhi_recv; /* receive timeout, in jiffies */
+	unsigned long link_reset_wait; /* link reset wait timeout, in jiffies */
 };
 
 /**
@@ -495,6 +497,7 @@ struct mei_dev_timeouts {
  *
  * @reset_count : number of consecutive resets
  * @dev_state   : device state
+ * @wait_dev_state: wait queue for device state change
  * @hbm_state   : state of host bus message protocol
  * @pxp_mode    : PXP device mode
  * @init_clients_timer : HBM init handshake timeout
@@ -548,9 +551,6 @@ struct mei_dev_timeouts {
  *
  * @dbgfs_dir   : debugfs mei root directory
  *
- * @saved_fw_status      : saved firmware status
- * @saved_dev_state      : saved device state
- * @saved_fw_status_flag : flag indicating that firmware status was saved
  * @gsc_reset_to_pxp     : state of reset to the PXP mode
  *
  * @ops:        : hw specific operations
@@ -587,6 +587,7 @@ struct mei_device {
 	 */
 	unsigned long reset_count;
 	enum mei_dev_state dev_state;
+	wait_queue_head_t wait_dev_state;
 	enum mei_hbm_state hbm_state;
 	enum mei_dev_pxp_mode pxp_mode;
 	u16 init_clients_timer;
@@ -650,9 +651,6 @@ struct mei_device {
 	struct dentry *dbgfs_dir;
 #endif /* CONFIG_DEBUG_FS */
 
-	struct mei_fw_status saved_fw_status;
-	enum mei_dev_state saved_dev_state;
-	bool saved_fw_status_flag;
 	enum mei_dev_reset_to_pxp gsc_reset_to_pxp;
 
 	const struct mei_hw_ops *ops;