tcg/aarch64: Use ADRP+ADD to compute target address

We use ADRP+ADD to compute the target address for goto_tb. This patch
introduces the NOP instruction which is used to align the above
instruction pair so that we can use one atomic instruction to patch
the destination offsets.

CC: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
Message-Id: <20170630143614.31059-2-bobby.prani@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg/aarch64: Introduce and use long branch to register

We can use a branch to register instruction for exit_tb for offsets
greater than 128MB.

CC: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Pranith Kumar <bobby.prani@gmail.com>
Message-Id: <20170630143614.31059-1-bobby.prani@gmail.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>

Merge remote-tracking branch 'remotes/borntraeger/tags/s390x-20170706' into staging

s390x/kvm/migration: fixes, enhancements and cleanups

- new email address for Cornelia
- Fixes: 3270, flic, virtio-scsi-ccw, ipl
- Enhancements, cpumodel, migration

# gpg: Signature made Thu 06 Jul 2017 08:18:19 BST
# gpg:                using RSA key 0x117BBC80B5A61C7C
# gpg: Good signature from "Christian Borntraeger (IBM) <borntraeger@de.ibm.com>"
# Primary key fingerprint: F922 9381 A334 08F9 DBAB  FBCA 117B BC80 B5A6 1C7C

* remotes/borntraeger/tags/s390x-20170706:
  hw/s390x/ipl: Fix endianness problem with netboot_start_addr
  virtio-scsi-ccw: use ioeventfd even when KVM is disabled
  s390x: return unavailable features via query-cpu-definitions
  s390x/MAINTAINERS: Update my email address
  s390x: fix realize inheritance for kvm-flic
  s390x: fix error propagation in kvm-flic's realize
  s390x/3270: fix instruction interception handler
  s390x: vmstatify config migration for virtio-ccw

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

Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* qemu-thread portability improvement (Fam)
* virtio-scsi IOMMU fix (Jason)
* poisoning and common-obj-y cleanups (Thomas)
* initial Hypervisor.framework refactoring (Sergio)
* x86 TCG interrupt injection fixes (Wu Xiang, me)
* --disable-tcg support for x86 (Yang Zhong, me)
* various other bugfixes and cleanups (Daniel, Peter, Thomas)

# gpg: Signature made Wed 05 Jul 2017 08:12:56 BST
# 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: (42 commits)
  target/i386: add the CONFIG_TCG into Makefiles
  target/i386: add the tcg_enabled() in target/i386/
  target/i386: move TLB refill function out of helper.c
  target/i386: split cpu_set_mxcsr() and make cpu_set_fpuc() inline
  target/i386: make cpu_get_fp80()/cpu_set_fp80() static
  target/i386: move cpu_sync_bndcs_hflags() function
  tcg: add the CONFIG_TCG into Makefiles
  tcg: add CONFIG_TCG guards in headers
  exec: elide calls to tb_lock and tb_unlock
  tcg: move tb_lock out of translate-all.h
  tcg: add the tcg-stub.c file into accel/stubs/
  vapic: use tcg_enabled
  monitor: disable "info jit" and "info opcount" if !TCG
  tcg: make tcg_allowed global
  cpu: move interrupt handling out of translate-common.c
  tcg: move page_size_init() function
  vl: add tcg_enabled() for tcg related code
  vl: convert -tb-size to qemu_strtoul
  configure: add --disable-tcg configure option
  configure: early test for supported targets
  ...

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

hw/s390x/ipl: Fix endianness problem with netboot_start_addr

The start address has to be stored in big endian byte order
in the iplb.ccw block for the guest.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1499268345-12552-1-git-send-email-thuth@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

virtio-scsi-ccw: use ioeventfd even when KVM is disabled

This patch is based on a similar patch from Stefan Hajnoczi -
commit c324fd0a39c ("virtio-pci: use ioeventfd even when KVM is disabled")

Do not check kvm_eventfds_enabled() when KVM is disabled since it
always returns 0.  Since commit 8c56c1a592b5092d91da8d8943c17777d6462a6f
("memory: emulate ioeventfd") it has been possible to use ioeventfds in
qtest or TCG mode.

This patch makes -device virtio-scsi-ccw,iothread=iothread0 work even
when KVM is disabled.
Currently we don't have an equivalent to "memory: emulate ioeventfd"
for ccw yet, but that this doesn't hurt and qemu-iotests 068 can pass with
skipping iothread arguments.

I have tested that virtio-scsi-ccw works under tcg both with and without
iothread.

This patch fixes qemu-iotests 068, which was accidentally merged early
despite the dependency on ioeventfd.

Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20170704132350.11874-2-haoqf@linux.vnet.ibm.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

s390x: return unavailable features via query-cpu-definitions

The response for query-cpu-definitions didn't include the
unavailable-features field, which is used by libvirt to figure
out whether a certain cpu model is usable on the host.

The unavailable features are now computed by obtaining the host CPU
model and comparing it against the known CPU models. The comparison
takes into account the generation, the GA level and the feature
bitmaps. In the case of a CPU generation/GA level mismatch
a feature called "type" is reported to be missing.

As a result, the output of virsh domcapabilities would change
from something like
 ...
     <mode name='custom' supported='yes'>
      <model usable='unknown'>z10EC-base</model>
      <model usable='unknown'>z9EC-base</model>
      <model usable='unknown'>z196.2-base</model>
      <model usable='unknown'>z900-base</model>
      <model usable='unknown'>z990</model>
 ...
to
 ...
     <mode name='custom' supported='yes'>
      <model usable='yes'>z10EC-base</model>
      <model usable='yes'>z9EC-base</model>
      <model usable='no'>z196.2-base</model>
      <model usable='yes'>z900-base</model>
      <model usable='yes'>z990</model>
 ...

Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
Message-Id: <1499082529-16970-1-git-send-email-mihajlov@linux.vnet.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

s390x/MAINTAINERS: Update my email address

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20170704092215.13742-2-cohuck@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

s390x: fix realize inheritance for kvm-flic

Commit f6f4ce4211 ("s390x: add property adapter_routes_max_batch",
2016-12-09) introduces a common realize (intended to be common for all
the subclasses) for flic, but fails to make sure the kvm-flic which had
its own is actually calling this common realize.

This omission fortunately does not result in a grave problem. The common
realize was only supposed to catch a possible programming mistake by
validating a value of a property set via the compat machine macros. Since
there was no programming mistake we don't need this fixed for stable.

Let's fix this problem by making sure kvm flic honors the realize of its
parent class.

Let us also improve on the error message we would hypothetically emit
when the validation fails.

Signed-off-by: Halil Pasic <pasic@linux.vnet.ibm.com>
Fixes: f6f4ce4211 ("s390x: add property adapter_routes_max_batch")
Reviewed-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Reviewed-by: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

s390x: fix error propagation in kvm-flic's realize

From the moment it was introduced by commit a2875e6f98 ("s390x/kvm:
implement floating-interrupt controller device", 2013-07-16) the kvm-flic
is not making realize fail properly in case it's impossible to create the
KVM device which basically serves as a backend and is absolutely
essential for having an operational kvm-flic.

Let's fix this by making sure we do proper error propagation in realize.

Signed-off-by: Halil Pasic <pasic@linux.vnet.ibm.com>
Fixes: a2875e6f98 "s390x/kvm: implement floating-interrupt controller device"
Reviewed-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Reviewed-by: Yi Min Zhao <zyimin@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

s390x/3270: fix instruction interception handler

Commit bab482d7405f ("s390x/css: ccw translation infrastructure")
introduced instruction interception handler for different types of
subchannels. For emulated 3270 devices, we should assign the virtual
subchannel handler to them during device realization process, or 3270
will not work.

Fixes: bab482d7405f ("s390x/css: ccw translation infrastructure")
Reviewed-by: Jing Liu <liujbjl@linux.vnet.ibm.com>
Reviewed-by: Halil Pasic <pasic@linux.vnet.ibm.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

s390x: vmstatify config migration for virtio-ccw

Let's vmstatify virtio_ccw_save_config and virtio_ccw_load_config for
flexibility (extending using subsections) and for fun.

To achieve this we need to hack the config_vector, which is VirtIODevice
(that is common virtio) state, in the middle of the VirtioCcwDevice state
representation.  This is somewhat ugly, but we have no choice because the
stream format needs to be preserved.

Almost no changes in behavior. Exception is everything that comes with
vmstate like extra bookkeeping about what's in the stream, and maybe some
extra checks and better error reporting.

Signed-off-by: Halil Pasic <pasic@linux.vnet.ibm.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Message-Id: <20170703213414.94298-1-pasic@linux.vnet.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>

target/i386: add the CONFIG_TCG into Makefiles

Add the CONFIG_TCG for frontend and backend's files in the related
Makefiles.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: add the tcg_enabled() in target/i386/

Add the tcg_enabled() where the x86 target needs to disable
TCG-specific code.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: move TLB refill function out of helper.c

This function calls tlb_set_page_with_attrs, which is not available
when TCG is disabled.  Move it to excp_helper.c.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: split cpu_set_mxcsr() and make cpu_set_fpuc() inline

Split the cpu_set_mxcsr() and make cpu_set_fpuc() inline with specific
tcg code.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: make cpu_get_fp80()/cpu_set_fp80() static

Move cpu_get_fp80()/cpu_set_fp80() from fpu_helper.c to
machine.c because fpu_helper.c will be disabled if tcg is
disabled in the build.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: move cpu_sync_bndcs_hflags() function

Move cpu_sync_bndcs_hflags() function from mpx_helper.c
to helper.c because mpx_helper.c need be disabled when
tcg is disabled.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tcg: add the CONFIG_TCG into Makefiles

Add the CONFIG_TCG for frontend and backend's files in the related
Makefiles.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tcg: add CONFIG_TCG guards in headers

Add CONFIG_TCG around TLB-related functions and structure declarations.
Some of these functions are defined in ./accel/tcg/cputlb.c, which will
not be linked in if TCG is disabled, and have no stubs; therefore, their
callers will also be compiled out for --disable-tcg.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

exec: elide calls to tb_lock and tb_unlock

Adding assertions fixes link errors.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tcg: move tb_lock out of translate-all.h

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tcg: add the tcg-stub.c file into accel/stubs/

If tcg is disabled, the functions in tcg-stub.c file will be called.
This file is target-independent file, do not include any platform
related stub functions into this file.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

vapic: use tcg_enabled

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

monitor: disable "info jit" and "info opcount" if !TCG

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tcg: make tcg_allowed global

Change the tcg_enabled() and make sure user build still enable tcg
even x86 softmmu disable tcg.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

cpu: move interrupt handling out of translate-common.c

translate-common.c will not be available anymore with --disable-tcg,
so we cannot leave cpu_interrupt_handler there.

Move the TCG-specific handler to accel/tcg/tcg-all.c, and adopt
KVM's handler as the default one, since it works just as well for
Xen and qtest.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tcg: move page_size_init() function

translate-all.c will be disabled if tcg is disabled in the build,
so page_size_init() function and related variables will be moved
to exec.c file.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

vl: add tcg_enabled() for tcg related code

Need to disable the tcg related code in the vl.c if the
disable-tcg option is added into ./configure command.

Signed-off-by: Yang Zhong <yang.zhong@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

vl: convert -tb-size to qemu_strtoul

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: add --disable-tcg configure option

This lets you build without TCG (hardware accelerationor qtest only).  When
this flag is passed to configure, it will automatically filter out the target
list to only those that support KVM or Xen or HAX.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: early test for supported targets

Check for unsupported targets in target_list, and print an
error early in the configuration process.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure: factor out list of supported Xen/KVM/HAX targets

This will be useful when the functions are called, early in the configure
process, to filter out targets that do not support hardware acceleration.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qemu-doc: do not refer to years-old version numbers

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qemu-thread: Assert locks are initialized before using

Not all platforms check whether a lock is initialized before used.  In
particular Linux seems to be more permissive than OSX.

Check initialization state explicitly in our code to catch such bugs
earlier.

Signed-off-by: Fam Zheng <famz@redhat.com>
Message-Id: <20170704122325.25634-1-famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

virtio-scsi: finalize IOMMU support

After converting to use DMA api for virtio devices, we should use
dma_as instead of address_space_memory. Otherwise it won't work if
IOMMU is enabled.

Fixes: commit 8607f5c3072c ("virtio: convert to use DMA api")
Cc: qemu-stable@nongnu.org
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <1499170866-9068-1-git-send-email-jasowang@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

checkpatch: should not use signal except for SIG_DFL or SIG_IGN

Using signal to establish a signal handler is not portable; on
SysV systems, the signal handler would be reset to SIG_DFL after
delivery, while BSD preserves the signal handler.  Daniel Berrange
reported that (to complicate matters further) the signal system call
has SysV behavior, but glibc signal() actually calls the sigaction
system call to provide BSD behavior.

However, using signal() to set a signal's disposition to SIG_DFL
or SIG_IGN is portable and is a relatively common occurrence in
QEMU source code, so allow that.

Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

main_loop: Make main_loop_wait() return void

The last users of main_loop_wait() that cared about the return value
have now been changed to no longer use it. Drop the now-useless return
value and make the function return void.

We avoid the awkwardness of ifdeffery to handle the 'ret'
variable in main_loop_wait() only being wanted if CONFIG_SLIRP
by simply dropping all the ifdefs. There are stub implementations
of slirp_pollfds_poll() and slirp_pollfds_fill() already in
stubs/slirp.c which do nothing, as required.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <1498584769-12439-3-git-send-email-peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

tests/test-char.c: Don't use main_loop_wait()'s return value

In QEMU's main_loop() we used to check whether we should do
a nonblocking call to main_loop(); this was deleted in commit e330c118f2a5,
because now that vCPUs always drop the I/O thread lock it is an unnecessary
optimization.

The loop in test-char.c copied the old QEMU main_loop() code, but
here the nonblocking check has never been necessary because this
standalone test case doesn't hold the I/O lock anyway. Remove it,
so we can drop the main_loop_wait() return value.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <1498584769-12439-2-git-send-email-peter.maydell@linaro.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

util/oslib-win32: Remove if conditional

The original ready < nhandles - 1 can be re-written as ready + 1 <
nhandles.  The check was actually incorrect because
WAIT_OBJECT_0 was not subtracted from ready; it worked because
WAIT_OBJECT_0 is zero.  After subtracting WAIT_OBJECT_0,
the result is the same condition that we are checking on the first
itteration of the for loop. This means we can remove the if statement
and let the for loop check the code.

Signed-off-by: Alistair Francis <alistair.francis@xilinx.com>
Message-Id: <a14083d681951f3999a0e9314605cb706381ae8d.1498756113.git.alistair.francis@xilinx.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

xsave_helper: pull xsave and xrstor out of kvm.c into helper function

This patch pulls out of kvm.c and into the new files the implementation
for the xsave and xrstor instructions. This so they can be shared by
kvm and hvf.

Signed-off-by: Sergio Andres Gomez Del Real <Sergio.G.DelReal@gmail.com>
Message-Id: <20170626200832.11058-1-Sergio.G.DelReal@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sergio Andres Gomez Del Real <sergio.g.delreal@gmail.com>

sockets: avoid formatting buffer that may not be NUL terminated

The 'sun_path' field in the sockaddr_un struct is not required
to be NUL termianted, so when reporting an error, we must use
the separate 'path' variable which is guaranteed terminated.

Fixes a bug spotted by coverity that was introduced in

  commit ad9579aaa16d5b385922d49edac2c96c79bcfb62
  Author: Daniel P. Berrange <berrange@redhat.com>
  Date:   Thu May 25 16:53:00 2017 +0100

    sockets: improve error reporting if UNIX socket path is too long

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-Id: <20170626103756.22974-1-berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

hw/misc/edu: Compile the edu device as common object

edu.c does not contain any target-specific code, so we can put
it into common-obj-y to compile it only once for all targets.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1498454578-18709-8-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Makefile: Move bootdevice.o to common-obj-y

There does not seem to be any target specific code in this file, so
we can put it into "common-obj" instead of "obj" to compile it only
once for all targets.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1498454578-18709-7-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

include/exec/poison: Mark CONFIG_SOFTMMU as poisoned

CONFIG_SOFTMMU should never be used in common code, so mark
it as poisoned, too.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1498454578-18709-6-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

cpu: Introduce a wrapper for tlb_flush() that can be used in common code

Commit 1f5c00cfdb8114c ("qom/cpu: move tlb_flush to cpu_common_reset")
moved the call to tlb_flush() from the target-specific reset handlers
into the common code qom/cpu.c file, and protected the call with
"#ifdef CONFIG_SOFTMMU" to avoid that it is called for linux-user
only targets. But since qom/cpu.c is common code, CONFIG_SOFTMMU is
*never* defined here, so the tlb_flush() was simply never executed
anymore. Fix it by introducing a wrapper for tlb_flush() in a file
that is re-compiled for each target, i.e. in translate-all.c.

Fixes: 1f5c00cfdb8114c1e3a13426588ceb64f82c9ddb
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1498454578-18709-5-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

include/exec/poison: Mark CONFIG_KVM as poisoned, too

CONFIG_KVM is only defined for target-specific code, so nobody should
use it by accident in common code. To avoid such subtle bugs,
CONFIG_KVM is now marked as poisoned in common code. The header
include/sysemu/kvm.h is somewhat special since it is included
all over the place from common code, too, so we need some extra
logic via "#ifdef NEED_CPU_H" here to make sure that we can
compile all files without problems.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1498454578-18709-4-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Move CONFIG_KVM related definitions to kvm_i386.h

pc.h and sysemu/kvm.h are also included from common code (where
CONFIG_KVM is not available), so the #defines that depend on CONFIG_KVM
should not be declared here to avoid that anybody is using them in a
wrong way. Since we're also going to poison CONFIG_KVM for common code,
let's move them to kvm_i386.h instead. Most of the dummy definitions
from sysemu/kvm.h are also unused since the code that uses them is
only compiled for CONFIG_KVM (e.g. target/i386/kvm.c), so the unused
defines are also simply dropped here instead of being moved.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1498454578-18709-3-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

include/exec/poison: Add some more missing TARGET and CONFIG defines

The defines of some *-linux-user targets were still missing.

Suggested-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1498454578-18709-2-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: simplify handling of conforming code segments on interrupt

Move the handling of conforming code segments before the handling
of stack switch.

Because dpl == cpl after the new "if", it's now unnecessary to check
the C bit when testing dpl < cpl.  Furthermore, dpl > cpl is checked
slightly above the modified code, so the final "else" is unreachable
and we can remove it.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386: fix interrupt CPL error when using ist in x86-64

In do_interrupt64(), when interrupt stack table(ist) is enabled
and the the target code segment is conforming(e2 & DESC_C_MASK), the
old implementation always set new CPL to 0, and SS.RPL to 0.

This is incorrect for when CPL3 code access a CPL0 conforming code
segment, the CPL should remain unchanged. Otherwise higher privileged
code can be compromised.

The patch fix this for always set dpl = cpl when the target code segment
is conforming, and modify the last parameter `flags`, which contains
correct new CPL, in cpu_x86_load_seg_cache().

Signed-off-by: Wu Xiang <willx8@gmail.com>
Message-Id: <20170621142152.GA18094@wxdeubuntu.ipads-lab.se.sjtu.edu.cn>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

nbd: fix NBD over TLS

When attaching the NBD QIOChannel to an AioContext, the TLS channel should
be used, not the underlying socket channel.  This is because, trivially,
the TLS channel will be the one that we read/write to and thus the one
that will get the qio_channel_yield() call.

Fixes: ff82911cd3f69f028f2537825c9720ff78bc3f19
Cc: qemu-stable@nongnu.org
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Tested-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qemu-doc: Add missing "@c man end" statements

Since commit 3f2ce724f1f1 ("Move the qemu-ga description into a
separate chapter"), the qemu.1 man page looks pretty much screwed
up, e.g. the title was "qemu-ga - QEMU Guest Agent" instead of
"qemu-doc - QEMU Emulator User Documentation". However, that movement
of the gemu-ga chapter is not the real problem, it just triggered
another bug in the qemu-doc.texi: There are some parts in the file
which introduce a "@c man begin OPTIONS" section, but never close
it again with "@c man end". After adding the proper end tags here,
the title of the man page is right again and the previously wrongly
tagged sections now also show up correctly in the man page, too.

Reported-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1497863771-24929-1-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

vcpu_dirty: share the same field in CPUState for all accelerators

This patch simply replaces the separate boolean field in CPUState that
kvm, hax (and upcoming hvf) have for keeping track of vcpu dirtiness
with a single shared field.

Signed-off-by: Sergio Andres Gomez Del Real <Sergio.G.DelReal@gmail.com>
Message-Id: <20170618191101.3457-1-Sergio.G.DelReal@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Merge remote-tracking branch 'remotes/edgar/tags/edgar/xilinx-next.for-upstream' into staging

edgar/xilinx-next.for-upstream

# gpg: Signature made Tue 04 Jul 2017 10:00:47 BST
# gpg:                using RSA key 0x29C596780F6BCA83
# gpg: Good signature from "Edgar E. Iglesias (Xilinx key) <edgar.iglesias@xilinx.com>"
# gpg:                 aka "Edgar E. Iglesias <edgar.iglesias@gmail.com>"
# Primary key fingerprint: AC44 FEDC 14F7 F1EB EDBF  4151 29C5 9678 0F6B CA83

* remotes/edgar/tags/edgar/xilinx-next.for-upstream:
  xilinx-dp: Add support for the yuy2 video format
  target-microblaze: Add CPU version 10.0
  target-microblaze: dec_barrel: Add BSIFI
  target-microblaze: dec_barrel: Add BSEFI
  target-microblaze: dec_barrel: Plug TCG temp leak
  target-microblaze: dec_barrel: Add braces around if-statements
  target-microblaze: dec_barrel: Use extract32
  target-microblaze: dec_barrel: Use bool instead of unsigned int
  target-microblaze: Introduce a use-pcmp-instr property
  target-microblaze: Introduce a use-msr-instr property
  target-microblaze: Introduce a use-hw-mul property
  target-microblaze: Introduce a use-div property
  target-microblaze: Introduce a use-barrel property
  target-microblaze: Add CPU versions 9.4, 9.5 and 9.6
  target-microblaze: Don't hard code 0xb as initial MB version
  target-microblaze: Correct bit shift for the PVR0 version field
  disas/microblaze: Add missing 'const' attributes

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

Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

pc, acpi, pci, virtio: fixes, cleanups, features, tests

Some fixes and cleanups. New tests.
Configurable tx queue size for virtio-net.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Mon 03 Jul 2017 20:43:17 BST
# gpg:                using RSA key 0x281F0DB8D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 0270 606B 6F3C DF3D 0B17  0970 C350 3912 AFBE 8E67
#      Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA  8A0D 281F 0DB8 D28D 5469

* remotes/mst/tags/for_upstream: (21 commits)
  i386/acpi: update expected acpi files
  virtio-net: fix tx queue size for !vhost-user
  tests: Add unit tests for the VM Generation ID feature
  vhost-user: unregister slave req handler at cleanup time
  vhost: ensure vhost_ops are set before calling iotlb callback
  intel_iommu: fix migration breakage on mr switch
  hw/acpi: remove dead acpi code
  fw_cfg: move setting of FW_CFG_VERSION_DMA bit to fw_cfg_init1()
  fw_cfg: don't map the fw_cfg IO ports in fw_cfg_io_realize()
  i386/kvm/pci-assign: Use errp directly rather than local_err
  i386/kvm/pci-assign: Fix return type of verify_irqchip_kernel()
  pci: Convert shpc_init() to Error
  pci: Convert to realize
  pci: Replace pci_add_capability2() with pci_add_capability()
  pci: Make errp the last parameter of pci_add_capability()
  pci: Fix the wrong assertion.
  pci: Add comment for pci_add_capability2()
  pci: Clean up error checking in pci_add_capability()
  intel_iommu: relax iq tail check on VTD_GCMD_QIE enable
  hw/pci-bridge/dec: Classify the DEC PCI bridge as bridge device
  ...

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

xilinx-dp: Add support for the yuy2 video format

Add support for the yuy2 video format.

Reviewed-by: KONRAD Frederic <frederic.konrad@adacore.com>
Acked-by: Sai Pavan Boddu <saipava@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Add CPU version 10.0

Add CPU version 10.0.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: dec_barrel: Add BSIFI

Add support for BSIFI.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: dec_barrel: Add BSEFI

Add support for BSEFI.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: dec_barrel: Plug TCG temp leak

Plug TCG temp leak.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: dec_barrel: Add braces around if-statements

Add braces around if-statements.
No functional change.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: dec_barrel: Use extract32

Use extract32 instead of opencoding the shifting and masking.
No functional change.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: dec_barrel: Use bool instead of unsigned int

Use bool instead of unsigned int to represent flags.
No functional change.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Introduce a use-pcmp-instr property

Introduce a use-pcmp-instr property making pcmp instructions
optional.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Introduce a use-msr-instr property

Introduce a use-msr-instr property making msr instructions
optional.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Introduce a use-hw-mul property

Introduce a use-div property making multiplication instructions
optional.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Introduce a use-div property

Introduce a use-div property making division instructions
optional.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Introduce a use-barrel property

Introduce a use-barrel property making barrel shifter instructions
optional.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Add CPU versions 9.4, 9.5 and 9.6

Add CPU versions 9.4, 9.5 and 9.6.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Don't hard code 0xb as initial MB version

Don't hard code 0xb as initial MB version.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

target-microblaze: Correct bit shift for the PVR0 version field

Correct bit shift for the PVR0 version field.

Reviewed-by: Alistair Francis <alistair.francis@xilinx.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

disas/microblaze: Add missing 'const' attributes

Making the opcode list 'const' saves memory.
Some function arguments and local variables needed 'const', too.

Add also 'static' to two local functions.

Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
[EI: Removed old prototypes to fix the build]
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>

i386/acpi: update expected acpi files

We dropped some dead code, update extected table binaries.

Fixes: 4d7e7f2702912 ("hw/acpi: remove dead acpi code")
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-net: fix tx queue size for !vhost-user

Current code segfaults when no nic peer is specified.
Fix it up - fall back to default queue size.

Fixes: 9b02e1618cf26a ("virtio-net: enable configurable tx queue size")
Cc: Wei Wang <wei.w.wang@intel.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tests: Add unit tests for the VM Generation ID feature

The following tests are implemented:
* test that a GUID passed in by command line is propagated to the guest.
  Read the GUID from guest memory
* test that the "auto" argument to the GUID generates a valid GUID, as
  seen by the guest.
* test that a GUID passed in can be queried from the monitor

  This patch is loosely based on a previous patch from:
  Gal Hammer <ghammer@redhat.com>  and Igor Mammedov <imammedo@redhat.com>

Signed-off-by: Ben Warren <ben@skyportsystems.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user: unregister slave req handler at cleanup time

If the backend sends a request just before closing the socket,
the aio dispatcher might schedule its reading after the vhost
device has been cleaned, leading to a NULL pointer dereference
in slave_read();

vhost_user_cleanup() already closes the socket but it is not
enough, the handler has to be unregistered.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost: ensure vhost_ops are set before calling iotlb callback

This patch fixes a crash that happens when vhost-user iommu
support is enabled and vhost-user socket is closed.

When it happens, if an IOTLB invalidation notification is sent
by the IOMMU, vhost_ops's NULL pointer is dereferenced.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

intel_iommu: fix migration breakage on mr switch

Migration is broken after the vfio integration work:

qemu-kvm: AHCI: Failed to start FIS receive engine: bad FIS receive buffer address
qemu-kvm: Failed to load ich9_ahci:ahci
qemu-kvm: error while loading state for instance 0x0 of device '0000:00:1f.2/ich9_ahci'
qemu-kvm: load of migration failed: Operation not permitted

The problem is that vfio work introduced dynamic memory region
switching (actually it is also used for future PT mode), and this memory
region layout is not properly delivered to destination when migration
happens. Solution is to rebuild the layout in post_load.

Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1459906
Fixes: 558e0024 ("intel_iommu: allow dynamic switch of IOMMU region")
Reviewed-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/acpi: remove dead acpi code

Signed-off-by: Aleksandr Bezzubikov <zuban32s@gmail.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

fw_cfg: move setting of FW_CFG_VERSION_DMA bit to fw_cfg_init1()

The setting of the FW_CFG_VERSION_DMA bit is the same across both the
TYPE_FW_CFG_MEM and TYPE_FW_CFG_IO devices, so unify the logic in
fw_cfg_init1().

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Gabriel Somlo <somlo@cmu.edu>

fw_cfg: don't map the fw_cfg IO ports in fw_cfg_io_realize()

As indicated by Laszlo it is a QOM bug for the realize() method to actually
map the device. Set up the IO regions within fw_cfg_io_realize() and defer
the mapping with sysbus_add_io() to the caller, as already done in
fw_cfg_init_mem_wide().

This makes the iobase and dma_iobase properties now obsolete so they can be
removed.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tested-by: Gabriel Somlo <somlo@cmu.edu>

i386/kvm/pci-assign: Use errp directly rather than local_err

In assigned_device_pci_cap_init(), first, error messages are filled
to a local_err variable, then through error_propagate() pass to
the parameter of errp. It leads to cumbersome code. In order to
avoid the extra local_err and error_propagate(), drop it and use
errp instead.

Cc: pbonzini@redhat.com
Cc: rth@twiddle.net
Cc: ehabkost@redhat.com
Cc: mst@redhat.com
Cc: armbru@redhat.com
Cc: marcel@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

i386/kvm/pci-assign: Fix return type of verify_irqchip_kernel()

When the function no success value to transmit, it usually make the
function return void. It has turned out not to be a success, because
it means that the extra local_err variable and error_propagate() will
be needed. It leads to cumbersome code, therefore, transmit success/
failure in the return value is worth. So fix the return type to avoid
it.

Cc: pbonzini@redhat.com
Cc: rth@twiddle.net
Cc: ehabkost@redhat.com
Cc: mst@redhat.com
Cc: armbru@redhat.com
Cc: marcel@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci: Convert shpc_init() to Error

In order to propagate error message better, convert shpc_init() to
Error also convert the pci_bridge_dev_initfn() to realize.

Cc: mst@redhat.com
Cc: marcel@redhat.com
Cc: armbru@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci: Convert to realize

Convert i82801b11, io3130_upstream, io3130_downstream and
pcie_root_port devices to realize.

Cc: mst@redhat.com
Cc: marcel@redhat.com
Cc: armbru@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci: Replace pci_add_capability2() with pci_add_capability()

After the patch 'Make errp the last parameter of pci_add_capability()',
pci_add_capability() and pci_add_capability2() now do exactly the same.
So drop the wrapper pci_add_capability() of pci_add_capability2(), then
replace the pci_add_capability2() with pci_add_capability() everywhere.

Cc: pbonzini@redhat.com
Cc: rth@twiddle.net
Cc: ehabkost@redhat.com
Cc: mst@redhat.com
Cc: dmitry@daynix.com
Cc: jasowang@redhat.com
Cc: marcel@redhat.com
Cc: alex.williamson@redhat.com
Cc: armbru@redhat.com
Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci: Make errp the last parameter of pci_add_capability()

Add Error argument for pci_add_capability() to leverage the errp
to pass info on errors. This way is helpful for its callers to
make a better error handling when moving to 'realize'.

Cc: pbonzini@redhat.com
Cc: rth@twiddle.net
Cc: ehabkost@redhat.com
Cc: mst@redhat.com
Cc: jasowang@redhat.com
Cc: marcel@redhat.com
Cc: alex.williamson@redhat.com
Cc: armbru@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci: Fix the wrong assertion.

pci_add_capability returns a strictly positive value on success,
correct asserts.

Cc: dmitry@daynix.com
Cc: jasowang@redhat.com
Cc: kraxel@redhat.com
Cc: alex.williamson@redhat.com
Cc: armbru@redhat.com
Cc: marcel@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci: Add comment for pci_add_capability2()

Comments for pci_add_capability2() to explain the return
value. This may help to make a correct return value check
for its callers.

Cc: mst@redhat.com
Cc: marcel@redhat.com
Cc: armbru@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pci: Clean up error checking in pci_add_capability()

On success, pci_add_capability2() returns a positive value. On
failure, it sets an error and return a negative value.

pci_add_capability() laboriously checks this behavior. No other
caller does. Drop the checks from pci_add_capability().

Cc: mst@redhat.com
Cc: marcel@redhat.com
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

intel_iommu: relax iq tail check on VTD_GCMD_QIE enable

The VT-d spec (section 6.5.2) prescribes software to zero the
Invalidation Queue Tail Register before enabling the VTD_GCMD_QIE
Global Command Register bit. Windows Server 2012 R2 and possibly
other older Windows versions violate the protocol and set a
non-zero queue tail first, which in effect makes them crash early
on boot with -device intel-iommu,intremap=on.

This commit relaxes the check and instead of failing to enable
VTD_GCMD_QIE with vtd_err_qi_enable, it behaves as if the tail
register was set just after enabling VTD_GCMD_QIE
(see vtd_handle_iqt_write).

Signed-off-by: Ladi Prosek <lprosek@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci-bridge/dec: Classify the DEC PCI bridge as bridge device

This way the bridge shows up in the correct section of the
"-device help" text.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Marcel Apfelbaum <marcel@redhat.com>

virtio-net: enable configurable tx queue size

This patch enables the virtio-net tx queue size to be configurable
between 256 (the default queue size) and 1024 by the user when the
vhost-user backend is used.

Currently, the maximum tx queue size for other backends is 512 due
to the following limitations:
- QEMU backend: the QEMU backend implementation in some cases may
send 1024+1 iovs to writev.
- Vhost_net backend: there are possibilities that the guest sends
a vring_desc of memory which crosses a MemoryRegion thereby
generating more than 1024 iovs after translation from guest-physical
address in the backend.

Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20170603' into staging

Queued TCG patches

# gpg: Signature made Fri 30 Jun 2017 20:03:53 BST
# gpg:                using RSA key 0xAD1270CC4DD0279B
# gpg: Good signature from "Richard Henderson <rth7680@gmail.com>"
# gpg:                 aka "Richard Henderson <rth@redhat.com>"
# gpg:                 aka "Richard Henderson <rth@twiddle.net>"
# Primary key fingerprint: 9CB1 8DDA F8E8 49AD 2AFC  16A4 AD12 70CC 4DD0 279B

* remotes/rth/tags/pull-tcg-20170603:
  tcg: consistently access cpu->tb_jmp_cache atomically
  gen-icount: use tcg_ctx.tcg_env instead of cpu_env
  gen-icount: add missing inline to gen_tb_end

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

tcg: consistently access cpu->tb_jmp_cache atomically

Some code paths can lead to atomic accesses racing with memset()
on cpu->tb_jmp_cache, which can result in torn reads/writes
and is undefined behaviour in C11.

These torn accesses are unlikely to show up as bugs, but from code
inspection they seem possible. For example, tb_phys_invalidate does:
    /* remove the TB from the hash list */
    h = tb_jmp_cache_hash_func(tb->pc);
    CPU_FOREACH(cpu) {
        if (atomic_read(&cpu->tb_jmp_cache[h]) == tb) {
            atomic_set(&cpu->tb_jmp_cache[h], NULL);
        }
    }
Here atomic_set might race with a concurrent memset (such as the
ones scheduled via "unsafe" async work, e.g. tlb_flush_page) and
therefore we might end up with a torn pointer (or who knows what,
because we are under undefined behaviour).

This patch converts parallel accesses to cpu->tb_jmp_cache to use
atomic primitives, thereby bringing these accesses back to defined
behaviour. The price to pay is to potentially execute more instructions
when clearing cpu->tb_jmp_cache, but given how infrequently they happen
and the small size of the cache, the performance impact I have measured
is within noise range when booting debian-arm.

Note that under "safe async" work (e.g. do_tb_flush) we could use memset
because no other vcpus are running. However I'm keeping these accesses
atomic as well to keep things simple and to avoid confusing analysis
tools such as ThreadSanitizer.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1497486973-25845-1-git-send-email-cota@braap.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

gen-icount: use tcg_ctx.tcg_env instead of cpu_env

We are relying on cpu_env being defined as a global, yet most
targets (i.e. all but arm/a64) have it defined as a local variable.
Luckily all of them use the same "cpu_env" name, but really
compilation shouldn't break if the name of that local variable
changed.

Fix it by using tcg_ctx.tcg_env, which all targets set in their
translate_init function. This change also helps paving the way
for the upcoming "translation loop common to all targets" work.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1497639397-19453-3-git-send-email-cota@braap.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

gen-icount: add missing inline to gen_tb_end

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1497639397-19453-2-git-send-email-cota@braap.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

Merge remote-tracking branch 'remotes/famz/tags/block-pull-request' into staging

# gpg: Signature made Fri 30 Jun 2017 15:08:45 BST
# gpg:                using RSA key 0xCA35624C6A9171C6
# gpg: Good signature from "Fam Zheng <famz@redhat.com>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 5003 7CB7 9706 0F76 F021  AD56 CA35 624C 6A91 71C6

* remotes/famz/tags/block-pull-request:
  block: Exploit BDRV_BLOCK_EOF for larger zero blocks
  block: Add BDRV_BLOCK_EOF to bdrv_get_block_status()

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

Merge remote-tracking branch 'remotes/vivier/tags/m68k-for-2.10-pull-request' into staging

# gpg: Signature made Fri 30 Jun 2017 13:30:44 BST
# gpg:                using RSA key 0xF30C38BD3F2FBE3C
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>"
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>"
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>"
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier/tags/m68k-for-2.10-pull-request:
  target/m68k: add fmovem
  target/m68k: add explicit single and double precision operations (part 2)
  target/m68k: add fsglmul and fsgldiv
  softfloat: define floatx80_round()
  target/m68k: add explicit single and double precision operations
  target/m68k: add fmovecr
  target/m68k: add fscc.

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