Right now, QEMU invokes aio_bh_poll before the "poll" phase
of aio_poll.  It is simpler to do it afterwards and skip the
"poll" phase altogether when the OS-dependent parts of AioContext
are invoked from GSource.  This way, AioContext behaves more
similarly when used as a GSource vs. when used as stand-alone.

As a start, take bottom halves into account when computing the
poll timeout.  If a bottom half is ready, do a non-blocking
poll.  As a side effect, this makes idle bottom halves work
with aio_poll; an improvement, but not really an important
one since they are deprecated.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

Name the 'granularity' parameter and give its expected value range.
Previously the device name was mistakenly reported as the parameter
name.

Note that the error class is unchanged from ERROR_CLASS_GENERIC_ERROR.
Reported-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>

block_job_sleep_ns is the only user. Since we are moving towards
AioContext aware code, it's better to use the explicit version and drop
the old one.
Signed-off-by: NFam Zheng <famz@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

This patch adds single read pattern to quorum driver and quorum vote is default
pattern.

For now we do a quorum vote on all the reads, it is designed for unreliable
underlying storage such as non-redundant NFS to make sure data integrity at the
cost of the read performance.

For some use cases as following:

        VM
  --------------
  |            |
  v            v
  A            B

Both A and B has hardware raid storage to justify the data integrity on its own.
So it would help performance if we do a single read instead of on all the nodes.
Further, if we run VM on either of the storage node, we can make a local read
request for better performance.

This patch generalize the above 2 nodes case in the N nodes. That is,

vm -> write to all the N nodes, read just one of them. If single read fails, we
try to read next node in FIFO order specified by the startup command.

The 2 nodes case is very similar to DRBD[1] though lack of auto-sync
functionality in the single device/node failure for now. But compared with DRBD
we still have some advantages over it:

- Suppose we have 20 VMs running on one(assume A) of 2 nodes' DRBD backed
storage. And if A crashes, we need to restart all the VMs on node B. But for
practice case, we can't because B might not have enough resources to setup 20 VMs
at once. So if we run our 20 VMs with quorum driver, and scatter the replicated
images over the data center, we can very likely restart 20 VMs without any
resource problem.

After all, I think we can build a more powerful replicated image functionality
on quorum and block jobs(block mirror) to meet various High Availibility needs.

E.g, Enable single read pattern on 2 children,

-drive driver=quorum,children.0.file.filename=0.qcow2,\
children.1.file.filename=1.qcow2,read-pattern=fifo,vote-threshold=1

[1] http://en.wikipedia.org/wiki/Distributed_Replicated_Block_Device

[Dropped \n from an error_setg() error message
--Stefan]

Cc: Benoit Canet <benoit@irqsave.net>
Cc: Eric Blake <eblake@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NLiu Yuan <namei.unix@gmail.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

Cc: Eric Blake <eblake@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>
Signed-off-by: NLiu Yuan <namei.unix@gmail.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

Recently, sheepdog revived its VDI locking functionality. This patch
updates sheepdog driver of QEMU for this feature. It changes an error
code for a case of failed locking. -EBUSY is a suitable one.
Reported-by: NValerio Pachera <sirio81@gmail.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Liu Yuan <namei.unix@gmail.com>
Cc: MORITA Kazutaka <morita.kazutaka@gmail.com>
Signed-off-by: NHitoshi Mitake <mitake.hitoshi@lab.ntt.co.jp>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

The update is required for supporting iSCSI multipath. It doesn't
affect behavior of QEMU driver but adding a new field to vdi request
struct is required.

Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Liu Yuan <namei.unix@gmail.com>
Cc: MORITA Kazutaka <morita.kazutaka@gmail.com>
Signed-off-by: NHitoshi Mitake <mitake.hitoshi@lab.ntt.co.jp>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

The out label has the qemu_progress_end() and other cleanup calls.
Always goto out in error paths so the cleanup happens.  These error
paths now return 1 instead of -1.

Note that bdrv_unref(NULL) is safe.  We just need to initialize bs to
NULL at the top of the function.

We can now remove the obsolete bs_old_backing = NULL and bs_new_backing
= NULL for safe mode.  Originally it was necessary in commit 3e85c6fd
("qemu-img rebase") but became useless in commit c2abccec ("qemu-img:
avoid calling exit(1) to release resources properly") because the
variables are already initialized during declaration.
Reported-by: NJohn Snow <jsnow@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>

If img_compare() fails to parse the cache flags the goto out3 code path
will call qemu_progress_end().  Make sure we actually call
qemu_progress_init() first.
Reported-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>

The img_commit() return value is a process exit code.  Use 1 for failure
instead of -1.  The other failure paths in this function already use 1.
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NMax Reitz <mreitz@redhat.com>

The curl hardcoded timeout (5 seconds) sometimes is not long
enough depending on the remote server configuration and network
traffic. The user should be able to set how much long he is
willing to wait for the connection.

Adding a new option to set this timeout gives the user this
flexibility. The previous default timeout of 5 seconds will be
used if this option is not present.
Reviewed-by: NFam Zheng <famz@redhat.com>
Signed-off-by: NDaniel Henrique Barboza <danielhb@linux.vnet.ibm.com>
Reviewed-by: NBenoit Canet <benoit.canet@nodalink.com>
Tested-by: NRichard W.M. Jones <rjones@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

We identify devices by their Open Firmware device paths.  The encoding
of bus numbers is incorrect: idebus_get_fw_dev_path() formats them in
decimal, while SeaBIOS uses hexadecimal.  With bus number > 9, SeaBIOS
will miss the bootindex (lucky case), or apply it to another device
(unlucky case).

Bug can't bite right now: ich9-ahci has six ports, and the sysbus-ahci
created by Calxeda Highbank has just one.

Fix it anyway, by changing %d to %x.

I couldn't find an Open Firmware spec covering this.  For what it's
worth, OVMF agrees with SeaBIOS.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

SCSI patches include bug fixes from Fam and Peter, improved error
reporting from Fam and a fix for DPRINTF bitrot.  Memory patches try
again to initialize name from the QOM name.

# gpg: Signature made Thu 28 Aug 2014 15:10:31 BST using RSA key ID 9B4D86F2
# gpg: Good signature from "Paolo Bonzini <pbonzini@redhat.com>"
# gpg:                 aka "Paolo Bonzini <bonzini@gnu.org>"

* remotes/bonzini/tags/for-upstream:
  memory: Lazy init name from QOM name as needed
  xen: hvm: Abstract away memory region name ref
  xen-hvm: Constify string
  virtio-scsi: Report error if num_queues is 0 or too large
  scsi-generic: remove superfluous DPRINTF avoid to break compiling
  block/iscsi: fix memory corruption on iscsi resize
  scsi-bus: Convert DeviceClass init to realize
  block: Pass errp in blkconf_geometry
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Mostly bugfixes + Alexey's interface-based implementation
of the NMI monitor command.

# gpg: Signature made Thu 28 Aug 2014 15:07:22 BST using RSA key ID 9B4D86F2
# gpg: Good signature from "Paolo Bonzini <pbonzini@redhat.com>"
# gpg:                 aka "Paolo Bonzini <bonzini@gnu.org>"

* remotes/kvm/tags/for-upstream:
  mc146818rtc: reinitialize irq_reinject_on_ack_count on reset
  target-i386: Add "tsc_adjust" CPU feature name
  target-i386: Add "mpx" CPU feature name
  vl: process -object after other backend options
  checkpatch.pl: adjust typedef definition to QEMU coding style
  x86: Clear MTRRs on vCPU reset
  x86: kvm: Add MTRR support for kvm_get|put_msrs()
  x86: Use common variable range MTRR counts
  target-i386: Don't forbid NX bit on PAE PDEs and PTEs
  spapr: Add support for new NMI interface
  s390x: Migrate to new NMI interface
  s390x: Convert QEMUMachine to MachineClass
  cpus: Define callback for QEMU "nmi" command
  kvm: run cpu state synchronization on target vcpu thread
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

To support name retrieval of MemoryRegions that were created
dynamically (that is, not via memory_region_init and friends). We
cache the name in MemoryRegion's state as
object_get_canonical_path_component mallocs the returned value
so it's not suitable for direct return to callers. Memory already
frees the name field, so this will be garbage collected along with
the MR object.
Signed-off-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The mr->name field is removed. This slipped through compile testing.
Fix.
Reviewed-by: NStefan Weil <sw@weilnetz.de>
Signed-off-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

It's constant, and sourced from existing const strings. Avoid dodgy
casts by converting to const.
Signed-off-by: NPeter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: NStefan Weil <sw@weilnetz.de>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Merge remote-tracking branch 'remotes/stefanha/tags/fix-buildbot-12082014-pull-request' into staging

Pull request

# gpg: Signature made Thu 28 Aug 2014 13:43:00 BST using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"

* remotes/stefanha/tags/fix-buildbot-12082014-pull-request:
  Revert "qemu-img: sort block formats in help message"
  block: sort formats alphabetically in bdrv_iterate_format()
  mirror: fix uninitialized variable delay_ns warnings
  trace: avoid Python 2.5 all() in tracetool
  libqtest: launch QEMU with QEMU_AUDIO_DRV=none
  qapi.py: avoid Python 2.5+ any() function
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

This reverts commit 1a443c1b and the
later commit 395071a7.

GSequence was introduced in glib 2.14.  RHEL 5 fails to compile since it
uses glib 2.12.3.

Now that bdrv_iterate_format() invokes the iteration callback in sorted
order these commits are unnecessary.
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>

Format names are best consumed in alphabetical order.  This makes
human-readable output easy to produce.

bdrv_iterate_format() already has an array of format strings.  Sort them
before invoking the iteration callback.
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>

The gcc 4.1.2 compiler warns that delay_ns may be uninitialized in
mirror_iteration().

There are two break statements in the do ... while loop that skip over
the delay_ns assignment.  These are probably the cause of the warning.
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>

Red Hat Enterprise Linux 5 ships Python 2.4.3.  The all() function was
added in Python 2.5 so we cannot use it.
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>

No test case actually uses the audio backend.  Disable audio to prevent
warnings on hosts with no sound hardware present:

  GTESTER check-qtest-aarch64
  sdl: SDL_OpenAudio failed
  sdl: Reason: No available audio device
  sdl: SDL_OpenAudio failed
  sdl: Reason: No available audio device
  audio: Failed to create voice `lm4549.out'
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NGerd Hoffmann <kraxel@redhat.com>

There is one instance of any() in qapi.py that breaks builds on older
distros that ship Python 2.4 (like RHEL5):

  GEN   qmp-commands.h
Traceback (most recent call last):
  File "build/scripts/qapi-commands.py", line 445, in ?
    exprs = parse_schema(input_file)
  File "build/scripts/qapi.py", line 329, in parse_schema
    schema = QAPISchema(open(input_file, "r"))
  File "build/scripts/qapi.py", line 110, in __init__
    if any(include_path == elem[1]
NameError: global name 'any' is not defined
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: NBenoît Canet <benoit.canet@nodalink.com>

This field was forgotten, and it makes the state after reset
non-deterministic.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

* remotes/mcayland/qemu-openbios:
  Update OpenBIOS images
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

tsc_adjust migration support is already implemented (commit
f28558d3), so we can add it to the list
of known feature names.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Migration support for MPX is already implemented (commit
79e9ebeb), so we can add it to the list
of known feature names.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Update OpenBIOS images to SVN r1316 built from submodule.
Signed-off-by: NMark Cave-Ayland <mark.cave-ayland@ilande.co.uk>

QOM backends can refer to chardevs, but not vice versa.  So
process -chardev and -fsdev options before -object

This fixes the rng-egd backend to virtio-rng.
Reported-by: NAmos Kong <akong@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Most QEMU typedefs are camelcase, starting with one uppercase letter
and containing at least one lowercase letter.  There are a few
all-uppercase types, add the most common too.

This fixes recognition of types in lines such as

    static __attribute__((unused)) inline void tcg_out8(TCGContext *s, uint8_t v)

(Example provided by Peter Maydell).
Reported-by: NAlexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: NMarkus Armbruster <armbru@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Stefan Weil <sw@weilnetz.de>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

No cmd vq surprises guest (Linux panics in virtscsi_probe), too many
queues abort qemu (in the following virtio_add_queue).
Signed-off-by: NFam Zheng <famz@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

variables lun and tag had been eliminated, break compiling
when enable debug switch. Meanwhile traces provide the same
information with this DPRINTF, so remove it.
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

bs->total_sectors is not yet updated at this point. resulting
in memory corruption if the volume has grown and data is written
to the newly availble areas.

CC: qemu-stable@nongnu.org
Signed-off-by: NPeter Lieven <pl@kamp.de>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Replace "init/destroy" with "realize/unrealize" in SCSIDeviceClass,
which has errp as a parameter. So all the implementations now use
error_setg instead of error_report for reporting error.

Also in scsi_bus_legacy_handle_cmdline, report the error when
initializing the if=scsi devices, before returning it, because in the
callee, error_report is changed to error_setg. And the callers don't
have the right locations (e.g. "-drive if=scsi").
Reviewed-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NFam Zheng <famz@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This allows us to pass error information to caller.
Reviewed-by: NAndreas Färber <afaerber@suse.de>
Reviewed-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NFam Zheng <famz@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

VFIO: Enable primary NVIDIA quirk regardless of VGA support

# gpg: Signature made Mon 25 Aug 2014 20:29:37 BST using RSA key ID 3BB08B22
# gpg: Can't check signature: public key not found

* remotes/awilliam/tags/vfio-pci-for-qemu-20140825.0:
  vfio: Enable NVIDIA 88000 region quirk regardless of VGA
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

If we make use of OVMF for the BIOS then we can use GPUs without VGA
space access, but we still need this quirk.  Disassociate it from the
x-vga option and enable it on all NVIDIA VGA display class devices.
Signed-off-by: NAlex Williamson <alex.williamson@redhat.com>

pci, pc fixes, features

A bunch of bugfixes - these will make sense for 2.1.1

ACPI support for TPM and partial ARI support for PCIE.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

# gpg: Signature made Sun 24 Aug 2014 23:16:35 BST using RSA key ID D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>"

* remotes/mst/tags/for_upstream:
  pcie: fix trailing whitespace
  ioh3420: Enable ARI forwarding
  ioh3420: Remove obsoleted, unused ioh3420_init function
  pcie: Rename the pcie_cap_ari_* functions to pcie_cap_arifwd_*
  pcie: Fix incorrect write to the ari capability next function field
  ssdt-tpm: add generated hex file to git
  Add ACPI tables for TPM
  pc: reserve more memory for ACPI for new machine types
  pcihp: fix possible array out of bounds
  pci_bridge: manually destroy memory regions within PCIBridgeWindows
  hostmem: set MPOL_MF_MOVE
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

The SDM specifies (June 2014 Vol3 11.11.5):

    On a hardware reset, the P6 and more recent processors clear the
    valid flags in variable-range MTRRs and clear the E flag in the
    IA32_MTRR_DEF_TYPE MSR to disable all MTRRs. All other bits in the
    MTRRs are undefined.

We currently do none of that, so whatever MTRR settings you had prior
to reset is what you have after reset.  Usually this doesn't matter
because KVM often ignores the guest mappings and uses write-back
anyway.  However, if you have an assigned device and an IOMMU that
allows NoSnoop for that device, KVM defers to the guest memory
mappings which are now stale after reset.  The result is that OVMF
rebooting on such a configuration takes a full minute to LZMA
decompress the firmware volume, a process that is nearly instant on
the initial boot.
Signed-off-by: NAlex Williamson <alex.williamson@redhat.com>
Reviewed-by: NLaszlo Ersek <lersek@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>