Drop duplicated code.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Drop the linux-specific virtio headers, use the copy from
standard-headers instead.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Drop duplicate code.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Drop duplicated code.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Acked-by: NPaolo Bonzini <pbonzini@redhat.com>

Drop duplicated code.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>

Drop duplicated code.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>

Drop duplicated code. Minor codechanges were required
as geometry is a sub-structure now.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>

Drop code duplicated from standard headers.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>

Drop code duplicated from standard headers.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>

Drop a bunch of code duplicated from virtio_config.h and virtio_ring.h.
This makes us rename event index accessors which conflict,
as reusing the ones from virtio_ring.h isn't trivial.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>

Switch to virtio_ring.h from standard headers.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Add files imported from linux-next (what will become linux 4.0) using
scripts/update-linux-headers.sh
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

It doesn't make sense to copy values manually:
the only issue with getting headers from linux
seems to be dealing with linux/types, we
can easily fix that automatically while importing.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NThomas Huth <thuth@linux.vnet.ibm.com>

For legacy machine types, rsdp is not in RAM, so we need a copy of rsdp
for fw cfg. We previously used g_array_free with false parameter,
but this seems to confuse people.
This also wastes a bit of memory as the buffer is unused for new
machine types.

Let's just use plain g_memdup, and free original memory together with
the array.

TODO: rationalize tcpalog memory management, and get rid of the mfre
parameter.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>

As comment in acpi-build.c notes, RSDP is not really immutable.  So it's
really a question of whether it's in RAM, name the variable accordingly.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>

This fixes multiple issues around ACPI RAM management:

RSDP and linker RAM aren't currently marked dirty
on update, so they won't be migrated correctly.

Let's handle all tables in the same way: set correct size (assert if
too big), update, mark RAM dirty.

This also drops assert checking that table size didn't change: table
size is fundamentally dynamic and depends on hw configuration,
just set the correct size and use that (memory core asserts if size is
too large).

This also means we can drop tracking table size, memory core does this
for us now.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Block size must fundamentally be a multiple of target page size.
Aligning automatically removes need to worry about the alignment
from callers.

Note: the only caller of qemu_ram_resize (acpi) already happens to have
size padded to a power of 2, but we would like to drop the padding in
ACPI core, and don't want to expose target page size knowledge to ACPI.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Acked-by: NPaolo Bonzini <ponzini@redhat.com>

Makes sure that RSDP stays the same
/i.e. matches ACPI tables blob in source/
if guest is migrated during RSDP reading or
has been already shadowed by firmware.

Fix applies only to new machine types starting
from 2.3, so it won't break migration for old
machine types.
Signed-off-by: NIgor Mammedov <imammedo@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NMarcel Apfelbaum <marcel@redhat.com>

Linker table is build only once, so if later during
tables rebuild sizes of other ACPI tables change
pointers will be patched incorrectly due to wrong
offsets in linker. Resulting in guest not being able
to find ACPI tables.
Fix it by updating 'linker' table with the rest of
tables when firmware reads it.
Signed-off-by: NIgor Mammedov <imammedo@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NMarcel Apfelbaum <marcel@redhat.com>

RSDT offset can change across reboots and that makes
immutable RSDP, which is build at startup, point to
incorrect place in ACPI table blob. That results in
BIOS corrupting tables and guest OS failing to find
ACPI tables.
We really should have put it in a ROM region, but
we can't change that for old machine types,
let's just set the callback and update it explicitly.
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NIgor Mammedov <imammedo@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NMarcel Apfelbaum <marcel@redhat.com>

If the maxram_size is not aligned and dimm devices were added on the
command line qemu would terminate with a rather unhelpful message:

ERROR:hw/mem/pc-dimm.c:150:pc_dimm_get_free_addr: assertion failed:
(QEMU_ALIGN_UP(address_space_size, align) == address_space_size)

In case no dimm device was originally added on the commandline qemu
exits on the assertion failure.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Acked-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>

Produce more human readable error messages and fix few spelling
mistakes.

Also remove a redundant check for the max memory size.
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>
Acked-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NPeter Krempa <pkrempa@redhat.com>

Memory and CPU hot unplug are both asynchronous procedures.
When the unplug operation happens, unplug request cb is called first.
And when guest OS finished handling unplug, unplug cb will be called
to do the real removal of device.

This patch adds hotunplug cb to piix4, which memory and CPU
hot unplug will use it.
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>
Signed-off-by: NTang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: NZhu Guihua <zhugh.fnst@cn.fujitsu.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Memory and CPU hot unplug are both asynchronous procedures.
When the unplug operation happens, unplug request cb is called first.
And when guest OS finished handling unplug, unplug cb will be called
to do the real removal of device.

This patch adds hotunplug cb to ich9, which memory and CPU
hot unplug will use it.
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>
Signed-off-by: NTang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: NZhu Guihua <zhugh.fnst@cn.fujitsu.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Memory and CPU hot unplug are both asynchronous procedures.
When the unplug operation happens, unplug request cb is called first.
And when guest OS finished handling unplug, unplug cb will be called
to do the real removal of device.

This patch adds hotunplug cb to pc machine, which memory and CPU
hot unplug will use it.
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>
Signed-off-by: NTang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: NZhu Guihua <zhugh.fnst@cn.fujitsu.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Memory and CPU hot unplug are both asynchronous procedures.
They both need unplug request cb when the unplug operation happens.

This patch adds hotunplug request cb for ich9, and memory and CPU
hot unplug will share it.
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>
Signed-off-by: NTang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: NZhu Guihua <zhugh.fnst@cn.fujitsu.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Memory and CPU hot unplug are both asynchronous procedures.
They both need unplug request callback to initiate unplug operation.

Add unplug handler to pc machine that will be used by following
CPU and memory unplug patches.
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>
Signed-off-by: NTang Chen <tangchen@cn.fujitsu.com>
Signed-off-by: NZhu Guihua <zhugh.fnst@cn.fujitsu.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

-global lets you set a nice booby-trap for yourself:

    $ qemu-system-x86_64 -nodefaults -S -display none -usb -monitor stdio -global usb-mouse.usb_version=l
    QEMU 2.1.94 monitor - type 'help' for more information
    (qemu) device_add usb-mouse
    Parameter 'usb_version' expects an int64 value or range
    $ echo $?
    1

Not nice.  Until commit 31962700 we even abort()ed.

The same error triggers if you manage to screw up a machine type's
compat_props.  To demonstrate, change HW_COMPAT_2_1's entry to

            .driver   = "usb-mouse",\
            .property = "usb_version",\
            .value    = "1", \

Then run

    $ qemu-system-x86_64 -usb -M pc-i440fx-2.1 -device usb-mouse
    upstream-qemu: -device usb-mouse: Parameter 'usb_version' expects an int64 value or range
    $ echo $?
    1

One of our creatively cruel error messages.

Since this is actually a coding error, we *should* abort() here.
Replace the error by an assertion failure in this case.

But turn the fatal error into a mere warning when the faulty
GlobalProperty comes from the user.  Looks like this:

    $ qemu-system-x86_64 -nodefaults -S -display none -usb -monitor stdio -global usb-mouse.usb_version=l
    QEMU 2.1.94 monitor - type 'help' for more information
    (qemu) device_add usb-mouse
    Warning: global usb-mouse.usb_version=l ignored (Parameter 'usb_version' expects an int64 value or range)
    (qemu)

This is consistent with how we handle similarly unusable -global in
qdev_prop_check_globals().

You could argue that the error should make device_add fail.  Would be
harder, because we're running within TypeInfo's instance_post_init()
method device_post_init(), which can't fail.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Acked-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Reviewed-by: NEduardo Habkost <ehabkost@redhat.com>

Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

The next commit will exploit the fact it never fails.  This one makes
it obvious.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Convert the device models where initialization obviously can't fail.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Call the new PCIDeviceClass method realize().  Default it to
pci_default_realize(), which calls old method init().

To convert a device model, make it implement realize() rather than
init().
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Implement DeviceClass methods realize() and unrealize() instead of
init() and exit().  The core's initialization errors now get
propagated properly, and QMP sends them instead of an unspecific
"Device initialization failed" error.  Unrealize can't fail, so no
change there.

PCIDeviceClass is unchanged: it still provides init() and exit().
Therefore, device models' errors are still not propagated.
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

Signed-off-by: NIgor Mammedov <imammedo@redhat.com>
Reviewed-by: NClaudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: NMarcel Apfelbaum <marcel@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>

Use build_append_namestring() instead of build_append_nameseg()
So user won't have to care whether name is NameSeg, NamePath or
NameString.

See for reference ACPI 5.0: 20.2.2 Name Objects Encoding
Signed-off-by: NIgor Mammedov <imammedo@redhat.com>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NMichael S. Tsirkin <mst@redhat.com>