On Thu, Jan 02, 2025 at 02:21:13PM -0500, Steven Sistare wrote:
> On 12/24/2024 2:24 PM, Peter Xu wrote:
> > On Tue, Dec 24, 2024 at 08:17:00AM -0800, Steve Sistare wrote:
> > > Add the cpr-transfer migration mode, which allows the user to transfer
> > > a guest to a new QEMU instance on the same host with minimal guest pause
> > > time, by preserving guest RAM in place, albeit with new virtual addresses
> > > in new QEMU, and by preserving device file descriptors. Pages that were
> > > locked in memory for DMA in old QEMU remain locked in new QEMU, because
> > > the
> > > descriptor of the device that locked them remains open.
> > >
> > > cpr-transfer preserves memory and devices descriptors by sending them to
> > > new QEMU over a unix domain socket using SCM_RIGHTS. Such CPR state
> > > cannot
> > > be sent over the normal migration channel, because devices and backends
> > > are created prior to reading the channel, so this mode sends CPR state
> > > over a second "cpr" migration channel. New QEMU reads the cpr channel
> > > prior to creating devices or backends. The user specifies the cpr channel
> > > in the channel arguments on the outgoing side, and in a second -incoming
> > > command-line parameter on the incoming side.
> > >
> > > The user must start old QEMU with the the '-machine aux-ram-share=on'
> > > option,
> > > which allows anonymous memory to be transferred in place to the new
> > > process
> > > by transferring a memory descriptor for each ram block. Memory-backend
> > > objects must have the share=on attribute, but memory-backend-epc is not
> > > supported.
> > >
> > > The user starts new QEMU on the same host as old QEMU, with command-line
> > > arguments to create the same machine, plus the -incoming option for the
> > > main migration channel, like normal live migration. In addition, the user
> > > adds a second -incoming option with channel type "cpr". The CPR channel
> > > address must be a type, such as unix socket, that supports SCM_RIGHTS.
> > >
> > > To initiate CPR, the user issues a migrate command to old QEMU, adding
> > > a second migration channel of type "cpr" in the channels argument.
> > > Old QEMU stops the VM, saves state to the migration channels, and enters
> > > the postmigrate state. New QEMU mmap's memory descriptors, and execution
> > > resumes.
> > >
> > > The implementation splits qmp_migrate into start and finish functions.
> > > Start sends CPR state to new QEMU, which responds by closing the CPR
> > > channel. Old QEMU detects the HUP then calls finish, which connects the
> > > main migration channel.
> > >
> > > In summary, the usage is:
> > >
> > > qemu-system-$arch -machine aux-ram-share=on ...
> > >
> > > start new QEMU with "-incoming <main-uri> -incoming <cpr-channel>"
> > >
> > > Issue commands to old QEMU:
> > > migrate_set_parameter mode cpr-transfer
> > >
> > > {"execute": "migrate", ...
> > > {"channel-type": "main"...}, {"channel-type": "cpr"...} ... }
> > >
> > > Signed-off-by: Steve Sistare <steven.sist...@oracle.com>
> >
> > Feel free to take:
> >
> > Reviewed-by: Peter Xu <pet...@redhat.com>
> >
> > I still have a few trivial comments.
> >
> > [...]
> >
> > > diff --git a/migration/cpr.c b/migration/cpr.c
> > > index 87bcfdb..584b0b9 100644
> > > --- a/migration/cpr.c
> > > +++ b/migration/cpr.c
> > > @@ -45,7 +45,7 @@ static const VMStateDescription vmstate_cpr_fd = {
> > > VMSTATE_UINT32(namelen, CprFd),
> > > VMSTATE_VBUFFER_ALLOC_UINT32(name, CprFd, 0, NULL, namelen),
> > > VMSTATE_INT32(id, CprFd),
> >
> > Could you remind me again on when id!=0 will start to be used?
>
> Each of vfio, iommufd, chardev, and tap will use id != 0.
I don't remember the details of the planned future series, but just to
mention that using integer ID can be error prone on device hot plug/unplug.
QEMU has a known bug even now on some device (e.g. slirp network backends)
that if the src QEMU originally has two devices (e.g. id=1,2), unplug
device id=1 (leaving id=2), then migrate, it could fail seeing dest only
has id=1 (dest QEMU starts with only one device), seeing a mismatched ID.
I recall PCIe frontend devices are not prone to such issue, that should
depend on whoever has ->get_id() (qdev_get_dev_path?) properly implemented
to generate a global unique ID that is not affected by order of device
realized / created.
It could boil down to how the IDs are allocated, anything that can be
allocated on the fly may not work well if there's no solid topology
information to fetch.
I wonder if CPR can be prone to this too when using IDs, just FYI. It
might be a good idea if ID integers can be avoided somehow. But you'll
definitely have the best picture of the whole thing, so it may or may not
apply.
Thanks,
--
Peter Xu
