qga: update guest-suspend-ram and guest-suspend-hybrid descriptions

This patch updates the descriptions of 'guest-suspend-ram' and
'guest-suspend-hybrid' to mention that both commands relies now
on the proper support for wake up from suspend, retrieved by the
'wakeup-suspend-support' attribute of the 'query-current-machine'
QMP command.

Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Message-Id: <20181205194701.17836-3-danielhb413@gmail.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qmp: query-current-machine with wakeup-suspend-support

When issuing the qmp/hmp 'system_wakeup' command, what happens in a
nutshell is:

- qmp_system_wakeup_request set runstate to RUNNING, sets a wakeup_reason
and notify the event
- in the main_loop, all vcpus are paused, a system reset is issued, all
subscribers of wakeup_notifiers receives a notification, vcpus are then
resumed and the wake up QAPI event is fired

Note that this procedure alone doesn't ensure that the guest will awake
from SUSPENDED state - the subscribers of the wake up event must take
action to resume the guest, otherwise the guest will simply reboot. At
this moment, only the ACPI machines via acpi_pm1_cnt_init and xen_hvm_init
have wake-up from suspend support.

However, only the presence of 'system_wakeup' is required for QGA to
support 'guest-suspend-ram' and 'guest-suspend-hybrid' at this moment.
This means that the user/management will expect to suspend the guest using
one of those suspend commands and then resume execution using system_wakeup,
regardless of the support offered in system_wakeup in the first place.

This patch creates a new API called query-current-machine [1], that holds
a new flag called 'wakeup-suspend-support' that indicates if the guest
supports wake up from suspend via system_wakeup. The machine is considered
to implement wake-up support if a call to a new 'qemu_register_wakeup_support'
is made during its init, as it is now being done inside acpi_pm1_cnt_init
and xen_hvm_init. This allows for any other machine type to declare wake-up
support regardless of ACPI state or wakeup_notifiers subscription, making easier
for newer implementations that might have their own mechanisms in the future.

This is the expected output of query-current-machine when running a x86
guest:

{"execute" : "query-current-machine"}
{"return": {"wakeup-suspend-support": true}}

Running the same x86 guest, but with the --no-acpi option:

{"execute" : "query-current-machine"}
{"return": {"wakeup-suspend-support": false}}

This is the output when running a pseries guest:

{"execute" : "query-current-machine"}
{"return": {"wakeup-suspend-support": false}}

With this extra tool, management can avoid situations where a guest
that does not have proper suspend/wake capabilities ends up in
inconsistent state (e.g.
https://github.com/open-power-host-os/qemu/issues/31).

[1] the decision of creating the query-current-machine API is based
on discussions in the QEMU mailing list where it was decided that
query-target wasn't a proper place to store the wake-up flag, neither
was query-machines because this isn't a static property of the
machine object. This new API can then be used to store other
dynamic machine properties that are scattered around the code
ATM. More info at:
https://lists.gnu.org/archive/html/qemu-devel/2018-05/msg04235.html

Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20181205194701.17836-2-danielhb413@gmail.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qmp: Split ShutdownCause host-qmp into quit and system-reset

It is interesting to know whether the shutdown cause was 'quit' or
'reset', especially when using "--no-reboot". In that case, a management
layer can now determine if the guest wanted a reboot or shutdown, and
can act accordingly.

Changes the output of the reason in the iotests from 'host-qmp' to
'host-qmp-quit'. This does not break compatibility because
the field was introduced in the same version.

Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Message-Id: <20181205110131.23049-4-d.csapak@proxmox.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qmp: Add reason to SHUTDOWN and RESET events

This makes it possible to determine what the exact reason was for
a RESET or a SHUTDOWN. A management layer might need the specific reason
of those events to determine which cleanups or other actions it needs to do.

This patch also updates the iotests to the new expected output that includes
the reason.

Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Message-Id: <20181205110131.23049-3-d.csapak@proxmox.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qapi: Turn ShutdownCause into QAPI enum

Needed so the patch after next can add ShutdownCause to QMP events
SHUTDOWN and RESET.

Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
Message-Id: <20181205110131.23049-2-d.csapak@proxmox.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

Merge remote-tracking branch 'remotes/rth/tags/pull-tcg-20181216' into staging

- Remove retranslation remenents
- Return success from patch_reloc
- Preserve 32-bit values as zero-extended on x86_64
- Make bswap during memory ops as optional
- Cleanup xxhash
- Revert constant pooling for tcg/sparc/

# gpg: Signature made Mon 17 Dec 2018 03:25:21 GMT
# 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-tcg-20181216: (33 commits)
  xxhash: match output against the original xxhash32
  include: move exec/tb-hash-xx.h to qemu/xxhash.h
  exec: introduce qemu_xxhash{2,4,5,6,7}
  qht-bench: document -p flag
  tcg: Drop nargs from tcg_op_insert_{before,after}
  tcg/mips: Improve the add2/sub2 command to use TCG_TARGET_REG_BITS
  tcg: Add TCG_TARGET_HAS_MEMORY_BSWAP
  tcg/optimize: Optimize bswap
  tcg: Clean up generic bswap64
  tcg: Clean up generic bswap32
  tcg/i386: Add setup_guest_base_seg for FreeBSD
  tcg/i386: Precompute all guest_base parameters
  tcg/i386: Assume 32-bit values are zero-extended
  tcg/i386: Implement INDEX_op_extr{lh}_i64_i32 for 32-bit guests
  tcg/i386: Propagate is64 to tcg_out_qemu_ld_slow_path
  tcg/i386: Propagate is64 to tcg_out_qemu_ld_direct
  tcg/s390x: Return false on failure from patch_reloc
  tcg/ppc: Return false on failure from patch_reloc
  tcg/arm: Return false on failure from patch_reloc
  tcg/aarch64: Return false on failure from patch_reloc
  ...

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

.shippable.yml: disable the win cross tests

The pkg.mxe.cc package repositories have been down for the last two
weeks causing the builds to fail when shippable re-builds the
containers.

This is really just a sticking plaster until we can get our own docker
hub images properly setup so we can avoid having dependencies on
external repos.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181214151718.5041-1-alex.bennee@linaro.org
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

xxhash: match output against the original xxhash32

Change the order in which we extract a/b and c/d to
match the output of the upstream xxhash32.

Tested with:
  https://github.com/cota/xxhash/tree/qemu

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

include: move exec/tb-hash-xx.h to qemu/xxhash.h

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

exec: introduce qemu_xxhash{2,4,5,6,7}

Before moving them all to include/qemu/xxhash.h.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

qht-bench: document -p flag

Which we forgot to do in bd224fce60 ("qht-bench: add -p flag
to precompute hash values", 2018-09-26).

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Drop nargs from tcg_op_insert_{before,after}

It's unused since 75e8b9b7aa0b95a761b9add7e2f09248b101a392.

Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <20181209193749.12277-9-cota@braap.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/mips: Improve the add2/sub2 command to use TCG_TARGET_REG_BITS

Instead of hard coding 31 for the shift right use TCG_TARGET_REG_BITS - 1.

Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <7dfbddf7014a595150aa79011ddb342c3cc17ec3.1544648105.git.alistair.francis@wdc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Add TCG_TARGET_HAS_MEMORY_BSWAP

For now, defined universally as true, since we previously required
backends to implement swapped memory operations.  Future patches
may now remove that support where it is onerous.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/optimize: Optimize bswap

Somehow we forgot these operations, once upon a time.
This will allow immediate stores to have their bswap
optimized away.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Clean up generic bswap64

Based on the only current user, Sparc:

New code uses 2 constants that take 2 insns to load from constant pool,
plus 13.  Old code used 6 constants that took 1 or 2 insns to create,
plus 21.  The result is a new total of 17 vs an old total of 29.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Clean up generic bswap32

Based on the only current user, Sparc:

New code uses 1 constant that takes 2 insns to create, plus 8.
Old code used 2 constants that took 2 insns to create, plus 9.
The result is a new total of 10 vs an old total of 13.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Add setup_guest_base_seg for FreeBSD

Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Precompute all guest_base parameters

These values are constant between all qemu_ld/st invocations;
there is no need to figure this out each time.  If we cannot
use a segment or an offset directly for guest_base, load the
value into a register in the prologue.

Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Assume 32-bit values are zero-extended

We now have an invariant that all TCG_TYPE_I32 values are
zero-extended, which means that we do not need to extend
them again during qemu_ld/st, either explicitly via a separate
tcg_out_ext32u or implicitly via P_ADDR32.

Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Implement INDEX_op_extr{lh}_i64_i32 for 32-bit guests

This preserves the invariant that all TCG_TYPE_I32 values are
zero-extended in the 64-bit host register.

Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Propagate is64 to tcg_out_qemu_ld_slow_path

This helps preserve the invariant that all TCG_TYPE_I32 values
are stored zero-extended in the 64-bit host registers.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Propagate is64 to tcg_out_qemu_ld_direct

This helps preserve the invariant that all TCG_TYPE_I32 values
are stored zero-extended in the 64-bit host registers.

Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/s390x: Return false on failure from patch_reloc

This does require an extra two checks within the slow paths
to replace the assert that we're moving.  Also add two checks
within existing functions that lacked any kind of assert for
out of range branch.

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

tcg/ppc: Return false on failure from patch_reloc

The reloc_pc{14,24}_val routines retain their asserts.
Use these directly within the slow paths.

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

tcg/arm: Return false on failure from patch_reloc

This does require an extra two checks within the slow paths
to replace the assert that we're moving.

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

tcg/aarch64: Return false on failure from patch_reloc

This does require an extra two checks within the slow paths
to replace the assert that we're moving.

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

tcg/i386: Return false on failure from patch_reloc

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

tcg: Return success from patch_reloc

This will move the assert for success from within (subroutines of)
patch_reloc into the callers.  It will also let new code do something
different when a relocation is out of range.

For the moment, all backends are trivially converted to return true.

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

tcg/mips: Remove retranslation code

There is no longer a need for preserving branch offset operands,
as we no longer re-translate.

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

tcg/sparc: Remove retranslation code

There is no longer a need for preserving branch offset operands,
as we no longer re-translate.

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

tcg/s390: Remove retranslation code

There is no longer a need for preserving branch offset operands,
as we no longer re-translate.

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

tcg/ppc: Fold away "noaddr" branch routines

There is no longer a need for preserving branch offset operands,
as we no longer re-translate.

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

tcg/arm: Fold away "noaddr" branch routines

There are one use apiece for these.  There is no longer a need for
preserving branch offset operands, as we no longer re-translate.

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

tcg/arm: Remove reloc_pc24_atomic

It is unused since 3fb53fb4d12f2e7833bd1659e6013237b130ef20.

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

tcg/aarch64: Fold away "noaddr" branch routines

There are one use apiece for these.  There is no longer a need for
preserving branch offset operands, as we no longer re-translate.

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

tcg/aarch64: Remove reloc_pc26_atomic

It is unused since b68686bd4bfeb70040b4099df993dfa0b4f37b03.

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

tcg/i386: Move TCG_REG_CALL_STACK from define to enum

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/i386: Always use %ebp for TCG_AREG0

For x86_64, this can remove a REX prefix resulting in smaller code
when manipulating globals of type i32, as we move them between backing
store via cpu_env, aka TCG_AREG0.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/sparc: Remove the constant pool

Partially reverts ab20bdc1162.  The 14-bit displacement that we
allowed to reach the constant pool is not always sufficient.
Retain the tb-relative addressing, as that is how most return
values from the tb are computed.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge remote-tracking branch 'remotes/pmaydell/tags/pull-misc-20181214' into staging

miscellaneous patches:
 * checkpatch.pl: Enforce multiline comment syntax
 * Rename cpu_physical_memory_write_rom() to address_space_write_rom()
 * disas, monitor, elf_ops: Use address_space_read() to read memory
 * Remove load_image() in favour of load_image_size()
 * Fix some minor memory leaks in arm boards/devices
 * virt: fix broken indentation

# gpg: Signature made Fri 14 Dec 2018 14:41:20 GMT
# 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-misc-20181214: (22 commits)
  virt: Fix broken indentation
  target/arm: Create timers in realize, not init
  tests/test-arm-mptimer: Don't leak string memory
  hw/sd/sdhci: Don't leak memory region in sdhci_sysbus_realize()
  hw/arm/mps2-tz.c: Free mscname string in make_dma()
  target/arm: Free name string in ARMCPRegInfo hashtable entries
  include/hw/loader.h: Document load_image_size()
  hw/core/loader.c: Remove load_image()
  device_tree.c: Don't use load_image()
  hw/block/tc58128.c: Don't use load_image()
  hw/i386/multiboot.c: Don't use load_image()
  hw/i386/pc.c: Don't use load_image()
  hw/pci/pci.c: Don't use load_image()
  hw/smbios/smbios.c: Don't use load_image()
  hw/ppc/ppc405_boards: Don't use load_image()
  hw/ppc/mac_newworld, mac_oldworld: Don't use load_image()
  elf_ops.h: Use address_space_write() to write memory
  monitor: Use address_space_read() to read memory
  disas.c: Use address_space_read() to read memory
  Rename cpu_physical_memory_write_rom() to address_space_write_rom()
  ...

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

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

Block layer patches:

- qcow2: Decompression worker threads
- dmg: lzfse compression support
- file-posix: Simplify delegation to worker thread
- Don't pass flags to bdrv_reopen_queue()
- iotests: make 235 work on s390 (and others)

# gpg: Signature made Fri 14 Dec 2018 10:55:09 GMT
# gpg:                using RSA key 7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kevin/tags/for-upstream: (42 commits)
  block/mirror: add missing coroutine_fn annotations
  iotests: make 235 work on s390 (and others)
  block: Assert that flags are up-to-date in bdrv_reopen_prepare()
  block: Remove assertions from update_flags_from_options()
  block: Stop passing flags to bdrv_reopen_queue_child()
  block: Remove flags parameter from bdrv_reopen_queue()
  block: Clean up reopen_backing_file() in block/replication.c
  qemu-io: Put flag changes in the options QDict in reopen_f()
  block: Drop bdrv_reopen()
  block: Use bdrv_reopen_set_read_only() in the mirror driver
  block: Use bdrv_reopen_set_read_only() in external_snapshot_commit()
  block: Use bdrv_reopen_set_read_only() in qmp_change_backing_file()
  block: Use bdrv_reopen_set_read_only() in stream_start/complete()
  block: Use bdrv_reopen_set_read_only() in bdrv_commit()
  block: Use bdrv_reopen_set_read_only() in commit_start/complete()
  block: Use bdrv_reopen_set_read_only() in bdrv_backing_update_filename()
  block: Add bdrv_reopen_set_read_only()
  file-posix: Avoid aio_worker() for QEMU_AIO_IOCTL
  file-posix: Switch to .bdrv_co_ioctl
  file-posix: Remove paio_submit_co()
  ...

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

Merge remote-tracking branch 'remotes/kraxel/tags/usb-20181214-pull-request' into staging

usb: fixes for mtp, ehci, usb-host and pvusb (xen).

# gpg: Signature made Fri 14 Dec 2018 10:38:33 GMT
# 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/usb-20181214-pull-request:
  usb-mtp: Limit filename to object information size
  usb-mtp: use O_NOFOLLOW and O_CLOEXEC.
  ehci: fix fetch qtd race
  usb-host: reset and close libusb_device_handle before qemu exit
  pvusb: set max grants only in initialise

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

Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2018-12-13-v2' into staging

QAPI patches for 2018-12-13

# gpg: Signature made Fri 14 Dec 2018 05:53:51 GMT
# 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-qapi-2018-12-13-v2: (32 commits)
  qapi: add conditions to REPLICATION type/commands on the schema
  qapi: add more conditions to SPICE
  qapi: add condition to variants documentation
  qapi: add 'If:' condition to struct members documentation
  qapi: add 'If:' condition to enum values documentation
  qapi: Add #if conditions to generated code members
  qapi: add 'if' to alternate members
  qapi: add 'if' to union members
  qapi: Add 'if' to implicit struct members
  qapi: add a dictionary form for TYPE
  qapi-events: add 'if' condition to implicit event enum
  qapi: add 'if' to enum members
  qapi: add a dictionary form with 'name' key for enum members
  qapi: improve reporting of unknown or missing keys
  qapi: factor out checking for keys
  tests: print enum type members more like object type members
  qapi: change enum visitor and gen_enum* to take QAPISchemaMember
  qapi: Do not define enumeration value explicitly
  qapi: break long lines at 'data' member
  qapi: rename QAPISchemaEnumType.values to .members
  ...

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

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

Most notable change in this PR is the full removal of the "handle" fsdev
backend.

# gpg: Signature made Wed 12 Dec 2018 13:20:42 GMT
# gpg:                using RSA key 71D4D5E5822F73D6
# gpg: Good signature from "Greg Kurz <groug@kaod.org>"
# gpg:                 aka "Gregory Kurz <gregory.kurz@free.fr>"
# gpg:                 aka "[jpeg image of size 3330]"
# Primary key fingerprint: B482 8BAF 9431 40CE F2A3  4910 71D4 D5E5 822F 73D6

* remotes/gkurz/tags/for-upstream:
  9p: remove support for the "handle" backend
  xen/9pfs: use g_new(T, n) instead of g_malloc(sizeof(T) * n)
  9p: use g_new(T, n) instead of g_malloc(sizeof(T) * n)

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

Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20181213' into staging

target-arm queue:
 * Convert various devices from sysbus init to instance_init
 * Remove the now unused sysbus init support entirely
 * Allow AArch64 processors to boot from a kernel placed over 4GB
 * hw: arm: musicpal: drop TYPE_WM8750 in object_property_set_link()
 * versal: minor fixes to virtio-mmio instantation
 * arm: Implement the ARMv8.1-HPD extension
 * arm: Implement the ARMv8.2-AA32HPD extension
 * arm: Implement the ARMv8.1-LOR extension (as the trivial
   "no limited ordering regions provided" minimum)

# gpg: Signature made Thu 13 Dec 2018 14:52:25 GMT
# 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-20181213: (37 commits)
  target/arm: Implement the ARMv8.1-LOR extension
  target/arm: Use arm_hcr_el2_eff more places
  target/arm: Introduce arm_hcr_el2_eff
  target/arm: Implement the ARMv8.2-AA32HPD extension
  target/arm: Implement the ARMv8.1-HPD extension
  target/arm: Tidy scr_write
  target/arm: Fix HCR_EL2.TGE check in arm_phys_excp_target_el
  target/arm: Add SCR_EL3 bits up to ARMv8.5
  target/arm: Add HCR_EL2 bits up to ARMv8.5
  target/arm: Move id_aa64mmfr* to ARMISARegisters
  hw/arm: versal: Correct the nr of IRQs to 192
  hw/arm: versal: Use IRQs 111 - 118 for virtio-mmio
  hw/arm: versal: Reduce number of virtio-mmio instances
  hw/arm: versal: Remove bogus virtio-mmio creation
  core/sysbus: remove the SysBusDeviceClass::init path
  xen_backend: remove xen_sysdev_init() function
  usb/tusb6010: Convert sysbus init function to realize function
  timer/puv3_ost: Convert sysbus init function to realize function
  timer/grlib_gptimer: Convert sysbus init function to realize function
  timer/etraxfs_timer: Convert sysbus init function to realize function
  ...

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

virt: Fix broken indentation

I introduced indentation using tabs instead of spaces in another
commit.  Peter reported the problem, and I failed to fix that
before sending my pull request.

Reported-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181212003147.29604-1-ehabkost@redhat.com
Fixes: 951597607696 ("virt: Eliminate separate instance_init functions")
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

target/arm: Create timers in realize, not init

The timer_new() function allocates memory; this means that
if we call it in the CPU's init method we would need
to provide an instance_finalize method to free it. Defer
the timer creation to the realize function instead.

This fixes a memory leak spotted by clang LeakSanitizer
when a CPU object is created for introspection.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20181204132952.2601-6-peter.maydell@linaro.org

tests/test-arm-mptimer: Don't leak string memory

The test-arm-mptimer setup creates a lot of test names using
g_strdup_printf() and never frees them. This is entirely
harmless since it's one-shot test code, but it clutters
up the output from clang's LeakSanitizer. Refactor to
use a helper function so we can free the memory.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181204132952.2601-5-peter.maydell@linaro.org

hw/sd/sdhci: Don't leak memory region in sdhci_sysbus_realize()

In sdhci_sysbus_realize() we override the initialization of
s->iomem that sdhci_common_realize() performs. However we
don't destroy the old memory region before reinitializing
it, which means that the memory allocated for mr->name in
memory_region_do_init() is leaked.

Since sdhci_initfn() already initializes s->io_ops to
&sdhci_mmio_ops, always use that in sdhci_common_realize()
and remove the now-unnecessary reinitialization of the
MMIO region from sdhci_sysbus_realize().

Spotted by clang's leak sanitizer.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181204132952.2601-4-peter.maydell@linaro.org

hw/arm/mps2-tz.c: Free mscname string in make_dma()

The clang leak sanitizer spots a (one-off, trivial) memory
leak in make_dma() due to a missing free.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181204132952.2601-3-peter.maydell@linaro.org

target/arm: Free name string in ARMCPRegInfo hashtable entries

When we add a new entry to the ARMCPRegInfo hash table in
add_cpreg_to_hashtable(), we allocate memory for tehe
ARMCPRegInfo struct itself, and we also g_strdup() the
name string. So the hashtable's value destructor function
must free the name string as well as the struct.

Spotted by clang's leak sanitizer. The leak here is a
small one-off leak at startup, because we don't support
CPU hotplug, and so the only time when we destroy
hash table entries is for the case where ARM_CP_OVERRIDE
means we register a wildcard entry and then override it later.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181204132952.2601-2-peter.maydell@linaro.org

include/hw/loader.h: Document load_image_size()

Add a documentation comment for load_image_size().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-11-peter.maydell@linaro.org

hw/core/loader.c: Remove load_image()

The load_image() function is now no longer used anywhere, so
we can remove it completely. (Use load_image_size() or
g_file_get_contents() instead.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-10-peter.maydell@linaro.org

device_tree.c: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Instead use load_image_size().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-9-peter.maydell@linaro.org

hw/block/tc58128.c: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Instead use load_image_size().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-8-peter.maydell@linaro.org

hw/i386/multiboot.c: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Instead use load_image_size().

While we are converting the code, add the missing error check.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-7-peter.maydell@linaro.org

hw/i386/pc.c: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Use the glib g_file_get_contents() function instead, which does
the whole "allocate memory for the file and read it in" operation.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-6-peter.maydell@linaro.org

hw/pci/pci.c: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Instead use load_image_size().

While we are converting this code, add an error-check
for read failure.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-5-peter.maydell@linaro.org

hw/smbios/smbios.c: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Instead use load_image_size().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20181130151712.2312-4-peter.maydell@linaro.org

hw/ppc/ppc405_boards: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Instead use load_image_size().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Message-id: 20181130151712.2312-3-peter.maydell@linaro.org

hw/ppc/mac_newworld, mac_oldworld: Don't use load_image()

The load_image() function is deprecated, as it does not let the
caller specify how large the buffer to read the file into is.
Use the glib g_file_get_contents() function instead, which does
the whole "allocate memory for the file and read it in" operation.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Message-id: 20181130151712.2312-2-peter.maydell@linaro.org

elf_ops.h: Use address_space_write() to write memory

Currently the load_elf function in elf_ops.h uses
cpu_physical_memory_write() to write the ELF file to
memory if it is not handling it as a ROM blob. This
means we ignore the AddressSpace that the function
is passed to define where it should be loaded.
Use address_space_write() instead.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181122172653.3413-4-peter.maydell@linaro.org

monitor: Use address_space_read() to read memory

Currently monitor.c reads physical memory using
cpu_physical_memory_read(). This effectively hard-codes
assuming that all CPUs have the same view of physical
memory. Switch to address_space_read() instead, which
lets us use the AddressSpace for the CPU we're
reading memory for (falling back to address_space_memory
if there is no CPU, as happens with the "none" board).
As a bonus, this allows us to detect failures to read memory.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181122172653.3413-3-peter.maydell@linaro.org

disas.c: Use address_space_read() to read memory

Currently disas.c reads physical memory using
cpu_physical_memory_read(). This effectively hard-codes
assuming that all CPUs have the same view of physical
memory. Switch to address_space_read() instead, which
lets us use the AddressSpace for the CPU we're
disassembling for.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20181122172653.3413-2-peter.maydell@linaro.org

Rename cpu_physical_memory_write_rom() to address_space_write_rom()

The API of cpu_physical_memory_write_rom() is odd, because it
takes an AddressSpace, unlike all the other cpu_physical_memory_*
access functions. Rename it to address_space_write_rom(), and
bring its API into line with address_space_write().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 20181122133507.30950-3-peter.maydell@linaro.org

exec.c: Rename cpu_physical_memory_write_rom_internal()

Rename cpu_physical_memory_write_rom_internal() to
address_space_write_rom_internal(), and make it take
MemTxAttrs and return a MemTxResult. This brings its
API into line with address_space_write().

This is an internal function to exec.c; fixing its API
will allow us to change the global function
cpu_physical_memory_write_rom().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 20181122133507.30950-2-peter.maydell@linaro.org

scripts/checkpatch.pl: Enforce multiline comment syntax

We now require Linux-kernel-style multiline comments:
    /*
     * line one
     * line two
     */

Enforce this in checkpatch.pl, by backporting the relevant
parts of the Linux kernel's checkpatch.pl. (The only changes
needed are that Linux's checkpatch.pl WARN() function takes
an extra argument that ours does not, and the kernel has a
special case for networking code we don't want.)"

The kernel's checkpatch does not enforce "leading /* on
a line of its own, so that part is unique to QEMU's checkpatch.

Sample warning output:

WARNING: Block comments use a leading /* on a separate line
#34: FILE: hw/intc/arm_gicv3_common.c:39:
+    /* Older versions of QEMU had a bug in the handling of state save/restore

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>

block/mirror: add missing coroutine_fn annotations

Marking a function coroutine_fn currently has no effect on the compiler,
but it documents that this function must be called from coroutine
context and it may yield.  This is important information for the
programmer.

Also, if we ever transition to a stackless coroutine implementation,
then it's likely that the annotation will become mandatory so the
compiler can use the correct calling convention for coroutine functions.

Cc: Max Reitz <mreitz@redhat.com>
Cc: John Snow <jsnow@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

iotests: make 235 work on s390 (and others)

"-machine pc" will not work all architectures. Lets fall back to the
default machine by not specifying it.

In addition we also need to specify -no-shutdown on s390 as qemu will
exit otherwise.

Cc: qemu-stable@nongnu.org
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Assert that flags are up-to-date in bdrv_reopen_prepare()

Towards the end of bdrv_reopen_queue_child(), before starting to
process the children, the update_flags_from_options() function is
called in order to have BDRVReopenState.flags in sync with the options
from the QDict.

This is necessary because during the reopen process flags must be
updated for all nodes in the queue so bdrv_is_writable_after_reopen()
and the permission checks work correctly.

Because of that, calling update_flags_from_options() again in
bdrv_reopen_prepare() doesn't really change the flags (they are
already up-to-date). But we need to call it in order to remove those
options from QemuOpts and that way indicate that they have been
processed.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Remove assertions from update_flags_from_options()

This function takes four options (cache.direct, cache.no-flush,
read-only and auto-read-only) from a QemuOpts object and updates the
flags accordingly.

If any of those options is not set (because it was missing from the
original QDict or because it had an invalid value) then the function
aborts with a failed assertion:

   $ qemu-io -c 'reopen -o read-only=foo' hd.qcow2
   block.c:1126: update_flags_from_options: Assertion `qemu_opt_find(opts, BDRV_OPT_CACHE_DIRECT)' failed.
   Aborted

This assertion is unnecessary, and it forces any caller of
bdrv_reopen() to pass all the aforementioned four options. This may
have made sense in order to remove ambiguity when bdrv_reopen() was
taking both flags and options, but that's not the case anymore.

It's also unnecessary if we want to validate the option values,
because bdrv_reopen_prepare() already takes care of that, as we can
see if we remove the assertions:

   $ qemu-io -c 'reopen -o read-only=foo' hd.qcow2
   Parameter 'read-only' expects 'on' or 'off'

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Stop passing flags to bdrv_reopen_queue_child()

Now that all callers are passing the new options using the QDict we no
longer need the 'flags' parameter.

This patch makes the following changes:

   1) The update_options_from_flags() call is no longer necessary
      so it can be removed.

   2) The update_flags_from_options() call is now used in all cases,
      and is moved down a few lines so it happens after the options
      QDict contains the final set of values.

   3) The flags parameter is removed. Now the flags are initialized
      using the current value (for the top-level node) or the parent
      flags (after inherit_options()). In both cases the initial
      values are updated to reflect the new options in the QDict. This
      happens in bdrv_reopen_queue_child() (as explained above) and in
      bdrv_reopen_prepare().

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Remove flags parameter from bdrv_reopen_queue()

Now that all callers are passing all flag changes as QDict options,
the flags parameter is no longer necessary, so we can get rid of it.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Clean up reopen_backing_file() in block/replication.c

This function is used to put the hidden and secondary disks in
read-write mode before launching the backup job, and back in read-only
mode afterwards.

This patch does the following changes:

  - Use an options QDict with the "read-only" option instead of
    passing the changes as flags only.

  - Simplify the code (it was unnecessarily complicated and verbose).

  - Fix a bug due to which the secondary disk was not being put back
    in read-only mode when writable=false (because in this case
    orig_secondary_flags always had the BDRV_O_RDWR flag set).

  - Stop clearing the BDRV_O_INACTIVE flag.

The flags parameter to bdrv_reopen_queue() becomes redundant and we'll
be able to get rid of it in a subsequent patch.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-io: Put flag changes in the options QDict in reopen_f()

When reopen_f() puts a block device in the reopen queue, some of the
new options are passed using a QDict, but others ("read-only" and the
cache options) are passed as flags.

This patch puts those flags in the QDict. This way the flags parameter
becomes redundant and we'll be able to get rid of it in a subsequent
patch.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Drop bdrv_reopen()

No one is using this function anymore, so we can safely remove it.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Use bdrv_reopen_set_read_only() in the mirror driver

The 'block-commit' QMP command is implemented internally using two
different drivers. If the source image is the active layer then the
mirror driver is used (commit_active_start()), otherwise the commit
driver is used (commit_start()).

In both cases the destination image must be put temporarily in
read-write mode. This is done correctly in the latter case, but what
commit_active_start() does is copy all flags instead.

This patch replaces the bdrv_reopen() calls in that function with
bdrv_reopen_set_read_only() so that only the read-only status is
changed.

A similar change is made in mirror_exit(), which is also used by the
'drive-mirror' and 'blockdev-mirror' commands.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Use bdrv_reopen_set_read_only() in external_snapshot_commit()

This patch replaces the bdrv_reopen() call that set and remove the
BDRV_O_RDWR flag with the new bdrv_reopen_set_read_only() function.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Use bdrv_reopen_set_read_only() in qmp_change_backing_file()

This patch replaces the bdrv_reopen() calls that set and remove the
BDRV_O_RDWR flag with the new bdrv_reopen_set_read_only() function.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Use bdrv_reopen_set_read_only() in stream_start/complete()

This patch replaces the bdrv_reopen() calls that set and remove the
BDRV_O_RDWR flag with the new bdrv_reopen_set_read_only() function.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Use bdrv_reopen_set_read_only() in bdrv_commit()

This patch replaces the bdrv_reopen() calls that set and remove the
BDRV_O_RDWR flag with the new bdrv_reopen_set_read_only() function.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Use bdrv_reopen_set_read_only() in commit_start/complete()

This patch replaces the bdrv_reopen() calls that set and remove the
BDRV_O_RDWR flag with the new bdrv_reopen_set_read_only() function.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Use bdrv_reopen_set_read_only() in bdrv_backing_update_filename()

This patch replaces the bdrv_reopen() calls that set and remove the
BDRV_O_RDWR flag with the new bdrv_reopen_set_read_only() function.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: Add bdrv_reopen_set_read_only()

Most callers of bdrv_reopen() only use it to switch a BlockDriverState
between read-only and read-write, so this patch adds a new function
that does just that.

We also want to get rid of the flags parameter in the bdrv_reopen()
API, so this function sets the "read-only" option and passes the
original flags (which will then be updated in bdrv_reopen_prepare()).

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Avoid aio_worker() for QEMU_AIO_IOCTL

aio_worker() doesn't add anything interesting, it's only a useless
indirection. Call the handler function directly instead.

As we know that this handler function is only called from coroutine
context and the coroutine stays around until the worker thread finishes,
we can keep RawPosixAIOData on the stack.

This was the last user of aio_worker(), so the function goes away now.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Switch to .bdrv_co_ioctl

No real reason to keep using the callback based mechanism here when the
rest of the file-posix driver is coroutine based. Changing it brings
ioctls more in line with how other request types work.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Remove paio_submit_co()

The function is not used any more, remove it.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Avoid aio_worker() for QEMU_AIO_READ/WRITE

aio_worker() doesn't add anything interesting, it's only a useless
indirection. Call the handler function directly instead.

As we know that this handler function is only called from coroutine
context and the coroutine stays around until the worker thread finishes,
we can keep RawPosixAIOData on the stack.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Move read/write operation logic out of aio_worker()

aio_worker() for reads and writes isn't boring enough yet. It still does
some postprocessing for handling short reads and turning the result into
the right return value.

However, there is no reason why handle_aiocb_rw() couldn't do the same,
and even without duplicating code between the read and write path. So
move the code there.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Avoid aio_worker() for QEMU_AIO_FLUSH

aio_worker() doesn't add anything interesting, it's only a useless
indirection. Call the handler function directly instead.

As we know that this handler function is only called from coroutine
context and the coroutine stays around until the worker thread finishes,
we can keep RawPosixAIOData on the stack.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Avoid aio_worker() for QEMU_AIO_DISCARD

aio_worker() doesn't add anything interesting, it's only a useless
indirection. Call the handler function directly instead.

As we know that this handler function is only called from coroutine
context and the coroutine stays around until the worker thread finishes,
we can keep RawPosixAIOData on the stack.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Avoid aio_worker() for QEMU_AIO_WRITE_ZEROES

aio_worker() doesn't add anything interesting, it's only a useless
indirection. Call the handler function directly instead.

As we know that this handler function is only called from coroutine
context and the coroutine stays around until the worker thread finishes,
we can keep RawPosixAIOData on the stack.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Avoid aio_worker() for QEMU_AIO_COPY_RANGE

aio_worker() doesn't add anything interesting, it's only a useless
indirection. Call the handler function directly instead.

As we know that this handler function is only called from coroutine
context and the coroutine stays around until the worker thread finishes,
we can keep RawPosixAIOData on the stack.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Avoid aio_worker() for QEMU_AIO_TRUNCATE

aio_worker() doesn't add anything interesting, it's only a useless
indirection. Call the handler function directly instead.

As we know that this handler function is only called from coroutine
context and the coroutine stays around until the worker thread finishes,
we can keep RawPosixAIOData on the stack.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Factor out raw_thread_pool_submit()

Getting the thread pool of the AioContext of a block node and scheduling
some work in it is an operation that is already done twice, and we'll
get more instances. Factor it out into a separate function.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

file-posix: Reorganise RawPosixAIOData

RawPosixAIOData contains a lot of fields for several separate operations
that are to be processed in a worker thread and that need different
parameters. The struct is currently rather unorganised, with unions that
cover some, but not all operations, and even one #define for field names
instead of a union.

Clean this up to have some common fields and a single union. As a side
effect, on x86_64 the struct shrinks from 72 to 48 bytes.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qcow2: do decompression in threads

Do decompression in threads, like it is already done for compression.
This improves asynchronous compressed reads performance.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qcow2: aio support for compressed cluster read

Allocate buffers locally and release qcow2 lock. Than, reads inside
qcow2_co_preadv_compressed may be done in parallel, however all
decompression is still done synchronously. Let's improve it in the
following commit.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qcow2: use byte-based read in qcow2_decompress_cluster

We are gradually moving away from sector-based interfaces, towards
byte-based.  Get rid of it here too.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>