namelen should be here, length is unrelated, and always 0 at this
point.  Broken in introduction in commit f37708f6, but mostly
harmless (replying with '' as the name does not violate protocol,
and does not confuse qemu as the nbd client since our implementation
does not ask for the name; but might confuse some other client that
does ask for the name especially if the default export is different
than the export name being queried).

Adding an assert makes it obvious that we are not skipping any bytes
in the client's message, as well as making it obvious that we were
using the wrong variable.
Signed-off-by: NVladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
CC: qemu-stable@nongnu.org
Message-Id: <20171101154204.27146-1-vsementsov@virtuozzo.com>
[eblake: improve commit message, squash in assert addition]
Signed-off-by: NEric Blake <eblake@redhat.com>

Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Pull IO 2017/11/06 v2

# gpg: Signature made Tue 07 Nov 2017 14:35:25 GMT
# gpg:                using RSA key 0xBE86EBB415104FDF
# gpg: Good signature from "Daniel P. Berrange <dan@berrange.com>"
# gpg:                 aka "Daniel P. Berrange <berrange@redhat.com>"
# Primary key fingerprint: DAF3 A6FD B26B 6291 2D0E  8E3F BE86 EBB4 1510 4FDF

* remotes/berrange/tags/pull-2017-11-06-2:
  sockets: avoid leak of listen file descriptor
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

If we iterate over the full port range without successfully binding+listening
on the socket, we'll try the next address, whereupon we overwrite the slisten
file descriptor variable without closing it.

Rather than having two places where we open + close socket FDs on different
iterations of nested for loops, re-arrange the code to always open+close
within the same loop iteration.
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

target-arm queue:
 * arm_gicv3_its: Don't abort on table save failure
 * arm_gicv3_its: Fix the VM termination in vm_change_state_handler()
 * translate.c: Fix usermode big-endian AArch32 LDREXD and STREXD
 * hw/arm: Mark the "fsl,imx31/25/6" devices with user_creatable = false
 * arm: implement cache/shareability attribute bits for PAR registers

# gpg: Signature made Tue 07 Nov 2017 13:33:58 GMT
# gpg:                using RSA key 0x3C2525ED14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>"
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>"
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-target-arm-20171107:
  hw/intc/arm_gicv3_its: Don't abort on table save failure
  hw/intc/arm_gicv3_its: Fix the VM termination in vm_change_state_handler()
  translate.c: Fix usermode big-endian AArch32 LDREXD and STREXD
  hw/arm: Mark the "fsl,imx31" device with user_creatable = false
  hw/arm: Mark the "fsl,imx25" device with user_creatable = false
  hw/arm: Mark the "fsl,imx6" device with user_creatable = false
  arm: implement cache/shareability attribute bits for PAR registers
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

The ITS is not fully properly reset at the moment. Caches are
not emptied.

After a reset, in case we attempt to save the state before
the bound devices have registered their MSIs and after the
1st level table has been allocated by the ITS driver
(device BASER is valid), the first level entries are still
invalid. If the device cache is not empty (devices registered
before the reset), vgic_its_save_device_tables fails with -EINVAL.
This causes a QEMU abort().

Cc: qemu-stable@nongnu.org
Signed-off-by: NEric Auger <eric.auger@redhat.com>
Reported-by: Nwanghaibin <wanghaibin.wang@huawei.com>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

The commit cddafd8f ("hw/intc/arm_gicv3_its: Implement state save
/restore") breaks the backward compatibility with the older kernels
where vITS save/restore support is not available. The vmstate function
vm_change_state_handler() should not be registered if the running kernel
doesn't support ITS save/restore feature. Otherwise VM instance will be
killed whenever vmstate callback function is invoked.

Observed a virtual machine shutdown with QEMU-2.10+linux-4.11 when testing
the reboot command "virsh reboot <domain> --mode acpi" instead of reboot.

KVM Error: 'KVM_SET_DEVICE_ATTR failed: Group 4 attr 0x00000000000001'
Signed-off-by: NShanker Donthineni <shankerd@codeaurora.org>
Reviewed-by: NEric Auger <eric.auger@redhat.com>
Message-id: 1509712671-16299-1-git-send-email-shankerd@codeaurora.org
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

For AArch32 LDREXD and STREXD, architecturally the 32-bit word at the
lowest address is always Rt and the one at addr+4 is Rt2, even if the
CPU is big-endian. Our implementation does these with a single
64-bit store, so if we're big-endian then we need to put the two
32-bit halves together in the opposite order to little-endian,
so that they end up in the right places. We were trying to do
this with the gen_aa32_frob64() function, but that is not correct
for the usermode emulator, because there there is a distinction
between "load a 64 bit value" (which does a BE 64-bit access
and doesn't need swapping) and "load two 32 bit values as one
64 bit access" (where we still need to do the swapping, like
system mode BE32).

Fixes: https://bugs.launchpad.net/qemu/+bug/1725267
Cc: qemu-stable@nongnu.org
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Reviewed-by: NRichard Henderson <richard.henderson@linaro.org>
Message-id: 1509622400-13351-1-git-send-email-peter.maydell@linaro.org

QEMU currently crashes when the user tries to instantiate the fsl,imx31
device manually:

$ aarch64-softmmu/qemu-system-aarch64 -M kzm -device fsl,,imx31
**
ERROR:/home/thuth/devel/qemu/tcg/tcg.c:538:tcg_register_thread:
 assertion failed: (n < max_cpus)
Aborted (core dumped)

The kzm board (which is the one that uses this CPU type) only supports
one CPU, and the realize function of the "fsl,imx31" device also uses
serial_hds[] directly, so this device clearly can not be instantiated
twice and thus we should mark it with user_creatable = false.
Signed-off-by: NThomas Huth <thuth@redhat.com>
Message-id: 1509519537-6964-4-git-send-email-thuth@redhat.com
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

QEMU currently crashes when the user tries to instantiate the fsl,imx25
device manually:

$ aarch64-softmmu/qemu-system-aarch64 -S -M imx25-pdk -device fsl,,imx25
**
ERROR:/home/thuth/devel/qemu/tcg/tcg.c:538:tcg_register_thread:
 assertion failed: (n < max_cpus)

The imx25-pdk board (which is the one that uses this CPU type) only
supports one CPU, and the realize function of the "fsl,imx25" device
also uses serial_hds[] directly, so this device clearly can not be
instantiated twice and thus we should mark it with user_creatable = 0.
Signed-off-by: NThomas Huth <thuth@redhat.com>
Message-id: 1509519537-6964-3-git-send-email-thuth@redhat.com
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

This device causes QEMU to abort if the user tries to instantiate it:

$ qemu-system-aarch64 -M sabrelite -smp 1,maxcpus=2 -device fsl,,imx6
Unexpected error in qemu_chr_fe_init() at chardev/char-fe.c:222:
qemu-system-aarch64: -device fsl,,imx6: Device 'serial0' is in use
Aborted (core dumped)

The device uses serial_hds[] directly in its realize function, so it
can not be instantiated again by the user.
Signed-off-by: NThomas Huth <thuth@redhat.com>
Message-id: 1509519537-6964-2-git-send-email-thuth@redhat.com
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

On a successful address translation instruction, PAR is supposed to
contain cacheability and shareability attributes determined by the
translation. We previously returned 0 for these bits (in line with the
general strategy of ignoring caches and memory attributes), but some
guest OSes may depend on them.

This patch collects the attribute bits in the page-table walk, and
updates PAR with the correct attributes for all LPAE translations.
Short descriptor formats still return 0 for these bits, as in the
prior implementation.
Signed-off-by: NAndrew Baumann <Andrew.Baumann@microsoft.com>
Message-id: 20171031223830.4608-1-Andrew.Baumann@microsoft.com
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

cocoa queue:
 * make scrolling work in GUI monitor windows
 * change ungrab to ctrl-alt-g (matching gtk)
 * pass unused ctrl-alt combos to guest

# gpg: Signature made Tue 07 Nov 2017 10:15:00 GMT
# gpg:                using RSA key 0x3C2525ED14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>"
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>"
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-cocoa-20171107:
  ui/cocoa.m: Send ctrl-alt key combos to guest if QEMU isn't using them
  ui/cocoa.m: move ungrab to ctrl-alt-g
  ui/cocoa.m: Make scrolling work again in GUI monitor windows
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Merge build 2017/11/07 v1

# gpg: Signature made Tue 07 Nov 2017 10:14:49 GMT
# gpg:                using RSA key 0xBE86EBB415104FDF
# gpg: Good signature from "Daniel P. Berrange <dan@berrange.com>"
# gpg:                 aka "Daniel P. Berrange <berrange@redhat.com>"
# Primary key fingerprint: DAF3 A6FD B26B 6291 2D0E  8E3F BE86 EBB4 1510 4FDF

* remotes/berrange/tags/pull-build-2017-11-07-1:
  build: remove use of MAKELEVEL optimization in submodule handling
  build: delay check for empty git submodule list
  build: don't fail if given a git submodule which does not exist
  build: allow automatic git submodule updates to be disabled
  build: don't create temporary files in source dir
  build: allow setting a custom GIT binary for transparent proxying
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

This fixes a bad errno returned to the guest and a trivial coding style nit.

# gpg: Signature made Mon 06 Nov 2017 18:09:24 GMT
# gpg:                using RSA key 0x71D4D5E5822F73D6
# gpg: Good signature from "Greg Kurz <groug@kaod.org>"
# gpg:                 aka "Gregory Kurz <gregory.kurz@free.fr>"
# gpg:                 aka "[jpeg image of size 3330]"
# Primary key fingerprint: B482 8BAF 9431 40CE F2A3  4910 71D4 D5E5 822F 73D6

* remotes/gkurz/tags/for-upstream:
  9pfs: fix v9fs_mark_fids_unreclaim() return value
  9pfs: drop one user of struct V9fsFidState
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Send those ctrl-alt key combos that QEMU doesn't treat specially to
the guest rather than ignoring them.

All the case where we do special handling of ctrl-alt-X exit the
event handling using a "return" statement, so we can simply allow
the rest to fall through into the normal key handling by deleting
the now-spurious "else".

We take the opportunity to clean up some oddly-formatted and
now rather uninformative comments by removing them.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Currently the cocoa user interface relis on the user pushing
control-alt to ungrab the mouse.  This is patch changes the key
combination to control-alt-g to be in line with the GTK user
interface.
Signed-off-by: NJohn Arbuckle <programmingkidx@gmail.com>
Message-id: 20171102213907.11443-1-programmingkidx@gmail.com
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Make scrolling in the monitor work, by correctly passing through
control+key combinations.
Signed-off-by: NJohn Arbuckle <programmingkidx@gmail.com>
Message-id: 20171101154607.1582-1-programmingkidx@gmail.com
[PMM: fixed coding style nits; cleaned up commit message]
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

The return value of v9fs_mark_fids_unreclaim() is then propagated to
pdu_complete(). It should be a negative errno, not -1.
Signed-off-by: NGreg Kurz <groug@kaod.org>
Reviewed-by: NEric Blake <eblake@redhat.com>

To comply with QEMU coding style.
Signed-off-by: NGreg Kurz <groug@kaod.org>

MIPS patches 2017-11-06

Changes:
Update email addresses of Yongbok Kim, James Hogan and Paul Burton.

# gpg: Signature made Mon 06 Nov 2017 15:38:58 GMT
# gpg:                using RSA key 0x2238EB86D5F797C2
# gpg: Good signature from "Yongbok Kim <yongbok.kim@mips.com>"
# gpg:                 aka "Yongbok Kim <yongbok.kim@imgtec.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: 8600 4CF5 3415 A5D9 4CFA  2B5C 2238 EB86 D5F7 97C2

* remotes/yongbok/tags/mips-20171106:
  MAINTAINERS: Update Paul Burton's email address
  MAINTAINERS: Update James Hogan's email address
  MAINTAINERS: Update Yongbok Kim's email address
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Update my email address from paul.burton@imgtec.com to
paul.burton@mips.com, since MIPS will soon no longer be a part of
Imagination Technologies & as such the @imgtec.com address will soon
cease to function.

A mapping is added in .mailmap such that git reports the new @mips.com
address, and get_maintainer.pl in turn reports it when examining git
history. Whilst here add a mapping for my also-defunct
paul@archlinuxmips.org email address too.
Signed-off-by: NPaul Burton <paul.burton@imgtec.com>
Cc: Laurent Vivier <laurent@vivier.eu>
Cc: Michael Tokarev <mjt@tls.msk.ru>
Cc: Yongbok Kim <yongbok.kim@imgtec.com>
Cc: Yongbok Kim <yongbok.kim@mips.com>
Cc: qemu-devel@nongnu.org
Cc: qemu-trivial@nongnu.org
Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
[Yongbok Kim:
  Updated message subject]
Signed-off-by: NYongbok Kim <yongbok.kim@imgtec.com>

Update my imgtec.com email address to my kernel.org one in MAINTAINERS
as MIPS will soon no longer be part of Imagination Technologies, and add
a mapping in .mailcap so get_maintainer.pl reports the right address.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>
Cc: Michael Tokarev <mjt@tls.msk.ru>
Cc: Laurent Vivier <laurent@vivier.eu>
Cc: qemu-devel@nongnu.org
Cc: qemu-trivial@nongnu.org
Reviewed-by: NLaurent Vivier <laurent@vivier.eu>
[Yongbok Kim:
  Updated message subject]
Signed-off-by: NYongbok Kim <yongbok.kim@imgtec.com>

Update my email address to mips.com
Signed-off-by: NYongbok Kim <yongbok.kim@imgtec.com>

The Makefile attempts to optimize the handling of submodules by using MAKELEVEL
to only check the submodule status when running from the top level make
invokation. This causes problems for people who are using a makefile of their
own to in turn invoke QEMU's makefile, as MAKELEVEL is already set to 1 (or
more) when QEMU's makefile runs.

This optimization should not really be needed, since the git-submodule.sh
script is already used to detect if a submodule update is required. This by
removing the MAKELEVEL check, we at most add an extra 'git-submodule.sh status'
call to each make level, the overhead of which is lost in noise of building
QEMU.
Reviewed-by: NGreg Kurz <groug@kaod.org>
Tested-by: NGreg Kurz <groug@kaod.org>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

# gpg: Signature made Mon 06 Nov 2017 11:20:18 GMT
# gpg:                using RSA key 0x9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/block-pull-request:
  aio-posix: drop QEMU_AIO_POLL_MAX_NS env var
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

We short circuit the git submodule update when passed an empty module list.
This accidentally causes the 'status' command to write to the status file. The
test needs to be delayed into the individual commands to avoid this premature
writing of the status file.
Tested-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

If going back in time in git history, across a commit that introduces a new
submodule, the 'git-submodule.sh' script will fail, causing rebuild to fail.

This is because config-host.mak contains a GIT_SUBMODULES variable that lists
a submodule that only exists in the later commit. config-host.mak won't get
repopulated until config.status is invoked, but make won't get this far due to
the submodule error.

This change makes 'git-submodule.sh' check whether each module is known to git
and drops any which are not present. A warning message will be printed when any
submodule is dropped in this manner.
Tested-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

This hunk should not have been merged but I forgot to remove it.  Let's
remove it before it slips into a QEMU release.

¯\_(ツ)_/¯
Reviewed-by: NThomas Huth <thuth@redhat.com>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Message-id: 20171103154041.12617-1-stefanha@redhat.com
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>

Some people building QEMU use VPATH builds where the source directory is on a
read-only volume. In such a case 'scripts/git-submodules.sh update' will always
fail and users are required to run it manually themselves on their original
writable source directory.

While this is already supported, it is nice to give users a command line flag
to configure to permanently disable automatic submodule updates, as it means
they won't get hard to diagnose failures from git-submodules.sh at an arbitrary
later date.

This patch thus introduces a flag '--disable-git-update' which will prevent
'make' from ever running 'scripts/git-submodules.sh update'. It will still run
the 'status' command to determine if a submodule update is needed, but when it
does this it'll simply stop and print a message instructing the developer what
todo. eg

$ ./configure  --target-list=x86_64-softmmu --disable-git-update
...snip...

$ make
  GEN     config-host.h
  GEN     trace/generated-tcg-tracers.h
  GEN     trace/generated-helpers-wrappers.h
  GEN     trace/generated-helpers.h
  GEN     trace/generated-helpers.c
  GEN     module_block.h

GIT submodule checkout is out of date. Please run
  scripts/git-submodule.sh update ui/keycodemapdb
from the source directory checkout /home/berrange/src/virt/qemu

make: *** [Makefile:31: git-submodule-update] Error 1
Tested-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

There are cases where users do VPATH builds with the source directory being on
a read-only volume. In such a case they have to manually run the command
'git-submodule.sh ...modules...' ahead of time. When checking for status we
should not then write into the source dir.
Tested-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

Some users can't run a bare 'git' command, due to need for a transparent
proxying solution such as 'tsocks'. This adds an argument to configure to
let users specify such a thing:

  ./configure --with-git="tsocks git"

The submodule script is also updated to give the user a hint about using this
flag, if we fail to checkout modules.
Tested-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>

* Remove legacy KVM PCI device assignment
* Fix qemu-pr-helper.c compilation on older compilers

# gpg: Signature made Sun 05 Nov 2017 13:52:36 GMT
# gpg:                using RSA key 0xBFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>"
# 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:
  build: disable -Wmissing-braces on older compilers
  pci-assign: Remove
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

GCC 4.9 and newer stopped warning for missing braces around the
"universal" C zero initializer {0}.  One such initializer sneaked
into scsi/qemu-pr-helper.c and is breaking the build with such
older GCC versions.

Detect the lack of support for the idiom, and disable the warning
in that case.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Legacy PCI device assignment has been removed from Linux in 4.12,
and had been deprecated 2 years ago there.  We can remove it from
QEMU as well.

The ROM loading code was shared with Xen PCI passthrough, so move
it to hw/xen.
Reviewed-by: NPhilippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Merge remote-tracking branch 'remotes/stsquad/tags/pull-ci-updates-for-softfreeze-021117-2' into staging

One fix for win32/64 cross compiles.

# gpg: Signature made Thu 02 Nov 2017 11:55:46 GMT
# gpg:                using RSA key 0xFBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>"
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* remotes/stsquad/tags/pull-ci-updates-for-softfreeze-021117-2:
  docker: add python stdlib dependency (required by keycodemapdb)
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Queued tcg patches

# gpg: Signature made Fri 03 Nov 2017 08:37:58 GMT
# gpg:                using RSA key 0x64DF38E8AF7E215F
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>"
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* remotes/rth/tags/pull-tcg-20171103:
  cpu-exec: Exit exclusive region on longjmp from step_atomic
  tcg/s390x: Use constant pool for prologue
  tcg: Allow constant pool entries in the prologue
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Commit ac03ee53 narrowed the scope of the exclusive
region so it only covers when we're executing the TB, not when
we're generating it. However it missed that there is more than
one execution path out of cpu_tb_exec -- if the atomic insn
causes an exception then the code will longjmp out, skipping
the code to end the exclusive region. This causes QEMU to hang
the next time the CPU calls start_exclusive(), waiting for
itself to exit the region.

Move the "end the region" code out to the end of the
function so that it is run for both normal exit and also
for exit-via-longjmp. We have to use a volatile bool flag
to decide whether we need to end the region, because we
can longjump out of the codegen as well as the execution.

(For some reason this only reproduces for me with a clang
optimized build, not a gcc debug build.)
Reviewed-by: NEmilio G. Cota <cota@braap.org>
Reviewed-by: NAlex Bennée <alex.bennee@linaro.org>
Reviewed-by: NRichard Henderson <richard.henderson@linaro.org>
Fixes: ac03ee53Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Message-Id: <1509640536-32160-1-git-send-email-peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <richard.henderson@linaro.org>

Rather than have separate code only used for guest_base,
rely on a recent change to handle constant pool entries.

Cc: qemu-s390x@nongnu.org
Signed-off-by: NRichard Henderson <richard.henderson@linaro.org>

Both ARMv6 and AArch64 currently may drop complex guest_base values
into the constant pool.  But generic code wasn't expecting that, and
the pool is not emitted.  Correct that.
Tested-by: NEmilio G. Cota <cota@braap.org>
Tested-by: NLaurent Desnogues <laurent.desnogues@gmail.com>
Reported-by: NLaurent Desnogues <laurent.desnogues@gmail.com>
Signed-off-by: NRichard Henderson <richard.henderson@linaro.org>