sdk/emulator/qemu.git
10 years agokeyboard: Maru keyboard device handles key events directly 83/19883/1
GiWoong Kim [Tue, 22 Apr 2014 10:33:22 +0000 (19:33 +0900)]
keyboard: Maru keyboard device handles key events directly

Change-Id: I18050322c30003656ecb73f0819547821fd196ba
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agotouch: Maru touchscreen device handles touch events directly 63/19863/1
GiWoong Kim [Tue, 22 Apr 2014 07:53:47 +0000 (16:53 +0900)]
touch: Maru touchscreen device handles touch events directly

Change-Id: I40f4af19103e7a2add265f3bcefe1f62d03fa2b7
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agobuild: fix a invalid merge
SeokYeon Hwang [Mon, 21 Apr 2014 08:11:52 +0000 (17:11 +0900)]
build: fix a invalid merge

Change-Id: I505dd07f471c3f2f6a974f1e06df90997cb5a62e
Signed-off-by: SeokYeon Hwang <syeon.hwang@samsung.com>
10 years agovigs: fix compilation errors
SeokYeon Hwang [Mon, 21 Apr 2014 07:03:05 +0000 (16:03 +0900)]
vigs: fix compilation errors

1. Some qemu function prototype modification.
graphic_console_init()
qemu_thread_create()

2. A "work_queue.c" is moved to "vigs".

Change-Id: I066734b08bb8ad59aeb5afb67049065048bcdd63
Signed-off-by: SeokYeon Hwang <syeon.hwang@samsung.com>
10 years agoMerge branch 'upstream_qemu_2.0' into tizen
SeokYeon Hwang [Mon, 21 Apr 2014 06:32:27 +0000 (15:32 +0900)]
Merge branch 'upstream_qemu_2.0' into tizen

Initial merge

Signed-off-by: SeokYeon Hwang <syeon.hwang@samsung.com>
Conflicts:
.gitignore
Makefile.target
block/raw-win32.c
configure
cpu-exec.c
cpus.c
hw/9pfs/virtio-9p-device.c
hw/i386/pc_piix.c
include/exec/cpu-defs.h
include/qemu-common.h
include/sysemu/sysemu.h
include/ui/console.h
tcg/i386/tcg-target.c
tcg/tcg.h
ui/console.c
ui/input.c
util/oslib-posix.c
vl.c

Change-Id: I52df9052942eea8d541bde75fd936e58a9f2f8ff

10 years agoUpdate version for v2.0.0 release
Peter Maydell [Thu, 17 Apr 2014 12:41:45 +0000 (13:41 +0100)]
Update version for v2.0.0 release

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoYaGL: Fix transport separate short command skip bug 51/19751/1
Stanislav Vorobiov [Thu, 17 Apr 2014 08:28:36 +0000 (12:28 +0400)]
YaGL: Fix transport separate short command skip bug

Separate short commands (1 unit long) could be skipped due
to incorrect check, this is now fixed

Change-Id: I8736abb171e35563580407a25cdffadfc589ed18
Signed-off-by: Stanislav Vorobiov <s.vorobiov@samsung.com>
10 years agoMerge "package: version up" into tizen
SeokYeon Hwang [Thu, 17 Apr 2014 06:18:58 +0000 (23:18 -0700)]
Merge "package: version up" into tizen

10 years agoMerge "display: do not call qemu_ds_sdl_update directly" into tizen
SeokYeon Hwang [Thu, 17 Apr 2014 06:18:41 +0000 (23:18 -0700)]
Merge "display: do not call qemu_ds_sdl_update directly" into tizen

10 years agoMerge "VIGS: Add framebuffer and temporary texture pool" into tizen
SeokYeon Hwang [Thu, 17 Apr 2014 06:18:11 +0000 (23:18 -0700)]
Merge "VIGS: Add framebuffer and temporary texture pool" into tizen

10 years agopackage: version up 75/19675/1
GiWoong Kim [Thu, 17 Apr 2014 05:34:21 +0000 (14:34 +0900)]
package: version up

1.7.35

Change-Id: I62d0e8dc3dd31955a2bb30f9e4db5dc24efe2f1b
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agodisplay: do not call qemu_ds_sdl_update directly 74/19674/1
GiWoong Kim [Thu, 17 Apr 2014 05:20:22 +0000 (14:20 +0900)]
display: do not call qemu_ds_sdl_update directly

qemu_update should be synchronized with hw_update.
Otherwise, image processing like a brightness compositing
can be multiple overlapped on display surface.

Change-Id: Icab65289a6f279b3e61b045319cb02383af591e2
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agoSENSOR: bug fixed to set gyroscope & light 76/19576/2
Jinhyung Choi [Wed, 16 Apr 2014 02:46:02 +0000 (11:46 +0900)]
SENSOR: bug fixed to set gyroscope & light

Change-Id: I6549f8432ad6b0c7aac80de6120555fce9d2a2df
Signed-off-by: Jinhyung Choi <jinhyung2.choi@samsung.com>
10 years agoVIGS: Add framebuffer and temporary texture pool 68/19568/1
Stanislav Vorobiov [Tue, 15 Apr 2014 11:24:31 +0000 (15:24 +0400)]
VIGS: Add framebuffer and temporary texture pool

In order not to waste host GPU resources we now
have 2 OpenGL object pools - for temporary textures
and for framebuffers. These used to be allocated for
every surface, but now we allocate only one object
per distinct format, i.e. one
object per (width, height, format) tuple

Change-Id: Ie63e571045a2477043e8e08da3a50765523e54fc
Signed-off-by: Stanislav Vorobiov <s.vorobiov@samsung.com>
10 years agoMerge "timer: fix qemu_poll_ns early timeout on windows" into tizen
SeokYeon Hwang [Tue, 15 Apr 2014 05:42:18 +0000 (22:42 -0700)]
Merge "timer: fix qemu_poll_ns early timeout on windows" into tizen

10 years agoMerge "display: Add dirty flag for display" into tizen
SeokYeon Hwang [Tue, 15 Apr 2014 02:07:27 +0000 (19:07 -0700)]
Merge "display: Add dirty flag for display" into tizen

10 years agoMerge "ecs : ecs log change to emulator log" into tizen
SeokYeon Hwang [Tue, 15 Apr 2014 02:06:22 +0000 (19:06 -0700)]
Merge "ecs : ecs log change to emulator log" into tizen

10 years agoUpdate version for v2.0.0-rc3 release
Peter Maydell [Mon, 14 Apr 2014 16:45:11 +0000 (17:45 +0100)]
Update version for v2.0.0-rc3 release

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoRevert "fix return check for KVM_GET_DIRTY_LOG ioctl"
Michael Tokarev [Mon, 14 Apr 2014 12:14:04 +0000 (16:14 +0400)]
Revert "fix return check for KVM_GET_DIRTY_LOG ioctl"

This reverts commit b533f658a98325d0e47b36113bd9f5bcc046fdae.

The original code was wrong, because effectively it ignored errors
from kernel, because kernel does not return -1 on error case but
returns -errno, and does not return -EPERM for this particular ioctl.
But in some cases kernel actually returned unsuccessful result,
namely, when the dirty bitmap in requested slot does not exist
it returns -ENOENT.  With new code this condition becomes an
error when it shouldn't be.

Revert that patch instead of fixing it properly this late in the
release process.  I disagree with this approach, but let's make
things move _somewhere_, instead of arguing endlessly whch of
the 2 proposed fixes is better.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Message-id: 1397477644-902-1-git-send-email-mjt@msgid.tls.msk.ru
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Mon, 14 Apr 2014 13:02:12 +0000 (14:02 +0100)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

acpi: SSDT update

This has a fix by Igor for a regression introduced by
bridge hotplug code.
Expected test files were updated accordingly.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Mon 14 Apr 2014 13:13:35 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-test: update expected files
  acpi: fix incorrect encoding for 0x{F-1}FFFF

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoide: Correct improper smart self test counter reset in ide core.
Benoît Canet [Sat, 12 Apr 2014 20:59:50 +0000 (22:59 +0200)]
ide: Correct improper smart self test counter reset in ide core.

The SMART self test counter was incorrectly being reset to zero,
not 1. This had the effect that on every 21st SMART EXECUTE OFFLINE:
 * We would write off the beginning of a dynamically allocated buffer
 * We forgot the SMART history
Fix this.

Signed-off-by: Benoit Canet <benoit@irqsave.net>
Message-id: 1397336390-24664-1-git-send-email-benoit.canet@irqsave.net
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Cc: qemu-stable@nongnu.org
Acked-by: Kevin Wolf <kwolf@redhat.com>
[PMM: tweaked commit message as per suggestions from Markus]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoacpi-test: update expected files
Michael S. Tsirkin [Mon, 14 Apr 2014 12:08:37 +0000 (15:08 +0300)]
acpi-test: update expected files

commit 58b035c7354afc0c5351ea62264c01d74196ec26
    acpi: fix incorrect encoding for 0x{F-1}FFFF
changes the SSDT, update expected files accordingly.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 years agoacpi: fix incorrect encoding for 0x{F-1}FFFF
Igor Mammedov [Sun, 13 Apr 2014 21:55:51 +0000 (23:55 +0200)]
acpi: fix incorrect encoding for 0x{F-1}FFFF

Fix typo in build_append_int() which causes integer
truncation when it's in range 0x{F-1}FFFF by packing it
as WordConst instead of required DWordConst.

In partucular this fixes a regression: hotplug in slots 16,17,18 and 19
didn't work, since SSDT had code like this:

                If (And (Arg0, 0x0000))
                {
                    Notify (S80, Arg1)
                }

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
10 years agoconfigure: Make stack-protector test check both compile and link
Peter Maydell [Fri, 11 Apr 2014 16:13:52 +0000 (17:13 +0100)]
configure: Make stack-protector test check both compile and link

Since we use the -fstack-protector argument at both compile and
link time in the build, we must check that it works with both
a compile and a link:
 * MacOSX only fails in the compile step, not linking
 * some gcc cross environments only fail at the link stage (if they
   require a libssp and it's not present for some reason)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1397232832-32301-1-git-send-email-peter.maydell@linaro.org
Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru>
10 years agovmxnet3: validate queues configuration read on migration
Dmitry Fleytman [Fri, 4 Apr 2014 09:45:22 +0000 (12:45 +0300)]
vmxnet3: validate queues configuration read on migration

CVE-2013-4544

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>
Reported-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-id: 1396604722-11902-5-git-send-email-dmitry@daynix.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agovmxnet3: validate interrupt indices read on migration
Dmitry Fleytman [Fri, 4 Apr 2014 09:45:21 +0000 (12:45 +0300)]
vmxnet3: validate interrupt indices read on migration

CVE-2013-4544

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>
Reported-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-id: 1396604722-11902-4-git-send-email-dmitry@daynix.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agovmxnet3: validate queues configuration coming from guest
Dmitry Fleytman [Fri, 4 Apr 2014 09:45:20 +0000 (12:45 +0300)]
vmxnet3: validate queues configuration coming from guest

CVE-2013-4544

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>
Reported-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-id: 1396604722-11902-3-git-send-email-dmitry@daynix.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agovmxnet3: validate interrupt indices coming from guest
Dmitry Fleytman [Fri, 4 Apr 2014 09:45:19 +0000 (12:45 +0300)]
vmxnet3: validate interrupt indices coming from guest

CVE-2013-4544

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>
Reported-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-id: 1396604722-11902-2-git-send-email-dmitry@daynix.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agodisplay: Add dirty flag for display 84/19484/2
GiWoong Kim [Fri, 11 Apr 2014 07:01:58 +0000 (16:01 +0900)]
display: Add dirty flag for display

At idle screen moment, there is no need to copy framebuffers
to the display.

Change-Id: I4006e5732de947ce1617161182ca901b179c9423
Signed-off-by: Stanislav Vorobiov <s.vorobiov@samsung.com>
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agoMerge changes I1b071d12,I55ce8eab into tizen
SeokYeon Hwang [Mon, 14 Apr 2014 07:12:14 +0000 (00:12 -0700)]
Merge changes I1b071d12,I55ce8eab into tizen

* changes:
  build: add enable-curl option
  package: version up

10 years agoecs : ecs log change to emulator log 30/19330/3
haken kim [Thu, 10 Apr 2014 10:32:51 +0000 (19:32 +0900)]
ecs : ecs log change to emulator log

Change-Id: I9bc28ff56764037a5045e3e6b1b6d55ad5a38332
Signed-off-by: haken kim <haken.kim@samsung.com>
10 years agotimer: fix qemu_poll_ns early timeout on windows 14/19414/2
Sangho Park [Fri, 11 Apr 2014 08:50:52 +0000 (12:50 +0400)]
timer: fix qemu_poll_ns early timeout on windows

g_poll has a problem on windows when using
timeouts < 10ms, in glib/gpoll.c:

/* If not, and we have a significant timeout, poll again with
 * timeout then. Note that this will return indication for only
 * one event, or only for messages. We ignore timeouts less than
 * ten milliseconds as they are mostly pointless on Windows, the
 * MsgWaitForMultipleObjectsEx() call will timeout right away
 * anyway.
 */
if (retval == 0 && (timeout == INFINITE || timeout >= 10))
  retval = poll_rest (poll_msgs, handles, nhandles, fds, nfds, timeout);

so whenever g_poll is called with timeout < 10ms it does
a quick poll instead of wait, this causes significant performance
degradation of qemu, thus we should use WaitForMultipleObjectsEx
directly

Change-Id: I95c7be6a57b9458e5ff2b82605aa5d7c4b6d303b
Signed-off-by: Stanislav Vorobiov <s.vorobiov@samsung.com>
10 years agoqom: Fix crash with qom-list and link properties
Cole Robinson [Thu, 10 Apr 2014 18:47:38 +0000 (14:47 -0400)]
qom: Fix crash with qom-list and link properties

Commit 9561fda8d90e176bef598ba87c42a1bd6ad03ef7 changed the type of
'opaque' for link properties, but missed updating this call site.
Reproducer:

./x86_64-softmmu/qemu-system-x86_64 -qmp unix:./qmp.sock,server &
./scripts/qmp/qmp-shell ./qmp.sock
(QEMU) qom-list path=//machine/i440fx/pci.0/child[2]

Reported-by: Marcin Gibuła <m.gibula@beyond.pl>
Signed-off-by: Cole Robinson <crobinso@redhat.com>
Message-id: 2f8f007ce2152ac3b65f0811199662799c509225.1397155389.git.crobinso@redhat.com
Acked-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agovirtio-net: fix guest-triggerable buffer overrun
Michael S. Tsirkin [Fri, 11 Apr 2014 12:18:08 +0000 (15:18 +0300)]
virtio-net: fix guest-triggerable buffer overrun

When VM guest programs multicast addresses for
a virtio net card, it supplies a 32 bit
entries counter for the number of addresses.
These addresses are read into tail portion of
a fixed macs array which has size MAC_TABLE_ENTRIES,
at offset equal to in_use.

To avoid overflow of this array by guest, qemu attempts
to test the size as follows:
-    if (in_use + mac_data.entries <= MAC_TABLE_ENTRIES) {

however, as mac_data.entries is uint32_t, this sum
can overflow, e.g. if in_use is 1 and mac_data.entries
is 0xffffffff then in_use + mac_data.entries will be 0.

Qemu will then read guest supplied buffer into this
memory, overflowing buffer on heap.

CVE-2014-0150

Cc: qemu-stable@nongnu.org
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 1397218574-25058-1-git-send-email-mst@redhat.com
Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoMerge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Peter Maydell [Fri, 11 Apr 2014 13:07:24 +0000 (14:07 +0100)]
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block patches for 2.0.0-rc3

# gpg: Signature made Fri 11 Apr 2014 13:37:34 BST using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"

* remotes/kevin/tags/for-upstream:
  block-commit: speed is an optional parameter
  iscsi: Remember to set ret for iscsi_open in error case
  bochs: Fix catalog size check
  bochs: Fix memory leak in bochs_open() error path

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-sdl-1' into staging
Peter Maydell [Fri, 11 Apr 2014 12:51:15 +0000 (13:51 +0100)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-sdl-1' into staging

sdl2 relative mouse mode fixes.

# gpg: Signature made Fri 11 Apr 2014 11:36:46 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-sdl-1:
  input: sdl2: Fix relative mode to match SDL1 behavior
  input: sdl2: Fix guest_cursor logic

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoblock-commit: speed is an optional parameter
Max Reitz [Thu, 10 Apr 2014 17:36:25 +0000 (19:36 +0200)]
block-commit: speed is an optional parameter

As speed is an optional parameter for the QMP block-commit command, it
should be set to 0 if not given (as it is undefined if has_speed is
false), that is, the speed should not be limited.

Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
10 years agoiscsi: Remember to set ret for iscsi_open in error case
Fam Zheng [Thu, 10 Apr 2014 01:33:55 +0000 (09:33 +0800)]
iscsi: Remember to set ret for iscsi_open in error case

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
10 years agobochs: Fix catalog size check
Kevin Wolf [Wed, 9 Apr 2014 10:10:34 +0000 (12:10 +0200)]
bochs: Fix catalog size check

The old check was off by a factor of 512 and didn't consider cases where
we don't get an exact division. This could lead to an out-of-bounds
array access in seek_to_sector().

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
10 years agobochs: Fix memory leak in bochs_open() error path
Kevin Wolf [Wed, 9 Apr 2014 09:19:04 +0000 (11:19 +0200)]
bochs: Fix memory leak in bochs_open() error path

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
10 years agoinput: sdl2: Fix relative mode to match SDL1 behavior
Cole Robinson [Tue, 1 Apr 2014 20:37:11 +0000 (16:37 -0400)]
input: sdl2: Fix relative mode to match SDL1 behavior

Right now relative mode accelerates too fast, and has the 'invisible wall'
problem. SDL2 added an explicit API to handle this use case, so let's use
it.

Signed-off-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agoinput: sdl2: Fix guest_cursor logic
Cole Robinson [Tue, 1 Apr 2014 20:37:10 +0000 (16:37 -0400)]
input: sdl2: Fix guest_cursor logic

Unbreaks relative mouse mode with sdl2, just like was done with sdl.c
in c3aa84b6.

Signed-off-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agobuild: add enable-curl option 82/19482/1
Kitae Kim [Fri, 11 Apr 2014 09:58:24 +0000 (18:58 +0900)]
build: add enable-curl option

libcurl is used for osutil. It has to be checked at configure time.

Change-Id: I1b071d129b394a4edd57f6e86a71f6e82e2f611f
Signed-off-by: Kitae Kim <kt920.kim@samsung.com>
10 years agocommunication: modified hb signal period 59/19359/1
GiWoong Kim [Wed, 9 Apr 2014 05:48:59 +0000 (14:48 +0900)]
communication: modified hb signal period

The initial period of HB signal needs a bit tuning
in case of low performance machine.

Change-Id: I08274b7ea6526aea9d356e5474bd0aa08236922a
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agopackage: version up 61/19361/1
Kitae Kim [Fri, 11 Apr 2014 06:12:37 +0000 (15:12 +0900)]
package: version up

1.7.34

Change-Id: I55ce8eaba4cdb53a1f6dbf9323e5336e6e57a5ae
Signed-off-by: Kitae Kim <kt920.kim@samsung.com>
10 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Thu, 10 Apr 2014 22:07:55 +0000 (23:07 +0100)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

acpi: DSDT update

Two fixes here:
- Test fix to avoid warning with make check.
- Hex file update so people building QEMU
  without installing iasl get exactly the same ACPI
  as with.

Both should help avoid user confusion.

As it's very easy to check that the produced ACPI
binary didn't change, I think these are very low risk.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Thu 10 Apr 2014 17:09:43 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: update generated hex files
  tests/acpi: update expected DSDT files

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoconfigure: use do_cc when checking for -fstack-protector support
Peter Maydell [Wed, 9 Apr 2014 11:04:47 +0000 (12:04 +0100)]
configure: use do_cc when checking for -fstack-protector support

MacOSX clang silently swallows unrecognized -f options when doing a link
with '-framework' also on the command line, so to detect support for
the various -fstack-protector options we must do a plain .c to .o compile,
not a complete compile-and-link.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 1397041487-28477-1-git-send-email-peter.maydell@linaro.org

10 years agoacpi: update generated hex files
Michael S. Tsirkin [Thu, 10 Apr 2014 16:03:18 +0000 (19:03 +0300)]
acpi: update generated hex files

commit f2ccc311df55ec026a8f8ea9df998f26314f22b2
    dsdt: tweak ACPI ID for hotplug resource device
changes the DSDT, update hex files to match

Otherwise the fix is only effective if QEMU is built
with iasl.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 years agopackage: version up 92/19292/1
GiWoong Kim [Thu, 10 Apr 2014 05:16:28 +0000 (14:16 +0900)]
package: version up

1.7.33

Change-Id: I6218a3e6b132ddd744888293a00aa8bdaabe5e77
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agoemulator: add kernel parameter 91/19291/2
GiWoong Kim [Thu, 10 Apr 2014 03:44:18 +0000 (12:44 +0900)]
emulator: add kernel parameter

In some cases, guest needs VM resolution value.
(e.g. touchscreen driver, model-config.xml)
In order to read that values from host easily,
I added resolution parameters for kernel.

Change-Id: I615ec909e517e2628cc727ec77d7d86a18d0c8c8
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agotests/acpi: update expected DSDT files
Michael S. Tsirkin [Wed, 9 Apr 2014 14:47:07 +0000 (17:47 +0300)]
tests/acpi: update expected DSDT files

commit f2ccc311df55ec026a8f8ea9df998f26314f22b2
    dsdt: tweak ACPI ID for hotplug resource device
changes the DSDT, update test expected files to match

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reported-by: Igor Mammedov <imammedo@redhat.com>
10 years agopackage : version up 31/19231/1
haken kim [Wed, 9 Apr 2014 06:55:11 +0000 (15:55 +0900)]
package : version up

1.7.32

Change-Id: I5024f5387e68792d6f97a99a1f03c5861cb26488
Signed-off-by: haken kim <haken.kim@samsung.com>
10 years agoMerge "skin: using sensor device" into tizen
SeokYeon Hwang [Wed, 9 Apr 2014 06:30:05 +0000 (23:30 -0700)]
Merge "skin: using sensor device" into tizen

10 years agoUpdate version for v2.0.0-rc2 release
Peter Maydell [Tue, 8 Apr 2014 17:52:06 +0000 (18:52 +0100)]
Update version for v2.0.0-rc2 release

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agohw/pci-host/prep: Don't reverse IO accesses on bigendian hosts
Peter Maydell [Tue, 8 Apr 2014 15:51:11 +0000 (16:51 +0100)]
hw/pci-host/prep: Don't reverse IO accesses on bigendian hosts

The raven_io_read() and raven_io_write() functions pass and
return values in little-endian format (since the IO op struct
is marked DEVICE_LITTLE_ENDIAN); however they were storing the
values in the buffer to pass to address_space_read/write()
in host-endian order, which meant that on big-endian hosts
the values were inadvertently reversed. Use the *_le_p()
accessors instead so that we are consistent regardless of
host endianness.

Strictly speaking the byte order of the buffer for
address_space_rw() is target byte order (which for PPC
will be BE) but it doesn't actually matter as long as we
are consistent about the marking on the IO op struct and
which stl_*_p().

This bug was probably introduced due to confusion caused by
the two different versions of ldl_p() and friends:
 bswap.h defines versions meaning "host endianness access"
 cpu-all.h defines versions meaning "target endianness access"
As a target-independent source file prep.c gets the bswap.h
versions; the very similar looking code in ioport.c is
compiled per-target and gets the cpu-all.h versions.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1396972271-22660-1-git-send-email-peter.maydell@linaro.org
Reviewed-by: Richard Henderson <rth@twiddle.net>
10 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Tue, 8 Apr 2014 12:59:28 +0000 (13:59 +0100)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

acpi bug fix

Here is a single last minute fix for 2.0

This changes the HID of the container used to claim
resources for CPU hotplug.
As a result, windows XP SP3 no longer brings up
an annoying "found new hardware" wizard on boot.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Tue 08 Apr 2014 13:23:30 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:
  dsdt: tweak ACPI ID for hotplug resource device

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agodsdt: tweak ACPI ID for hotplug resource device
Michael S. Tsirkin [Sun, 6 Apr 2014 09:47:37 +0000 (12:47 +0300)]
dsdt: tweak ACPI ID for hotplug resource device

ACPI0004 seems too new:
Windows XP complains about an unrecognized device.
This is a regression since 1.7.
Use PNP0A06 instead - Generic Container Device.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-By: Igor Mammedov <imammedo@redhat.com>
10 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-gtk-5' into staging
Peter Maydell [Tue, 8 Apr 2014 12:05:25 +0000 (13:05 +0100)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-gtk-5' into staging

gtk: Implement grab-on-click behavior in relative mode

# gpg: Signature made Tue 08 Apr 2014 12:58:49 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-5:
  gtk: Implement grab-on-click behavior in relative mode

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agogtk: Implement grab-on-click behavior in relative mode
Takashi Iwai [Tue, 8 Apr 2014 09:26:45 +0000 (11:26 +0200)]
gtk: Implement grab-on-click behavior in relative mode

This patch changes the behavior in the relative mode to be compatible
with other UIs, namely, grabbing the input at the first left click.
It improves the usability a lot; otherwise you have to press ctl-alt-G
or select from menu at each time you want to move the pointer.  Also,
the input grab is cleared when the current mode is switched to the
absolute mode.

The automatic reset of the implicit grabbing is needed since the
switching to the absolute mode happens always after the click even on
Gtk.  That is, we cannot check whether the absolute mode is already
available at the first click time even though it should have been
switched in X11 input driver side.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agolog bug fix for 93/19193/1
haken kim [Tue, 8 Apr 2014 10:28:12 +0000 (19:28 +0900)]
log bug fix for
sdcard attach/detach support

Change-Id: I1c5ae6015c576c729c050da0ccebdd09d3fab6d9
Signed-off-by: haken kim <haken.kim@samsung.com>
10 years agoMerge remote-tracking branch 'remotes/agraf/tags/signed-ppc-for-upstream' into staging
Peter Maydell [Tue, 8 Apr 2014 09:58:31 +0000 (10:58 +0100)]
Merge remote-tracking branch 'remotes/agraf/tags/signed-ppc-for-upstream' into staging

Patch queue for ppc - 2014-04-08

This is the final queue for 2.0! It fixes a lot of bugs people have
seen during testing:

  - Fix e500 SMP
  - Fix book3s_64 DEC
  - Fix VSX (new feature in 2.0) for LE hosts
  - Fix PR KVM on top of pHyp (SLOF update)

# gpg: Signature made Tue 08 Apr 2014 10:24:18 BST using RSA key ID 03FEDC60
# gpg: Can't check signature: public key not found

* remotes/agraf/tags/signed-ppc-for-upstream:
  PPC: Add l1 cache sizes for 970 and above systems
  ppce500_spin: Initialize struct properly
  PPC: Only enter MSR_POW when no interrupts pending
  PPC: Clean up DECR implementation
  target-ppc: Correct VSX Integer to FP Conversion
  target-ppc: Correct VSX FP to Integer Conversion
  target-ppc: Correct VSX FP to FP Conversions
  target-ppc: Correct VSX Scalar Compares
  target-ppc: Correct Simple VSR LE Host Inversions
  target-ppc: Correct LE Host Inversion of Lower VSRs
  target-ppc: Define Endian-Correct Accessors for VSR Field Access
  target-ppc: Bug: VSX Convert to Integer Should Truncate
  softfloat: Introduce float32_to_uint64_round_to_zero
  pseries: Update SLOF firmware image to qemu-slof-20140404
  PPC: E500: Set PIR default reset value rather than SPR value

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoMerge remote-tracking branch 'remotes/mdroth/qga-pull-2014-4-7' into staging
Peter Maydell [Tue, 8 Apr 2014 09:41:30 +0000 (10:41 +0100)]
Merge remote-tracking branch 'remotes/mdroth/qga-pull-2014-4-7' into staging

* remotes/mdroth/qga-pull-2014-4-7:
  vss-win32: Fix build with mingw64-headers-3.1.0
  Makefile: add qga-vss-dll-obj-y to nested variables

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoskin: using sensor device 86/19186/1
GiWoong Kim [Tue, 8 Apr 2014 09:22:12 +0000 (18:22 +0900)]
skin: using sensor device

When emulator window is rotated, sensor value must be updated by its device

Change-Id: Ia3201b3ea6c6c06aa3fbcce652c8fbdfe8ab4888
Signed-off-by: GiWoong Kim <giwoong.kim@samsung.com>
10 years agoPPC: Add l1 cache sizes for 970 and above systems
Alexander Graf [Mon, 7 Apr 2014 23:42:53 +0000 (01:42 +0200)]
PPC: Add l1 cache sizes for 970 and above systems

Book3s_64 guests expect the L1 cache size in device tree, so let's give
them proper values for all CPU types we support.

This fixes a "not compliant" warning with sles11 guests on -M pseries for me.

Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agoppce500_spin: Initialize struct properly
Alexander Graf [Mon, 7 Apr 2014 14:48:42 +0000 (16:48 +0200)]
ppce500_spin: Initialize struct properly

The spinning struct is in guest endianness, so we need to initialize
its variables in guest endianness too.

This fixes booting e500 guests with SMP on x86 for me.

Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agoPPC: Only enter MSR_POW when no interrupts pending
Alexander Graf [Sun, 6 Apr 2014 20:40:47 +0000 (22:40 +0200)]
PPC: Only enter MSR_POW when no interrupts pending

We were entering the power saving state even when interrupts (like an
external interrupt or a decrementer interrupt) were still in flight.

In case we find a pending interrupt, don't enter power saving state.

Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Tom Musta <tmusta@gmail.com>
10 years agoPPC: Clean up DECR implementation
Alexander Graf [Sat, 5 Apr 2014 23:32:06 +0000 (01:32 +0200)]
PPC: Clean up DECR implementation

There are 3 different variants of the decrementor for BookE and BookS.

The BookE variant sets TSR[DIS] to 1 when the DEC value becomes 1 or 0. TSR[DIS]
is then the indicator whether the decrementor interrupt line is asserted or not.

The old BookS variant treats DEC as an edge interrupt that gets triggered when
the DEC value's top bit turns 1 from 0.

The new BookS variant maintains the assertion bit inside DEC itself. Whenever
the DEC value becomes negative (top bit set) the DEC interrupt line is asserted.

So far we implemented mostly the old BookS variant. Let's do them all properly.

This fixes booting pseries ppc64 guest images in TCG mode for me.

Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Correct VSX Integer to FP Conversion
Tom Musta [Mon, 31 Mar 2014 21:04:03 +0000 (16:04 -0500)]
target-ppc: Correct VSX Integer to FP Conversion

This patch corrects the VSX integer to floating point conversion instructions
by using the endian correct accessors.  The auxiliary "j" index used by the
existing macros is now obsolete and is removed.  The JOFFSET preprocessor
macro is also obsolete and removed.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Correct VSX FP to Integer Conversion
Tom Musta [Mon, 31 Mar 2014 21:04:02 +0000 (16:04 -0500)]
target-ppc: Correct VSX FP to Integer Conversion

This patch corrects the VSX floating point to integer conversion
instructions by using the endian correct accessors.  The auxiliary
"j" index used by the existing macros is now obsolete and is removed.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Correct VSX FP to FP Conversions
Tom Musta [Mon, 31 Mar 2014 21:04:01 +0000 (16:04 -0500)]
target-ppc: Correct VSX FP to FP Conversions

This change corrects the VSX double precision to single precision and
single precision to double precisions conversion routines.  The endian
correct accessors are now used.  The auxiliary "j" index is no longer
necessary and is eliminated.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Correct VSX Scalar Compares
Tom Musta [Mon, 31 Mar 2014 21:04:00 +0000 (16:04 -0500)]
target-ppc: Correct VSX Scalar Compares

This change fixes the VSX scalar compare instructions.  The existing usage of "x.f64[0]"
is changed to "x.VsrD(0)".

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Correct Simple VSR LE Host Inversions
Tom Musta [Mon, 31 Mar 2014 21:03:59 +0000 (16:03 -0500)]
target-ppc: Correct Simple VSR LE Host Inversions

A common pattern in the VSX helper code macros is the use of "x.fld[i]" where
"x" is a VSR and "fld" is an argument to a macro ("f64" or "f32" is passed).
This is not always correct on LE hosts.

This change addresses all instances of this pattern to be "x.fld" where "fld" is:

  - "VsrD(0)" for scalar instructions accessing 64-bit numbers
  - "VsrD(i)" for vector instructions accessing 64-bit numbers
  - "VsrW(i)" for vector instructions accessing 32-bit numbers

Note that there are no instances of this pattern where a scalar instruction
accesses a 32-bit number.

Note also that it would be correct to use "VsrD(i)" for scalar instructions since
the loop index is only ever "0".  I have choosen to use "VsrD(0)" instead ... it
seems a little clearer.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Correct LE Host Inversion of Lower VSRs
Tom Musta [Mon, 31 Mar 2014 21:03:58 +0000 (16:03 -0500)]
target-ppc: Correct LE Host Inversion of Lower VSRs

This change properly orders the doublewords of the VSRs 0-31.  Because these
registers are constructed from separate doublewords, they must be inverted
on Little Endian hosts.  The inversion is performed both when the VSR is read
and when it is written.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Define Endian-Correct Accessors for VSR Field Access
Tom Musta [Mon, 31 Mar 2014 21:03:57 +0000 (16:03 -0500)]
target-ppc: Define Endian-Correct Accessors for VSR Field Access

This change defines accessors for VSR doubleword and word fields that
are correct from a host Endian perspective.  This allows code to
use the Power ISA indexing numbers in code.

For example, the xscvdpsxws instruction has a target VSR that looks
like this:

  0           32       64                    127
  +-----------+--------+-----------+-----------+
  | undefined | SW     | undefined | undefined |
  +-----------+--------+-----------+-----------+

VSX helper code will use VsrW(1) to access this field.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agotarget-ppc: Bug: VSX Convert to Integer Should Truncate
Tom Musta [Mon, 31 Mar 2014 21:03:56 +0000 (16:03 -0500)]
target-ppc: Bug: VSX Convert to Integer Should Truncate

The various VSX Convert to Integer instructions should truncate the
floating point number to an integer value, which is equivalent to
a round-to-zero rounding mode.  The existing VSX floating point to
integer conversion helpers are erroneously using the rounding mode set
int the PowerPC Floating Point Status and Control Register (FPSCR).
This change corrects this defect by using the appropriate
float*_to_*_round_to_zero() routines fro the softfloat library.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agosoftfloat: Introduce float32_to_uint64_round_to_zero
Tom Musta [Mon, 31 Mar 2014 21:03:55 +0000 (16:03 -0500)]
softfloat: Introduce float32_to_uint64_round_to_zero

This change adds the float32_to_uint64_round_to_zero function to the softfloat
library.  This function fills out the complement of float32 to INT round-to-zero
conversion rountines, where INT is {int32_t, uint32_t, int64_t, uint64_t}.

This contribution can be licensed under either the softfloat-2a or -2b
license.

Signed-off-by: Tom Musta <tommusta@gmail.com>
Tested-by: Tom Musta <tommusta@gmail.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agopseries: Update SLOF firmware image to qemu-slof-20140404
Alexey Kardashevskiy [Fri, 4 Apr 2014 00:57:35 +0000 (11:57 +1100)]
pseries: Update SLOF firmware image to qemu-slof-20140404

The change log is:
  > Isolate sc 1 detection logic
  > build: auto-detect ppc64 architecture
  > cas: increase hcall buffer size to accomodate 256 cpus
  > usb: change device tree naming
  > usb-core: adjust port numbers in set_address
  > virtio-scsi: correct srplun comment
  > Fix kernel loading
  > Workaround to make grub2 assign server ip from dhcp ack packet only
  > ELF: Enter LE binary in LE mode
  > ELF loading should fail for virt != phys

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Alexander Graf <agraf@suse.de>
10 years agoPPC: E500: Set PIR default reset value rather than SPR value
Alexander Graf [Thu, 3 Apr 2014 18:45:27 +0000 (20:45 +0200)]
PPC: E500: Set PIR default reset value rather than SPR value

We now reset SPRs to their reset values on CPU reset. So if we want
to have an SPR persistently changed, we need to change its default
reset value rather than the value itself manually.

Do this for SPR_BOOKE_PIR, fixing e500v2 SMP boot.

Reported-by: Frederic Konrad <fred.konrad@greensocs.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
Tested-by: KONRAD Frederic <fred.konrad@greensocs.com>
10 years agoMerge "ecs : sdcard attach/detach support" into tizen
Sangho Park [Tue, 8 Apr 2014 05:45:33 +0000 (22:45 -0700)]
Merge "ecs : sdcard attach/detach support" into tizen

10 years agovss-win32: Fix build with mingw64-headers-3.1.0
Tomoki Sekiyama [Wed, 26 Mar 2014 18:28:51 +0000 (14:28 -0400)]
vss-win32: Fix build with mingw64-headers-3.1.0

In mingw64-headers-3.1.0, definition of _com_issue_error() is added, which
conflicts with definition in install.cpp. This adds version checking for
mingw headers to disable the definition when the headers>=3.1 is used.

Signed-off-by: Tomoki Sekiyama <tomoki.sekiyama@hds.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
10 years agoMakefile: add qga-vss-dll-obj-y to nested variables
Tomoki Sekiyama [Wed, 26 Mar 2014 18:28:45 +0000 (14:28 -0400)]
Makefile: add qga-vss-dll-obj-y to nested variables

The build rule for qga/vss-win32/qga-vss.dll is broken by commit
ba1183da9a10b94611cad88c44a5c6df005f9b55, because it misses
qga-vss-dll-obj-y in the list of nested variables.
This fixes build of qga-vss.dll by adding qga-vss-dll-obj-y to the list.

Signed-off-by: Tomoki Sekiyama <tomoki.sekiyama@hds.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
10 years agoMerge remote-tracking branch 'remotes/afaerber/tags/qom-devices-for-2.0' into staging
Peter Maydell [Mon, 7 Apr 2014 16:57:23 +0000 (17:57 +0100)]
Merge remote-tracking branch 'remotes/afaerber/tags/qom-devices-for-2.0' into staging

QOM/QTest infrastructure fixes

* Relicensing of FWPathProvider interface
* Clean up all targets' qtests

# gpg: Signature made Mon 07 Apr 2014 17:56:13 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-2.0:
  tests: Update check-clean rule
  fw-path-provider: Change GPL version to 2+

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agotests: Update check-clean rule
Andreas Färber [Mon, 7 Apr 2014 16:33:22 +0000 (18:33 +0200)]
tests: Update check-clean rule

Only i386, x86_64, sparc and sparc64 qtests were cleaned up.
Make this more generic to not miss any newly tested targets.

Reported-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Andreas Färber <afaerber@suse.de>
10 years agoMakefile: remove bashism
Michael Tokarev [Sat, 5 Apr 2014 14:25:46 +0000 (18:25 +0400)]
Makefile: remove bashism

When installing modules (when --enable-modules is specified for
./configure), Makefile uses the following construct to replace all
slashes with dashes in module name:

 ${s//\//-}

This is a bash-specific substitution mechanism.  POSIX does not
have it, and some operating systems (for example Debian) does not
implement this construct in default shell (for example dash).

Use more traditional way to perform the substitution: use `tr' tool.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Message-id: 1396707946-21351-1-git-send-email-mjt@msgid.tls.msk.ru
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agochar/serial: Fix emptyness handling
Don Slutz [Tue, 18 Mar 2014 16:29:34 +0000 (12:29 -0400)]
char/serial: Fix emptyness handling

The commit 88c1ee73d3231c74ff90bcfc084a7589670ec244
char/serial: Fix emptyness check

Still causes extra NULL byte(s) to be sent.

So if the fifo is empty, do not send an extra NULL byte.

Reviewed-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Signed-off-by: Don Slutz <dslutz@verizon.com>
Message-id: 1395160174-16006-1-git-send-email-dslutz@verizon.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agofw-path-provider: Change GPL version to 2+
Alexey Kardashevskiy [Wed, 26 Mar 2014 14:13:02 +0000 (01:13 +1100)]
fw-path-provider: Change GPL version to 2+

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
10 years agoMerge remote-tracking branch 'remotes/spice/tags/pull-spice-6' into staging
Peter Maydell [Mon, 7 Apr 2014 11:48:34 +0000 (12:48 +0100)]
Merge remote-tracking branch 'remotes/spice/tags/pull-spice-6' into staging

spice: monitors_config: check pointer before dereferencing

# gpg: Signature made Mon 07 Apr 2014 11:19:19 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/spice/tags/pull-spice-6:
  spice: monitors_config: check pointer before dereferencing

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-gtk-4' into staging
Peter Maydell [Mon, 7 Apr 2014 11:27:10 +0000 (12:27 +0100)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-gtk-4' into staging

gtk: pointer fixes from Takashi Iwai.

# gpg: Signature made Mon 07 Apr 2014 09:51:52 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-4:
  ui: Update MAINTAINERS entry.
  gtk: Remember the last grabbed pointer position
  gtk: Fix the relative pointer tracking mode
  gtk: Use gtk generic event signal instead of motion-notify-event

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agobuild: make configure better, especially on windows 01/19101/2
Kitae Kim [Mon, 7 Apr 2014 10:30:37 +0000 (19:30 +0900)]
build: make configure better, especially on windows

Change the way to search libraries and headers such as libpng and libcurl.

Change-Id: Ic7743d4196968b0dc702d858dcce52fc4d0671fb
Signed-off-by: Kitae Kim <kt920.kim@samsung.com>
10 years agospice: monitors_config: check pointer before dereferencing
Gerd Hoffmann [Mon, 7 Apr 2014 10:15:44 +0000 (12:15 +0200)]
spice: monitors_config: check pointer before dereferencing

Reported-by: Fabio Fantoni <fabio.fantoni@m2r.biz>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agoui: Update MAINTAINERS entry.
Gerd Hoffmann [Mon, 7 Apr 2014 08:42:03 +0000 (10:42 +0200)]
ui: Update MAINTAINERS entry.

With Amazon eating Anthonys time status "Maintained" certainly isn't
true any more.  Update entry accordingly.

Also add myself, so scripts/get_maintainer.pl will Cc: me, to reduce
the chance ui patches fall through the cracks on our pretty loaded
qemu-devel mailing list.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agogtk: Remember the last grabbed pointer position
Takashi Iwai [Fri, 4 Apr 2014 10:41:23 +0000 (12:41 +0200)]
gtk: Remember the last grabbed pointer position

It's pretty annoying that the pointer reappears at a random place once
after grabbing and ungrabbing the input.  Better to restore to the
original position where the pointer was grabbed.

Reference: https://bugzilla.novell.com/show_bug.cgi?id=849587
Tested-by: Cole Robinson <crobinso@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agogtk: Fix the relative pointer tracking mode
Takashi Iwai [Fri, 4 Apr 2014 10:41:22 +0000 (12:41 +0200)]
gtk: Fix the relative pointer tracking mode

The relative pointer tracking mode was still buggy even after the
previous fix of the motion-notify-event since the events are filtered
out when the pointer moves outside the drawing window due to the
boundary check for the absolute mode.

This patch fixes the issue by moving the unnecessary boundary check
into the if block of absolute mode, and keep the coordinate in the
relative mode even if it's outside the drawing area.  But this makes
the coordinate (last_x, last_y) possibly pointing to (-1,-1),
introduce a new flag to indicate the last coordinate has been
updated.

Reference: https://bugzilla.novell.com/show_bug.cgi?id=849587
Tested-by: Cole Robinson <crobinso@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agogtk: Use gtk generic event signal instead of motion-notify-event
Takashi Iwai [Fri, 4 Apr 2014 10:41:21 +0000 (12:41 +0200)]
gtk: Use gtk generic event signal instead of motion-notify-event

The GDK motion-notify-event isn't generated when the pointer goes out
of the target window even if the pointer is grabbed, which essentially
means to lose the pointer tracking in gtk-ui.

Meanwhile the generic "event" signal is sent when the pointer is
grabbed, so we can use this and pick the motion notify events manually
there instead.

Reference: https://bugzilla.novell.com/show_bug.cgi?id=849587
Tested-by: Cole Robinson <crobinso@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Tested-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 years agoecs : sdcard attach/detach support 13/19013/3
haken kim [Fri, 4 Apr 2014 03:04:44 +0000 (12:04 +0900)]
ecs : sdcard attach/detach support

Change-Id: Id64f007a89678eb9b517c61fab71d6bb5665e341
Signed-off-by: haken kim <haken.kim@samsung.com>
10 years agotarget-i386: reorder fields in cpu/msr_hyperv_hypercall subsection
Paolo Bonzini [Wed, 2 Apr 2014 15:33:02 +0000 (17:33 +0200)]
target-i386: reorder fields in cpu/msr_hyperv_hypercall subsection

The subsection already exists in one well-known enterprise Linux
distribution, but for some strange reason the fields were swapped
when forward-porting the patch to upstream.

Limit headaches for said enterprise Linux distributor when the
time will come to rebase their version of QEMU.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1396452782-21473-1-git-send-email-pbonzini@redhat.com
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agoMerge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Peter Maydell [Fri, 4 Apr 2014 23:18:19 +0000 (00:18 +0100)]
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block patches for 2.0.0

# gpg: Signature made Fri 04 Apr 2014 20:25:08 BST using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"

* remotes/kevin/tags/for-upstream:
  dataplane: replace iothread object_add() with embedded instance
  iothread: make IOThread struct definition public
  dma-helpers: Initialize DMAAIOCB in_cancel flag
  block: Check bdrv_getlength() return value in bdrv_append_temp_snapshot()
  block: Fix snapshot=on for protocol parsed from filename
  qemu-iotests: Remove CR line endings in reference output
  block: Don't parse 'filename' option
  qcow2: Put cache reference in error case
  qcow2: Flush metadata during read-only reopen
  iscsi: Don't set error if already set in iscsi_do_inquiry

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 years agodataplane: replace iothread object_add() with embedded instance
Stefan Hajnoczi [Thu, 20 Mar 2014 14:06:32 +0000 (15:06 +0100)]
dataplane: replace iothread object_add() with embedded instance

Before IOThread was its own object, each virtio-blk device would create
its own internal thread.  We need to preserve this behavior for
backwards compatibility when users do not specify -device
virtio-blk-pci,iothread=<id>.

This patch changes how the internal IOThread object is created.
Previously we used the monitor object_add() function, which is really a
layering violation.  The problem is that this needs to assign a name but
we don't have a name for this internal object.

Generating names for internal objects is a pain but even worse is that
they may collide with user-defined names.

Paolo Bonzini <pbonzini@redhat.com> suggested that the internal IOThread
object should not be named.  This way the conflict cannot happen and we
no longer need object_add().

One gotcha is that internal IOThread objects will not be listed by the
query-iothreads command since they are not named.  This is okay though
because query-iothreads is new and the internal IOThread is just for
backwards compatibility.  New users should explicitly define IOThread
objects.

Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
10 years agoiothread: make IOThread struct definition public
Stefan Hajnoczi [Thu, 20 Mar 2014 14:06:31 +0000 (15:06 +0100)]
iothread: make IOThread struct definition public

Make the IOThread struct definition public so objects can be embedded in
parent structs.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
10 years agodma-helpers: Initialize DMAAIOCB in_cancel flag
Peter Maydell [Fri, 28 Mar 2014 14:22:49 +0000 (14:22 +0000)]
dma-helpers: Initialize DMAAIOCB in_cancel flag

Initialize the dbs->in_cancel flag in dma_bdrv_io(), since qemu_aio_get()
does not return zero-initialized memory. Spotted by the clang sanitizer
(which complained when the value loaded in dma_complete() was not valid
for a bool type); this might have resulted in leaking the AIO block.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>