pseries: Remove redundant calls to spapr_machine_initfn()

The instance_init() functions for several of the pseries-x.y versioned
machine types explicitly call spapr_machine_initfn().  But that's the
instance_init function for the common parent of all those machine types,
so will already have been called beforehand by the QOM infrastructure.

Remove the redundant calls.

Signed-off-by: David Gibson <[email protected]>
Reviewed-by: Thomas Huth <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>

pseries: Rearrange versioned machine type code

hw/ppc/spapr.c has a number of definitions related to the various versioned
machine types ("pseries-2.1" .. "pseries-2.5") it defines.  These are
mostly arranged by type of function first, then machine version second, and
it's not consistent about whether it goes in increasing or decreasing
version order.

This rearranges the code to keep all the definitions for a particular
machine version together, and arrange then consistently in order most
recent to least recent.

This brings us closer to matching the way PC does things, and makes later
cleanups easier to follow.

Apart from adding some comments marking each section, this is a pure
mechanical rearrangement with no semantic changes.

Signed-off-by: David Gibson <[email protected]>
Reviewed-by: Thomas Huth <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>

pseries: Remove redundant setting of mc->name for pseries-2.5 machine

98cec76 "machine: Set MachineClass::name automatically" removed the setting
of mc->name for the pseries machine types, since it can be derived
automatically from the type names constructed with MACHINE_TYPE_NAME().

Unfortunately fb0fc8f "spapr: Create pseries-2.5 machine" went in later and
brought one of them back.

This removes it again.

Signed-off-by: David Gibson <[email protected]>
Reviewed-by: Thomas Huth <[email protected]>
Reviewed-by: Alexey Kardashevskiy <[email protected]>

spapr: Add /system-id

Section B.6.2.1 Root Node Properties of PAPR specification defines
a set of properties which shall be present in the device tree root,
one of these properties is "system-id" which "should be unique across
all systems and all manufacturers". Since UUID is meant to be unique,
it makes sense to use it as "system-id".

This adds "system-id" property to the device tree root when not empty.

Signed-off-by: Alexey Kardashevskiy <[email protected]>
Signed-off-by: David Gibson <[email protected]>

target-ppc: Define kvmppc_read_int_dt()

Extract code from the function kvmppc_read_int_cpu_dt() that actually
reads the file into a separate function, so it can be called from
other places.

Signed-off-by: Sukadev Bhattiprolu <[email protected]>
Signed-off-by: David Gibson <[email protected]>

hw/ppc/spapr_rtc: Remove bad class_size value

class_size = sizeof(XICSStateClass) does not make much sense
in the RTC code and likely was just a copy-n-paste error.
Let's simply remove it.

Signed-off-by: Thomas Huth <[email protected]>
Signed-off-by: David Gibson <[email protected]>

Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20160108-1' into staging

sdl2/opengl: add opengl context and scanout support
ui/curses: Fix color attribute of monitor for curses

# gpg: Signature made Fri 08 Jan 2016 12:42:02 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <[email protected]>"
# gpg:                 aka "Gerd Hoffmann <[email protected]>"
# gpg:                 aka "Gerd Hoffmann (private) <[email protected]>"

* remotes/kraxel/tags/pull-ui-20160108-1:
  sdl2/opengl: add opengl context and scanout support
  ui/curses: Fix color attribute of monitor for curses

Signed-off-by: Peter Maydell <[email protected]>

Merge remote-tracking branch 'remotes/kraxel/tags/pull-usb-20160108-1' into staging

usb: mtp and ohci fixes.

# gpg: Signature made Fri 08 Jan 2016 10:14:59 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <[email protected]>"
# gpg:                 aka "Gerd Hoffmann <[email protected]>"
# gpg:                 aka "Gerd Hoffmann (private) <[email protected]>"

* remotes/kraxel/tags/pull-usb-20160108-1:
  ohci: clear pending SOF on suspend
  ohci: delay first SOF interrupt
  usb-mtp: fix call to trace function
  usb-mtp: use safe variant when cleaning events list
  ohci: fix command HostControllerReset
  ohci: fix Host Controller USBRESET
  ohci: split reset method in 3 parts

Signed-off-by: Peter Maydell <[email protected]>

sdl2/opengl: add opengl context and scanout support

This allows virtio-gpu to render in 3d mode.

Signed-off-by: Gerd Hoffmann <[email protected]>
Reviewed-by: Max Reitz <[email protected]>

ui/curses: Fix color attribute of monitor for curses

Current text_console_update() writes totally broken color attributes
to console_write_ch(). The format now is writing,

[WRONG]
bold << 21 | fg << 12 | bg << 8 | char
fg == 3bits curses color number
bg == 3bits curses color number

I can't see this format is where come from. Anyway, this doesn't work
at all.

What curses expects is actually (and vga.c is using),

[RIGHT]
bold << 21 | bg << 11 | fg << 8 | char
fg == 3bits vga color number
bg == 3bits vga color number

And curses set COLOR_PAIR() up to match this format, and curses's
chtype. I.e,

bold | color_pair | char
color_pair == (bg << 3 | fg)

To fix, this simply uses VGA color number everywhere except curses.c
internal. Then, convert it to above [RIGHT] format to write by
console_write_ch(). And as bonus, this reduces to expose curses define
to other parts (removes COLOR_* from console.c).

[Tested the first line is displayed as white on blue back for monitor
in curses console]

Signed-off-by: OGAWA Hirofumi <[email protected]>
Message-id: [email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

Merge remote-tracking branch 'remotes/maxreitz/tags/pull-block-for-peter-2016-01-07' into staging

Block patches from 2015-12-23 until 2016-01-07.

# gpg: Signature made Thu 07 Jan 2016 22:46:08 GMT using RSA key ID E838ACAD
# gpg: Good signature from "Max Reitz <[email protected]>"

* remotes/maxreitz/tags/pull-block-for-peter-2016-01-07: (21 commits)
  iotests: Add test cases for blockdev-mirror
  qmp: Add blockdev-mirror command
  block: Add check on mirror target
  block: Extract blockdev part of qmp_drive_mirror
  block: Rename BLOCK_OP_TYPE_MIRROR to BLOCK_OP_TYPE_MIRROR_SOURCE
  qemu-iotests: s390x: fix test 051
  iotests: 095: Filter _img_info output
  iotests: 095: Use TEST_IMG override instead of "mv"
  iotests: 050: Use TEST_IMG override instead of "mv"
  iotests: 038: Use TEST_IMG override instead of "mv"
  iotests: 037: Use TEST_IMG override instead of "mv"
  iotests: 034: Use TEST_IMG override instead of "mv"
  iotests: 028: Use TEST_IMG override instead of "mv"
  iotests: 024: Use TEST_IMG override instead of "mv"
  iotests: 020: Use TEST_IMG override instead of "mv"
  iotests: 019: Use TEST_IMG override instead of "mv"
  iotests: 018: Use TEST_IMG override instead of "mv"
  block/qapi: Clear err for further error
  block: use drained section in bdrv_close
  qemu-iotests: make check-block.sh work on out-of-tree builds
  ...

Signed-off-by: Peter Maydell <[email protected]>

ohci: clear pending SOF on suspend

On overcommitted CPU, kernel can be so slow that an interrupt can
be triggered by the device whereas the driver is not ready to receive
it. This drives us into an infinite loop.

On suspend, if a SOF interrupt is raised between the stop of the
device processing and the change of the device internal state to
OHCI_USB_SUSPEND (QEMU stops SOF timer on this state change), this
interrupt is never acknowledged.

This patch clears pending SOF interrupt on OHCI_USB_SUSPEND setting.

Some details:

- ohci_irq(): the OHCI interrupt handler, acknowledges the SOF IRQ
  only if the state of the driver (rh_state) is OHCI_STATE_RUNNING.
  So if this interrupt happens and the driver is not in this state,
  the function is called again and again, moving the system to a
  CPU starvation.

- ohci_rh_suspend(): the function stop the operation and acknowledge
  pending interrupts (but doesn't disable it). Later in the function,
  the device is moved to OHCI_SUSPEND_STATE, and the driver to
  OHCI_RH_SUSPENDED. If between the moment when the interrupt is
  acknowledged and the moment when the device is suspended a new
  interrupt is raised, it will be never acknowledged because the
  driver is now not in OHCI_RH_RUNNING state.

Signed-off-by: Laurent Vivier <[email protected]>
Reviewed-by: Thomas Huth <[email protected]>
Message-id: 1452109525[email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

ohci: delay first SOF interrupt

On overcommitted CPU, kernel can be so slow that an interrupt can
be triggered by the device whereas the driver is not ready to receive
it. This drives us into an infinite loop.

This does not happen on real hardware because real hardware never send
interrupt immediately after the controller has been moved to OPERATION state.

This patch tries to delay the first SOF interrupt to let driver exits from
the critical section (which is not protected against interrupts...)

Some details:

- ohci_irq(): the OHCI interrupt handler, acknowledges the SOF IRQ
  only if the state of the driver (rh_state) is OHCI_STATE_RUNNING.
  So if this interrupt happens and the driver is not in this state,
  the function is called again and again, moving the system to a
  CPU starvation.

- ohci_rh_resume(): the driver re-enables operation with OHCI_USB_OPER.
  In QEMU this start the SOF timer and QEMU starts to send IRQs. As
  the driver is not in OHCI_STATE_RUNNING and not protected against IRQ,
  the ohci_irq() can be called and the driver never moved to
  OHCI_STATE_RUNNING.

Suggested-by: Gerd Hoffmann <[email protected]>
Signed-off-by: Laurent Vivier <[email protected]>
Reviewed-by: Thomas Huth <[email protected]>
Message-id: 1452109525[email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

usb-mtp: fix call to trace function

trace_usb_mtp_inotify_event() was being called after the object was
being freed.

Signed-off-by: Bandan Das <[email protected]>
Message-id: 1450861787[email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

usb-mtp: use safe variant when cleaning events list

usb_mtp_inotify_cleanup uses QLIST_FOREACH to pick events
from a list and free them which is incorrect. Use QLIST_FOREACH_SAFE
instead.

Signed-off-by: Bandan Das <[email protected]>
Message-id: 1450861787[email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

ohci: fix command HostControllerReset

Specification says that: "This bit is set by HCD to initiate a software reset of HC."

Signed-off-by: Hervé Poussineau <[email protected]>
Tested-by: Mark Cave-Ayland <[email protected]>
Message-id: 1450567431[email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

ohci: fix Host Controller USBRESET

Specification says that, when entering this state, "the contents of the registers
(except Root Hub registers) are preserved by the HC. [...] The Root Hub is being reset,
which causes the Root Hub's downstream ports to be reset and possibly powered off."

Signed-off-by: Hervé Poussineau <[email protected]>
Tested-by: Mark Cave-Ayland <[email protected]>
Message-id: 1450567431[email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

ohci: split reset method in 3 parts

The three parts are:
- root hub reset (ohci_roothub_reset)
- host controller soft reset (ohci_soft_reset)
- host controller hard reset (ohci_hard_reset)

Signed-off-by: Hervé Poussineau <[email protected]>
Tested-by: Mark Cave-Ayland <[email protected]>
Message-id: 1450567431[email protected]
Signed-off-by: Gerd Hoffmann <[email protected]>

iotests: Add test cases for blockdev-mirror

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450932306[email protected]
Signed-off-by: Max Reitz <[email protected]>

qmp: Add blockdev-mirror command

This will start a mirror job from a named device to another named
device, its relation with drive-mirror is similar with blockdev-backup
to drive-backup.

In blockdev-mirror, the target node should be prepared by blockdev-add,
which will be responsible for assigning a name to the new node, so
we don't have 'node-name' parameter.

Signed-off-by: Fam Zheng <[email protected]>
Acked-by: Markus Armbruster <[email protected]>
Message-id: 1450932306[email protected]
Reviewed-by: Max Reitz <[email protected]>
Signed-off-by: Max Reitz <[email protected]>

block: Add check on mirror target

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450932306[email protected]
Signed-off-by: Max Reitz <[email protected]>

block: Extract blockdev part of qmp_drive_mirror

This is the part that will be reused by blockdev-mirror.

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450932306[email protected]
Signed-off-by: Max Reitz <[email protected]>

block: Rename BLOCK_OP_TYPE_MIRROR to BLOCK_OP_TYPE_MIRROR_SOURCE

It's necessary to distinguish source and target before we can add
blockdev-mirror, because we would want a concrete type of operation to
check on target bs before starting.

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450932306[email protected]
Signed-off-by: Max Reitz <[email protected]>

qemu-iotests: s390x: fix test 051

Replace the remaining "-drive file..."
by "-drive file=...,if=none,id=$device_id", then x86 and s390x
can get the common output.
"if=ide, if=floppy, if=scsi" are not supported by s390x,
so these test cases are not executed for s390x platform.

Signed-off-by: Bo Tu <[email protected]>
Message-id: 1451885360[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 095: Filter _img_info output

Signed-off-by: Fam Zheng <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 095: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 050: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 038: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 037: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 034: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 028: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 024: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 020: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 019: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 018: Use TEST_IMG override instead of "mv"

Signed-off-by: Fam Zheng <[email protected]>
Reviewed-by: Max Reitz <[email protected]>
Message-id: 1450752561[email protected]
Signed-off-by: Max Reitz <[email protected]>

block/qapi: Clear err for further error

Since a5002d5 (block/qapi: allow best-effort query) we don't return at
this error, however err must be cleared before passing to
bdrv_query_snapshot_info_list below, as required by error API.

Signed-off-by: Fam Zheng <[email protected]>
Message-id: 1450779107[email protected]
Reviewed-by: John Snow <[email protected]>
Signed-off-by: Max Reitz <[email protected]>

block: use drained section in bdrv_close

bdrv_close is used when ejecting a medium.  Use a drained section to ensure
that all I/O goes to either the old medium or the bitbucket.

Signed-off-by: Paolo Bonzini <[email protected]>
Message-id: 1450867706[email protected]
Signed-off-by: Max Reitz <[email protected]>

qemu-iotests: make check-block.sh work on out-of-tree builds

Since check-block.sh, the "check" script has learnt to find the source
path.  On the other hand, it expects common.env to be in the build tree
(both changes made in commit 76c7560, "configure: Enable out-of-tree
iotests", 2014-05-24).  So, it is wrong to invoke "check" from the source
path like check-block.sh does.  Fix it.

Signed-off-by: Paolo Bonzini <[email protected]>
Message-id: 1450867341[email protected]
Signed-off-by: Max Reitz <[email protected]>

iotests: 086: Add raw format

Raw is as qualified as qcow2 for this test case, add it for more
coverage.

Signed-off-by: Fam Zheng <[email protected]>
Message-id: 1450851979[email protected]
Signed-off-by: Max Reitz <[email protected]>

petalogix-ml605: Set the MicroBlaze CPU version to 8.10.a

Set the MicroBlaze CPU version to 8.10.a avoiding a runtime
warning due to an unset CPU version.

Reviewed-by: Peter Crosthwaite <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>

s3adsp1800: Set the MicroBlaze CPU version to 7.10.d

Set the MicroBlaze CPU version to 7.10.d avoiding a runtime
warning due to an unset CPU version.

Reviewed-by: Peter Crosthwaite <[email protected]>
Signed-off-by: Edgar E. Iglesias <[email protected]>

Merge remote-tracking branch 'remotes/mcayland/tags/qemu-sparc-signed' into staging

qemu-sparc update

# gpg: Signature made Thu 07 Jan 2016 13:20:13 GMT using RSA key ID AE0F321F
# gpg: Good signature from "Mark Cave-Ayland <[email protected]>"

* remotes/mcayland/tags/qemu-sparc-signed:
  target-sparc: implement NPT timer bit
  sun4u: split NPT and INT_DIS accesses between timer and compare registers
  sun4u: split out NPT and INT_DIS into separate CPUTimer fields

Signed-off-by: Peter Maydell <[email protected]>

target-sparc: implement NPT timer bit

If the NPT bit is set in the timer register, all non-supervisor read accesses
to the register should fail with a privilege exception.

Signed-off-by: Mark Cave-Ayland <[email protected]>
Reviewed-By: Artyom Tarasenko <[email protected]>
Signed-off-by: Mark Cave-Ayland <[email protected]>

sun4u: split NPT and INT_DIS accesses between timer and compare registers

Accesses to the timer register high bit should only set NPT, whilst accesses
to the timer compare register high bit should only set INT_DIS. This fixes
issues with the timer being unexpectedly disabled whilst trying to boot
FreeBSD SPARC64.

Signed-off-by: Mark Cave-Ayland <[email protected]>
Reviewed-By: Artyom Tarasenko <[email protected]>
Signed-off-by: Mark Cave-Ayland <[email protected]>

sun4u: split out NPT and INT_DIS into separate CPUTimer fields

Currently there is confusion between use of these bits for the timer and timer
compare registers (while they both have the same value, the behaviour is
different). Split into two separate CPUTimer fields so we can always reference
the correct value.

Signed-off-by: Mark Cave-Ayland <[email protected]>
Reviewed-By: Artyom Tarasenko <[email protected]>
Signed-off-by: Mark Cave-Ayland <[email protected]>

Merge remote-tracking branch 'remotes/kraxel/tags/pull-seabios-20160105-1' into staging

seabios: update to release 1.9.0

# gpg: Signature made Tue 05 Jan 2016 12:07:22 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <[email protected]>"
# gpg:                 aka "Gerd Hoffmann <[email protected]>"
# gpg:                 aka "Gerd Hoffmann (private) <[email protected]>"

* remotes/kraxel/tags/pull-seabios-20160105-1:
  seabios: update binaries to release 1.9.0
  seabios: stop updating aml files
  seabios: update 128k bios config
  seabios: use new EXTRAVERSION to tag qemu builds
  seabios: update submodule to release 1.9.0

Signed-off-by: Peter Maydell <[email protected]>

Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging

# gpg: Signature made Thu 07 Jan 2016 09:13:22 GMT using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <[email protected]>"
# gpg:                 aka "Stefan Hajnoczi <[email protected]>"

* remotes/stefanha/tags/tracing-pull-request:
  trace: add make dependencies on tracetool source
  trace: fix make foo-timestamp rules
  trace: fix PRIx64 constants in trace-events
  trace: reflect the file name change

Signed-off-by: Peter Maydell <[email protected]>

trace: add make dependencies on tracetool source

Patches that change tracetool can break the build if old build output
files are lying around.

This happens because the Makefile does not specify dependencies on
tracetool.  The build will use old object files that do not match the
current source code.

Signed-off-by: Stefan Hajnoczi <[email protected]>

trace: fix make foo-timestamp rules

The Makefile uses intermediate timestamp files to avoid rebuilding if
tracetool output is unchanged.

Timestamps are implemented incorrectly.  This was fixed for rules.mak in
commit 4b25966ab976f3a7fd9008193b2defcc82f8f04d ("rules.mak: cleanup
config generation rules") but never fixed in trace/Makefile.objs.

The problem with the old timestamp implementation was that make doesn't
notice the updated file modification time until the next time it is run.
It was necessary to run make twice in a row to achieve a full rebuild.

Signed-off-by: Stefan Hajnoczi <[email protected]>

trace: fix PRIx64 constants in trace-events

Commit c8ee0a4 introduced new events containing PRIx64 constants without
including the % prefix in the preceding string. This results in a compile
error during build if --enable-trace-backends is passed to configure.

Signed-off-by: Mark Cave-Ayland <[email protected]>
Message-id: 1450566522[email protected]
Signed-off-by: Stefan Hajnoczi <[email protected]>

trace: reflect the file name change

Some functions was moved from block.c to block/io.c, so the trace-events file should reflect that change.

Signed-off-by: Qinghua Jin <[email protected]>
Signed-off-by: Stefan Hajnoczi <[email protected]>

seabios: update binaries to release 1.9.0

Signed-off-by: Gerd Hoffmann <[email protected]>

seabios: stop updating aml files

ACPI aml files traditionally have been managed in the seabios repo.
In qemu version 2.0 we've switched over to have qemu generate the
acpi tables and provide them to the firmware via fw_cfg.

The old aml files are still there and used for old machine types.
Well, actually the q35 file only, the piix4 version is compiled into
seabios (unless built with CONFIG_ACPI_DSDT=n) and is there for
reference only.

The aml files havn't been touched for a long time, and given that
new features requiring acpi changes are typically only added to new
machine types this is unlikely to change in the future.  So stop
updating them.

That allows to cleanup things a bit on the seabios side in the future.

Signed-off-by: Gerd Hoffmann <[email protected]>

seabios: update 128k bios config

Turn off OHCI + TPM support to keep the size below 128k.

Signed-off-by: Gerd Hoffmann <[email protected]>

seabios: use new EXTRAVERSION to tag qemu builds

Signed-off-by: Gerd Hoffmann <[email protected]>

seabios: update submodule to release 1.9.0

Highlights / user visible changes in seabios:
 * boot menu key is ESC now.
 * virtio 1.0 support.
 * sdcard support.
 * fw_cfg dma suport.
 * usual share of bugfixes ;)

In vgabios:
 * Emulates leal instruction.  Works around a bug in old x86emu versions,
   which makes old xorg vesa drivers work (RHEL-5 for example).

full shortlog rel-1.8.2..rel-1.9.0
----------------------------------

Ameya Palande (1):
      x86: add barrier to read{b,w,l} and write{b,w,l} functions

Andreas Färber (1):
      checkrom: Fix typo in error message

Chen Fan (1):
      pci: enable SERR# for error forwarding in bridge control register

Gerd Hoffmann (28):
      vga: simplify vga builds
      vga: rework virtio-vga support
      vga: add virtio-vga to kconfig
      pci: allow to loop over capabilities
      virtio: run drivers in 32bit mode
      virtio: add struct vp_device
      virtio: pass struct pci_device to vp_init_simple
      virtio: add version 1.0 structs and #defines
      virtio: add version 0.9.5 struct
      virtio: find version 1.0 virtio capabilities
      virtio: create vp_cap struct for legacy bar
      virtio: add read/write functions and macros
      virtio: make features 64bit, support version 1.0 features
      virtio: add version 1.0 support to vp_{get,set}_status
      virtio: add version 1.0 support to vp_get_isr
      virtio: add version 1.0 support to vp_reset
      virtio: add version 1.0 support to vp_notify
      virtio: remove unused vp_del_vq
      virtio: add version 1.0 support to vp_find_vq
      virtio-scsi: fix initialization for version 1.0
      virtio-blk: fix initialization for version 1.0
      virtio: use version 1.0 if available (flip the big switch)
      virtio: also probe version 1.0 pci ids
      virtio: legacy cleanup
      virtio-blk: 32bit cleanup
      virtio-scsi: 32bit cleanup
      virtio-ring: 32bit cleanup
      virtio-pci: use high memory for rings

Julius Werner (1):
      xhci: Count new Max Scratchpad Bufs bits from XHCI 1.1

Kevin O'Connor (126):
      docs: add page for SeaVGABIOS
      docs: Add page describing the patch contribution process
      docs: Add page on available CBFS/fw_cfg runtime config files
      docs: Prefer triple backticks to multiple lines with single backticks
      smp: Fix smp race introduced in 0673b787
      docs: Note release date of 1.8.1
      vgabios: On bda_save_restore() the saved vbe_mode also has flags in it
      vgabios: Don't use extra stack if it appears a modern OS is in use
      docs: Clarify that pci-optionrom-exec doesn't apply to roms in cbfs
      checkstack: Replace function information tuple with class
      checkstack: Simplify yield calculations
      checkstack: Prefer passing "function" class instead of function address
      smbios: Use integer signature instead of string signature
      vgabios: Don't use "smsww" instruction - it confuses x86emu
      vgabios: Add config option for assembler fixups
      vgabios: Emulate "leal" instruction
      checkstack: Minor - continue if not a regular asm line
      Don't forward declare functions with "inline" in headers
      build: Support "make VERSION=xyz" to override the default build version
      tcg: Use seabios setup()/prepboot() calling convention for tcg
      build: CONFIG_VGA_FIXUP_ASM should depend on CONFIG_BUILD_VGABIOS
      bootorder: Update "extra pci root" buses bootorder format to match qemu
      Make sure all code checks for malloc failures
      docs: Note release date of 1.8.2
      block: Split process_op() command dispatch up into multiple functions
      block: Introduce default_process_op() with common command handling codes
      block: Route scsi style commands through 'struct disk_op_s'
      blockcmd: Introduce scsi_fill_cmd()
      ata: Handle ATA ATAPI drives directly via 'struct disk_op_s' requests
      ahci: Handle AHCI ATAPI drives directly via 'struct disk_op_s' requests
      usb-msc: Handle USB drives directly via 'struct disk_op_s' requests
      usb-uas: Handle USB drives directly via 'struct disk_op_s' requests
      lsi-scsi: Handle LSI drives directly via 'struct disk_op_s' requests
      esp-scsi: Handle ESP drives directly via 'struct disk_op_s' requests
      megasas: Handle Megasas drives directly via 'struct disk_op_s' requests
      virtio-scsi: Handle virtio drives directly via 'struct disk_op_s' requests
      pvscsi: Move pvscsi_fill_req() code into pvscsi_cmd()
      pvscsi: Handle pvscsi drives directly via 'struct disk_op_s' requests
      blockcmd: Remove unused scsi_process_op() and cdb_cmd_data()
      blockcmd: Convert cdb_is_read() to scsi_is_read()
      block: Rename process_XXX_op() functions to XXX_process_op()
      coreboot: Try to auto-detect if the CBFS anchor pointer is a relative pointer
      ps2: Support mode for polling the PS2 port instead of using irqs
      ata: Make sure "chanid" is relative to PCI device for bootorder file
      Don't enable interrupts prior to IVT and PIC setup
      ps2: Don't wait 100ms to discard possible extra reset receive byte
      timer: Delay timestamp counter init until after pmtimer is probed
      timer: Add CONFIG_TSC_TIMER build option to disable the CPU TSC timer
      ramdisk: Allow ramdisk support (CONFIG_FLASH_FLOPPY) under QEMU
      Minor - move declaration of CDRom_locks to code that uses it
      smm: ignore bits 16,18-31 of SMM revision ID at runtime too
      vgafb: Minor - move gfx_common() variables outside of switch statement
      sdcard: Check if card is present before sending commands to card
      sdcard: Implement controller frequency setting according to sdhci spec
      sdcard: Make sure controller support 3.3V before enabling it
      sdcard: Set timeout control register during init (to max allowed timeout)
      sdcard: Improve SD card initialization command sequence
      sdcard: Add proper delays during card power up
      mptable: Don't create mptable if it is very large
      optionroms: Don't run option rom on PCI bar if CBFS/fw_cfg version exists
      edd: Pass the segment/offset from int 1348 calls using a 'struct segoff_s'
      edd: Reduce parameters to fill_generic_edd()
      Move CanInterrupt check to check_irqs()
      Call cpu_relax() if yielding prior to interrupts being enabled
      sdcard: Fix typo - use sdcard_pio() instead of sdcard_pio_app()
      sdcard: Fill command bits according to spec
      sdcard: Support SDHCI v3.00 spec clock setting
      sdcard: Move power setup to new function sdcard_set_power()
      sdcard: Power controller up to maximum voltage supported
      sdcard: Power down controller on failure
      sdcard: The card should never be in a busy state at start of sdcard_pio()
      sdcard: Implement timeout on every block read in sdcard_pio_transfer()
      sdcard: Rename waitw() to sdcard_waitw() and simplify
      sdcard: Perform a controller reset at start of init
      sdcard: Check for error events during sdcard_pio()
      sdcard: Initial support for MMC cards
      sdcard: Allow the sdcard driver to run on real hardware
      rtc: Support disabling the RTC timer irq support
      Add minimal support for machines without hardware interrupts
      ps2: Eliminate "etc/ps2-poll-only"; use CONFIG_HARDWARE_IRQ instead
      sdcard: Allow sdcard addresses to be specified in CBFS files
      xhci: Minor - add USB port type comments to xhci_hub_reset()
      docs: Don't use an add-symbol-file offset when describing gdb debugging
      rtc: Disable NMI in rtc_mask()
      sdcard: Move sdcard_set_frequency()/sdcard_set_power() in sdcard.c
      sdcard: Move frequency setting into sdcard_card_setup()
      sdcard: Move drive registration to sdcard_card_setup()
      sdcard: Turn card_type into a bitmap and store if card is MMC type
      sdcard: Display sdcard product name in boot menu
      sdcard: Obtain card capacity and report it on the boot menu
      megasas: Use outl() on MFI_IDB register
      minor - correct spelling error in comment
      Simplify transition16/32 assembler code
      docs: Minor - add "code relocation" link to "Execution and code flow" document
      Unify smm/sloppy variants of call32_prep/post and call16_helper
      Rename Call32Data to Call16Data
      Unify inline assembler in variants of call16 functions
      Unify call32_sloppy() and call32()
      Use transition32_nmi_off from call32() and call16_back()
      Consolidate code16*() functions
      Always enable caching on transition32; backup/restore cr0 on call32
      e820: Introduce e820_remove() and avoid exporting E820_HOLE
      e820: Rename memmap.c to e820map.c and use consistent "e820_" prefix
      e820: Update debugging messages to report 64bit values
      virtio: Simplify vring alignment code
      virtio: Move standard definitions from virtio-ring.h to standard headers
      malloc: Use consistent naming for internal low-level "alloc" functions
      malloc: Introduce common helper alloc_new_detail()
      malloc: Add warning if free() called on invalid memory
      malloc: Don't mix virtual and physical addresses
      memmap: Introduce SYMBOL() macro to access linker script symbols
      build: Rework version generation; don't allow make version override
      build: Report gcc and binutils versions in debug log
      build: Generate "reproducible" version strings on "clean" builds
      stacks: Use macro wrappers for call32() and stack_hop_back()
      malloc: Rename csm_malloc_preinit() to malloc_csm_preinit()
      build: Be more permissive in buildversion.py tool version scan
      docs: Document 'make EXTRAVERSION=xyz' and scripts/tarball.sh
      build: Allow official tarball builds to be considered "clean"
      coreboot: Minor - avoid K&R style function declaration
      biostables: Minor - fix incorrect indentation
      virtio: Minor - replace tab characters with space
      docs: Minor - replace seavgabios text in Build_overview.md with link
      buildversion: Avoid subprocess.check_output() as that requires python2.7
      buildversion: Add debugging messages
      docs: Note v1.9.0 release

Kyösti Mälkki (1):
      PCI SDHCI driver: Fix base address

Magnus Granberg (1):
      build: use -fstack-check=no when available

Marc Marí (1):
      Add QEMU fw_cfg DMA interface

Marcel Apfelbaum (2):
      fw/pci: scan all buses if extraroots romfile is present
      fw/pci: map memory and IO regions for multiple pci root buses

Paolo Bonzini (4):
      boot.c: delay exiting boot if menu key is ESC
      boot: switch default menu key to ESC
      smm: ignore bits 16,18-31 of SMM revision ID
      smm: fix outl argument order

Paulo Alcantara (1):
      ich9: initialise RCBA register through LPC interface

Quan Xu (1):
      make SeaBios compatible with Xen vTPM.

Stefan Berger (9):
      Add an implementation of a TPM TIS driver
      Implementation of the TCG BIOS extensions
      Support for BIOS interrupt handler
      Add 'measurement' code to the BIOS
      tpm: Introduce a #define for command tag
      tpm: Be consistent with array sizes in tcgbios.c
      tpm: clean up parameters to build_and_send_cmd
      tpm: Clean up in tcgbios.h
      tpm: Move call to tpm_option_rom into init_optionrom

Stefan Weil (2):
      megasas: Fix outw, outl argument order
      Fix typos found by codespell

Vladimir Serbinenko (3):
      ahci: Ignore max_ports.
      Link rom.o with -N option.
      Add multiboot support.

[email protected] (1):
      Add an option to only execute option ROMs contained in CBFS

Signed-off-by: Gerd Hoffmann <[email protected]>

Merge remote-tracking branch 'remotes/berrange/tags/pull-crypto-fixes-2015-12-23-1' into staging

Merge misc crypto changes & fixes

# gpg: Signature made Wed 23 Dec 2015 11:11:54 GMT using RSA key ID 15104FDF
# gpg: Good signature from "Daniel P. Berrange <[email protected]>"
# gpg:                 aka "Daniel P. Berrange <[email protected]>"

* remotes/berrange/tags/pull-crypto-fixes-2015-12-23-1:
  crypto: fix transposed arguments in cipher error message
  crypto: ensure qapi/crypto.json is listed in qapi-modules
  crypto: move QCryptoCipherAlgorithm/Mode enum definitions into QAPI
  crypto: move QCryptoHashAlgorithm enum definition into QAPI
  crypto: add ability to query hash digest len
  crypto: add additional query accessors for cipher instances

Signed-off-by: Peter Maydell <[email protected]>

Merge remote-tracking branch 'remotes/gkurz/tags/for-upstream' into staging

Fix a 2.5 regression.

# gpg: Signature made Wed 23 Dec 2015 10:57:00 GMT using DSA key ID 0101DBC2
# gpg: Good signature from "Greg Kurz <[email protected]>"
# gpg:                 aka "Greg Kurz <[email protected]>"
# gpg:                 aka "Greg Kurz <[email protected]>"
# gpg:                 aka "Gregory Kurz (Groug) <[email protected]>"
# gpg:                 aka "Gregory Kurz (Cimai Technology) <[email protected]>"
# gpg:                 aka "Gregory Kurz (Meiosys Technology) <[email protected]>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 2BD4 3B44 535E C0A7 9894  DBA2 02FC 3AEB 0101 DBC2

* remotes/gkurz/tags/for-upstream:
  virtio-9p: use accessor to get thread_pool

Signed-off-by: Peter Maydell <[email protected]>

Merge remote-tracking branch 'remotes/berrange/tags/pull-io-fixes-2015-12-23-1' into staging

Merge misc I/O channel fixes

# gpg: Signature made Wed 23 Dec 2015 10:54:52 GMT using RSA key ID 15104FDF
# gpg: Good signature from "Daniel P. Berrange <[email protected]>"
# gpg:                 aka "Daniel P. Berrange <[email protected]>"

* remotes/berrange/tags/pull-io-fixes-2015-12-23-1:
  io: fix stack allocation when sending of file descriptors
  io: fix setting of QIO_CHANNEL_FEATURE_FD_PASS on server connections
  io: bind to loopback IP addrs in test suite

Signed-off-by: Peter Maydell <[email protected]>

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

acpi, pc features

pxb support for q35
nvdimm support
most of ipmi support
part of DSDT rewrite

Signed-off-by: Michael S. Tsirkin <[email protected]>
# gpg: Signature made Tue 22 Dec 2015 16:47:18 GMT using RSA key ID D28D5469
# gpg: Good signature from "Michael S. Tsirkin <[email protected]>"
# gpg:                 aka "Michael S. Tsirkin <[email protected]>"

* remotes/mst/tags/for_upstream: (55 commits)
  acpi: extend aml_and() to accept target argument
  acpi: extend aml_or() to accept target argument
  acpi add aml_dma()
  acpi: add aml_to_buffer()
  acpi: add aml_to_hexstring()
  acpi: extend aml_field() to support LockRule
  acpi: add aml_lgreater()
  acpi: add aml_lor()
  acpi: add aml_sleep()
  acpi: add aml_alias()
  acpi: extend aml_shiftright() to accept target argument
  acpi: add aml_to_integer()
  acpi: add aml_call0() helper
  acpi: add aml_decrement() and aml_subtract()
  acpi: extend aml_add() to accept target argument
  acpi: aml: add helper for Opcode Arg2 Arg2 [Dst] AML pattern
  acpi: add aml_create_qword_field()
  acpi: add aml_mutex(), aml_acquire(), aml_release()
  acpi: add aml_lgreater_equal()
  acpi: add aml_sizeof
  ...

Signed-off-by: Peter Maydell <[email protected]>

Merge remote-tracking branch 'remotes/sstabellini/tags/xen-2015-12-22' into staging

Xen 2015/12/22

# gpg: Signature made Tue 22 Dec 2015 16:17:57 GMT using RSA key ID 70E1AE90
# gpg: Good signature from "Stefano Stabellini <[email protected]>"

* remotes/sstabellini/tags/xen-2015-12-22:
  xen_disk: treat "vhd" as "vpc"
  xen/pass-through: correctly deal with RW1C bits
  xen/MSI-X: really enforce alignment
  xen/MSI-X: latch MSI-X table writes

Signed-off-by: Peter Maydell <[email protected]>

crypto: fix transposed arguments in cipher error message

When reporting an incorrect key length for a cipher, we
mixed up the actual vs expected arguments.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

crypto: ensure qapi/crypto.json is listed in qapi-modules

The rebuild of qapi-types.c/h is not correctly triggered
when qapi/crypto.json is changed because it was missing
from the list of files in the qapi-modules variable.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

crypto: move QCryptoCipherAlgorithm/Mode enum definitions into QAPI

The QCryptoCipherAlgorithm and QCryptoCipherMode enums are
defined in the crypto/cipher.h header. In the future some
QAPI types will want to reference the hash enums, so move
the enum definition into QAPI too.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

crypto: move QCryptoHashAlgorithm enum definition into QAPI

The QCryptoHashAlgorithm enum is defined in the crypto/hash.h
header. In the future some QAPI types will want to reference
the hash enums, so move the enum definition into QAPI too.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

crypto: add ability to query hash digest len

Add a qcrypto_hash_digest_len() method which allows querying of
the raw digest size for a given hash algorithm.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

crypto: add additional query accessors for cipher instances

Adds new methods to allow querying the length of the cipher
key, block size and initialization vectors.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

io: fix stack allocation when sending of file descriptors

When sending file descriptors over a socket, we have to
allocate a data buffer to hold the FDs in the scmsghdr.
Unfortunately we allocated the buffer on the stack inside
an if () {} block, but called sendmsg() outside the block.
So the stack bytes holding the FDs were liable to be
overwritten with other data. By luck this was not a problem
when sending 1 FD, but if sending 2 or more then it would
fail.

The fix is to simply move the variables outside the nested
'if' block. To keep valgrind quiet we also zero-initialize
the 'control' buffer.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

virtio-9p: use accessor to get thread_pool

The aio_context_new() function does not allocate a thread pool. This is
deferred to the first call to the aio_get_thread_pool() accessor. It is
hence forbidden to access the thread_pool field directly, as it may be
NULL. The accessor *must* be used always.

Fixes: ebac1202c95a4f1b76b6ef3f0f63926fa76e753e
Reviewed-by: Michael Tokarev <[email protected]>
Tested-by: Michael Tokarev <[email protected]>
Cc: [email protected]
Signed-off-by: Greg Kurz <[email protected]>

io: fix setting of QIO_CHANNEL_FEATURE_FD_PASS on server connections

The QIO_CHANNEL_FEATURE_FD_PASS feature flag is set in the
qio_channel_socket_set_fd() method, however, this only deals
with client side connections.

To ensure server side connections also have the feature flag
set, we must set it in qio_channel_socket_accept() too. This
also highlighted a typo fix where the code updated the
sockaddr struct in the wrong object instance.

Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

io: bind to loopback IP addrs in test suite

The test suite currently binds to 0.0.0.0 or ::, which covers
all interfaces of the machine. It is bad practice for test
suite to open publically accessible ports on a machine, so
switch to use loopback addrs 127.0.0.1 or ::1.

Reported-by: Peter Maydell <[email protected]>
Reviewed-by: Eric Blake <[email protected]>
Signed-off-by: Daniel P. Berrange <[email protected]>

acpi: extend aml_and() to accept target argument

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

acpi: extend aml_or() to accept target argument

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi add aml_dma()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

acpi: add aml_to_buffer()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_to_hexstring()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: extend aml_field() to support LockRule

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

acpi: add aml_lgreater()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_lor()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_sleep()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_alias()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: extend aml_shiftright() to accept target argument

it allows to express ShiftRight(A,B,C) syntax

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_to_integer()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_call0() helper

it will help to call a method with 0 arguments

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_decrement() and aml_subtract()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: extend aml_add() to accept target argument

it allows to express following ASL expression:
 Add(arg1, arg2, result)

usecases that do not need to store result
should pass NULL as 3rd arg that would express
 Add(arg1, arg2,)
construct.

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: aml: add helper for Opcode Arg2 Arg2 [Dst] AML pattern

Currently AML API doesn't compose terms in form of
following pattern:

   Opcode Arg2 Arg2 [Dst]

but ASL used in piix4/q35 DSDT ACPI tables uses that
form, so for clean conversion of it, AML API should
be able to handle an optional 'Dst' argumet used there.

Since above pattern is used by arithmetic/bit ops,
introduce helper that they could reuse.
It reduces code duplication in existing 5 aml_foo()
functions and also will prevent more duplication
when exiting functions are extended to support
optional 'Dst' argument.

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>
Reviewed-by: Shannon Zhao <[email protected]>

acpi: add aml_create_qword_field()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

acpi: add aml_mutex(), aml_acquire(), aml_release()

Signed-off-by: Xiao Guangrong <[email protected]>
Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

acpi: add aml_lgreater_equal()

Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>
Reviewed-by: Shannon Zhao <[email protected]>

acpi: add aml_sizeof

Implement SizeOf term which is used by NVDIMM _DSM method in later patch

Reviewed-by: Igor Mammedov <[email protected]>
Signed-off-by: Xiao Guangrong <[email protected]>
Signed-off-by: Igor Mammedov <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

acpi: add aml_derefof

Implement DeRefOf term which is used by NVDIMM _DSM method in later patch

Reviewed-by: Igor Mammedov <[email protected]>
Signed-off-by: Xiao Guangrong <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

nvdimm: add maintain info

Add NVDIMM maintainer

Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Xiao Guangrong <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

nvdimm acpi: build ACPI nvdimm devices

NVDIMM devices is defined in ACPI 6.0 9.20 NVDIMM Devices

There is a root device under \_SB and specified NVDIMM devices are under the
root device. Each NVDIMM device has _ADR which returns its handle used to
associate MEMDEV structure in NFIT

Currently, we do not support any function on _DSM, that means, NVDIMM
label data has not been supported yet

Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Xiao Guangrong <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

nvdimm acpi: build ACPI NFIT table

NFIT is defined in ACPI 6.0: 5.2.25 NVDIMM Firmware Interface Table (NFIT)

Currently, we only support PMEM mode. Each device has 3 structures:
- SPA structure, defines the PMEM region info

- MEM DEV structure, it has the @handle which is used to associate specified
  ACPI NVDIMM  device we will introduce in later patch.
  Also we can happily ignored the memory device's interleave, the real
  nvdimm hardware access is hidden behind host

- DCR structure, it defines vendor ID used to associate specified vendor
  nvdimm driver. Since we only implement PMEM mode this time, Command
  window and Data window are not needed

The NVDIMM functionality is controlled by the parameter, 'nvdimm', which
is introduced for the machine, there is a example to enable it:
-machine pc,nvdimm -m 8G,maxmem=100G,slots=100  -object \
memory-backend-file,id=mem1,share,mem-path=/tmp/nvdimm1,size=10G -device \
nvdimm,memdev=mem1,id=nv1

It is disabled on default

Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Xiao Guangrong <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

acpi: support specified oem table id for build_header

Let build_header() support specified OEM table id so that we can build
multiple SSDT later

If the oem table id is not specified (aka, NULL), we use the default id
instead as the previous behavior

Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Xiao Guangrong <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

nvdimm: implement NVDIMM device abstract

Introduce "nvdimm" device which is based on pc-dimm device type

Currently, nothing is specific for nvdimm but hotplug is disabled

Reviewed-by: Stefan Hajnoczi <[email protected]>
Signed-off-by: Xiao Guangrong <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>

docs/pci_expander_bridge: fix typo

Signed-off-by: Cao jin <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Marcel Apfelbaum <[email protected]>

hw/compat.h: Change indentation of HW_COMPAT_* to 4 spaces

Cosmetic change only.

Signed-off-by: Eduardo Habkost <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Thomas Huth <[email protected]>

pc: Change indentation of PC_COMPAT_* to 4 spaces

Cosmetic change only.

Signed-off-by: Eduardo Habkost <[email protected]>
Reviewed-by: Michael S. Tsirkin <[email protected]>
Signed-off-by: Michael S. Tsirkin <[email protected]>
Reviewed-by: Thomas Huth <[email protected]>