Enable live migration only once the firmware supports PRE_COPY.
PRE_COPY has been supported by the firmware for a long time already [1]
and is required to achieve a low downtime upon live migration.
This lets us clean up some old code that is not applicable those days
while PRE_COPY is fully supported by the firmware.
[1] The minimum firmware version that supports PRE_COPY is 28.36.1010,
it was released in January 2023.
No firmware without PRE_COPY support ever available to users.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20240306105624.114830-1-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Let firmware knows upon leaving PRE_COPY back to RUNNING as of some
error in the target/migration cancellation.
This will let firmware cleaning its internal resources that were turned
on upon PRE_COPY.
The flow is based on the device specification in this area.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Leon Romanovsky <leon@kernel.org>
Link: https://lore.kernel.org/r/20240205124828.232701-6-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Add support for tracker object change event by referring to its
MLX5_EVENT_TYPE_OBJECT_CHANGE event when occurs.
This lets the driver recognize whether the firmware moved the tracker
object to an error state.
In that case, the driver will skip/block any usage of that object
including an early exit in case the object was previously marked with an
error.
This functionality also covers the case when no CQE is delivered as of
the error state.
The driver was adapted to the device specification to handle the above.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Leon Romanovsky <leon@kernel.org>
Link: https://lore.kernel.org/r/20240205124828.232701-3-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Add support for SAVING in chunk mode, it includes running a work
that will fill the next chunk from the device.
In case the number of available chunks will reach the MAX_NUM_CHUNKS,
the next chunk SAVING will be delayed till the reader will consume one
chunk.
The next patch from the series will add the reader part of the chunk
mode.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20230911093856.81910-8-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
This patch is another preparation step towards working in chunk mode.
It pre-allocates chunks for the STOP_COPY phase to let the driver use
them immediately and prevent an extra allocation upon that phase.
Before that patch we had a single large buffer that was dedicated for
the STOP_COPY phase as there was a single SAVE in the source for the
last image.
Once we'll move to chunk mode the idea is to have some small buffers
that will be used upon the STOP_COPY phase.
The driver will read-ahead from the firmware the full state in
small/optimized chunks while letting QEMU/user space read in parallel
the available data.
Each buffer holds its chunk number to let it be recognized down the road
in the coming patches.
The chunk buffer size is picked-up based on the minimum size that
firmware requires, the total full size and some max value in the driver
code which was set to 8MB to achieve some optimized downtime in the
general case.
As the chunk mode is applicable even if we move directly to STOP_COPY
the buffers preparation and some other related stuff is done
unconditionally with regards to STOP/PRE-COPY.
Note:
In that phase in the series we still didn't activate the chunk mode and
the first buffer will be used in all the places.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20230911093856.81910-7-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Upon chunk mode there may be multiple images that will be read from the
device upon STOP_COPY.
This patch is some preparation for that mode by replacing the relevant
stuff to a better matching name.
As part of that, be stricter to recognize PRE_COPY error only when it
didn't occur on a STOP_COPY chunk.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20230911093856.81910-6-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Once the device supports 'chunk mode' the driver can support state size
which is larger than 4GB.
In that case the device has the capability to split a single image to
multiple chunks as long as the software provides a buffer in the minimum
size reported by the device.
The driver should query for the minimum buffer size required using
QUERY_VHCA_MIGRATION_STATE command with the 'chunk' bit set in its
input, in that case, the output will include both the minimum buffer
size (i.e. required_umem_size) and also the remaining total size to be
reported/used where that it will be applicable.
At that point in the series the 'chunk' bit is off, the last patch will
activate the feature once all pieces will be ready.
Note:
Before this change we were limited to 4GB state size as of 4 bytes max
value based on the device specification for the query/save/load
commands.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20230911093856.81910-5-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Improve the target side flow to reduce downtime as of below.
- Support reading an optional record which includes the expected
stop_copy size.
- Once the source sends this record data, which expects to be sent as
part of the pre_copy flow, prepare the data buffers that may be large
enough to hold the final stop_copy data.
The above reduces the migration downtime as the relevant stuff that is
needed to load the image data is prepared ahead as part of pre_copy.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20230124144955.139901-4-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Improve the source side flow upon pre_copy as of below.
- Prepare the stop_copy buffers as part of moving to pre_copy.
- Send to the target a record that includes the expected
stop_copy size to let it optimize its stop_copy flow as well.
As for sending the target this new record type (i.e.
MLX5_MIGF_HEADER_TAG_STOP_COPY_SIZE) we split the current 64 header
flags bits into 32 flags bits and another 32 tag bits, each record may
have a tag and a flag whether it's optional or mandatory. Optional
records will be ignored in the target.
The above reduces the downtime upon stop_copy as the relevant data stuff
is prepared ahead as part of pre_copy.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20230124144955.139901-3-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Add a check whether VF is migratable. Only if VF is migratable, mark the
VFIO device as migration capable.
Signed-off-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20230124144955.139901-2-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Before a SAVE command is issued, a QUERY command is issued in order to
know the device data size.
In case PRE_COPY is used, the above commands are issued while the device
is running. Thus, it is possible that between the QUERY and the SAVE
commands the state of the device will be changed significantly and thus
the SAVE will fail.
Currently, if a SAVE command is failing, the driver will fail the
migration. In the above case, don't fail the migration, but don't allow
for new SAVEs to be executed while the device is in a RUNNING state.
Once the device will be moved to STOP_COPY, SAVE can be executed again
and the full device state will be read.
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-14-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
In order to support PRE_COPY, mlx5 driver transfers multiple states
(images) of the device. e.g.: the source VF can save and transfer
multiple states, and the target VF will load them by that order.
This patch implements the changes for the target VF to decompose the
header for each state and to write and load multiple states.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-13-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
During PRE_COPY the migration data FD may have a temporary "end of
stream" that is reached when the initial_bytes were read and no other
dirty data exists yet.
For instance, this may indicate that the device is idle and not
currently dirtying any internal state. When read() is done on this
temporary end of stream the kernel driver should return ENOMSG from
read(). Userspace can wait for more data or consider moving to
STOP_COPY.
To not block the user upon read() and let it get ENOMSG we add a new
state named MLX5_MIGF_STATE_PRE_COPY on the migration file.
In addition, we add the MLX5_MIGF_STATE_SAVE_LAST state to block the
read() once we call the last SAVE upon moving to STOP_COPY.
Any further error will be marked with MLX5_MIGF_STATE_ERROR and the user
won't be blocked.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-12-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
As mentioned in the previous patches, mlx5 is transferring multiple
states when the PRE_COPY protocol is used. This states mechanism
requires the target VM to know the states' size in order to execute
multiple loads. Therefore, add SW header, with the needed information,
for each saved state the source VM is transferring to the target VM.
This patch implements the source VM handling of the headers, following
patch will implement the target VM handling of the headers.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-10-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
In order to support PRE_COPY, mlx5 driver is transferring multiple
states (images) of the device. e.g.: the source VF can save and transfer
multiple states, and the target VF will load them by that order.
The device is saving three kinds of states:
1) Initial state - when the device moves to PRE_COPY state.
2) Middle state - during PRE_COPY phase via VFIO_MIG_GET_PRECOPY_INFO.
There can be multiple states of this type.
3) Final state - when the device moves to STOP_COPY state.
After moving to PRE_COPY state, user is holding the saving migf FD and
can use it. For example: user can start transferring data via read()
callback. Also, user can switch from PRE_COPY to STOP_COPY whenever he
sees it fits. This will invoke saving of final state.
This means that mlx5 VFIO device can be switched to STOP_COPY without
transferring any data in PRE_COPY state. Therefore, when the device
moves to STOP_COPY, mlx5 will store the final state on a dedicated queue
entry on the list.
Co-developed-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-9-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Refactor to use queue based data chunks on the migration file.
The SAVE command adds a chunk to the tail of the queue while the read()
API finds the required chunk and returns its data.
In case the queue is empty but the state of the migration file is
MLX5_MIGF_STATE_COMPLETE, read() may not be blocked but will return 0 to
indicate end of file.
This is a step towards maintaining multiple images and their meta data
(i.e. headers) on the migration file as part of next patches from the
series.
Note:
At that point, we still use a single chunk on the migration file but
becomes ready to support multiple.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-8-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Refactor migration file state to be an emum which is mutual exclusive.
As of that dropped the 'disabled' state as 'error' is the same from
functional point of view.
Next patches from the series will extend this enum for other relevant
states.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-7-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
This patch refactors MKEY usage such as its life cycle will be as of the
migration file instead of allocating/destroying it upon each
SAVE/LOAD command.
This is a preparation step towards the PRE_COPY series where multiple
images will be SAVED/LOADED.
We achieve it by having a new struct named mlx5_vhca_data_buffer which
holds the mkey and its related stuff as of sg_append_table,
allocated_length, etc.
The above fields were taken out from the migration file main struct,
into mlx5_vhca_data_buffer dedicated struct with the proper helpers in
place.
For now we have a single mlx5_vhca_data_buffer per migration file.
However, in coming patches we'll have multiple of them to support
multiple images.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-6-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
This patch refactors PD usage such as its life cycle will be as of the
migration file instead of allocating/destroying it upon each SAVE/LOAD
command.
This is a preparation step towards the PRE_COPY series where multiple
images will be SAVED/LOADED and a single PD can be simply reused.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-5-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Enforce a single SAVE command at a time.
As the SAVE command is an asynchronous one, we must enforce running only
a single command at a time.
This will preserve ordering between multiple calls and protect from
races on the migration file data structure.
This is a must for the next patches from the series where as part of
PRE_COPY we may have multiple images to be saved and multiple SAVE
commands may be issued from different flows.
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20221206083438.37807-4-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Now that everything is ready set the driver DMA logging callbacks if
supported by the device.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20220908183448.195262-11-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Handle async error events and health/recovery flow to safely stop the
tracker upon error scenarios.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20220908183448.195262-10-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Report dirty pages from tracker.
It includes:
Querying for dirty pages in a given IOVA range, this is done by
modifying the tracker into the reporting state and supplying the
required range.
Using the CQ event completion mechanism to be notified once data is
ready on the CQ/QP to be processed.
Once data is available turn on the corresponding bits in the bit map.
This functionality will be used as part of the 'log_read_and_clear'
driver callback in the next patches.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20220908183448.195262-9-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Add support for creating and destroying page tracker object.
This object is used to control/report the device dirty pages.
As part of creating the tracker need to consider the device capabilities
for max ranges and adapt/combine ranges accordingly.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20220908183448.195262-8-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Init QP based resources for dirty tracking to be used upon start
logging.
It includes:
Creating the host and firmware RC QPs, move each of them to its expected
state based on the device specification, etc.
Creating the relevant resources which are needed by both QPs as of UAR,
PD, etc.
Creating the host receive side resources as of MKEY, CQ, receive WQEs,
etc.
The above resources are cleaned-up upon stop logging.
The tracker object that will be introduced by next patches will use
those resources.
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20220908183448.195262-7-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
vfio core checks whether the driver sets some migration op (e.g.
set_state/get_state) and accordingly calls its op.
However, currently mlx5 driver sets the above ops without regards to its
migration caps.
This might lead to unexpected usage/Oops if user space may call to the
above ops even if the driver doesn't support migration. As for example,
the migration state_mutex is not initialized in that case.
The cleanest way to manage that seems to split the migration ops from
the main device ops, this will let the driver setting them separately
from the main ops when it's applicable.
As part of that, validate ops construction on registration and include a
check for VFIO_MIGRATION_STOP_COPY since the uAPI claims it must be set
in migration_flags.
HISI driver was changed as well to match this scheme.
This scheme may enable down the road to come with some extra group of
ops (e.g. DMA log) that can be set without regards to the other options
based on driver caps.
Fixes: 6fadb02126 ("vfio/mlx5: Implement vfio_pci driver for mlx5 devices")
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20220628155910.171454-3-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Protect mlx5vf_disable_fds() upon close device to be called under the
state mutex as done in all other places.
This will prevent a race with any other flow which calls
mlx5vf_disable_fds() as of health/recovery upon
MLX5_PF_NOTIFY_DISABLE_VF event.
Encapsulate this functionality in a separate function named
mlx5vf_cmd_close_migratable() to consider migration caps and for further
usage upon close device.
Fixes: 6fadb02126 ("vfio/mlx5: Implement vfio_pci driver for mlx5 devices")
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Link: https://lore.kernel.org/r/20220628155910.171454-2-yishaih@nvidia.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Use the PF asynchronous command mode for the SAVE state command.
This enables returning earlier to user space upon issuing successfully
the command and improve latency by let things run in parallel.
Link: https://lore.kernel.org/r/20220510090206.90374-5-yishaih@nvidia.com
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Refactor to enable different VFs to run their commands over the PF
command interface in parallel and to not block one each other.
This is done by not using the global PF lock that was used before but
relying on the VF attach/detach mechanism to sync.
Link: https://lore.kernel.org/r/20220510090206.90374-4-yishaih@nvidia.com
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Manage the VF attach/detach callback from the PF.
This lets the driver to enable parallel VFs migration as will be
introduced in the next patch.
As part of this, reorganize the VF is migratable code to be in a
separate function and rename it to be set_migratable() to match its
functionality.
Link: https://lore.kernel.org/r/20220510090206.90374-3-yishaih@nvidia.com
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
This patch adds support for vfio_pci driver for mlx5 devices.
It uses vfio_pci_core to register to the VFIO subsystem and then
implements the mlx5 specific logic in the migration area.
The migration implementation follows the definition from uapi/vfio.h and
uses the mlx5 VF->PF command channel to achieve it.
This patch implements the suspend/resume flows.
Link: https://lore.kernel.org/all/20220224142024.147653-14-yishaih@nvidia.com
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Expose migration commands over the device, it includes: suspend, resume,
get vhca id, query/save/load state.
As part of this adds the APIs and data structure that are needed to manage
the migration data.
Link: https://lore.kernel.org/all/20220224142024.147653-13-yishaih@nvidia.com
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
