There are two ways to check for I/O limits in a BlockDriverState:
- bs->throttle_state: if this pointer is not NULL, it means that this
BDS is member of a throttling group, its ThrottleTimers structure
has been initialized and its I/O limits are ready to be applied.
- bs->io_limits_enabled: if true it means that the throttle_state
pointer is valid _and_ the limits are currently enabled.
The latter is used in several places to check whether a BDS has I/O
limits configured, but what it really checks is whether requests
are being throttled or not. For example, io_limits_enabled can be
temporarily set to false in cases like bdrv_read_unthrottled() without
otherwise touching the throtting configuration of that BDS.
This patch replaces bs->io_limits_enabled with bs->throttle_state in
all cases where what we really want to check is the existence of I/O
limits, not whether they are currently enabled or not.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
throttle_group_unregister_bs() removes a BlockDriverState from its
throttling group and destroys the timers. This means that there must
be no pending throttled requests at that point (because it would be
impossible to complete them), so the caller has to drain them first.
At the moment throttle_group_unregister_bs() is only called from
bdrv_io_limits_disable(), which already takes care of draining the
requests, so there's nothing to worry about, but this patch makes
this invariant explicit in the documentation and adds the relevant
assertions.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The mirror job doesn't update its total length until
it has already started running, so we should translate
a zero-length job-len as meaning 0%.
Otherwise, we may get divide-by-zero faults.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
If we create a buffer directly on the stack by using 12 bytes, there's
no guarantee the 64bit value we want to swap will be aligned, which
could cause errors with undefined behavior.
Spotted with clang -fsanitize=undefined and observed in iotests 15, 26,
44, 115 and 121.
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The 'block-commit' command needs the overlay image of 'top' to
be opened in read-write mode in order to update the backing file
string. If 'top' is not the active layer or its backing file then its
overlay needs to be reopened during the block job.
This is a test case for that scenario.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
'block-commit' needs write access to two different nodes of the chain:
- 'base', because that's where the data is written to.
- the overlay of 'top', because it needs to update the backing file
string to point to 'base' after the operation.
Both images have to be opened in read-write mode, and commit_start()
takes care of reopening them if necessary.
With the current implementation, however, when overlay_bs is reopened
in read-write mode it has the side effect of making 'base' read-only
again, eventually making 'block-commit' fail.
This needs to be fixed in bdrv_reopen(), but until we get to that it
can be worked around simply by swapping the order of base and
overlay_bs in the reopen queue.
In order to reproduce this bug, overlay_bs needs to be initially in
read-only mode. That is: the 'top' parameter of 'block-commit' cannot
be the active layer nor its immediate backing chain.
Cc: qemu-stable@nongnu.org
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
One of the limitations of the 'blockdev-snapshot-sync' command is that
it does not allow passing BlockdevOptions to the newly created
snapshots, so they are always opened using the default values.
Extending the command to allow passing options is not a practical
solution because there is overlap between those options and some of
the existing parameters of the command.
This patch introduces a new 'blockdev-snapshot' command with a simpler
interface: it just takes two references to existing block devices that
will be used as the source and target for the snapshot.
Since the main difference between the two commands is that one of them
creates and opens the target image, while the other uses an already
opened one, the bulk of the implementation is shared.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Passing an empty string allows opening an image but not its backing
file. This was already described in the API documentation, only the
implementation was missing.
This is useful for creating snapshots using images opened with
blockdev-add, since they are not supposed to have a backing image
before the operation.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
We will introduce the 'blockdev-snapshot' command that will require
its own struct for the parameters, so we need to rename this one in
order to avoid name clashes.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The 'snapshot-node-name' parameter of blockdev-snapshot-sync allows
setting the node name of the image that is going to be created.
Before creating the image, external_snapshot_prepare() checks that the
name is not already being used. The check is however incomplete since
it only considers existing node names, but node names must not clash
with device IDs either because they share the same namespace.
If the user attempts to create a snapshot using the name of an
existing device for the 'snapshot-node-name' parameter the operation
will eventually fail, but only after the new image has been created.
This patch replaces bdrv_find_node() with bdrv_lookup_bs() to extend
the check to existing device IDs, and thus detect possible name
clashes before the new image is created.
Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Expose the new read-only-mode option of 'blockdev-change-medium' for the
'change' HMP command.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add an option to qmp_blockdev_change_medium() which allows changing the
read-only status of the block device whose medium is changed.
Some drives do not have a inherently fixed read-only status; for
instance, floppy disks can be set read-only or writable independently of
the drive. Some users may find it useful to be able to therefore change
the read-only status of a block device when changing the medium.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Use separate code paths for the two overloaded functions of the 'change'
HMP command, and invoke the 'blockdev-change-medium' QMP command if used
on a block device (by calling qmp_blockdev_change_medium()).
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Introduce a new QMP command 'blockdev-change-medium' which is intended
to replace the 'change' command for block devices. The existing function
qmp_change_blockdev() is accordingly renamed to
qmp_blockdev_change_medium().
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
blk_dev_change_media_cb() is called for all potential tray movements;
however, it is possible to request closing the tray but nothing actually
happening (on a floppy disk drive without a medium).
Thus, the actual tray status should be inquired before sending a
tray-moved event (and an event should be sent whenever the status
changed).
Checking @load is now superfluous; it was necessary because it was
possible to change a medium without having explicitly opened the tray
and closed it again (or it might have been possible, at least). This is
no longer possible, though.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Implement 'change' on block devices by calling blockdev-open-tray,
blockdev-remove-medium, blockdev-insert-medium (a variation of that
which does not need a node-name) and blockdev-close-tray.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Implement 'eject' by calling blockdev-open-tray and
blockdev-remove-medium.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
And a helper function for that, which directly takes a pointer to the
BDS to be inserted instead of its node-name (which will be used for
implementing 'change' using blockdev-insert-medium).
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
In order to open a BDS which inherits a BB's root state,
blk_get_open_flags_from_root_state() is used to inquire the flags to be
passed to bdrv_open(), and blk_apply_root_state() is used to apply the
remaining state after the BDS has been opened.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
When inserting a BDS tree into a BB, we will need to add the root BDS to
this list. Since we will want to do that in the blockdev-insert-medium
implementation in blockdev.c, we will need access to it there.
This patch is not exactly elegant, but bdrv_states will be removed in
the future anyway because we no longer need it since we have BBs.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This function removes the BlockDriverState associated with the given
BlockBackend from that BB and sets the BDS pointer in the BB to NULL.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
KVM_PPC_ALLOCATE_HTAB ioctl can return -ENOMEM for KVM guests and QEMU
never handled this correctly. But this didn't cause any problems till
now as KVM_PPC_ALLOCATE_HTAB ioctl returned with smaller than requested
HTAB when enough contiguous memory wasn't available in the host.
After the proposed kernel change: https://patchwork.ozlabs.org/patch/530501/,
KVM_PPC_ALLOCATE_HTAB ioctl will not fallback to lower sized HTAB
allocation and will fail if requested HTAB size can't be met.
Check for such failures in QEMU and abort appropriately. This will
prevent guest kernel from hanging/freezing during early boot by doing
graceful exit when host is unable to allocate requested HTAB.
Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The !CONFIG_KVM implementation of kvmppc_reset_htab() returns -1
by default. Change this to return 0 so that we fall back to user space
HTAB allocation for emulated guests.
This fixes the make check failures for ppc64 emulated target.
Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The changes are:
1. supports recent binutils;
2. 64bit BARs behind PCI bridges supported;
3. Many fixes for USB keyboard support - keys, XHCI;
4. virtio-vga support.
This image was built with:
gcc version 4.8.3 20140911 (Red Hat 4.8.3-7) (GCC)
GNU ld version 2.23.2
The full changelog is:
> version: update to 20151103
> documentation: Add a clause about signing off
> qemu/js2x/client: Support binutils >= 2.25.1
> Fix special keys on USB
> Fix function keys on USB
> pci-scan: program 64-bit mem bar range in pci-bridge bar
> Allow to build SLOF on Little Endian host
> usb-xhci: add keyboard support
> usb-xhci: ready the link trb early
> usb-xhci: scan usb high speed ports
> usb-xhci: bulk improve event handling loop
> usb-xhci: return on allocation failure
> usb-xhci: add delay in shutdown path
> usb-xhci: event trbs does not need link trb
> usb-hid: refactor usb key reading
> takeover: Fix header includes
> board-js2x: Add missing file dma-function.fs
> vga: Add support for virtio-vga
> qemu-vga: Use MMIO BAR instead of legacy IO ports
> slof: Change call_c() function to a proper assembler function
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Commit aa4bb58752 (ppc: Add mmu_model defines for arch 2.03 and 2.07)
removed the mmu_model definition POWERPC_MMU_2_06a which is needed by
PR KVM. Reintroduce it and also add POWERPC_MMU_2_07a.
This fixes QEMU crash (qemu: fatal: Unknown MMU model) during booting
of PR KVM guest.
Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
- Make Windows happy with vfio-pci devices exposed on conventional
PCI buses on q35 by hiding PCIe capability (Alex Williamson)
- Convert to g_new() where appropriate (Markus Armbruster)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQIcBAABAgAGBQJWQkmhAAoJECObm247sIsizL0P+wSaFCeng85KNmzp6LWoqg8P
/qqwokR32HdJ5NeLXQd1k6inEvI8Toz0K4ZUgOehscGAh8fJTj2MMhdk1B3vueCF
g/gnW2EKxDfbJSqEG1+erOrIAyn1SPjZH1hm0jCcpTvXtjM81mIpiRRNaQ0XWLSp
YdRvf8p6q/cjF4f5JAUVPA8ZIl1r1ZN3FbCMJt2xYKK4cZbhNAs4M1khfrpGXzHX
1RJkZx1x+phE604cplSDuVGHtacfHefX4D2qkYIXgeUQ+8stR2xx7dm5S30Q16Y+
u91TMmwcTo5kzHmxXmGur6UcO9dJax6hk2ENudr7qZAIwJPjyH2qBOznCwa1y9+q
eK3gxrVsGyno0+t10XL8nauXP1CUYHm9M1d6uhii/o5eBtWHIofYQVtjotmqMKHQ
kwDK8v06R+jlEIxhjMhn1qBf8+q50C1Z/7s5qTqRPNFZ9jurQA3OtvdC+3psIzHj
ra70M4skvRKBEE82f0AoMT2ehEaYMEtMYlCpzOSo5OME+JmJIbCxvLzWO8XQs0im
RYRc+ZszeGaaP2j+IqWzSdz+sCIlboQaqtqNW2REGB/ADHbBO2dD4OEQ5ZiA8Ze9
ZtXXEMW/BcwP1wkldj007YBTHHyU6eAmRNEBMp1A58BIlDyM2O8tHg0ICLLH6SCS
MgaXsfeSVnwNt/bA86XV
=qJ77
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/awilliam/tags/vfio-update-20151110.0' into staging
VFIO updates 2015-11-10
- Make Windows happy with vfio-pci devices exposed on conventional
PCI buses on q35 by hiding PCIe capability (Alex Williamson)
- Convert to g_new() where appropriate (Markus Armbruster)
# gpg: Signature made Tue 10 Nov 2015 19:46:41 GMT using RSA key ID 3BB08B22
# gpg: Good signature from "Alex Williamson <alex.williamson@redhat.com>"
# gpg: aka "Alex Williamson <alex@shazbot.org>"
# gpg: aka "Alex Williamson <alwillia@redhat.com>"
# gpg: aka "Alex Williamson <alex.l.williamson@gmail.com>"
* remotes/awilliam/tags/vfio-update-20151110.0:
vfio: Use g_new() & friends where that makes obvious sense
vfio/pci: Hide device PCIe capability on non-express buses for PCIe VMs
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
g_new(T, n) is neater than g_malloc(sizeof(T) * n). It's also safer,
for two reasons. One, it catches multiplication overflowing size_t.
Two, it returns T * rather than void *, which lets the compiler catch
more type errors.
This commit only touches allocations with size arguments of the form
sizeof(T). Same Coccinelle semantic patch as in commit b45c03f.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
When we have a PCIe VM, such as Q35, guests start to care more about
valid configurations of devices relative to the VM view of the PCI
topology. Windows will error with a Code 10 for an assigned device if
a PCIe capability is found for a device on a conventional bus. We
also have the possibility of IOMMUs, like VT-d, where the where the
guest may be acutely aware of valid express capabilities on physical
hardware.
Some devices, like tg3 are adversely affected by this due to driver
dependencies on the PCIe capability. The only solution for such
devices is to attach them to an express capable bus in the VM.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=tlQv
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/juanquintela/tags/migration/20151110' into staging
migration/next for 20151110
# gpg: Signature made Tue 10 Nov 2015 14:23:26 GMT using RSA key ID 5872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>"
# gpg: aka "Juan Quintela <quintela@trasno.org>"
* remotes/juanquintela/tags/migration/20151110: (57 commits)
migration: qemu_savevm_state_cleanup becomes mandatory operation
Inhibit ballooning during postcopy
Disable mlock around incoming postcopy
End of migration for postcopy
Postcopy: Mark nohugepage before discard
postcopy: Wire up loadvm_postcopy_handle_ commands
Start up a postcopy/listener thread ready for incoming page data
Postcopy; Handle userfault requests
Round up RAMBlock sizes to host page sizes
Host page!=target page: Cleanup bitmaps
Don't iterate on precopy-only devices during postcopy
Don't sync dirty bitmaps in postcopy
postcopy: Check order of received target pages
Postcopy: Use helpers to map pages during migration
postcopy_ram.c: place_page and helpers
Page request: Consume pages off the post-copy queue
Page request: Process incoming page request
Page request: Add MIG_RP_MSG_REQ_PAGES reverse command
Postcopy: End of iteration
Postcopy: Postcopy startup in migration thread
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
since commit
commit 94f5a43704
Author: Liang Li <liang.z.li@intel.com>
Date: Mon Nov 2 15:37:00 2015 +0800
migration: defer migration_end & blk_mig_cleanup
when actual .cleanup callbacks calling was removed from complete operations.
The patch fixes regression introduced by the commit above results in
100% reliable assert for virtio-scsi VM with iothreads enabled during
'virsh create-snapshot' operation:
assert(i != mr->ioeventfd_nb);
memory_region_del_eventfd
virtio_pci_set_host_notifier_internal
virtio_pci_set_host_notifier
virtio_scsi_dataplane_start
virtio_scsi_handle_cmd
virtio_queue_notify_vq
virtio_queue_host_notifier_read
aio_dispatch
Signed-off-by: Denis V. Lunev <den@openvz.org>
Reviewed-by: Liang Li <liang.z.li@intel.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
CC: Paolo Bonzini <pbonzini@redhat.com>
CC: Juan Quintela <quintela@redhat.com>
CC: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Postcopy detects accesses to pages that haven't been transferred yet
using userfaultfd, and it causes exceptions on pages that are 'not
present'.
Ballooning also causes pages to be marked as 'not present' when the
guest inflates the balloon.
Potentially a balloon could be inflated to discard pages that are
currently inflight during postcopy and that may be arriving at about
the same time.
To avoid this confusion, disable ballooning during postcopy.
When disabled we drop balloon requests from the guest. Since ballooning
is generally initiated by the host, the management system should avoid
initiating any balloon instructions to the guest during migration,
although it's not possible to know how long it would take a guest to
process a request made prior to the start of migration.
Guest initiated ballooning will not know if it's really freed a page
of host memory or not.
Queueing the requests until after migration would be nice, but is
non-trivial, since the set of inflate/deflate requests have to
be compared with the state of the page to know what the final
outcome is allowed to be.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Userfault doesn't work with mlock; mlock is designed to nail down pages
so they don't move, userfault is designed to tell you when they're not
there.
munlock the pages we userfault protect before postcopy.
mlock everything again at the end if mlock is enabled.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Amit Shah <amit.shah@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Tweak the end of migration cleanup; we don't want to close stuff down
at the end of the main stream, since the postcopy is still sending pages
on the other thread.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Prior to servicing userfault requests we must ensure we've not got
huge pages in the area that might include non-transferred memory,
since a hugepage could incorrectly mark the whole huge page as present.
We mark the area as non-huge page (nhp) just before we perform
discards; the discard code now tells us to discard any areas
that haven't been sent (as well as any that are redirtied);
any already formed transparent-huge-pages get fragmented
by this discard process if they cotnain any discards.
Transparent huge pages that have been entirely transferred
and don't contain any discards are not broken by this mechanism;
they stay as huge pages.
By starting postcopy after a full precopy pass, many of the pages
then stay as huge pages; this is important for maintaining performance
after the end of the migration.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Wire up more of the handlers for the commands on the destination side,
in particular loadvm_postcopy_handle_run now has enough to start the
guest running.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
The loading of a device state (during postcopy) may access guest
memory that's still on the source machine and thus might need
a page fill; split off a separate thread that handles the incoming
page data so that the original incoming migration code can finish
off the device data.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
userfaultfd is a Linux syscall that gives an fd that receives a stream
of notifications of accesses to pages registered with it and allows
the program to acknowledge those stalls and tell the accessing
thread to carry on.
We convert the requests from the kernel into messages back to the
source asking for the pages.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
RAMBlocks that are not a multiple of host pages in length
cause problems for postcopy (I've seen an ACPI table on aarch64
be 5k in length - i.e. 5x target-page), so round RAMBlock sizes
up to a host-page.
This potentially breaks migration compatibility due to changes
in RAMBlock sizes; however:
1) x86 and s390 I think always have host=target page size
2) When I've tried on Power the block sizes already seem aligned.
3) I don't think there's anything else that maintains per-version
machine-types for compatibility.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Prior to the start of postcopy, ensure that everything that will
be transferred later is a whole host-page in size.
This is accomplished by discarding partially transferred host pages
and marking any that are partially dirty as fully dirty.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
During the postcopy phase we must not call the iterate method on
precopy-only devices, since they may have done some cleanup during
the _complete call at the end of the precopy phase.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Once we're in postcopy the source processors are stopped and memory
shouldn't change any more, so there's no need to look at the dirty
map.
There are two notes to this:
1) If we do resync and a page had changed then the page would get
sent again, which the destination wouldn't allow (since it might
have also modified the page)
2) Before disabling this I'd seen very rare cases where a page had been
marked dirtied although the memory contents are apparently identical
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Ensure that target pages received within a host page are in order.
This shouldn't trigger, but in the cases where the sender goes
wrong and sends stuff out of order it produces a corruption that's
really nasty to debug.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
In postcopy, the destination guest is running at the same time
as it's receiving pages; as we receive new pages we must put
them into the guests address space atomically to avoid a running
CPU accessing a partially written page.
Use the helpers in postcopy-ram.c to map these pages.
qemu_get_buffer_in_place is used to avoid a copy out of qemu_file
in the case that postcopy is going to do a copy anyway.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
