target/i386: Add support for CMPCCXADD in CPUID enumeration

CMPccXADD is a new set of instructions in the latest Intel platform
Sierra Forest. This new instruction set includes a semaphore operation
that can compare and add the operands if condition is met, which can
improve database performance.

The bit definition:
CPUID.(EAX=7,ECX=1):EAX[bit 7]

Add CPUID definition for CMPCCXADD.

Signed-off-by: Jiaxi Chen <jiaxi.chen@linux.intel.com>
Signed-off-by: Tao Su <tao1.su@linux.intel.com>
Reviewed-by: Xiaoyao Li <xiaoyao.li@intel.com>
Message-Id: <20230303065913.1246327-2-tao1.su@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

i386/cpu: Update how the EBX register of CPUID 0x8000001F is set

Update the setting of CPUID 0x8000001F EBX to clearly document the ranges
associated with fields being set.

Fixes: 6cb8f2a663 ("cpu/i386: populate CPUID 0x8000_001F when SEV is active")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <5822fd7d02b575121380e1f493a8f6d9eba2b11a.1664550870.git.thomas.lendacky@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

i386/sev: Update checks and information related to reduced-phys-bits

The value of the reduced-phys-bits parameter is propogated to the CPUID
information exposed to the guest. Update the current validation check to
account for the size of the CPUID field (6-bits), ensuring the value is
in the range of 1 to 63.

Maintain backward compatibility, to an extent, by allowing a value greater
than 1 (so that the previously documented value of 5 still works), but not
allowing anything over 63.

Fixes: d8575c6c02 ("sev/i386: add command to initialize the memory encryption context")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <cca5341a95ac73f904e6300f10b04f9c62e4e8ff.1664550870.git.thomas.lendacky@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qemu-options.hx: Update the reduced-phys-bits documentation

A guest only ever experiences, at most, 1 bit of reduced physical
addressing. Update the documentation to reflect this as well as change
the example value on the reduced-phys-bits option.

Fixes: a9b4942f48 ("target/i386: add Secure Encrypted Virtualization (SEV) object")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <13a62ced1808546c1d398e2025cf85f4c94ae123.1664550870.git.thomas.lendacky@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

qapi, i386/sev: Change the reduced-phys-bits value from 5 to 1

A guest only ever experiences, at most, 1 bit of reduced physical
addressing. Change the query-sev-capabilities json comment to use 1.

Fixes: 31dd67f684 ("sev/i386: qmp: add query-sev-capabilities command")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <cb96d8e09154533af4b4e6988469bc0b32390b65.1664550870.git.thomas.lendacky@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Merge tag 'migration-20230427-pull-request' of https://gitlab.com/juan.quintela/qemu into staging

Migration Pull request (20230427 edition)

Hi

Everything that has been reviewed:
- stat64_set() by paolo
- atomic_counters series fully reviewed (juan)
- move capabilities to options.c fully reviewed (juan)
- fix the channels_ready semaphore (juan)
- multifd flush optimization reviewed (juan)

Please, apply.

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEGJn/jt6/WMzuA0uC9IfvGFhy1yMFAmRKi3YACgkQ9IfvGFhy
# 1yPtERAAgkBt4g3tpn90q+QeFLBPyo0KGxX4UiAmh8aiGRAJzmSJfjaQG3RgK0vM
# Iyat60Tx6lpHta7xRE+pDF7tXmoi0631B5FJjfuyTIFGWdrOpjGHihuxWdnIjK0H
# ZhR3LfQvHqybMfiyZN/7T5/qICB0zZ+erIeoaurRU3RVrO8yRPM+VBwsYr1RW1S4
# DxUeXEJF09tat7TgiQYp2w5cMPcQ2AvzgViR2r7yc0WjY2yUscyMjuPbHPqAsYWB
# UkbNuH7M/quvHqGvgxGYEQ+a6wyBzKlissjxWe8KTuSLNML2RVR6IrpuEPe6XyGe
# FWi+XRB5114EMy7YjjgTWzcLMC9EmhpzKTgJi0NEtL+cGRb9ZaSaSLE5yO7GLF9b
# efhCXT9jIef9/3Hi+Kr0juJ3nX6c7pINTAsx7wxKU0phz5kYx37IqIioMlVDh1ew
# ECpH+8Syrp/zYytc/HIpyOwIR9VdZrRi6jr2yqcc9lLLcuNUzl+E+9mAxWzO/gef
# +cYPnQZC0rjIzJSBZjq7TXYFNblD6HreONGL8UdFN0B7Js5mPu2PU3449jYNSOxi
# 1aj/eHNXEcqKcVZ3GDGgtus3ARvFsbEba0UZYIVboKp5i96f8O9YqKvRdBjAjpq1
# IDT/fvI2Fj1mjKi8KuAq8fK5utm+g4NR9WkQX0d661dQNZ3XqVc=
# =NE9v
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 27 Apr 2023 03:49:26 PM BST
# gpg:                using RSA key 1899FF8EDEBF58CCEE034B82F487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>" [undefined]
# gpg:                 aka "Juan Quintela <quintela@trasno.org>" [undefined]
# 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: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* tag 'migration-20230427-pull-request' of https://gitlab.com/juan.quintela/qemu:
  migration: Make dirty_bytes_last_sync atomic
  migration: Make dirty_pages_rate atomic
  stat64: Add stat64_set() operation
  multifd: Only flush once each full round of memory
  multifd: Protect multifd_send_sync_main() calls
  multifd: Create property multifd-flush-after-each-section
  migration: Move migration_properties to options.c
  migration: Create migrate_block_bitmap_mapping() function
  migration: Create migrate_tls_hostname() function
  migration: Create migrate_tls_authz() function
  migration: Create migrate_tls_creds() function
  migration: Remove MigrationState from block_cleanup_parameters()
  migration: Move block_cleanup_parameters() to options.c
  migration: Move migrate_set_block_incremental() to options.c
  migration: Create migrate_downtime_limit() function
  migration: Make all functions check have the same format
  migration: Create migrate_params_init() function
  multifd: Fix the number of channels ready

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

Merge tag 'pull-testing-docs-270423-1' of https://gitlab.com/stsquad/qemu into staging

Testing and documentation updates:

  - bump avocado to 101.0
  - use snapshots for tuxrun baseline tests
  - add sbda-ref test to avocado
  - avoid spurious re-configure in gitlab
  - better description of blockdev options
  - drop FreeBSD 12 from Cirrus CI
  - fix up the ast2[56]00 tests to be more stable
  - improve coverage of ppc64 tests in tuxrun baselines
  - limit plugin tests to just the generic multiarch binaries

# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCgAdFiEEZoWumedRZ7yvyN81+9DbCVqeKkQFAmRKgI0ACgkQ+9DbCVqe
# KkQG9Qf/fvSaiNvkttmb0OFDf3+Qz1uQ33YzLZxETCwI1cSqpGZbssQUrTJZWgpu
# c7FHzdOnTem3Q3PsBY9UN5oqm5IraXEu+nZRO+QyCwHEZzdL9DBfJs46La4BkWG6
# 9vXbNtXLUPd6qJy9ntcZzRShrYy0x0KeszDq7371LS/fng+zMtaIhm4ck0fVWKnj
# htrZEN6nn+CqEnvOc06ICmxiysUVLGRScWKgAHCS9ORGyOtZsj3vWafBoIC6hwzi
# oM3kebhFsVOdbGdL0ZiBdHZqUGAEq3gr+3CpX/48bQ0pYnuYMX8iHk1FhqEK7Adk
# H9ZLnpYUXVyt70rJUjImPyIHpdyq0A==
# =fCif
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 27 Apr 2023 03:02:53 PM BST
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [undefined]
# 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: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* tag 'pull-testing-docs-270423-1' of https://gitlab.com/stsquad/qemu:
  docs/style: call out the use of GUARD macros
  docs/devel: mention the spacing requirement for QOM
  docs/devel: make a statement about includes
  docs/system: remove excessive punctuation from guest-loader docs
  qemu-options.hx: Update descriptions of memory options for NUMA node
  tests/tcg: limit the scope of the plugin tests
  tests/avocado/tuxrun_baselines.py: improve code coverage for ppc64
  avocado_qemu/__init__.py: factor out the qemu-img finding
  MAINTAINERS: Cover tests/avocado/machine_aspeed.py
  tests/avocado/machine_aspeed: Fix the broken ast2[56]00_evb_sdk tests
  tests/avocado: Make ssh_command_output_contains() globally available
  .gitlab-ci.d/cirrus: Drop the CI job for compiling with FreeBSD 12
  qemu-options: finesse the recommendations around -blockdev
  scripts/device-crash-test: Add a parameter to run with TCG only
  gitlab-ci: Avoid to re-run "configure" in the device-crash-test jobs
  tests/avocado: Add set of boot tests on SBSA-ref
  tests/avocado: use the new snapshots for testing
  tests/requirements.txt: bump up avocado-framework version to 101.0

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

migration: Make dirty_bytes_last_sync atomic

As we set its value, it needs to be operated with atomics.
We rename it from remaining to better reflect its meaning.

Statistics always return the real reamaining bytes.  This was used to
store how much pages where dirty on the previous generation, so we can
calculate the expected downtime as: dirty_bytes_last_sync /
current_bandwith.

If we use the actual remaining bytes, we would see a very small value
at the end of the iteration.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
---

I am open to use ram_bytes_remaining() in its only use and be more
"optimistic" about the downtime.

Don't use __nocheck() functions.
Use stat64_get() now that it exists.

migration: Make dirty_pages_rate atomic

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
---

Don't use __nocheck() variants
Use stat64_get()

stat64: Add stat64_set() operation

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Juan Quintela <quintela@redhat.com>

multifd: Only flush once each full round of memory

We need to add a new flag to mean to flush at that point.
Notice that we still flush at the end of setup and at the end of
complete stages.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
---

Add missing qemu_fflush(), now it passes all tests always.
In the previous version, the check that changes the default value to
false got lost in some rebase.  Get it back.

multifd: Protect multifd_send_sync_main() calls

We only need to do that on the ram_save_iterate() call on sending and
on destination when we get a RAM_SAVE_FLAG_EOS.

In setup() and complete() we need to synch in both new and old cases,
so don't add a check there.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
---

Remove the wrappers that we take out on patch 5.

multifd: Create property multifd-flush-after-each-section

We used to flush all channels at the end of each RAM section
sent.  That is not needed, so preparing to only flush after a full
iteration through all the RAM.

Default value of the property is false.  But we return "true" in
migrate_multifd_flush_after_each_section() until we implement the code
in following patches.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
---

Rename each-iteration to after-each-section
Rename multifd-sync-after-each-section to
       multifd-flush-after-each-section
Move to machine-8.0 (peter)

migration: Move migration_properties to options.c

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

migration: Create migrate_block_bitmap_mapping() function

Notice that we changed the test of ->has_block_bitmap_mapping
for the test that block_bitmap_mapping is not NULL.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
---

Make it return const (vladimir)

migration: Create migrate_tls_hostname() function

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
---

Moved the type to const char * (vladimir)

migration: Create migrate_tls_authz() function

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
---

Moved the type to const char * (vladimir)

migration: Create migrate_tls_creds() function

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
---

Moved the type to const char * (vladimir)

migration: Remove MigrationState from block_cleanup_parameters()

This makes the function more regular with everything else.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

migration: Move block_cleanup_parameters() to options.c

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

migration: Move migrate_set_block_incremental() to options.c

Once there, make it more regular and remove the need for
MigrationState parameter.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

migration: Create migrate_downtime_limit() function

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

migration: Make all functions check have the same format

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

migration: Create migrate_params_init() function

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

multifd: Fix the number of channels ready

We don't wait in the sem when we are doing a sync_main.  Make it wait
there.  To make things clearer, we mark the channel ready at the
begining of the thread loop.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Fabiano Rosas <farosas@suse.de>

docs/style: call out the use of GUARD macros

There use makes our code safer so we should mention them.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Message-Id: <20230424092249.58552-19-alex.bennee@linaro.org>

docs/devel: mention the spacing requirement for QOM

We have a more complete document on QOM but we should at least mention
the style requirements in the style guide.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Message-Id: <20230424092249.58552-18-alex.bennee@linaro.org>

docs/devel: make a statement about includes

While not explicitly disallowing header macro abuse (because that
would make us hypocrites) lets at least address some things to think
about.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20230424092249.58552-17-alex.bennee@linaro.org>

docs/system: remove excessive punctuation from guest-loader docs

A possessive its needs no ' whereas the contraction of it is does.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230424092249.58552-16-alex.bennee@linaro.org>

qemu-options.hx: Update descriptions of memory options for NUMA node

This commit adds the following description:
1. `memdev` option is recommended over `mem` option (see [1,2])
2. users must specify memory for all NUMA nodes (see [2])

This commit also separates descriptions for `mem` and `memdev` into two
paragraphs. The old doc describes legacy `mem` option first, and it was
a bit confusing.

Related documentation:
[1] https://wiki.qemu.org/ChangeLog/5.1#Incompatible_changes
[2] https://www.qemu.org/docs/master/about/removed-features.html

Signed-off-by: Yohei Kojima <y-koj@outlook.jp>
Message-Id: <TYZPR06MB5418D6B0175A49E8E76988439D8E9@TYZPR06MB5418.apcprd06.prod.outlook.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
[AJB: fix documentation in commit message]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-15-alex.bennee@linaro.org>

tests/tcg: limit the scope of the plugin tests

Running every plugin with every test is getting excessive as well as
not really improving coverage that much. Restrict the plugin tests to
just the MULTIARCH_TESTS which are shared between most architecture
for both system and user-mode. For those that aren't we need to squash
MULTIARCH_TESTS so we don't add them when they are not part of the
TESTS global.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230424092249.58552-14-alex.bennee@linaro.org>

tests/avocado/tuxrun_baselines.py: improve code coverage for ppc64

Commit c0c8687ef0fd990db8db1655a8a6c5a5e35dd4bb disabled the
boot_linux.py test-case due to which the code coverage for ppc
decreased by around 2%. As per the discussion on
https://lore.kernel.org/qemu-devel/87sfdpqcy4.fsf@linaro.org/ it
was mentioned that the baseline test for ppc64 could be modified
to make up this 2% code coverage. This patch attempts to achieve
this 2% code coverage by adding various device command line
arguments (to ./qemu-system-ppc64) in the tuxrun_baselines.py
test-case.

The code coverage report with boot_linux.py, without it and finally
with these tuxrun_baselines.py changes is as follows:

With boot_linux.py
------------------
  lines......: 13.8% (58006 of 420997 lines)
  functions..: 20.7% (7675 of 36993 functions)
  branches...: 9.2% (22146 of 240611 branches)
Without boot_linux.py (without this patch changes)
--------------------------------------------------
  lines......: 11.9% (50174 of 420997 lines)
  functions..: 18.8% (6947 of 36993 functions)
  branches...: 7.4% (17580 of 239017 branches)
Without boot_linux.py (with this patch changes)
-----------------------------------------------
  lines......: 13.8% (58287 of 420997 lines)
  functions..: 20.7% (7640 of 36993 functions)
  branches...: 8.4% (20223 of 240611 branches)

Rebased on Alex Benee's testing/next branch:
https://gitlab.com/stsquad/qemu/-/tree/testing/next

Signed-off-by: Kautuk Consul <kconsul@linux.vnet.ibm.com>
Reported-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424041830.1275636-1-kconsul@linux.vnet.ibm.com>
Message-Id: <20230424092249.58552-13-alex.bennee@linaro.org>

avocado_qemu/__init__.py: factor out the qemu-img finding

Factor out the code that finds the qemu-img binary in the
QemuSystemTest class and create a new get_qemu_img() function
with it. This function will get called also from the new code
in tuxrun_baselines.py avocado test-case.

Signed-off-by: Kautuk Consul <kconsul@linux.vnet.ibm.com>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Suggested-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230421042322.684093-2-kconsul@linux.vnet.ibm.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-12-alex.bennee@linaro.org>

MAINTAINERS: Cover tests/avocado/machine_aspeed.py

tests/avocado/machine_aspeed.py should belong to the ASPEED section
in the maintainers file. Improve the wildcards here a little bit,
so that it is covered, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20230421110345.1294131-4-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-11-alex.bennee@linaro.org>

tests/avocado/machine_aspeed: Fix the broken ast2[56]00_evb_sdk tests

test_arm_ast2500_evb_sdk and test_arm_ast2600_evb_sdk are currently
failing. The problem is that they are trying to look for the login
prompt that does not have a newline at the end - but the logic in
_console_interaction() only handles full lines. It used to work by
accident in the past since there were sometimes kernel (warning and
error) messages popping up that finally provided a newline character
in the output, but since the tests have been changed to run with the
"quiet" kernel parameter, this is not working anymore.

To make this work reliably, we must not look for the "login:" prompt,
but have to use some text ending with a newline instead. And in the
ast2600 test, switch to ssh instead of trying to log into the serial
console - this works much more reliable and also has the benefit of
excercising the network interface here a little bit, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20230421110345.1294131-3-thuth@redhat.com>
[AJB: remove stray debug log]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-10-alex.bennee@linaro.org>

tests/avocado: Make ssh_command_output_contains() globally available

This function will be useful in other tests, too, so move it to the
core LinuxSSHMixIn class.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20230421110345.1294131-2-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-9-alex.bennee@linaro.org>

.gitlab-ci.d/cirrus: Drop the CI job for compiling with FreeBSD 12

FreeBSD 13.0 has been released in April 2021:

 https://www.freebsd.org/releases/13.0R/announce/

According to QEMU's support policy, we stop supporting the previous
major release two years after the the new major release has been
published. So we can stop testing FreeBSD 12 in our CI now.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230418160225.529172-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Warner Losh <imp@bsdimp.com>
Message-Id: <20230424092249.58552-8-alex.bennee@linaro.org>

qemu-options: finesse the recommendations around -blockdev

We are a bit premature in recommending -blockdev/-device as the best
way to configure block devices. It seems there are times the more
human friendly -drive still makes sense especially when -snapshot is
involved.

Improve the language to hopefully make things clearer.

Suggested-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Cc: Markus Armbruster <armbru@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230424092249.58552-7-alex.bennee@linaro.org>

scripts/device-crash-test: Add a parameter to run with TCG only

We're currently facing the problem that the device-crash-test script
runs twice as long in the CI when a runner supports KVM - which sometimes
results in a timeout of the CI job. To get a more deterministic runtime
here, add an option to the script that allows to run it with TCG only.

Reported-by: Eldon Stegall <eldon-qemu@eldondev.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230414145845.456145-3-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-6-alex.bennee@linaro.org>

gitlab-ci: Avoid to re-run "configure" in the device-crash-test jobs

After "make check-venv" had been added to these jobs, they started
to re-run "configure" each time since our logic in the makefile
thinks that some files are out of date here. Avoid it with the same
trick that we are using in buildtest-template.yml already by disabling
the up-to-date check via NINJA=":".

Fixes: 1d8cf47e5b ("tests: run 'device-crash-test' from tests/venv")
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230414145845.456145-2-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-5-alex.bennee@linaro.org>

tests/avocado: Add set of boot tests on SBSA-ref

This change adds set of boot tests on SBSA-ref machine:

1. boot firmware up to the EDK2 banner
2. boot Alpine Linux

Prebuilt flash volumes are included, built using upstream documentation.

To unify tests for AArch64/virt and AArch64/sbsa-ref we boot
the same Alpine Linux image on both.

Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20230323082813.971535-1-marcin.juszkiewicz@linaro.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Message-Id: <20230328171426.14258-1-philmd@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230424092249.58552-4-alex.bennee@linaro.org>

tests/avocado: use the new snapshots for testing

The tuxboot images now have a stable snapshot URL so we can enable the
checksums and remove the avocado warnings. We will have to update as
old snapshots retire but that won't be too frequent.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Acked-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20230424092249.58552-3-alex.bennee@linaro.org>

tests/requirements.txt: bump up avocado-framework version to 101.0

Avocado version 101.0 has a fix to re-compute the checksum
of an asset file if the algorithm used in the *-CHECKSUM
file isn't the same as the one being passed to it by the
avocado user (i.e. the avocado_qemu python module).
In the earlier avocado versions this fix wasn't there due
to which if the checksum wouldn't match the earlier
checksum (calculated by a different algorithm), the avocado
code would start downloading a fresh image from the internet
URL thus making the test-cases take longer to execute.

Bump up the avocado-framework version to 101.0.

Signed-off-by: Kautuk Consul <kconsul@linux.vnet.ibm.com>
Tested-by: Hariharan T S <hariharan.ts@linux.vnet.ibm.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20230327115030.3418323-2-kconsul@linux.vnet.ibm.com>
Message-Id: <20230424092249.58552-2-alex.bennee@linaro.org>

Merge tag 'migration-20230426-pull-request' of https://gitlab.com/juan.quintela/qemu into staging

Migration Pull request (take2)

With respect to the last PULL request:
- fix compilation on hosts without userfaultd.

Please, apply.

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEGJn/jt6/WMzuA0uC9IfvGFhy1yMFAmRKMYYACgkQ9IfvGFhy
# 1yNoIBAAu4nyERXe7LuRcG02mXQ7N8NtKJXxlhvFGivmlZEzPPkI/SKSXVmzP47E
# uIq1+N+3couYlIof6mY3sjVDwv9Rk7AaGUNUsGv5ygLbXEd/3/qBnqcGV9pUBr3e
# Ls07l2bdb6QPSUGEfu3YATrYzEvWQOAA6AHEBJeqplPlIprgUM3qTlX3bSXrBb26
# J7+p+6A3xdH2koCwmBcWt7cO6IzM7t8pazOS+rJepFri/ZD9aVNr7E3QjXgMUeCs
# 4htBrFaCEXzb80n8O7yPCk90ei0PvGiYBdxI2MGCAy96MN0I5rfMLC/DDFQ+uMH2
# wMVpB5S5YiKrCkoP/ftiYZsSmfRMmQD9aVS6YHR8yt82kEbz616lkXIpvC5Zeads
# J9zTmFuI6fdaF/2PEDzFUACz3YDMtlqQZqKy+e9Ca8gVeoX+ZL4e5yQGNe5nx4BN
# wMr7tPjZenXUpDhuDMcw3/APqaPPZIPVKwYwSgiEL3mqVpU67pre5F4phERCVVVu
# GP/h9yhRa+dkBVNtkqIkK48vQvr7XXisukwSCWmFDA9kfDyDolN6IDaIjv5B7TdL
# DOAW+oz+O4QLVHQhbiQN9WGIc2OenhpKOJ9Rd6cUph6s9tBppaJTrTRRBTYNwPLv
# 8X5h/e7Q1Ee8xGWI1RhAsmD54/e8bu8zxczgf9iMIT8GskLoPco=
# =Ye0/
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 27 Apr 2023 09:25:42 AM BST
# gpg:                using RSA key 1899FF8EDEBF58CCEE034B82F487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>" [undefined]
# gpg:                 aka "Juan Quintela <quintela@trasno.org>" [undefined]
# 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: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* tag 'migration-20230426-pull-request' of https://gitlab.com/juan.quintela/qemu:
  vmstate-static-checker: Recognize "num" field
  migration/vmstate-dump: Dump array size too as "num"
  migration: Allow postcopy_ram_supported_by_host() to report err
  migration: Move qmp_migrate_set_parameters() to options.c
  migration: Move migrate_use_tls() to options.c
  MAINTAINERS: Add Leonardo and Peter as reviewers
  migration: Disable postcopy + multifd migration

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

vmstate-static-checker: Recognize "num" field

Recognize this field for VMS_ARRAY typed vmsd fields, then we can do proper
size matching with previous patch.

Note that this is compatible with old -dump-vmstate output, because when
"num" is not there we'll still use the old "size" only.

Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration/vmstate-dump: Dump array size too as "num"

For VMS_ARRAY typed vmsd fields, also dump the number of entries in the
array in -vmstate-dump.

Without such information, vmstate static checker can report false negatives
of incompatible vmsd on VMS_ARRAY typed fields, when the src/dst do not
have the same type of array defined.  It's because in the checker we only
check against size of fields within a VMSD field.

One example: e1000e used to have a field defined as a boolean array with 5
entries, then removed it and replaced it with UNUSED (in 31e3f318c8b535):

-        VMSTATE_BOOL_ARRAY(core.eitr_intr_pending, E1000EState,
-                           E1000E_MSIX_VEC_NUM),
+        VMSTATE_UNUSED(E1000E_MSIX_VEC_NUM),

It's a legal replacement but vmstate static checker is not happy with it,
because it checks only against the "size" field between the two
fields (here one is BOOL_ARRAY, the other is UNUSED):

For BOOL_ARRAY:

      {
        "field": "core.eitr_intr_pending",
        "version_id": 0,
        "field_exists": false,
        "size": 1
      },

For UNUSED:

      {
        "field": "unused",
        "version_id": 0,
        "field_exists": false,
        "size": 5
      },

It's not the script to blame because there's just not enough information
dumped to show the total size of the entry for an array.  Add it.

Note that this will not break old vmstate checker because the field will
just be ignored.

Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration: Allow postcopy_ram_supported_by_host() to report err

Instead of print it to STDERR, bring the error upwards so that it can be
reported via QMP responses.

E.g.:

{ "execute": "migrate-set-capabilities" ,
  "arguments": { "capabilities":
  [ { "capability": "postcopy-ram", "state": true } ] } }

{ "error":
  { "class": "GenericError",
    "desc": "Postcopy is not supported: Host backend files need to be TMPFS
    or HUGETLBFS only" } }

Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

migration: Move qmp_migrate_set_parameters() to options.c

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>

migration: Move migrate_use_tls() to options.c

Once there, rename it to migrate_tls() and make it return bool for
consistency.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
---

Fix typos found by fabiano

MAINTAINERS: Add Leonardo and Peter as reviewers

Now that David has stepped down with Migration maintainership,
Leonardo and Peter has volunteer to review the migration patches.
This way they got CC'd on every migration patch.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
Acked-by: Leonardo Bras <leobras@redhat.com>

migration: Disable postcopy + multifd migration

Since the introduction of multifd, it's possible to perform a multifd
migration and finish it using postcopy.

A bug introduced by yank (fixed on cfc3bcf373) was previously preventing
a successful use of this migration scenario, and now thing should be
working on most scenarios.

But since there is not enough testing/support nor any reported users for
this scenario, we should disable this combination before it may cause any
problems for users.

Suggested-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Leonardo Bras <leobras@redhat.com>
Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>

Merge tag 'pull-qapi-2023-04-26' of https://repo.or.cz/qemu/armbru into staging

QAPI patches patches for 2023-04-26

# -----BEGIN PGP SIGNATURE-----
#
# iQJGBAABCAAwFiEENUvIs9frKmtoZ05fOHC0AOuRhlMFAmRIvOkSHGFybWJydUBy
# ZWRoYXQuY29tAAoJEDhwtADrkYZTX3MQAIqrKQbOzQ81/cDZ7aeLOroDoGYf1Cs0
# 0NiEVlyoblWNzL3HraGgXiNRTP+zaG/TcFKza1nz8qjdkxWxBjdbfF5Lm6mQf5Zo
# tcHUjksmnUlPkLYSOyEjfY9SNvcS6g7djE/NF5lbJtzYGZScZpLarELR7oUvrcXR
# AEiw8N5FZXp+j6cTeWvrLzSqz9qBsFJUCGcGER0T/Mt5MlUwDexE1xe7g8oD5l+b
# s0jeQr1PTZm5k6ehajQtgbHvAkgH8xVTKqbB/U5iz4VhYriH+IPEOtfCFt6/1soz
# pVkYikJpazCvQMjqnWu9dE1onthgSsEIOV29kFU0Kr8ATZuJBQMuLVp4hSsbKANj
# BUVyL2/fUsIp7gd+KikXUOjKYajxek6Q2YLAPpL+1pBCTql/PBQ7td8CECdiv/9e
# Xh50q+BGvyEiyoyf4EEpaLXUZog605WHEaODj9uPtNHJP9x6Rqt93FUsdWUtt/k9
# hJ8RSKy8njr0vxGoJkj89m2XfCwtuX3VQ5IXvv/If4U5Y4+JhcLtiqW+Njh8fAM4
# ZwIrlUYG7inLUKFVcQ3sEGpaj611i5ixIxctUvEiggZX+fPeSFKYUr+Rq8WXM8gv
# suLXz7VF6H4Sw30lCvdQ4LSungbzlYAtQYpmdEQGoM8iasIi4PoDf0cTYBbMYHDX
# +pZvWC50cVtf
# =wLx6
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 26 Apr 2023 06:55:53 AM BST
# gpg:                using RSA key 354BC8B3D7EB2A6B68674E5F3870B400EB918653
# gpg:                issuer "armbru@redhat.com"
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>" [undefined]
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>" [undefined]
# 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: 354B C8B3 D7EB 2A6B 6867  4E5F 3870 B400 EB91 8653

* tag 'pull-qapi-2023-04-26' of https://repo.or.cz/qemu/armbru:
  qapi: allow unions to contain further unions
  qapi: Improve specificity of type/member descriptions
  qapi: support updating expected test output via make
  qapi: Require boxed for conditional command and event arguments
  qapi: Fix code generated for optional conditional struct member
  tests/qapi-schema: Cover optional conditional struct member
  tests/qapi-schema: Clean up positive test for conditionals
  tests/qapi-schema: Rename a few conditionals
  tests/qapi-schema: Improve union discriminator coverage
  qapi: Fix to reject 'data': 'mumble' in struct
  qapi: Fix error message when type name or array is expected
  qapi: Simplify code a bit after previous commits
  qapi: Improve error message for unexpected array types
  qapi: Split up check_type()
  qapi: Clean up after removal of simple unions
  qapi/schema: Use super()
  qapi: Fix error message format regression

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

Merge tag 'for-upstream' of https://repo.or.cz/qemu/kevin into staging

Block layer patches

- Protect BlockBackend.queued_requests with its own lock
- Switch to AIO_WAIT_WHILE_UNLOCKED() where possible
- AioContext removal: LinuxAioState/LuringState/ThreadPool
- Add more coroutine_fn annotations, use bdrv/blk_co_*
- Fix crash when execute hmp_commit

# -----BEGIN PGP SIGNATURE-----
#
# iQJFBAABCAAvFiEE3D3rFZqa+V09dFb+fwmycsiPL9YFAmRH0b0RHGt3b2xmQHJl
# ZGhhdC5jb20ACgkQfwmycsiPL9Y0yw/6A/vzA4TGgFUP3WIvH/sQri4/V3gyR+PT
# u3hOQUCYZ99nioTpKV91TSuUPuU/Mdspy/0NKM+K92yIXqxa9172A2zLOsGOu21l
# qKpse+nBf1zqEgB8YzUHyCBdetPz916C/f9RS26SNUCW85GCHYGHA3u7nKvWLMyV
# oKIoTlA8QOglOuEKlRoYh7hCFm7ET51NOSEftm8GsYbsW/I2Vzl8a1SHN1lHufjd
# We3+898zUrmFqNMp6Rjdhn+yZmmoGzoZqV4YQi83z7xjiv+Ms4VHVVW7X8d20xRX
# 5BLFiLHAuZ/1d26HyVhgBUr7KHyf94odocz8BylWKXGl5SXMCZun1Td1vgVKlGK+
# GRxzB2cWGWqzC2UmqSTc0Z0aIWbXukKwvcX76uBKsQZ+kB2A7jFobxHiaoQEDJ8B
# WRNEMH2+CqCAu9rsrNRinnJKhT2nXcr9F9YfwRIlagdAePGWin+EUW8huf14dDBm
# Z2Y34aKW4RQibF8xirMHeRBbOLmcq2VpKLKwNfBHUDgZB8iuD7bLn4n9nwWXMG1w
# zgNsTybkv46vLPamTpEaUoNTHfuRDTAuE7Z7lkcc7jF41Z0V1DC/DCCWcL/0LvhP
# GIxFdkYug3hetdF2U/OZhUoEfxvkqcuBnrr55LFzqheKEllQpPwPpt7UF0aH8bg3
# i/YpjHsf3xU=
# =mpYX
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 25 Apr 2023 02:12:29 PM BST
# gpg:                using RSA key DC3DEB159A9AF95D3D7456FE7F09B272C88F2FD6
# gpg:                issuer "kwolf@redhat.com"
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]

* tag 'for-upstream' of https://repo.or.cz/qemu/kevin: (25 commits)
  block/monitor: Fix crash when executing HMP commit
  vmdk: make vmdk_is_cid_valid a coroutine_fn
  qcow2: mark various functions as coroutine_fn and GRAPH_RDLOCK
  tests: mark more coroutine_fns
  qemu-pr-helper: mark more coroutine_fns
  9pfs: mark more coroutine_fns
  nbd: mark more coroutine_fns, do not use co_wrappers
  mirror: make mirror_flush a coroutine_fn, do not use co_wrappers
  blkdebug: add missing coroutine_fn annotation
  vvfat: mark various functions as coroutine_fn
  thread-pool: avoid passing the pool parameter every time
  thread-pool: use ThreadPool from the running thread
  io_uring: use LuringState from the running thread
  linux-aio: use LinuxAioState from the running thread
  block: add missing coroutine_fn to bdrv_sum_allocated_file_size()
  include/block: fixup typos
  monitor: convert monitor_cleanup() to AIO_WAIT_WHILE_UNLOCKED()
  hmp: convert handle_hmp_command() to AIO_WAIT_WHILE_UNLOCKED()
  block: convert bdrv_drain_all_begin() to AIO_WAIT_WHILE_UNLOCKED()
  block: convert bdrv_graph_wrlock() to AIO_WAIT_WHILE_UNLOCKED()
  ...

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

qapi: allow unions to contain further unions

This extends the QAPI schema validation to permit unions inside unions,
provided the checks for clashing fields pass.

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20230420102619.348173-4-berrange@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qapi: Improve specificity of type/member descriptions

Error messages describe object members, enumeration values, features,
and variants like ROLE 'NAME', where ROLE is "member", "value",
"feature", or "branch", respectively.  When the member is defined in
another type, e.g. inherited from a base type, we add "of type
'TYPE'".  Example: test case struct-base-clash-deep reports a member
of type 'Sub' clashing with a member of its base type 'Base' as

    struct-base-clash-deep.json: In struct 'Sub':
    struct-base-clash-deep.json:10: member 'name' collides with member 'name' of type 'Base'

Members of implicitly defined types need special treatment.  We don't
want to add "of type 'TYPE'" for them, because their named are made up
and mean nothing to the user.  Instead, we describe members of an
implicitly defined base type as "base member 'NAME'", and command and
event parameters as "parameter 'NAME'".  Example: test case
union-bad-base reports member of a variant's type clashing with a
member of its implicitly defined base type as

    union-bad-base.json: In union 'TestUnion':
    union-bad-base.json:8: member 'string' of type 'TestTypeA' collides with base member 'string'

The next commit will permit unions as variant types.  "base member
'NAME' would then be ambigious: is it the union's base, or is it the
union's variant's base?  One of its test cases would report a clash
between two such bases as "base member 'type' collides with base
member 'type'".  Confusing.

Refine the special treatment: add "of TYPE" even for implicitly
defined types, but massage TYPE and ROLE so they make sense for the
user.

Message-Id: <20230420102619.348173-3-berrange@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

qapi: support updating expected test output via make

It is possible to pass --update to tests/qapi-schema/test-qapi.py
to make it update the output files on error. This is inconvenient
to achieve though when test-qapi.py is run indirectly by make/meson.

Instead simply allow for an env variable to be set:

 $ QAPI_TEST_UPDATE= make check-qapi-schema

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20230420102619.348173-2-berrange@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>

block/monitor: Fix crash when executing HMP commit

hmp_commit() calls blk_is_available() from a non-coroutine context (and
in the main loop). blk_is_available() is a co_wrapper_mixed_bdrv_rdlock
function, and in the non-coroutine context it calls AIO_WAIT_WHILE(),
which crashes if the aio_context lock is not taken before.

Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1615
Signed-off-by: Wang Liang <wangliangzz@inspur.com>
Message-Id: <20230424103902.45265-1-wangliangzz@126.com>
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

vmdk: make vmdk_is_cid_valid a coroutine_fn

Functions that can do I/O are prime candidates for being coroutine_fns.  Make the
change for the one that is itself called only from coroutine_fns.  Unfortunately
vmdk does not use a coroutine_fn for the bulk of the open (like qcow2 does) so
vmdk_read_cid cannot have the same treatment.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-10-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qcow2: mark various functions as coroutine_fn and GRAPH_RDLOCK

Functions that can do I/O (including calling bdrv_is_allocated
and bdrv_block_status functions) are prime candidates for being
coroutine_fns.  Make the change for those that are themselves called
only from coroutine_fns.  Also annotate that they are called with the
graph rdlock taken, thus allowing them to call bdrv_co_*() functions
for I/O.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-9-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

tests: mark more coroutine_fns

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-8-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu-pr-helper: mark more coroutine_fns

do_sgio can suspend via the coroutine function thread_pool_submit_co, so it
has to be coroutine_fn as well---and the same is true of all its direct and
indirect callers.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-7-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

9pfs: mark more coroutine_fns

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-6-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

nbd: mark more coroutine_fns, do not use co_wrappers

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

mirror: make mirror_flush a coroutine_fn, do not use co_wrappers

mirror_flush calls a mixed function blk_flush but it is only called
from mirror_run; so call the coroutine version and make mirror_flush
a coroutine_fn too.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-4-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

blkdebug: add missing coroutine_fn annotation

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-3-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

vvfat: mark various functions as coroutine_fn

Functions that can do I/O are prime candidates for being coroutine_fns.  Make the
change for those that are themselves called only from coroutine_fns.

In addition, coroutine_fns should do I/O using bdrv_co_*() functions, for
which it is required to hold the BlockDriverState graph lock.  So also nnotate
functions on the I/O path with TSA attributes, making it possible to
switch them to use bdrv_co_*() functions.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230309084456.304669-2-pbonzini@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

thread-pool: avoid passing the pool parameter every time

thread_pool_submit_aio() is always called on a pool taken from
qemu_get_current_aio_context(), and that is the only intended
use: each pool runs only in the same thread that is submitting
work to it, it can't run anywhere else.

Therefore simplify the thread_pool_submit* API and remove the
ThreadPool function parameter.

Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20230203131731.851116-5-eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

thread-pool: use ThreadPool from the running thread

Use qemu_get_current_aio_context() where possible, since we always
submit work to the current thread anyways.

We want to also be sure that the thread submitting the work is
the same as the one processing the pool, to avoid adding
synchronization to the pool list.

Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20230203131731.851116-4-eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

io_uring: use LuringState from the running thread

Remove usage of aio_context_acquire by always submitting asynchronous
AIO to the current thread's LuringState.

In order to prevent mistakes from the caller side, avoid passing LuringState
in luring_io_{plug/unplug} and luring_co_submit, and document the functions
to make clear that they work in the current thread's AioContext.

Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20230203131731.851116-3-eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

linux-aio: use LinuxAioState from the running thread

Remove usage of aio_context_acquire by always submitting asynchronous
AIO to the current thread's LinuxAioState.

In order to prevent mistakes from the caller side, avoid passing LinuxAioState
in laio_io_{plug/unplug} and laio_co_submit, and document the functions
to make clear that they work in the current thread's AioContext.

Signed-off-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20230203131731.851116-2-eesposit@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: add missing coroutine_fn to bdrv_sum_allocated_file_size()

Not a coroutine_fn, you say?

  static int64_t bdrv_sum_allocated_file_size(BlockDriverState *bs)
  {
      BdrvChild *child;
      int64_t child_size, sum = 0;

      QLIST_FOREACH(child, &bs->children, next) {
          if (child->role & (BDRV_CHILD_DATA | BDRV_CHILD_METADATA |
                             BDRV_CHILD_FILTERED))
          {
              child_size = bdrv_co_get_allocated_file_size(child->bs);
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Well what do we have here?!

I rest my case, your honor.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230308211435.346375-1-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

include/block: fixup typos

Fixup a few minor typos

Signed-off-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Message-Id: <20230313003744.55476-1-wilfred.mallawa@opensource.wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

monitor: convert monitor_cleanup() to AIO_WAIT_WHILE_UNLOCKED()

monitor_cleanup() is called from the main loop thread. Calling
AIO_WAIT_WHILE(qemu_get_aio_context(), ...) from the main loop thread is
equivalent to AIO_WAIT_WHILE_UNLOCKED(NULL, ...) because neither unlocks
the AioContext and the latter's assertion that we're in the main loop
succeeds.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230309190855.414275-7-stefanha@redhat.com>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

hmp: convert handle_hmp_command() to AIO_WAIT_WHILE_UNLOCKED()

The HMP monitor runs in the main loop thread. Calling
AIO_WAIT_WHILE(qemu_get_aio_context(), ...) from the main loop thread is
equivalent to AIO_WAIT_WHILE_UNLOCKED(NULL, ...) because neither unlocks
the AioContext and the latter's assertion that we're in the main loop
succeeds.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230309190855.414275-6-stefanha@redhat.com>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: convert bdrv_drain_all_begin() to AIO_WAIT_WHILE_UNLOCKED()

Since the AioContext argument was already NULL, AIO_WAIT_WHILE() was
never going to unlock the AioContext. Therefore it is possible to
replace AIO_WAIT_WHILE() with AIO_WAIT_WHILE_UNLOCKED().

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230309190855.414275-5-stefanha@redhat.com>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: convert bdrv_graph_wrlock() to AIO_WAIT_WHILE_UNLOCKED()

The following conversion is safe and does not change behavior:

     GLOBAL_STATE_CODE();
     ...
  -  AIO_WAIT_WHILE(qemu_get_aio_context(), ...);
  +  AIO_WAIT_WHILE_UNLOCKED(NULL, ...);

Since we're in GLOBAL_STATE_CODE(), qemu_get_aio_context() is our home
thread's AioContext. Thus AIO_WAIT_WHILE() does not unlock the
AioContext:

  if (ctx_ && in_aio_context_home_thread(ctx_)) {                \
      while ((cond)) {                                           \
          aio_poll(ctx_, true);                                  \
          waited_ = true;                                        \
      }                                                          \

And that means AIO_WAIT_WHILE_UNLOCKED(NULL, ...) can be substituted.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230309190855.414275-4-stefanha@redhat.com>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: convert blk_exp_close_all_type() to AIO_WAIT_WHILE_UNLOCKED()

There is no change in behavior. Switch to AIO_WAIT_WHILE_UNLOCKED()
instead of AIO_WAIT_WHILE() to document that this code has already been
audited and converted. The AioContext argument is already NULL so
aio_context_release() is never called anyway.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230309190855.414275-3-stefanha@redhat.com>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: don't acquire AioContext lock in bdrv_drain_all()

There is no need for the AioContext lock in bdrv_drain_all() because
nothing in AIO_WAIT_WHILE() needs the lock and the condition is atomic.

AIO_WAIT_WHILE_UNLOCKED() has no use for the AioContext parameter other
than performing a check that is nowadays already done by the
GLOBAL_STATE_CODE()/IO_CODE() macros. Set the ctx argument to NULL here
to help us keep track of all converted callers. Eventually all callers
will have been converted and then the argument can be dropped entirely.

Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230309190855.414275-2-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: protect BlockBackend->queued_requests with a lock

The CoQueue API offers thread-safety via the lock argument that
qemu_co_queue_wait() and qemu_co_enter_next() take. BlockBackend
currently does not make use of the lock argument. This means that
multiple threads submitting I/O requests can corrupt the CoQueue's
QSIMPLEQ.

Add a QemuMutex and pass it to CoQueue APIs so that the queue is
protected. While we're at it, also assert that the queue is empty when
the BlockBackend is deleted.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-Id: <20230307210427.269214-4-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: make BlockBackend->disable_request_queuing atomic

This field is accessed by multiple threads without a lock. Use explicit
qatomic_read()/qatomic_set() calls. There is no need for acquire/release
because blk_set_disable_request_queuing() doesn't provide any
guarantees (it helps that it's used at BlockBackend creation time and
not when there is I/O in flight).

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-Id: <20230307210427.269214-3-stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

block: make BlockBackend->quiesce_counter atomic

The main loop thread increments/decrements BlockBackend->quiesce_counter
when drained sections begin/end. The counter is read in the I/O code
path. Therefore this field is used to communicate between threads
without a lock.

Acquire/release are not necessary because the BlockBackend->in_flight
counter already uses sequentially consistent accesses and running I/O
requests hold that counter when blk_wait_while_drained() is called.
qatomic_read() can be used.

Use qatomic_fetch_inc()/qatomic_fetch_dec() for modifications even
though sequentially consistent atomic accesses are not strictly required
here. They are, however, nicer to read than multiple calls to
qatomic_read() and qatomic_set(). Since beginning and ending drain is
not a hot path the extra cost doesn't matter.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230307210427.269214-2-stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

Merge tag 'for_upstream' of https://git./virt/kvm/mst/qemu into staging

virtio,pc,pci: fixes, features, cleanups

Mostly just fixes, cleanups all over the place.
Some optimizations.
More control over slot_reserved_mask.
More feature bits supported for SVQ.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# -----BEGIN PGP SIGNATURE-----
#
# iQFDBAABCAAtFiEEXQn9CHHI+FuUyooNKB8NuNKNVGkFAmRHQvAPHG1zdEByZWRo
# YXQuY29tAAoJECgfDbjSjVRpQc0H/RD+RXy7IAnmhkdCyjj0hM8pftPTwCJfrSCW
# DLHP4c5jiKO5ngUoAv3YJdM77TBCXlJn6gceeKBrzhGUTtJ7dTLC+Udeq/jW43EF
# /E2ldLLbTNFyUqW8yX7D+EVio7Jy4zXTHpczKCF5vO7MaVWS/b3QdCpmjXpEHLNb
# janv24vQHHgmRwK96uIdIauJJT8aqYW0arn1po8anxuFS8ok9Tf8LTEF5uBHokJP
# MriTwMaqMgRK+4rzh+b6wc7QC5GqIr44gFrsfFYuNOUY0+BizvGvUAtMt+B/XZwt
# OF4RSShUh2bhsQoYwgvShfEsR/vWwOl3yMAhcsB+wMgMzMG8MUQ=
# =e8DF
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 25 Apr 2023 04:03:12 AM BST
# gpg:                using RSA key 5D09FD0871C8F85B94CA8A0D281F0DB8D28D5469
# gpg:                issuer "mst@redhat.com"
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [undefined]
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>" [undefined]
# 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: 0270 606B 6F3C DF3D 0B17  0970 C350 3912 AFBE 8E67
#      Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA  8A0D 281F 0DB8 D28D 5469

* tag 'for_upstream' of https://git.kernel.org/pub/scm/virt/kvm/mst/qemu: (31 commits)
  hw/pci-bridge: Make PCIe and CXL PXB Devices inherit from TYPE_PXB_DEV
  hw/pci-bridge: pci_expander_bridge fix type in pxb_cxl_dev_reset()
  docs/specs: Convert pci-testdev.txt to rst
  docs/specs: Convert pci-serial.txt to rst
  docs/specs/pci-ids: Convert from txt to rST
  acpi: pcihp: allow repeating hot-unplug requests
  virtio: i2c: Check notifier helpers for VIRTIO_CONFIG_IRQ_IDX
  docs: Remove obsolete descriptions of SR-IOV support
  intel_iommu: refine iotlb hash calculation
  docs/cxl: Fix sentence
  MAINTAINERS: Add Eugenio Pérez as vhost-shadow-virtqueue reviewer
  tests: bios-tables-test: replace memset with initializer
  hw/acpi: limit warning on acpi table size to pc machines older than version 2.3
  Add my old and new work email mapping and use work email to support acpi
  vhost-user-blk-server: notify client about disk resize
  pci: avoid accessing slot_reserved_mask directly outside of pci.c
  hw: Add compat machines for 8.1
  hw/i386/amd_iommu: Factor amdvi_pci_realize out of amdvi_sysbus_realize
  hw/i386/amd_iommu: Set PCI static/const fields via PCIDeviceClass
  hw/i386/amd_iommu: Move capab_offset from AMDVIState to AMDVIPCIState
  ...

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

Merge tag 'block-pull-request' of https://gitlab.com/stefanha/qemu into staging

Pull request (v2)

I dropped the zoned storage patches that had CI failures. This pull request
only contains fixes now.

# -----BEGIN PGP SIGNATURE-----
#
# iQEzBAABCAAdFiEEhpWov9P5fNqsNXdanKSrs4Grc8gFAmRGwoIACgkQnKSrs4Gr
# c8jyeAf+JqYgJ0sG7rkYwGNUnmdyzjcZh8sokmIoR0a1VbFjpG54QqvvsSdrtZ5V
# +XdYPgn3ASqyw6/737chw8BBQ1YQEtALCXCk9vx2/Vpmmd6NUkJcxTLvV17o8wit
# asMO14R8p4b/9NUbRnMA+OcnJ1R0y0nKaw0fK4v2DjCN3Uy7b7zJ5Xb5Mg1lkrN6
# RCB6uobWwN58LqY90lN2m1EMvBy0hhI5P1wDjcuZjG1/GnnvVzrIVMOC0ddiLgRH
# 0vh9MvdSjOwkdM1J9eJyeijfyEjKS9e/E/c4WrjUs+gH5wcb9zNMNFOGeG7Ih2U0
# wEUIpe2y3YjSHTSTGZLYGguCGmJPZQ==
# =MKCk
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 24 Apr 2023 06:55:14 PM BST
# gpg:                using RSA key 8695A8BFD3F97CDAAC35775A9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>" [full]
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>" [full]

* tag 'block-pull-request' of https://gitlab.com/stefanha/qemu:
  tracetool: use relative paths for '#line' preprocessor directives
  block/dmg: Declare a type definition for DMG uncompress function

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

hw/pci-bridge: Make PCIe and CXL PXB Devices inherit from TYPE_PXB_DEV

Previously, PXB_CXL_DEVICE, PXB_PCIE_DEVICE and PXB_DEVICE all
have PCI_DEVICE as their direct parent but share a common state
struct PXBDev. convert_to_pxb() is used to get the PXBDev
instance from which ever of these types it is called on.

This patch switches to an explicit hierarchy based on shared
functionality.  To allow use of OBJECT_DECLARE_SIMPLE_TYPE()
whilst minimizing code changes, all types are renamed to have
the postfix _DEV rather than _DEVICE.  The new heirarchy
has PXB_CXL_DEV with parent PXB_PCIE_DEV which in turn
has parent PXB_DEV which continues to have parent PCI_DEVICE.

This allows simple use of PXB_DEV() etc rather than a custom function
+ removal of duplicated properties and moving the CXL specific
elements out of struct PXBDev.

Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20230420142750.6950-3-Jonathan.Cameron@huawei.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/pci-bridge: pci_expander_bridge fix type in pxb_cxl_dev_reset()

Reproduce issue with

configure --enable-qom-cast-debug ...

qemu-system-x86_64 -display none -machine q35,cxl=on -device pxb-cxl,bus=pcie.0

  hw/pci-bridge/pci_expander_bridge.c:54:PXB_DEV: Object 0x5570e0b1ada0 is not an instance of type pxb
  Aborted

The type conversion results in the right state structure, but PXB_DEV is
not a parent of PXB_CXL_DEV hence the error. Rather than directly
cleaning up the inheritance, this is the minimal fix which will be
followed by the cleanup.

Fixes: 154070eaf6 ("hw/pxb-cxl: Support passthrough HDM Decoders unless overridden")
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Message-Id: <20230420142750.6950-2-Jonathan.Cameron@huawei.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

docs/specs: Convert pci-testdev.txt to rst

Convert pci-testdev.txt to reStructuredText. Includes
some minor wordsmithing.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20230420160334.1048224-4-peter.maydell@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

docs/specs: Convert pci-serial.txt to rst

Convert pci-serial.txt to reStructuredText. This includes
some wordsmithing, and the correction of the docs to note
that the Windows inf file includes 2x and 4x support
(as it has done since commit dc9528fdf9f61 in 2014).

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20230420160334.1048224-3-peter.maydell@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>

docs/specs/pci-ids: Convert from txt to rST

Convert the pci-ids document from plain text to reStructuredText.

I opted to use definition-lists here because rST tables are
super-clunky, and actually formatting these as tables didn't
seem necessary.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20230420160334.1048224-2-peter.maydell@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

acpi: pcihp: allow repeating hot-unplug requests

with Q35 using ACPI PCI hotplug by default, user's request to unplug
device is ignored when it's issued before guest OS has been booted.
And any additional attempt to request device hot-unplug afterwards
results in following error:

  "Device XYZ is already in the process of unplug"

arguably it can be considered as a regression introduced by [2],
before which it was possible to issue unplug request multiple
times.

Accept new uplug requests after timeout (1ms). This brings ACPI PCI
hotplug on par with native PCIe unplug behavior [1] and allows user
to repeat unplug requests at propper times.
Set expire timeout to arbitrary 1msec so user won't be able to
flood guest with SCI interrupts by calling device_del in tight loop.

PS:
ACPI spec doesn't mandate what OSPM can do with GPEx.status
bits set before it's booted => it's impl. depended.
Status bits may be retained (I tested with one Windows version)
or cleared (Linux since 2.6 kernel times) during guest's ACPI
subsystem initialization.
Clearing status bits (though not wrong per se) hides the unplug
event from guest, and it's upto user to repeat device_del later
when guest is able to handle unplug requests.

1) 18416c62e3 ("pcie: expire pending delete")
2)
Fixes: cce8944cc9ef ("qdev-monitor: Forbid repeated device_del")
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
CC: mst@redhat.com
CC: anisinha@redhat.com
CC: jusual@redhat.com
CC: kraxel@redhat.com
Message-Id: <20230418090449.2155757-1-imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Ani Sinha <anisinha@redhat.com>

virtio: i2c: Check notifier helpers for VIRTIO_CONFIG_IRQ_IDX

Since the driver doesn't support interrupts, we must return early when
index is set to VIRTIO_CONFIG_IRQ_IDX.

Fixes: 544f0278afca ("virtio: introduce macro VIRTIO_CONFIG_IRQ_IDX")
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Message-Id: <d53ec8bc002001eafac597f6bd9a8812df989257.1681790067.git.viresh.kumar@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

docs: Remove obsolete descriptions of SR-IOV support

The documentation used to say there is no device implemented with
SR-IOV, but igb and nvme support SR-IOV today.

Signed-off-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-Id: <20230414090441.23156-1-akihiko.odaki@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

intel_iommu: refine iotlb hash calculation

Commit 1b2b12376c8 ("intel-iommu: PASID support") takes PASID into
account when calculating iotlb hash like:

static guint vtd_iotlb_hash(gconstpointer v)
{
    const struct vtd_iotlb_key *key = v;

    return key->gfn | ((key->sid) << VTD_IOTLB_SID_SHIFT) |
           (key->level) << VTD_IOTLB_LVL_SHIFT |
           (key->pasid) << VTD_IOTLB_PASID_SHIFT;
}

This turns out to be problematic since:

- the shift will lose bits if not converting to uint64_t
- level should be off by one in order to fit into 2 bits
- VTD_IOTLB_PASID_SHIFT is 30 but PASID is 20 bits which will waste
  some bits
- the hash result is uint64_t so we will lose bits when converting to
  guint

So this patch fixes them by

- converting the keys into uint64_t before doing the shift
- off level by one to make it fit into two bits
- change the sid, lvl and pasid shift to 26, 42 and 44 in order to
  take the full width of uint64_t
- perform an XOR to the top 32bit with the bottom 32bit for the final
  result to fit guint

Fixes: Coverity CID 1508100
Fixes: 1b2b12376c8 ("intel-iommu: PASID support")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230412073510.7158-1-jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>

docs/cxl: Fix sentence

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Message-Id: <20230409201828.1159568-1-sw@weilnetz.de>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

MAINTAINERS: Add Eugenio Pérez as vhost-shadow-virtqueue reviewer

I'd like to be notified on SVQ patches and review them.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20230331150410.2627214-1-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>

tests: bios-tables-test: replace memset with initializer

Coverity complains that memset() writes over a const field.  Use
an initializer instead, so that the const field is left to zero.
Tests that have to write the const field already use an initializer
for the whole struct, here I am choosing the smallest possible
patch (which is not that small already).

Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20230330131109.47856-1-pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>

hw/acpi: limit warning on acpi table size to pc machines older than version 2.3

i440fx machine versions 2.3 and newer supports dynamic ram
resizing. See commit a1666142db6233 ("acpi-build: make ROMs RAM blocks resizeable") .
Currently supported all q35 machine types (versions 2.4 and newer) supports
resizable RAM/ROM blocks.Therefore the warning generated when the ACPI table
size exceeds a pre-defined value does not apply to those machine versions.
Add a check limiting the warning message to only those machines that does not
support expandable ram blocks (that is, i440fx machines with version 2.2
and older).

Signed-off-by: Ani Sinha <anisinha@redhat.com>
Message-Id: <20230329045726.14028-1-anisinha@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Add my old and new work email mapping and use work email to support acpi

Updating mailmap to indicate ani@anisinha.ca and anisinha@redhat.com are one
and the same person. Also updating my email in MAINTAINERS for all my acpi work
(reviewing patches and biosbits) to my work email. Also doing the same for
bios bits test framework documentation.

Signed-off-by: Ani Sinha <anisinha@redhat.com>
Message-Id: <20230329040834.11973-1-anisinha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tracetool: use relative paths for '#line' preprocessor directives

The event filename is an absolute path. Convert it to a relative path when
writing '#line' directives, to preserve reproducibility of the generated
output when different base paths are used.

Signed-off-by: Thomas De Schampheleire <thomas.de_schampheleire@nokia.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230406080045.21696-1-thomas.de_schampheleire@nokia.com>

block/dmg: Declare a type definition for DMG uncompress function

Introduce the BdrvDmgUncompressFunc type defintion. To emphasis
dmg_uncompress_bz2 and dmg_uncompress_lzfse are pointer to functions,
declare them using this new typedef.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-id: 20230320152610.32052-1-philmd@linaro.org
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Merge tag 'migration-20230424-pull-request' of https://gitlab.com/juan.quintela/qemu into staging

Migration Pull request

Everything that was reviewed since last PULL request:
- fix to control flow (eric)
- rearrange of hmp commands (juan)
- Make capabilities more consistent and coherent (juan)
  Not all of them reviewed yet, so only the ones reviewed.

Later, Juan.

PD.  I am waiting to finish review of the compression fixes to send
them.

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEGJn/jt6/WMzuA0uC9IfvGFhy1yMFAmRGf0kACgkQ9IfvGFhy
# 1yNojRAAhvOZoYRpTizbSo7wOr2TpO4F1R/opWECubBvw3/yTo3ifsVjovRo/nxe
# zmCXnY/ykqxskb/OMNyLZQYG4MPT24nyf7dhT3VxcFjTYsaNU59yvJoBtvy7Oq5h
# 3Yk0459eKmsP39IXFlAMMRBwX+Hu3aE/8xAAHOowAhkmDkMFYjf7I1Lxpuarbhp+
# k9O5eqdNchG21YUzWXUe4ivAWqZmvzXtkwCp+XJ/KizjIEKgsm8HO1nwm5mtpmnu
# SS9Kkf957jYHqK73YXQhUV+iQ0kCVpclBPfZc2KuzudPi/aMG6LEVKfV//z5KIAz
# amME/6D1oSBfpgtqoCCPELdNfZOz+ZIa+XJzXlWkuiBDu9yNpUP339EVClmStwFu
# 1UAOJIs8VUjPr9zTItgCDjZ17nh4Q0I04aMGuxgQIu82e8CTgS/QrnH2Tr2lUoMO
# QLgYAetVIDGVVFAA9Clba4C7AbS5hBeWMnd9Qd4cP93d6Z/C65xUv0a9mI7edpMb
# RNbvg73ZZCb6tye25cPSr07VaGTS+TELVMEh9RX3KZrfMTsHfGQ/ZHZv9wqJrQ04
# 0wCidqBIbBk+BN8AtJ9vwqtPpL/Nf/BwDKPiwOVuZHCcrP+veXtlKb00SwNpJwkN
# x3Ld4cq22ZLeqO4dMueK16ij0ZpuXsF7jM/ptEvxrw6oxh/6xYQ=
# =g5gx
# -----END PGP SIGNATURE-----
# gpg: Signature made Mon 24 Apr 2023 02:08:25 PM BST
# gpg:                using RSA key 1899FF8EDEBF58CCEE034B82F487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>" [undefined]
# gpg:                 aka "Juan Quintela <quintela@trasno.org>" [undefined]
# 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: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* tag 'migration-20230424-pull-request' of https://gitlab.com/juan.quintela/qemu: (30 commits)
  migration: Create migrate_max_bandwidth() function
  migration: Move migrate_postcopy() to options.c
  migration: Create migrate_cpu_throttle_tailslow() function
  migration: Create migrate_cpu_throttle_increment() function
  migration: Create  migrate_cpu_throttle_initial() to option.c
  migration: Move migrate_announce_params() to option.c
  migration: Create migrate_max_cpu_throttle()
  migration: Create migrate_checkpoint_delay()
  migration: Create migrate_throttle_trigger_threshold()
  migration: Move migrate_use_block_incremental() to option.c
  migration: Use migrate_max_postcopy_bandwidth()
  migration: Move parameters functions to option.c
  migration: Move migrate_cap_set() to options.c
  migration: Move qmp_migrate_set_capabilities() to options.c
  migration: Move qmp_query_migrate_capabilities() to options.c
  migration: Move migrate_caps_check() to options.c
  migration: Create migrate_rdma_pin_all() function
  migration: Move migrate_use_return() to options.c
  migration: Move migrate_use_block() to options.c
  migration: Move migrate_use_xbzrle() to options.c
  ...

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