Updates for testing, plugins, gdbstub

  - restore some 32 bit host builds and testing
  - move some physmem tracepoint definitions
  - use --userns keep-id for podman builds
  - cleanup check-tcg compiler flag checking for Arm
  - fix some casting in fcvt test
  - tweak check-tcg inline asm for clang
  - suppress some invalid clang warnings
  - disable KVM for the TCI builds
  - improve the insn tracking plugin
  - cleanups to the lockstep plugin
  - free plugin data on cpu finalise
  - assert cpu->index assigned
  - move qemu_plugin_vcpu_init__async into plugin code
  - add support for dynamic gdb command tables
  - allow targets to extend gdb capabilities
  - enable user-mode MTE support

maintainer updates (plugins, gdbstub):

  - add missing include guard comment to gdbstub.h
  - move gdbstub enums into separate header
  - move qtest_[get|set]_virtual_clock functions
  - allow plugins to manipulate the virtual clock
  - introduce an Instructions Per Second plugin
  - fix inject_mem_cb rw mask tests
  - allow qemu_plugin_vcpu_mem_cb to shortcut when no memory cbs

testing cleanups (ci, vm, lcitool, ansible):

  - clean up left over Centos 8 references
  - use -fno-sanitize=function to avoid non-useful errors
  - bump lcitool and update images (alpine, fedora)
  - make sure we have mingw-w64-tools for windows builds
  - drive ansible scripts with lcitool package lists

plugin and testing updates

 - don't duplicate options for microbit test
 - don't spam the linux source tree when importing headers
 - add STORE_U64 inline op to TCG plugins
 - add conditional callback op to TCG plugins

plugin and testing updates

 - don't duplicate options for microbit test
 - don't spam the linux source tree when importing headers
 - add STORE_U64 inline op to TCG plugins
 - add conditional callback op to TCG plugins

maintainer updates (gitlab, avocado):

  - avoid extra git data on gitlab checkouts
  - update sbsa-ref tests

final updates for 9.0 (testing, gdbstub):

  - fix the over rebuilding of test VMs
  - support Xfer:siginfo:read in gdbstub
  - fix double close() in gdbstub

final updates for 9.0 (testing, gdbstub):

  - avoid transferring pointless git data
  - fix the over rebuilding of test VMs
  - support Xfer:siginfo:read in gdbstub
  - fix double close() in gdbstub

maintainer updates (tests, gdbstub, plugins):

  - expand QOS_PATH_MAX_ELEMENT_SIZE to avoid LTO issues
  - support fork-follow-mode in gdbstub
  - new thread-safe scoreboard API for TCG plugins
  - suppress showing opcodes in plugin disassembly

Testing, gdbstub and plugin updates:

  - fix some test/tcg license headers to GPLv2+
  - bump up check-tcg timeout to 120s
  - avoid re-building VM images too often
  - update OpenBSD to 7.4
  - use GDBFeature to build gdbstub XML
  - unify plugin vcpu count under qemu_plugin_num_vcpus
  - avoid spurious idle/resume callbacks on new vCPUs
  - ensure nios2-linux-user processes async work
  - call vcpu_init plugin callback through async work
  - define plugin helpers when registers being read
  - add plugin API for reading register values
  - add support for register tracking to execlog
  - update plugin docs with assumptions
  - mention plugins can trigger tb_flush in mttcg design doc

testing, doc and gdbstub updates:

  - add sqlite3 to openSUSE image
  - mark CRIS as deprecated
  - re-enable the TCG plugin tests
  - use select for semihosting
  - implement syscall catching in gdbstub

maintainer updates for 9.0 (docker, plugin tests, deprecation, elf, semihosting)

A fairly random collection of fixes in this tree. I've still got a
report of openbsd rebuilding which is confusing me but I didn't want
to hold up getting eyes on real fixes.

The plugins register support may get added to the PR if it gets any
review.

The following need review:

  Revert "hw/elf_ops: Ignore loadable segments with zero size"
  docs: mark CRIS support as deprecated

Alex.

plugin updates (register access) for 9.0 (pre-PR?)

Akihiko requested the register support not be merged in its current
state so it's time for another round of review. I've made a few tweaks
to simplify the register and CPU tracking code in execlog and removed
some stale API functions. However from my point of view its ready to
merge.

v3
--
  - split from testing bits (merged)
  - removed unused api funcs
  - keep CPUs in a GArray instead of doing by hand

v2
--

 - Review feedback for register API
 - readthedocs update
 - add expectation docs for plugins

The following still need review:

  contrib/plugins: extend execlog to track register changes
  gdbstub: expose api to find registers

read-only register access for plugins:

  - move misa_mxl to CPU class for riscv
  - use GDBFeature for arm XML
  - use GDBFeature for ppc XML
  - use GDBFeature for riscv XML
  - unify gdb code to use GDBFeature
  - move dynamic XML generation to core GDB code
  - provide introspection APIs for rest of QEMU
  - expose a plugin API to access registers
  - fix memory re-use in execlog
  - extend execlog to track registers
  - optimise instrumentation based on disassembly
  - tweak API docs and expand on assumptions

read-only register access for plugins:

  - move misa_mxl to CPU class for riscv
  - use GDBFeature for arm XML
  - use GDBFeature for ppc XML
  - use GDBFeature for riscv XML
  - unify gdb code to use GDBFeature
  - move dynamic XML generation to core GDB code
  - provide introspection APIs for rest of QEMU
  - expose a plugin API to access registers
  - fix memory re-use in execlog
  - extend execlog to track registers
  - optimise instrumentation based on disassembly
  - tweak API docs and expand on assumptions

testing and misc updates

  - add LE microblaze test to avocado
  - use modern snapshot=on to avoid trashing disk image
  - use plain bool for fe_is_open
  - various updates to qtest timeouts
  - enable meson test timeouts
  - tweak the readthedocs environment
  - partially revert un-flaking x86_64

testing and misc updates

  - add LE microblaze test to avocado
  - use modern snapshot=on to avoid trashing disk image
  - use plain bool for fe_is_open
  - various updates to qtest timeouts
  - enable meson test timeouts
  - tweak the readthedocs environment
  - revert the revert of flaky replay tests

Record/replay fixes for replay_kernel tests

  - add a 32 bit x86 replay test case
  - fix some typos
  - use modern snapshot setting for tests
  - update replay_dump for current ABI
  - remove stale replay variables
  - improve kdoc for ReplayState
  - introduce common error path for replay
  - always fully drain chardevs when in replay
  - catch unexpected waitio on playback
  - remove flaky tags from replay_kernel tests

testing and plugin updates for 9.0 (pre-PR)

This brings in the first batch of testing updates for the next
release. The main bulk of these is Daniel and Thomas' cleanups of the
qtest timeouts and allowing meson control them. There are a few minor
tweaks I've made to some avocado and gitlab tests.

The big update is support for reading register values in TCG plugins.
After feedback from Akihiko I've left all the smarts to the plugin and
made the interface a simple "all the registers" dump. There is a
follow on patch to make the register code a little more efficient by
checking disassembly. However we can leave the door open for future
API enhancements if the translator ever learns to reliably know when
registers might be touched.

v2
--

 - Review feedback for register API
 - readthedocs update
 - add expectation docs for plugins

The following still need review:

  docs/devel: document some plugin assumptions
  docs/devel: lift example and plugin API sections up
  contrib/plugins: optimise the register value tracking
  contrib/plugins: extend execlog to track register changes
  contrib/plugins: fix imatch
  plugins: add an API to read registers
  gdbstub: expose api to find registers
  readthodocs: fully specify a build environment
  gitlab: include microblazeel in testing
  tests/avocado: use snapshot=on in kvm_xen_guest

testing and plugin updates for 9.0

This brings in the first batch of testing updates for the next
release. The main bulk of these is Daniel and Thomas' cleanups of the
qtest timeouts and allowing meson control them. There are a few minor
tweaks I've made to some avocado and gitlab tests.

The big update is support for reading register values in TCG plugins.
After feedback from Akihiko I've left all the smarts to the plugin and
made the interface a simple "all the registers" dump. There is a
follow on patch to make the register code a little more efficient by
checking disassembly. However we can leave the door open for future
API enhancements if the translator ever learns to reliably know when
registers might be touched.

The following still need review:

  contrib/plugins: optimise the register value tracking
  contrib/plugins: extend execlog to track register changes
  contrib/plugins: fix imatch
  plugins: add an API to read registers
  gdbstub: expose api to find registers
  gitlab: include microblazeel in testing
  tests/avocado: use snapshot=on in kvm_xen_guest