Date:   Wed, 1 Feb 2023 09:36:44 -0700
From:   Alex Williamson <alex.williamson@redhat.com>
To:     liulongfang <liulongfang@huawei.com>
Cc:     <jgg@nvidia.com>, <shameerali.kolothum.thodi@huawei.com>,
        <jonathan.cameron@huawei.com>, <cohuck@redhat.com>,
        <linux-kernel@vger.kernel.org>, <linuxarm@openeuler.org>
Subject: Re: [PATCH v7 1/5] vfio/migration: Add debugfs to live migration
 driver
Message-ID: <20230201093644.4983db0d.alex.williamson@redhat.com>
In-Reply-To: <add8f068-3aaa-0707-bfe6-27fd8bf9e10c@huawei.com>
References: <20230120032930.43608-1-liulongfang@huawei.com>
        <20230120032930.43608-2-liulongfang@huawei.com>
        <20230126122759.5aa4d84c.alex.williamson@redhat.com>
        <add8f068-3aaa-0707-bfe6-27fd8bf9e10c@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Wed, 1 Feb 2023 11:20:08 +0800
liulongfang <liulongfang@huawei.com> wrote:

> On 2023/1/27 3:27, Alex Williamson wrote:
> > On Fri, 20 Jan 2023 11:29:26 +0800
> > Longfang Liu <liulongfang@huawei.com> 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   |
> >>  +-------+--------+        +-------+--------+
> >>
> >> The entire debugfs directory will be based on the definition of
> >> the CONFIG_DEBUG_FS macro. If this macro is not enabled, the
> >> interfaces in vfio_pci_debugfs.h will be empty definitions,
> >> and the creation and initialization of the debugfs directory
> >> will not be executed.
> >>
> >> debugfs will create a public root directory "vfio_mig" file.
> >> then create a "debug_root" file for each live migration VF device.  
> > 
> > s/VF // In practice it may be VFs, but I don't think we want to
> > artificially imply that the device must be a VF.  In fact, I'm not sure  
> 
> Yes, the description should be modified here, and it should not be limited
> to only support VF.
> 
> > why any of this code lives in vfio/pci/ vs vfio/, there's nothing PCI
> > specific here.  Why isn't all of this done in vfio_main to be shared
> > with both pci and non-pci vfio drivers and devices?
> >  
> 
> Combined with your observations below, I think it is more appropriate to
> create a "vfio-pci" directory here, because the current live migration
> function is still used on pci devices.
> 
> For non-pci devices, if there is a device that needs to add debugfs,
> then it can create a similar directory.

But live migration itself is not specific to vfio-pci, it's device
agnostic.  We'd certainly expect to see mdev devices supporting
migration in the near term.  Thanks,

Alex