lkml.org 
[lkml]   [2023]   [Jan]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH v5 1/5] vfio/migration: Add debugfs to live migration driver
From
Date
On 2023/1/3 23:05, Jason Gunthorpe Wrote:
> On Sat, Dec 17, 2022 at 02:32:44PM +0800, Longfang Liu wrote:
>> 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 |
>> +-------+-+------+ +-------+-+------+
>> |Debugfs|state | |Debugfs|state |
>> +-------+--------+ +-------+--------+
>> | debug_root | | debug_root |
>> +-------+--------+ +-------+--------+
>>
>> debugfs will create a public root directory "vfio_mig" file.
>> then create a "debug_root" file for each live migration VF device.
>> state file: used to get the status of the live migration device.
>> debug_root: used to create your own custom debug files for live
>> migration device drivers.
>>
>> The live migration function of the current device can be tested by
>> operating the debug files, 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 files.
>>
>> 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 | 76 +++++++++++++++++++++++++++++
>> drivers/vfio/pci/vfio_pci_debugfs.h | 30 ++++++++++++
>> include/linux/vfio.h | 6 +++
>> 5 files changed, 119 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 ($(CONFIG_DEBUG_FS),)
>> + vfio-pci-core-y += vfio_pci_debugfs.o
>> +endif
>
> This is now how makefile should be written
>
> vfio-pci-core-$(CONFIG_DEBUG_FS) += vfio_pci_debugfs.o
>

Ok, here it will be changed on the next patch version.

>> +static int vfio_pci_vf_state_read(struct seq_file *seq, void *data)
>> +{
>> + struct device *vf_dev = seq->private;
>> + struct vfio_pci_core_device *core_device = dev_get_drvdata(vf_dev);
>> + struct vfio_device *vdev = &core_device->vdev;
>> + enum vfio_device_mig_state state;
>> +
>> + (void)vdev->mig_ops->migration_get_state(vdev, &state);
>
> Do not use (void) like this, and why not handle the error?
>

Ok, here I will add an error handling.

>> + switch (state) {
>> + case VFIO_DEVICE_STATE_RUNNING:
>> + seq_printf(seq, "%s\n", "RUNNING");
>> + break;
>> + case VFIO_DEVICE_STATE_STOP_COPY:
>> + seq_printf(seq, "%s\n", "STOP_COPY");
>> + break;
>> + case VFIO_DEVICE_STATE_STOP:
>> + seq_printf(seq, "%s\n", "STOP");
>> + break;
>> + case VFIO_DEVICE_STATE_RESUMING:
>> + seq_printf(seq, "%s\n", "RESUMING");
>> + break;
>> + case VFIO_DEVICE_STATE_RUNNING_P2P:
>> + seq_printf(seq, "%s\n", "RESUMING_P2P");
>> + break;
>> + case VFIO_DEVICE_STATE_ERROR:
>> + seq_printf(seq, "%s\n", "ERROR");
>> + break;
>> + default:
>> + seq_printf(seq, "%s\n", "Invalid");
>> + }
>> +
>> + return 0;
>> +}
>> +
>> +void vfio_pci_vf_debugfs_init(struct vfio_device *vdev)
>> +{
>> + struct device *dev = vdev->dev;
>> +
>> + vdev->debug_root = debugfs_create_dir(dev_name(dev), vfio_pci_debugfs_root);
>> +
>> + debugfs_create_devm_seqfile(dev, "state", vdev->debug_root,
>> + vfio_pci_vf_state_read);
>> +}
>> +EXPORT_SYMBOL_GPL(vfio_pci_vf_debugfs_init);
>> +
>> +void vfio_pci_vf_debugfs_exit(struct vfio_device *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);
>
> Why are these exported symbols? The core code should handle this on
> its own.
>

I included this set of interfaces through the #ifdef CONFIG_DEBUG_FS macro
in vfio_pci_debugfs.h.
If EXPORT_SYMBOL_GPL is not added here, a compilation error will appear:

ERROR: modpost: "vfio_pci_vf_debugfs_exit" [drivers/vfio/pci/hisilicon/hisi-acc-vfio-pci.ko] undefined!
ERROR: modpost: "vfio_pci_vf_debugfs_init" [drivers/vfio/pci/hisilicon/hisi-acc-vfio-pci.ko] undefined!
scripts/Makefile.modpost:126: recipe for target 'Module.symvers' failed

> If the device defines migration_get_state() then the debugfs should be
> created.
>

The trigger strategy created by the current debugfs is determined in the device
driver according to the setting method of vfio_device_ops.
If vfio_device_ops is enabled with live migration, then its migration_get_state()
will also be enabled.

> Mabye you need an op to populate the driver's debugfs to also trigger
> creating the directory
>

It is also possible to put this trigger strategy in the corresponding device driver, right?
Putting this strategy in this public vfio_pci_debugfs.c will also increase the implementation
complexity of this function.

> Jason
> .
>
Thanks,
Longfang.

\
 
 \ /
  Last update: 2023-03-26 23:26    [W:0.083 / U:0.164 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site