Commit Graph

648 Commits

Author SHA1 Message Date
Paolo Bonzini
db10ca9057 piolist: add owner argument to initialization functions and pass devices
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:47 +02:00
Paolo Bonzini
3cd2cf4375 pam: pass device to init_pam and use it to set owner
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:47 +02:00
Paolo Bonzini
84af6d9f97 spapr_iommu: pass device to spapr_tce_new_table and use it to set owner
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:47 +02:00
Paolo Bonzini
1b5ec23467 memory: return MemoryRegion from qemu_ram_addr_from_host
It will be needed in the next patch.

Reviewed-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:46 +02:00
Paolo Bonzini
7443b43758 exec: move qemu_ram_addr_from_host_nofail to cputlb.c
After the next patch it would not be used elsewhere anyway.  Also,
the _nofail and the standard versions of this function return different
things, which is confusing.  Removing the function from the public headers
limits the confusion.

Reviewed-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:45 +02:00
Paolo Bonzini
dfde4e6e1a memory: add ref/unref calls
Add ref/unref calls at the following places:

- places where memory regions are stashed by a listener and
  used outside the BQL (including in Xen or KVM).

- memory_region_find callsites

- creation of aliases and containers (only the aliased/contained
  region gets a reference to avoid loops)

- around calls to del_subregion/add_subregion, where the region
  could disappear after the first call

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:45 +02:00
Paolo Bonzini
3ce10901ca memory: introduce memory_region_present
This new API will avoid having too many memory_region_ref/unref
in paths that currently use memory_region_find.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:45 +02:00
Paolo Bonzini
46637be269 memory: add ref/unref
Reviewed-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:45 +02:00
Paolo Bonzini
803c0816a7 memory: add getter for owner
Whenever memory regions are accessed outside the BQL, they need to be
preserved against hot-unplug.  MemoryRegions actually do not have their
own reference count; they piggyback on a QOM object, their "owner".
The owner is set at creation time, and there is a function to retrieve
the owner.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:45 +02:00
Paolo Bonzini
2c9b15cab1 memory: add owner argument to initialization functions
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:44 +02:00
Jan Kiszka
5767e4e198 ioport: Move portio types to ioport.h
This decouples memory.h from ioport.h, concentrating all portio related
types in a single header.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:44 +02:00
Jan Kiszka
d67f679d99 vmport: Disentangle read handler type from portio
In case the latter may vanish one day, make sure the vmport read handler
type will remain unaffected. This is also conceptually cleaner.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:44 +02:00
Jan Kiszka
0659097de2 ioport: Remove unused old dispatching services
Remove unused ioport_register and isa_unassign_ioport along with
everything that only those services used.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:44 +02:00
Jan Kiszka
b40acf99be ioport: Switch dispatching to memory core layer
The current ioport dispatcher is a complex beast, mostly due to the
need to deal with old portio interface users. But we can overcome it
without converting all portio users by embedding the required base
address of a MemoryRegionPortio access into that data structure. That
removes the need to have the additional MemoryRegionIORange structure
in the loop on every access.

To handle old portio memory ops, we simply install dispatching handlers
for portio memory regions when registering them with the memory core.
This removes the need for the old_portio field.

We can drop the additional aliasing of ioport regions and also the
special address space listener. cpu_in and cpu_out now simply call
address_space_read/write. And we can concentrate portio handling in a
single source file.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:44 +02:00
Jan Kiszka
b6f3296292 isa: implement isa_is_ioport_assigned via memory_region_find
Open-code isa_is_ioport_assigned via a memory region lookup. As all IO
ports are now directly or indirectly registered via the memory API, this
becomes possible and will finally allow us to drop the ioport tables.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:43 +02:00
Jan Kiszka
8ab9b41876 Privatize register_ioport_read/write
No more users outside of ioport.c.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:43 +02:00
Paolo Bonzini
f487b677c2 dma: keep a device alive while it has SGLists
Reviewed-by: Anthony Liguori <aliguori@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-04 17:42:30 +02:00
Paolo Bonzini
6046c62086 int128: optimize and add test cases
For add, the carry only requires checking one of the arguments.
For sub and neg, we can similarly optimize computation of the
carry.

For ge, we can just do lexicographic order.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-07-02 08:17:01 +02:00
Alexander Graf
a1014f25ef PPC: Add clock-frequency export for Mac machines
Support in fwcfg has been around for exposure of the clock-frequency
CPU property. OpenBIOS reads it, we just never exposed it.

Since Mac OS X is very picky about its clock frequency values, let's
just take a known good value and always expose that.

Reported-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:17 +02:00
Anthony Liguori
210b580b10 spapr-rtas: add CPU argument to RTAS calls
RTAS is a hypervisor provided binary blob that a guest loads and
calls into to execute certain functions.  It's similar to the
vsyscall page in Linux or the short lived VMCI paravirt interface
from VMware.

The QEMU implementation of the RTAS blob is simply a passthrough
that proxies all RTAS calls to the hypervisor via an hypercall.

While we pass a CPU argument for hypercall handling in QEMU, we
don't pass it for RTAS calls.  Since some RTAs calls require
making hypercalls (normally RTAS is implemented as guest code) we
have nasty hacks to allow that.

Add a CPU argument to RTAS call handling so we can more easily
invoke hypercalls just as guest code would.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:16 +02:00
Andreas Färber
dd49c038c3 intc/openpic_kvm: Fix QOM and build issues
Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:15 +02:00
Andreas Färber
e1766344fd intc/openpic: QOM'ify
Introduce type constant and cast macro.

Signed-off-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:14 +02:00
Scott Wood
d85937e683 kvm/openpic: in-kernel mpic support
Enables support for the in-kernel MPIC that thas been merged into the
KVM next branch.  This includes irqfd/KVM_IRQ_LINE support from Alex
Graf (along with some other improvements).

Note from Alex regarding kvm_irqchip_create():

  On x86, one would call kvm_irqchip_create() to initialize an
  in-kernel interrupt controller.  That function then goes ahead and
  initializes global capability variables as well as the default irq
  routing table.

  On ppc, we can't call kvm_irqchip_create() because we can have
  different types of interrupt controllers.  So we want to do all the
  things that function would do for us in the in-kernel device init
  handler.

Signed-off-by: Scott Wood <scottwood@freescale.com>
[agraf: squash in kvm_irqchip_commit_routes patch, fix non-kvm build,
        fix ppcemb]
Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:14 +02:00
Alexander Graf
cb925cf923 KVM: PIC: Only commit irq routing when necessary
The current logic updates KVM's view of our interrupt map every time we
change it. While this is nice and bullet proof, it slows things down
badly for me. QEMU spends about 3 seconds on every start telling KVM what
news it has on its routing maps.

Instead, let's just synchronize the whole irq routing map as a whole when
we're done constructing it. For things that change during runtime, we can
still update the routing table on demand.

Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:14 +02:00
Scott Wood
8935a442cd openpic: factor out some common defines into openpic.h
...for use by the KVM in-kernel irqchip stub.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:14 +02:00
Alexander Graf
7b7745930b KVM: Export kvm_init_irq_routing
On PPC, we can have different types of interrupt controllers, so we really
only know that we are going to use one when we created it.

Export kvm_init_irq_routing() to common code, so that we don't have to call
kvm_irqchip_create().

Signed-off-by: Alexander Graf <agraf@suse.de>
2013-07-01 01:11:14 +02:00
Alexander Graf
215e79c01c KVM: Don't assume that mpstate exists with in-kernel PIC always
On PPC, we don't support MP state. So far it's not necessary and I'm
not convinced yet that we really need to support it ever.

However, the current idle logic in QEMU assumes that an in-kernel PIC
also means we support MP state. This assumption is not true anymore.

Let's split up the two cases into two different variables. That way
PPC can expose an in-kernel PIC, while not implementing MP state.

Signed-off-by: Alexander Graf <agraf@suse.de>
CC: Jan Kiszka <jan.kiszka@siemens.com>
2013-07-01 01:11:13 +02:00
Anthony Liguori
ffeec223b5 Merge remote-tracking branch 'mjt/trivial-patches' into staging
# By Gerd Hoffmann (13) and Michael Tokarev (1)
# Via Michael Tokarev
* mjt/trivial-patches:
  doc: we use seabios, not bochs bios
  qemu-socket: don't leak opts on error
  qemu-char: report udp backend errors
  qemu-char: add -chardev mux support
  qemu-char: minor mux chardev fixes
  qemu-char: use ChardevBackendKind in CharDriver
  qemu-char: don't leak opts on error
  qemu-char: fix documentation for telnet+wait socket flags
  qemu-char: print notification to stderr
  qemu-char: use more specific error_setg_* variants
  qemu-char: check optional fields using has_*
  qemu-socket: catch monitor_get_fd failures
  qemu-socket: drop pointless allocation
  qemu-socket: zero-initialize SocketAddress

Message-id: 1372443465-22384-1-git-send-email-mjt@msgid.tls.msk.ru
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2013-06-28 15:48:35 -05:00
Gerd Hoffmann
99aec0123c qemu-char: use ChardevBackendKind in CharDriver
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
2013-06-28 22:10:33 +04:00
Anthony Liguori
8a9c98aedc Merge remote-tracking branch 'afaerber/qom-cpu' into staging
# By Andreas Färber
# Via Andreas Färber
* afaerber/qom-cpu: (24 commits)
  cpu: Turn cpu_unassigned_access() into a CPUState hook
  hwaddr: Make hwaddr type usable beyond softmmu
  cpu: Change qemu_init_vcpu() argument to CPUState
  cpus: Change qemu_dummy_start_vcpu() argument to CPUState
  cpus: Change qemu_kvm_start_vcpu() argument to CPUState
  cpus: Change cpu_handle_guest_debug() argument to CPUState
  gdbstub: Set gdb_set_stop_cpu() argument to CPUState
  kvm: Change kvm_cpu_exec() argument to CPUState
  kvm: Change kvm_handle_internal_error() argument to CPUState
  cpu: Turn cpu_dump_{state,statistics}() into CPUState hooks
  cpus: Change qemu_kvm_init_cpu_signals() argument to CPUState
  kvm: Change kvm_set_signal_mask() argument to CPUState
  cpus: Change qemu_kvm_wait_io_event() argument to CPUState
  cpus: Change cpu_thread_is_idle() argument to CPUState
  cpu: Change cpu_exit() argument to CPUState
  kvm: Change cpu_synchronize_state() argument to CPUState
  kvm: Change kvm_cpu_synchronize_state() argument to CPUState
  gdbstub: Simplify find_cpu()
  cpu: Guard cpu_{save,load}() definitions
  target-openrisc: Register VMStateDescription for OpenRISCCPU
  ...
2013-06-28 11:48:09 -05:00
Peter Lieven
3ac216270a block: change default of .has_zero_init to 0
.has_zero_init defaults to 1 for all formats and protocols.

this is a dangerous default since this means that all
new added drivers need to manually overwrite it to 0 if
they do not ensure that a device is zero initialized
after bdrv_create().

if a driver needs to explicitly set this value to
1 its easier to verify the correctness in the review process.

during review of the existing drivers it turned out
that ssh and gluster had a wrong default of 1.
both protocols support host_devices as backend
which are not by default zero initialized. this
wrong assumption will lead to possible corruption
if qemu-img convert is used to write to such a backend.

vpc and vmdk also defaulted to 1 altough they support
fixed respectively flat extends. this has to be addresses
in separate patches. both formats as well as the mentioned
ssh and gluster are turned to the default of 0 with this
patch for safety.

a similar problem with the wrong default existed for
iscsi most likely because the driver developer did
oversee the default value of 1.

Signed-off-by: Peter Lieven <pl@kamp.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2013-06-28 13:52:35 +02:00
Andreas Färber
c658b94f6e cpu: Turn cpu_unassigned_access() into a CPUState hook
Use it for all targets, but be careful not to pass invalid CPUState.
cpu_single_env can be NULL, e.g. on Xen.

Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:13 +02:00
Andreas Färber
ce927ed9e4 hwaddr: Make hwaddr type usable beyond softmmu
While not normally needed for *-user, it can safely be used there since
always based on uint64_t, to avoid ifdeffery.

To avoid accidental uses, move the guards from exec/hwaddr.h to its
inclusion sites.  No need for them in include/hw/.

Prepares for hwaddr use in qom/cpu.h.

Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:13 +02:00
Andreas Färber
c643bed99f cpu: Change qemu_init_vcpu() argument to CPUState
This allows to move the call into CPUState's realizefn.
Therefore move the stub into libqemustub.a.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:13 +02:00
Andreas Färber
64f6b3469d gdbstub: Set gdb_set_stop_cpu() argument to CPUState
Use CPUState::env_ptr for now.

Prepares for changing cpu_handle_guest_debug() argument to CPUState.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:13 +02:00
Andreas Färber
1458c36313 kvm: Change kvm_cpu_exec() argument to CPUState
It no longer uses CPUArchState.

Prepares for changing qemu_kvm_cpu_thread_fn() opaque to CPUState.

Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:12 +02:00
Andreas Färber
878096eeb2 cpu: Turn cpu_dump_{state,statistics}() into CPUState hooks
Make cpustats monitor command available unconditionally.

Prepares for changing kvm_handle_internal_error() and kvm_cpu_exec()
arguments to CPUState.

Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:12 +02:00
Andreas Färber
491d6e80ea kvm: Change kvm_set_signal_mask() argument to CPUState
CPUArchState is no longer needed.

Prepares for changing qemu_kvm_init_cpu_signals() argument to CPUState.

Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:12 +02:00
Andreas Färber
60a3e17a46 cpu: Change cpu_exit() argument to CPUState
It no longer depends on CPUArchState, so move it to qom/cpu.c.

Prepares for changing GDBState::c_cpu to CPUState.

Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:12 +02:00
Andreas Färber
cb446ecab7 kvm: Change cpu_synchronize_state() argument to CPUState
Change Monitor::mon_cpu to CPUState as well.

Reviewed-by: liguang <lig.fnst@cn.fujitsu.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:12 +02:00
Andreas Färber
dd1750d798 kvm: Change kvm_cpu_synchronize_state() argument to CPUState
It no longer relies on CPUArchState since 20d695a.

Reviewed-by: liguang <lig.fnst@cn.fujitsu.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:12 +02:00
Andreas Färber
8d0f2baebe cpu: Guard cpu_{save,load}() definitions
A few targets already managed to implement cpu_save() and cpu_load()
without defining CPU_SAVE_VERSION that causes them to be registered.

Guard the prototypes with CPU_SAVE_VERSION to avoid this happening again
until all targets are converted to VMState (or QIDL).

Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:12 +02:00
Andreas Färber
1a1562f5ea cpu: Introduce VMSTATE_CPU() macro for CPUState
To be used to embed common CPU state into CPU subclasses.

Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:11 +02:00
Andreas Färber
19e3835cff cpu: Introduce device_class_set_vmsd() helper
It's the equivalent to cpu_class_set_vmsd(), to assign
DeviceClass::vmsd. It wasn't needed before since only static,
unmigratable VMStateDescriptions were assigned so far.

Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:11 +02:00
Andreas Färber
76d5f0294a cpu: Fix cpu_class_set_vmsd() documentation
It's CPUClass::vmsd, not CPUState::vmsd.

Reviewed-by: liguang <lig.fnst@cn.fujitsu.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
2013-06-28 13:25:11 +02:00
Dietmar Maurer
98d2c6f2cd block: add basic backup support to block driver
backup_start() creates a block job that copies a point-in-time snapshot
of a block device to a target block device.

We call backup_do_cow() for each write during backup. That function
reads the original data from the block device before it gets
overwritten.  The data is then written to the target device.

Currently backup cluster size is hardcoded to 65536 bytes.

[I made a number of changes to Dietmar's original patch and folded them
in to make code review easy.  Here is the full list:

 * Drop BackupDumpFunc interface in favor of a target block device
 * Detect zero clusters with buffer_is_zero() and use bdrv_co_write_zeroes()
 * Use 0 delay instead of 1us, like other block jobs
 * Unify creation/start functions into backup_start()
 * Simplify cleanup, free bitmap in backup_run() instead of cb
 * function
 * Use HBitmap to avoid duplicating bitmap code
 * Use bdrv_getlength() instead of accessing ->total_sectors
 * directly
 * Delete the backup.h header file, it is no longer necessary
 * Move ./backup.c to block/backup.c
 * Remove #ifdefed out code
 * Coding style and whitespace cleanups
 * Use bdrv_add_before_write_notifier() instead of blockjob-specific hooks
 * Keep our own in-flight CowRequest list instead of using block.c
   tracked requests.  This means a little code duplication but is much
   simpler than trying to share the tracked requests list and use the
   backup block size.
 * Add on_source_error and on_target_error error handling.
 * Use trace events instead of DPRINTF()

-- stefanha]

Signed-off-by: Dietmar Maurer <dietmar@proxmox.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2013-06-28 09:20:26 +02:00
Stefan Hajnoczi
d616b22474 block: add bdrv_add_before_write_notifier()
The bdrv_add_before_write_notifier() function installs a callback that
is invoked before a write request is processed.  This will be used to
implement copy-on-write point-in-time snapshots where we need to copy
out old data before overwriting it.

Note that BdrvTrackedRequest is moved to block_int.h since it is passed
to .notify() functions.

Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2013-06-28 09:20:26 +02:00
Stefan Hajnoczi
5dae8e5fb8 notify: add NotiferWithReturn so notifier list can abort
notifier_list_notify() has no return value.  This is fine when we just
want to invoke side-effects.

Sometimes it's useful for notifiers to produce a return value.  This
allows notifiers to "veto" an operation and will be used by the block
layer before-write notifier.

Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2013-06-28 09:20:26 +02:00
Peter Maydell
ec3f8c9913 linux-user: Fix compilation failure
Fix compilation failures for linux-user targets following recent
migration related commits bd2fa51fcd and 43487c67.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1372362818-4740-1-git-send-email-peter.maydell@linaro.org
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2013-06-27 15:38:35 -05:00
Michael R. Hines
60d9222c8f rdma: introduce capability x-rdma-pin-all
This capability allows you to disable dynamic chunk registration
for better throughput on high-performance links.

For example, using an 8GB RAM virtual machine with all 8GB of memory in
active use and the VM itself is completely idle using a 40 gbps infiniband link:

1. x-rdma-pin-all disabled total time: approximately 7.5 seconds @ 9.5 Gbps
2. x-rdma-pin-all enabled total time: approximately 4 seconds @ 26 Gbps

These numbers would of course scale up to whatever size virtual machine
you have to migrate using RDMA.

Enabling this feature does *not* have any measurable affect on
migration *downtime*. This is because, without this feature, all of the
memory will have already been registered already in advance during
the bulk round and does not need to be re-registered during the successive
iteration rounds.

Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Chegu Vinod <chegu_vinod@hp.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Tested-by: Chegu Vinod <chegu_vinod@hp.com>
Tested-by: Michael R. Hines <mrhines@us.ibm.com>
Signed-off-by: Michael R. Hines <mrhines@us.ibm.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
2013-06-27 02:38:36 +02:00