pc-bios/s390-ccw: Move string arrays from bootmap header to .c file

bootmap.h can currently only be included once - otherwise the linker
complains about multiple definitions of the "magic" strings. It's a
bad style to define string arrays in header files, so let's better
move these to the bootmap.c file instead where they are used.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1520317081-5341-1-git-send-email-thuth@redhat.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390x/sclp: clean up sclp masks

Introduce an sccb_mask_t to be used for SCLP event masks instead of just
unsigned int or uint32_t. This will allow later to extend the mask with
more ease.

Signed-off-by: Claudio Imbrenda <imbrenda@linux.vnet.ibm.com>
Message-Id: <1519407778-23095-3-git-send-email-imbrenda@linux.vnet.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390x/sclp: proper support of larger send and receive masks

Until 67915de9f0383ccf4a ("s390x/event-facility: variable-length event masks")
we only supported sclp event masks with a size of exactly 4 bytes, even
though the architecture allows the guests to set up sclp event masks
from 1 to 1021 bytes in length.
After that patch, the behaviour was almost compliant, but some issues
were still remaining, in particular regarding the handling of selective
reads and migration.

When setting the sclp event mask, a mask size is also specified. Until
now we only considered the size in order to decide which bits to save
in the internal state. On the other hand, when a guest performs a
selective read, it sends a mask, but it does not specify a size; the
implied size is the size of the last mask that has been set.

Specifying bits in the mask of selective read that are not available in
the internal mask should return an error, and bits past the end of the
mask should obviously be ignored. This can only be achieved by keeping
track of the lenght of the mask.

The mask length is thus now part of the internal state that needs to be
migrated.

This patch fixes the handling of selective reads, whose size will now
match the length of the event mask, as per architecture.

While the default behaviour is to be compliant with the architecture,
when using older machine models the old broken behaviour is selected
(allowing only masks of size exactly 4), in order to be able to migrate
toward older versions.

Fixes: 67915de9f0383ccf4a ("s390x/event-facility: variable-length event masks")
Signed-off-by: Claudio Imbrenda <imbrenda@linux.vnet.ibm.com>
Message-Id: <1519407778-23095-2-git-send-email-imbrenda@linux.vnet.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

vfio-ccw: license text should indicate GPL v2 or later

The license text currently specifies "any version" of the GPL. It
is unlikely that GPL v1 was ever intended; change this to the
standard "or any later version" text.

Cc: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Cc: Xiao Feng Ren <renxiaof@linux.vnet.ibm.com>
Cc: Pierre Morel <pmorel@linux.vnet.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Dong Jia Shi <bjsdjshi@linux.vnet.ibm.com>
Acked-by: Pierre Morel <pmorel@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390x/sclpconsole: Remove dead code - remove exit handlers

The other event handlers (quiesce and cpu) do not define these
handlers, and this one does nothing, so it can be removed.

Signed-off-by: Nia Alarie <nia.alarie@gmail.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20180306100721.19419-1-nia.alarie@gmail.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

numa: we don't implement NUMA for s390x

Right now it is possible to crash QEMU for s390x by providing e.g.
    -numa node,nodeid=0,cpus=0-1

Problem is, that numa.c uses mc->cpu_index_to_instance_props as an
indicator whether NUMA is supported by a machine type. We don't
implement NUMA for s390x ("topology") yet. However we need
mc->cpu_index_to_instance_props for query-cpus.

So let's fix this case by also checking for mc->get_default_cpu_node_id,
which will be needed by machine_set_cpu_numa_node().

qemu-system-s390x: -numa node,nodeid=0,cpus=0-1: NUMA is not supported by
                   this machine-type

While at it, make s390_cpu_index_to_props() look like on other
architectures.

Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180227110255.20999-1-david@redhat.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

hw/s390x: Add the possibility to specify the netboot image on the command line

The file name of the netboot binary is currently hard-coded to
"s390-netboot.img", without a possibility for the user to select
an alternative firmware image here. That's unfortunate, especially
since the basics are already there: The filename is a property of
the s390-ipl device. So we just have to add a check whether the user
already provided the property and only set the default if the string
is still empty. Now it is possible to select a different firmware
image with "-global s390-ipl.netboot_fw=/path/to/s390-netboot.img".

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1519731154-3127-1-git-send-email-thuth@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

target/s390x: Remove leading underscores from #defines

We should not use leading underscores followed by a capital letter
in #defines since such identifiers are reserved by the C standard.

For ASCE_ORIGIN, REGION_ENTRY_ORIGIN and SEGMENT_ENTRY_ORIGIN I also
added parentheses around the value to silence an error message from
checkpatch.pl.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1520227018-4061-1-git-send-email-thuth@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

s390/ipl: only print boot menu error if -boot menu=on was specified

It is possible that certain QEMU configurations may not
create an IPLB (such as when -kernel is provided). In
this case, a misleading error message will be printed
stating that the "boot menu is not supported for this
device type".

To amend this, only print this message iff boot menu=on
was provided on the commandline. Otherwise, return silently.

While we're at it, remove trailing periods from error
messages.

Signed-off-by: Collin L. Walling <walling@linux.vnet.ibm.com>
Message-Id: <1519760121-24594-1-git-send-email-walling@linux.vnet.ibm.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

hw/s390x/ipl: Bail out if the network bootloader can not be found

If QEMU fails to load 's390-netboot.img', the guest firmware currently
loops forever and just floods the console with "Network boot device
detected" messages. The code in ipl.c apparently already tried to stop
the VM with vm_stop() in this case, but this is in vain since the run
state is later reset due to a call to vm_start() from vl.c again.
To avoid the ugly firmware loop, let's simply exit QEMU directly instead
since it just does not make sense to continue if the required firmware
image can not be loaded. While we're at it, also add the file name of
the netboot binary to the error message, so that the user has a better
hint about what is missing.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1519725913-24852-1-git-send-email-thuth@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Farhan Ali <alifm@linux.vnet.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>

Merge remote-tracking branch 'remotes/mcayland/tags/qemu-sparc-signed' into staging

qemu-sparc update

# gpg: Signature made Thu 08 Mar 2018 07:23:01 GMT
# gpg:                using RSA key 5BC2C56FAE0F321F
# gpg: Good signature from "Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>"
# Primary key fingerprint: CC62 1AB9 8E82 200D 915C  C9C4 5BC2 C56F AE0F 321F

* remotes/mcayland/tags/qemu-sparc-signed:
  sparc: fix leon3 casa instruction when MMU is disabled
  hw/sparc/sun4m: Fix implicit creation of "-drive if=scsi" devices

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

Merge remote-tracking branch 'remotes/stefanberger/tags/pull-tpm-2018-03-07-1' into staging

Merge tpm 2018/03/07

# gpg: Signature made Wed 07 Mar 2018 12:42:13 GMT
# gpg:                using RSA key 75AD65802A0B4211
# gpg: Good signature from "Stefan Berger <stefanb@linux.vnet.ibm.com>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: B818 B9CA DF90 89C2 D5CE  C66B 75AD 6580 2A0B 4211

* remotes/stefanberger/tags/pull-tpm-2018-03-07-1:
  tpm: convert tpm_tis.c to use trace-events
  tpm: convert tpm_emulator.c to use trace-events
  tpm: convert tpm_util.c to use trace-events
  tpm: convert tpm_passthrough.c to use trace-events
  tpm: convert tpm_crb.c to use trace-events

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

Merge remote-tracking branch 'remotes/berrange/tags/qio-next-pull-request' into staging

# gpg: Signature made Wed 07 Mar 2018 11:24:41 GMT
# gpg:                using RSA key BE86EBB415104FDF
# gpg: Good signature from "Daniel P. Berrange <dan@berrange.com>"
# gpg:                 aka "Daniel P. Berrange <berrange@redhat.com>"
# Primary key fingerprint: DAF3 A6FD B26B 6291 2D0E  8E3F BE86 EBB4 1510 4FDF

* remotes/berrange/tags/qio-next-pull-request:
  qio: non-default context for TLS handshake
  qio: non-default context for async conn
  qio: non-default context for threaded qtask
  qio: store gsources for net listeners
  qio: introduce qio_channel_add_watch_{full|source}
  qio: rename qio_task_thread_result

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

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

Multiboot patches

# gpg: Signature made Wed 07 Mar 2018 11:15:17 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:
  multiboot: fprintf(stderr...) -> error_report()
  multiboot: Use header names when displaying fields
  multiboot: Remove unused variables from multiboot.c
  multiboot: bss_end_addr can be zero

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

sparc: fix leon3 casa instruction when MMU is disabled

Since the commit af7a06bac7d3abb2da48ef3277d2a415772d2ae8:
`casa [..](10), .., ..` (and probably others alternate space instructions)
triggers a data access exception when the MMU is disabled.

When we enter get_asi(...) dc->mem_idx is set to MMU_PHYS_IDX when the MMU
is disabled. Just keep mem_idx unchanged in this case so we passthrough the
MMU when it is disabled.

Signed-off-by: KONRAD Frederic <frederic.konrad@adacore.com>
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>

hw/sparc/sun4m: Fix implicit creation of "-drive if=scsi" devices

The global hack for creating SCSI devices has recently been removed,
but this apparently broke SCSI devices on some boards that were not
ready for this change yet. For the sun4m machines you now get:

$ sparc-softmmu/qemu-system-sparc -boot d -cdrom x.iso
qemu-system-sparc: -cdrom x.iso: machine type does not support if=scsi,bus=0,unit=2

Fix it by calling scsi_bus_legacy_handle_cmdline() after creating the
corresponding SCSI controller.

Reported-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Fixes: 1454509726719e0933c800fad00d6999752688ea
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>

multiboot: fprintf(stderr...) -> error_report()

Change all fprintf(stderr...) calls in hw/i386/multiboot.c to call
error_report() instead, including the mb_debug macro.  Remove the "\n"
from strings passed to all modified calls, since error_report() appends
one.

Signed-off-by: Jack Schwartz <jack.schwartz@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

multiboot: Use header names when displaying fields

Refer to field names when displaying fields in printf and debug statements.

Signed-off-by: Jack Schwartz <jack.schwartz@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

multiboot: Remove unused variables from multiboot.c

Remove unused variables: mh_mode_type, mh_width, mh_height, mh_depth

Signed-off-by: Jack Schwartz <jack.schwartz@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

multiboot: bss_end_addr can be zero

The multiboot spec (https://www.gnu.org/software/grub/manual/multiboot/),
section 3.1.3, allows for bss_end_addr to be zero.

A zero bss_end_addr signifies there is no .bss section.

Suggested-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Jack Schwartz <jack.schwartz@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

tpm: convert tpm_tis.c to use trace-events

Leave the DEBUG_TIS for more debugging and convert to use if (DEBUG_TIS)
rather than #if DEBUG_TIS where it is being used.

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

tpm: convert tpm_emulator.c to use trace-events

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

tpm: convert tpm_util.c to use trace-events

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

tpm: convert tpm_passthrough.c to use trace-events

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

tpm: convert tpm_crb.c to use trace-events

Signed-off-by: Stefan Berger <stefanb@linux.vnet.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

misc: don't use hwaddr as a type in trace events

Use types that are defined by QEMU in trace events caused build failures
for the UST trace backend:

  In file included from trace-ust-all.c:13:0:
  trace-ust-all.h:11844:206: error: unknown type name ‘hwaddr’

It only knows about C built-in types, and any types that are pulled in
from includs of qemu-common.h and lttng/tracepoint.h. This does not
include the 'hwaddr' type, so replace it with a uint64_t which is what
exec/hwaddr.h defines 'hwaddr' as. This fixes the build failure
introduced by

  commit 9eb8040c2d2b38e1a40bb6129b1b668fa178fcab
  Author: Peter Maydell <peter.maydell@linaro.org>
  Date:   Fri Mar 2 10:45:39 2018 +0000

    hw/misc/tz-ppc: Model TrustZone peripheral protection controller

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20180306134317.836-1-berrange@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

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

* new QMP command qom-list-properties (Alexey)
* TCG cleanups (David)
* use g_path_get_basename/g_path_get_dirname when useful (Julia)
* WHPX fixes (Justin)
* ASAN fixes (Marc-André)
* g364fb memory leak fix, address_space_to_flatview RCU fixes (me)
* chardev memory leak fix (Peter)
* checkpatch improvements (Julia, Su Hang)
* next round of deprecation patches (Thomas)

# gpg: Signature made Tue 06 Mar 2018 13:11:58 GMT
# gpg:                using RSA key BFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>"
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini/tags/for-upstream: (34 commits)
  use g_path_get_basename instead of basename
  balloon: Fix documentation of the --balloon parameter and deprecate it
  WHPX improve interrupt notification registration
  WHXP Removes the use of WHvGetExitContextSize
  Fix WHPX issue leaking tpr values
  Fix WHPX typo in 'mmio'
  Fix WHPX additional lock acquisition
  Remove unnecessary WHPX __debugbreak();
  Resolves WHPX breaking changes in SDK 17095
  Fixing WHPX casing to match SDK
  Revert "build-sys: compile with -Og or -O1 when --enable-debug"
  checkpatch: add check for `while` and `for`
  checkpatch: add a warning for basename/dirname
  address_space_rw: address_space_to_flatview needs RCU lock
  address_space_map: address_space_to_flatview needs RCU lock
  address_space_access_valid: address_space_to_flatview needs RCU lock
  address_space_read: address_space_to_flatview needs RCU lock
  address_space_write: address_space_to_flatview needs RCU lock
  memory: inline some performance-sensitive accessors
  openpic_kvm: drop address_space_to_flatview call
  ...

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

use g_path_get_basename instead of basename

basename(3) and dirname(3) modify their argument and may return
pointers to statically allocated memory which may be overwritten by
subsequent calls.
g_path_get_basename and g_path_get_dirname have no such issues, and
therefore more preferable.

Signed-off-by: Julia Suvorova <jusual@mail.ru>
Message-Id: <1519888086-4207-1-git-send-email-jusual@mail.ru>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

balloon: Fix documentation of the --balloon parameter and deprecate it

There are two issues with the documentation of the --balloon parameter:
First, "--balloon none" is simply doing nothing. Even if a machine had a
balloon device by default, this option is not disabling anything, it is
simply ignored. Thus let's simply drop this option from the documentation
to avoid to confuse the users (but keep the code in vl.c for backward
compatibility).
Second, the documentation claims that "--balloon virtio" is the default
mode, but this is not true anymore since commit 382f074371f7dc32a34.
Since that commit, the option also has no real use case anymore, since
you can simply use "--device virtio-balloon" nowadays instead. Thus to
simplify our complex parameter zoo a little bit, let's deprecate the
the parameter now and tell the user to use "--device virtio-balloon"
instead.

Fixes: 382f074371f7dc32a34c944c845b1698e83d8c36
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1519796303-13257-1-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

WHPX improve interrupt notification registration

Improves the usage of the InterruptNotification registration by skipping the
additional call to WHvSetVirtualProcessorRegisters if we have already
registered for the window exit.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-9-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

WHXP Removes the use of WHvGetExitContextSize

The use of WHvGetExitContextSize will break ABI compatibility if the platform
changes the context size while a qemu compiled executable does not recompile.
To avoid this we now use sizeof and let the platform determine which version
of the struction was passed for ABI compatibility.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-8-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

Fix WHPX issue leaking tpr values

Fixes an issue where if the tpr is assigned to the array but not a different
value from what is already expected on the vp the code will skip incrementing
the reg_count. In this case its possible that we set an invalid memory section
of the next call for DeliverabilityNotifications that was not expected.

The fix is to use a local variable to store the temporary tpr and only update
the array if the local tpr value is different than the vp context.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-7-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

Fix WHPX typo in 'mmio'

Renames the usage of 'memio' to 'mmio' in the emulator callbacks.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-6-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

Fix WHPX additional lock acquisition

The code already is holding the qemu_mutex for the IO thread. We do not need
to additionally take the lock again in this case.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-5-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

Remove unnecessary WHPX __debugbreak();

Minor code cleanup. The calls to __debugbreak() are not required and should
no longer be used to prevent unnecessary breaks.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-4-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

Resolves WHPX breaking changes in SDK 17095

1. Fixes the changes required to the WHvTryMmioEmulation, WHvTryIoEmulation, and
WHvEmulatorCreateEmulator based on the new VpContext forwarding.
2. Removes the WHvRunVpExitReasonAlerted case.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-3-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

Fixing WHPX casing to match SDK

Fixes an issue where the SDK that was releases had a different casing for the
*.h and *.lib files causing a build break if linked directly from Windows Kits.

Signed-off-by: Justin Terry (VM) <juterry@microsoft.com>
Message-Id: <1519665216-1078-2-git-send-email-juterry@microsoft.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Justin Terry (VM) via Qemu-devel <qemu-devel@nongnu.org>

Revert "build-sys: compile with -Og or -O1 when --enable-debug"

This reverts commit 906548689e37ab6cca1e93b3f8d9327a4e17e8af.
Even with -Og, the debug experience is noticeably worse
because gdb shows a lot more "<optimised out>" variables and
function arguments.

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

checkpatch: add check for `while` and `for`

Adding check for `while` and `for` statements, which condition has more than
one line.

The former checkpatch.pl can check `if` statement, which condition has more
than one line, whether block misses brace round, like this:
'''
if (cond1 ||
    cond2)
    statement;
'''
But it doesn't do the same check for `for` and `while` statements.

Using `(?:...)` instead of `(...)` in regex pattern catch.
Because `(?:...)` is faster and avoids unwanted side-effect.

Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
Suggested-by: Eric Blake <eblake@redhat.com>
Suggested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Su Hang <suhang16@mails.ucas.ac.cn>
Message-Id: <1520319890-19761-1-git-send-email-suhang16@mails.ucas.ac.cn>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

checkpatch: add a warning for basename/dirname

g_path_get_* do the same as g_strdup(basename/dirname(...)) but
without modifying the argument.

Signed-off-by: Julia Suvorova <jusual@mail.ru>
Message-Id: <1519987399-19160-1-git-send-email-jusual@mail.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

address_space_rw: address_space_to_flatview needs RCU lock

address_space_rw is calling address_space_to_flatview but it can
be called outside the RCU lock.  To fix it, transform flatview_rw
into address_space_rw, since flatview_rw is otherwise unused.

Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

address_space_map: address_space_to_flatview needs RCU lock

address_space_map is calling address_space_to_flatview but it can
be called outside the RCU lock.  The function itself is calling
rcu_read_lock/rcu_read_unlock, just in the wrong place, so the
fix is easy.

Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

address_space_access_valid: address_space_to_flatview needs RCU lock

address_space_access_valid is calling address_space_to_flatview but it can
be called outside the RCU lock.  To fix it, push the rcu_read_lock/unlock
pair up from flatview_access_valid to address_space_access_valid.

Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

address_space_read: address_space_to_flatview needs RCU lock

address_space_read is calling address_space_to_flatview but it can
be called outside the RCU lock.  To fix it, push the rcu_read_lock/unlock
pair up from flatview_read_full to address_space_read's constant size
fast path and address_space_read_full.

Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

address_space_write: address_space_to_flatview needs RCU lock

address_space_write is calling address_space_to_flatview but it can
be called outside the RCU lock.  To fix it, push the rcu_read_lock/unlock
pair up from flatview_write to address_space_write.

Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

memory: inline some performance-sensitive accessors

These accessors are called from inlined functions, and the call sequence
is much more expensive than just inlining the access.  Move the
struct declaration to memory-internal.h so that exec.c and memory.c
can both use an inline function.

Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

openpic_kvm: drop address_space_to_flatview call

The MemoryListener is registered on address_space_memory, there is
not much to assert.  This currently works because the callback
is invoked only once when the listener is registered, but section->fv
is the _new_ FlatView, not the old one on later calls and that
would break.

This confines address_space_to_flatview to exec.c and memory.c.

Acked-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

chardev: fix leak in tcp_chr_telnet_init_io()

Need to free TCPChardevTelnetInit when session established.

Since at it, switch to use G_SOURCE_* macros.

Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180301084438.13594-2-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

sdhci-test: fix leaks

Fix the following ASAN reports:

==20125==ERROR: LeakSanitizer: detected memory leaks

Direct leak of 24 byte(s) in 1 object(s) allocated from:
    #0 0x7f0faea03a38 in __interceptor_calloc (/lib64/libasan.so.4+0xdea38)
    #1 0x7f0fae450f75 in g_malloc0 ../glib/gmem.c:124
    #2 0x562fffd526fc in machine_start /home/elmarco/src/qemu/tests/sdhci-test.c:180

Indirect leak of 152 byte(s) in 1 object(s) allocated from:
    #0 0x7f0faea03850 in malloc (/lib64/libasan.so.4+0xde850)
    #1 0x7f0fae450f0c in g_malloc ../glib/gmem.c:94
    #2 0x562fffd5d21d in qpci_init_pc /home/elmarco/src/qemu/tests/libqos/pci-pc.c:122

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180215212552.26997-7-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

ahci-test: fix opts leak of skip tests

Fixes the following ASAN report:

Direct leak of 128 byte(s) in 8 object(s) allocated from:
    #0 0x7fefce311850 in malloc (/lib64/libasan.so.4+0xde850)
    #1 0x7fefcdd5ef0c in g_malloc ../glib/gmem.c:94
    #2 0x559b976faff0 in create_ahci_io_test /home/elmarco/src/qemu/tests/ahci-test.c:1810

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180215212552.26997-6-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

lockable: workaround GCC link issue with ASAN

Current GCC has an optimization bug when compiling with ASAN.

See also GCC bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84307

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180215212552.26997-3-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

build-sys: fix -fsanitize=address check

Since 218bb57dd79d6843e0592c30a82ea8c1fddc74a5, the -fsanitize=address
check fails with:
config-temp/qemu-conf.c:3:20: error: integer overflow in expression [-Werror=overflow]
   return INT32_MIN / -1;

Interestingly, UBSAN check doesn't produce a compile time warning.
Use a test that doesn't have compile time warnings, and make it
specific to UBSAN check.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180215212552.26997-2-marcandre.lureau@redhat.com>
Reviewed-by: Emilio G. Cota <cota@braap.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qmp: Add qom-list-properties to list QOM object properties

There is already 'device-list-properties' which does most of the job,
however it does not handle everything returned by qom-list-types such
as machines as they inherit directly from TYPE_OBJECT and not TYPE_DEVICE.
It does not handle abstract classes either.

This adds a new qom-list-properties command which prints properties
of a specific class and its instance. It is pretty much a simplified copy
of the device-list-properties handler.

Since it creates an object instance, device properties should appear
in the output as they are copied to QOM properties at the instance_init
hook.

This adds a object_class_property_iter_init() helper to allow class
properties enumeration uses it in the new QMP command to allow properties
listing for abstract classes.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Message-Id: <20180301130939.15875-3-aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qmp: Merge ObjectPropertyInfo and DevicePropertyInfo

ObjectPropertyInfo is more generic and only missing @description.
This adds a description to ObjectPropertyInfo and removes
DevicePropertyInfo so the resulting ObjectPropertyInfo can be used
elsewhere.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Message-Id: <20180301130939.15875-2-aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Document --rtc-td-hack, --localtime and --startdate as deprecated

These options have been marked in a comment in qemu-options.hx as
deprecated in 2009 already (see commit 1ed2fc1fa35fadc0d6), but we
never informed the users about these deprecations. Let's catch up
on that omission now.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1519138892-12836-1-git-send-email-thuth@redhat.com>
[Fix messages. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

cpus: CPU threads are always created initially for one CPU only

It can never happen for single-threaded TCG that we have more than one
CPU in the list, while the first one has not been marked as "created".

Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180209195239.16048-4-david@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

cpus: wait for CPU creation at central place

We can now also wait for the CPU creation for single-threaded TCG, so we
can move the waiting bits further out.

Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180209195239.16048-3-david@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

cpus: properly inititalize CPU > 1 under single-threaded TCG

All but the first CPU are currently not fully inititalized (e.g.
cpu->created is never set).

Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20180209195239.16048-2-david@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

scsi: Remove automatic creation of SCSI controllers with -drive if=scsi

Automatic creation of SCSI controllers for "-drive if=scsi" for x86
machines was quite a bad idea (see description of commit f778a82f0c179
for details). This is marked as deprecated since QEMU v2.9.0, and as
far as I know, nobody complained that this is still urgently required
anymore. Time to remove this now.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1519123357-13225-1-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Remove the deprecated -tdf option

It's been marked as deprecated since a very long time already, and
the parameter is not doing anything useful anymore except for printing
a warning, so it's now time to finally get rid of this option.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1519071820-4062-1-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

g364fb: fix DirtyBitmapSnapshot leak

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

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

Block layer patches

# gpg: Signature made Mon 05 Mar 2018 17:45:51 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: (38 commits)
  block: Fix NULL dereference on empty drive error
  qcow2: Replace align_offset() with ROUND_UP()
  block/ssh: Add basic .bdrv_truncate()
  block/ssh: Make ssh_grow_file() blocking
  block/ssh: Pull ssh_grow_file() from ssh_create()
  qemu-img: Make resize error message more general
  qcow2: make qcow2_co_create2() a coroutine_fn
  block: rename .bdrv_create() to .bdrv_co_create_opts()
  Revert "IDE: Do not flush empty CDROM drives"
  block: test blk_aio_flush() with blk->root == NULL
  block: add BlockBackend->in_flight counter
  block: extract AIO_WAIT_WHILE() from BlockDriverState
  aio: rename aio_context_in_iothread() to in_aio_context_home_thread()
  docs: document how to use the l2-cache-entry-size parameter
  specs/qcow2: Fix documentation of the compressed cluster descriptor
  iotest 033: add misaligned write-zeroes test via truncate
  block: fix write with zero flag set and iovector provided
  block: Drop unused .bdrv_co_get_block_status()
  vvfat: Switch to .bdrv_co_block_status()
  vpc: Switch to .bdrv_co_block_status()
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
# Conflicts:
# include/block/block.h

Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.12-20180306' into staging

ppc patch queue 2018-03-06

This pull request supersedes ppc-for-2.12-20180302 which had compile
problems with some gcc versions.  It also contains a few additional
patches.

Highlights are:
    * New Sam460ex machine type
    * Yet more fixes related to vcpu id allocation for spapr
    * Numerous macio cleanupsr
    * Some enhancements to the Spectre/Meltdown fixes for pseries,
      allowing use of a better mitigation for indirect branch based
      exploits
    * New pseries machine types with Spectre/Meltdown mitigations
      enabled (stop gap until libvirt and management understands the
      machine options)
    * A handful of other fixes

# gpg: Signature made Tue 06 Mar 2018 04:01:00 GMT
# gpg:                using RSA key 6C38CACA20D9B392
# gpg: Good signature from "David Gibson <david@gibson.dropbear.id.au>"
# gpg:                 aka "David Gibson (Red Hat) <dgibson@redhat.com>"
# gpg:                 aka "David Gibson (ozlabs.org) <dgibson@ozlabs.org>"
# gpg:                 aka "David Gibson (kernel.org) <dwg@kernel.org>"
# Primary key fingerprint: 75F4 6586 AE61 A66C C44E  87DC 6C38 CACA 20D9 B392

* remotes/dgibson/tags/ppc-for-2.12-20180306: (30 commits)
  PowerPC: Add TS bits into msr_mask
  adb: add trace-events for monitoring keyboard/mouse during bus enumeration
  PPC: e500: Fix duplicate kernel load and device tree overlap
  hw/ppc/spapr,e500: Use new property "stdout-path" for boot console
  ppc/spapr-caps: Define the pseries-2.12-sxxm machine type
  ppc/spapr-caps: Convert cap-ibs to custom spapr-cap
  ppc/spapr-caps: Convert cap-sbbc to custom spapr-cap
  ppc/spapr-caps: Convert cap-cfpc to custom spapr-cap
  ppc/spapr-caps: Add support for custom spapr_capabilities
  target/ppc: Check mask when setting cap_ppc_safe_indirect_branch
  macio: remove macio_init() function
  macio: move setting of CUDA timebase frequency to macio_common_realize()
  mac_newworld: use object link to pass OpenPIC object to macio
  openpic: move OpenPIC state and related definitions to openpic.h
  openpic: move KVM-specific declarations into separate openpic_kvm.h file
  mac_oldworld: use object link to pass heathrow PIC object to macio
  macio: move macio related structures and defines into separate macio.h file
  heathrow: change heathrow_pic_init() to return the heathrow device
  heathrow: convert to trace-events
  heathrow: QOMify heathrow PIC
  ...

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

qio: non-default context for TLS handshake

A new parameter "context" is added to qio_channel_tls_handshake() is to
allow the TLS to be run on a non-default context.  Still, no functional
change.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qio: non-default context for async conn

We have worked on qio_task_run_in_thread() already.  Further, let
all the qio channel APIs use that context.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qio: non-default context for threaded qtask

qio_task_run_in_thread() allows main thread to run blocking operations
in the background. However it has an assumption on that it's always
working with the default context. This patch tries to allow the threaded
QIO task framework to run with non-default gcontext.

Currently no functional change so far, so the QIOTasks are still always
running on main context.

Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qio: store gsources for net listeners

Originally we were storing the GSources tag IDs.  That'll be not enough
if we are going to support non-default gcontext for QIO code.  Switch to
GSources without changing anything real.  Now we still always pass in
NULL, which means the default gcontext.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qio: introduce qio_channel_add_watch_{full|source}

Firstly, introduce an internal qio_channel_add_watch_full(), which
enhances qio_channel_add_watch() that context can be specified.

Then add a new API wrapper qio_channel_add_watch_source() to return a
GSource pointer rather than a tag ID.

Note that the _source() call will keep a reference of GSource so that
callers need to unref them explicitly when finished using the GSource.

Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

qio: rename qio_task_thread_result

It is strange that it was called gio_task_thread_result.  Rename it to
follow the naming rule of the file.

Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

PowerPC: Add TS bits into msr_mask

During migration, after MSR bits is synced, cpu_post_load() will use
msr_mask to determine which PPC MSR bits will be applied into the target
side. Hardware Transaction Memory(HTM) has been supported since Power8,
but TS0/TS1 bit was not in msr_mask yet. That will prevent target KVM
from loading TM checkpointed values.

This patch adds TS bits into msr_mask for Power8, so that transactional
application can be migrated across qemu.

Signed-off-by: Simon Guo <wei.guo.simon@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

adb: add trace-events for monitoring keyboard/mouse during bus enumeration

This is useful to help diagnose problems related to address clashes during
MacOS 9 boot.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

PPC: e500: Fix duplicate kernel load and device tree overlap

This patch fixes an incorrect behavior when the -kernel argument has been
specified without -bios. In this case the kernel was loaded twice. At address
32M as a raw image and afterwards by load_elf/load_uimage at the
corresponding load address. In this case the region for the device tree and
the raw kernel image may overlap.

The patch fixes the behavior by loading the kernel image once with
load_elf/load_uimage and skips loading the raw image.

When here do not use bios_name/size for the kernel and use a more generic
name called payload_name/size.

New in v3: dtb must be stored between kernel and initrd because Linux can
           handle the dtb only within the first 64MB. Add a comment to
           clarify the behavior.

Signed-off-by: David Engraf <david.engraf@sysgo.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

hw/ppc/spapr,e500: Use new property "stdout-path" for boot console

Linux kernel commit 2a9d832cc9aae21ea827520fef635b6c49a06c6d
(of: Add bindings for chosen node, stdout-path) deprecated chosen property
"linux,stdout-path" and "stdout".

Introduce the new property "stdout-path" and continue supporting the older
property to remain compatible with existing/older firmware. This older property
can be deprecated after 5 years.

Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc/spapr-caps: Define the pseries-2.12-sxxm machine type

The sxxm (speculative execution exploit mitigation) machine type is a
variant of the 2.12 machine type with workarounds for speculative
execution vulnerabilities enabled by default.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc/spapr-caps: Convert cap-ibs to custom spapr-cap

Convert cap-ibs (indirect branch speculation) to a custom spapr-cap
type.

All tristate caps have now been converted to custom spapr-caps, so
remove the remaining support for them.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Don't explicitly list "?"/help option, trust convention]
[dwg: Fold tristate removal into here, to not break bisect]
[dwg: Fix minor style problems]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc/spapr-caps: Convert cap-sbbc to custom spapr-cap

Convert cap-sbbc (speculation barrier bounds checking) to a custom
spapr-cap type.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Removed trailing whitespace]
[dwg: Don't explicitly list "?"/help option, trust convention]
[dwg: Fix some minor style problems]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc/spapr-caps: Convert cap-cfpc to custom spapr-cap

Convert cap-cfpc (cache flush on privilege change) to a custom spapr-cap
type.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Don't explicitly list "?"/help option, trusting convention]
[dwg: Strip no-longer-necessary ATTRIBUTE_UNUSED back off]
[dwg: Fix some minor style problems]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc/spapr-caps: Add support for custom spapr_capabilities

There are currently 2 implemented types of spapr-caps, boolean and
tristate. However there may be a need for caps which don't fit either of
these options. Add a custom capability type for which a list of custom
valid strings can be specified and implement the get/set functions for
these. Also add a field for help text to describe the available options.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
[dwg: Change "help" option to "?" matching qemu conventions]
[dwg: Add ATTRIBUTE_UNUSED to avoid breaking bisect]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

target/ppc: Check mask when setting cap_ppc_safe_indirect_branch

Check the character and character_mask field when setting
cap_ppc_safe_indirect_branch based on the hypervisor response
to KVM_PPC_GET_CPU_CHAR. Previously the mask field wasn't checked
which was incorrect.

Fixes: 8acc2ae5 (target/ppc/kvm: Add cap_ppc_safe_[cache/bounds_check/indirect_branch])

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

macio: remove macio_init() function

Move the remaining comment into macio.c for reference, then remove the
macio_init() function and instantiate the macio devices for both Old World
and New World machines via qdev_init_nofail() directly.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

macio: move setting of CUDA timebase frequency to macio_common_realize()

This removes the last of the functionality from macio_init() in preparation
for its subsequent removal.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

mac_newworld: use object link to pass OpenPIC object to macio

Also switch macio_newworld_realize() over to use it rather than using the pic_mem
memory region directly.

Now that both Old World and New World macio devices no longer make use of the
pic_mem memory region directly, we can remove it.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

openpic: move OpenPIC state and related definitions to openpic.h

This is to faciliate access to OpenPICState when wiring up the PIC to the macio
controller.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

openpic: move KVM-specific declarations into separate openpic_kvm.h file

This is needed before the next patch because the target-dependent kvm stub
uses the existing kvm_openpic_connect_vcpu() declaration, making it impossible
to move the device-specific declarations into the same file without breaking
ppc-linux-user compilation.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

mac_oldworld: use object link to pass heathrow PIC object to macio

Also switch macio_oldworld_realize() over to use it rather than using the pic_mem
memory region directly.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

macio: move macio related structures and defines into separate macio.h file

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

heathrow: change heathrow_pic_init() to return the heathrow device

This enables the device to be made available during the setup of the Old World
machine. In order to pass back the previous set of IRQs we temporarily introduce
a new pic_irqs parameter until it can be removed.

An additional benefit of this change is that it is also possible to remove the
pic_mem pointer used for macio by accessing the memory region via sysbus.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

heathrow: convert to trace-events

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

heathrow: QOMify heathrow PIC

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

macio: move ESCC device within the macio device

Now that the ESCC device is instantiated directly via qdev, move it to within
the macio device and wire up the IRQs and memory regions using the sysbus API.

This enables to remove the now-obsolete escc_mem parameter to the macio_init()
function.

(Note this patch also contains small touch-ups to the formatting in
macio_escc_legacy_setup() and ppc_heathrow_init() in order to keep checkpatch
happy)

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

macio: embed DBDMA device directly within macio

The current recommendation is to embed subdevices directly within their container
device, so do this for the DBDMA device.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

spapr: harden code that depends on VSMT

VSMT must be set in order to compute VCPU ids. This means that the
following functions must not be called before spapr_set_vsmt_mode()
was called:
- spapr_vcpu_id()
- spapr_is_thread0_in_vcore()
- xics_max_server_number()

We had a recent regression where the latter would be called before VSMT
was set, and broke migration of some old machine types. This patch
adds assert() in the above functions to avoid problems in the future.

Also, since VSMT is really a CPU related thing, spapr_set_vsmt_mode() is
now called from spapr_init_cpus(), just before the first VSMT user.

Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

spapr: register dummy ICPs later

Some older machine types create more ICPs than needed. We hence
need to register up to xics_max_server_number() dummy ICPs to
accomodate the migration of these machine types.

Recent VSMT rework changed xics_max_server_number() to return

    DIV_ROUND_UP(max_cpus * spapr->vsmt, smp_threads)

instead of

    DIV_ROUND_UP(max_cpus * kvmppc_smt_threads(), smp_threads);

The change is okay but it requires spapr->vsmt to be set, which
isn't the case with the current code. This causes the formula to
return zero and we don't create dummy ICPs. This breaks migration
of older guests as reported here:

    https://bugzilla.redhat.com/show_bug.cgi?id=1549087

The dummy ICP workaround doesn't really have a dependency on XICS
itself. But it does depend on proper VCPU id numbering and it must
be applied before creating vCPUs (ie, creating real ICPs). So this
patch moves the workaround to spapr_init_cpus(), which already
assumes VSMT to be set.

Fixes: 72194664c8a1 ("spapr: use spapr->vsmt to compute VCPU ids")
Reported-by: Lukas Doktor <ldoktor@redhat.com>
Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc: Add aCube Sam460ex board

Add emulation of aCube Sam460ex board based on AMCC 460EX embedded SoC.
This is not a complete implementation yet with a lot of components
still missing but enough for the U-Boot firmware to start and to boot
a Linux kernel or AROS.

Signed-off-by: François Revol <revol@free.fr>
Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

pc-bios: Added u-boot-sam460 firmware binary

Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

roms: Added git submodule for u-boot-sam460 (firmware for sam460ex)

Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc440: Add emulation of plb-pcix controller found in some 440 SoCs

This is the PCIX controller found in newer 440 core SoCs e.g. the
AMMC 460EX. The device tree refers to this as plb-pcix compared to
the plb-pci controller in older 440 SoCs.

Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
[dwg: Remove hwaddr from trace-events, that doesn't work with some
 trace backends]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

ppc440_uc: Fix unintialized variable warning with older gcc

Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

spapr: fix missing CPU core nodes in DT when running with TCG

Commit 5d0fb1508e2d "spapr: consolidate the VCPU id numbering logic
in a single place" introduced a helper to detect thread0 of a virtual
core based on its VCPU id. This is used to create CPU core nodes in
the DT, but it is broken in TCG.

$ qemu-system-ppc64 -nographic -accel tcg -machine dumpdtb=dtb.bin \
                    -smp cores=16,maxcpus=16,threads=1
$ dtc -f -O dts dtb.bin | grep POWER8
                PowerPC,POWER8@0 {
                PowerPC,POWER8@8 {

instead of the expected 16 cores that we get with KVM:

$ dtc -f -O dts dtb.bin | grep POWER8
                PowerPC,POWER8@0 {
                PowerPC,POWER8@8 {
                PowerPC,POWER8@10 {
                PowerPC,POWER8@18 {
                PowerPC,POWER8@20 {
                PowerPC,POWER8@28 {
                PowerPC,POWER8@30 {
                PowerPC,POWER8@38 {
                PowerPC,POWER8@40 {
                PowerPC,POWER8@48 {
                PowerPC,POWER8@50 {
                PowerPC,POWER8@58 {
                PowerPC,POWER8@60 {
                PowerPC,POWER8@68 {
                PowerPC,POWER8@70 {
                PowerPC,POWER8@78 {

This happens because spapr_get_vcpu_id() maps VCPU ids to
cs->cpu_index in TCG mode. This confuses the code in
spapr_is_thread0_in_vcore(), since it assumes thread0 VCPU
ids to have a spapr->vsmt spacing.

    spapr_get_vcpu_id(cpu) % spapr->vsmt == 0

Actually, there's no real reason to expose cs->cpu_index instead
of the VCPU id, since we also generate it with TCG. Also we already
set it explicitly in spapr_set_vcpu_id(), so there's no real reason
either to call kvm_arch_vcpu_id() with KVM.

This patch unifies spapr_get_vcpu_id() to always return the computed
VCPU id both in TCG and KVM. This is one step forward towards KVM<->TCG
migration.

Fixes: 5d0fb1508e2d
Reported-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>

block: Fix NULL dereference on empty drive error

blk_error_action() sends a BLOCK_IO_ERROR QMP event which includes the
node name of its root node. If the BlockBackend represents an empty
drive, there is no root node, so we should not try to access its node
name. Make the field optional in the event and include it only when
the BlockBackend isn't empty.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>