sdk/emulator/qemu.git
9 years agovl: Avoid qerror_report() outside QMP command handlers
Markus Armbruster [Fri, 27 Feb 2015 08:47:12 +0000 (09:47 +0100)]
vl: Avoid qerror_report() outside QMP command handlers

qerror_report() is a transitional interface to help with converting
existing monitor commands to QMP.  It should not be used elsewhere.
Replace by error_report() in initial startup helpers parse_sandbox()
and parse_add_fd().

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Luiz Capitulino <lcapitulino@redhat.com>
9 years agoQemuOpts: Wean off qerror_report_err()
Markus Armbruster [Fri, 13 Feb 2015 11:50:26 +0000 (12:50 +0100)]
QemuOpts: Wean off qerror_report_err()

qerror_report_err() is a transitional interface to help with
converting existing monitor commands to QMP.  It should not be used
elsewhere.

The only remaining user in qemu-option.c is qemu_opts_parse().  Is it
used in QMP context?  If not, we can simply replace
qerror_report_err() by error_report_err().

The uses in qemu-img.c, qemu-io.c, qemu-nbd.c and under tests/ are
clearly not in QMP context.

The uses in vl.c aren't either, because the only QMP command handlers
there are qmp_query_status() and qmp_query_machines(), and they don't
call it.

Remaining uses:

* drive_def(): Command line -drive and such, HMP drive_add and pci_add

* hmp_chardev_add(): HMP chardev-add

* monitor_parse_command(): HMP core

* tmp_config_parse(): Command line -tpmdev

* net_host_device_add(): HMP host_net_add

* net_client_parse(): Command line -net and -netdev

* qemu_global_option(): Command line -global

* vnc_parse_func(): Command line -display, -vnc, default display, HMP
  change, QMP change.  Bummer.

* qemu_pci_hot_add_nic(): HMP pci_add

* usb_net_init(): Command line -usbdevice, HMP usb_add

Propagate errors through qemu_opts_parse().  Create a convenience
function qemu_opts_parse_noisily() that passes errors to
error_report_err().  Switch all non-QMP users outside tests to it.

That leaves vnc_parse_func().  Propagate errors through it.  Since I'm
touching it anyway, rename it to vnc_parse().

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Luiz Capitulino <lcapitulino@redhat.com>
9 years agoqdev-monitor: Propagate errors through qdev_device_add()
Markus Armbruster [Thu, 12 Mar 2015 13:00:41 +0000 (14:00 +0100)]
qdev-monitor: Propagate errors through qdev_device_add()

Also polish an error message while I'm touching the line anyway,

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Andreas Färber <afaerber@suse.de>
9 years agoqdev-monitor: Propagate errors through set_property()
Markus Armbruster [Thu, 12 Mar 2015 12:58:02 +0000 (13:58 +0100)]
qdev-monitor: Propagate errors through set_property()

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Andreas Färber <afaerber@suse.de>
9 years agoqdev-monitor: Convert qbus_find() to Error
Markus Armbruster [Wed, 11 Mar 2015 18:16:04 +0000 (19:16 +0100)]
qdev-monitor: Convert qbus_find() to Error

As usual, the conversion breaks printing explanatory messages after
the error: actual printing of the error gets delayed, so the
explanations precede rather than follow it.

Pity.  Disable them for now.  See also commit 7216ae3.

While there, eliminate QERR_BUS_NOT_FOUND, and clean up unusual
spelling in the error message.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqdev-monitor: Fix check for full bus
Markus Armbruster [Wed, 11 Mar 2015 17:39:16 +0000 (18:39 +0100)]
qdev-monitor: Fix check for full bus

Property bus has always been too screwed up to be really usable for
values other than plain bus IDs.  This just fixes a bug that crept in
in commit 1395af6 "qdev: add a maximum device allowed field for the
bus."

It doesn't always fail when it should:

    $ qemu-system-x86_64 -nodefaults -device virtio-serial-pci -device virtio-rng-device,bus=pci.0/virtio-serial-pci/virtio-bus

Happily plugs the virtio-rng-device into the virtio-bus provided by
virtio-serial-pci, even though its only slot is already occupied by a
virtio-serial-device.

And sometimes fails when it shouldn't:

    $ qemu-system-x86_64 -nodefaults -device virtio-serial-pci -device virtserialport,bus=virtio-bus/virtio-serial-device

Yes, the virtio-bus is full, but the virtio-serial-bus provided by
virtio-serial-device isn't, and that's the one we're trying to use.

Root cause: we check "bus full" when we resolve the first element of
the path.  That's the correct one only when it's also the last one.

Fix by moving the "bus full" check to right before we return a bus.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqdev-monitor: Stop error avalanche in qbus_find_recursive()
Markus Armbruster [Wed, 11 Mar 2015 16:26:31 +0000 (17:26 +0100)]
qdev-monitor: Stop error avalanche in qbus_find_recursive()

Reproducer:

    $ qemu-system-x86_64 -nodefaults -device virtio-rng-pci -device virtio-rng-pci -device virtio-rng-device,bus=virtio-bus
    qemu-system-x86_64: -device virtio-rng-device,bus=virtio-bus: Bus 'virtio-bus' is full
    qemu-system-x86_64: -device virtio-rng-device,bus=virtio-bus: Bus 'virtio-bus' is full
    qemu-system-x86_64: -device virtio-rng-device,bus=virtio-bus: Bus 'virtio-bus' not found

qbus_find_recursive() reports the "is full" error itself, and leaves
reporting "not found" to its caller.  The result is confusion.  Write
it a function contract that permits leaving all error reporting to the
caller, and implement it.  Update callers to detect and report "is
full".

Screwed up when commit 1395af6 added the max_dev limit and the "is
full" error condition to enforce it.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agodisas: Remove uses of CPU env
Peter Crosthwaite [Sun, 24 May 2015 21:20:41 +0000 (14:20 -0700)]
disas: Remove uses of CPU env

disas does not need to access the CPU env for any reason. Change the
APIs to accept CPU pointers instead. Small change pattern needs to be
applied to all target translate.c. This brings us closer to making
disas.o a common-obj and less architecture specific in general.

Cc: Richard Henderson <rth@twiddle.net>
Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: "Edgar E. Iglesias" <edgar.iglesias@gmail.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Michael Walle <michael@walle.cc>
Cc: Aurelien Jarno <aurelien@aurel32.net>
Cc: Leon Alrae <leon.alrae@imgtec.com>
Cc: Jia Liu <proljc@gmail.com>
Cc: Alexander Graf <agraf@suse.de>
Cc: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Cc: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Cc: Guan Xuetao <gxt@mprc.pku.edu.cn>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Peter Crosthwaite <crosthwaite.peter@gmail.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agomonitor: Split mon_get_cpu fn to remove ENV_GET_CPU
Peter Crosthwaite [Sun, 24 May 2015 21:20:40 +0000 (14:20 -0700)]
monitor: Split mon_get_cpu fn to remove ENV_GET_CPU

The monitor currently has one helper, mon_get_cpu() which will return
an env pointer. The target specific users of this API want an env, but
all the target agnostic users really just want the cpu pointer. These
users then need to use the target-specifically defined ENV_GET_CPU to
navigate back up to the CPU from the ENV. Split the API for the two
uses cases to remove all need for ENV_GET_CPU.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Peter Crosthwaite <crosthwaite.peter@gmail.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agomonitor: Fix failure path for "S" argument
Bandan Das [Wed, 3 Jun 2015 22:38:10 +0000 (18:38 -0400)]
monitor: Fix failure path for "S" argument

Since the "S" argument type is only used with the "?" flag,
the bug can't bite.

Signed-off-by: Bandan Das <bsd@redhat.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agomonitor: Point to "help" command on syntax error
Bandan Das [Wed, 3 Jun 2015 22:38:09 +0000 (18:38 -0400)]
monitor: Point to "help" command on syntax error

When a command fails due to incorrect syntax or input, suggest using
the "help" command to get more information about the command.  This
is only applicable for HMP.

Signed-off-by: Bandan Das <bsd@redhat.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agomonitor: cleanup parsing of cmd name and cmd arguments
Bandan Das [Wed, 3 Jun 2015 22:38:08 +0000 (18:38 -0400)]
monitor: cleanup parsing of cmd name and cmd arguments

There's too much going on in monitor_parse_command().
Split up the arguments parsing bits into a separate function
monitor_parse_arguments(). Let the original function check for
command validity and sub-commands if any and return data (*cmd)
that the newly introduced function can process and return a
QDict. Also, pass a pointer to the cmdline to track current
parser location.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Bandan Das <bsd@redhat.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agomonitor: remove debug prints
Bandan Das [Wed, 3 Jun 2015 22:38:07 +0000 (18:38 -0400)]
monitor: remove debug prints

The preferred solution is to use tracepoints and there
is good chance of bitrot with the debug prints not being
enabled at compile time. Remove them.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Bandan Das <bsd@redhat.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agoqobject: Use 'bool' inside qdict
Eric Blake [Fri, 15 May 2015 22:25:00 +0000 (16:25 -0600)]
qobject: Use 'bool' inside qdict

Now that qbool is fixed, let's fix getting and setting a bool
value to a qdict member to also use C99 bool rather than int.

I audited all callers to ensure that the changed return type
will not cause any changed semantics.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agoqobject: Use 'bool' for qbool
Eric Blake [Fri, 15 May 2015 22:24:59 +0000 (16:24 -0600)]
qobject: Use 'bool' for qbool

We require a C99 compiler, so let's use 'bool' instead of 'int'
when dealing with boolean values.  There are few enough clients
to fix them all in one pass.

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Acked-by: Luiz Capitulino <lcapitulino@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
9 years agoMerge remote-tracking branch 'remotes/afaerber/tags/qom-devices-for-peter' into staging
Peter Maydell [Mon, 22 Jun 2015 11:50:30 +0000 (12:50 +0100)]
Merge remote-tracking branch 'remotes/afaerber/tags/qom-devices-for-peter' into staging

QOM infrastructure fixes and device conversions

* Changes to name string ownership for alias properties
* Improvements around enum properties
* Cleanups around -object handling
* New helper functions
* Cleanups of qdev init helper functions
* Add path argument to qom-tree script
* QTest cleanup to use new qtest_add_data_func() consistently

# gpg: Signature made Fri Jun 19 18:14:38 2015 BST using RSA key ID 3E7E013F
# gpg: Good signature from "Andreas Färber <afaerber@suse.de>"
# gpg:                 aka "Andreas Färber <afaerber@suse.com>"

* remotes/afaerber/tags/qom-devices-for-peter:
  qdev: Un-deprecate qdev_init_nofail()
  qdev: Deprecated qdev_init() is finally unused, drop
  qom: Don't pass string table to object_get_enum() function
  qom: Add an object_property_add_enum() helper function
  qom: Make enum string tables const-correct
  qom: Add object_new_with_props() / object_new_withpropv() helpers
  qom: Add helper function for getting user objects root
  vl: Create (most) objects before creating chardev backends
  doc: Document user creatable object types in help text
  backends: Fix typename of 'policy' enum property in hostmem obj
  scripts: Add support for path as argument of qom-tree
  tests: Use qtest_add_data_func() consistently
  qdev: Free property names after registering gpio aliases
  qom: strdup() target property name on object_property_add_alias()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
Peter Maydell [Mon, 22 Jun 2015 10:50:07 +0000 (11:50 +0100)]
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* i8254 security fix
* Avoid long 100% CPU wait after restarting guests that use the periodic timer
* Fixes for access clamping (WinXP, MIPS)
* wixl/.msi support for qemu-ga on Windows

# gpg: Signature made Fri Jun 19 11:30:53 2015 BST using RSA key ID 78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>"
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini/tags/for-upstream:
  exec: clamp accesses against the MemoryRegionSection
  exec: do not clamp accesses to MMIO regions
  mc146818rtc: Reset the periodic timer on load
  qemu-timer: Call clock reset notifiers on forward jumps
  tests: virtio-scsi: Add test for unaligned WRITE SAME
  tests: virtio-scsi: Move start/stop to individual test functions
  libqos: Complete virtio device ID definition list
  libqos: Allow calling guest_free on NULL pointer
  tests: Link libqos virtio object to virtio-scsi-test
  i8254: fix out-of-bounds memory access in pit_ioport_read()
  qemu-ga: Building Windows MSI installation with configure/Makefile
  qemu-ga: Introduce Windows MSI script
  qemu-ga: debug printouts to help troubleshoot installation
  qemu-ga: adding vss-[un]install options
  qemu-log: Open file for logging when specified

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-microblaze: Remove dead code
Alistair Francis [Fri, 19 Jun 2015 04:16:52 +0000 (21:16 -0700)]
target-microblaze: Remove dead code

This code is already being run in the mb_cpu_realizefn()
function. As PVR registers are preserved on reset this
code is not required.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agos3adsp1800: Remove the hardcoded values from the reset
Alistair Francis [Fri, 19 Jun 2015 04:16:48 +0000 (21:16 -0700)]
s3adsp1800: Remove the hardcoded values from the reset

Remove the hardcoded values from the machine specific reset
function, as the same values are already set in the standard
MicroBlaze reset.

This also allows the entire reset function to be deleted, as
PVR registers are now preserved on reset.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agoml605_mmu: Move the hardcoded values to the init function
Alistair Francis [Fri, 19 Jun 2015 04:16:45 +0000 (21:16 -0700)]
ml605_mmu: Move the hardcoded values to the init function

Move the hard coded register values to the init function.
This also allows the entire reset function to be deleted, as
PVR registers are now preserved on reset.

The hardcoded PVR0 values can be removed as they are setting
the endianness and stack protection, which is already done
or invalid.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Convert pvr-full to a CPU property
Alistair Francis [Fri, 19 Jun 2015 04:16:42 +0000 (21:16 -0700)]
target-microblaze: Convert pvr-full to a CPU property

Originally the pvr-full PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Convert version_mask to a CPU property
Alistair Francis [Fri, 19 Jun 2015 04:16:38 +0000 (21:16 -0700)]
target-microblaze: Convert version_mask to a CPU property

Originally the version_mask PVR bits were manually set for each
machine. This is a hassle and difficult to read, instead set them
based on the CPU properties.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Convert endi to a CPU property
Alistair Francis [Fri, 19 Jun 2015 04:16:35 +0000 (21:16 -0700)]
target-microblaze: Convert endi to a CPU property

Originally the endi PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Convert dcache-writeback to a CPU property
Alistair Francis [Fri, 19 Jun 2015 04:16:32 +0000 (21:16 -0700)]
target-microblaze: Convert dcache-writeback to a CPU property

Originally  the dcache-writeback PVR bits were manually set for each machine.
This is a hassle and difficult to read, instead set them based on the CPU
properties.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Convert use-mmu to a CPU property
Alistair Francis [Fri, 19 Jun 2015 04:16:29 +0000 (21:16 -0700)]
target-microblaze: Convert use-mmu to a CPU property

Originally the use-mmu PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Rename the usefpu variable
Alistair Francis [Fri, 19 Jun 2015 04:16:25 +0000 (21:16 -0700)]
target-microblaze: Rename the usefpu variable

Rename the usefpu variable to use_fpu.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Disable stack protection by default
Alistair Francis [Fri, 29 May 2015 06:32:35 +0000 (16:32 +1000)]
target-microblaze: Disable stack protection by default

Stack protection is not available when the MMU is enabled.
As the MMU is enabled by default, disable stack protection
by default.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Convert use-fpu to a CPU property
Alistair Francis [Fri, 29 May 2015 06:31:58 +0000 (16:31 +1000)]
target-microblaze: Convert use-fpu to a CPU property

Originally the use-fpu PVR bits were manually set for each machine. This
is a hassle and difficult to read, instead set them based on the CPU
properties.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Tidy up the base-vectors property
Alistair Francis [Fri, 29 May 2015 06:31:20 +0000 (16:31 +1000)]
target-microblaze: Tidy up the base-vectors property

Rename the "xlnx.base-vectors" string to "base-vectors" and
move the base_vectors variable into the cfg struct.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Allow the stack protection to be disabled
Alistair Francis [Fri, 29 May 2015 06:30:43 +0000 (16:30 +1000)]
target-microblaze: Allow the stack protection to be disabled

Microblaze stack protection is configurable and isn't always enabled.
This patch allows the stack protection to be disabled from the
CPU properties.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Preserve the pvr registers during reset
Alistair Francis [Fri, 29 May 2015 06:30:05 +0000 (16:30 +1000)]
target-microblaze: Preserve the pvr registers during reset

Move the Microblaze PVR registers to the end of the CPUMBState
and preserve them during reset. This is similar to what the
QEMU ARM model does with some of it's registers.

This allows the Microblaze PVR registers to only be set once
at realise instead of constantly at reset.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agotarget-microblaze: Fix up indentation
Alistair Francis [Fri, 29 May 2015 06:29:28 +0000 (16:29 +1000)]
target-microblaze: Fix up indentation

Fix up the incorrect indentation level in the helper_stackprot() function.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agomicroblaze: s3adsp: Instantiate CPU using QOM
Peter Crosthwaite [Mon, 25 May 2015 03:31:40 +0000 (20:31 -0700)]
microblaze: s3adsp: Instantiate CPU using QOM

Instantiate and realise the CPU directly, rather than using
cpu_mb_init. Microblazes cpu_model argument is a dummy so remove the
default cpu_model set logic.

Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: Peter Crosthwaite <crosthwaite.peter@gmail.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
9 years agoqdev: Un-deprecate qdev_init_nofail()
Markus Armbruster [Fri, 19 Jun 2015 14:17:23 +0000 (16:17 +0200)]
qdev: Un-deprecate qdev_init_nofail()

It's a perfectly sensible helper function.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqdev: Deprecated qdev_init() is finally unused, drop
Markus Armbruster [Fri, 19 Jun 2015 14:17:22 +0000 (16:17 +0200)]
qdev: Deprecated qdev_init() is finally unused, drop

qdev_init() is a wrapper around setting property "realized" to true,
plus error handling that passes errors to qerror_report_err().
qerror_report_err() is a transitional interface to help with
converting existing monitor commands to QMP.  It should not be used
elsewhere.

All code has been modernized to avoid qdev_init() and its
inappropriate error handling.  We can finally drop it.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqom: Don't pass string table to object_get_enum() function
Daniel P. Berrange [Wed, 27 May 2015 15:07:56 +0000 (16:07 +0100)]
qom: Don't pass string table to object_get_enum() function

Now that properties can be explicitly registered as an enum
type, there is no need to pass the string table to the
object_get_enum() function. The object property registration
already has a pointer to the string table.

In changing this method signature, the hostmem backend object
has to be converted to use the new enum property registration
code, which simplifies it somewhat.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqom: Add an object_property_add_enum() helper function
Daniel P. Berrange [Wed, 13 May 2015 16:14:08 +0000 (17:14 +0100)]
qom: Add an object_property_add_enum() helper function

A QOM property can be parsed as enum using the visit_type_enum()
helper function, but this forces callers to use the more complex
generic object_property_add() method when registering it. It
also requires that users of that object have access to the
string map when they want to read the property value.

This patch introduces a specialized object_property_add_enum()
method which simplifies the use of enum properties, so the
setters/getters directly get passed the int value.

  typedef enum {
     MYDEV_TYPE_FROG,
     MYDEV_TYPE_ALLIGATOR,
     MYDEV_TYPE_PLATYPUS,

     MYDEV_TYPE_LAST
  } MyDevType;

Then provide a table of enum <-> string mappings

  static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = {
     [MYDEV_TYPE_FROG] = "frog",
     [MYDEV_TYPE_ALLIGATOR] = "alligator",
     [MYDEV_TYPE_PLATYPUS] = "platypus",
     [MYDEV_TYPE_LAST] = NULL,
  };

Assuming an object struct of

   typedef struct {
      Object parent_obj;
      MyDevType devtype;
      ...other fields...
   } MyDev;

The property can then be registered as follows:

   static int mydev_prop_get_devtype(Object *obj,
                                     Error **errp G_GNUC_UNUSED)
   {
       MyDev *dev = MYDEV(obj);

       return dev->devtype;
   }

   static void mydev_prop_set_devtype(Object *obj,
                                      int value,
                                      Error **errp G_GNUC_UNUSED)
   {
       MyDev *dev = MYDEV(obj);

       dev->devtype = value;
   }

   object_property_add_enum(obj, "devtype",
                            mydevtypemap, "MyDevType",
                            mydev_prop_get_devtype,
                            mydev_prop_set_devtype,
                            NULL);

Note there is no need to check the range of 'value' in
the setter, because the string->enum conversion code will
have already done that and reported an error as required.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqom: Make enum string tables const-correct
Daniel P. Berrange [Wed, 13 May 2015 16:14:07 +0000 (17:14 +0100)]
qom: Make enum string tables const-correct

The enum string table parameters in various QOM/QAPI methods
are declared 'const char *strings[]'. This results in const
warnings if passed a variable that was declared as

   static const char * const strings[] = { .... };

Add the extra const annotation to the parameters, since
neither the string elements, nor the array itself should
ever be modified.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqom: Add object_new_with_props() / object_new_withpropv() helpers
Daniel P. Berrange [Wed, 13 May 2015 16:14:06 +0000 (17:14 +0100)]
qom: Add object_new_with_props() / object_new_withpropv() helpers

It is reasonably common to want to create an object, set a
number of properties, register it in the hierarchy and then
mark it as complete (if a user creatable type). This requires
quite a lot of error prone, verbose, boilerplate code to achieve.

First a pair of functions object_set_props() / object_set_propv()
are added which allow for a list of objects to be set in
one single API call.

Then object_new_with_props() / object_new_with_propv() constructors
are added which simplify the sequence of calls to create an
object, populate properties, register in the object composition
tree and mark the object complete, into a single method call.

Usage would be:

   Error *err = NULL;
   Object *obj;
   obj = object_new_with_propv(TYPE_MEMORY_BACKEND_FILE,
                               object_get_objects_root(),
                               "hostmem0",
                               &err,
                               "share", "yes",
                               "mem-path", "/dev/shm/somefile",
                               "prealloc", "yes",
                               "size", "1048576",
                               NULL);

Note all property values are passed in string form and will
be parsed into their required data types, using normal QOM
semantics for parsing from string format.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqom: Add helper function for getting user objects root
Daniel P. Berrange [Wed, 13 May 2015 16:14:05 +0000 (17:14 +0100)]
qom: Add helper function for getting user objects root

Add object_get_objects_root() function which is a convenience for
obtaining the Object * located at /objects in the object
composition tree. Convert existing code over to use the new
API where appropriate.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agovl: Create (most) objects before creating chardev backends
Daniel P. Berrange [Wed, 13 May 2015 16:14:04 +0000 (17:14 +0100)]
vl: Create (most) objects before creating chardev backends

Some types of object must be created before chardevs, other types of
object must be created after chardevs. As such there is no option but
to create objects in two phases.

This takes the decision to create as many object types as possible
right away before anyother backends are created, and only delay
creation of those few which have an explicit dependency on the
chardevs. Hopefully the set which need delaying will remain small
over time.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agodoc: Document user creatable object types in help text
Daniel P. Berrange [Wed, 13 May 2015 16:14:03 +0000 (17:14 +0100)]
doc: Document user creatable object types in help text

The QEMU help for -object is essentially useless, just giving users
the generic syntax. Move it down into its own section and introduce
a nested table where each user creatable object can be documented.
The existing memory-backend-file, rng-random and rng-egd object
types are documented.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agobackends: Fix typename of 'policy' enum property in hostmem obj
Daniel P. Berrange [Wed, 13 May 2015 16:14:02 +0000 (17:14 +0100)]
backends: Fix typename of 'policy' enum property in hostmem obj

The 'policy' property was being registered with a typename of
'str', but it is in fact an enum of the 'HostMemPolicy' type.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20150619' into...
Peter Maydell [Fri, 19 Jun 2015 16:05:15 +0000 (17:05 +0100)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20150619' into staging

target-arm queue:
 * support --semihosting-config,arg=value
 * Cortex-R5 support (including implementing them on the Zynq board)
 * Cortex-M4 support (without FPU)
 * enable vfio-calxeda-xgmac
 * don't reset ALIAS sysregs

# gpg: Signature made Fri Jun 19 14:41:54 2015 BST using RSA key ID 14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"

* remotes/pmaydell/tags/pull-target-arm-20150619:
  semihosting: add --semihosting-config arg sub-argument
  semihosting: create SemihostingConfig structure and semihost.h
  arm: xlnx-zynqmp: Add 2xCortexR5 CPUs
  arm: xlnx-zynqmp: Add boot-cpu property
  arm: xlnx-zynqmp: Preface CPU variables with "apu"
  target-arm: Add support for Cortex-R5
  target-arm: Implement PMSAv7 MPU
  target-arm: Add registers for PMSAv7
  target-arm/helper.c: define MPUIR register
  target-arm: Do not reset sysregs marked as ALIAS
  hw/arm/sysbus-fdt: enable vfio-calxeda-xgmac dynamic instantiation
  target-arm: Add the Cortex-M4 CPU

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agosemihosting: add --semihosting-config arg sub-argument
Leon Alrae [Fri, 19 Jun 2015 13:17:45 +0000 (14:17 +0100)]
semihosting: add --semihosting-config arg sub-argument

Add new "arg" sub-argument to the --semihosting-config allowing the user
to pass multiple input arguments separately. It is required for example
by UHI semihosting to construct argc and argv.

Also, update ARM semihosting to support new option (at the moment it is
the only target which cares about arguments).

If the semihosting is enabled and no semihosting args have been specified,
then fall back to -kernel/-append. The -append string is split on whitespace
before initializing semihosting.argv[1..n]; this is different from what
QEMU MIPS machines' pseudo-bootloaders do (i.e. argv[1] contains the whole
-append), but is more intuitive from UHI user's point of view and Linux
kernel just does not care as it concatenates argv[1..n] into single cmdline
string anyway.

Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
Message-id: 1434643256-16858-3-git-send-email-leon.alrae@imgtec.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agosemihosting: create SemihostingConfig structure and semihost.h
Leon Alrae [Fri, 19 Jun 2015 13:17:45 +0000 (14:17 +0100)]
semihosting: create SemihostingConfig structure and semihost.h

Remove semihosting_enabled and semihosting_target and replace them with
SemihostingConfig structure containing equivalent fields. The structure
is defined in vl.c where it is actually set.

Also introduce separate header file include/exec/semihost.h allowing to
access semihosting config related stuff from target specific semihosting
code.

Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1434643256-16858-2-git-send-email-leon.alrae@imgtec.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoarm: xlnx-zynqmp: Add 2xCortexR5 CPUs
Peter Crosthwaite [Fri, 19 Jun 2015 13:17:45 +0000 (14:17 +0100)]
arm: xlnx-zynqmp: Add 2xCortexR5 CPUs

Add the 2xCortexR5 CPUs to zynqmp board. They are powered off on reset
(this is true of real hardware) by default or selectable as the boot
processor.

Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: da34128c73ca13fc4f8c3293e1a33d1e1e345655.1434501320.git.peter.crosthwaite@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoarm: xlnx-zynqmp: Add boot-cpu property
Peter Crosthwaite [Fri, 19 Jun 2015 13:17:45 +0000 (14:17 +0100)]
arm: xlnx-zynqmp: Add boot-cpu property

Add a string property that specifies the primary boot cpu. All CPUs
except the one selected will start-powered-off. This allows for elf
boots on any CPU, which prepares support for booting R5 elfs directly
on the R5 processors.

Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 53331c00d80c7ce9c6a83712348773f1b38fae2b.1434501320.git.peter.crosthwaite@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoarm: xlnx-zynqmp: Preface CPU variables with "apu"
Peter Crosthwaite [Fri, 19 Jun 2015 13:17:45 +0000 (14:17 +0100)]
arm: xlnx-zynqmp: Preface CPU variables with "apu"

The CPUs currently supported by zynqmp are the APU (application
processing unit) CPUs. There are other CPUs in Zynqmp so unqualified
"cpus" in ambiguous. Preface the variables with "APU" accordingly, to
prepare support adding the RPU (realtime processing unit) processors.

Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: ce32287fc365aea898465e981da3546a227e0811.1434501320.git.peter.crosthwaite@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add support for Cortex-R5
Peter Crosthwaite [Fri, 19 Jun 2015 13:17:45 +0000 (14:17 +0100)]
target-arm: Add support for Cortex-R5

Introduce a CPU model for the Cortex R5 processor. ARMv7 with MPU,
and both thumb and ARM div instructions.

Also implement dummy ATCM and BTCM. These CPs are defined for R5 but
don't have a lot of meaning in QEMU yet. Raz them so the guest can
proceed if they are read. The TCM registers will return a size of 0,
indicating no TCM.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: efe213163e6800578494aba864ac30329de4d396.1434501320.git.peter.crosthwaite@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Implement PMSAv7 MPU
Peter Crosthwaite [Fri, 19 Jun 2015 13:17:45 +0000 (14:17 +0100)]
target-arm: Implement PMSAv7 MPU

Unified MPU only. Uses ARM architecture major revision to switch
between PMSAv5 and v7 when ARM_FEATURE_MPU is set. PMSA v6 remains
unsupported and is asserted against.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: dcb03cda6dd754c5cc6a962fa11f25089811e954.1434501320.git.peter.crosthwaite@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add registers for PMSAv7
Peter Crosthwaite [Fri, 19 Jun 2015 13:17:44 +0000 (14:17 +0100)]
target-arm: Add registers for PMSAv7

Define the arm CP registers for PMSAv7 and their accessor functions.
RGNR serves as a shared index that indexes into arrays storing the
DRBAR, DRSR and DRACR registers. DRBAR and friends have to be VMSDd
separately from the CP interface using a new PMSA specific VMSD
subsection.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 172cf135fbd8f5cea413c00e71cc1c3cac704744.1434501320.git.peter.crosthwaite@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm/helper.c: define MPUIR register
Peter Crosthwaite [Fri, 19 Jun 2015 13:17:44 +0000 (14:17 +0100)]
target-arm/helper.c: define MPUIR register

Define the MPUIR register for MPU supporting ARMv6 and onwards.
Currently we only support unified MPU.

The size of the unified MPU is defined via the number of "dregions".
So just a single config is added to specify this size. (When split MPU
is implemented we will add an extra iregions config).

Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Message-id: 9f248950b803a08c8b3c978931663182f7e882e7.1434501320.git.peter.crosthwaite@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Do not reset sysregs marked as ALIAS
Sergey Fedorov [Fri, 19 Jun 2015 13:17:44 +0000 (14:17 +0100)]
target-arm: Do not reset sysregs marked as ALIAS

cp_reg_reset() is called from g_hash_table_foreach() which does not
define a specific ordering of the hash table iteration. Thus doing reset
for registers marked as ALIAS would give an ambiguous result when
resetvalue is different for original and alias registers. Exit
cp_reg_reset() early when passed an alias register. Then clean up alias
register definitions from needless resetvalue and resetfn.

In particular, this fixes a bug in the handling of the PMCR register,
which had different resetvalues for its 32 and 64-bit views.

Signed-off-by: Sergey Fedorov <serge.fdrv@gmail.com>
Message-id: 1434554713-10220-1-git-send-email-serge.fdrv@gmail.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agohw/arm/sysbus-fdt: enable vfio-calxeda-xgmac dynamic instantiation
Eric Auger [Fri, 19 Jun 2015 13:17:44 +0000 (14:17 +0100)]
hw/arm/sysbus-fdt: enable vfio-calxeda-xgmac dynamic instantiation

This patch allows the instantiation of the vfio-calxeda-xgmac device
from the QEMU command line (-device vfio-calxeda-xgmac,host="<device>").

A specialized device tree node is created for the guest, containing
compat, dma-coherent, reg and interrupts properties.

Signed-off-by: Eric Auger <eric.auger@linaro.org>
Acked-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1434455898-17895-1-git-send-email-eric.auger@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agotarget-arm: Add the Cortex-M4 CPU
Aurelio C. Remonda [Fri, 19 Jun 2015 13:17:44 +0000 (14:17 +0100)]
target-arm: Add the Cortex-M4 CPU

This patch adds the Cortex-M4 CPU. The M4 is basically the same as
the M3, the main differences being the DSP instructions and an
optional FPU.  Only no-FPU cortex-M4 is implemented here, cortex-M4F
is not because the core target-arm code doesn't support the M-profile
FPU model yet.

Signed-off-by: Aurelio C. Remonda <aurelioremonda@gmail.com>
Message-id: 1434461850-4104-1-git-send-email-aurelioremonda@gmail.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-cocoa-20150619-1' into staging
Peter Maydell [Fri, 19 Jun 2015 11:54:08 +0000 (12:54 +0100)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-cocoa-20150619-1' into staging

cocoa queue:
 * Add Machine menu, with entries for pause, resume, reset, power down, and
   media change and eject for removable drives

# gpg: Signature made Fri Jun 19 11:24:11 2015 BST using RSA key ID 14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"

* remotes/pmaydell/tags/pull-cocoa-20150619-1:
  ui/cocoa.m: Add machine menu items to change and eject removable drive media
  ui/cocoa.m: Add Reset and Power Down menu items to Machine menu
  ui/cocoa.m: Add Machine menu with pause and resume menu items

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Fri, 19 Jun 2015 10:30:57 +0000 (11:30 +0100)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

virtio, pci fixes, enhancements

Most notably this includes virtio cross-endian patches.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Fri Jun 19 11:18:05 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:
  vhost: enable vhost without without MSI-X
  pci: Don't register a specialized 'config_write' if default behavior is intended
  hw/core: rebase sysbus_get_fw_dev_path() to g_strdup_printf()
  vhost_net: re-enable when cross endian
  vhost-net: tell tap backend about the vnet endianness
  tap: fix non-linux build
  tap: add VNET_LE/VNET_BE operations
  vhost: set vring endianness for legacy virtio
  virtio: introduce virtio_legacy_is_cross_endian()
  linux-headers: sync vhost.h
  vhost-user: part of virtio

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoexec: clamp accesses against the MemoryRegionSection
Paolo Bonzini [Wed, 17 Jun 2015 08:36:54 +0000 (10:36 +0200)]
exec: clamp accesses against the MemoryRegionSection

Because the clamping was done against the MemoryRegion,
address_space_rw was effectively broken if a write spanned
multiple sections that are not linear in underlying memory
(with the memory not being under an IOMMU).

This is visible with the MIPS rc4030 IOMMU, which is implemented
as a series of alias memory regions that point to the actual RAM.

Tested-by: Hervé Poussineau <hpoussin@reactos.org>
Tested-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agoexec: do not clamp accesses to MMIO regions
Paolo Bonzini [Wed, 17 Jun 2015 08:40:27 +0000 (10:40 +0200)]
exec: do not clamp accesses to MMIO regions

It is common for MMIO registers to overlap, for example a 4 byte register
at 0xcf8 (totally random choice... :)) and a 1 byte register at 0xcf9.
If these registers are implemented via separate MemoryRegions, it is
wrong to clamp the accesses as the value written would be truncated.

Hence for these regions the effects of commit 23820db (exec: Respect
as_translate_internal length clamp, 2015-03-16, previously applied as
commit c3c1bb99) must be skipped.

Tested-by: Hervé Poussineau <hpoussin@reactos.org>
Tested-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agomc146818rtc: Reset the periodic timer on load
Paul Donohue [Fri, 12 Jun 2015 14:10:14 +0000 (10:10 -0400)]
mc146818rtc: Reset the periodic timer on load

When loading a VM from a snapshot or migration, clock changes can cause
the periodic timer to stall or loop rapidly.

qemu-timer has a reset notifier mechanism that is used to avoid timer
stalls or loops if the host clock changes while the VM is running when
using QEMU_CLOCK_HOST.  However, when loading a snapshot or migration,
qemu-timer is initialized and fires the reset notifier before
mc146818rtc is initialized and has registered its reset handler.  In
addition, this mechanism isn't used when using QEMU_CLOCK_REALTIME,
which might also change when loading a snapshot or migration.

To correct that problem, this commit resets the periodic timer after
loading from a snapshot or migration if the clock has either jumped
backward or has jumped forward by more than the clock jump limit that
is used by the reset notifier code in qemu-timer.

Signed-off-by: Paul Donohue <qemu-git@PaulSD.com>
Message-Id: <20150612141013.GE2749@TopQuark.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agoqemu-timer: Call clock reset notifiers on forward jumps
Paul Donohue [Fri, 12 Jun 2015 14:08:45 +0000 (10:08 -0400)]
qemu-timer: Call clock reset notifiers on forward jumps

Commit 691a0c9c introduced a mechanism by which QEMU_CLOCK_HOST can
notify other parts of the emulator when the host clock has jumped
backward.  This is used to avoid stalling timers that were scheduled
based on the host clock.

However, if the host clock jumps forward, then timers that were
scheduled based on the host clock may fire rapidly and cause other
problems.  For example, the mc146818rtc periodic timer will block
execution of the VM and consume host CPU while firing every interrupt
for the time period that was skipped by the host clock.

To correct that problem, this commit fires the reset notification if the
host clock jumps forward by more than a hard-coded limit.  The limit is
currently set to a value of 60 seconds, which should be small enough to
prevent excessive timer loops, but large enough to avoid frequent resets
in idle VMs.

Signed-off-by: Paul Donohue <qemu-git@PaulSD.com>
Message-Id: <20150612140845.GD2749@TopQuark.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agotests: virtio-scsi: Add test for unaligned WRITE SAME
Fam Zheng [Fri, 24 Apr 2015 11:35:20 +0000 (19:35 +0800)]
tests: virtio-scsi: Add test for unaligned WRITE SAME

This is an exercise for virtio-scsi tests using the libqos virtio
library. A few common routines are added to facilitate future extensions
of the test set.

The added test case is a regression test for the bug in d7f4b1999e.

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
9 years agoui/cocoa.m: Add machine menu items to change and eject removable drive media
John Arbuckle [Fri, 19 Jun 2015 09:53:27 +0000 (10:53 +0100)]
ui/cocoa.m: Add machine menu items to change and eject removable drive media

Adds all removable devices to the Machine menu as a Change and Eject menu
item pair. ide-cd0 would have a "Change ide-cd0..." and "Eject ide-cd0"
menu items.

Signed-off-by: John Arbuckle <programmingkidx@gmail.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoui/cocoa.m: Add Reset and Power Down menu items to Machine menu
John Arbuckle [Fri, 19 Jun 2015 09:53:27 +0000 (10:53 +0100)]
ui/cocoa.m: Add Reset and Power Down menu items to Machine menu

Add "Reset" and "Power Down" menu items to Machine menu.

Signed-off-by: John Arbuckle <programmingkidx@gmail.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agovhost: enable vhost without without MSI-X
Pankaj Gupta [Tue, 16 Jun 2015 08:18:59 +0000 (13:48 +0530)]
vhost: enable vhost without without MSI-X

We use vhostforce to enable vhost even if Guests don't have MSI-X
support and we fall back to QEMU virtio-net.

This gives a very small performance gain, but the disadvantage
is that guest now controls which virtio code is running
(qemu or vhost) so our attack surface is doubled.

This patch will enable vhost unconditionally whenever it's requested.
For compatibility, enable vhost when vhostforce is set, as well.

Signed-off-by: Pankaj Gupta <pagupta@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Jason Wang <jasowang@redhat.com>
9 years agopci: Don't register a specialized 'config_write' if default behavior is intended
Shmulik Ladkani [Tue, 16 Jun 2015 08:24:39 +0000 (11:24 +0300)]
pci: Don't register a specialized 'config_write' if default behavior is intended

Few devices have their specialized 'config_write' methods which simply
call 'pci_default_write_config' followed by a 'msix_write_config' or
'msi_write_config' calls, using exact same arguments.

This is unnecessary as 'pci_default_write_config' already invokes
'msi_write_config' and 'msix_write_config'.

Also, since 'pci_default_write_config' is the default 'config_write'
handler, we can simply avoid the registration of these specialized
versions.

Cc: Leonid Shatz <leonid.shatz@ravellosystems.com>
Signed-off-by: Shmulik Ladkani <shmulik.ladkani@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agohw/core: rebase sysbus_get_fw_dev_path() to g_strdup_printf()
Laszlo Ersek [Wed, 17 Jun 2015 12:45:03 +0000 (14:45 +0200)]
hw/core: rebase sysbus_get_fw_dev_path() to g_strdup_printf()

This is done mainly for improving readability, and in preparation for the
next patch, but Markus pointed out another bonus for the string being
returned:

"No arbitrary length limit. Before the patch, it's 39 characters, and the
code breaks catastrophically when qdev_fw_name() is longer: the second
snprintf() is called with its first argument pointing beyond path[], and
its second argument underflowing to a huge size."

Cc: qemu-stable@nongnu.org
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovhost_net: re-enable when cross endian
Cédric Le Goater [Wed, 17 Jun 2015 13:23:54 +0000 (15:23 +0200)]
vhost_net: re-enable when cross endian

Cross-endianness is now checked by the core vhost code.

revert 371df9f5e0f1 "vhost-net: disable when cross-endian"

Signed-off-by: Cédric Le Goater <clg@fr.ibm.com>
[ added commit message, Greg Kurz <gkurz@linux.vnet.ibm.com> ]
Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovhost-net: tell tap backend about the vnet endianness
Greg Kurz [Wed, 17 Jun 2015 13:23:49 +0000 (15:23 +0200)]
vhost-net: tell tap backend about the vnet endianness

The default behaviour for TAP/MACVTAP is to consider vnet as native endian.

This patch handles the cases when this is not true:
- virtio 1.0: always little-endian
- legacy cross-endian

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agotap: fix non-linux build
Michael S. Tsirkin [Thu, 18 Jun 2015 14:52:23 +0000 (16:52 +0200)]
tap: fix non-linux build

tap_fd_set_vnet_le/tap_fd_set_vnet_be was missing,
fix it up.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
9 years agoui/cocoa.m: Add Machine menu with pause and resume menu items
John Arbuckle [Fri, 19 Jun 2015 09:53:27 +0000 (10:53 +0100)]
ui/cocoa.m: Add Machine menu with pause and resume menu items

Add Machine menu to the Macintosh interface with pause
and resume menu items. These items can either pause or
resume execution of the guest operating system.

Signed-off-by: John Arbuckle <programmingkidx@gmail.com>
Message-id: 6D7AE6AA-0595-4FAD-AACF-9DFAB87248F0@gmail.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoq35: Re-enable FDC on pc-q35-2.3 and older
Eduardo Habkost [Thu, 18 Jun 2015 16:49:28 +0000 (13:49 -0300)]
q35: Re-enable FDC on pc-q35-2.3 and older

commit ea96bc629cbd52be98b2967a4b4f72e91dfc3ee4 doesn't match the patch
submitted by Laszlo to qemu-devel. We reuse pc_q35_2_4_machine_options()
inside pc_q35_2_3_machine_options(), so we need to undo the no_floppy
change in pc_q35_2_3_machine_options().

(This discrepancy was due to a bad merge.)

This restores the previous behavior where all the 2.3 and older machines
had no_floppy=0.

Reported-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-id: 1434646168-3100-1-git-send-email-ehabkost@redhat.com
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
[PMM: mention that this was a merge issue, not a review issue]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoscripts: Add support for path as argument of qom-tree
Martin Cerveny [Wed, 13 May 2015 12:14:54 +0000 (14:14 +0200)]
scripts: Add support for path as argument of qom-tree

Add processing of optional argument path as "tree base".

Signed-off-by: Martin Cerveny <M.Cerveny@computer.org>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agotests: Use qtest_add_data_func() consistently
Andreas Färber [Wed, 25 Mar 2015 17:40:15 +0000 (18:40 +0100)]
tests: Use qtest_add_data_func() consistently

Replace uses of g_test_add_data_func() for QTest test cases.

It is still valid to use it for any non-QTest test cases,
which are not run for multiple target binaries.

Suggested-by: John Snow <jsnow@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqdev: Free property names after registering gpio aliases
Eduardo Habkost [Thu, 9 Apr 2015 19:57:30 +0000 (16:57 -0300)]
qdev: Free property names after registering gpio aliases

Now that object_property_add_alias() strdup()s target_name, we can free
the property names in qdev_pass_gpios().

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoqom: strdup() target property name on object_property_add_alias()
Eduardo Habkost [Thu, 9 Apr 2015 19:57:29 +0000 (16:57 -0300)]
qom: strdup() target property name on object_property_add_alias()

With this, object_property_add_alias() callers can safely free the
target property name, like what already happens with the 'name' argument
to all object_property_add*() functions.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
9 years agoMerge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2015-06-18' into staging
Peter Maydell [Thu, 18 Jun 2015 12:32:39 +0000 (13:32 +0100)]
Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2015-06-18' into staging

QAPI patches

# gpg: Signature made Thu Jun 18 13:20:00 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-qapi-2015-06-18:
  qapi-types: Bury code dead since commit 6b5abc7
  qapi-types: Split generate_fwd_builtin() off generate_fwd_struct()
  qapi-types: Drop unused members parameters
  qapi-types: Don't filter out expressions with 'gen'
  qapi: Catch and reject flat union branch of array type
  tests/qapi-schema: New flat union array branch test case
  qapi: Better separate the different kinds of helpers
  qapi: Move exprs checking from parse_schema() to check_exprs()
  qapi: Fix to reject stray 't', 'f' and 'n'
  qapi: Simplify inclusion cycle detection
  qapi: Fix file name in error messages for included files
  qapi: Improve a couple of confusing variable names
  qapi: Eliminate superfluous QAPISchema attribute input_dir
  qapi: Drop bogus command from docs
  MAINTAINERS: Fix up QAPI and QAPI schema file patterns

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoqapi-types: Bury code dead since commit 6b5abc7
Markus Armbruster [Fri, 12 Jun 2015 07:45:55 +0000 (09:45 +0200)]
qapi-types: Bury code dead since commit 6b5abc7

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi-types: Split generate_fwd_builtin() off generate_fwd_struct()
Markus Armbruster [Fri, 12 Jun 2015 07:22:32 +0000 (09:22 +0200)]
qapi-types: Split generate_fwd_builtin() off generate_fwd_struct()

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi-types: Drop unused members parameters
Markus Armbruster [Fri, 12 Jun 2015 08:40:17 +0000 (10:40 +0200)]
qapi-types: Drop unused members parameters

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi-types: Don't filter out expressions with 'gen'
Markus Armbruster [Fri, 12 Jun 2015 06:32:51 +0000 (08:32 +0200)]
qapi-types: Don't filter out expressions with 'gen'

Useless, because it can only occur in commands, and we're not dealing
with commands here.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Catch and reject flat union branch of array type
Markus Armbruster [Wed, 10 Jun 2015 11:07:43 +0000 (13:07 +0200)]
qapi: Catch and reject flat union branch of array type

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agotests/qapi-schema: New flat union array branch test case
Markus Armbruster [Wed, 10 Jun 2015 11:03:04 +0000 (13:03 +0200)]
tests/qapi-schema: New flat union array branch test case

The new test demonstrates another generator crash.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Better separate the different kinds of helpers
Markus Armbruster [Wed, 10 Jun 2015 08:04:36 +0000 (10:04 +0200)]
qapi: Better separate the different kinds of helpers

Insert comments to separate sections dealing with parsing, semantic
analysis, code generation, and so forth.

Move helpers to their proper section.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Move exprs checking from parse_schema() to check_exprs()
Markus Armbruster [Wed, 10 Jun 2015 06:55:21 +0000 (08:55 +0200)]
qapi: Move exprs checking from parse_schema() to check_exprs()

To have expression semantic analysis in one place rather than two.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Fix to reject stray 't', 'f' and 'n'
Markus Armbruster [Wed, 10 Jun 2015 06:24:58 +0000 (08:24 +0200)]
qapi: Fix to reject stray 't', 'f' and 'n'

Screwed up in commit e53188a.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Simplify inclusion cycle detection
Markus Armbruster [Tue, 9 Jun 2015 14:54:09 +0000 (16:54 +0200)]
qapi: Simplify inclusion cycle detection

We maintain a stack of filenames in include_hist for convenient cycle
detection.

As error_path() demonstrates, the same information is readily
available in the expr_info, so just use that, and drop include_hist.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Fix file name in error messages for included files
Markus Armbruster [Tue, 9 Jun 2015 16:32:29 +0000 (18:32 +0200)]
qapi: Fix file name in error messages for included files

We print the name as it appears in the include expression.  Tools
processing error messages want it relative to the working directory.
Make it so.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Improve a couple of confusing variable names
Markus Armbruster [Tue, 9 Jun 2015 14:22:45 +0000 (16:22 +0200)]
qapi: Improve a couple of confusing variable names

old name      new name
----------------------------
input_file    fname
input_relname fname
input_fname   abs_fname
include_path  incl_abs_fname
parent_info   incl_info

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Eliminate superfluous QAPISchema attribute input_dir
Markus Armbruster [Tue, 9 Jun 2015 14:49:13 +0000 (16:49 +0200)]
qapi: Eliminate superfluous QAPISchema attribute input_dir

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoqapi: Drop bogus command from docs
Markus Armbruster [Tue, 9 Jun 2015 12:38:58 +0000 (14:38 +0200)]
qapi: Drop bogus command from docs

Commit 87a560c4 added it in the wrong place.  Commit 59a2c4ce added it
in the right place, but didn't remove it from the wrong place.  Do
that now.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoMAINTAINERS: Fix up QAPI and QAPI schema file patterns
Markus Armbruster [Fri, 12 Jun 2015 13:15:54 +0000 (15:15 +0200)]
MAINTAINERS: Fix up QAPI and QAPI schema file patterns

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
9 years agoMerge remote-tracking branch 'remotes/mcayland/tags/qemu-openbios-signed' into staging
Peter Maydell [Thu, 18 Jun 2015 10:36:42 +0000 (11:36 +0100)]
Merge remote-tracking branch 'remotes/mcayland/tags/qemu-openbios-signed' into staging

Update OpenBIOS images

# gpg: Signature made Wed Jun 17 20:06:06 2015 BST using RSA key ID AE0F321F
# gpg: Good signature from "Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>"

* remotes/mcayland/tags/qemu-openbios-signed:
  Update OpenBIOS images

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agovfio: fix build error on CentOS 5.7
Leon Alrae [Wed, 17 Jun 2015 12:35:00 +0000 (13:35 +0100)]
vfio: fix build error on CentOS 5.7

Include linux/vfio.h after sys/ioctl.h, just like in hw/vfio/common.c.

Signed-off-by: Leon Alrae <leon.alrae@imgtec.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Message-id: 1434544500-22405-1-git-send-email-leon.alrae@imgtec.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
9 years agoUpdate OpenBIOS images
Mark Cave-Ayland [Wed, 17 Jun 2015 19:02:15 +0000 (20:02 +0100)]
Update OpenBIOS images

Update OpenBIOS images to SVN r1340 built from submodule.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
9 years agotap: add VNET_LE/VNET_BE operations
Greg Kurz [Wed, 17 Jun 2015 13:23:44 +0000 (15:23 +0200)]
tap: add VNET_LE/VNET_BE operations

The linux tap and macvtap backends can be told to parse vnet headers
according to little or big endian. This is done through the TUNSETVNETLE
and TUNSETVNETBE ioctls.

This patch brings all the plumbing for QEMU to use these APIs.

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovhost: set vring endianness for legacy virtio
Greg Kurz [Wed, 17 Jun 2015 13:23:39 +0000 (15:23 +0200)]
vhost: set vring endianness for legacy virtio

Legacy virtio is native endian: if the guest and host endianness differ,
we have to tell vhost so it can swap bytes where appropriate. This is
done through a vhost ring ioctl.

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agovirtio: introduce virtio_legacy_is_cross_endian()
Greg Kurz [Wed, 17 Jun 2015 13:23:34 +0000 (15:23 +0200)]
virtio: introduce virtio_legacy_is_cross_endian()

This helper will be used by vhost and tap to detect cross-endianness in
the legacy virtio case.

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
9 years agolinux-headers: sync vhost.h
Greg Kurz [Wed, 17 Jun 2015 13:23:29 +0000 (15:23 +0200)]
linux-headers: sync vhost.h

This patch brings the cross-endian vhost API to QEMU.

Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>