[PATCH v3 0/5] add debugfs to migration driver

[PATCH v3 0/5] add debugfs to migration driver

[Longfang Liu]

Add a debugfs function to the migration driver in VFIO to provide
a step-by-step test function for the migration driver.

When the execution of live migration fails, the user can view the
status and data during the migration process separately from the
source and the destination, which is convenient for users to analyze
and locate problems.

Changes v2 -> v3
	Extend the debugfs function from hisilicon device to vfio.

Changes v1 -> v2
	Change the registration method of root_debugfs to register
	with module initialization.

Longfang Liu (5):
  vfio/migration: Add debugfs to live migration driver
  hisi_acc_vfio_pci: extract public functions for container_of
  hisi_acc_vfio_pci: register debugfs for hisilicon migration driver
  Documentation: add debugfs description for hisi_acc_vfio_pci
  vfio: update live migration device status

 .../ABI/testing/debugfs-hisi-migration        |  32 +++
 drivers/vfio/pci/Makefile                     |   4 +
 .../vfio/pci/hisilicon/hisi_acc_vfio_pci.c    | 192 ++++++++++++++-
 .../vfio/pci/hisilicon/hisi_acc_vfio_pci.h    |   6 +
 drivers/vfio/pci/mlx5/main.c                  |   3 +
 drivers/vfio/pci/vfio_pci_core.c              |   3 +
 drivers/vfio/pci/vfio_pci_debugfs.c           | 230 ++++++++++++++++++
 drivers/vfio/pci/vfio_pci_debugfs.h           |  30 +++
 include/linux/vfio.h                          |  16 ++
 9 files changed, 506 insertions(+), 10 deletions(-)
 create mode 100644 Documentation/ABI/testing/debugfs-hisi-migration
 create mode 100644 drivers/vfio/pci/vfio_pci_debugfs.c
 create mode 100644 drivers/vfio/pci/vfio_pci_debugfs.h

-- 
2.24.0

[PATCH v3 1/5] vfio/migration: Add debugfs to live migration driver

[Longfang Liu]

There are multiple devices, software and operational steps involved
in the process of live migration. An error occurred on any node may
cause the live migration operation to fail.
This complex process makes it very difficult to locate and analyze
the cause when the function fails.

In order to quickly locate the cause of the problem when the
live migration fails, I added a set of debugfs to the vfio
live migration driver.

 +------------------------------------------+
 |                                          |
 |                                          |
 |                 QEMU                     |
 |                                          |
 |                                          |
 +-----+-------------------------+----------+
       |   ^                     |   ^
       v   |                     v   |
     +-----+--+                +-----+--+
     | src VF |                | dst VF |
     +-+------+                +-+------+
       |   ^                     |   ^
       v   |                     v   |
 +-------+-+------+        +-------+-+------+
 |       |state   |        |       |state   |
 |       +--------+        |       +--------+
 |       |mig_data|        |       |mig_data|
 |Debugfs+--------+        |Debugfs+--------+
 |       |mig_attr|        |       |mig_attr|
 |       +--------+        |       +--------+
 |       |debug   |        |       |debug   |
 +-------+--------+        +-------+--------+

debugfs will create a public root directory vfio_mig.
Then create four debug files for each live migration driven VF.
state file: used to get the status of the live migration device.
mig_data file: used to get the migration data of the live migration device
mig_attr file: used to get device attributes of the live migration device
debug file: used to test migration-related functions, specific test
commands and test functions are added by device drivers.

The live migration function of the current device can be tested by
operating the debug file, and the functional status of the equipment
and software at each stage can be tested step by step without
performing the complete live migration function. And after the live
migration is performed, the migration device data of the live migration
can be obtained through the debug file.

Signed-off-by: Longfang Liu <liulongfang@huawei.com>
---
 drivers/vfio/pci/Makefile           |   4 +
 drivers/vfio/pci/vfio_pci_core.c    |   3 +
 drivers/vfio/pci/vfio_pci_debugfs.c | 230 ++++++++++++++++++++++++++++
 drivers/vfio/pci/vfio_pci_debugfs.h |  30 ++++
 include/linux/vfio.h                |  16 ++
 5 files changed, 283 insertions(+)
 create mode 100644 drivers/vfio/pci/vfio_pci_debugfs.c
 create mode 100644 drivers/vfio/pci/vfio_pci_debugfs.h

diff --git a/drivers/vfio/pci/Makefile b/drivers/vfio/pci/Makefile
index 24c524224da5..efe032623664 100644
--- a/drivers/vfio/pci/Makefile
+++ b/drivers/vfio/pci/Makefile
@@ -4,6 +4,10 @@ vfio-pci-core-y := vfio_pci_core.o vfio_pci_intrs.o vfio_pci_rdwr.o vfio_pci_con
 vfio-pci-core-$(CONFIG_VFIO_PCI_ZDEV_KVM) += vfio_pci_zdev.o
 obj-$(CONFIG_VFIO_PCI_CORE) += vfio-pci-core.o
 
+ifneq ($(CONFIG_DEBUG_FS),)
+	vfio-pci-core-y += vfio_pci_debugfs.o
+endif
+
 vfio-pci-y := vfio_pci.o
 vfio-pci-$(CONFIG_VFIO_PCI_IGD) += vfio_pci_igd.o
 obj-$(CONFIG_VFIO_PCI) += vfio-pci.o
diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c
index 189d4930c276..1d6dd61aef7d 100644
--- a/drivers/vfio/pci/vfio_pci_core.c
+++ b/drivers/vfio/pci/vfio_pci_core.c
@@ -28,6 +28,7 @@
 #include <linux/nospec.h>
 #include <linux/sched/mm.h>
 
+#include "vfio_pci_debugfs.h"
 #include "vfio_pci_priv.h"
 
 #define DRIVER_AUTHOR   "Alex Williamson <alex.williamson@redhat.com>"
@@ -2548,11 +2549,13 @@ EXPORT_SYMBOL_GPL(vfio_pci_core_set_params);
 
 static void vfio_pci_core_cleanup(void)
 {
+	vfio_pci_debugfs_remove_root();
 	vfio_pci_uninit_perm_bits();
 }
 
 static int __init vfio_pci_core_init(void)
 {
+	vfio_pci_debugfs_create_root();
 	/* Allocate shared config space permission data used by all devices */
 	return vfio_pci_init_perm_bits();
 }
diff --git a/drivers/vfio/pci/vfio_pci_debugfs.c b/drivers/vfio/pci/vfio_pci_debugfs.c
new file mode 100644
index 000000000000..8faf609906c7
--- /dev/null
+++ b/drivers/vfio/pci/vfio_pci_debugfs.c
@@ -0,0 +1,230 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2022, HiSilicon Ltd.
+ */
+
+#include <linux/pci.h>
+#include <linux/vfio_pci_core.h>
+
+#include "vfio_pci_debugfs.h"
+
+#define VFIO_DEV_DBG_LEN		256
+static struct dentry *vfio_pci_debugfs_root;
+
+static int vfio_pci_vf_debug_create(struct vfio_device *core_vdev, size_t migf_size)
+{
+	void *migf;
+
+	migf = kzalloc(migf_size, GFP_KERNEL);
+	if (!migf)
+		return -ENOMEM;
+
+	core_vdev->vf_migration_file = migf;
+
+	return 0;
+}
+
+static void vfio_pci_vf_debug_release(struct vfio_device *core_vdev)
+{
+	if (core_vdev)
+		kfree(core_vdev->vf_migration_file);
+}
+
+static int vfio_pci_vf_debug_check(struct vfio_device *vdev)
+{
+	enum vfio_device_mig_state state;
+
+	if (!vdev->mig_ops || !vdev->mig_ops->migration_debug_operate) {
+		dev_err(vdev->dev, "device not support debugfs!\n");
+		return -EINVAL;
+	}
+
+	/* If device not opened, the debugfs operation will trigger calltrace */
+	(void)vdev->mig_ops->migration_get_state(vdev, &state);
+	if (state == VFIO_DEVICE_STATE_ERROR ||
+	     state == VFIO_DEVICE_STATE_STOP) {
+		dev_err(vdev->dev, "device not opened!\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static ssize_t vfio_pci_vf_debug_write(struct file *filp, const char __user *buffer,
+			    size_t count, loff_t *pos)
+{
+	struct vfio_device	*vdev = filp->private_data;
+	char tbuf[VFIO_DEV_DBG_LEN];
+	unsigned long cmd;
+	int len, ret;
+
+	if (*pos)
+		return 0;
+
+	if (count >= VFIO_DEV_DBG_LEN)
+		return -ENOSPC;
+
+	ret = vfio_pci_vf_debug_check(vdev);
+	if (ret)
+		return -EINVAL;
+
+	len = simple_write_to_buffer(tbuf, VFIO_DEV_DBG_LEN - 1,
+					pos, buffer, count);
+	if (len < 0 || len > VFIO_DEV_DBG_LEN - 1)
+		return -EINVAL;
+	tbuf[len] = '\0';
+	if (kstrtoul(tbuf, 0, &cmd))
+		return -EFAULT;
+
+	ret = vdev->mig_ops->migration_debug_operate(vdev, cmd);
+	if (ret) {
+		dev_err(vdev->dev, "device debug test failed!\n");
+		return -EINVAL;
+	}
+
+	return count;
+}
+
+static const struct file_operations vfio_vf_debug_fops = {
+	.owner = THIS_MODULE,
+	.open = simple_open,
+	.write = vfio_pci_vf_debug_write,
+};
+
+static ssize_t vfio_pci_vf_data_dump(struct file *filp, char __user *buffer,
+			   size_t count, loff_t *pos)
+{
+	struct vfio_device	*vdev = filp->private_data;
+	int  len;
+
+	if (!vdev->mig_ops || !vdev->mig_ops->migration_get_data) {
+		dev_err(vdev->dev, "device not support get data!\n");
+		return -EINVAL;
+	}
+
+	len = vdev->mig_ops->migration_get_data(vdev, buffer);
+
+	return len;
+}
+
+static const struct file_operations vfio_vf_data_fops = {
+	.owner = THIS_MODULE,
+	.open = simple_open,
+	.read = vfio_pci_vf_data_dump,
+};
+
+static ssize_t vfio_pci_vf_attr_show(struct file *filp, char __user *buffer,
+			   size_t count, loff_t *pos)
+{
+	struct vfio_device	*vdev = filp->private_data;
+	int  len;
+
+	if (!vdev->mig_ops || !vdev->mig_ops->migration_get_attr) {
+		dev_err(vdev->dev, "device not support get attribute!\n");
+		return -EINVAL;
+	}
+
+	len = vdev->mig_ops->migration_get_attr(vdev, buffer);
+
+	return len;
+}
+
+static const struct file_operations vfio_vf_attr_fops = {
+	.owner = THIS_MODULE,
+	.open = simple_open,
+	.read = vfio_pci_vf_attr_show,
+};
+
+static ssize_t vfio_pci_vf_state_read(struct file *filp, char __user *buffer,
+			   size_t count, loff_t *pos)
+{
+	struct vfio_device	*vdev = filp->private_data;
+	char buf[VFIO_DEV_DBG_LEN];
+	enum vfio_device_mig_state state;
+	int  len;
+
+	(void)vdev->mig_ops->migration_get_state(vdev, &state);
+	switch (state) {
+	case VFIO_DEVICE_STATE_RUNNING:
+		len = scnprintf(buf, VFIO_DEV_DBG_LEN, "%s\n",
+			"RUNNING\n");
+		break;
+	case VFIO_DEVICE_STATE_STOP_COPY:
+		len = scnprintf(buf, VFIO_DEV_DBG_LEN, "%s\n",
+			"STOP_COPY\n");
+		break;
+	case VFIO_DEVICE_STATE_STOP:
+		len = scnprintf(buf, VFIO_DEV_DBG_LEN, "%s\n",
+			"STOP\n");
+		break;
+	case VFIO_DEVICE_STATE_RESUMING:
+		len = scnprintf(buf, VFIO_DEV_DBG_LEN, "%s\n",
+			"RESUMING\n");
+		break;
+	case VFIO_DEVICE_STATE_RUNNING_P2P:
+		len = scnprintf(buf, VFIO_DEV_DBG_LEN, "%s\n",
+			"RESUMING_P2P\n");
+		break;
+	case VFIO_DEVICE_STATE_ERROR:
+		len = scnprintf(buf, VFIO_DEV_DBG_LEN, "%s\n",
+			"ERROR\n");
+		break;
+	default:
+		len = scnprintf(buf, VFIO_DEV_DBG_LEN, "%s\n",
+			"Invalid\n");
+	}
+
+	return simple_read_from_buffer(buffer, count, pos, buf, len);
+}
+
+static const struct file_operations vfio_vf_state_fops = {
+	.owner = THIS_MODULE,
+	.open = simple_open,
+	.read = vfio_pci_vf_state_read,
+};
+
+void vfio_pci_vf_debugfs_init(struct vfio_device *vdev, size_t migf_size)
+{
+	struct device *dev = vdev->dev;
+	int ret;
+
+	vdev->debug_root = debugfs_create_dir(dev_name(dev), vfio_pci_debugfs_root);
+	debugfs_create_file("state", 0444, vdev->debug_root,
+				  vdev, &vfio_vf_state_fops);
+
+	debugfs_create_file("mig_data", 0444, vdev->debug_root,
+				  vdev, &vfio_vf_data_fops);
+
+	debugfs_create_file("mig_attr", 0444, vdev->debug_root,
+				  vdev, &vfio_vf_attr_fops);
+
+	ret = vfio_pci_vf_debug_create(vdev, migf_size);
+	if (ret) {
+		dev_err(dev, "failed to alloc migration debug node\n");
+		return;
+	}
+	debugfs_create_file("debug", 0200, vdev->debug_root,
+				  vdev, &vfio_vf_debug_fops);
+}
+EXPORT_SYMBOL_GPL(vfio_pci_vf_debugfs_init);
+
+void vfio_pci_vf_debugfs_exit(struct vfio_device *vdev)
+{
+	vfio_pci_vf_debug_release(vdev);
+	debugfs_remove_recursive(vdev->debug_root);
+}
+EXPORT_SYMBOL_GPL(vfio_pci_vf_debugfs_exit);
+
+void  vfio_pci_debugfs_create_root(void)
+{
+	vfio_pci_debugfs_root = debugfs_create_dir("vfio_mig", NULL);
+}
+EXPORT_SYMBOL_GPL(vfio_pci_debugfs_create_root);
+
+void  vfio_pci_debugfs_remove_root(void)
+{
+	debugfs_remove_recursive(vfio_pci_debugfs_root);
+	vfio_pci_debugfs_root = NULL;
+}
+EXPORT_SYMBOL_GPL(vfio_pci_debugfs_remove_root);
+
diff --git a/drivers/vfio/pci/vfio_pci_debugfs.h b/drivers/vfio/pci/vfio_pci_debugfs.h
new file mode 100644
index 000000000000..34ae6de41a2f
--- /dev/null
+++ b/drivers/vfio/pci/vfio_pci_debugfs.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * vfio_pci_debugfs.h - VFIO migration debugfs interface
+ *
+ * Copyright (c) 2022, HiSilicon Ltd.
+ *
+ * Author: Longfang Liu <liulongfang@huawei.com>
+ */
+
+#ifndef __VFIO_PCI_DEBUGFS_H
+#define __VFIO_PCI_DEBUGFS_H
+
+#include <linux/debugfs.h>
+#include <linux/vfio.h>
+
+#ifdef CONFIG_DEBUG_FS
+void  vfio_pci_debugfs_create_root(void);
+void  vfio_pci_debugfs_remove_root(void);
+
+void vfio_pci_vf_debugfs_init(struct vfio_device *vdev, size_t migf_size);
+void vfio_pci_vf_debugfs_exit(struct vfio_device *vdev);
+#else
+static inline void vfio_pci_debugfs_create_root(void) { }
+static inline void vfio_pci_debugfs_remove_root(void) { }
+
+static inline void vfio_pci_vf_debugfs_init(struct vfio_device *vdev, size_t migf_size) { }
+static inline void vfio_pci_vf_debugfs_exit(struct vfio_device *vdev) { }
+#endif /* CONFIG_DEBUG_FS */
+
+#endif /* __VFIO_PCI_DEBUGFS_H */
diff --git a/include/linux/vfio.h b/include/linux/vfio.h
index 43b67e46a2cb..dd1299acb69a 100644
--- a/include/linux/vfio.h
+++ b/include/linux/vfio.h
@@ -54,6 +54,13 @@ struct vfio_device {
 	struct completion comp;
 	struct list_head group_next;
 	struct list_head iommu_entry;
+
+	/*
+	 * debug_root is a static property of the vfio_device
+	 * which must be set prior to registering the vfio_device.
+	 */
+	struct dentry *debug_root;
+	void *vf_migration_file;
 };
 
 /**
@@ -110,6 +117,12 @@ struct vfio_device_ops {
  * @migration_get_data_size: Optional callback to get the estimated data
  *          length that will be required to complete stop copy. It's mandatory for
  *          VFIO_DEVICE_FEATURE_MIGRATION migration support.
+ * @migration_get_data: Optional callback to get the migration data
+ *         that will be required by vfio debugfs.
+ * @migration_get_attr: Optional callback to get the device attributes
+ *         that will be required by vfio debugfs.
+ * @migration_debug_operate: Optional callback for step-by-step debugging
+ *         of the live migration function.
  */
 struct vfio_migration_ops {
 	struct file *(*migration_set_state)(
@@ -119,6 +132,9 @@ struct vfio_migration_ops {
 				   enum vfio_device_mig_state *curr_state);
 	int (*migration_get_data_size)(struct vfio_device *device,
 				       unsigned long *stop_copy_length);
+	int (*migration_get_data)(struct vfio_device *device, char *buffer);
+	int (*migration_get_attr)(struct vfio_device *device, char *buffer);
+	int (*migration_debug_operate)(struct vfio_device *device, unsigned int cmd);
 };
 
 /**
-- 
2.24.0

[PATCH v3 2/5] hisi_acc_vfio_pci: extract public functions for container_of

[Longfang Liu]

In the current driver, vdev is obtained from struct
hisi_acc_vf_core_device through the container_of function.
This method is used in many places in the driver. In order to
reduce this repetitive operation, I extracted a public function
to replace it.

Signed-off-by: Longfang Liu <liulongfang@huawei.com>
---
 .../vfio/pci/hisilicon/hisi_acc_vfio_pci.c    | 21 ++++++++++---------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
index 0c0c0c7f0521..eb18edffbd5f 100644
--- a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
+++ b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
@@ -633,6 +633,12 @@ static void hisi_acc_vf_disable_fds(struct hisi_acc_vf_core_device *hisi_acc_vde
 	}
 }
 
+static struct hisi_acc_vf_core_device *hisi_acc_get_vf_dev(struct vfio_device *vdev)
+{
+	return container_of(vdev, struct hisi_acc_vf_core_device,
+			    core_device.vdev);
+}
+
 /*
  * This function is called in all state_mutex unlock cases to
  * handle a 'deferred_reset' if exists.
@@ -927,8 +933,7 @@ static struct file *
 hisi_acc_vfio_pci_set_device_state(struct vfio_device *vdev,
 				   enum vfio_device_mig_state new_state)
 {
-	struct hisi_acc_vf_core_device *hisi_acc_vdev = container_of(vdev,
-			struct hisi_acc_vf_core_device, core_device.vdev);
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(vdev);
 	enum vfio_device_mig_state next_state;
 	struct file *res = NULL;
 	int ret;
@@ -969,8 +974,7 @@ static int
 hisi_acc_vfio_pci_get_device_state(struct vfio_device *vdev,
 				   enum vfio_device_mig_state *curr_state)
 {
-	struct hisi_acc_vf_core_device *hisi_acc_vdev = container_of(vdev,
-			struct hisi_acc_vf_core_device, core_device.vdev);
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(vdev);
 
 	mutex_lock(&hisi_acc_vdev->state_mutex);
 	*curr_state = hisi_acc_vdev->mig_state;
@@ -1186,8 +1190,7 @@ static long hisi_acc_vfio_pci_ioctl(struct vfio_device *core_vdev, unsigned int
 
 static int hisi_acc_vfio_pci_open_device(struct vfio_device *core_vdev)
 {
-	struct hisi_acc_vf_core_device *hisi_acc_vdev = container_of(core_vdev,
-			struct hisi_acc_vf_core_device, core_device.vdev);
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(core_vdev);
 	struct vfio_pci_core_device *vdev = &hisi_acc_vdev->core_device;
 	int ret;
 
@@ -1210,8 +1213,7 @@ static int hisi_acc_vfio_pci_open_device(struct vfio_device *core_vdev)
 
 static void hisi_acc_vfio_pci_close_device(struct vfio_device *core_vdev)
 {
-	struct hisi_acc_vf_core_device *hisi_acc_vdev = container_of(core_vdev,
-			struct hisi_acc_vf_core_device, core_device.vdev);
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(core_vdev);
 	struct hisi_qm *vf_qm = &hisi_acc_vdev->vf_qm;
 
 	iounmap(vf_qm->io_base);
@@ -1226,8 +1228,7 @@ static const struct vfio_migration_ops hisi_acc_vfio_pci_migrn_state_ops = {
 
 static int hisi_acc_vfio_pci_migrn_init_dev(struct vfio_device *core_vdev)
 {
-	struct hisi_acc_vf_core_device *hisi_acc_vdev = container_of(core_vdev,
-			struct hisi_acc_vf_core_device, core_device.vdev);
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(core_vdev);
 	struct pci_dev *pdev = to_pci_dev(core_vdev->dev);
 	struct hisi_qm *pf_qm = hisi_acc_get_pf_qm(pdev);
 
-- 
2.24.0

[PATCH v3 3/5] hisi_acc_vfio_pci: register debugfs for hisilicon migration driver

[Longfang Liu]

On the debugfs framework of VFIO, register the debug function
for the live migration driver of the hisilicon accelerator device.

Signed-off-by: Longfang Liu <liulongfang@huawei.com>
---
 .../vfio/pci/hisilicon/hisi_acc_vfio_pci.c    | 168 ++++++++++++++++++
 .../vfio/pci/hisilicon/hisi_acc_vfio_pci.h    |   6 +
 2 files changed, 174 insertions(+)

diff --git a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
index eb18edffbd5f..d683d48911d1 100644
--- a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
+++ b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
@@ -15,6 +15,7 @@
 #include <linux/anon_inodes.h>
 
 #include "hisi_acc_vfio_pci.h"
+#include "../vfio_pci_debugfs.h"
 
 /* Return 0 on VM acc device ready, -ETIMEDOUT hardware timeout */
 static int qm_wait_dev_not_ready(struct hisi_qm *qm)
@@ -609,6 +610,18 @@ hisi_acc_check_int_state(struct hisi_acc_vf_core_device *hisi_acc_vdev)
 	}
 }
 
+static void hisi_acc_vf_migf_save(struct hisi_acc_vf_migration_file *src_migf,
+	struct hisi_acc_vf_migration_file *dst_migf)
+{
+	if (!dst_migf)
+		return;
+
+	dst_migf->disabled = false;
+	dst_migf->total_length = src_migf->total_length;
+	memcpy(&dst_migf->vf_data, &src_migf->vf_data,
+		    sizeof(struct acc_vf_data));
+}
+
 static void hisi_acc_vf_disable_fd(struct hisi_acc_vf_migration_file *migf)
 {
 	mutex_lock(&migf->lock);
@@ -620,13 +633,19 @@ static void hisi_acc_vf_disable_fd(struct hisi_acc_vf_migration_file *migf)
 
 static void hisi_acc_vf_disable_fds(struct hisi_acc_vf_core_device *hisi_acc_vdev)
 {
+	struct vfio_device	*vdev = &hisi_acc_vdev->core_device.vdev;
+
 	if (hisi_acc_vdev->resuming_migf) {
+		hisi_acc_vf_migf_save(hisi_acc_vdev->resuming_migf,
+						vdev->vf_migration_file);
 		hisi_acc_vf_disable_fd(hisi_acc_vdev->resuming_migf);
 		fput(hisi_acc_vdev->resuming_migf->filp);
 		hisi_acc_vdev->resuming_migf = NULL;
 	}
 
 	if (hisi_acc_vdev->saving_migf) {
+		hisi_acc_vf_migf_save(hisi_acc_vdev->saving_migf,
+						vdev->vf_migration_file);
 		hisi_acc_vf_disable_fd(hisi_acc_vdev->saving_migf);
 		fput(hisi_acc_vdev->saving_migf->filp);
 		hisi_acc_vdev->saving_migf = NULL;
@@ -1188,6 +1207,145 @@ static long hisi_acc_vfio_pci_ioctl(struct vfio_device *core_vdev, unsigned int
 	return vfio_pci_core_ioctl(core_vdev, cmd, arg);
 }
 
+static int hisi_acc_vf_debug_io(struct vfio_device *vdev)
+{
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(vdev);
+	struct hisi_qm *vf_qm = &hisi_acc_vdev->vf_qm;
+	struct device *dev = vdev->dev;
+	u64 data;
+	int ret;
+
+	data = readl(vf_qm->io_base + QM_MB_CMD_SEND_BASE);
+	dev_info(dev, "debug mailbox val: 0x%llx\n", data);
+
+	ret = qm_wait_dev_not_ready(vf_qm);
+	if (ret)
+		dev_err(dev, "VF device not ready!\n");
+
+	return ret;
+}
+
+static int hisi_acc_vf_debug_resume(struct vfio_device *vdev)
+{
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(vdev);
+	struct hisi_acc_vf_migration_file *migf =
+		(struct hisi_acc_vf_migration_file *)vdev->vf_migration_file;
+	struct device *dev = vdev->dev;
+	int ret;
+
+	ret = vf_qm_state_save(hisi_acc_vdev, migf);
+	if (ret) {
+		dev_err(dev, "failed to save device data!\n");
+		return -EINVAL;
+	}
+
+	ret = vf_qm_check_match(hisi_acc_vdev, migf);
+	if (ret) {
+		dev_err(dev, "failed to match the VF!\n");
+		return -EINVAL;
+	}
+
+	ret = vf_qm_load_data(hisi_acc_vdev, migf);
+	if (ret) {
+		dev_err(dev, "failed to recover the VF!\n");
+		return -EINVAL;
+	}
+
+	vf_qm_fun_reset(&hisi_acc_vdev->vf_qm);
+	dev_info(dev, "successful to resume device data!\n");
+
+	return 0;
+}
+
+static int hisi_acc_vf_debug_save(struct vfio_device *vdev)
+{
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(vdev);
+	struct hisi_acc_vf_migration_file *migf =
+		(struct hisi_acc_vf_migration_file *)vdev->vf_migration_file;
+	struct device *dev = vdev->dev;
+	int ret;
+
+	ret = vf_qm_state_save(hisi_acc_vdev, migf);
+	if (ret) {
+		dev_err(dev, "failed to save device data!\n");
+		return -EINVAL;
+	}
+	dev_info(dev, "successful to save device data!\n");
+
+	return 0;
+}
+
+static int
+hisi_acc_vfio_pci_debug_operate(struct vfio_device *vdev, unsigned int cmd)
+{
+	int ret;
+
+	switch (cmd) {
+	case STATE_SAVE:
+		ret = hisi_acc_vf_debug_save(vdev);
+		if (ret)
+			return ret;
+		break;
+	case STATE_RESUME:
+		ret = hisi_acc_vf_debug_resume(vdev);
+		if (ret)
+			return ret;
+		break;
+	case RW_IO_TEST:
+		ret = hisi_acc_vf_debug_io(vdev);
+		if (ret)
+			return ret;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int
+hisi_acc_vfio_pci_get_device_data(struct vfio_device *vdev, char *buff)
+{
+	struct hisi_acc_vf_migration_file *debug_migf =
+		(struct hisi_acc_vf_migration_file *)vdev->vf_migration_file;
+	size_t vf_data_sz = offsetofend(struct acc_vf_data, padding);
+	struct device *dev = vdev->dev;
+
+	if (debug_migf && debug_migf->total_length) {
+		memcpy(buff, &debug_migf->vf_data, vf_data_sz);
+		return vf_data_sz;
+	}
+	dev_info(dev, "device not migrated!\n");
+
+	return 0;
+}
+
+static int
+hisi_acc_vfio_pci_get_device_attr(struct vfio_device *vdev, char *buff)
+{
+	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(vdev);
+	struct hisi_acc_vf_migration_file *debug_migf =
+		(struct hisi_acc_vf_migration_file *)vdev->vf_migration_file;
+	struct device *dev = vdev->dev;
+
+	if (debug_migf && debug_migf->total_length) {
+		scnprintf(buff, PAGE_SIZE,
+			 "acc device:\n"
+			 "device  state: %d\n"
+			 "device  ready: %u\n"
+			 "data    valid: %d\n"
+			 "data     size: %lu\n",
+			 hisi_acc_vdev->mig_state,
+			 hisi_acc_vdev->vf_qm_state,
+			 debug_migf->disabled,
+			 debug_migf->total_length);
+		return PAGE_SIZE;
+	}
+	dev_info(dev, "device not migrated!\n");
+
+	return 0;
+}
+
 static int hisi_acc_vfio_pci_open_device(struct vfio_device *core_vdev)
 {
 	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(core_vdev);
@@ -1224,6 +1382,9 @@ static const struct vfio_migration_ops hisi_acc_vfio_pci_migrn_state_ops = {
 	.migration_set_state = hisi_acc_vfio_pci_set_device_state,
 	.migration_get_state = hisi_acc_vfio_pci_get_device_state,
 	.migration_get_data_size = hisi_acc_vfio_pci_get_data_size,
+	.migration_get_data = hisi_acc_vfio_pci_get_device_data,
+	.migration_get_attr = hisi_acc_vfio_pci_get_device_attr,
+	.migration_debug_operate = hisi_acc_vfio_pci_debug_operate,
 };
 
 static int hisi_acc_vfio_pci_migrn_init_dev(struct vfio_device *core_vdev)
@@ -1295,6 +1456,10 @@ static int hisi_acc_vfio_pci_probe(struct pci_dev *pdev, const struct pci_device
 	if (IS_ERR(hisi_acc_vdev))
 		return PTR_ERR(hisi_acc_vdev);
 
+	if (ops == &hisi_acc_vfio_pci_migrn_ops)
+		vfio_pci_vf_debugfs_init(&hisi_acc_vdev->core_device.vdev,
+						sizeof(struct hisi_acc_vf_migration_file));
+
 	dev_set_drvdata(&pdev->dev, &hisi_acc_vdev->core_device);
 	ret = vfio_pci_core_register_device(&hisi_acc_vdev->core_device);
 	if (ret)
@@ -1302,6 +1467,8 @@ static int hisi_acc_vfio_pci_probe(struct pci_dev *pdev, const struct pci_device
 	return 0;
 
 out_put_vdev:
+	if (ops == &hisi_acc_vfio_pci_migrn_ops)
+		vfio_pci_vf_debugfs_exit(&hisi_acc_vdev->core_device.vdev);
 	vfio_put_device(&hisi_acc_vdev->core_device.vdev);
 	return ret;
 }
@@ -1310,6 +1477,7 @@ static void hisi_acc_vfio_pci_remove(struct pci_dev *pdev)
 {
 	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_drvdata(pdev);
 
+	vfio_pci_vf_debugfs_exit(&hisi_acc_vdev->core_device.vdev);
 	vfio_pci_core_unregister_device(&hisi_acc_vdev->core_device);
 	vfio_put_device(&hisi_acc_vdev->core_device.vdev);
 }
diff --git a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h
index 67343325b320..64b52fed69f2 100644
--- a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h
+++ b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.h
@@ -49,6 +49,12 @@
 #define QM_EQC_DW0		0X8000
 #define QM_AEQC_DW0		0X8020
 
+enum mig_debug_cmd {
+	STATE_SAVE,
+	STATE_RESUME,
+	RW_IO_TEST,
+};
+
 struct acc_vf_data {
 #define QM_MATCH_SIZE offsetofend(struct acc_vf_data, qm_rsv_state)
 	/* QM match information */
-- 
2.24.0

[PATCH v3 4/5] Documentation: add debugfs description for hisi_acc_vfio_pci

[Longfang Liu]

Add a debugfs document description file to help users understand
how to use the accelerator live migration driver's debugfs.

Signed-off-by: Longfang Liu <liulongfang@huawei.com>
---
 .../ABI/testing/debugfs-hisi-migration        | 32 +++++++++++++++++++
 1 file changed, 32 insertions(+)
 create mode 100644 Documentation/ABI/testing/debugfs-hisi-migration

diff --git a/Documentation/ABI/testing/debugfs-hisi-migration b/Documentation/ABI/testing/debugfs-hisi-migration
new file mode 100644
index 000000000000..a8cfc7688b88
--- /dev/null
+++ b/Documentation/ABI/testing/debugfs-hisi-migration
@@ -0,0 +1,32 @@
+What:		/sys/kernel/debug/vfio_mig/<bdf>/state
+Date:		Nov 2022
+Contact:	linux-kernel@vger.kernel.org
+Description:	Read the live migration status of the VF device.
+		The status of these live migrations includes:
+		ERROR, RUNNING, STOP, STOP_COPY, RESUMING.
+
+What:		/sys/kernel/debug/vfio_mig/<bdf>/mig_data
+Date:		Nov 2022
+Contact:	linux-kernel@vger.kernel.org
+Description:	Read the live migration data of the VF device.
+		The output format of the data is defined by the live
+		migration driver.
+
+What:		/sys/kernel/debug/vfio_mig/<bdf>/mig_attr
+Date:		Nov 2022
+Contact:	linux-kernel@vger.kernel.org
+Description:	Read the live migration attributes of the VF device.
+		The output format of the attributes is defined by the live
+		migration driver.
+
+What:		/sys/kernel/debug/vfio_mig/<bdf>/debug
+Date:		Nov 2022
+Contact:	linux-kernel@vger.kernel.org
+Description:	This debug file supports "echo" command	write operations.
+		The command is actually represented by an integer number, and the
+		specific meaning is defined by the corresponding device driver.
+		For example, three commands are defined in the HiSilicon accelerator
+		live migration driver:
+		0: State save
+		1: State resume
+		2: IO test
-- 
2.24.0

[PATCH v3 5/5] vfio: update live migration device status

[Longfang Liu]

migration debugfs needs to perform debug operations based on the
status of the current device. If the device is not loaded or has
stopped, debugfs does not allow operations.

so, after the live migration function is executed and the device is
turned off, the device no longer needs to be accessed. At this time,
the status of the device needs to be set to stop.

Signed-off-by: Longfang Liu <liulongfang@huawei.com>
---
 drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c | 3 +++
 drivers/vfio/pci/mlx5/main.c                   | 3 +++
 2 files changed, 6 insertions(+)

diff --git a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
index d683d48911d1..16faa7632035 100644
--- a/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
+++ b/drivers/vfio/pci/hisilicon/hisi_acc_vfio_pci.c
@@ -1374,6 +1374,9 @@ static void hisi_acc_vfio_pci_close_device(struct vfio_device *core_vdev)
 	struct hisi_acc_vf_core_device *hisi_acc_vdev = hisi_acc_get_vf_dev(core_vdev);
 	struct hisi_qm *vf_qm = &hisi_acc_vdev->vf_qm;
 
+	if (core_vdev->mig_ops)
+		hisi_acc_vdev->mig_state = VFIO_DEVICE_STATE_STOP;
+
 	iounmap(vf_qm->io_base);
 	vfio_pci_core_close_device(core_vdev);
 }
diff --git a/drivers/vfio/pci/mlx5/main.c b/drivers/vfio/pci/mlx5/main.c
index 6e9cf2aacc52..0be038a8bf39 100644
--- a/drivers/vfio/pci/mlx5/main.c
+++ b/drivers/vfio/pci/mlx5/main.c
@@ -587,6 +587,9 @@ static void mlx5vf_pci_close_device(struct vfio_device *core_vdev)
 	struct mlx5vf_pci_core_device *mvdev = container_of(
 		core_vdev, struct mlx5vf_pci_core_device, core_device.vdev);
 
+	if (mvdev->migrate_cap)
+		mvdev->mig_state = VFIO_DEVICE_STATE_STOP;
+
 	mlx5vf_cmd_close_migratable(mvdev);
 	vfio_pci_core_close_device(core_vdev);
 }
-- 
2.24.0

Re: [PATCH v3 1/5] vfio/migration: Add debugfs to live migration driver

[Jason Gunthorpe]

On Mon, Nov 28, 2022 at 11:11:56AM +0800, Longfang Liu wrote:

> @@ -119,6 +132,9 @@ struct vfio_migration_ops {
>  				   enum vfio_device_mig_state *curr_state);
>  	int (*migration_get_data_size)(struct vfio_device *device,
>  				       unsigned long *stop_copy_length);
> +	int (*migration_get_data)(struct vfio_device *device, char *buffer);
> +	int (*migration_get_attr)(struct vfio_device *device, char *buffer);
> +	int (*migration_debug_operate)(struct vfio_device *device,
>  				   unsigned int cmd);

Please don't do wrappers like this, give the variant driver a chance
to create its own files natively under the debugfs directory

Jason

Re: [PATCH v3 1/5] vfio/migration: Add debugfs to live migration driver

[liulongfang]

On 2022/11/28 21:39, Jason Gunthorpe wrote:
> On Mon, Nov 28, 2022 at 11:11:56AM +0800, Longfang Liu wrote:
> 
>> @@ -119,6 +132,9 @@ struct vfio_migration_ops {
>>  				   enum vfio_device_mig_state *curr_state);
>>  	int (*migration_get_data_size)(struct vfio_device *device,
>>  				       unsigned long *stop_copy_length);
>> +	int (*migration_get_data)(struct vfio_device *device, char *buffer);
>> +	int (*migration_get_attr)(struct vfio_device *device, char *buffer);
>> +	int (*migration_debug_operate)(struct vfio_device *device,
>>  				   unsigned int cmd);
> 
> Please don't do wrappers like this, give the variant driver a chance
> to create its own files natively under the debugfs directory
> 

Using this public debugfs file does not need to deal with file
creation and destruction.
The driver only needs to define its debug operation command,
isn't it simpler?

Thanks,
Longfang.
> Jason
> .
> 

Re: [PATCH v3 1/5] vfio/migration: Add debugfs to live migration driver

[Jason Gunthorpe]

On Tue, Nov 29, 2022 at 10:02:46AM +0800, liulongfang wrote:
> On 2022/11/28 21:39, Jason Gunthorpe wrote:
> > On Mon, Nov 28, 2022 at 11:11:56AM +0800, Longfang Liu wrote:
> > 
> >> @@ -119,6 +132,9 @@ struct vfio_migration_ops {
> >>  				   enum vfio_device_mig_state *curr_state);
> >>  	int (*migration_get_data_size)(struct vfio_device *device,
> >>  				       unsigned long *stop_copy_length);
> >> +	int (*migration_get_data)(struct vfio_device *device, char *buffer);
> >> +	int (*migration_get_attr)(struct vfio_device *device, char *buffer);
> >> +	int (*migration_debug_operate)(struct vfio_device *device,
> >>  				   unsigned int cmd);
> > 
> > Please don't do wrappers like this, give the variant driver a chance
> > to create its own files natively under the debugfs directory
> > 
>
> Using this public debugfs file does not need to deal with file
> creation and destruction.
> The driver only needs to define its debug operation command,
> isn't it simpler?

No, it is obfuscating the normal kernel apis, what if a driver wants
two files?

debugfs is not a heavy burden to implement files for.

Jason

Re: [PATCH v3 1/5] vfio/migration: Add debugfs to live migration driver

[liulongfang]

On 2022/11/29 21:15, Jason Gunthorpe wrote:
> On Tue, Nov 29, 2022 at 10:02:46AM +0800, liulongfang wrote:
>> On 2022/11/28 21:39, Jason Gunthorpe wrote:
>>> On Mon, Nov 28, 2022 at 11:11:56AM +0800, Longfang Liu wrote:
>>>
>>>> @@ -119,6 +132,9 @@ struct vfio_migration_ops {
>>>>  				   enum vfio_device_mig_state *curr_state);
>>>>  	int (*migration_get_data_size)(struct vfio_device *device,
>>>>  				       unsigned long *stop_copy_length);
>>>> +	int (*migration_get_data)(struct vfio_device *device, char *buffer);
>>>> +	int (*migration_get_attr)(struct vfio_device *device, char *buffer);
>>>> +	int (*migration_debug_operate)(struct vfio_device *device,
>>>>  				   unsigned int cmd);
>>>
>>> Please don't do wrappers like this, give the variant driver a chance
>>> to create its own files natively under the debugfs directory
>>>
>>
>> Using this public debugfs file does not need to deal with file
>> creation and destruction.
>> The driver only needs to define its debug operation command,
>> isn't it simpler?
> 
> No, it is obfuscating the normal kernel apis, what if a driver wants
> two files?
> 

Yes, if there are special devices that require multiple debugfs files
and use the same operation, using this migration_debug_operate() will
cause the function to be impossible.

Thanks,
Longfang.

> debugfs is not a heavy burden to implement files for.
>
> Jason
> .
>