migration: Export ram_load_postcopy()

Will be reused in postcopy fast load thread.

Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-6-peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

migration: Move migrate_allow_multifd and helpers into migration.c

This variable, along with its helpers, is used to detect whether multiple
channel will be supported for migration.  In follow up patches, there'll be
other capability that requires multi-channels.  Hence move it outside multifd
specific code and make it public.  Meanwhile rename it from "multifd" to
"multi_channels" to show its real meaning.

Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-5-peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

migration: Add pss.postcopy_requested status

This boolean flag shows whether the current page during migration is triggered
by postcopy or not.  Then in ram_save_host_page() and deeper stack we'll be
able to have a reference on the priority of this page.

Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-4-peterx@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

migration: Drop multifd tls_hostname cache

The hostname is cached N times, N equals to the multifd channels.

Drop that cache because after previous patch we've got s->hostname
being alive for the whole lifecycle of migration procedure.

Cc: Juan Quintela <quintela@redhat.com>
Cc: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-3-peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

migration: Postpone releasing MigrationState.hostname

We used to release it right after migrate_fd_connect().  That's not good
enough when there're more than one socket pair required, because it'll be
needed to establish TLS connection for the rest channels.

One example is multifd, where we copied over the hostname for each channel
but that's actually not needed.

Keeping the hostname until the cleanup phase of migration.

Cc: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20220331150857.74406-2-peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
  dgilbert: Fixup checkpatch error; don't need to check for NULL
  around g_free

tests: expand the migration precopy helper to support failures

The migration precopy testing helper function always expects the
migration to run to a completion state. There will be test scenarios
for TLS where expect either the client or server to fail the migration.
This expands the helper to cope with these scenarios.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-12-berrange@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests: switch migration FD passing test to use common precopy helper

The combination of the start and finish hooks allow the FD passing
code to use the precopy helper

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-11-berrange@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests: introduce ability to provide hooks for migration precopy test

There are alot of different scenarios to test with migration due to the
wide number of parameters and capabilities available. To enable sharing
of the basic precopy test scenario, we need to be able to set arbitrary
parameters and capabilities before the migration is initiated, but don't
want to have all this logic in the common helper function. Solve this
by defining two hooks that can be provided by the test case, one before
migration starts and one after migration finishes.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-10-berrange@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests: merge code for UNIX and TCP migration pre-copy tests

The test cases differ only in the URI they provide to the migration
commands, and the ability to set the dirty_ring mode. This code is
trivially merged into a common helper.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-9-berrange@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests: switch MigrateStart struct to be stack allocated

There's no compelling reason why the MigrateStart struct needs to be
heap allocated. Using stack allocation and static initializers is
simpler.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-8-berrange@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

migration: fix use of TLS PSK credentials with a UNIX socket

The migration TLS code has a check mandating that a hostname be
available when starting a TLS session. This is expected when using
x509 credentials, but is bogus for PSK and anonymous credentials
as neither involve hostname validation.

The TLS crdentials object gained suitable error reporting in the
case of TLS with x509 credentials, so there is no longer any need
for the migration code to do its own (incorrect) validation.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-7-berrange@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests: print newline after QMP response in qtest logs

The QMP commands have a trailing newline, but the response does not.
This makes the qtest logs hard to follow as the next QMP command
appears in the same line as the previous QMP response.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-5-berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests: support QTEST_TRACE env variable

When debugging failing qtests it is useful to be able to turn on trace
output to stderr. The QTEST_TRACE env variable contents get injected
as a '-trace <str>' command line arg

Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-4-berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

tests: improve error message when saving TLS PSK file fails

Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220310171821.3724080-3-berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>

Merge tag 'misc-pull-request' of gitlab.com:marcandre.lureau/qemu into staging

Misc cleanups

# -----BEGIN PGP SIGNATURE-----
#
# iQJQBAABCAA6FiEEh6m9kz+HxgbSdvYt2ujhCXWWnOUFAmJhYIscHG1hcmNhbmRy
# ZS5sdXJlYXVAcmVkaGF0LmNvbQAKCRDa6OEJdZac5W0jD/43n8PL2cv42lq6OaIS
# OYN9vfW9xgA9THZiUH4xEHYreZh+KofmY1PbJn1n7q+v6DecBiqM4fZr1LY8X3PM
# xRUH0l4gjXwYwX2cSSo5UTZ/PF248Uoo3tUE3vgMFkYghHMjKcTtaSkYEPGHf2nR
# t3m1qLG9w9YPhVg7PNCntjUKi+w2TtcrRVzP7V7XyFc1HrAoT0ys6KaBBrXMbcjz
# SxTRbcwSq+6aPjQIn0RWp8Hp1HkdNjegB98dkyqRLlVaugHZWPYDXDQTgVziQlX8
# dU8YrlvTOtDWwsNP6awWnW6/IjKuJjGR0wT3QKwi8JAZ0YV3egwEKoQRUAyHtnn2
# FkSMYgmJcF0ai1aIJFAx+3PIzCfS49lKXA0t303DtY3hRR9JKGMwaV2do9Wm2irt
# o7T1lKKN7R7R8Q3U4OsatYMYm7KYL07NEDiQCPloGvCo27ezkAWCKXAw1mRUkxKF
# jKwJPcnOUq21Jp6tpjsR8ifSw70jBSEWQSGqhXnDhZhx2C2/Qqkg2I8DagLiPger
# kYxbQ13LTG0R25YHa1r3UmzuD+HpZOM8XoLJc5yun/1UrwyR9ghHrOoxkSnRT2Ks
# QFn//xQ2SzUnGBNzNSMfTk8vzludxSWfFnOjkviF6E2Elnw3p8f/kOQRAft5dMBY
# ftgoy2yLone3HpKfjuOriicIzg==
# =0GLo
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 21 Apr 2022 06:47:55 AM PDT
# gpg:                using RSA key 87A9BD933F87C606D276F62DDAE8E10975969CE5
# gpg:                issuer "marcandre.lureau@redhat.com"
# gpg: Good signature from "Marc-André Lureau <marcandre.lureau@redhat.com>" [full]
# gpg:                 aka "Marc-André Lureau <marcandre.lureau@gmail.com>" [full]

* tag 'misc-pull-request' of gitlab.com:marcandre.lureau/qemu: (30 commits)
  qga: use fixed-length and GDateTime for log timestamp
  tests/fuzz: fix warning
  qga: remove need for QEMU atomic.h
  util: replace qemu_get_local_state_pathname()
  util: use qemu_create() in qemu_write_pidfile()
  util: use qemu_write_full() in qemu_write_pidfile()
  util: simplify write in signal handler
  qtest: simplify socket_send()
  qga: move qga_get_host_name()
  Move error_printf_unless_qmp() with monitor unit
  tests: run-time skip test-qga if TSAN is enabled
  compiler.h: add QEMU_SANITIZE_{ADDRESS,THREAD}
  tests: remove block/qdict checks from check-qobject.c
  include: move qdict_{crumple,flatten} declarations
  include: add qemu/keyval.h
  include: move qemu_fdatasync() to osdep
  include: move qemu_msync() to osdep
  compiler.h: replace QEMU_NORETURN with G_NORETURN
  osdep.h: move qemu_build_not_reached()
  doc/style: CLang -> Clang
  ...

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

Merge tag 'pull-target-arm-20220421' of https://git.linaro.org/people/pmaydell/qemu-arm into staging

target-arm queue:
 * hw/arm/virt: Check for attempt to use TrustZone with KVM or HVF
 * versal: Add the Cortex-R5s in the Real-Time Processing Unit (RPU) subsystem
 * versal: model enough of the Clock/Reset Low-power domain (CRL) to allow control of the Cortex-R5s
 * xlnx-zynqmp: Connect 4 TTC timers
 * exynos4210: Refactor GIC/combiner code to stop using qemu_split_irq
 * realview: replace 'qemu_split_irq' with 'TYPE_SPLIT_IRQ'
 * stellaris: replace 'qemu_split_irq' with 'TYPE_SPLIT_IRQ'
 * hw/core/irq: remove unused 'qemu_irq_split' function
 * npcm7xx: use symbolic constants for PWRON STRAP bit fields
 * virt: document impact of gic-version on max CPUs

# -----BEGIN PGP SIGNATURE-----
#
# iQJNBAABCAA3FiEE4aXFk81BneKOgxXPPCUl7RQ2DN4FAmJhPSUZHHBldGVyLm1h
# eWRlbGxAbGluYXJvLm9yZwAKCRA8JSXtFDYM3hAsD/4qzZK6LFL4kFH6E4z3tWIn
# ErHrfPGUt/SEfHLP+stQP/loFgkR1SNzcrIZ/HiDCB/W+IqQKuP+tHin2lMhO1tR
# KM6suUO1In2hoxfzimVta4F4GVN8ifY69qUYhaRxcBYSUpRXDNFJGsRIeT5JeUMd
# SArZUifRs7JUo25rIkg5Y+YZE37dmiA5gcuswtoLPa/UlDVqRxihLnItySmeutsc
# /Y8d/iym/ydlhvtL1OUt1KKYeg4ykrPzJCfvopsT2xgkwwB0PYci8//fa5IrRVlp
# Uw6yDssZrDIcXfVz88rdriILaszicCv8lOhTH6I74oXCatiyvi4BEzW8uGqVS8wt
# ff+AaKvGqb5t4GKKhCdpL2NzDwKBGWZHuruACs9IfvMkz62HE12Vr99qAKdQ3s93
# QnFIyUKg90mGkvKy8336zX3hnWjPH8wTASOXbNrgnt6GVLkqwy9ibug5kS+n77eJ
# BnkE5p3OfMVJ5a4o+iZbbDJKfzhNUHDSMIBbG1jRNzax1RgxOBtHFSqP5jmbpm+S
# agyr8h+Md0Tx1dwZKxdCGyvcbSZiG2WxRnci3dyT4MqYY1t1GEpOfcs1EN+CYKwG
# iuezZzJopjOFGaXQaB3OvbvCKxuroHKG61SmDmx+5OkfAxhrqe4ulwYij4jhsyhH
# t8zGzDOKLakv3li90xCX/w==
# =Rke9
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 21 Apr 2022 04:16:53 AM PDT
# gpg:                using RSA key E1A5C593CD419DE28E8315CF3C2525ED14360CDE
# gpg:                issuer "peter.maydell@linaro.org"
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>" [full]
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>" [full]
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>" [full]

* tag 'pull-target-arm-20220421' of https://git.linaro.org/people/pmaydell/qemu-arm: (31 commits)
  hw/arm: Use bit fields for NPCM7XX PWRON STRAPs
  hw/misc: Add PWRON STRAP bit fields in GCR module
  hw/arm/virt: impact of gic-version on max CPUs
  hw/core/irq: remove unused 'qemu_irq_split' function
  hw/arm/stellaris: replace 'qemu_split_irq' with 'TYPE_SPLIT_IRQ'
  hw/arm/realview: replace 'qemu_split_irq' with 'TYPE_SPLIT_IRQ'
  hw/arm/exynos4210: Drop Exynos4210Irq struct
  hw/arm/exynos4210: Put combiners into state struct
  hw/arm/exynos4210: Fold combiner splits into exynos4210_init_board_irqs()
  hw/arm/exynos4210: Don't connect multiple lines to external GIC inputs
  hw/arm/exynos4210: Connect MCT_G0 and MCT_G1 to both combiners
  hw/arm/exynos4210: Fill in irq_table[] for internal-combiner-only IRQ lines
  hw/arm/exynos4210: Use TYPE_SPLIT_IRQ in exynos4210_init_board_irqs()
  hw/arm/exynos4210: Delete unused macro definitions
  hw/arm/exynos4210: Move exynos4210_combiner_get_gpioin() into exynos4210.c
  hw/arm/exynos4210: Drop ext_gic_irq[] from Exynos4210Irq struct
  hw/arm/exynos4210: Put external GIC into state struct
  hw/arm/exynos4210: Move exynos4210_init_board_irqs() into exynos4210.c
  hw/arm/exynos4210: Fix code style nit in combiner_grp_to_gic_id[]
  hw/arm/exynos4210: Coalesce board_irqs and irq_table
  ...

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

qga: use fixed-length and GDateTime for log timestamp

The old code is kind of wrong. Say it's 1649309843.000001 seconds past
the epoch. Prints "1649309843.1". 9us later, it prints "1649309843.10".
Should really use %06lu for the microseconds part.

Use GDateTime instead, as suggested by Daniel.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>

tests/fuzz: fix warning

../tests/qtest/fuzz/generic_fuzz.c:746:17: warning: variable 'name' set but not used [-Wunused-but-set-variable]

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220420132624.2439741-42-marcandre.lureau@redhat.com>

qga: remove need for QEMU atomic.h

Since the introduction of guest-exec in/out/err redirections in commit
a1853dca74 ("qga: guest-exec simple stdin/stdout/stderr redirection"),
some execution state variables are handled with atomic ops. However,
there are no threads involved in this code (and glib sources are
dispatched in the same thread), and no other obvious reason to use them.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-40-marcandre.lureau@redhat.com>

util: replace qemu_get_local_state_pathname()

Simplify the function to only return the directory path. Callers are
adjusted to use the GLib function to build paths, g_build_filename().

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-39-marcandre.lureau@redhat.com>

util: use qemu_create() in qemu_write_pidfile()

qemu_open_old(O_CREATE) should be replaced with qemu_create() which
handles Error reporting.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-38-marcandre.lureau@redhat.com>

util: use qemu_write_full() in qemu_write_pidfile()

Mostly for correctness.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-37-marcandre.lureau@redhat.com>

util: simplify write in signal handler

Use qemu_write_full() instead of open-coding a write loop.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-36-marcandre.lureau@redhat.com>

qtest: simplify socket_send()

Reuse qemu_write_full().

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20220420132624.2439741-33-marcandre.lureau@redhat.com>

qga: move qga_get_host_name()

The function is specific to qemu-ga, no need to share it in QEMU.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Konstantin Kostiuk <kkostiuk@redhat.com>
Message-Id: <20220420132624.2439741-32-marcandre.lureau@redhat.com>

Move error_printf_unless_qmp() with monitor unit

Since it depends on monitor code, and error_vprintf_unless_qmp() is
already there.

This will help to move error-report in a common subproject.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-31-marcandre.lureau@redhat.com>

tests: run-time skip test-qga if TSAN is enabled

This allows to make sure the test is still built, and gives more
accurate report details.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-30-marcandre.lureau@redhat.com>

compiler.h: add QEMU_SANITIZE_{ADDRESS,THREAD}

Simplify a bit pre-compiler conditions.

For TSAN, QEMU already has CONFIG_TSAN, but it is only set when the
fiber API is present. (I wonder whether supporting TSAN without the
fiber API is really relevant)

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-27-marcandre.lureau@redhat.com>

tests: remove block/qdict checks from check-qobject.c

The functions are already covered in check-block-qdict.c.
This will help moving QAPI-related tests in a common subproject.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-26-marcandre.lureau@redhat.com>

include: move qdict_{crumple,flatten} declarations

Move them where they belong, since the functions are implemented in block-qdict.c.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-25-marcandre.lureau@redhat.com>

include: add qemu/keyval.h

Do not require the whole option machinery to handle keyval, as it is
used by QAPI alone, without the option API. And match the associated
unit name.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-24-marcandre.lureau@redhat.com>

include: move qemu_fdatasync() to osdep

Move QEMU-specific code to util/osdep.c, so cutils can become a common
subproject.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-22-marcandre.lureau@redhat.com>

include: move qemu_msync() to osdep

The implementation depends on the OS. (and longer-term goal is to move
cutils to a common subproject)

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-21-marcandre.lureau@redhat.com>

compiler.h: replace QEMU_NORETURN with G_NORETURN

G_NORETURN was introduced in glib 2.68, fallback to G_GNUC_NORETURN in
glib-compat.

Note that this attribute must be placed before the function declaration
(bringing a bit of consistency in qemu codebase usage).

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Warner Losh <imp@bsdimp.com>
Message-Id: <20220420132624.2439741-20-marcandre.lureau@redhat.com>

osdep.h: move qemu_build_not_reached()

Move the macro and declaration so it can use glib in the following
patch (it already depends on glib anyway for !optimize)

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-19-marcandre.lureau@redhat.com>

doc/style: CLang -> Clang

It's not the way it is usually written (see https://clang.llvm.org/).

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Damien Hedde <damien.hedde@greensocs.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-17-marcandre.lureau@redhat.com>

intc/exynos4210_gic: replace snprintf() with g_strdup_printf()

While at it, replace '%x' with '%u' as suggested by Philippe Mathieu-Daudé.

Also fixes a GCC 12.0.1 -Wformat-overflow false-positive.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-16-marcandre.lureau@redhat.com>

arm/allwinner-a10: replace snprintf() with g_strdup_printf()

Also fixes a GCC 12.0.1 false-positive:
../hw/arm/allwinner-a10.c: In function ‘aw_a10_realize’:
../hw/arm/allwinner-a10.c:135:35: error: ‘%d’ directive writing between 1 and 11 bytes into a region of size 8 [-Werror=format-overflow=]
  135 |             sprintf(bus, "usb-bus.%d", i);
      |                                   ^~

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-15-marcandre.lureau@redhat.com>

arm/digic: replace snprintf() with g_strdup_printf()

Also fixes a GCC 12.0.1 false-positive:
../hw/arm/digic.c: In function ‘digic_init’:
../hw/arm/digic.c:45:54: error: ‘%d’ directive output may be truncated writing between 1 and 11 bytes into a region of size 5 [-Werror=format-truncation=]
   45 |         snprintf(name, DIGIC_TIMER_NAME_MLEN, "timer[%d]", i);
      |                                                      ^~

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-14-marcandre.lureau@redhat.com>

docs: trace-events-all is installed without renaming

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-13-marcandre.lureau@redhat.com>

qga: replace usleep() with g_usleep()

The latter simply requires glib.h, while the former is not in the
Windows API (but provided by mingw header & CRT)

Also simplify the expression for 1/10s.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-12-marcandre.lureau@redhat.com>

include: rename qemu-common.h qemu/help-texts.h

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Warner Losh <imp@bsdimp.com>
Message-Id: <20220420132624.2439741-7-marcandre.lureau@redhat.com>

hw/hyperv: remove needless qemu-common.h include

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-6-marcandre.lureau@redhat.com>

Simplify softmmu/main.c

Move qemu_main() declaration to a new header.

Simplify main.c since both cocoa & sdl cannot be enabled together.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@gmail.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-5-marcandre.lureau@redhat.com>

scripts/analyze-inclusions: drop qemu-common.h from analysis

The header is no longer commonly included.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-4-marcandre.lureau@redhat.com>

glib-compat: isolate g_date_time_format_iso8601 version-bypass

The solution was discussed with Markus Armbruster during the review:
https://patchew.org/QEMU/20220323155743.1585078-1-marcandre.lureau@redhat.com/20220323155743.1585078-14-marcandre.lureau@redhat.com/

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20220420132624.2439741-3-marcandre.lureau@redhat.com>

hw/arm: Use bit fields for NPCM7XX PWRON STRAPs

This patch uses the defined fields to describe PWRON STRAPs for
better readability.

Signed-off-by: Hao Wu <wuhaotsh@google.com>
Reviewed-by: Patrick Venture <venture@google.com>
Message-id: 20220411165842.3912945-3-wuhaotsh@google.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/misc: Add PWRON STRAP bit fields in GCR module

Similar to the Aspeed code in include/misc/aspeed_scu.h, we define
the PWRON STRAP fields in their corresponding module for NPCM7XX.

Signed-off-by: Hao Wu <wuhaotsh@google.com>
Reviewed-by: Patrick Venture <venture@google.com>
Message-id: 20220411165842.3912945-2-wuhaotsh@google.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/virt: impact of gic-version on max CPUs

Describe that the gic-version influences the maximum number of CPUs.

Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Message-id: 20220413231456.35811-1-heinrich.schuchardt@canonical.com
[PMM: minor punctuation tweaks]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/core/irq: remove unused 'qemu_irq_split' function

Signed-off-by: Zongyuan Li <zongyuan.li@smartx.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20220324181557.203805-5-zongyuan.li@smartx.com
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/811
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/stellaris: replace 'qemu_split_irq' with 'TYPE_SPLIT_IRQ'

Signed-off-by: Zongyuan Li <zongyuan.li@smartx.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20220324181557.203805-3-zongyuan.li@smartx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/realview: replace 'qemu_split_irq' with 'TYPE_SPLIT_IRQ'

Signed-off-by: Zongyuan Li <zongyuan.li@smartx.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20220324181557.203805-2-zongyuan.li@smartx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/exynos4210: Drop Exynos4210Irq struct

The only time we use the int_combiner_irq[] and ext_combiner_irq[]
arrays in the Exynos4210Irq struct is during realize of the SoC -- we
initialize them with the input IRQs of the combiner devices, and then
connect those to outputs of other devices in
exynos4210_init_board_irqs().  Now that the combiner objects are
easily accessible as s->int_combiner and s->ext_combiner we can make
the connections directly from one device to the other without going
via these arrays.

Since these are the only two remaining elements of Exynos4210Irq,
we can remove that struct entirely.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-19-peter.maydell@linaro.org

hw/arm/exynos4210: Put combiners into state struct

Switch the creation of the combiner devices to the new-style
"embedded in state struct" approach, so we can easily refer
to the object elsewhere during realize.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-18-peter.maydell@linaro.org

hw/arm/exynos4210: Fold combiner splits into exynos4210_init_board_irqs()

At this point, the function exynos4210_init_board_irqs() splits input
IRQ lines to connect them to the input combiner, output combiner and
external GIC.  The function exynos4210_combiner_get_gpioin() splits
some of the combiner input lines further to connect them to multiple
different inputs on the combiner.

Because (unlike qemu_irq_split()) the TYPE_SPLIT_IRQ device has a
configurable number of outputs, we can do all this in one place, by
making exynos4210_init_board_irqs() add extra outputs to the splitter
device when it must be connected to more than one input on each
combiner.

We do this with a new data structure, the combinermap, which is an
array each of whose elements is a list of the interrupt IDs on the
combiner which must be tied together.  As we loop through each
interrupt ID, if we find that it is the first one in one of these
lists, we configure the splitter device with eonugh extra outputs and
wire them up to the other interrupt IDs in the list.

Conveniently, for all the cases where this is necessary, the
lowest-numbered interrupt ID in each group is in the range of the
external combiner, so we only need to code for this in the first of
the two loops in exynos4210_init_board_irqs().

The old code in exynos4210_combiner_get_gpioin() which is being
deleted here had several problems which don't exist in the new code
in its handling of the multi-core timer interrupts:
 (1) the case labels specified bits 4 ... 8, but bit '8' doesn't
     exist; these should have been 4 ... 7
 (2) it used the input irq[EXYNOS4210_COMBINER_GET_IRQ_NUM(1, bit + 4)]
     multiple times as the input of several different splitters,
     which isn't allowed
 (3) in an apparent cut-and-paste error, the cases for all the
     multi-core timer inputs used "bit + 4" even though the
     bit range for the case was (intended to be) 4 ... 7, which
     meant it was looking at non-existent bits 8 ... 11.
None of these exist in the new code.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-17-peter.maydell@linaro.org

hw/arm/exynos4210: Don't connect multiple lines to external GIC inputs

The combiner_grp_to_gic_id[] array includes the EXT_GIC_ID_MCT_G0
and EXT_GIC_ID_MCT_G1 multiple times. This means that we will
connect multiple IRQs up to the same external GIC input, which
is not permitted. We do the same thing in the code in
exynos4210_init_board_irqs() because the conditionals selecting
an irq_id in the first loop match multiple interrupt IDs.

Overall we do this for interrupt IDs
(1, 4), (12, 4), (35, 4), (51, 4), (53, 4) for EXT_GIC_ID_MCT_G0
and
(1, 5), (12, 5), (35, 5), (51, 5), (53, 5) for EXT_GIC_ID_MCT_G1

These correspond to the cases for the multi-core timer that we are
wiring up to multiple inputs on the combiner in
exynos4210_combiner_get_gpioin().  That code already deals with all
these interrupt IDs being the same input source, so we don't need to
connect the external GIC interrupt for any of them except the first
(1, 4) and (1, 5). Remove the array entries and conditionals which
were incorrectly causing us to wire up extra lines.

This bug didn't cause any visible effects, because we only connect
up a device to the "primary" ID values (1, 4) and (1, 5), so the
extra lines would never be set to a level.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-16-peter.maydell@linaro.org

hw/arm/exynos4210: Connect MCT_G0 and MCT_G1 to both combiners

Currently for the interrupts MCT_G0 and MCT_G1 which are
the only ones in the input range of the external combiner
and which are also wired to the external GIC, we connect
them only to the internal combiner and the external GIC.
This seems likely to be a bug, as all other interrupts
which are in the input range of both combiners are
connected to both combiners. (The fact that the code in
exynos4210_combiner_get_gpioin() is also trying to wire
up these inputs on both combiners also suggests this.)

Wire these interrupts up to both combiners, like the rest.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-15-peter.maydell@linaro.org

hw/arm/exynos4210: Fill in irq_table[] for internal-combiner-only IRQ lines

In exynos4210_init_board_irqs(), the loop that handles IRQ lines that
are in a range that applies to the internal combiner only creates a
splitter for those interrupts which go to both the internal combiner
and to the external GIC, but it does nothing at all for the
interrupts which don't go to the external GIC, leaving the
irq_table[] array element empty for those.  (This will result in
those interrupts simply being lost, not in a QEMU crash.)

I don't have a reliable datasheet for this SoC, but since we do wire
up one interrupt line in this category (the HDMI I2C device on
interrupt 16,1), this seems like it must be a bug in the existing
QEMU code.  Fill in the irq_table[] entries where we're not splitting
the IRQ to both the internal combiner and the external GIC with the
IRQ line of the internal combiner.  (That is, these IRQ lines go to
just one device, not multiple.)

This bug didn't have any visible guest effects because the only
implemented device that was affected was the HDMI I2C controller,
and we never connect any I2C devices to that bus.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-14-peter.maydell@linaro.org

hw/arm/exynos4210: Use TYPE_SPLIT_IRQ in exynos4210_init_board_irqs()

In exynos4210_init_board_irqs(), use the TYPE_SPLIT_IRQ device
instead of qemu_irq_split().

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-13-peter.maydell@linaro.org

hw/arm/exynos4210: Delete unused macro definitions

Delete a couple of #defines which are never used.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-12-peter.maydell@linaro.org

hw/arm/exynos4210: Move exynos4210_combiner_get_gpioin() into exynos4210.c

The function exynos4210_combiner_get_gpioin() currently lives in
exynos4210_combiner.c, but it isn't really part of the combiner
device itself -- it is a function that implements the wiring up of
some interrupt sources to multiple combiner inputs.  Move it to live
with the other SoC-level code in exynos4210.c, along with a few
macros previously defined in exynos4210.h which are now used only
in exynos4210.c.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-11-peter.maydell@linaro.org

hw/arm/exynos4210: Drop ext_gic_irq[] from Exynos4210Irq struct

The only time we use the ext_gic_irq[] array in the Exynos4210Irq
struct is during realize of the SoC -- we initialize it with the
input IRQs of the external GIC device, and then connect those to
outputs of other devices further on in realize (including in the
exynos4210_init_board_irqs() function).  Now that the ext_gic object
is easily accessible as s->ext_gic we can make the connections
directly from one device to the other without going via this array.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-10-peter.maydell@linaro.org

hw/arm/exynos4210: Put external GIC into state struct

Switch the creation of the external GIC to the new-style "embedded in
state struct" approach, so we can easily refer to the object
elsewhere during realize.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-9-peter.maydell@linaro.org

hw/arm/exynos4210: Move exynos4210_init_board_irqs() into exynos4210.c

The function exynos4210_init_board_irqs() currently lives in
exynos4210_gic.c, but it isn't really part of the exynos4210.gic
device -- it is a function that implements (some of) the wiring up of
interrupts between the SoC's GIC and combiner components.  This means
it fits better in exynos4210.c, which is the SoC-level code.  Move it
there. Similarly, exynos4210_git_irq() is used almost only in the
SoC-level code, so move it too.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-8-peter.maydell@linaro.org

hw/arm/exynos4210: Fix code style nit in combiner_grp_to_gic_id[]

Fix a missing set of spaces around '-' in the definition of
combiner_grp_to_gic_id[]. We're about to move this code, so
fix the style issue first to keep checkpatch happy with the
code-motion patch.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-7-peter.maydell@linaro.org

hw/arm/exynos4210: Coalesce board_irqs and irq_table

The exynos4210 code currently has two very similar arrays of IRQs:

 * board_irqs is a field of the Exynos4210Irq struct which is filled
   in by exynos4210_init_board_irqs() with the appropriate qemu_irqs
   for each IRQ the board/SoC can assert
 * irq_table is a set of qemu_irqs pointed to from the
   Exynos4210State struct.  It's allocated in exynos4210_init_irq,
   and the only behaviour these irqs have is that they pass on the
   level to the equivalent board_irqs[] irq

The extra indirection through irq_table is unnecessary, so coalesce
these into a single irq_table[] array as a direct field in
Exynos4210State which exynos4210_init_board_irqs() fills in.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-6-peter.maydell@linaro.org

hw/arm/exynos4210: Drop int_gic_irq[] from Exynos4210Irq struct

The only time we use the int_gic_irq[] array in the Exynos4210Irq
struct is in the exynos4210_realize() function: we initialize it with
the GPIO inputs of the a9mpcore device, and then a bit later on we
connect those to the outputs of the internal combiner.  Now that the
a9mpcore object is easily accessible as s->a9mpcore we can make the
connection directly from one device to the other without going via
this array.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-5-peter.maydell@linaro.org

hw/arm/exynos4210: Put a9mpcore device into state struct

The exynos4210 SoC mostly creates its child devices as if it were
board code.  This includes the a9mpcore object.  Switch that to a
new-style "embedded in the state struct" creation, because in the
next commit we're going to want to refer to the object again further
down in the exynos4210_realize() function.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-4-peter.maydell@linaro.org

hw/intc/exynos4210_gic: Remove unused TYPE_EXYNOS4210_IRQ_GATE

Now we have removed the only use of TYPE_EXYNOS4210_IRQ_GATE we can
delete the device entirely.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Francisco Iglesias <frasse.iglesias@gmail.com>
Message-id: 20220404154658.565020-3-peter.maydell@linaro.org

hw/arm/exynos4210: Use TYPE_OR_IRQ instead of custom OR-gate device

The Exynos4210 SoC device currently uses a custom device
"exynos4210.irq_gate" to model the OR gate that feeds each CPU's IRQ
line.  We have a standard TYPE_OR_IRQ device for this now, so use
that instead.

(This is a migration compatibility break, but that is OK for this
machine type.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404154658.565020-2-peter.maydell@linaro.org

hw/arm: versal: Connect the CRL

Connect the CRL (Clock Reset LPD) to the Versal SoC.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Frederic Konrad <fkonrad@amd.com>
Reviewed-by: Francisco Iglesias <francisco.iglesias@amd.com>
Message-id: 20220406174303.2022038-5-edgar.iglesias@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/misc: Add a model of the Xilinx Versal CRL

Add a model of the Xilinx Versal CRL.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Frederic Konrad <fkonrad@amd.com>
Reviewed-by: Francisco Iglesias <francisco.iglesias@amd.com>
Message-id: 20220406174303.2022038-4-edgar.iglesias@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm: versal: Add the Cortex-R5Fs

Add the Cortex-R5Fs of the Versal RPU (Real-time Processing Unit)
subsystem.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Francisco Iglesias <francisco.iglesias@amd.com>
Message-id: 20220406174303.2022038-3-edgar.iglesias@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm: versal: Create an APU CPU Cluster

Create an APU CPU Cluster. This is in preparation to add the RPU.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Francisco Iglesias <francisco.iglesias@amd.com>
Message-id: 20220406174303.2022038-2-edgar.iglesias@xilinx.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/xlnx-zynqmp: Connect 4 TTC timers

Connect the 4 TTC timers on the ZynqMP.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Luc Michel <luc@lmichel.fr>
Reviewed-by: Francisco Iglesias <frasse.iglesias@gmail.com>
Message-id: 20220331222017.2914409-3-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

timer: cadence_ttc: Break out header file to allow embedding

Break out header file to allow embedding of the the TTC.

Signed-off-by: Edgar E. Iglesias <edgar.iglesias@amd.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Luc Michel <luc@lmichel.fr>
Reviewed-by: Francisco Iglesias <frasse.iglesias@gmail.com>
Message-id: 20220331222017.2914409-2-edgar.iglesias@gmail.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

hw/arm/virt: Check for attempt to use TrustZone with KVM or HVF

It's not possible to provide the guest with the Security extensions
(TrustZone) when using KVM or HVF, because the hardware
virtualization extensions don't permit running EL3 guest code.
However, we weren't checking for this combination, with the result
that QEMU would assert if you tried it:

$ qemu-system-aarch64 -enable-kvm -machine virt,secure=on -cpu host -display none
Unexpected error in object_property_find_err() at ../../qom/object.c:1304:
qemu-system-aarch64: Property 'host-arm-cpu.secure-memory' not found
Aborted

Check for this combination of options and report an error, in the
same way we already do for attempts to give a KVM or HVF guest the
Virtualization or MTE extensions. Now we will report:

qemu-system-aarch64: mach-virt: KVM does not support providing Security extensions (TrustZone) to the guest CPU

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/961
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220404155301.566542-1-peter.maydell@linaro.org

Merge tag 'pull-ppc-20220420-2' of https://gitlab.com/danielhb/qemu into staging

ppc patch queue for 2022-04-20

First batch of ppc patches for QEMU 7.1:

- skiboot firmware version bump
- pseries: add 2M DDW pagesize
- pseries: make virtual hypervisor code TCG only
- powernv: introduce GPIO lines for PSIHB device
- powernv: remove PCIE root bridge LSI
- target/ppc: alternative softfloat 128 bit integer support
- assorted fixes

# -----BEGIN PGP SIGNATURE-----
#
# iHUEABYKAB0WIQQX6/+ZI9AYAK8oOBk82cqW3gMxZAUCYmB/ngAKCRA82cqW3gMx
# ZE10AP4wPeJQ3fxXb5ylVtL4qkJaLWy6VrJBQSKSb5YEA0fhegEA9ZufpnENQePU
# gZF0eFAQK/DbSnDyvRQVpGcJM0K1UgI=
# =nVRw
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 20 Apr 2022 02:48:14 PM PDT
# gpg:                using EDDSA key 17EBFF9923D01800AF2838193CD9CA96DE033164
# gpg: Can't check signature: No public key

* tag 'pull-ppc-20220420-2' of https://gitlab.com/danielhb/qemu: (23 commits)
  hw/ppc: change indentation to spaces from TABs
  target/ppc: Add two missing register callbacks on POWER10
  ppc/pnv: Remove LSI on the PCIE host bridge
  pcie: Don't try triggering a LSI when not defined
  ppc/vof: Fix uninitialized string tracing
  hw/ppc/ppc405_boards: Initialize g_autofree pointer
  target/ppc: implement xscvqp[su]qz
  target/ppc: implement xscv[su]qqp
  softfloat: add float128_to_int128
  softfloat: add float128_to_uint128
  softfloat: add int128_to_float128
  softfloat: add uint128_to_float128
  qemu/int128: add int128_urshift
  target/ppc: Improve KVM hypercall trace
  spapr: Move nested KVM hypercalls under a TCG only config.
  spapr: Move hypercall_register_softmmu
  ppc/pnv: Remove useless checks in set_irq handlers
  ppc/pnv: Remove PnvPsiClas::irq_set
  ppc/pnv: Remove PnvOCC::psi link
  ppc/pnv: Remove PnvLpcController::psi link
  ...

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

Merge tag 'pull-tcg-20220420' of https://gitlab.com/rth7680/qemu into staging

Cleanup sysemu/tcg.h usage.
Fix indirect lowering vs cond branches
Remove ATOMIC_MMU_IDX
Add tcg_constant_ptr

# -----BEGIN PGP SIGNATURE-----
#
# iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmJgW38dHHJpY2hhcmQu
# aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV8tpggApfg2CDI0bRMDBh0g
# 04/xwNnzHuSa84/ocMOMUfD5pvBblUmeTH8fAwqcAPDM/EEZwWZl2V1bYzuIrbmR
# 8zV+r1cOenDF5Tz8PWfy8XssinTVtTWh/TE0XNV9R/SbEM9eMsjHNu5osKVuLuq1
# rnHWZf8LuY7xGsy4GYqPN0dLE6HtQOfpj/eLGRAj9mZ7re0jKeWg3GdxYoiYDmks
# NKmNHYcWD+SjjFvXlOafniQsHbBZmQc/qp7AShG/+VcYY9o1VfncWD6I2dV13RdB
# N7++ZhGyQR4NOVo6CN1zLKhfuJqzH2q+qJ7vQ3xtXNAk53LGQ91zjoE+3KaJTrcy
# dmnLUw==
# =aKdS
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 20 Apr 2022 12:14:07 PM PDT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [ultimate]

* tag 'pull-tcg-20220420' of https://gitlab.com/rth7680/qemu:
  tcg: Add tcg_constant_ptr
  accel/tcg: Remove ATOMIC_MMU_IDX
  tcg: Fix indirect lowering vs TCG_OPF_COND_BRANCH
  Don't include sysemu/tcg.h if it is not necessary

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

hw/ppc: change indentation to spaces from TABs

There are still some files in the QEMU PPC code base that use TABs for
indentation instead of using  spaces. The TABs should be replaced so
that we have a consistent coding style.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/374
Signed-off-by: Guo Zhi <qtxuning1999@sjtu.edu.cn>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220412021240.2080218-1-qtxuning1999@sjtu.edu.cn>
[danielhb: trimmed commit msg to 72 chars per line]
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc: Add two missing register callbacks on POWER10

This patch adds tcg accessors for 2 SPRs which were missing on P10:

- the TBU40 register is used to write the upper 40 bits of the
timebase register. It is used by kvm to update the timebase when
entering/exiting the guest on P9 and above. The missing definition was
causing erratic decrementer interrupts in a pseries/kvm guest running
in a powernv10/tcg host, typically resulting in hangs.

- the missing DPDES SPR was found through code inspection. It exists
unchanged on P10.

Both existed on previous versions of the processor and a bit of git
archaeology hints that they were added while the P10 model was already
being worked on so they may have simply fallen through the cracks.

Signed-off-by: Frederic Barrat <fbarrat@linux.ibm.com>
Reviewed-by: Fabiano Rosas <farosas@linux.ibm.com>
Message-Id: <20220411125900.352028-1-fbarrat@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

ppc/pnv: Remove LSI on the PCIE host bridge

The phb3/phb4/phb5 root ports inherit from the default PCIE root port
implementation, which requests a LSI interrupt (#INTA). On real
hardware (POWER8/POWER9/POWER10), there is no such LSI. This patch
corrects it so that it matches the hardware.

As a consequence, the device tree previously generated was bogus, as
the root bridge LSI was not properly mapped. On some
implementation (powernv9), it was leading to inconsistent interrupt
controller (xive) data. With this patch, it is now clean.

Signed-off-by: Frederic Barrat <fbarrat@linux.ibm.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220408131303.147840-3-fbarrat@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

pcie: Don't try triggering a LSI when not defined

This patch skips [de]asserting a LSI interrupt if the device doesn't
have any LSI defined. Doing so would trigger an assert in
pci_irq_handler().

The PCIE root port implementation in qemu requests a LSI (INTA), but a
subclass may want to change that behavior since it's a valid
configuration. For example on the POWER8/POWER9/POWER10 systems, the
root bridge doesn't request any LSI.

Signed-off-by: Frederic Barrat <fbarrat@linux.ibm.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20220408131303.147840-2-fbarrat@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

ppc/vof: Fix uninitialized string tracing

There are error paths which do not initialize propname but the trace_exit
label prints it anyway. This initializes the problem string.

Spotted by Coverity CID 1487241.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220406045013.3610172-1-aik@ozlabs.ru>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/ppc405_boards: Initialize g_autofree pointer

Resolves the only compiler warning when building a full QEMU under Arch Linux:

  Compiling C object libqemu-ppc-softmmu.fa.p/hw_ppc_ppc405_boards.c.o
  In file included from /usr/include/glib-2.0/glib.h:114,
                   from qemu/include/glib-compat.h:32,
                   from qemu/include/qemu/osdep.h:132,
                   from ../src/hw/ppc/ppc405_boards.c:25:
  ../src/hw/ppc/ppc405_boards.c: In function ‘ref405ep_init’:
  /usr/include/glib-2.0/glib/glib-autocleanups.h:28:3: warning: ‘filename’ may be used uninitialized in this function [-Wmaybe-uninitialized]
     28 |   g_free (*pp);
        |   ^~~~~~~~~~~~
  ../src/hw/ppc/ppc405_boards.c:265:26: note: ‘filename’ was declared here
    265 |         g_autofree char *filename;
        |                          ^~~~~~~~

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20220405123534.3395-1-shentey@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc: implement xscvqp[su]qz

Implement the following PowerISA v3.1 instructions:
xscvqpsqz: VSX Scalar Convert with round to zero Quad-Precision to
           Signed Quadword
xscvqpuqz: VSX Scalar Convert with round to zero Quad-Precision to
           Unsigned Quadword

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220330175932.6995-9-matheus.ferst@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc: implement xscv[su]qqp

Implement the following PowerISA v3.1 instructions:
xscvsqqp: VSX Scalar Convert with round Signed Quadword to
          Quad-Precision
xscvuqqp: VSX Scalar Convert with round Unsigned Quadword to
          Quad-Precision format

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220330175932.6995-8-matheus.ferst@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

softfloat: add float128_to_int128

Implements float128_to_int128 based on parts_float_to_int logic.

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220330175932.6995-7-matheus.ferst@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

softfloat: add float128_to_uint128

Implements float128_to_uint128 based on parts_float_to_uint logic.

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220330175932.6995-6-matheus.ferst@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

softfloat: add int128_to_float128

Based on parts_sint_to_float, implements int128_to_float128 to convert a
signed 128-bit value received through an Int128 argument.

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Message-Id: <20220330175932.6995-5-matheus.ferst@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

softfloat: add uint128_to_float128

Based on parts_uint_to_float, implements uint128_to_float128 to convert
an unsigned 128-bit value received through an Int128 argument.

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220330175932.6995-4-matheus.ferst@eldorado.org.br>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

qemu/int128: add int128_urshift

Implement an unsigned right shift for Int128 values and add the same
tests cases of int128_rshift in the unit test.

Signed-off-by: Matheus Ferst <matheus.ferst@eldorado.org.br>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220330175932.6995-3-matheus.ferst@eldorado.org.br>
[danielhb: fixed long lines in test_urshift()]
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc: Improve KVM hypercall trace

Before:

  kvm_handle_papr_hcall handle PAPR hypercall
  kvm_handle_papr_hcall handle PAPR hypercall
  kvm_handle_papr_hcall handle PAPR hypercall
  kvm_handle_papr_hcall handle PAPR hypercall
  kvm_handle_papr_hcall handle PAPR hypercall
  kvm_handle_papr_hcall handle PAPR hypercall

After:

  kvm_handle_papr_hcall 0x3a8
  kvm_handle_papr_hcall 0x3ac
  kvm_handle_papr_hcall 0x108
  kvm_handle_papr_hcall 0x104
  kvm_handle_papr_hcall 0x104
  kvm_handle_papr_hcall 0x108

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20220325223316.276494-1-farosas@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

spapr: Move nested KVM hypercalls under a TCG only config.

These are the spapr virtual hypervisor implementation of the nested
KVM API. They only make sense when running with TCG.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20220325221113.255834-3-farosas@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

spapr: Move hypercall_register_softmmu

I'm moving this because next patch will add more code under the ifdef
and it will be cleaner if we keep them together.

Also switch the ifdef branches to make it more convenient to add code
under CONFIG_TCG in the next patch.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Message-Id: <20220325221113.255834-2-farosas@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

ppc/pnv: Remove useless checks in set_irq handlers

Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220323072846.1780212-6-clg@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

ppc/pnv: Remove PnvPsiClas::irq_set

All devices raising PSI interrupts are now converted to use GPIO lines
and the pnv_psi_irq_set() routines have become useless. Drop them.

Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220323072846.1780212-5-clg@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

ppc/pnv: Remove PnvOCC::psi link

Use an anonymous output GPIO line to connect the OCC device with the
PSIHB device and raise the appropriate PSI IRQ line depending on the
processor model.

Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220323072846.1780212-4-clg@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

ppc/pnv: Remove PnvLpcController::psi link

Create an anonymous output GPIO line to connect the LPC device with
the PSIHB device and raise the appropriate PSI IRQ line depending on
the processor model.

A temporary __pnv_psi_irq_set() routine is introduced to handle the
transition. It will be removed when all devices raising PSI interrupts
are converted to use GPIOs.

Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220323072846.1780212-3-clg@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

ppc/pnv: Fix PSI IRQ definition

On HW, the PSI and FSP interrupt levels are muxed under the same
interrupt number. For coding reasons, an extra IRQ number was
introduced to index register values in an array. It increased the
count of IRQs which do not fit in the PSI IRQ range anymore.

The PSI and FSP interrupts should be modeled with an extra level of
GPIO lines but since QEMU does not support them, simply drop the extra
number to stay within the IRQ range.

Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20220323072846.1780212-2-clg@kaod.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>