iotests: Make 191 work with qcow2 options

In order for 191 to work with an explicit refcount_bits or compat=0.10,
we should strip format-specific information from the output--and we can
do so by using _filter_img_info.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-17-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Make 184 image-less

184 does not need an image, so don't use one.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-16-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Make 089 compatible with compat=0.10

The only thing that is missing is a _filter_img_info after the
"$QEMU_IO -c info" invocations.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-15-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Fix 067 for compat=0.10

067 works very well with compat=0.10 once you remove format-specific
information from the QMP output.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-14-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Fix 059's reference output

As of commit 9877860e7bd1e26ee70ab9bb5ebc34c92bf23bf5, vmdk fails
differently when opening the sample image.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-13-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Fix 051 for compat=0.10

051 has both compat=1.1 and compat=0.10 tests (once it uses
lazy_refcounts, once it tests that setting them does not work).
For the compat=0.10 tests, it already explicitly creates a suitable
image.  So let's just ignore the user-specified compat level for the
lazy_refcounts test and explicitly create a compat=1.1 image there, too.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-12-mreitz@redhat.com
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Fix 020 for vmdk

vmdk cannot work with anything but vmdk backing files, so make the
backing file be the same format as the overlay.

Reported-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-11-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Skip 103 for refcount_bits=1

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-9-mreitz@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Forbid 020 for non-file protocols

This test does funny things like TEST_IMG="TEST_IMG.base" _make_test_img
that usually only work with the file protocol.  More specifically, they
do not work with the most interesting non-file protocols, so we might as
well skip this for anything but file.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-8-mreitz@redhat.com
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Drop format-specific in _filter_img_info

_filter_img_info should remove format-specific information, too.  We
already have such a filter in _img_info, and it is very useful for
query-block-named-block-nodes (etc.), too.

However, in 198 we need that information (but we still want the rest of
the filter), so make that filtering optional.  Note that "the rest of
the filter" includes filtering of the test directory, so we can drop the
_filter_testdir from 198 at the same time.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-7-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Fix _img_info for backslashes

read without -r eats backslashes.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-6-mreitz@redhat.com
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

block/vmdk: Add blkdebug events

This is certainly not complete, but it includes at least write_aio and
read_aio.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-5-mreitz@redhat.com
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

block/qcow: Add blkdebug events

This is not necessarily complete, but it should include the most
important places.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-4-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

qcow2: No persistent dirty bitmaps for compat=0.10

Persistent dirty bitmaps require a properly functioning
autoclear_features field, or we cannot track when an unsupporting
program might overwrite them.  Therefore, we cannot support them for
compat=0.10 images.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-3-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

block/vmdk: Fix , instead of ; at end of line

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171123020832.8165-2-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

qemu-iotests: Fix locking issue in 102

102 truncates a qcow2 file (the raw file) on purpose while a VM is
running.  However, image locking will usually prevent exactly this.

The fact that most people have not noticed until now (I suppose you may
have seen sporadic failures, but not taken them too seriously, like me)
further shows that this truncation is actually not really done
concurrently, but that the VM is still starting up by this point and has
not yet opened the image.  Remedy this by waiting for the monitor shell
to appear before the qemu-img invocation so we know the VM is up.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171129185102.29390-1-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

blockdev: Mark BD-{remove,insert}-medium stable

Now that iotest 093 test proves that the throttling configuration
survives a blockdev-remove-medium/blockdev-insert-medium pair, the
original reason for declaring these commands experimental is gone
(see commit 6e0abc251dd4f8eba1f53656dfede12e5840e83b).

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171110224302.14424-5-mreitz@redhat.com
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

blockdev: Drop BD-{remove,insert}-medium's @device

This is an incompatible change, which is fine as the commands are
experimental.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171110224302.14424-4-mreitz@redhat.com
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

tests/ahci: Switch tray and medium commands to @id

Currently, the tray and medium commands in the AHCI test use the
deprecated @device parameter.  This patch switches all invocations over
to use @id.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171110224302.14424-3-mreitz@redhat.com
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

iotests: Make BD-{remove,insert}-medium use @id

In some cases, these commands still use the deprecated @device
parameter.  Fix that so we can later drop that parameter from their
interface.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171110224302.14424-2-mreitz@redhat.com
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

qcow2: Repair unaligned preallocated zero clusters

We can easily repair unaligned preallocated zero clusters by discarding
them, so why not do it?

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20171110203759.14018-2-mreitz@redhat.com
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>

tests/qemu-iotests: adding savevm/loadvm with postcopy flag test

This patch implements a test case for the scenario that was failing
prior to the patch "migration/ram.c: do not set 'postcopy_running' in
POSTCOPY_INCOMING_END", commit acab30b85d.

This new test file 201 was derived from the test file 181 authored
by Kevin Wolf.

CC: Kevin Wolf <kwolf@redhat.com>
CC: Max Reitz <mreitz@redhat.com>
CC: Cleber Rosa <crosa@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

tests/.gitignore: Add test-bdrv-drain

Commit 881cfd17 added a new test binary, include it in .gitignore.

Reported-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>

hw/block: Fix pin-based interrupt behaviour of NVMe

Pin-based interrupt of NVMe controller did not work properly
because using an obsolated function pci_irq_pulse().
To fix this, change to use pci_irq_assert() / pci_irq_deassert()
instead of pci_irq_pulse().

Signed-off-by: Hikaru Nishida <hikarupsp@gmail.com>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

osdep: Retry SETLK upon EINTR

We could hit lock failure if there is a signal that makes fcntl return
-1 and errno set to EINTR. In this case we should retry.

Cc: qemu-stable@nongnu.org
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

migration: Revert postcopy-blocktime commit set

This reverts commits
ca6011c migration: add postcopy total blocktime into query-migrate
5f32dc8 migration: add blocktime calculation into migration-test
2f7dae9 migration: postcopy_blocktime documentation
3be98be migration: calculate vCPU blocktime on dst side
01a87f0 migration: add postcopy blocktime ctx into MigrationIncomingState
31bf06a migration: introduce postcopy-blocktime capability

as they don't build on ppc32 due to trying to do atomic accesses
on types that are larger than the host pointer type.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.12-20180121' into staging

ppc patch queue 2018-01-21

This request supersedes the one from 2018-01-19.  The only difference
is that the patch deprecating ppcemb-softmmu, and thereby creating
many annying warnings from make check has been removed.

Highlights are:
  * Significant TCG speedup by optimizing cmp generation
  * Fix a regression caused by recent change to set compat mode on
    hotplugged cpus
  * Cleanup of default configs
  * Some implementation of msgsnd/msgrcv instructions for server chips

# gpg: Signature made Sun 21 Jan 2018 05:30:54 GMT
# gpg:                using RSA key 0x6C38CACA20D9B392
# gpg: Good signature from "David Gibson <david@gibson.dropbear.id.au>"
# gpg:                 aka "David Gibson (Red Hat) <dgibson@redhat.com>"
# gpg:                 aka "David Gibson (ozlabs.org) <dgibson@ozlabs.org>"
# gpg:                 aka "David Gibson (kernel.org) <dwg@kernel.org>"
# Primary key fingerprint: 75F4 6586 AE61 A66C C44E  87DC 6C38 CACA 20D9 B392

* remotes/dgibson/tags/ppc-for-2.12-20180121:
  target/ppc/spapr_caps: Add macro to generate spapr_caps migration vmstate
  target/ppc: add support for hypervisor doorbells on book3s CPUs
  sii3112: Add explicit type casts to avoid unintended sign extension
  sm501: Add missing break to case
  target-ppc: optimize cmp translation
  spapr: fix device tree properties when using compatibility mode
  spapr: drop duplicate variable in spapr_core_plug()
  target/ppc: msgsnd and msgclr instructions need hypervisor privilege
  target/ppc: fix doorbell and hypervisor doorbell definitions
  hw/ppc/Makefile: Add a way to disable the PPC4xx boards
  default-configs/ppc-softmmu: Restructure the switches according to the machines
  default-configs/ppc64-softmmu: Include 32-bit configs instead of copying them

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/elmarco/tags/dump-pull-request' into staging

# gpg: Signature made Sat 20 Jan 2018 20:03:47 GMT
# gpg:                using RSA key 0xDAE8E10975969CE5
# gpg: Good signature from "Marc-André Lureau <marcandre.lureau@redhat.com>"
# gpg:                 aka "Marc-André Lureau <marcandre.lureau@gmail.com>"
# Primary key fingerprint: 87A9 BD93 3F87 C606 D276  F62D DAE8 E109 7596 9CE5

* remotes/elmarco/tags/dump-pull-request:
  dump-guest-memory.py: fix python 2 support

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/jnsnow/tags/ide-pull-request' into staging

# gpg: Signature made Fri 19 Jan 2018 22:31:27 GMT
# gpg:                using RSA key 0x7DEF8106AAFC390E
# gpg: Good signature from "John Snow (John Huston) <jsnow@redhat.com>"
# Primary key fingerprint: FAEB 9711 A12C F475 812F  18F2 88A9 064D 1835 61EB
#      Subkey fingerprint: F9B7 ABDB BCAC DF95 BE76  CBD0 7DEF 8106 AAFC 390E

* remotes/jnsnow/tags/ide-pull-request:
  hw/ide: Remove duplicated definitions from ahci_internal.h
  ide: abort TRIM operation for invalid range
  ide: move ide_sect_range_ok() up
  ide: pass IDEState to trim AIO callback

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

dump-guest-memory.py: fix python 2 support

Python GDB support may use Python 2 or 3.

Inferior.read_memory() may return a 'buffer' with Python 2 or a
'memoryview' with Python 3 (see also
https://sourceware.org/gdb/onlinedocs/gdb/Inferiors-In-Python.html)

The elf.add_vmcoreinfo_note() method expects a "bytes" object. Wrap
the returned memory with bytes(), which works with both 'memoryview'
and 'buffer'.

Fixes a regression introduced with commit
d23bfa91b7789534d16ede6cb7d925bfac3f3c4c ("add vmcoreinfo").

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>

target/ppc/spapr_caps: Add macro to generate spapr_caps migration vmstate

The vmstate description and the contained needed function for migration
of spapr_caps is the same for each cap, with the name of the cap
substituted. As such introduce a macro to allow for easier generation of
these.

Convert the three existing spapr_caps (htm, vsx, and dfp) to use this
macro.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

target/ppc: add support for hypervisor doorbells on book3s CPUs

The hypervisor doorbells are used by skiboot and Linux on POWER9
processors to wake up secondaries.

This adds processor control support to the Server architecture by
reusing the Embedded support. They are very similar, only the bits
definition of the CPU identifier differ.

Still to be done is message broadcast to all threads of the same
processor.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

sii3112: Add explicit type casts to avoid unintended sign extension

Noticed by Coverity

Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

sm501: Add missing break to case

Noticed by Coverity, forgotten in 5690d9ece

Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

target-ppc: optimize cmp translation

We know that only one bit (in addition to SO) is going to be set in
the condition register, so do two movconds instead of three setconds,
three shifts and two ORs.

For ppc64-linux-user, the code size reduction is around 5% and the
performance improvement slightly less than 10%.  For softmmu, the
improvement is around 5%.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

spapr: fix device tree properties when using compatibility mode

Commit 51f84465dd98 changed the compatility mode setting logic:
- machine reset only sets compatibility mode for the boot CPU
- compatibility mode is set for other CPUs when they are put online
  by the guest with the "start-cpu" RTAS call

This causes a regression for machines started with max-compat-cpu:
the device tree nodes related to secondary CPU cores contain wrong
"cpu-version" and "ibm,pa-features" values, as shown below.

Guest started on a POWER8 host with:
     -smp cores=2 -machine pseries,max-cpu-compat=compat7

                        ibm,pa-features = [18 00 f6 3f c7 c0 80 f0 80 00
 00 00 00 00 00 00 00 00 80 00 80 00 80 00 00 00];
                        cpu-version = <0x4d0200>;

                               ^^^
                        second CPU core

                        ibm,pa-features = <0x600f63f 0xc70080c0>;
                        cpu-version = <0xf000003>;

                               ^^^
                          boot CPU core

The second core is advertised in raw POWER8 mode. This happens because
CAS assumes all CPUs to have the same compatibility mode. Since the
boot CPU already has the requested compatibility mode, the CAS code
does not set it for the secondary one, and exposes the bogus device
tree properties in in the CAS response to the guest.

A similar situation is observed when hot-plugging a CPU core. The
related device tree properties are generated and exposed to guest
with the "ibm,configure-connector" RTAS before "start-cpu" is called.
The CPU core is advertised to the guest in raw mode as well.

It both cases, it boils down to the fact that "start-cpu" happens too
late. This can be fixed globally by propagating the compatibility mode
of the boot CPU to the other CPUs during reset.  For this to work, the
compatibility mode of the boot CPU must be set before the machine code
actually resets all CPUs.

It is not needed to set the compatibility mode in "start-cpu" anymore,
so the code is dropped.

Fixes: 51f84465dd98
Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

spapr: drop duplicate variable in spapr_core_plug()

A variable is already defined at the begining of the function to
hold a pointer to the CPU core object:

    sPAPRCPUCore *core = SPAPR_CPU_CORE(OBJECT(dev));

No need to define it again in the pre-2.10 compatibility code snipplet.

Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

target/ppc: msgsnd and msgclr instructions need hypervisor privilege

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

target/ppc: fix doorbell and hypervisor doorbell definitions

commit f03a1af581b9 ("ppc: Fix POWER7 and POWER8 exception definitions")
introduced definitions for the server doorbell exceptions by reusing
the embedded definitions but this adds complexity in the powerpc_excp()
routine. Let's introduce specific definitions for the Server doorbells
exception.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

hw/ppc/Makefile: Add a way to disable the PPC4xx boards

We've got the config switch CONFIG_PPC4XX, so we should use it
in the Makefile accordingly and only include the PPC4xx boards
if this switch has been enabled. (Note: Unfortunately, the files
ppc4xx_devs.c and ppc405_uc.c still have to be included in the
build anyway to fulfil some complicated linker dependencies ...
so these are subject to a more thourough clean-up later)

Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

default-configs/ppc-softmmu: Restructure the switches according to the machines

Order the CONFIG switches in ppc-softmmu.mak according to the machine
classes where they are used (embedded, Mac or PReP), so that it is
easier for the users to disable a set of switches completely if they
are not needed.

Also add the missing CONFIG_IDE_SII3112 switch to the embedded section
which was previously only added to ppcemb-softmmu.mak.

And while we're at it, also remove the CONFIG_IDE_CMD646 switch since
this controller does not seem to be used by any ppc machine in QEMU.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

default-configs/ppc64-softmmu: Include 32-bit configs instead of copying them

qemu-softmmu-ppc64 is supposed to be a superset of qemu-softmmu-ppc.
However, instead of simply including the 32-bit config file, we've
duplicated all CONFIG_xxx settings there instead. This way, we've missed
some CONFIG switches in ppc64-softmmu.mak which were only added to the
32-bit config file (e.g. CONFIG_SUNGEM). Let's fix this problem by
including the 32-bit config file into the 64-bit config file instead
of duplicating all the CONFIG switches there.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

hw/ide: Remove duplicated definitions from ahci_internal.h

The same definitions can also be found in include/hw/ide/ahci.h
so let's remove these #defines from ahci_internal.h.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-id: 1512457825-3847-1-git-send-email-thuth@redhat.com
[Maintainer edit: publicize object names, privatize object macros.]
Signed-off-by: John Snow <jsnow@redhat.com>

ide: abort TRIM operation for invalid range

ATA8-ACS3, 7.9 DATA SET MANAGEMENT - 06h, DMA

    7.9.5 Error Outputs
    If the Trim bit is set to one and:
      a) the device detects an invalid LBA Range Entry; or
      b) count is greater than IDENTIFY DEVICE data word 105
         (see 7.16.7.55),
    then the device shall return command aborted.
    A device may trim one or more LBA Range Entries before it returns
    command aborted. See table 209.

This check is not in the common ide_dma_cb() as the range for TRIM
is harder to reach: it is not in LBA/count registers and the buffer has
to be parsed first.

Signed-off-by: Anton Nefedov <anton.nefedov@virtuozzo.com>
Message-id: 1512735034-35327-4-git-send-email-anton.nefedov@virtuozzo.com
Signed-off-by: John Snow <jsnow@redhat.com>

ide: move ide_sect_range_ok() up

to use it without a forward declaration in the commit to follow

Signed-off-by: Anton Nefedov <anton.nefedov@virtuozzo.com>
Message-id: 1512735034-35327-3-git-send-email-anton.nefedov@virtuozzo.com
Signed-off-by: John Snow <jsnow@redhat.com>

ide: pass IDEState to trim AIO callback

It will be needed to handle invalid requests

Signed-off-by: Anton Nefedov <anton.nefedov@virtuozzo.com>
Message-id: 1512735034-35327-2-git-send-email-anton.nefedov@virtuozzo.com
Signed-off-by: John Snow <jsnow@redhat.com>

Merge remote-tracking branch 'remotes/ehabkost/tags/machine-next-pull-request' into staging

machine queue, 2018-01-19

# gpg: Signature made Fri 19 Jan 2018 16:30:19 GMT
# gpg:                using RSA key 0x2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>"
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/machine-next-pull-request:
  fw_cfg: fix memory corruption when all fw_cfg slots are used
  possible_cpus: add CPUArchId::type field
  nvdimm: add 'unarmed' option
  nvdimm: add a macro for property "label-size"
  hostmem-file: add "align" option
  scripts: Remove fixed entries from the device-crash-test
  qdev: Check for the availability of a hotplug controller before adding a device
  qdev_monitor: Simplify error handling in qdev_device_add()
  q35: Allow only supported dynamic sysbus devices
  xen: Add only xen-sysdev to dynamic sysbus device list
  spapr: Allow only supported dynamic sysbus devices
  ppc: e500: Allow only supported dynamic sysbus devices
  hw/arm/virt: Allow only supported dynamic sysbus devices
  machine: Replace has_dynamic_sysbus with list of allowed devices
  numa: fix missing '-numa cpu' in '-help' output
  qemu-options: document memory-backend-ram
  qemu-options: document missing memory-backend-file options
  memfd: remove needless include
  memfd: split qemu_memfd_alloc()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

fw_cfg: fix memory corruption when all fw_cfg slots are used

When all the fw_cfg slots are used, a write is made outside the
bounds of the fw_cfg files array as part of the sort algorithm.

Fix it by avoiding an unnecessary array element move.
Fix also an assert while at it.

Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Message-Id: <20180108215007.46471-1-marcel@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

possible_cpus: add CPUArchId::type field

Remove dependency of possible_cpus on 1st CPU instance,
which decouples configuration data from CPU instances that
are created using that data.

Also later it would be used for enabling early cpu to numa node
configuration at runtime qmp_query_hotpluggable_cpus() should
provide a list of available cpu slots at early stage,
before machine_init() is called and the 1st cpu is created,
so that mgmt might be able to call it and use output to set
numa mapping.

Use MachineClass::possible_cpu_arch_ids() callback to set
cpu type info, along with the rest of possible cpu properties,
to let machine define which cpu type* will be used.

* for SPAPR it will be a spapr core type and for ARM/s390x/x86
  a respective descendant of CPUClass.

Move parse_numa_opts() in vl.c after cpu_model is parsed into
cpu_type so that possible_cpu_arch_ids() would know which
cpu_type to use during layout initialization.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <1515597770-268979-1-git-send-email-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

nvdimm: add 'unarmed' option

Currently the only vNVDIMM backend can guarantee the guest write
persistence is device DAX on Linux, because no host-side kernel cache
is involved in the guest access to it. The approach to detect whether
the backend is device DAX needs to access sysfs, which may not work
with SELinux.

Instead, we add the 'unarmed' option to device 'nvdimm', so that users
or management utils, which have enough knowledge about the backend,
can control the unarmed flag in guest ACPI NFIT via this option. The
guest Linux NVDIMM driver, for example, will mark the corresponding
vNVDIMM device read-only if the unarmed flag in guest NFIT is set.

The default value of 'unarmed' option is 'off' in order to keep the
backwards compatibility.

Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
Message-Id: <20171211072806.2812-4-haozhong.zhang@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

nvdimm: add a macro for property "label-size"

Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20171211072806.2812-3-haozhong.zhang@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

hostmem-file: add "align" option

When mmap(2) the backend files, QEMU uses the host page size
(getpagesize(2)) by default as the alignment of mapping address.
However, some backends may require alignments different than the page
size. For example, mmap a device DAX (e.g., /dev/dax0.0) on Linux
kernel 4.13 to an address, which is 4K-aligned but not 2M-aligned,
fails with a kernel message like

[617494.969768] dax dax0.0: qemu-system-x86: dax_mmap: fail, unaligned vma (0x7fa37c579000 - 0x7fa43c579000, 0x1fffff)

Because there is no common approach to get such alignment requirement,
we add the 'align' option to 'memory-backend-file', so that users or
management utils, which have enough knowledge about the backend, can
specify a proper alignment via this option.

Signed-off-by: Haozhong Zhang <haozhong.zhang@intel.com>
Message-Id: <20171211072806.2812-2-haozhong.zhang@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
[ehabkost: fixed typo, fixed error_setg() format string]
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

scripts: Remove fixed entries from the device-crash-test

These are crashes / errors which have been fixed already in the past
months. We can remove these from the device-crash-test script now.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1513613438-11017-1-git-send-email-thuth@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

qdev: Check for the availability of a hotplug controller before adding a device

The qdev_unplug() function contains a g_assert(hotplug_ctrl) statement,
so QEMU crashes when the user tries to device_add + device_del a device
that does not have a corresponding hotplug controller. This could be
provoked for a couple of devices in the past (see commit 4c93950659487c7ad
or 84ebd3e8c7d4fe955 for example), and can currently for example also be
triggered like this:

$ s390x-softmmu/qemu-system-s390x -M none -nographic
QEMU 2.10.50 monitor - type 'help' for more information
(qemu) device_add qemu-s390x-cpu,id=x
(qemu) device_del x
**
ERROR:qemu/qdev-monitor.c:872:qdev_unplug: assertion failed: (hotplug_ctrl)
Aborted (core dumped)

So devices clearly need a hotplug controller when they should be usable
with device_add.
The code in qdev_device_add() already checks whether the bus has a proper
hotplug controller, but for devices that do not have a corresponding bus,
there is no appropriate check available yet. In that case we should check
whether the machine itself provides a suitable hotplug controller and
refuse to plug the device if none is available.

Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1509617407-21191-3-git-send-email-thuth@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

qdev_monitor: Simplify error handling in qdev_device_add()

Instead of doing the clean-ups on errors multiple times, introduce
a jump label at the end of the function that can be used by all
error paths that need this cleanup.

Suggested-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1509617407-21191-2-git-send-email-thuth@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

q35: Allow only supported dynamic sysbus devices

The only user-creatable sysbus devices in qemu-system-x86_64 are
amd-iommu, intel-iommu, and xen-backend.  xen-backend is handled
by xen_set_dynamic_sysbus(), so we only need to add amd-iommu and
intel-iommu.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Marcel Apfelbaum <marcel.a@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20171125151610.20547-7-ehabkost@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

xen: Add only xen-sysdev to dynamic sysbus device list

There's no need to make the machine allow every possible sysbus
device.  We can now just add xen-sysdev to the allowed list.

Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Anthony Perard <anthony.perard@citrix.com>
Cc: xen-devel@lists.xenproject.org
Cc: Juergen Gross <jgross@suse.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20171125151610.20547-6-ehabkost@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

spapr: Allow only supported dynamic sysbus devices

TYPE_SPAPR_PCI_HOST_BRIDGE is the only dynamic sysbus device not
rejected by ppc_spapr_reset(), so it can be the only entry on the
allowed list.

Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Alexander Graf <agraf@suse.de>
Cc: qemu-ppc@nongnu.org
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20171125151610.20547-5-ehabkost@redhat.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

ppc: e500: Allow only supported dynamic sysbus devices

platform_bus_create_devtree() already rejects all dynamic sysbus
devices except TYPE_ETSEC_COMMON, so register it as the only
allowed dynamic sysbus device for the ppce500 machine-type.

Cc: Alexander Graf <agraf@suse.de>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: qemu-ppc@nongnu.org
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20171125151610.20547-4-ehabkost@redhat.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

hw/arm/virt: Allow only supported dynamic sysbus devices

Replace the TYPE_SYS_BUS_DEVICE entry in the allowed sysbus
device list with the two device types that are really supported
by the virt machine: vfio-amd-xgbe and vfio-calxeda-xgmac.

Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: qemu-arm@nongnu.org
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20171125151610.20547-3-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

machine: Replace has_dynamic_sysbus with list of allowed devices

The existing has_dynamic_sysbus flag makes the machine accept
every user-creatable sysbus device type on the command-line.
Replace it with a list of allowed device types, so machines can
easily accept some sysbus devices while rejecting others.

To keep exactly the same behavior as before, the existing
has_dynamic_sysbus=true assignments are replaced with a
TYPE_SYS_BUS_DEVICE entry on the allowed list.  Other patches
will replace the TYPE_SYS_BUS_DEVICE entries with more specific
lists of devices.

Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Marcel Apfelbaum <marcel@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Alexander Graf <agraf@suse.de>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Stefano Stabellini <sstabellini@kernel.org>
Cc: Anthony Perard <anthony.perard@citrix.com>
Cc: qemu-arm@nongnu.org
Cc: qemu-ppc@nongnu.org
Cc: xen-devel@lists.xenproject.org
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20171125151610.20547-2-ehabkost@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

numa: fix missing '-numa cpu' in '-help' output

commit 419fcdec3c (numa: add '-numa cpu,...' option for property based node mapping)
added '-numa cpu' option but forgot to update appropriate section for '--help'.
Add '-numa cpu' description to '-help' output

Reported-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <1511880838-56509-1-git-send-email-imammedo@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

qemu-options: document memory-backend-ram

The documentation should mention -object memory-backend-ram.

Suggested-by: Yumei Huang <yuhuang@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20171128161529.3025-3-stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

qemu-options: document missing memory-backend-file options

This patch adds undocumented memory-backend-file options to the
documentation.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20171128161529.3025-2-stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

memfd: remove needless include

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20171023141815.17709-3-marcandre.lureau@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

memfd: split qemu_memfd_alloc()

Add a function to only create a memfd, without mmap. The function is
used in the following memory backend.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20171023141815.17709-2-marcandre.lureau@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>

Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

pc, pci, virtio: features, fixes, cleanups

A bunch of fixes, cleanus and new features all over the place.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Thu 18 Jan 2018 20:41:03 GMT
# gpg:                using RSA key 0x281F0DB8D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>"
# Primary key fingerprint: 0270 606B 6F3C DF3D 0B17  0970 C350 3912 AFBE 8E67
#      Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA  8A0D 281F 0DB8 D28D 5469

* remotes/mst/tags/for_upstream: (29 commits)
  vhost: remove assertion to prevent crash
  vhost-user: fix misaligned access to payload
  vhost-user: factor out msg head and payload
  tests: acpi: add comments to fetch_rsdt_referenced_tables/data->tables usage
  tests: acpi: rename test_acpi_tables()/test_dst_table() to reflect its usage
  tests: acpi: init table descriptor in test_dst_table()
  tests: acpi: move tested tables array allocation outside of test_acpi_dsdt_table()
  x86_iommu: check if machine has PCI bus
  x86_iommu: Move machine check to x86_iommu_realize()
  vhost-user-test: use init_virtio_dev in multiqueue test
  vhost-user-test: make features mask an init_virtio_dev() argument
  vhost-user-test: setup virtqueues in all tests
  vhost-user-test: extract read-guest-mem test from main loop
  vhost-user-test: fix features mask
  hw/acpi-build: Make next_base easy to follow
  ACPI/unit-test: Add a testcase for RAM allocation in numa node
  hw/pci-bridge: fix QEMU crash because of pcie-root-port
  intel-iommu: Extend address width to 48 bits
  intel-iommu: Redefine macros to enable supporting 48 bit address width
  vhost-user: fix multiple queue specification
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

vhost: remove assertion to prevent crash

QEMU will assert on vhost-user backed virtio device hotplug if QEMU is
using more RAM regions than VHOST_MEMORY_MAX_NREGIONS (for example if
it were started with a lot of DIMM devices).

Fix it by returning error instead of asserting and let callers of
vhost_set_mem_table() handle error condition gracefully.

Cc: qemu-stable@nongnu.org
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Jay Zhou <jianjay.zhou@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user: fix misaligned access to payload

We currently take a pointer to a misaligned field of a packed structure.
clang reports this as a build warning.
A fix is to keep payload in a separate structure, and access is it
from there using a vectored write.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user: factor out msg head and payload

split header and payload into separate structures,
to enable easier handling of alignment issues.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tests: acpi: add comments to fetch_rsdt_referenced_tables/data->tables usage

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tests: acpi: rename test_acpi_tables()/test_dst_table() to reflect its usage

Main purpose of test_dst_table() is loading a table from QEMU
with checking that checksum in header matches actual one,
rename it reflect main action it performs.

Likewise test_acpi_tables() name is to broad, while the function
only loads tables referenced by RSDT, rename it to reflect it.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tests: acpi: init table descriptor in test_dst_table()

remove code duplication and make sure that table descriptor
passed in for initialization is in expected state.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tests: acpi: move tested tables array allocation outside of test_acpi_dsdt_table()

at best it's confusing that array for list of tables to be tested
against reference tables is allocated within test_acpi_dsdt_table()
and at worst it would just overwrite list of tables if they were
added before test_acpi_dsdt_table().
Move array initialization to test_acpi_one() before we start
processing tables.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

x86_iommu: check if machine has PCI bus

Starting qemu with
qemu-system-x86_64 -S -M isapc -device {amd|intel}-iommu
leads to a segfault. The code assume PCI bus is present and
tries to access the bus structure without checking.

Since Intel VT-d and AMDVI should only work with PCI, add a
check for PCI bus and return error if not present.

Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Mohammed Gamal <mgamal@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>

x86_iommu: Move machine check to x86_iommu_realize()

Instead of having the same error checks in vtd_realize()
and amdvi_realize(), move that over to the generic
x86_iommu_realize().

Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Mohammed Gamal <mgamal@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>

vhost-user-test: use init_virtio_dev in multiqueue test

Now that init_virtio_dev() has been generalized to all cases,
use it in test_multiqueue() to avoid code duplication.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

vhost-user-test: make features mask an init_virtio_dev() argument

The goal is to generalize the use of [un]init_virtio_dev() to
all tests, which does not necessarily expose the same features
set.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

vhost-user-test: setup virtqueues in all tests

Only the multiqueue test setups the virtqueues.
This patch generalizes the setup of virtqueues for all tests.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

vhost-user-test: extract read-guest-mem test from main loop

This patch makes read-guest-test consistent with other tests,
i.e. create the test server in the test function.

Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user-test: fix features mask

VIRTIO_NET_F_MAC is a bit position, not a bit mask.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>

hw/acpi-build: Make next_base easy to follow

It may be hard to read the assignment statement of "next_base", so

S/next_base += (1ULL << 32) - pcms->below_4g_mem_size;
 /next_base = mem_base + mem_len;

... for readability.

No functionality change.

Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

ACPI/unit-test: Add a testcase for RAM allocation in numa node

As QEMU supports the memory-less node, it is possible that there is
no RAM in the first numa node(also be called as node0). eg:
  ... \
  -m 128,slots=3,maxmem=1G \
  -numa node -numa node,mem=128M \

But, this makes it hard for QEMU to build a known-to-work ACPI SRAT
table. Only fixing it is not enough.

Add a testcase for this situation to make sure the ACPI table is
correct for guest.

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci-bridge: fix QEMU crash because of pcie-root-port

If we try to use more pcie_root_ports then available slots
and an IO hint is passed to the port, QEMU crashes because
we try to init the "IO hint" capability even if the device
is not created.
Fix it by checking for error before adding the capability,
so QEMU can fail gracefully.

Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

intel-iommu: Extend address width to 48 bits

The current implementation of Intel IOMMU code only supports 39 bits
iova address width. This patch provides a new parameter (x-aw-bits)
for intel-iommu to extend its address width to 48 bits but keeping the
default the same (39 bits). The reason for not changing the default
is to avoid potential compatibility problems with live migration of
intel-iommu enabled QEMU guest. The only valid values for 'x-aw-bits'
parameter are 39 and 48.

After enabling larger address width (48), we should be able to map
larger iova addresses in the guest. For example, a QEMU guest that
is configured with large memory ( >=1TB ). To check whether 48 bits
aw is enabled, we can grep in the guest dmesg output with line:
"DMAR: Host address width 48".

Signed-off-by: Prasad Singamsetty <prasad.singamsety@oracle.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

intel-iommu: Redefine macros to enable supporting 48 bit address width

The current implementation of Intel IOMMU code only supports 39 bits
host/iova address width so number of macros use hard coded values based
on that. This patch is to redefine them so they can be used with
variable address widths. This patch doesn't add any new functionality
but enables adding support for 48 bit address width.

Signed-off-by: Prasad Singamsetty <prasad.singamsety@oracle.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user: fix multiple queue specification

The number of queues supported by the slave is queried with
message VHOST_USER_GET_QUEUE_NUM, not with message
VHOST_USER_GET_PROTOCOL_FEATURES.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci/shpc: Move function to generic header file

This function should be declared in generic header file so we can
utilize it.

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio: improve virtio devices initialization time

The loading time of a VM is quite significant when its virtio
devices use a large amount of virt-queues (e.g. a virtio-serial
device with max_ports=511). Most of the time is spend in the
creation of all the required event notifiers (ioeventfd and memory
regions).

This patch pack all the changes to the memory regions in a
single memory transaction.

Reported-by: Sitong Liu <siliu@redhat.com>
Reported-by: Xiaoling Gao <xiagao@redhat.com>
Signed-off-by: Gal Hammer <ghammer@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio: postpone the execution of event_notifier_cleanup function

Use the EventNotifier's cleanup callback function to execute the
event_notifier_cleanup function after kvm unregistered the eventfd.

This change supports running the virtio_bus_set_host_notifier
function inside a memory region transaction. Otherwise, a closed
fd is sent to kvm, which results in a failure.

Signed-off-by: Gal Hammer <ghammer@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

qemu: add a cleanup callback function to EventNotifier

Adding a cleanup callback function to the EventNotifier struct
which allows users to execute event_notifier_cleanup in a
different context.

Signed-off-by: Gal Hammer <ghammer@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

contrib/vhost-user-blk: introduce a vhost-user-blk sample application

This commit introduces a vhost-user-blk backend device, it uses UNIX
domain socket to communicate with QEMU. The vhost-user-blk sample
application should be used with QEMU vhost-user-blk-pci device.

To use it, complie with:
make vhost-user-blk

and start like this:
vhost-user-blk -b /dev/sdb -s /path/vhost.socket

Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

contrib/libvhost-user: enable virtio config space messages

Enable VHOST_USER_GET_CONFIG/VHOST_USER_SET_CONFIG messages in
libvhost-user library, users can implement their own I/O target
based on the library. This enable the virtio config space delivered
between QEMU host device and the I/O target.

Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user-blk: introduce a new vhost-user-blk host device

This commit introduces a new vhost-user device for block, it uses a
chardev to connect with the backend, same with Qemu virito-blk device,
Guest OS still uses the virtio-blk frontend driver.

To use it, start QEMU with command line like this:

qemu-system-x86_64 \
    -chardev socket,id=char0,path=/path/vhost.socket \
    -device vhost-user-blk-pci,chardev=char0,num-queues=2, \
            bootindex=2... \

Users can use different parameters for `num-queues` and `bootindex`.

Different with exist Qemu virtio-blk host device, it makes more easy
for users to implement their own I/O processing logic, such as all
user space I/O stack against hardware block device. It uses the new
vhost messages(VHOST_USER_GET_CONFIG) to get block virtio config
information from backend process.

Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user: add new vhost user messages to support virtio config space

Add VHOST_USER_GET_CONFIG/VHOST_USER_SET_CONFIG messages which can be
used for live migration of vhost user devices, also vhost user devices
can benefit from the messages to get/set virtio config space from/to the
I/O target. For the purpose to support virtio config space change,
VHOST_USER_SLAVE_CONFIG_CHANGE_MSG message is added as the event notifier
in case virtio config space change in the slave I/O target.

Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

MAINTAINERS: Add myself as maintainer to X86 machines

Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2018-01-17' into staging

pull-nbd-2018-01-17

- Vladimir Sementsov-Ogievskiy/Eric Blake: 0/6 NBD server refactoring

# gpg: Signature made Thu 18 Jan 2018 02:21:55 GMT
# gpg:                using RSA key 0xA7A16B4A2527436A
# gpg: Good signature from "Eric Blake <eblake@redhat.com>"
# gpg:                 aka "Eric Blake (Free Software Programmer) <ebb9@byu.net>"
# gpg:                 aka "[jpeg image of size 6874]"
# Primary key fingerprint: 71C2 CC22 B1C4 6029 27D2  F3AA A7A1 6B4A 2527 436A

* remotes/ericb/tags/pull-nbd-2018-01-17:
  nbd/server: structurize option reply sending
  nbd/server: Add helper functions for parsing option payload
  nbd/server: Add va_list form of nbd_negotiate_send_rep_err()
  nbd/server: Better error for NBD_OPT_EXPORT_NAME failure
  nbd/server: refactor negotiation functions parameters
  nbd/server: Hoist nbd_reject_length() earlier

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/ehabkost/tags/x86-pull-request' into staging

x86 queue, 2018-01-17

Highlight: new CPU models that expose CPU features that guests
can use to mitigate CVE-2017-5715 (Spectre variant #2).

# gpg: Signature made Thu 18 Jan 2018 02:00:03 GMT
# gpg:                using RSA key 0x2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>"
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/x86-pull-request:
  i386: Add EPYC-IBPB CPU model
  i386: Add new -IBRS versions of Intel CPU models
  i386: Add FEAT_8000_0008_EBX CPUID feature word
  i386: Add spec-ctrl CPUID bit
  i386: Add support for SPEC_CTRL MSR
  i386: Change X86CPUDefinition::model_id to const char*
  target/i386: add clflushopt to "Skylake-Server" cpu model
  pc: add 2.12 machine types

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.12-20180117' into staging

ppc patch queue 2017-01-17

Another pull request for ppc related patches.  The most interesting
thing here is the new capabilities framework for the pseries machine
type.  This gives us better handling of several existing
incompatibilities between TCG, PR and HV KVM, as well as new ones that
arise with POWER9.  Further, it will allow reasonable handling of the
advertisement of features necessary to mitigate the recent CVEs
(Spectre and Meltdown).

In addition there's:
     * Improvide handling of different "vsmt" modes
     * Significant enhancements to the "pnv" machine type
     * Assorted other bugfixes

# gpg: Signature made Wed 17 Jan 2018 02:21:50 GMT
# gpg:                using RSA key 0x6C38CACA20D9B392
# gpg: Good signature from "David Gibson <david@gibson.dropbear.id.au>"
# gpg:                 aka "David Gibson (Red Hat) <dgibson@redhat.com>"
# gpg:                 aka "David Gibson (ozlabs.org) <dgibson@ozlabs.org>"
# gpg:                 aka "David Gibson (kernel.org) <dwg@kernel.org>"
# Primary key fingerprint: 75F4 6586 AE61 A66C C44E  87DC 6C38 CACA 20D9 B392

* remotes/dgibson/tags/ppc-for-2.12-20180117: (22 commits)
  target-ppc: Fix booke206 tlbwe TLB instruction
  target/ppc: add support for POWER9 HILE
  ppc/pnv: change initrd address
  ppc/pnv: fix XSCOM core addressing on POWER9
  ppc/pnv: introduce pnv*_is_power9() helpers
  ppc/pnv: change core mask for POWER9
  ppc/pnv: use POWER9 DD2 processor
  tests/boot-serial-test: fix powernv support
  ppc/pnv: Update skiboot firmware image
  spapr: Adjust default VSMT value for better migration compatibility
  spapr: Allow some cases where we can't set VSMT mode in the kernel
  target/ppc: Clarify compat mode max_threads value
  ppc: Change Power9 compat table to support at most 8 threads/core
  spapr: Remove unnecessary 'options' field from sPAPRCapabilityInfo
  hw/ppc/spapr_caps: Rework spapr_caps to use uint8 internal representation
  spapr: Handle Decimal Floating Point (DFP) as an optional capability
  spapr: Handle VMX/VSX presence as an spapr capability flag
  target/ppc: Clean up probing of VMX, VSX and DFP availability on KVM
  spapr: Validate capabilities on migration
  spapr: Treat Hardware Transactional Memory (HTM) as an optional capability
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

cocoa.m: Fix scroll wheel support

When using a mouse's scroll wheel in a guest with
the cocoa front-end, the mouse pointer moves up
and down instead of scrolling the window. This
patch fixes this problem.

Signed-off-by: John Arbuckle <programmingkidx@gmail.com>
Message-id: 20180108180707.7976-1-programmingkidx@gmail.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>