sdk/emulator/qemu.git
12 years agoqom: add test tools
Anthony Liguori [Mon, 20 Feb 2012 22:28:30 +0000 (16:28 -0600)]
qom: add test tools

Tested-by: Andreas F=E4rber <afaerber@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqmp: make qmp.py easier to use
Anthony Liguori [Mon, 20 Feb 2012 22:28:29 +0000 (16:28 -0600)]
qmp: make qmp.py easier to use

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoMakefile: Add dependency to fix linux-user-only build
Peter Maydell [Tue, 21 Feb 2012 14:44:40 +0000 (14:44 +0000)]
Makefile: Add dependency to fix linux-user-only build

Make qemu-bridge-helper explicitly depend on $(GENERATED_HEADERS)
so that it doesn't fail to build when we configured for linux-user
targets only. (Build breakage introduced in commit 7b93fad.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoMake kernel, initrd and append be machine_opts
Peter Maydell [Wed, 8 Feb 2012 05:41:39 +0000 (05:41 +0000)]
Make kernel, initrd and append be machine_opts

Make kernel, initrd, append be machine opts (ie -machine kernel=foo)
with the old plain command line arguments as legacy/convenience
equivalents.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agojazz-led: compile it only twice
Hervé Poussineau [Fri, 17 Feb 2012 19:27:17 +0000 (20:27 +0100)]
jazz-led: compile it only twice

Signed-off-by: Hervé Poussineau <hpoussin@reactos.org>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agojazz-led: convert to QOM
Hervé Poussineau [Fri, 17 Feb 2012 19:27:16 +0000 (20:27 +0100)]
jazz-led: convert to QOM

Some simplifications in I/O functions are possible because
Jazz LED only registers one byte of I/O.

Signed-off-by: Hervé Poussineau <hpoussin@reactos.org>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agojazz-led: use trace framework
Hervé Poussineau [Fri, 17 Feb 2012 19:27:15 +0000 (20:27 +0100)]
jazz-led: use trace framework

Signed-off-by: Hervé Poussineau <hpoussin@reactos.org>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqom: Fix object_initialize_with_type() assertion
Andreas Färber [Sun, 19 Feb 2012 23:49:07 +0000 (00:49 +0100)]
qom: Fix object_initialize_with_type() assertion

Assert the object is at least sizeof(Object), not sizeof(ObjectClass).

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agopci: set memory type for memory behind the bridge
Michael S. Tsirkin [Tue, 21 Feb 2012 13:57:58 +0000 (15:57 +0200)]
pci: set memory type for memory behind the bridge

As we make upper bits in IO and prefetcheable memory
registers writeable, we should declare support
for 64 bit prefetcheable memory and 32 bit io
in the bridge.

This changes the default for apb, dec, but I'm guessing
they got the defaults wrong by accident.
Alternatively, we could let bridges declare lack of
64 bit support and make the upper bits read-only zero.

With this applied, we can drop these bits
from express code.

Reported-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Could someone familiar with apb,dec ack this please?
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agopci: add accessors to get/set registers by mask
Michael S. Tsirkin [Tue, 21 Feb 2012 13:41:30 +0000 (15:41 +0200)]
pci: add accessors to get/set registers by mask

pci_regs.h specifies many registers by mask +
shifted register values.
There's always some duplication when using such:
for example to override device type, we would need:

pci_word_test_and_clear_mask(cap + PCI_EXP_FLAGS,
     PCI_EXP_FLAGS_TYPE);
pci_word_test_and_set_mask(cap + PCI_EXP_FLAGS,
    PCI_EXP_TYPE_ENDPOINT << (ffs(PCI_EXP_FLAGS_TYPE) - 1));

Getting such registers also uses some duplication:

word = pci_get_word(cap + PCI_EXP_FLAGS) & PCI_EXP_FLAGS_TYPE;
if ((word >> ffs((PCI_EXP_FLAGS_TYPE) - 1)) == PCI_EXP_TYPE_ENDPOINT)

Add API to access such registers in one line:
pci_set_word_by_mask(cap + PCI_EXP_FLAGS, PCI_EXP_FLAGS_TYPE,
     PCI_EXP_TYPE_ENDPOINT)

and
word = pci_get_word_by_mask(cap + PCI_EXP_FLAGS, PCI_EXP_FLAGS_TYPE)
if (word == PCI_EXP_TYPE_ENDPOINT)

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqom: Document ways to retrieve child object added by object_property_add_child()
Alexander Barabash [Tue, 21 Feb 2012 10:14:22 +0000 (12:14 +0200)]
qom: Document ways to retrieve child object added by object_property_add_child()

object_property_add_child() creates a property whose values as a string is
the child object's canonical path.

Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alexander Barabash <alexander_barabash@mentor.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agopc_piix/pc_sysfw: enable flash by default
Jordan Justen [Wed, 22 Feb 2012 07:18:56 +0000 (23:18 -0800)]
pc_piix/pc_sysfw: enable flash by default

Now, the pc-sysfw:rom_only property will default
to false which enables flash by default.

All pc types below pc-1.1 set rom_only to true.
This prevents flash from being enabled on these
pc machine types.

For pc-1.1 rom_only will use the default (false),
which will allow flash to be used for pc-1.1.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agohw/pc_piix: add pc-1.1
Jordan Justen [Wed, 22 Feb 2012 07:18:55 +0000 (23:18 -0800)]
hw/pc_piix: add pc-1.1

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agohw/pc_piix: remove is_default for pc-0.15
Jordan Justen [Wed, 22 Feb 2012 07:18:54 +0000 (23:18 -0800)]
hw/pc_piix: remove is_default for pc-0.15

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agohw/pc_sysfw: support system flash memory with pflash
Jordan Justen [Wed, 22 Feb 2012 07:18:53 +0000 (23:18 -0800)]
hw/pc_sysfw: support system flash memory with pflash

Flash can be enabled by calling pc_system_firmware_init
with the system_flash_enabled parameter being non-zero.
If system_flash_enabled is zero, then the older qemu
rom creation method will be used.

If flash is enabled and a pflash image is found, then
it is used for the system firmware image.

If flash is enabled and a pflash image is not initially
found, then a read-only pflash device is created using
the -bios filename.

KVM cannot execute from a pflash region currently.
Therefore, when KVM is enabled, the old rom based
initialization method is used.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agohw/pc_sysfw: enable pc-sysfw as a qdev
Jordan Justen [Wed, 22 Feb 2012 07:18:52 +0000 (23:18 -0800)]
hw/pc_sysfw: enable pc-sysfw as a qdev

Setup a pc-sysfw device type.  It contains a single
property of 'rom_only' which is defaulted to enabled.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agohw/pc: move rom init to pc_sysfw.c
Jordan Justen [Wed, 22 Feb 2012 07:18:51 +0000 (23:18 -0800)]
hw/pc: move rom init to pc_sysfw.c

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agovl: make find_default_machine externally visible
Jordan Justen [Wed, 22 Feb 2012 07:18:50 +0000 (23:18 -0800)]
vl: make find_default_machine externally visible

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agopflash_cfi01/02: support read-only pflash devices
Jordan Justen [Wed, 22 Feb 2012 07:18:49 +0000 (23:18 -0800)]
pflash_cfi01/02: support read-only pflash devices

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoblockdev: allow read-only pflash devices
Jordan Justen [Wed, 22 Feb 2012 07:18:48 +0000 (23:18 -0800)]
blockdev: allow read-only pflash devices

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoMerge remote-tracking branch 'bonzini/qdev-props-for-anthony' into staging
Anthony Liguori [Wed, 22 Feb 2012 13:35:03 +0000 (07:35 -0600)]
Merge remote-tracking branch 'bonzini/qdev-props-for-anthony' into staging

* bonzini/qdev-props-for-anthony:
  qdev: drop unnecessary parse/print methods
  qdev: use built-in QOM string parser
  qdev: accept hex properties only if prefixed by 0x
  qdev: accept both strings and integers for PCI addresses
  qom: add generic string parsing/printing
  qapi: add tests for string-based visitors
  qapi: add string-based visitors
  qapi: drop qmp_input_end_optional
  qapi: allow sharing enum implementation across visitors

12 years agoqdev: drop unnecessary parse/print methods
Paolo Bonzini [Thu, 9 Feb 2012 09:12:21 +0000 (10:12 +0100)]
qdev: drop unnecessary parse/print methods

More qdev printers could have been removed in the previous series, and
object_property_parse also made several parsers unnecessary.  In fact,
the new code is even more robust with respect to overflows, so clean
them up!

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqdev: use built-in QOM string parser
Paolo Bonzini [Thu, 9 Feb 2012 09:12:08 +0000 (10:12 +0100)]
qdev: use built-in QOM string parser

object_property_parse lets us drop the legacy setters when their task
is done just as well by the string visitors.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqdev: accept hex properties only if prefixed by 0x
Paolo Bonzini [Thu, 9 Feb 2012 09:07:38 +0000 (10:07 +0100)]
qdev: accept hex properties only if prefixed by 0x

Hex properties are an obstacle to removal of old qdev string parsing, but
even here we can lay down the foundations for future simplification.  In
general, they are rarely used and their printed form is more interesting
than the parsing.  For example you'd usually set isa-serial.index
instead of isa-serial.iobase.  And luckily our main client, libvirt
only cares about few of these, and always sets them with a 0x prefix.
So the series stops accepting bare hexadecimal numbers, preparing for
making legacy properties read-only in 1.3 or so.  The read side will
stay as long as "info qtree" is with us.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqdev: accept both strings and integers for PCI addresses
Paolo Bonzini [Thu, 9 Feb 2012 08:53:32 +0000 (09:53 +0100)]
qdev: accept both strings and integers for PCI addresses

Visitors allow a limited form of polymorphism.  Exploit it to support
setting the non-legacy PCI address property both as a DD.F string
and as an 8-bit integer.

The 8-bit integer form is just too clumsy, it is unlikely that we will
ever drop it.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqom: add generic string parsing/printing
Paolo Bonzini [Thu, 9 Feb 2012 08:52:59 +0000 (09:52 +0100)]
qom: add generic string parsing/printing

Add generic property accessors that take a string and parse it
appropriately for the property type.  All the magic here is done
by the new string-based visitors.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqxl: add user-friendly bar size properties
Gerd Hoffmann [Fri, 17 Feb 2012 14:03:24 +0000 (15:03 +0100)]
qxl: add user-friendly bar size properties

Add two properties to specify bar sizes in megabytes instead of bytes,
which is alot more user-friendly.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoqxl: move ram size init to new function
Gerd Hoffmann [Fri, 17 Feb 2012 14:02:40 +0000 (15:02 +0100)]
qxl: move ram size init to new function

Factor memory bar sizing bits out to a separate function.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoqxl: drop vram bar minimum size
Gerd Hoffmann [Fri, 17 Feb 2012 13:40:01 +0000 (14:40 +0100)]
qxl: drop vram bar minimum size

There is no reason to require a minimum size of 16 MB for the vram.
Lower the limit to 4096 (one page).  Make it disapper completely would
break guests.

12 years agospice: support ipv6 channel address in monitor events and in spice info
Yonit Halperin [Wed, 8 Feb 2012 13:40:15 +0000 (15:40 +0200)]
spice: support ipv6 channel address in monitor events and in spice info

RHBZ #788444

CC: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Yonit Halperin <yhalperi@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoAdd SPICE support to add_client monitor command
Daniel P. Berrange [Mon, 13 Feb 2012 13:43:08 +0000 (13:43 +0000)]
Add SPICE support to add_client monitor command

With the acceptance of some new APIs to libspice-server.so it
is possible to add support for SPICE to the 'add_client'
monitor command, bringing parity with VNC. Since SPICE can
use TLS or plain connections, the command also gains a new
'tls' parameter to specify whether TLS should be attempted
on the injected client sockets.

This new feature is only enabled if building against a
libspice-server >= 0.10.1

* qmp-commands.hx: Add 'tls' parameter & missing doc for
  'skipauth' parameter
* monitor.c: Wire up SPICE for 'add_client' command
* ui/qemu-spice.h, ui/spice-core.c: Add qemu_spice_display_add_client
  API to wire up from monitor

[1] http://cgit.freedesktop.org/spice/spice/commit/server/spice.h?id=d55b68b6b44f2499278fa860fb47ff22f5011faa
    http://cgit.freedesktop.org/spice/spice/commit/server/spice.h?id=bd07dde530d9504e1cfe7ed5837fc00c26f36716

Changes in v3:
 - Added 'optional' flag to new parameters documented
 - Added no-op impl of qemu_spice_display_add_client when
   SPICE is disabled during build

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoqxl: make sure primary surface is saved on migration also in compat mode
Yonit Halperin [Wed, 15 Feb 2012 09:22:16 +0000 (11:22 +0200)]
qxl: make sure primary surface is saved on migration also in compat mode

RHBZ #790083

Signed-off-by: Yonit Halperin <yhalperi@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoqxl: set only off-screen surfaces dirty instead of the whole vram
Yonit Halperin [Wed, 15 Feb 2012 09:22:15 +0000 (11:22 +0200)]
qxl: set only off-screen surfaces dirty instead of the whole vram

We used to assure the guest surfaces were saved before migration by
setting the whole vram dirty. This patch sets dirty only the areas
that are actually used in the vram.

Signed-off-by: Yonit Halperin <yhalperi@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoqxl: don't render stuff when the vm is stopped.
Gerd Hoffmann [Wed, 15 Feb 2012 13:04:44 +0000 (14:04 +0100)]
qxl: don't render stuff when the vm is stopped.

This patch fixes the local qxl renderer to not kick spice-server
in case the vm is stopped.  First it is largely pointless because
we ask spice-server to process all not-yet processed commands when
the vm is stopped, so there isn't much do do anyway.  Second we
avoid triggering an assert in spice-server.

The patch makes sure we still honor redraw requests, even if we don't
ask spice-server for updates.  This is needed to handle displaysurface
changes with a stopped vm correctly.

With this patch applied it is possible to take screen shots (via
screendump monitor command) from a qxl gpu even in case the guest
is stopped.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoqxl: fix warnings on 32bit
Gerd Hoffmann [Wed, 8 Feb 2012 14:58:35 +0000 (15:58 +0100)]
qxl: fix warnings on 32bit

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
12 years agoqapi: add tests for string-based visitors
Paolo Bonzini [Thu, 9 Feb 2012 10:21:03 +0000 (11:21 +0100)]
qapi: add tests for string-based visitors

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqapi: add string-based visitors
Paolo Bonzini [Thu, 9 Feb 2012 08:36:37 +0000 (09:36 +0100)]
qapi: add string-based visitors

String based visitors provide a consistent interface for parsing
strings to C values, as well as consuming C values as strings.
They will be used to parse command-line options.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqapi: drop qmp_input_end_optional
Paolo Bonzini [Thu, 9 Feb 2012 08:14:00 +0000 (09:14 +0100)]
qapi: drop qmp_input_end_optional

This method is optional, do not implement it if it is empty.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoqapi: allow sharing enum implementation across visitors
Paolo Bonzini [Thu, 9 Feb 2012 08:11:52 +0000 (09:11 +0100)]
qapi: allow sharing enum implementation across visitors

Most visitors will use the same code for enum parsing.  Move it to
the core.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoinput: send kbd+mouse events only to running guests.
Gerd Hoffmann [Wed, 15 Feb 2012 08:15:37 +0000 (09:15 +0100)]
input: send kbd+mouse events only to running guests.

Trying to interact with a stopped guest will queue up the events,
then send them all at once when the guest continues running, with
a high chance to have them cause unwanted actions.

Avoid that by only injecting the input events only when the guest
is in running state.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoi8259: Do not clear level-triggered lines in IRR on init
Jan Kiszka [Tue, 24 Jan 2012 15:29:29 +0000 (16:29 +0100)]
i8259: Do not clear level-triggered lines in IRR on init

When an input line is handled as level-triggered, it will immediately
raise an IRQ on the output of a PIC again that goes through an init
reset. So only clear the edge-triggered inputs from IRR in that
scenario.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoi8254: Factor out pit_get_channel_info
Jan Kiszka [Wed, 1 Feb 2012 19:31:43 +0000 (20:31 +0100)]
i8254: Factor out pit_get_channel_info

Instead of providing 4 individual query functions for mode, gate, output
and initial counter state, introduce a service that queries all
information at once. This comes with tiny additional costs for
pcspk_callback but with a much cleaner interface. Also, it will simplify
the implementation of the KVM in-kernel PIT model.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agopcspk: Convert to qdev
Jan Kiszka [Fri, 17 Feb 2012 10:24:34 +0000 (11:24 +0100)]
pcspk: Convert to qdev

Convert the PC speaker device to a qdev ISA model. Move the public
interface to a dedicated header file at this chance.

CC: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoi8254: Rework & fix interaction with HPET in legacy mode
Jan Kiszka [Wed, 1 Feb 2012 19:31:41 +0000 (20:31 +0100)]
i8254: Rework & fix interaction with HPET in legacy mode

When the HPET enters legacy mode, the IRQ output of the PIT is
suppressed and replaced by the HPET timer 0. But the current code to
emulate this was broken in many ways. It reset the PIT state after
re-enabling, it worked against a stale static PIT structure, and it did
not properly saved/restored the IRQ output mask in the PIT vmstate.

This patch solves the PIT IRQ control in a different way. On x86, it
both redirects the PIT IRQ to the HPET, just like the RTC. But it also
keeps the control line from the HPET to the PIT. This allows to disable
the PIT QEMU timer when it is not needed. The PIT's view on the control
line state is now saved in the same format that qemu-kvm is already
using.

Note that, in contrast to the suppressed RTC IRQ line, we do not need to
save/restore the PIT line state in the HPET. As we trigger a PIT IRQ
update via the control line, the line state is reconstructed on mode
switch.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoi8254: Pass alternative IRQ output object on initialization
Jan Kiszka [Wed, 1 Feb 2012 19:31:40 +0000 (20:31 +0100)]
i8254: Pass alternative IRQ output object on initialization

HPET legacy emulation will require control over the PIT IRQ output. To
enable this, add support for an alternative IRQ output object to the PIT
factory function. If the isa_irq number is < 0, this object will be
used.

This also removes the IRQ number property from the PIT class as we now
use a generic GPIO output pin that is connected by the factory function.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoi8254: Factor out interface header
Jan Kiszka [Wed, 1 Feb 2012 19:31:39 +0000 (20:31 +0100)]
i8254: Factor out interface header

Move the public interface of the PIT into its own header file and update
all users.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agohpet: Save/restore cached RTC IRQ level
Jan Kiszka [Wed, 1 Feb 2012 19:31:38 +0000 (20:31 +0100)]
hpet: Save/restore cached RTC IRQ level

In legacy mode, the HPET suppresses the RTC interrupt delivery via IRQ
8 but keeps track of the RTC output level and applies it when legacy
mode is turned off again. This value has to be preserved across save/
restore as it cannot be reconstructed otherwise.

To document that a raised rtc_irq_level won't survive a vmload without
a hpet/rtc_irq_level subsection, add an explicit clearing to the reset
handler.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoi8254: Do not raise IRQ level on reset
Jan Kiszka [Wed, 1 Feb 2012 19:31:37 +0000 (20:31 +0100)]
i8254: Do not raise IRQ level on reset

Avoid changing the IRQ level to high on reset as it may trigger spurious
events. Instead, open-code the effects of pit_load_count(0) in the reset
handler.

Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agovnc: Don't demote authentication scheme when changing password/disabling login
Daniel P. Berrange [Tue, 14 Feb 2012 12:37:29 +0000 (12:37 +0000)]
vnc: Don't demote authentication scheme when changing password/disabling login

Currently when disabling login in VNC, the password is cleared out and the
authentication protocol is forced to AUTH_VNC.  If you're using a stronger
authentication protocol, this has the effect of downgrading your security
protocol.

Fix this by only changing the authentication protocol if the current
authentication protocol is AUTH_NONE.  That ensures we're never downgrading.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
--
NB. This patch is derived from one posted by Anthony last year, which got
accidentally lost after Luiz took over the QMP series work

  https://lists.gnu.org/archive/html/qemu-devel/2011-09/msg00392.html

 v1 -> v2
 - Make sure to not demote when changing password (Daniel)
 v2 -> v3
 - Rebase to latest GIT master wrt QMP changes

12 years agoqdev: Fix qdev_try_create() semantics
Andreas Färber [Fri, 17 Feb 2012 01:47:44 +0000 (02:47 +0100)]
qdev: Fix qdev_try_create() semantics

Since QOM'ification, qdev_try_create() uses object_new() internally,
which asserts "type != NULL" when the type is not registered.
This was revealed by the combination of kvmclock's kvm_enabled() check
and early QOM type registration.

Check whether the class exists before calling object_new(), so that
the caller (e.g., qdev_create) can fail gracefully, telling us which
device could not be created.

Signed-off-by: Andreas Färber <afaerber@suse.de>
Cc: Anthony Liguori <aliguori@codemonkey.ws>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqom: Fix identifiers in documentation
Andreas Färber [Thu, 16 Feb 2012 17:03:19 +0000 (18:03 +0100)]
qom: Fix identifiers in documentation

Fixes gtk-doc warnings.

Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqom: Fix typo in Object's documentation
Andreas Färber [Thu, 16 Feb 2012 17:03:18 +0000 (18:03 +0100)]
qom: Fix typo in Object's documentation

Fixes a warning from gtk-doc.

Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoconfigure: Remove OpenBSD workaround for curses probe
Brad Smith [Fri, 10 Feb 2012 20:59:38 +0000 (15:59 -0500)]
configure: Remove OpenBSD workaround for curses probe

Remove the OpenBSD workaround for the curses probe. This has not been
necessary for 5 releases now.

Signed-off-by: Brad Smith <brad@comstyle.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agolibcacard configure fixes
Paul Brook [Thu, 9 Feb 2012 19:05:29 +0000 (19:05 +0000)]
libcacard configure fixes

libcacard is only used by system emulation.
Only define libcacard_libs/cflags once.

Signed-off-by: Paul Brook <paul@codesourcery.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agocoroutine: switch to QSLIST
Paolo Bonzini [Fri, 13 Jan 2012 16:34:04 +0000 (17:34 +0100)]
coroutine: switch to QSLIST

QSLIST can be used for a free list, do it.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqemu-queue: drop QCIRCLEQ
Paolo Bonzini [Fri, 13 Jan 2012 16:34:03 +0000 (17:34 +0100)]
qemu-queue: drop QCIRCLEQ

The main advantage of circular lists (the fact that the head node
has the same memory layout as any other node) is completely negated
by the implementation in qemu-queue.h.  Not surprisingly, nobody
uses QCIRCLEQ.  While this might change if RCU is ever adopted by
QEMU, the QLIST is also RCU-friendly and in fact it is used in a
RCU-like manner by 9pfs already.  So, just kill QCIRCLEQ.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqemu-queue: add QSLIST
Paolo Bonzini [Fri, 13 Jan 2012 16:34:02 +0000 (17:34 +0100)]
qemu-queue: add QSLIST

Based on http://cvsweb.netbsd.org/bsdweb.cgi/src/sys/sys/queue.h?rev=1.53
with only the prefix change.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agonotifier: switch to QLIST
Paolo Bonzini [Fri, 13 Jan 2012 16:34:01 +0000 (17:34 +0100)]
notifier: switch to QLIST

Notifiers do not need to access both ends of the list, and using
a QLIST also simplifies the API.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoMerge remote-tracking branch 'bonzini/nbd-for-anthony' into staging
Anthony Liguori [Fri, 17 Feb 2012 12:50:58 +0000 (06:50 -0600)]
Merge remote-tracking branch 'bonzini/nbd-for-anthony' into staging

* bonzini/nbd-for-anthony:
  nbd: add git tree to MAINTAINERS
  open /dev/nbd in nbd_client_thread
  do not chdir(/) in qemu-nbd before opening all files

12 years agoMerge remote-tracking branch 'pmaydell/arm-devs.for-upstream' into staging
Anthony Liguori [Fri, 17 Feb 2012 12:50:07 +0000 (06:50 -0600)]
Merge remote-tracking branch 'pmaydell/arm-devs.for-upstream' into staging

* pmaydell/arm-devs.for-upstream: (22 commits)
  hw/pl031: Actually raise interrupt on timer expiry
  MAINTAINERS: Add hw/highbank.c maintainer
  Remove unnecessary includes of primecell.h
  hw/primecell.h: Remove obsolete pl080_init() declaration
  hw/arm_sysctl: Drop legacy init function
  hw/vexpress.c: Add vexpress-a15 machine
  arm_boot: Pass base address of GIC CPU interface, not whole GIC
  hw/vexpress.c: Instantiate the motherboard CLCD
  hw/vexpress.c: Factor out daughterboard-specific initialization
  hw/vexpress.c: Move secondary CPU boot code to SRAM
  hw/vexpress.c: Make motherboard peripheral memory map table-driven
  hw/a15mpcore.c: Add Cortex-A15 private peripheral model
  MAINTAINERS: Add maintainers for Exynos SOC.
  Exynos4210: added display controller implementation
  hw/exynos4210.c: Add LAN support for SMDKC210.
  hw/lan9118: Add basic 16-bit mode support.
  ARM: exynos4210: MCT support.
  ARM: exynos4210: basic Power Management Unit implementation
  ARM: exynos4210: PWM support.
  ARM: exynos4210: UART support
  ...

12 years agoMerge remote-tracking branch 'stefanha/trivial-patches' into staging
Anthony Liguori [Fri, 17 Feb 2012 12:48:47 +0000 (06:48 -0600)]
Merge remote-tracking branch 'stefanha/trivial-patches' into staging

* stefanha/trivial-patches:
  linux-user: brk() debugging
  virtio: Remove unneeded g_free() check in virtio_cleanup()
  net: remove extra spaces in help messages
  fmopl: Fix typo in function name
  vl.c: Fix typo in variable name
  ide: fix compilation errors when DEBUG_IDE is set
  cpu-exec.c: Correct comment about this file and indentation cleanup
  CODING_STYLE: Clarify style for enum and function type names
  linux-user: fail execve() if env/args too big

12 years agonbd: add git tree to MAINTAINERS
Paolo Bonzini [Wed, 18 Jan 2012 11:48:19 +0000 (12:48 +0100)]
nbd: add git tree to MAINTAINERS

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agoopen /dev/nbd in nbd_client_thread
Paolo Bonzini [Tue, 6 Dec 2011 08:07:00 +0000 (09:07 +0100)]
open /dev/nbd in nbd_client_thread

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agodo not chdir(/) in qemu-nbd before opening all files
Michael Tokarev [Mon, 16 Jan 2012 14:37:44 +0000 (18:37 +0400)]
do not chdir(/) in qemu-nbd before opening all files

When qemu-nbd becomes a daemon it calls daemon(3) with
nochdir=0, so daemon(3) changes current directory to /.
But at this time, qemu-nbd did not open any user-specified
files yet, so by changing current directory, all non-absolute
paths becomes wrong.  The solution is to pass nochdir=1 to
daemon(3) function, and to chdir("/") after all init has
been performed, before entering the main loop, -- just like
a good daemon should do.

This patch is applicable for -stable.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
12 years agohw/pl031: Actually raise interrupt on timer expiry
Peter Maydell [Thu, 16 Feb 2012 09:56:10 +0000 (09:56 +0000)]
hw/pl031: Actually raise interrupt on timer expiry

Fix a typo in pl031_interrupt() which meant we were setting a bit
in the interrupt mask rather than the interrupt status register
and thus not actually raising an interrupt. This fix allows the
rtctest program from the kernel's Documentation/rtc.txt to pass
rather than hanging.

Reported-by: Daniel Forsgren <daniel.forsgren@enea.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Andreas Färber <afaerber@suse.de>
12 years agoMAINTAINERS: Add hw/highbank.c maintainer
Mark Langsdorf [Thu, 16 Feb 2012 09:56:10 +0000 (09:56 +0000)]
MAINTAINERS: Add hw/highbank.c maintainer

Signed-off-by: Mark Langsdorf <mark.langsdorf@calxeda.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoRemove unnecessary includes of primecell.h
Peter Maydell [Thu, 16 Feb 2012 09:56:10 +0000 (09:56 +0000)]
Remove unnecessary includes of primecell.h

The primecell.h header now only has the definitions of constants
indicating the usage of the arm_sysctl GPIO lines; remove obsolete
includes of it from source files which don't care about those GPIO
lines.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agohw/primecell.h: Remove obsolete pl080_init() declaration
Peter Maydell [Thu, 16 Feb 2012 09:56:10 +0000 (09:56 +0000)]
hw/primecell.h: Remove obsolete pl080_init() declaration

Remove an obsolete declaration of pl080_init(), which has been
incorrect since the conversion of pl080 to qdev back in 2009.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agohw/arm_sysctl: Drop legacy init function
Peter Maydell [Thu, 16 Feb 2012 09:56:09 +0000 (09:56 +0000)]
hw/arm_sysctl: Drop legacy init function

Drop the legacy init function arm_sysctl_init(), since it has no
users left any more. This allows us to drop the awkward '1' from
the actual device init function.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Andreas Färber <afaerber@suse.de>
12 years agohw/vexpress.c: Add vexpress-a15 machine
Peter Maydell [Thu, 16 Feb 2012 09:56:09 +0000 (09:56 +0000)]
hw/vexpress.c: Add vexpress-a15 machine

Add the vexpress-a15 machine, and the A-Series memory map it uses.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Andreas Färber <afaerber@suse.de>
12 years agoarm_boot: Pass base address of GIC CPU interface, not whole GIC
Peter Maydell [Thu, 16 Feb 2012 09:56:09 +0000 (09:56 +0000)]
arm_boot: Pass base address of GIC CPU interface, not whole GIC

The arm_boot secondary boot loader code needs the address of
the GIC CPU interface. Obtaining this from the base address
of the private peripheral region was possible for A9 and 11MPcore,
but the A15 puts the GIC CPU interface in a different place.
So make boards pass in the GIC CPU interface address directly.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agohw/vexpress.c: Instantiate the motherboard CLCD
Peter Maydell [Thu, 16 Feb 2012 09:56:09 +0000 (09:56 +0000)]
hw/vexpress.c: Instantiate the motherboard CLCD

Instantiate the CLCD on the vexpress motherboard as well as one on
the daughterboard -- the A15 daughterboard does not have a CLCD
and so relies on the motherboard one.

At the moment QEMU doesn't provide infrastructure for selecting
which display device gets to actually show graphics -- the first
one registered is it. Fortunately this works for the major use
case (Linux): if the daughterboard has a CLCD it will come first
and be used, otherwise we fall back to the motherboard CLCD.
So we don't (currently) need to implement the control register
which allows software to tell the mux which video output to pass
through to the outside world.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Andreas Färber <afaerber@suse.de>
12 years agohw/vexpress.c: Factor out daughterboard-specific initialization
Peter Maydell [Thu, 16 Feb 2012 09:56:09 +0000 (09:56 +0000)]
hw/vexpress.c: Factor out daughterboard-specific initialization

Factor out daughterboard specifics into a data structure and
daughterboard initialization function, in preparation for adding
vexpress-a15 support.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agohw/vexpress.c: Move secondary CPU boot code to SRAM
Peter Maydell [Thu, 16 Feb 2012 09:56:08 +0000 (09:56 +0000)]
hw/vexpress.c: Move secondary CPU boot code to SRAM

On real Versatile Express hardware, the boot ROM puts the secondary
CPU bootcode/holding pen in SRAM. We can therefore rely on Linux not
trashing this memory until secondary CPUs have booted up, and can
put our QEMU-specific pen code in the same place. This allows us to
drop the odd "hack" RAM page we were using before.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agohw/vexpress.c: Make motherboard peripheral memory map table-driven
Peter Maydell [Thu, 16 Feb 2012 09:56:08 +0000 (09:56 +0000)]
hw/vexpress.c: Make motherboard peripheral memory map table-driven

Pull the addresses used for mapping motherboard peripherals into
memory out into a table. This will allow us to simply provide a
second table to implement the "Cortex-A Series" memory map used by
the A15 variant of Versatile Express, as well as the current
"Legacy" map used by A9.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Andreas Färber <afaerber@suse.de>
12 years agohw/a15mpcore.c: Add Cortex-A15 private peripheral model
Peter Maydell [Thu, 16 Feb 2012 09:56:07 +0000 (09:56 +0000)]
hw/a15mpcore.c: Add Cortex-A15 private peripheral model

Add a model of the Cortex-A15 memory mapped private peripheral
space. This is fairly simple because the only memory mapped
bit of the A15 is the GIC.

Note that we don't currently model a VGIC and therefore don't
map the VGIC related bits of the GIC.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoMAINTAINERS: Add maintainers for Exynos SOC.
Evgeny Voevodin [Thu, 16 Feb 2012 09:56:06 +0000 (09:56 +0000)]
MAINTAINERS: Add maintainers for Exynos SOC.

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Acked-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoExynos4210: added display controller implementation
Mitsyanko Igor [Thu, 16 Feb 2012 09:56:06 +0000 (09:56 +0000)]
Exynos4210: added display controller implementation

Exynos4210 display controller (FIMD) has 5 hardware windows with alpha and
chroma key blending functions.

Signed-off-by: Mitsyanko Igor <i.mitsyanko@samsung.com>
Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agohw/exynos4210.c: Add LAN support for SMDKC210.
Evgeny Voevodin [Thu, 16 Feb 2012 09:56:06 +0000 (09:56 +0000)]
hw/exynos4210.c: Add LAN support for SMDKC210.

SMDKC210 uses lan9215 chip, but lan9118 in 16-bit mode seems to
be enough.

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agohw/lan9118: Add basic 16-bit mode support.
Evgeny Voevodin [Thu, 16 Feb 2012 09:56:06 +0000 (09:56 +0000)]
hw/lan9118: Add basic 16-bit mode support.

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoARM: exynos4210: MCT support.
Evgeny Voevodin [Thu, 16 Feb 2012 09:56:05 +0000 (09:56 +0000)]
ARM: exynos4210: MCT support.

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoARM: exynos4210: basic Power Management Unit implementation
Maksim Kozlov [Thu, 16 Feb 2012 09:56:05 +0000 (09:56 +0000)]
ARM: exynos4210: basic Power Management Unit implementation

Patch adds basic model for Exynos4210 SoC PMU.
This model implements PMU registers just as a bulk of memory. Currently,
the only reason this device exists is that secondary CPU boot loader
uses PMU INFORM5 register as a holding pen.

Signed-off-by: Maksim Kozlov <m.kozlov@samsung.com>
Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoARM: exynos4210: PWM support.
Evgeny Voevodin [Thu, 16 Feb 2012 09:56:05 +0000 (09:56 +0000)]
ARM: exynos4210: PWM support.

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoARM: exynos4210: UART support
Maksim Kozlov [Thu, 16 Feb 2012 09:56:05 +0000 (09:56 +0000)]
ARM: exynos4210: UART support

Add basic support of exynos4210 UART

Signed-off-by: Maksim Kozlov <m.kozlov@samsung.com>
Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoARM: Samsung exynos4210-based boards emulation
Evgeny Voevodin [Thu, 16 Feb 2012 09:56:05 +0000 (09:56 +0000)]
ARM: Samsung exynos4210-based boards emulation

Add initial support of NURI and SMDKC210 boards

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoARM: exynos4210: IRQ subsystem support.
Evgeny Voevodin [Thu, 16 Feb 2012 09:56:04 +0000 (09:56 +0000)]
ARM: exynos4210: IRQ subsystem support.

Signed-off-by: Evgeny Voevodin <e.voevodin@samsung.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
12 years agoMake -machine/-enable-kvm options merge into a single list
Peter Maydell [Wed, 8 Feb 2012 05:41:38 +0000 (05:41 +0000)]
Make -machine/-enable-kvm options merge into a single list

Make the "machine" option list use list merging, so that multiple
-machine arguments (and the -enable-kvm argument) all merge together
into a single list. Drop the calls to qemu_opts_reset() which meant
that only the last -machine or -enable-kvm option had any effect.

This fixes the bug where "-enable-kvm -machine foo" would ignore
the '-enable-kvm' option, and "-machine foo -enable-kvm" would
ignore the '-machine foo' option.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
12 years agoqemu-option: Add support for merged QemuOptsLists
Peter Maydell [Wed, 8 Feb 2012 05:41:37 +0000 (05:41 +0000)]
qemu-option: Add support for merged QemuOptsLists

Add support for option lists which are merged together, so that
"-listname foo=bar -listname bar=baz" is equivalent to "-listname
foo=bar,bar=baz" rather than generating two separate lists of options.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
12 years agotarget-arm/helper.c: tb_flush() on CPU reset
Peter Maydell [Wed, 1 Feb 2012 17:23:04 +0000 (17:23 +0000)]
target-arm/helper.c: tb_flush() on CPU reset

Since target-arm has some CPUState fields for which we take the approach
of baking assumptions about them into translated code and then calling
tb_flush() when the fields change, we must also tb_flush on CPU reset,
because reset is a change of those fields.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
12 years agohw/arm: Remove redundant arguments from set_kernel_args*
Stefan Weil [Sun, 29 Jan 2012 07:52:15 +0000 (08:52 +0100)]
hw/arm: Remove redundant arguments from set_kernel_args*

The parameters initrd_size and base are already included
in the info parameter, so there is no need to pass them
separately.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>,
12 years agopxa2xx_lcd: SRAM is valid location for the framebuffer
Vasily Khoruzhick [Tue, 24 Jan 2012 19:32:30 +0000 (22:32 +0300)]
pxa2xx_lcd: SRAM is valid location for the framebuffer

Signed-off-by: Vasily Khoruzhick <anarsoul@gmail.com>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
12 years agotarget-arm/helper.c: Correct FPSID value for Cortex-A9
Peter Maydell [Tue, 7 Feb 2012 17:56:57 +0000 (17:56 +0000)]
target-arm/helper.c: Correct FPSID value for Cortex-A9

The correct FPSID for the Cortex-A9 (according to the TRM) is
0x41033090 for the r0p0 that we claim to model.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
12 years agonseries: attach monitor powerdown request to menelaus
Paolo Bonzini [Fri, 20 Jan 2012 11:10:34 +0000 (12:10 +0100)]
nseries: attach monitor powerdown request to menelaus

I noticed some unused code in the twl92230, probably from before
qdev-ification.  This patch makes the machine use the chip's pwrbtn
signal.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
12 years agoMerge remote-tracking branch 'qemu-kvm/uq/master' into staging
Anthony Liguori [Thu, 16 Feb 2012 00:41:28 +0000 (18:41 -0600)]
Merge remote-tracking branch 'qemu-kvm/uq/master' into staging

* qemu-kvm/uq/master:
  apic: Fix legacy vmstate loading for KVM
  kvm: Implement kvm_irqchip_in_kernel like kvm_enabled
  kvm: Allow to set shadow MMU size

12 years agoMerge remote-tracking branch 'kraxel/vnc.2' into staging
Anthony Liguori [Thu, 16 Feb 2012 00:40:53 +0000 (18:40 -0600)]
Merge remote-tracking branch 'kraxel/vnc.2' into staging

* kraxel/vnc.2:
  vnc: lift modifier keys on client disconnect.
  vnc: implement shared flag handling.
  vnc: fix ctrl key in vnc terminal emulation
  Fix vnc memory corruption with width = 1400

12 years agoMerge remote-tracking branch 'kiszka/queues/slirp' into staging
Anthony Liguori [Thu, 16 Feb 2012 00:40:26 +0000 (18:40 -0600)]
Merge remote-tracking branch 'kiszka/queues/slirp' into staging

* kiszka/queues/slirp:
  slirp: Prevent sending ICMP error replies to source-only addresses
  slirp: Remove unused variable and unused code

12 years agoMerge remote-tracking branch 'kraxel/usb.38' into staging
Anthony Liguori [Wed, 15 Feb 2012 23:25:25 +0000 (17:25 -0600)]
Merge remote-tracking branch 'kraxel/usb.38' into staging

* kraxel/usb.38: (28 commits)
  xhci: handle USB_RET_NAK
  xhci: remote wakeup support
  xhci: kill port arg from xhci_setup_packet
  xhci: stop on errors
  xhci: add trb type name lookup support.
  xhci: signal low- and fullspeed support
  usb: add USBBusOps->wakeup_endpoint
  usb: pass USBEndpoint to usb_wakeup
  usb: maintain async packet list per endpoint
  usb: Set USBEndpoint in usb_packet_setup().
  usb: add USBEndpoint->{nr,pid}
  usb: USBPacket: add status, rename owner -> ep
  usb: fold usb_generic_handle_packet into usb_handle_packet
  usb: kill handle_packet callback
  usb-xhci: switch to usb_find_device()
  usb-musb: switch to usb_find_device()
  usb-ohci: switch to usb_find_device()
  usb-ehci: switch to usb_find_device()
  usb-uhci: switch to usb_find_device()
  usb: handle dev == NULL in usb_handle_packet()
  ...

12 years agoMerge remote-tracking branch 'kwolf/for-anthony' into staging
Anthony Liguori [Wed, 15 Feb 2012 23:18:04 +0000 (17:18 -0600)]
Merge remote-tracking branch 'kwolf/for-anthony' into staging

* kwolf/for-anthony:
  AHCI: Masking of IRQs actually masks them
  sheepdog: fix co_recv coroutine context
  AHCI: Fix port reset race
  rewrite QEMU_BUILD_BUG_ON
  qcow2: Keep unknown header extension when rewriting header
  qcow2: Update whole header at once
  vpc: Round up image size during fixed image creation
  vpc: Add support for Fixed Disk type
  iSCSI: add configuration variables for iSCSI
  qemu-io: add write -z option for bdrv_co_write_zeroes
  qed: add .bdrv_co_write_zeroes() support
  qed: replace is_write with flags field
  block: perform zero-detection during copy-on-read
  block: add .bdrv_co_write_zeroes() interface
  cutils: extract buffer_is_zero() from qemu-img.c

12 years agodevice_add: don't add a /peripheral link until init is complete
Anthony Liguori [Sun, 12 Feb 2012 17:36:24 +0000 (11:36 -0600)]
device_add: don't add a /peripheral link until init is complete

Otherwise we end up with a dangling reference which causes qdev_free() to fail.

Reported-by: Michael Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
12 years agoqom: Unify type registration
Andreas Färber [Thu, 9 Feb 2012 14:20:55 +0000 (15:20 +0100)]
qom: Unify type registration

Replace device_init() with generalized type_init().

While at it, unify naming convention: type_init([$prefix_]register_types)
Also, type_init() is a function, so add preceding blank line where
necessary and don't put a semicolon after the closing brace.

Signed-off-by: Andreas Färber <afaerber@suse.de>
Cc: Anthony Liguori <anthony@codemonkey.ws>
Cc: malc <av1474@comtv.ru>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>