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: NPeter Maydell <peter.maydell@linaro.org>

* 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: NPeter Maydell <peter.maydell@linaro.org>

The start address has to be stored in big endian byte order
in the iplb.ccw block for the guest.
Signed-off-by: NThomas Huth <thuth@redhat.com>
Message-Id: <1499268345-12552-1-git-send-email-thuth@redhat.com>
Reviewed-by: NCornelia Huck <cohuck@redhat.com>
Signed-off-by: NChristian Borntraeger <borntraeger@de.ibm.com>

This patch is based on a similar patch from Stefan Hajnoczi -
commit c324fd0a ("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 8c56c1a5
("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: NQingFeng Hao <haoqf@linux.vnet.ibm.com>
Reviewed-by: NCornelia Huck <cohuck@redhat.com>
Message-Id: <20170704132350.11874-2-haoqf@linux.vnet.ibm.com>
Reviewed-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NChristian Borntraeger <borntraeger@de.ibm.com>

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

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

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: NHalil Pasic <pasic@linux.vnet.ibm.com>
Fixes: f6f4ce4211 ("s390x: add property adapter_routes_max_batch")
Reviewed-by: NDong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Reviewed-by: NYi Min Zhao <zyimin@linux.vnet.ibm.com>
Reviewed-by: NCornelia Huck <cohuck@redhat.com>
Signed-off-by: NChristian Borntraeger <borntraeger@de.ibm.com>

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

Commit bab482d7 ("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: bab482d7 ("s390x/css: ccw translation infrastructure")
Reviewed-by: NJing Liu <liujbjl@linux.vnet.ibm.com>
Reviewed-by: NHalil Pasic <pasic@linux.vnet.ibm.com>
Reviewed-by: NCornelia Huck <cohuck@redhat.com>
Signed-off-by: NDong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Signed-off-by: NChristian Borntraeger <borntraeger@de.ibm.com>

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: NHalil Pasic <pasic@linux.vnet.ibm.com>
Reviewed-by: NDr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: NJuan Quintela <quintela@redhat.com>
Reviewed-by: NCornelia Huck <cornelia.huck@de.ibm.com>
Message-Id: <20170703213414.94298-1-pasic@linux.vnet.ibm.com>
Signed-off-by: NChristian Borntraeger <borntraeger@de.ibm.com>

Add the CONFIG_TCG for frontend and backend's files in the related
Makefiles.
Signed-off-by: NYang Zhong <yang.zhong@intel.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Add the tcg_enabled() where the x86 target needs to disable
TCG-specific code.
Signed-off-by: NYang Zhong <yang.zhong@intel.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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: NPaolo Bonzini <pbonzini@redhat.com>

Split the cpu_set_mxcsr() and make cpu_set_fpuc() inline with specific
tcg code.
Signed-off-by: NYang Zhong <yang.zhong@intel.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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

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

Add the CONFIG_TCG for frontend and backend's files in the related
Makefiles.
Signed-off-by: NYang Zhong <yang.zhong@intel.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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

Adding assertions fixes link errors.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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

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

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

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

Change the tcg_enabled() and make sure user build still enable tcg
even x86 softmmu disable tcg.
Signed-off-by: NYang Zhong <yang.zhong@intel.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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: NPaolo Bonzini <pbonzini@redhat.com>

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

Need to disable the tcg related code in the vl.c if the
disable-tcg option is added into ./configure command.
Signed-off-by: NYang Zhong <yang.zhong@intel.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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

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

Check for unsupported targets in target_list, and print an
error early in the configuration process.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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: NPaolo Bonzini <pbonzini@redhat.com>

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

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

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 8607f5c3 ("virtio: convert to use DMA api")
Cc: qemu-stable@nongnu.org
Signed-off-by: NJason Wang <jasowang@redhat.com>
Message-Id: <1499170866-9068-1-git-send-email-jasowang@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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: NDaniel P. Berrange <berrange@redhat.com>
Reviewed-by: NRichard W.M. Jones <rjones@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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

In QEMU's main_loop() we used to check whether we should do
a nonblocking call to main_loop(); this was deleted in commit e330c118,
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: NPeter Maydell <peter.maydell@linaro.org>
Message-Id: <1498584769-12439-2-git-send-email-peter.maydell@linaro.org>
Reviewed-by: NMarc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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

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

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 ad9579aa
  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: NThomas Huth <thuth@redhat.com>
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Message-Id: <20170626103756.22974-1-berrange@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>