COLO: notify net filters about checkpoint/failover event

Notify all net filters about the checkpoint and failover event.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

filter-rewriter: handle checkpoint and failover event

After one round of checkpoint, the states between PVM and SVM
become consistent, so it is unnecessary to adjust the sequence
of net packets for old connections, besides, while failover
happens, filter-rewriter will into failover mode that needn't
handle the new TCP connection.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

filter: Add handle_event method for NetFilterClass

Filter needs to process the event of checkpoint/failover or
other event passed by COLO frame.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

COLO: flush host dirty ram from cache

Don't need to flush all VM's ram from cache, only
flush the dirty pages since last checkpoint

Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

savevm: split the process of different stages for loadvm/savevm

There are several stages during loadvm/savevm process. In different stage,
migration incoming processes different types of sections.
We want to control these stages more accuracy, it will benefit COLO
performance, we don't have to save type of QEMU_VM_SECTION_START
sections everytime while do checkpoint, besides, we want to separate
the process of saving/loading memory and devices state.

So we add three new helper functions: qemu_load_device_state() and
qemu_savevm_live_state() to achieve different process during migration.

Besides, we make qemu_loadvm_state_main() and qemu_save_device_state()
public, and simplify the codes of qemu_save_device_state() by calling the
wrapper qemu_savevm_state_header().

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

qapi: Add new command to query colo status

Libvirt or other high level software can use this command query colo status.
You can test this command like that:
{'execute':'query-colo-status'}

Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

qapi/migration.json: Rename COLO unknown mode to none mode.

Suggested by Markus Armbruster rename COLO unknown mode to none mode.

Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

qmp event: Add COLO_EXIT event to notify users while exited COLO

If some errors happen during VM's COLO FT stage, it's important to
notify the users of this event. Together with 'x-colo-lost-heartbeat',
Users can intervene in COLO's failover work immediately.
If users don't want to get involved in COLO's failover verdict,
it is still necessary to notify users that we exited COLO mode.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

COLO: Flush memory data from ram cache

During the time of VM's running, PVM may dirty some pages, we will transfer
PVM's dirty pages to SVM and store them into SVM's RAM cache at next checkpoint
time. So, the content of SVM's RAM cache will always be same with PVM's memory
after checkpoint.

Instead of flushing all content of PVM's RAM cache into SVM's MEMORY,
we do this in a more efficient way:
Only flush any page that dirtied by PVM since last checkpoint.
In this way, we can ensure SVM's memory same with PVM's.

Besides, we must ensure flush RAM cache before load device state.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

ram/COLO: Record the dirty pages that SVM received

We record the address of the dirty pages that received,
it will help flushing pages that cached into SVM.

Here, it is a trick, we record dirty pages by re-using migration
dirty bitmap. In the later patch, we will start the dirty log
for SVM, just like migration, in this way, we can record both
the dirty pages caused by PVM and SVM, we only flush those dirty
pages from RAM cache while do checkpoint.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

COLO: Load dirty pages into SVM's RAM cache firstly

We should not load PVM's state directly into SVM, because there maybe some
errors happen when SVM is receving data, which will break SVM.

We need to ensure receving all data before load the state into SVM. We use
an extra memory to cache these data (PVM's ram). The ram cache in secondary side
is initially the same as SVM/PVM's memory. And in the process of checkpoint,
we cache the dirty pages of PVM into this ram cache firstly, so this ram cache
always the same as PVM's memory at every checkpoint, then we flush this cached ram
to SVM after we receive all PVM's state.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

COLO: Remove colo_state migration struct

We need to know if migration is going into COLO state for
incoming side before start normal migration.

Instead by using the VMStateDescription to send colo_state
from source side to destination side, we use MIG_CMD_ENABLE_COLO
to indicate whether COLO is enabled or not.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

COLO: Add block replication into colo process

Make sure master start block replication after slave's block
replication started.

Besides, we need to activate VM's blocks before goes into
COLO state.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Li Zhijian <lizhijian@cn.fujitsu.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

COLO: integrate colo compare with colo frame

For COLO FT, both the PVM and SVM run at the same time,
only sync the state while it needs.

So here, let SVM runs while not doing checkpoint, change
DEFAULT_MIGRATE_X_CHECKPOINT_DELAY to 200*100.

Besides, we forgot to release colo_checkpoint_semd and
colo_delay_timer, fix them here.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

colo-compare: use notifier to notify packets comparing result

It's a good idea to use notifier to notify COLO frame of
inconsistent packets comparing.

Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

colo-compare: implement the process of checkpoint

While do checkpoint, we need to flush all the unhandled packets,
By using the filter notifier mechanism, we can easily to notify
every compare object to do this process, which runs inside
of compare threads as a coroutine.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

filter-rewriter: Add TCP state machine and fix memory leak in connection_track_table

We add almost full TCP state machine in filter-rewriter, except
TCPS_LISTEN and some simplify in VM active close FIN states.
The reason for this simplify job is because guest kernel will track
the TCP status and wait 2MSL time too, if client resend the FIN packet,
guest will resend the last ACK, so we needn't wait 2MSL time in filter-rewriter.

After a net connection is closed, we didn't clear its related resources
in connection_track_table, which will lead to memory leak.

Let's track the state of net connection, if it is closed, its related
resources will be cleared up.

Signed-off-by: zhanghailiang <zhang.zhanghailiang@huawei.com>
Signed-off-by: Zhang Chen <zhangckid@gmail.com>
Signed-off-by: Zhang Chen <chen.zhang@intel.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>

Merge remote-tracking branch 'remotes/huth-gitlab/tags/pull-request-2018-10-17' into staging

- Updates for qtest entries in test/Makefile.include
- Simple updates for some shell scripts
- Misc simple patches for files without regular subsystem pull requests

# gpg: Signature made Wed 17 Oct 2018 08:20:35 BST
# gpg:                using RSA key 2ED9D774FE702DB5
# gpg: Good signature from "Thomas Huth <th.huth@gmx.de>"
# gpg:                 aka "Thomas Huth <thuth@redhat.com>"
# gpg:                 aka "Thomas Huth <huth@tuxfamily.org>"
# gpg:                 aka "Thomas Huth <th.huth@posteo.de>"
# Primary key fingerprint: 27B8 8847 EEE0 2501 18F3  EAB9 2ED9 D774 FE70 2DB5

* remotes/huth-gitlab/tags/pull-request-2018-10-17:
  configure: remove glib_subprocess check
  hw/core/generic-loader: Compile only once, not for each target
  cpu: Provide a proper prototype for target_words_bigendian() in a header
  hw/core/generic-loader: Set a category for the generic-loader device
  qemu/compiler: Wrap __attribute__((flatten)) in a macro
  mailmap: Fix Reimar Döffinger name
  show-fixed-bugs.sh: Modern shell scripting (use $() instead of ``)
  git-submodule.sh: Modern shell scripting (use $() instead of ``)
  archive-source.sh: Modern shell scripting (use $() instead of ``)
  MAINTAINERS: update block/sheepdog maintainers
  gdbstub: Remove unused include
  tests: remove gcov-files- variables
  tests: Prevent more accidental test disabling
  target/cris/translate: Get rid of qemu_log_separate()
  qemu-common.h: update copyright date to 2018

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

Merge remote-tracking branch 'remotes/rth/tags/pull-hppa-20181016' into staging

Queued hppa patch

# gpg: Signature made Tue 16 Oct 2018 23:45:46 BST
# gpg:                using RSA key 64DF38E8AF7E215F
# 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-hppa-20181016:
  target/hppa: Raise exception 26 on emulated hardware

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

configure: remove glib_subprocess check

This should have been removed as part of commit
692fbdf9f4c6f6bafd0b3a4d4f94973effd3bbae.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

hw/core/generic-loader: Compile only once, not for each target

The generic-loader is currently compiled target specific due to one
single "#ifdef TARGET_WORDS_BIGENDIAN" in the file. We have already a
function called target_words_bigendian() for this instead, so we can
put the generic-loader into common-obj to save some compilation time.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

cpu: Provide a proper prototype for target_words_bigendian() in a header

We've got three places already that provide a prototype for this
function in a .c file - that's ugly. Let's provide a proper prototype
in a header instead, with a proper description why this function should
not be used in most cases.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

hw/core/generic-loader: Set a category for the generic-loader device

Each device that is instantiatable by the users should be marked with
a category. Since the generic-loader does not fit anywhere else, put
it into the MISC category.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

qemu/compiler: Wrap __attribute__((flatten)) in a macro

Older versions of Clang (before 3.5) and GCC (before 4.1) do not
support the "__attribute__((flatten))" yet. We don't care about
such old versions of GCC anymore, but since Clang 3.4 is still
used in EPEL for RHEL7 / CentOS 7, we should not use this attribute
directly but with a wrapper macro instead.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

mailmap: Fix Reimar Döffinger name

This probably happened when interpreting the utf8 name as latin1.

Fixes dbbaaff6867 and f4e94dfefb6.

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>

target/hppa: Raise exception 26 on emulated hardware

On PCXS chips (PA7000, pa 1.1a), trap #18 is raised on memory faults,
while all later chips (>= PA7100) generate either trap #26, #27 or #28
(depending on the fault type).

Since the current qemu emulation emulates a B160L machine (with a
PA7300LC PCX-L2 chip, we should raise trap #26 (EXCP_DMAR) instead
of #18 (EXCP_DMP) on access faults by the Linux kernel to page zero.

With the patch we now get the correct output (I tested against real
hardware):
 Kernel Fault: Code=26 (Data memory access rights trap)
instead of:
 Kernel Fault: Code=18 (Data memory protection/unaligned access trap)

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20181007205153.GA30270@ls3530.fritz.box>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20181016-1' into staging

target-arm queue:
 * hw/arm/virt: add DT property /secure-chosen/stdout-path indicating secure UART
 * target/arm: Fix aarch64_sve_change_el wrt EL0
 * target/arm: Define fields of ISAR registers
 * target/arm: Align cortex-r5 id_isar0
 * target/arm: Fix cortex-a7 id_isar0
 * net/cadence_gem: Fix various bugs, add support for new
   features that will be used by the Xilinx Versal board
 * target-arm: powerctl: Enable HVC when starting CPUs to EL2
 * target/arm: Add the Cortex-A72
 * target/arm: Mark PMINTENCLR and PMINTENCLR_EL1 accesses as possibly doing IO
 * target/arm: Mask PMOVSR writes based on supported counters
 * target/arm: Initialize ARMMMUFaultInfo in v7m_stack_read/write
 * coccinelle: new inplace-byteswaps.cocci to remove inplace-byteswapping calls

# gpg: Signature made Tue 16 Oct 2018 17:42:01 BST
# gpg:                using RSA key 3C2525ED14360CDE
# 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-20181016-1:
  coccinelle: new inplace-byteswaps.cocci to remove inplace-byteswapping calls
  target/arm: Initialize ARMMMUFaultInfo in v7m_stack_read/write
  target/arm: Mask PMOVSR writes based on supported counters
  target/arm: Mark PMINTENCLR and PMINTENCLR_EL1 accesses as possibly doing IO
  target/arm: Add the Cortex-A72
  target-arm: powerctl: Enable HVC when starting CPUs to EL2
  net: cadence_gem: Implement support for 64bit descriptor addresses
  net: cadence_gem: Add support for selecting the DMA MemoryRegion
  net: cadence_gem: Add support for extended descriptors
  net: cadence_gem: Add macro with max number of descriptor words
  net: cadence_gem: Use uint32_t for 32bit descriptor words
  net: cadence_gem: Disable TSU feature bit
  target/arm: Fix cortex-a7 id_isar0
  target/arm: Align cortex-r5 id_isar0
  target/arm: Define fields of ISAR registers
  target/arm: Fix aarch64_sve_change_el wrt EL0
  hw/arm/virt: add DT property /secure-chosen/stdout-path indicating secure UART

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

show-fixed-bugs.sh: Modern shell scripting (use $() instead of ``)

Various shell files contain a mix between obsolete ``
and modern $(); It would be nice to convert to using $()
everywhere.

Signed-off-by: Mao Zhongyi <maozhongyi@cmss.chinamobile.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

git-submodule.sh: Modern shell scripting (use $() instead of ``)

Various shell files contain a mix between obsolete ``
and modern $(); It would be nice to convert to using $()
everywhere.

Signed-off-by: Mao Zhongyi <maozhongyi@cmss.chinamobile.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

archive-source.sh: Modern shell scripting (use $() instead of ``)

Various shell files contain a mix between obsolete ``
and modern $(); It would be nice to convert to using $()
everywhere.

Signed-off-by: Mao Zhongyi <maozhongyi@cmss.chinamobile.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

coccinelle: new inplace-byteswaps.cocci to remove inplace-byteswapping calls

Add a new Coccinelle script which replaces uses of the inplace
byteswapping functions *_to_cpus() and cpu_to_*s() with their
not-in-place equivalents. This is useful for where the swapping
is done on members of a packed struct -- taking the address
of the member to pass it to an inplace function is undefined
behaviour in C.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181009181612.10633-1-peter.maydell@linaro.org

target/arm: Initialize ARMMMUFaultInfo in v7m_stack_read/write

The get_phys_addr() functions take a pointer to an ARMMMUFaultInfo
struct, which they fill in only if a fault occurs. This means that
the caller must always zero-initialize the struct before passing
it in. We forgot to do this in v7m_stack_read() and v7m_stack_write().
Correct the error.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181011172057.9466-1-peter.maydell@linaro.org

target/arm: Mask PMOVSR writes based on supported counters

This is an amendment to my earlier patch:
    commit 7ece99b17e832065236c07a158dfac62619ef99b
    Author: Aaron Lindsay <alindsay@codeaurora.org>
    Date:   Thu Apr 26 11:04:39 2018 +0100

target/arm: Mask PMU register writes based on PMCR_EL0.N

Signed-off-by: Aaron Lindsay <alindsay@codeaurora.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181010203735.27918-3-aclindsa@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Mark PMINTENCLR and PMINTENCLR_EL1 accesses as possibly doing IO

I previously fixed this for PMINTENSET_EL1, but missed these.

Signed-off-by: Aaron Lindsay <alindsay@codeaurora.org>
Signed-off-by: Aaron Lindsay <aclindsa@gmail.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181010203735.27918-2-aclindsa@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Add the Cortex-A72

Add the ARM Cortex-A72.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 20181011021931.4249-11-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target-arm: powerctl: Enable HVC when starting CPUs to EL2

When QEMU provides the equivalent of the EL3 firmware, we
need to enable HVCs in scr_el3 when turning on CPUs that
target EL2.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 20181011021931.4249-10-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

net: cadence_gem: Implement support for 64bit descriptor addresses

Implement support for 64bit descriptor addresses.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 20181011021931.4249-8-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

net: cadence_gem: Add support for selecting the DMA MemoryRegion

Add support for selecting the Memory Region that the GEM
will do DMA to.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20181011021931.4249-7-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

net: cadence_gem: Add support for extended descriptors

Add support for extended descriptors with optional 64bit
addressing and timestamping. QEMU will not yet provide
timestamps (always leaving the valid timestamp bit as zero).

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 20181011021931.4249-6-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

net: cadence_gem: Add macro with max number of descriptor words

Add macro with max number of DMA descriptor words.
No functional change.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 20181011021931.4249-5-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

net: cadence_gem: Use uint32_t for 32bit descriptor words

Use uint32_t instead of unsigned to describe 32bit descriptor words.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 20181011021931.4249-4-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

MAINTAINERS: update block/sheepdog maintainers

E-mail to one of block/sheepdog maintainers Mitake Hitoshi bounces

<mitake.hitoshi@lab.ntt.co.jp>: unknown user: "mitake.hitoshi"

and no current address is known. So just remove it.

Signed-off-by: Liu Yuan <liuyuan1@cmiot.chinamobile.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

gdbstub: Remove unused include

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests: remove gcov-files- variables

Commit 31d2dda ("build-system: remove per-test GCOV reporting", 2018-06-20)
removed users of the variables, since those uses can be replaced by a simple
overall report produced by gcovr.  However, the variables were never removed.
Do it now.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
[thuth: Fixed up contextual conflicts with the patch from Eric]
Signed-off-by: Thomas Huth <thuth@redhat.com>

tests: Prevent more accidental test disabling

GNU make is perfectly happy to use 'check-FOO-y += bar' to
initialize check-FOO-y.  (GNU Automake strictly insists that
you cannot use += until after an initial = per variable, but
thankfully we aren't using automake).

As we have had more than one instance where copy-and-paste of
'check-FOO-y = bar' from a first test under category FOO into
an additional test, which ends up disabling the first (see
commits 992159c7 and 4429532b), it's better to just always use
the form that survives copy-and-paste, even for categories that
don't currently add more than one test.

Done with s/^\(check-[a-z]*-y \)=/\1+=/g

Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

target/cris/translate: Get rid of qemu_log_separate()

The gen_BUG() function calls already cpu_abort(), which prints the
information to stderr and the log already. So instead of additionally
printing the dc->pc via fprintf() and qemu_log here, too, we can
simply pass this information to cpu_abort() instead.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>

qemu-common.h: update copyright date to 2018

Currently the copyright date is set to 2017. Update the date to say
2018.

Signed-off-by: John Arbuckle <programmingkidx@gmail.com>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Thomas Huth <thuth@redhat.com>

net: cadence_gem: Disable TSU feature bit

Disable the Timestamping Unit feature bit since QEMU does not
yet support it. This allows guest SW to correctly probe for
its existance.

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Message-id: 20181011021931.4249-2-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Fix cortex-a7 id_isar0

The incorrect value advertised only thumb2 div without arm div.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181008212205.17752-6-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Align cortex-r5 id_isar0

The missing nibble made it more difficult to read.

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181008212205.17752-5-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Define fields of ISAR registers

Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181008212205.17752-3-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Fix aarch64_sve_change_el wrt EL0

At present we assert:

  arm_el_is_aa64: Assertion `el >= 1 && el <= 3' failed.

The comment in arm_el_is_aa64 explains why asking about EL0 without
extra information is impossible.  Add an extra argument to provide
it from the surrounding context.

Fixes: 0ab5953b00b3
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181008212205.17752-2-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/virt: add DT property /secure-chosen/stdout-path indicating secure UART

Bindings for /secure-chosen and /secure-chosen/stdout-path have been
proposed 1.5 years ago [1] and implemented in OP-TEE at the same time [2].
They've now been officially agreed on, so we can implement them
in QEMU.

This patch creates the property when the machine is secure.

[1] https://patchwork.kernel.org/patch/9602401/
[2] https://github.com/OP-TEE/optee_os/commit/4dc31c52544a

Signed-off-by: Jerome Forissier <jerome.forissier@linaro.org>
Message-id: 20181005080729.6480-1-jerome.forissier@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
[PMM: commit message tweak]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge remote-tracking branch 'remotes/awilliam/tags/vfio-updates-20181015.0' into staging

VFIO updates 2018-10-15

 - ramfb support for vfio-pci via new -nohotplug device variant
   (Gerd Hoffmann)

 - Preparation for generic DT pass-through in vfio-platform
   (Geert Uytterhoeven & Eric Auger)

 - vfio-pci QOM fixups (Li Qiang)

# gpg: Signature made Mon 15 Oct 2018 18:26:29 BST
# gpg:                using RSA key 239B9B6E3BB08B22
# gpg: Good signature from "Alex Williamson <alex.williamson@redhat.com>"
# gpg:                 aka "Alex Williamson <alex@shazbot.org>"
# gpg:                 aka "Alex Williamson <alwillia@redhat.com>"
# gpg:                 aka "Alex Williamson <alex.l.williamson@gmail.com>"
# Primary key fingerprint: 42F6 C04E 540B D1A9 9E7B  8A90 239B 9B6E 3BB0 8B22

* remotes/awilliam/tags/vfio-updates-20181015.0:
  vfio-pci: make vfio-pci device more QOM conventional
  hw/arm/virt: Allow dynamic vfio-platform devices again
  hw/arm/sysbus-fdt: Allow device matching with DT compatible value
  vfio/platform: Make the vfio-platform device non-abstract
  hw/vfio/display: add ramfb support
  stubs: add ramfb

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

vfio-pci: make vfio-pci device more QOM conventional

Define a TYPE_VFIO_PCI and drop DO_UPCAST.

Signed-off-by: Li Qiang <liq3ea@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

hw/arm/virt: Allow dynamic vfio-platform devices again

Allow the instantation of generic dynamic vfio-platform devices again,
without the need to create a new device-specific vfio type.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

hw/arm/sysbus-fdt: Allow device matching with DT compatible value

Up to now we have relied on the device type to identify a device tree
node creation function.  Since we would like the vfio-platform device to
be instantiable with different compatible strings we introduce the
capability to specialize the node creation depending on actual
compatible value.

NodeCreationPair is renamed into BindingEntry. The struct is enhanced
with compat and match_fn() fields.  We introduce a new matching function
adapted to the vfio-platform generic device.

Soon, the AMD XGBE can be instantiated with either manner, i.e.:

    -device vfio-amd-xgbe,host=e0900000.xgmac

or using the new option line:

    -device vfio-platform,host=e0900000.xgmac

Signed-off-by: Eric Auger <eric.auger@redhat.com>
[geert: Match using compatible values in sysfs instead of user-supplied
manufacturer/model options, reword]
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

vfio/platform: Make the vfio-platform device non-abstract

Up to now the vfio-platform device has been abstract and could not be
instantiated.  The integration of a new vfio platform device required
creating a dummy derived device which only set the compatible string.

Following the few vfio-platform device integrations we have seen the
actual requested adaptation happens on device tree node creation
(sysbus-fdt).

Hence remove the abstract setting, and read the list of compatible
values from sysfs if not set by a derived device.

Update the amd-xgbe and calxeda-xgmac drivers to fill in the number of
compatible values, as there can now be more than one.

Note that sysbus-fdt does not support the instantiation of the
vfio-platform device yet.

Signed-off-by: Eric Auger <eric.auger@redhat.com>
[geert: Rebase, set user_creatable=true, use compatible values in sysfs
instead of user-supplied manufacturer/model options, reword]
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

hw/vfio/display: add ramfb support

So we have a boot display when using a vgpu as primary display.

ramfb depends on a fw_cfg file.  fw_cfg files can not be added and
removed at runtime, therefore a ramfb-enabled vfio device can't be
hotplugged.

Add a nohotplug variant of the vfio-pci device (as child class).  Add
the ramfb property to the nohotplug variant only.  So to enable the vgpu
display with boot support use this:

  -device vfio-pci-nohotplug,display=on,ramfb=on,sysfsdev=...

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

stubs: add ramfb

Needed to make sure code using ramfb (vfio) compiles properly even on
platforms without fw_cfg (and therefore no ramfb) support.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>

Merge remote-tracking branch 'remotes/kraxel/tags/vga-20181015-pull-request' into staging

vga: config tweaks, edid updates, qxl bugfix, install new vgabios blobs.

# gpg: Signature made Mon 15 Oct 2018 10:05:03 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/vga-20181015-pull-request:
  Makefile: Install new vgabios binaries
  hw/display/cirrus_vga: Move "isa-cirrus-vga" device into a separate file
  i2c: switch ddc to use the new edid generator
  qxl: check qxl_phys2virt return value
  bochs-display: wire up edid support
  display: add separate config option for bochs-display
  edid: fix alignment issues

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

Makefile: Install new vgabios binaries

Difficult to make use of if not installed

Fixes: cd1bfd5ef336 ("seabios: update bios and vgabios binaries")
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 153936155938.28040.11513367417790075721.stgit@gimli.home
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

hw/display/cirrus_vga: Move "isa-cirrus-vga" device into a separate file

In downstream distros like RHEL we'd like to disable some of the "legacy"
devices of QEMU. The ISA version of the Cirrus VGA device is one of these
legacy devices. So let's make the build process a little bit more flexible
here by putting the Cirrus ISA code into a separate file which is only
included if both, CONFIG_VGA_CIRRUS and CONFIG_VGA_ISA are set.

Note that this disables "isa-cirrus-vga" for the ppc-softmmu and the
alpha-softmmu target since CONFIG_VGA_ISA is not set there. But I think
this is OK since these targets are only interested in the PCI variant
anyway.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-id: 1539339106-32427-1-git-send-email-thuth@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

i2c: switch ddc to use the new edid generator

This also makes the default display resolution configurable,
via xres and yres properties.  The default is 1024x768.

The old code had a hard-coded resolution of 1600x1200.

Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20181005110837.28209-1-kraxel@redhat.com

qxl: check qxl_phys2virt return value

Fixes: CID 1395986
Fixes: 979f7ef8966bc4495a710ed9e4af42098f92ee79
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181005134608.1251-1-kraxel@redhat.com

bochs-display: wire up edid support

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20181005160147.892-3-kraxel@redhat.com

display: add separate config option for bochs-display

This allows modern architectures which don't care about vga
compatibility (risc-v for example) build bochs-display without
including all vga emulation too.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20181005160147.892-2-kraxel@redhat.com

edid: fix alignment issues

Use stl_le_p() & friends instead of casts.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20181015055333.32030-1-kraxel@redhat.com

Merge remote-tracking branch 'remotes/kraxel/tags/ui2-20181012-pull-request' into staging

ui: drop gtk2 support.

# gpg: Signature made Fri 12 Oct 2018 15:05:25 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/ui2-20181012-pull-request:
  ui: increase min required GTK3 version to 3.14.0
  ui: remove support for GTK2 in favour of GTK3

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

Merge remote-tracking branch 'remotes/kraxel/tags/ui-20181012-pull-request' into staging

ui: gtk fixes, support more formats in sdl2.

# gpg: Signature made Fri 12 Oct 2018 14:21:24 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/ui-20181012-pull-request:
  gtk: fix uninitialized variable
  sdl2: Support all virtio-gpu formats
  gtk: Don't vte_terminal_set_encoding() on new VTE versions

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

ui: increase min required GTK3 version to 3.14.0

Per supported platforms doc[1], the various min GTK3 on relevant distros is:

  RHEL-7.0: 3.8.8
  RHEL-7.2: 3.14.13
  RHEL-7.4: 3.22.10
  RHEL-7.5: 3.22.26
  Debian (Stretch): 3.22.11
  Debian (Jessie): 3.14.5
  OpenBSD (Ports): 3.22.30
  FreeBSD (Ports): 3.22.29
  OpenSUSE Leap 15: 3.22.30
  SLE12-SP2: Unknown
  Ubuntu (Xenial): 3.18.9
  macOS (Homebrew): 3.22.30

This suggests that a minimum GTK3 of 3.14.0 is a reasonable target,
as users are unlikely to be stuck on RHEL-7.0/7.1 still

[1] https://qemu.weilnetz.de/doc/qemu-doc.html#Supported-build-platforms

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20180822131554.3398-3-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

ui: remove support for GTK2 in favour of GTK3

GTK2 was deprecated in the 2.12.0 release with:

  commit b7715af2b31f47060cc5b4be930d16c13be93fa9
  Author: Daniel P. Berrange <berrange@redhat.com>
  Date:   Tue Dec 12 11:34:40 2017 +0000

    ui: deprecate use of GTK 2.x in favour of 3.x series

    The GTK 3.0 release was made in Feb, 2011:

      https://blog.gtk.org/2011/02/10/gtk-3-0-released/

    That will soon be 7 years ago, which is enough time to consider
    the 3.x series widely supported.

    Thus we deprecate the GTK 2.x support, which will allow us to
    delete it in the last release of 2018. By this time, GTK 3.x
    will be almost 8 years old.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
    Message-id: 20171212113440.16483-1-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
It is thus able to be removed in the 3.1.0 release.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20180822131554.3398-2-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

gtk: fix uninitialized variable

zoom_to_fit is never initialized to false, Coverity complains
(not sure why GCC does not).

Fixes: e8b1386ea1719525a1a92df03377764703fe8c64
Cc: kraxel@redhat.com
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181003121138.22037-1-pbonzini@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

sdl2: Support all virtio-gpu formats

There are some 2D resource formats that can be used through virtio-gpu,
but which are not supported by SDL2 when used for a scanout; these are
all alpha-channel formats and also XBGR (RGBX in non-BE pixman).

Add these formats in the switch converting pixman to SDL format
constants so a guest cannot crash the VM by triggering the
g_assert_not_reached() with an unsupported format.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 20181008185013.19371-1-mreitz@redhat.com

[ kraxel: also update sdl2_2d_check_format() ]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

gtk: Don't vte_terminal_set_encoding() on new VTE versions

The function vte_terminal_set_encoding() is deprecated since VTE 0.54,
so stop calling it from that version on. This fixes a build error
because of our use of warning flags [-Werror=deprecated-declarations].

Fixes: https://bugs.launchpad.net/bugs/1794939
Reported-by: Bastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-id: 20181011153039.2324-1-kwolf@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Merge remote-tracking branch 'remotes/cohuck/tags/s390x-20181012' into staging

More s390x updates:
- introduce support for vfio-ap (s390 crypto devices), including a
  Linux headers update to get the new interfaces
- the usual fixing + cleanup

# gpg: Signature made Fri 12 Oct 2018 10:54:38 BST
# gpg:                using RSA key DECF6B93C6F02FAF
# gpg: Good signature from "Cornelia Huck <conny@cornelia-huck.de>"
# gpg:                 aka "Cornelia Huck <huckc@linux.vnet.ibm.com>"
# gpg:                 aka "Cornelia Huck <cornelia.huck@de.ibm.com>"
# gpg:                 aka "Cornelia Huck <cohuck@kernel.org>"
# gpg:                 aka "Cornelia Huck <cohuck@redhat.com>"
# Primary key fingerprint: C3D0 D66D C362 4FF6 A8C0  18CE DECF 6B93 C6F0 2FAF

* remotes/cohuck/tags/s390x-20181012:
  hw/s390x: Include the tod-qemu also for builds with --disable-tcg
  s390: doc: detailed specifications for AP virtualization
  s390x/vfio: ap: Introduce VFIO AP device
  s390x/ap: base Adjunct Processor (AP) object model
  s390x/kvm: enable AP instruction interpretation for guest
  s390x/cpumodel: Set up CPU model for AP device support
  linux-headers: update
  target/s390x/excp_helper: Remove DPRINTF() macro

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

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

Block patches

One fix from Paolo on nvme:// driver.
One fix from Marc-Andre for iothread.

# gpg: Signature made Fri 12 Oct 2018 02:49:42 BST
# gpg:                using RSA key CA35624C6A9171C6
# gpg: Good signature from "Fam Zheng <famz@redhat.com>"
# Primary key fingerprint: 5003 7CB7 9706 0F76 F021  AD56 CA35 624C 6A91 71C6

* remotes/famz/tags/block-pull-request:
  nvme: correct locking around completion
  iothread: fix crash with invalid properties

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

Merge remote-tracking branch 'remotes/dgilbert/tags/pull-migration-20181011a' into staging

Migration pull 2018-10-11

With one bonus HMP fix.

# gpg: Signature made Thu 11 Oct 2018 20:23:12 BST
# gpg:                using RSA key 0516331EBC5BFDE7
# gpg: Good signature from "Dr. David Alan Gilbert (RH2) <dgilbert@redhat.com>"
# Primary key fingerprint: 45F5 C71B 4A0C B7FB 977A  9FA9 0516 331E BC5B FDE7

* remotes/dgilbert/tags/pull-migration-20181011a:
  migration-test: Only generate a single target architecture
  qmp, hmp: make subsystem/system-vendor identities optional
  vhost-user: Don't ask for reply on postcopy mem table set
  vhost-user: Fix userfaultfd leak
  migration: Stop postcopy fault thread before notifying
  tests/migration: Enable the migration test on s390x, too
  tests: Add migration test for aarch64

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

hw/s390x: Include the tod-qemu also for builds with --disable-tcg

The device is required for running qtests, see hw/s390x/tod.c:

void s390_init_tod(void)
{
    Object *obj;

    if (kvm_enabled()) {
        obj = object_new(TYPE_KVM_S390_TOD);
    } else {
        obj = object_new(TYPE_QEMU_S390_TOD);
    }
    [...]
  }

During qtests, we're running without kvm, so TYPE_QEMU_S390_TOD is
required to avoid that QEMU aborts here.

Fixes: 8046f374a6 ("s390x/tod: factor out TOD into separate device")
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1539264723-741-1-git-send-email-thuth@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390: doc: detailed specifications for AP virtualization

This patch provides documentation describing the AP architecture and
design concepts behind the virtualization of AP devices. It also
includes an example of how to configure AP devices for exclusive
use of KVM guests.

Signed-off-by: Tony Krowiak <akrowiak@linux.ibm.com>
Reviewed-by: Pierre Morel <pmorel@linux.ibm.com>
Tested-by: Pierre Morel <pmorel@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20181010170309.12045-7-akrowiak@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390x/vfio: ap: Introduce VFIO AP device

Introduces a VFIO based AP device. The device is defined via
the QEMU command line by specifying:

    -device vfio-ap,sysfsdev=<path-to-mediated-matrix-device>

There may be only one vfio-ap device configured for a guest.

The mediated matrix device is created by the VFIO AP device
driver by writing a UUID to a sysfs attribute file (see
docs/vfio-ap.txt). The mediated matrix device will be named
after the UUID. Symbolic links to the $uuid are created in
many places, so the path to the mediated matrix device $uuid
can be specified in any of the following ways:

/sys/devices/vfio_ap/matrix/$uuid
/sys/devices/vfio_ap/matrix/mdev_supported_types/vfio_ap-passthrough/devices/$uuid
/sys/bus/mdev/devices/$uuid
/sys/bus/mdev/drivers/vfio_mdev/$uuid

When the vfio-ap device is realized, it acquires and opens the
VFIO iommu group to which the mediated matrix device is
bound. This causes a VFIO group notification event to be
signaled. The vfio_ap device driver's group notification
handler will get called at which time the device driver
will configure the the AP devices to which the guest will
be granted access.

Signed-off-by: Tony Krowiak <akrowiak@linux.ibm.com>
Tested-by: Pierre Morel <pmorel@linux.ibm.com>
Acked-by: Halil Pasic <pasic@linux.ibm.com>
Tested-by: Pierre Morel <pmorel@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20181010170309.12045-6-akrowiak@linux.ibm.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
[CH: added missing g_free and device category]
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390x/ap: base Adjunct Processor (AP) object model

Introduces the base object model for virtualizing AP devices.

Signed-off-by: Tony Krowiak <akrowiak@linux.ibm.com>
Tested-by: Pierre Morel <pmorel@linux.ibm.com>
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20181010170309.12045-5-akrowiak@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390x/kvm: enable AP instruction interpretation for guest

Let's use the KVM_SET_DEVICE_ATTR ioctl to enable hardware
interpretation of AP instructions executed on the guest.
If the S390_FEAT_AP feature is switched on for the guest,
AP instructions must be interpreted by default; otherwise,
they will be intercepted.

This attribute setting may be overridden by a device. For example,
a device may want to provide AP instructions to the guest (i.e.,
S390_FEAT_AP turned on), but it may want to emulate them. In this
case, the AP instructions executed on the guest must be
intercepted; so when the device is realized, it must disable
interpretation.

Signed-off-by: Tony Krowiak <akrowiak@linux.ibm.com>
Tested-by: Pierre Morel <pmorel@linux.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: Halil Pasic <pasic@linux.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20181010170309.12045-4-akrowiak@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390x/cpumodel: Set up CPU model for AP device support

A new CPU model feature and two new CPU model facilities are
introduced to support AP devices for a KVM guest.

CPU model features:

1. The S390_FEAT_AP CPU model feature indicates whether AP
   instructions are available to the guest. This feature will
   be enabled only if the AP instructions are available on the
   linux host as determined by the availability of the
   KVM_S390_VM_CRYPTO_ENABLE_APIE VM attribute which is exposed
   by KVM only if the AP instructions are available on the
   host.

   This feature must be turned on from userspace to execute AP
   instructions on the KVM guest. The QEMU command line to turn
   this feature on looks something like this:

qemu-system-s390x ... -cpu xxx,ap=on ...

   This feature will be supported for zEC12 and newer CPU models.
   The feature will not be supported for older models because
   there are few older systems on which to test and the older
   crypto cards will be going out of service in the relatively
   near future.

CPU model facilities:

1. The S390_FEAT_AP_QUERY_CONFIG_INFO feature indicates whether the
   AP Query Configuration Information (QCI) facility is available
   to the guest as determined by whether the facility is available
   on the host. This feature will be exposed by KVM only if the
   QCI facility is installed on the host.

2. The S390_FEAT_AP_FACILITY_TEST feature indicates whether the AP
   Facility Test (APFT) facility is available to the guest as
   determined by whether the facility is available on the host.
   This feature will be exposed by KVM only if APFT is installed
   on the host.

Signed-off-by: Tony Krowiak <akrowiak@linux.ibm.com>
Tested-by: Pierre Morel <pmorel@linux.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Halil Pasic <pasic@linux.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20181010170309.12045-3-akrowiak@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

linux-headers: update

Update to kvm/next commit dd5bd0a65ff6 ("Merge tag 'kvm-s390-next-4.20-1'
of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD")

Signed-off-by: Cornelia Huck <cohuck@redhat.com>

target/s390x/excp_helper: Remove DPRINTF() macro

Debug macros that are disabled by default should be avoided (since the
code bit-rots quite easily). Thus turn these debug prints into proper
qemu_log_mask(CPU_LOG_xxx, ...) statements instead. The DPRINTF statements
in do_[ext|io|mchk]_interrupt can even be removed completely since we can
log the information in a central place, s390_cpu_do_interrupt, instead.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1538751601-7433-1-git-send-email-thuth@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

nvme: correct locking around completion

nvme_poll_queues is already protected by q->lock, and
AIO callbacks are invoked outside the AioContext lock.
So remove the acquire/release pair in nvme_handle_event.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20180814062739.19640-1-pbonzini@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>

iothread: fix crash with invalid properties

-object iothread,id=foo,? will crash qemu:

qemu-system-x86_64:qemu-thread-posix.c:128: qemu_cond_destroy: Assertion `cond->initialized' failed.

Use thread_id != -1 to check if iothread_complete() finished
successfully and the mutex/cond have been initialized.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180821100716.13803-1-marcandre.lureau@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>

migration-test: Only generate a single target architecture

Several changes:
- We only allow generate header "inside" the tree.  Why?  Because we
  need to connit the result, so it makes no sense to generate them on
  the build dir.
- We only generate a single target each time.  Getting all the
  cross-compilers correctly is an impossible task.  So know you do:
     make -C tests/migration $target (native)
     make CROSS_PREFIX=foo- -C tests/migratiion $target (cross)
  And you are done.

- If we are building out of tree, we have no data about if we are
  cross-compile or whatever.  So instead of guess what is happening,
  just do what I pointed on previous point.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20180913132313.11370-1-quintela@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

qmp, hmp: make subsystem/system-vendor identities optional

According to PCI specification, subsystem id and subsystem vendor id
are present only in type 0 and type 2 headers (at different offsets),
but not in type 1 headers.

Thus we should make this data optional in struct PciDeviceId and skip
reporting them via HMP if the information is not available.

Additional (wrong information) about PCI bridges (Type1 devices) has been
added in 5383a705 and fortunately not released. This patch fixes that
problem. The problem was spotted by Markus.

Signed-off-by: Denis V. Lunev <den@openvz.org>
CC: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
CC: Eric Blake <eblake@redhat.com>
CC: Markus Armbruster <armbru@redhat.com>
Message-Id: <20181002135538.12113-1-den@openvz.org>
Reported-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

vhost-user: Don't ask for reply on postcopy mem table set

According to documentation, NEED_REPLY_MASK should not be set
for VHOST_USER_SET_MEM_TABLE request in postcopy mode.
This restriction was mistakenly applied to 'reply_supported'
variable, which is local and used only for non-postcopy case.

CC: Dr. David Alan Gilbert <dgilbert@redhat.com>
Fixes: 9bb38019942c ("vhost+postcopy: Send address back to qemu")
Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
Message-Id: <20181002140947.4107-1-i.maximets@samsung.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

vhost-user: Fix userfaultfd leak

'fd' received from the vhost side is never freed.
Also, everything (including 'postcopy_listen' state) should be
cleaned up on vhost cleanup.

Fixes: 46343570c06e ("vhost+postcopy: Wire up POSTCOPY_END notify")
Fixes: f82c11165ffa ("vhost+postcopy: Register shared ufd with postcopy")
Cc: qemu-stable@nongnu.org
Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
Message-Id: <20181008160536.6332-3-i.maximets@samsung.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

migration: Stop postcopy fault thread before notifying

POSTCOPY_NOTIFY_INBOUND_END handlers will remove userfault fds
from the postcopy_remote_fds array which could be still in
use by the fault thread. Let's stop the thread before
notification to avoid possible accessing wrong memory.

Fixes: 46343570c06e ("vhost+postcopy: Wire up POSTCOPY_END notify")
Cc: qemu-stable@nongnu.org
Signed-off-by: Ilya Maximets <i.maximets@samsung.com>
Message-Id: <20181008160536.6332-2-i.maximets@samsung.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests/migration: Enable the migration test on s390x, too

We can re-use the s390-ccw bios code to implement a small firmware
for a s390x guest which prints out the "A" and "B" characters and
modifies the memory, as required for the migration test.

[quintela: Converted the compile script to Makefile rules]
Signed-off-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1539078677-25396-1-git-send-email-thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
  Fixed up Makefile since the aarch patch sneaked in first

tests: Add migration test for aarch64

This patch adds migration test support for aarch64. The test code, which
implements the same functionality as x86, is booted as a kernel in qemu.
Here are the design choices we make for aarch64:

 * We choose this -kernel approach because aarch64 QEMU doesn't provide a
   built-in fw like x86 does. So instead of relying on a boot loader, we
   use -kernel approach for aarch64.
 * The serial output is sent to PL011 directly.
 * The physical memory base for mach-virt machine is 0x40000000. We change
   the start_address and end_address for aarch64.

In addition to providing the binary, this patch also includes the source
code and the build script in tests/migration/aarch64. So users can change
the source and/or re-compile the binary as they wish.

Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Wei Huang <wei@redhat.com>
Message-Id: <1538669326-28135-1-git-send-email-wei@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

Merge remote-tracking branch 'remotes/armbru/tags/pull-misc-2018-10-10' into staging

Miscellaneous patches for 2018-10-10

# gpg: Signature made Wed 10 Oct 2018 07:03:10 BST
# gpg:                using RSA key 3870B400EB918653
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>"
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>"
# Primary key fingerprint: 354B C8B3 D7EB 2A6B 6867  4E5F 3870 B400 EB91 8653

* remotes/armbru/tags/pull-misc-2018-10-10:
  ivshmem: Fix unplug of device "ivshmem-plain"
  libqtest: Inline g_assert_no_errno()
  tests: Restore check-qdict unit test
  qapi/misc.json: Remove superfluous words in CpuModelExpansionType
  tests/check-qjson: fix a leak

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

ivshmem: Fix unplug of device "ivshmem-plain"

Commit 2aece63c8a "hostmem: detect host backend memory is being used
properly" fixed "ivshmem-plain" to reject memory backends that are
already in use, and to block their deletion while in use.  Two bugs
escaped review:

* New ivshmem_plain_exit() fails to call ivshmem_exit().  This breaks
  unplug.  Reproducer: migration after unplug still fails with
  "Migration is disabled when using feature 'peer mode' in device
  'ivshmem'".

* It failed to update legacy "ivshmem".  Harmless, because it creates
  the memory backend itself, and nothing else should use it.

Fix by moving the two host_memory_backend_set_mapped() calls into
ivshmem_common_realize() and ivshmem_exit(), guarded by s->hostmem.

Fixes: 2aece63c8a9d2c3a8ff41d2febc4cdeff2633331
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20180926163709.22876-1-armbru@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>

libqtest: Inline g_assert_no_errno()

Macro g_assert_no_errno() intrudes into GLib's namespace.  It's also
pretty pointless.  Inline.

At one call site, its redundancy is now obvious.  Delete it there.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20180926122933.3858-1-armbru@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>

tests: Restore check-qdict unit test

Commit 0bcc8e5bd8d accidentally dropped check-qdict from the list of
unit tests (again, see commit 4429532b48a).  Put it back, and fix up
the test.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20180926122309.30631-1-armbru@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>

qapi/misc.json: Remove superfluous words in CpuModelExpansionType

Remove repetition of 'independent of'.  While at it, s/QMU/QEMU in
@CpuDefinitionInfo.

Signed-off-by: Kashyap Chamarthy <kchamart@redhat.com>
Message-Id: <20181003104605.8477-1-kchamart@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
[Commit message updated as requested by Kashyap]
Signed-off-by: Markus Armbruster <armbru@redhat.com>