sdk/emulator/qemu.git
9 years agomacio: switch pmac_dma_write() over to new offset/len implementation
Mark Cave-Ayland [Thu, 4 Jun 2015 21:59:35 +0000 (22:59 +0100)]
macio: switch pmac_dma_write() over to new offset/len implementation

In particular, this fixes a bug whereby chains of overlapping head/tail chains
would incorrectly write over each other's remainder cache. This is the access
pattern used by OS X/Darwin and fixes an issue with a corrupt Darwin
installation in my local tests.

While we are here, rename the DBDMA_io struct property remainder to
head_remainder for clarification.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1433455177-21243-3-git-send-email-mark.cave-ayland@ilande.co.uk
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agomacio: switch pmac_dma_read() over to new offset/len implementation
Mark Cave-Ayland [Thu, 4 Jun 2015 21:59:34 +0000 (22:59 +0100)]
macio: switch pmac_dma_read() over to new offset/len implementation

For better handling of unaligned block device accesses.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1433455177-21243-2-git-send-email-mark.cave-ayland@ilande.co.uk
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc-test: Test state for existing cases more thoroughly
Kevin Wolf [Thu, 21 May 2015 13:19:38 +0000 (15:19 +0200)]
fdc-test: Test state for existing cases more thoroughly

This just adds a few additional checks of the MSR and interrupt pin to
the already existing test cases.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-9-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc: Fix MSR.RQM flag
Kevin Wolf [Thu, 21 May 2015 13:19:37 +0000 (15:19 +0200)]
fdc: Fix MSR.RQM flag

The RQM bit in MSR should be set whenever the guest is supposed to
access the FIFO, and it should be cleared in all other cases. This is
important so the guest can't continue writing/reading the FIFO beyond
the length that it's suppossed to access (see CVE-2015-3456).

Commit e9077462 fixed the CVE by adding code that avoids the buffer
overflow; however it doesn't correct the wrong behaviour of the floppy
controller which should already have cleared RQM.

Currently, RQM stays set all the time and during all phases while a
command is being processed. This is error-prone because the command has
to explicitly clear the flag if it doesn't need data (and indeed, the
two buggy commands that are the culprits for the CVE just forgot to do
that).

This patch clears RQM immediately as soon as all bytes that are expected
have been received. If the the FIFO is used in the next phase, the flag
has to be set explicitly there.

It also clear RQM after receiving all bytes even if the phase transition
immediately sets it again. While it's technically not necessary at the
moment because the state between clearing and setting RQM is not
observable by the guest, this is more explicit and matches how real
hardware works. It will actually become necessary in qemu once
asynchronous code paths are introduced.

This alone should have been enough to fix the CVE, but now we have two
lines of defense - even better.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-8-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc: Disentangle phases in fdctrl_read_data()
Kevin Wolf [Thu, 21 May 2015 13:19:36 +0000 (15:19 +0200)]
fdc: Disentangle phases in fdctrl_read_data()

This commit makes similar improvements as have already been made to the
write function: Instead of relying on a flag in the MSR to distinguish
controller phases, use the explicit phase that we store now. Assertions
of the right MSR flags are added.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-7-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc: Code cleanup in fdctrl_write_data()
Kevin Wolf [Thu, 21 May 2015 13:19:35 +0000 (15:19 +0200)]
fdc: Code cleanup in fdctrl_write_data()

Factor out a few common lines of code, reformat, improve comments.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-6-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc: Use phase in fdctrl_write_data()
Kevin Wolf [Thu, 21 May 2015 13:19:34 +0000 (15:19 +0200)]
fdc: Use phase in fdctrl_write_data()

Instead of relying on a flag in the MSR to distinguish controller phases,
use the explicit phase that we store now. Assertions of the right MSR
flags are added.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-5-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc: Introduce fdctrl->phase
Kevin Wolf [Thu, 21 May 2015 13:19:33 +0000 (15:19 +0200)]
fdc: Introduce fdctrl->phase

The floppy controller spec describes three different controller phases,
which are currently not explicitly modelled in our emulation. Instead,
each phase is represented by a combination of flags in registers.

This patch makes explicit in which phase the controller currently is.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-4-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc: Rename fdctrl_set_fifo() to fdctrl_to_result_phase()
Kevin Wolf [Thu, 21 May 2015 13:19:32 +0000 (15:19 +0200)]
fdc: Rename fdctrl_set_fifo() to fdctrl_to_result_phase()

What callers really do with this function is to switch from execution
phase (including data transfers) to result phase where the guest can
read out one or more status bytes from the FIFO (the number depends on
the command).

Rename the function accordingly.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-3-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agofdc: Rename fdctrl_reset_fifo() to fdctrl_to_command_phase()
Kevin Wolf [Thu, 21 May 2015 13:19:31 +0000 (15:19 +0200)]
fdc: Rename fdctrl_reset_fifo() to fdctrl_to_command_phase()

What all callers of fdctrl_reset_fifo() really want to do is to start
the command phase, where writes to the data port initiate a new command.

The function doesn't only clear the FIFO, but also sets up the state so
that a new command can be received. Rename it to reflect this.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1432214378-31891-2-git-send-email-kwolf@redhat.com
Signed-off-by: John Snow <jsnow@redhat.com>
9 years agoMerge remote-tracking branch 'remotes/sstabellini/tags/xen-15-06-02-tag' into staging
Peter Maydell [Tue, 2 Jun 2015 15:47:31 +0000 (16:47 +0100)]
Merge remote-tracking branch 'remotes/sstabellini/tags/xen-15-06-02-tag' into staging

XSA 128 129 130 131

# gpg: Signature made Tue Jun  2 16:46:38 2015 BST using RSA key ID 70E1AE90
# gpg: Good signature from "Stefano Stabellini <stefano.stabellini@eu.citrix.com>"

* remotes/sstabellini/tags/xen-15-06-02-tag:
  xen/pt: unknown PCI config space fields should be read-only
  xen/pt: add a few PCI config space field descriptions
  xen/pt: mark reserved bits in PCI config space fields
  xen/pt: mark all PCIe capability bits read-only
  xen/pt: split out calculation of throughable mask in PCI config space handling
  xen/pt: correctly handle PM status bit
  xen/pt: consolidate PM capability emu_mask
  xen/MSI: don't open-code pass-through of enable bit modifications
  xen/MSI-X: limit error messages
  xen: don't allow guest to control MSI mask register
  xen: properly gate host writes of modified PCI CFG contents

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoxen/pt: unknown PCI config space fields should be read-only
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/pt: unknown PCI config space fields should be read-only

... by default. Add a per-device "permissive" mode similar to pciback's
to allow restoring previous behavior (and hence break security again,
i.e. should be used only for trusted guests).

This is part of XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>)
9 years agoxen/pt: add a few PCI config space field descriptions
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/pt: add a few PCI config space field descriptions

Since the next patch will turn all not explicitly described fields
read-only by default, those fields that have guest writable bits need
to be given explicit descriptors.

This is a preparatory patch for XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
9 years agoxen/pt: mark reserved bits in PCI config space fields
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/pt: mark reserved bits in PCI config space fields

The adjustments are solely to make the subsequent patches work right
(and hence make the patch set consistent), namely if permissive mode
(introduced by the last patch) gets used (as both reserved registers
and reserved fields must be similarly protected from guest access in
default mode, but the guest should be allowed access to them in
permissive mode).

This is a preparatory patch for XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
9 years agoxen/pt: mark all PCIe capability bits read-only
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/pt: mark all PCIe capability bits read-only

xen_pt_emu_reg_pcie[]'s PCI_EXP_DEVCAP needs to cover all bits as read-
only to avoid unintended write-back (just a precaution, the field ought
to be read-only in hardware).

This is a preparatory patch for XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
9 years agoxen/pt: split out calculation of throughable mask in PCI config space handling
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/pt: split out calculation of throughable mask in PCI config space handling

This is just to avoid having to adjust that calculation later in
multiple places.

Note that including ->ro_mask in get_throughable_mask()'s calculation
is only an apparent (i.e. benign) behavioral change: For r/o fields it
doesn't matter > whether they get passed through - either the same flag
is also set in emu_mask (then there's no change at all) or the field is
r/o in hardware (and hence a write won't change it anyway).

This is a preparatory patch for XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
9 years agoxen/pt: correctly handle PM status bit
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/pt: correctly handle PM status bit

xen_pt_pmcsr_reg_write() needs an adjustment to deal with the RW1C
nature of the not passed through bit 15 (PCI_PM_CTRL_PME_STATUS).

This is a preparatory patch for XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
9 years agoxen/pt: consolidate PM capability emu_mask
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/pt: consolidate PM capability emu_mask

There's no point in xen_pt_pmcsr_reg_{read,write}() each ORing
PCI_PM_CTRL_STATE_MASK and PCI_PM_CTRL_NO_SOFT_RESET into a local
emu_mask variable - we can have the same effect by setting the field
descriptor's emu_mask member suitably right away. Note that
xen_pt_pmcsr_reg_write() is being retained in order to allow later
patches to be less intrusive.

This is a preparatory patch for XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
9 years agoxen/MSI: don't open-code pass-through of enable bit modifications
Jan Beulich [Tue, 2 Jun 2015 15:07:01 +0000 (15:07 +0000)]
xen/MSI: don't open-code pass-through of enable bit modifications

Without this the actual XSA-131 fix would cause the enable bit to not
get set anymore (due to the write back getting suppressed there based
on the OR of emu_mask, ro_mask, and res_mask).

Note that the fiddling with the enable bit shouldn't really be done by
qemu, but making this work right (via libxc and the hypervisor) will
require more extensive changes, which can be postponed until after the
security issue got addressed.

This is a preparatory patch for XSA-131.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
9 years agoxen/MSI-X: limit error messages
Jan Beulich [Tue, 2 Jun 2015 15:07:00 +0000 (15:07 +0000)]
xen/MSI-X: limit error messages

Limit error messages resulting from bad guest behavior to avoid allowing
the guest to cause the control domain's disk to fill.

The first message in pci_msix_write() can simply be deleted, as this
is indeed bad guest behavior, but such out of bounds writes don't
really need to be logged.

The second one is more problematic, as there guest behavior may only
appear to be wrong: For one, the old logic didn't take the mask-all bit
into account. And then this shouldn't depend on host device state (i.e.
the host may have masked the entry without the guest having done so).
Plus these writes shouldn't be dropped even when an entry is unmasked.
Instead, if they can't be made take effect right away, they should take
effect on the next unmasking or enabling operation - the specification
explicitly describes such caching behavior. Until we can validly drop
the message (implementing such caching/latching behavior), issue the
message just once per MSI-X table entry.

Note that the log message in pci_msix_read() similar to the one being
removed here is not an issue: "addr" being of unsigned type, and the
maximum size of the MSI-X table being 32k, entry_nr simply can't be
negative and hence the conditonal guarding issuing of the message will
never be true.

This is XSA-130.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
9 years agoxen: don't allow guest to control MSI mask register
Jan Beulich [Tue, 2 Jun 2015 15:07:00 +0000 (15:07 +0000)]
xen: don't allow guest to control MSI mask register

It's being used by the hypervisor. For now simply mimic a device not
capable of masking, and fully emulate any accesses a guest may issue
nevertheless as simple reads/writes without side effects.

This is XSA-129.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
9 years agoxen: properly gate host writes of modified PCI CFG contents
Jan Beulich [Tue, 2 Jun 2015 15:07:00 +0000 (15:07 +0000)]
xen: properly gate host writes of modified PCI CFG contents

The old logic didn't work as intended when an access spanned multiple
fields (for example a 32-bit access to the location of the MSI Message
Data field with the high 16 bits not being covered by any known field).
Remove it and derive which fields not to write to from the accessed
fields' emulation masks: When they're all ones, there's no point in
doing any host write.

This fixes a secondary issue at once: We obviously shouldn't make any
host write attempt when already the host read failed.

This is XSA-128.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
9 years agotarget-arm: Correct check for non-EL3
Edgar E. Iglesias [Fri, 29 May 2015 06:43:13 +0000 (16:43 +1000)]
target-arm: Correct check for non-EL3

This fixes a compile warning from clang 3.5 (the assertion
could never fire).

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 1432881807-18164-2-git-send-email-edgar.iglesias@gmail.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: John Snow <jsnow@redhat.com>
[PMM: added note in commit message that this is fixing a build warning]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-input-20150602-1' into staging
Peter Maydell [Tue, 2 Jun 2015 10:25:12 +0000 (11:25 +0100)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-input-20150602-1' into staging

virtio-input: two small fixups

# gpg: Signature made Tue Jun  2 09:32:51 2015 BST using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/kraxel/tags/pull-input-20150602-1:
  virtio-input: make virtio devices follow usual naming convention
  virtio-input: const_le16 and const_le32 not build time constant

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Tue, 2 Jun 2015 09:20:03 +0000 (10:20 +0100)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

pc build fix

My last pull breaks build on systems with iasl.
Fix this up.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Mon Jun  1 20:41:08 2015 BST using RSA key ID D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>"

* remotes/mst/tags/for_upstream:
  acpi: add missing ssdt

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agovirtio-input: make virtio devices follow usual naming convention
Gerd Hoffmann [Tue, 2 Jun 2015 08:31:29 +0000 (10:31 +0200)]
virtio-input: make virtio devices follow usual naming convention

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agovirtio-input: const_le16 and const_le32 not build time constant
Michael Mueller [Mon, 1 Jun 2015 13:51:56 +0000 (15:51 +0200)]
virtio-input: const_le16 and const_le32 not build time constant

As the implementation of const_le16 and const_le32 is not build time constant
on big endian systems this need to be fixed.

  CC    hw/input/virtio-input-hid.o
hw/input/virtio-input-hid.c:340:13: error: initializer element is not constant
hw/input/virtio-input-hid.c:340:13: error: (near initialization for ‘virtio_keyboard_config[1].u.ids.bustype’)
...

Signed-off-by: Michael Mueller <mimu@linux.vnet.ibm.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
9 years agoacpi: add missing ssdt
Michael S. Tsirkin [Mon, 1 Jun 2015 19:03:59 +0000 (21:03 +0200)]
acpi: add missing ssdt

commit 5cb18b3d7bff2a83275ee98af2a14eb9e21c93ab
    TPM2 ACPI table support

was missing a file, so build with iasl fails
(build without iasl works since it uses the generated
 hex files).

Reported-by: "Daniel P. Berrange" <berrange@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Mon, 1 Jun 2015 14:22:46 +0000 (15:22 +0100)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

pc, pci, tpm, virtio, vhost enhancements and fixes

A bunch of cleanups and fixes all over the place,
enhancements in TPM, virtio and vhost.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Mon Jun  1 13:19:48 2015 BST using RSA key ID D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>"

* remotes/mst/tags/for_upstream: (60 commits)
  vhost-user: add multi queue support
  virtio: make features 64bit wide
  qdev: add 64bit properties
  virtio-mmio: ioeventfd support
  hw/acpi/aml-build: Fix memory leak
  acpi: add aml_while() term
  acpi: add aml_increment() term
  acpi: add aml_shiftright() term
  acpi: add aml_shiftleft() term
  acpi: add aml_index() term
  acpi: add aml_lless() term
  acpi: add aml_add() term
  TPM2 ACPI table support
  tpm: Probe for connected TPM 1.2 or TPM 2
  Extend TPM TIS interface to support TPM 2
  Add stream ID to MSI write
  acpi: Simplify printing to dynamic string
  i386: drop FDC in pc-q35-2.4+ if neither it nor floppy drives are wanted
  i386/pc_q35: don't insist on board FDC if there's no default floppy
  i386/pc: '-drive if=floppy' should imply a board-default FDC
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agovhost-user: add multi queue support
Ouyang Changchun [Thu, 28 May 2015 01:23:06 +0000 (09:23 +0800)]
vhost-user: add multi queue support

Based on patch by Nikolay Nikolaev:
Vhost-user will implement the multi queue support in a similar way
to what vhost already has - a separate thread for each queue.
To enable the multi queue functionality - a new command line parameter
"queues" is introduced for the vhost-user netdev.

Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com>
Signed-off-by: Changchun Ouyang <changchun.ouyang@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio: make features 64bit wide
Gerd Hoffmann [Mon, 1 Jun 2015 08:45:40 +0000 (10:45 +0200)]
virtio: make features 64bit wide

Make features 64bit wide everywhere.

On migration a full 64bit guest_features field is sent if one of the
high bits is set, in addition to the lower 32bit guest_features field
which must stay for compatibility reasons.  That way we send the lower
32 feature bits twice, but the code is simpler because we don't have
to split and compose the 64bit features into two 32bit fields.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoqdev: add 64bit properties
Gerd Hoffmann [Mon, 1 Jun 2015 08:45:39 +0000 (10:45 +0200)]
qdev: add 64bit properties

Needed for virtio features which go from 32bit to 64bit with virtio 1.0

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio-mmio: ioeventfd support
Ying-Shiuan Pan [Tue, 12 May 2015 08:10:50 +0000 (11:10 +0300)]
virtio-mmio: ioeventfd support

set_host_notifier and set_guest_notifiers supported by virtio-mmio now.
Most code copied from virtio-pci.

This makes it possible to use vhost-net with virtio-mmio,
improving performance by about 30%.

The kvm-arm does not yet support irqfd, need to fix the hard-coded part after
kvm-arm gets irqfd support.

Signed-off-by: Ying-Shiuan Pan <yingshiuan.pan@gmail.com>
Signed-off-by: Pavel Fedin <p.fedin@samsung.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agohw/acpi/aml-build: Fix memory leak
Shannon Zhao [Mon, 25 May 2015 07:14:37 +0000 (15:14 +0800)]
hw/acpi/aml-build: Fix memory leak

Signed-off-by: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
9 years agoacpi: add aml_while() term
Marcel Apfelbaum [Mon, 25 May 2015 15:33:46 +0000 (18:33 +0300)]
acpi: add aml_while() term

Add encoding for ACPI DefWhile Opcode.

Reviewed-by: Shannon Zhao <zhaoshenglong@huawei.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoacpi: add aml_increment() term
Marcel Apfelbaum [Mon, 25 May 2015 15:33:45 +0000 (18:33 +0300)]
acpi: add aml_increment() term

Add encoding for ACPI DefIncrement Opcode.

Reviewed-by: Shannon Zhao <zhaoshenglong@huawei.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoacpi: add aml_shiftright() term
Marcel Apfelbaum [Mon, 25 May 2015 15:33:44 +0000 (18:33 +0300)]
acpi: add aml_shiftright() term

Add encoding for ACPI DefShiftRight Opcode.

Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Shannon Zhao <shannon.zhao@linaro.org>
9 years agoacpi: add aml_shiftleft() term
Marcel Apfelbaum [Mon, 25 May 2015 15:33:43 +0000 (18:33 +0300)]
acpi: add aml_shiftleft() term

Add encoding for ACPI DefShiftLeft Opcode.

Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Shannon Zhao <shannon.zhao@linaro.org>
9 years agoacpi: add aml_index() term
Marcel Apfelbaum [Mon, 25 May 2015 15:33:42 +0000 (18:33 +0300)]
acpi: add aml_index() term

Add encoding for ACPI DefIndex Opcode.

Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Shannon Zhao <shannon.zhao@linaro.org>
9 years agoacpi: add aml_lless() term
Marcel Apfelbaum [Mon, 25 May 2015 15:33:41 +0000 (18:33 +0300)]
acpi: add aml_lless() term

Add encoding for ACPI DefLLess Opcode.

Reviewed-by: Shannon Zhao <zhaoshenglong@huawei.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoacpi: add aml_add() term
Marcel Apfelbaum [Mon, 25 May 2015 15:33:40 +0000 (18:33 +0300)]
acpi: add aml_add() term

Add encoding for ACPI DefAdd Opcode.

Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Shannon Zhao <shannon.zhao@linaro.org>
9 years agoTPM2 ACPI table support
Stefan Berger [Tue, 26 May 2015 20:51:07 +0000 (16:51 -0400)]
TPM2 ACPI table support

Add a TPM2 ACPI table if a TPM 2 is used in the backend.
Also add an SSDT for the TPM 2.

Rename tpm_find() to tpm_get_version() and have this function
return the version of the TPM found, TPMVersion_Unspec if
no TPM is found. Use the version number to build version
specific ACPI tables.

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoMerge remote-tracking branch 'remotes/bkoppelmann/tags/pull-tricore-20150530' into...
Peter Maydell [Mon, 1 Jun 2015 10:29:37 +0000 (11:29 +0100)]
Merge remote-tracking branch 'remotes/bkoppelmann/tags/pull-tricore-20150530' into staging

TriCore bugfixes

# gpg: Signature made Sat May 30 15:50:49 2015 BST using RSA key ID 6B69CA14
# gpg: Good signature from "Bastian Koppelmann <kbastian@mail.uni-paderborn.de>"

* remotes/bkoppelmann/tags/pull-tricore-20150530:
  target-tricore: fix BOL_ST_H_LONGOFF using ld
  target-tricore: fix msub32_q producing the wrong overflow bit
  target-tricore: fix OPC2_32_RR_DVINIT_HU having write before use on the result

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotpm: Probe for connected TPM 1.2 or TPM 2
Stefan Berger [Tue, 26 May 2015 20:51:06 +0000 (16:51 -0400)]
tpm: Probe for connected TPM 1.2 or TPM 2

In the TPM passthrough backend driver, modify the probing code so
that we can check whether a TPM 1.2 or TPM 2 is being used
and adapt the behavior of the TPM TIS accordingly.

Move the code that tested for a TPM 1.2 into tpm_utils.c
and extend it with test for probing for TPM 2. Have the
function return the version of TPM found.

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoExtend TPM TIS interface to support TPM 2
Stefan Berger [Tue, 26 May 2015 20:51:05 +0000 (16:51 -0400)]
Extend TPM TIS interface to support TPM 2

Following the recent upgrade to version 1.3, extend the TPM TIS
interface with capabilities introduced for support of a TPM 2.

TPM TIS for TPM 2 introduced the following extensions beyond the
TPM TIS 1.3 (used for TPM 1.2):

- A new 32bit interface Id register was introduced.
- New flags for the status (STS) register were defined.
- New flags for the capability flags were defined.

Support the above if a TPM TIS 1.3 for TPM 2 is used with a TPM 2
on the backend side. Support the old TPM TIS 1.3 configuration if a
TPM 1.2 is being used. A subsequent patch will then determine which
TPM version is being used in the backend.

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoAdd stream ID to MSI write
Pavel Fedin [Wed, 27 May 2015 12:59:59 +0000 (15:59 +0300)]
Add stream ID to MSI write

GICv3 ITS distinguishes between devices by using hardwired device IDs passed on the bus.
This patch implements passing these IDs in qemu.
SMMU is also known to use stream IDs, therefore this addition can also be useful for
implementing platforms with SMMU.

Signed-off-by: Pavel Fedin <p.fedin@samsung.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
 Changes from v1:
- Added bus number to the stream ID
- Added stream ID not only to MSI-X, but also to plain MSI. Some common code was made into
msi_send_message() function.
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agoacpi: Simplify printing to dynamic string
Markus Armbruster [Wed, 27 May 2015 17:55:55 +0000 (19:55 +0200)]
acpi: Simplify printing to dynamic string

build_append_namestringv() and aml_string() first calculate the
resulting string's length with vsnprintf(NULL, ...), then allocate,
then print for real.  Simply use g_strdup_vprintf() or g_vasprintf()
instead.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
9 years agoi386: drop FDC in pc-q35-2.4+ if neither it nor floppy drives are wanted
Laszlo Ersek [Thu, 28 May 2015 20:04:11 +0000 (22:04 +0200)]
i386: drop FDC in pc-q35-2.4+ if neither it nor floppy drives are wanted

It is Very annoying to carry forward an outdatEd coNtroller with a mOdern
Machine type.

Hence, let us not instantiate the FDC when all of the following apply:
- the machine type is pc-q35-2.4 or later,
- "-device isa-fdc" is not passed on the command line (nor in the config
  file),
- no "-drive if=floppy,..." is requested.

Cc: Markus Armbruster <armbru@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: John Snow <jsnow@redhat.com>
Cc: "Gabriel L. Somlo" <gsomlo@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: qemu-block@nongnu.org
Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agoi386/pc_q35: don't insist on board FDC if there's no default floppy
Laszlo Ersek [Thu, 28 May 2015 20:04:10 +0000 (22:04 +0200)]
i386/pc_q35: don't insist on board FDC if there's no default floppy

The "no_floppy = 1" machine class setting causes "default_floppy" in
main() to become zero. Consequently, default_drive() will not call
drive_add() and drive_new() for IF_FLOPPY, index=0, meaning that no
default floppy drive will be created for the virtual machine. In that
case, board code should also not insist on the creation of the
board-default FDC.

The board-default FDC will still be created if the user requests a floppy
drive with "-drive if=floppy".

Additionally, separate FDCs can be specified manually with "-device
isa-fdc". They allow the

  -device isa-fdc,driveA=...

syntax that is more flexible than the one required by the board-default
FDC:

  -global isa-fdc.driveA=...

This patch doesn't change the behavior observably, as all Q35 machine
types have "no_floppy = 0".

Cc: Markus Armbruster <armbru@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: John Snow <jsnow@redhat.com>
Cc: "Gabriel L. Somlo" <gsomlo@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: qemu-block@nongnu.org
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agoi386/pc: '-drive if=floppy' should imply a board-default FDC
Laszlo Ersek [Thu, 28 May 2015 20:04:09 +0000 (22:04 +0200)]
i386/pc: '-drive if=floppy' should imply a board-default FDC

Even if board code decides not to request the creation of the FDC (keyed
off board-level factors, to be determined later), we should create the FDC
nevertheless if the user passes '-drive if=floppy' on the command line.

Otherwise '-drive if=floppy' would break without explicit '-device
isa-fdc' on such boards.

Cc: Markus Armbruster <armbru@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: John Snow <jsnow@redhat.com>
Cc: "Gabriel L. Somlo" <gsomlo@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: qemu-block@nongnu.org
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agoi386/pc: pc_basic_device_init(): delegate FDC creation request
Laszlo Ersek [Thu, 28 May 2015 20:04:08 +0000 (22:04 +0200)]
i386/pc: pc_basic_device_init(): delegate FDC creation request

This patch introduces no observable change, but it allows the callers of
pc_basic_device_init(), ie. pc_init1() and pc_q35_init(), to request (or
not request) the creation of the FDC explicitly.

At the moment both callers pass constant create_fdctrl=true (hence no
observable change).

Assuming a board passes create_fdctrl=false, "floppy" will be NULL on
output, and (beyond the FDC not being created) that NULL will be passed on
to pc_cmos_init(). Luckily, pc_cmos_init() already handles that case.

Cc: Markus Armbruster <armbru@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: John Snow <jsnow@redhat.com>
Cc: "Gabriel L. Somlo" <gsomlo@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: qemu-block@nongnu.org
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
9 years agovirtio: increase the queue limit to 1024
Jason Wang [Fri, 29 May 2015 06:15:32 +0000 (14:15 +0800)]
virtio: increase the queue limit to 1024

Increase the queue limit to 1024. But virtio-ccw and s390-virtio won't
support this, this is done through failing device_plugged() for those
two transports if the number of virtqueues is greater than 64.

Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio: rename VIRTIO_PCI_QUEUE_MAX to VIRTIO_QUEUE_MAX
Jason Wang [Fri, 29 May 2015 06:15:31 +0000 (14:15 +0800)]
virtio: rename VIRTIO_PCI_QUEUE_MAX to VIRTIO_QUEUE_MAX

VIRTIO_PCI_QUEUE_MAX is not only used for pci, so rename it be generic.

Cc: Amit Shah <amit.shah@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio-s390: introduce virtio_s390_device_plugged()
Jason Wang [Fri, 29 May 2015 06:15:30 +0000 (14:15 +0800)]
virtio-s390: introduce virtio_s390_device_plugged()

This patch introduce a virtio-s390 specific device_plugged() function
and doing the number of virtqueue validation inside.

Cc: Alexander Graf <agraf@suse.de>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio-s390: introduce virito s390 queue limit
Jason Wang [Fri, 29 May 2015 06:15:29 +0000 (14:15 +0800)]
virtio-s390: introduce virito s390 queue limit

Cc: Alexander Graf <agraf@suse.de>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio-ccw: validate the number of queues against bus limitation
Jason Wang [Fri, 29 May 2015 06:15:28 +0000 (14:15 +0800)]
virtio-ccw: validate the number of queues against bus limitation

Cc: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Alexander Graf <agraf@suse.de>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio-ccw: introduce ccw specific queue limit
Jason Wang [Fri, 29 May 2015 06:15:27 +0000 (14:15 +0800)]
virtio-ccw: introduce ccw specific queue limit

Cc: Alexander Graf <agraf@suse.de>
Cc: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio: introduce virtio_get_num_queues()
Jason Wang [Fri, 29 May 2015 06:15:26 +0000 (14:15 +0800)]
virtio: introduce virtio_get_num_queues()

This patch introduces virtio_get_num_queues() which iterates the vqs
array and return the number of virtqueues used by device.

Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio: device_plugged() can fail
Jason Wang [Fri, 29 May 2015 06:15:25 +0000 (14:15 +0800)]
virtio: device_plugged() can fail

This patch passes error pointer to transport specific device_plugged()
callback. Through this way, device_plugged() can do some transport
specific check and fail. This will be uesd by following patches that
check the number of virtqueues against the transport limitation.

Cc: Cornelia Huck <cornelia.huck@de.ibm.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Alexander Graf <agraf@suse.de>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio-net: adding all queues in .realize()
Jason Wang [Fri, 29 May 2015 06:15:24 +0000 (14:15 +0800)]
virtio-net: adding all queues in .realize()

Instead of adding queues for multiqueue during feature set. This patch
did this in .realize(), this will help the following patches that
count the number of virtqueues used in .device_plugged() callback.

Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio: move VIRTIO_F_NOTIFY_ON_EMPTY into core
Cornelia Huck [Fri, 29 May 2015 09:29:40 +0000 (11:29 +0200)]
virtio: move VIRTIO_F_NOTIFY_ON_EMPTY into core

Nearly all transports have been offering VIRTIO_F_NOTIFY_ON_EMPTY,
s390-virtio being the exception. There's no reason why it shouldn't
offer it as well, though (handling is done in core anyway), so let's
move it to the common virtio features.

While we're changing it anyway, fix the indentation for the
DEFINE_VIRTIO_COMMON_FEATURES macro.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio-ccw: Don't advertise VIRTIO_F_BAD_FEATURE
Cornelia Huck [Fri, 29 May 2015 09:29:39 +0000 (11:29 +0200)]
virtio-ccw: Don't advertise VIRTIO_F_BAD_FEATURE

This was copied from virtio-pci, but it doesn't make much sense for
ccw, as it doesn't have to handle the broken implementations this bit
is supposed to deal with. Remove it.

Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio: move host_features
Cornelia Huck [Tue, 26 May 2015 14:34:47 +0000 (16:34 +0200)]
virtio: move host_features

Move host_features from the individual transport proxies into
the virtio device. Transports may continue to add feature bits
during device plugging.

This should it make easier to offer different sets of host features
for virtio-1/transitional support.

Tested-by: Shannon Zhao <shannon.zhao@linaro.org>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: acpi: fix pvpanic for buggy guests
Radim Krčmář [Fri, 29 May 2015 19:57:32 +0000 (21:57 +0200)]
pc: acpi: fix pvpanic for buggy guests

In the old times, we always had pvpanic in ACPI and a _STA method told
the guest not to use it.  Automatic generation dropped the _STA method
as the specification says that missing _STA means enabled and working.
Some guests (Linux) had buggy drivers and this change made them unable
to utilize pvpanic.

A Linux patch is posted as well, but I think it's worth to make pvpanic
useable on old guests at the price of three lines and few bytes of SSDT.

The old _STA method was
  Method (_STA, 0, NotSerialized) {
      Store (PEST, Local0)
      If (LEqual (Local0, Zero)) {
          Return (Zero) }
      Else {
          Return (0x0F) }}

Igor pointed out that we don't need to use a method to return a constant
and that 0xB (don't show in UI) is the common definition now.

Also, the device used to be PEVT.  (PEVT as in "panic event"?)

Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Generate init functions with a macro
Eduardo Habkost [Fri, 15 May 2015 17:19:01 +0000 (14:19 -0300)]
pc: Generate init functions with a macro

All pc-i440fx and pc-q35 init functions simply call the corresponding
compat function and then call the main init function. Use a macro to
generate that code.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopiix: Eliminate pc_init_pci()
Eduardo Habkost [Fri, 15 May 2015 17:19:00 +0000 (14:19 -0300)]
piix: Eliminate pc_init_pci()

The function is not needed anymore, we can simply call pc_init1()
directly.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopiix: Add kvmclock_enabled, pci_enabled globals
Eduardo Habkost [Fri, 15 May 2015 17:18:59 +0000 (14:18 -0300)]
piix: Add kvmclock_enabled, pci_enabled globals

This looks like a step backwards, but it will allow pc-0.1[0123] and
isapc to follow the same compat+init pattern used by the other
machine-types, allowing us to generate all init function using the same
macro later.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agomachine: Remove unused fields from QEMUMachine
Eduardo Habkost [Fri, 15 May 2015 17:18:58 +0000 (14:18 -0300)]
machine: Remove unused fields from QEMUMachine

This removes the following fields from QEMUMachine: family, alias,
reset, hot_add_cpu, units_per_default_bus, no_serial, no_parallel,
use_virtcon, use_sclp, no_floppy, no_cdrom, default_display,
compat_props, and hw_version.

The only users of those fields were already converted to use QOM and
MachineClass directly, so they are not needed anymore.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Remove qemu_register_pc_machine() function
Eduardo Habkost [Fri, 15 May 2015 17:18:57 +0000 (14:18 -0300)]
pc: Remove qemu_register_pc_machine() function

The helper is not needed anymore, as the PC machine classes are
registered using QOM directly.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Don't use QEMUMachine anymore
Eduardo Habkost [Fri, 15 May 2015 17:18:56 +0000 (14:18 -0300)]
pc: Don't use QEMUMachine anymore

Now that we have a DEFINE_PC_MACHINE helper macro that just requires an
initialization function, it is trivial to convert them to register a QOM
machine class directly, instead of using QEMUMachine.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Move compat_props setting inside *_machine_options() functions
Eduardo Habkost [Fri, 15 May 2015 17:18:55 +0000 (14:18 -0300)]
pc: Move compat_props setting inside *_machine_options() functions

This will simplify the DEFINE_PC_MACHINE macro, and will help us to
implement reuse of PC_COMPAT_* macros through class_init function reuse,
in the future.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Convert *_MACHINE_OPTIONS macros into functions
Eduardo Habkost [Fri, 15 May 2015 17:18:54 +0000 (14:18 -0300)]
pc: Convert *_MACHINE_OPTIONS macros into functions

By now the new functions will get QEMUMachine as argument, but they will
be later converted to initialize a MachineClass struct directly.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Define machines using a DEFINE_PC_MACHINE macro
Eduardo Habkost [Fri, 15 May 2015 17:18:53 +0000 (14:18 -0300)]
pc: Define machines using a DEFINE_PC_MACHINE macro

This will automatically generate the existing QEMUMachine structs based
on the *_MACHINE_OPTIONS macros, and automatically add registration code
for them.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Define MACHINE_OPTIONS macros consistently for all machines
Eduardo Habkost [Fri, 15 May 2015 17:18:52 +0000 (14:18 -0300)]
pc: Define MACHINE_OPTIONS macros consistently for all machines

Define a MACHINE_OPTIONS macro for each PC machine, and move every field
inside the QEMUMachine structs to the macros, except for name, init, and
compat_props.

This also ensures that all MACHINE_OPTIONS inherit the fields from the
next version, so their definitions carry only the changes that exist
between one version and the next one.

Comments about specific cases:

pc-*-2.1:

  Existing PC_*_2_1_MACHINE_OPTIONS macros were defined as:
      PC_*_MACHINE_OPTIONS,
      .default_machine_opts = "firmware=bios-256k.bin"

  PC_*_2_2_MACHINE_OPTIONS is:
      PC_*_2_3_MACHINE_OPTIONS
  which is expanded to:
      PC_*_MACHINE_OPTIONS,
      .default_machine_opts = "firmware=bios-256k.bin",
      .default_display = "std"

  The only difference between 2_1 and 2_2 is .default_display, that's why
  we didn't reuse PC_*_2_2_MACHINE_OPTIONS. The good news is that having
  multiple initializers for a field is allowed by C99, and the last
  initializer overrides the previous ones.

  So we can reuse the 2_2 macro in 2_1 and define PC_*_2_1_MACHINE_OPTIONS
  as:
      PC_*_2_2_MACHINE_OPTIONS,
      .default_display = NULL

pc-*-1.7:

  PC_*_1_7_MACHINE_OPTIONS was defined as:
      PC_*_MACHINE_OPTIONS

  PC_*_2_0_MACHINE_OPTIONS is defined as:
      PC_*_2_1_MACHINE_OPTIONS
  which is expanded to:
      PC_*_2_2_MACHINE_OPTIONS,
      .default_display = NULL
  which is expanded to:
      PC_*_2_3_MACHINE_OPTIONS,
      .default_display = NULL
  which is expanded to:
      PC_*_MACHINE_OPTIONS,
      .default_machine_opts = "firmware=bios-256k.bin",
      .default_display = "std",
      .default_display = NULL  /* overrides the previous line */

  So, the only difference between PC_*_1_7_MACHINE_OPTIONS and
  PC_*_2_0_MACHINE_OPTIONS is .default_machine_opts (as .default_display
  is not explicitly set by PC_*_MACHINE_OPTIONS so it is NULL).

  So we can keep the macro reuse pattern and define
  PC_*_2_0_MACHINE_OPTIONS as:
      PC_*_2_0_MACHINE_OPTIONS,
      .default_machine_opts = NULL

pc-*-2.4 (alias and is_default fields):

  Set alias and is_default fields inside the 2.4 MACHINE_OPTIONS macro,
  and clear it in the 2.3 macro (that reuses the 2.4 macro).

hw_machine:

  As all the machines older than v1.0 set hw_version explicitly, we can
  safely move the field to the MACHINE_OPTIONS macros without affecting
  the other versions that reuse them.

init function:

  Some machines had the init function set inside the MACHINE_OPTIONS
  macro. Move it to the QEMUMachine declaration, to keep it consistent
  with the other machines.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopiix: Define PC_COMPAT_0_10
Eduardo Habkost [Thu, 14 May 2015 18:53:10 +0000 (15:53 -0300)]
piix: Define PC_COMPAT_0_10

Move compat_props from pc-0.10 to the macro, to make it consistent with
the other machines.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopiix: Move pc-0.1[23] rombar compat props to PC_COMPAT_0_13
Eduardo Habkost [Thu, 14 May 2015 18:53:09 +0000 (15:53 -0300)]
piix: Move pc-0.1[23] rombar compat props to PC_COMPAT_0_13

The VGA and vmware-svga rombar compat properties were added by commit
281a26b15b4adcecb8604216738975abd754bea8, but only to pc-0.13 and
pc-0.12. This breaks the PC_COMPAT_* nesting pattern we currently
follow.

The new variables will now be inherited by pc-0.11 and older, but
pc-0.11 and pc-0.10 already have PCI.rombar=0 on compat_props, so they
shouldn't be affected at all.

Cc: Stefan Weil <sw@weilnetz.de>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopiix: Move pc-0.13 virtio-9p-pci compat to PC_COMPAT_0_13
Eduardo Habkost [Thu, 14 May 2015 18:53:08 +0000 (15:53 -0300)]
piix: Move pc-0.13 virtio-9p-pci compat to PC_COMPAT_0_13

The compat property was added by commit
9dbcca5aa13cb9ab40788ac4c56bc227d94ca920, and the pc-0.12 and older
machine-types were not changed because virtio-9p-pci was introduced on QEMU
0.13 (commit 9f10751365b26b13b8a9b67e0e90536ae3d282df). The only problem is
that this breaks the PC_COMPAT_* nesting pattern we currently use.

So, move the property to PC_COMPAT_0_13. This make pc-0.12 and older inherit
it, but that shouldn't be an issue as QEMU 0.12 didn't have virtio-9p-pci.

Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopiix: Move pc-0.11 drive version compat props to PC_COMPAT_0_11
Eduardo Habkost [Thu, 14 May 2015 18:53:07 +0000 (15:53 -0300)]
piix: Move pc-0.11 drive version compat props to PC_COMPAT_0_11

The current code setting ide-drive.ver and scsi-disk.ver on pc-0.11
breaks the PC_COMPAT_* nesting pattern we currently use.

As those variables are overwritten in pc-0.10 too, they can be inherited
by pc-0.10 with no side-effects at all.

Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopiix: Move pc-0.14 qxl compat properties to PC_COMPAT_0_14
Eduardo Habkost [Thu, 14 May 2015 18:53:06 +0000 (15:53 -0300)]
piix: Move pc-0.14 qxl compat properties to PC_COMPAT_0_14

Those properties were introduced by commit
3827cdb1c3aa17a792d1658161195b9d7173c26b. They were not duplicated into
pc-0.13 and older because 0.14 was the first QEMU version supporting
qxl. The only problem is that this breaks the PC_COMPAT_* nesting
pattern we currently use.

So, move the properties to PC_COMPAT_0_14. This makes pc-0.13 and older
inherit them, but that shouldn't be an issue as QEMU 0.13 didn't support
qxl.

Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agospapr: define SPAPR_COMPAT_2_3
Eduardo Habkost [Thu, 14 May 2015 18:53:05 +0000 (15:53 -0300)]
spapr: define SPAPR_COMPAT_2_3

Don't add the pseries-2.3 machine yet, but define the corresponding
SPAPR_COMPAT macro to make sure both pseries-2.2 and pseries-2.1 will
inherit HW_COMPAT_2_3.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agospapr: Use HW_COMPAT_* inside SPAPR_COMPAT_* macros
Eduardo Habkost [Thu, 14 May 2015 18:53:04 +0000 (15:53 -0300)]
spapr: Use HW_COMPAT_* inside SPAPR_COMPAT_* macros

SPAPR_COMPAT_2_1 will need to include both HW_COMPAT_2_2 and
HW_COMPAT_2_1, so include HW_COMPAT_2_1 inside SPAPR_COMPAT_2_1 and
HW_COMPAT_2_2 inside SPAPR_COMPAT_2_2.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Define PC_COMPAT_2_[123] macros
Eduardo Habkost [Thu, 14 May 2015 18:53:03 +0000 (15:53 -0300)]
pc: Define PC_COMPAT_2_[123] macros

Once we start adding compat code for pc-2.3, the usage of HW_COMPAT_2_1
in pc-*-2.2 won't be enough, as it also has to include PC_COMPAT_2_3
inside it. To ensure that, define PC_COMPAT_2_3, PC_COMPAT_2_2, and
PC_COMPAT_2_1 macros.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agohw: Define empty HW_COMPAT_2_[23] macros
Eduardo Habkost [Thu, 14 May 2015 18:53:02 +0000 (15:53 -0300)]
hw: Define empty HW_COMPAT_2_[23] macros

Now we can make everything consistent and define the macros even if they
are still empty.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agospapr: Move commas inside SPAPR_COMPAT_* macros
Eduardo Habkost [Thu, 14 May 2015 18:53:01 +0000 (15:53 -0300)]
spapr: Move commas inside SPAPR_COMPAT_* macros

Changing the convention to include commas inside the macros will allow
macros containing empty lists to be defined and used without compilation
errors.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Move commas inside PC_COMPAT_* macros
Eduardo Habkost [Thu, 14 May 2015 18:53:00 +0000 (15:53 -0300)]
pc: Move commas inside PC_COMPAT_* macros

Changing the convention to include commas inside the macros will allow
macros containing empty lists to be defined and used without compilation
errors.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agohw: Move commas inside HW_COMPAT_2_1 macro
Eduardo Habkost [Thu, 14 May 2015 18:52:59 +0000 (15:52 -0300)]
hw: Move commas inside HW_COMPAT_2_1 macro

Changing the convention to include commas inside the macros will allow
macros containing empty lists to be defined and used without compilation
errors.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agopc: Replace tab with spaces
Eduardo Habkost [Thu, 14 May 2015 18:52:58 +0000 (15:52 -0300)]
pc: Replace tab with spaces

Coding style change only.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agohw/s390x/virtio-ccw: use alias property for virtio-balloon-ccw
Shannon Zhao [Mon, 11 May 2015 09:34:07 +0000 (17:34 +0800)]
hw/s390x/virtio-ccw: use alias property for virtio-balloon-ccw

Signed-off-by: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agohw/virtio/virtio-pci: use alias property for virtio-balloon-pci
Shannon Zhao [Mon, 11 May 2015 09:34:06 +0000 (17:34 +0800)]
hw/virtio/virtio-pci: use alias property for virtio-balloon-pci

Signed-off-by: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agohw/virtio/virtio-balloon: move adding property to virtio_balloon_instance_init
Shannon Zhao [Mon, 11 May 2015 09:34:05 +0000 (17:34 +0800)]
hw/virtio/virtio-balloon: move adding property to virtio_balloon_instance_init

This is in preparation for using alias property in virtio-balloon-pci
and virtio-balloon-ccw.

Signed-off-by: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agotarget-tricore: fix BOL_ST_H_LONGOFF using ld
Bastian Koppelmann [Fri, 22 May 2015 10:15:58 +0000 (12:15 +0200)]
target-tricore: fix BOL_ST_H_LONGOFF using ld

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Message-Id: <1432289758-6250-4-git-send-email-kbastian@mail.uni-paderborn.de>

9 years agotarget-tricore: fix msub32_q producing the wrong overflow bit
Bastian Koppelmann [Fri, 22 May 2015 10:15:57 +0000 (12:15 +0200)]
target-tricore: fix msub32_q producing the wrong overflow bit

The inversion of the overflow bit as a special case, which was needed for the
madd32_q instructions, does not apply for msub32_q instructions. So remove it.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Message-Id: <1432289758-6250-3-git-send-email-kbastian@mail.uni-paderborn.de>

9 years agotarget-tricore: fix OPC2_32_RR_DVINIT_HU having write before use on the result
Bastian Koppelmann [Fri, 22 May 2015 10:15:56 +0000 (12:15 +0200)]
target-tricore: fix OPC2_32_RR_DVINIT_HU having write before use on the result

If the argument r1 was the same as the extended result register r3+1, we would
overwrite r1 and then use it.

Signed-off-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Message-Id: <1432289758-6250-2-git-send-email-kbastian@mail.uni-paderborn.de>

9 years agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20150529' into...
Peter Maydell [Fri, 29 May 2015 16:10:57 +0000 (17:10 +0100)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20150529' into staging

target-arm:
 * Support ACPI for ARMv8 systems using the 'virt' board
   (and a UEFI boot image, typically)
 * avoid buffer overrun in some UNPREDICTABLE ldrd/strd cases
 * further work preparing for 64-bit EL2/EL3 support

# gpg: Signature made Fri May 29 12:14:06 2015 BST using RSA key ID 14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"

* remotes/pmaydell/tags/pull-target-arm-20150529: (39 commits)
  target-arm: Avoid buffer overrun on UNPREDICTABLE ldrd/strd
  hw/arm/virt: Enable dynamic generation of ACPI v5.1 tables
  ACPI: split CONFIG_ACPI into 4 pieces
  hw/arm/virt-acpi-build: Add PCIe controller in ACPI DSDT table
  hw/acpi/aml-build: Add Unicode macro
  hw/acpi/aml-build: Add aml_dword_io() term
  hw/acpi/aml-build: Add aml_create_dword_field() term
  hw/acpi/aml-build: Add aml_else() term
  hw/acpi/aml-build: Add aml_lnot() term
  hw/acpi/aml-build: Add aml_or() term
  hw/acpi/aml-build: Add ToUUID macro
  hw/acpi/aml-build: Make aml_buffer() definition consistent with the spec
  hw/arm/virt-acpi-build: Generate MCFG table
  hw/arm/virt-acpi-build: Generate RSDP table
  hw/arm/virt-acpi-build: Generate RSDT table
  hw/arm/virt-acpi-build: Generate GTDT table
  hw/arm/virt-acpi-build: Generate MADT table
  hw/arm/virt-acpi-build: Generate FADT table and update ACPI headers
  hw/arm/virt-acpi-build: Generation of DSDT table for virt devices
  hw/acpi/aml-build: Add aml_interrupt() term
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/armbru/tags/pull-block-2015-05-29' into staging
Peter Maydell [Fri, 29 May 2015 14:32:15 +0000 (15:32 +0100)]
Merge remote-tracking branch 'remotes/armbru/tags/pull-block-2015-05-29' into staging

Block QAPI, monitor, command line patches

# gpg: Signature made Fri May 29 12:02:32 2015 BST using RSA key ID EB918653
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>"
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>"

* remotes/armbru/tags/pull-block-2015-05-29:
  qapi: add dirty bitmap status

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-gtk-20150529-1' into staging
Peter Maydell [Fri, 29 May 2015 13:24:35 +0000 (14:24 +0100)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-gtk-20150529-1' into staging

gtk: add opengl rendering support.
small bugfixes for gtk and opengl ui code.

# gpg: Signature made Fri May 29 10:44:54 2015 BST using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"

* remotes/kraxel/tags/pull-gtk-20150529-1:
  gtk: Replace gdk_cursor_new()
  gtk: add opengl support, using egl
  ui: add egl-helpers
  ui: shader.h protect against double inclusion
  ui: use libexpoxy

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoqapi: add dirty bitmap status
John Snow [Tue, 12 May 2015 19:53:01 +0000 (15:53 -0400)]
qapi: add dirty bitmap status

Bitmaps can be in a handful of different states with potentially
more to come as we tool around with migration and persistence patches.

Management applications may need to know why certain bitmaps are
unavailable for various commands, e.g. busy in another operation,
busy being migrated, etc.

Right now, all we offer is BlockDirtyInfo's boolean member 'frozen'.
Instead of adding more booleans, replace it by an enumeration member
'status' with values 'active' and 'frozen'.  Then add new value
'disabled'.

Incompatible change.  Fine because the changed part hasn't been
released so far.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[Commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agotarget-arm: Avoid buffer overrun on UNPREDICTABLE ldrd/strd
Peter Maydell [Fri, 29 May 2015 10:29:00 +0000 (11:29 +0100)]
target-arm: Avoid buffer overrun on UNPREDICTABLE ldrd/strd

A LDRD or STRD where rd is not an even number is UNPREDICTABLE.
We were letting this fall through, which is OK unless rd is 15,
in which case we would attempt to do a load_reg or store_reg
to a nonexistent r16 for the second half of the double-word.
Catch the odd-numbered-rd cases and UNDEF them instead.

To do this we rearrange the structure of the code a little
so we can put the UNDEF catches at the top before we've
allocated TCG temporaries.

Cc: qemu-stable@nongnu.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1431348973-21315-1-git-send-email-peter.maydell@linaro.org

9 years agohw/arm/virt: Enable dynamic generation of ACPI v5.1 tables
Shannon Zhao [Fri, 29 May 2015 10:28:59 +0000 (11:28 +0100)]
hw/arm/virt: Enable dynamic generation of ACPI v5.1 tables

Initialize VirtGuestInfoState and register a machine_init_done notify to
call virt_acpi_build().

Signed-off-by: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Message-id: 1432522520-8068-25-git-send-email-zhaoshenglong@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoACPI: split CONFIG_ACPI into 4 pieces
Shannon Zhao [Fri, 29 May 2015 10:28:59 +0000 (11:28 +0100)]
ACPI: split CONFIG_ACPI into 4 pieces

As core.c, piix4.c, ich9.c and pcihp.c are for x86, add CONFIG_ACPI_X86
to make it only for x86. ARM doesn't support cpu and memory hotplug, add
CONFIG_ACPI_CPU_HOTPLUG and CONFIG_ACPI_MEMORY_HOTPLUG to exclude them
for target-arm.

Signed-off-by: Shannon Zhao <zhaoshenglong@huawei.com>
Signed-off-by: Shannon Zhao <shannon.zhao@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 1432522520-8068-24-git-send-email-zhaoshenglong@huawei.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>