tcg: Move TCG_{LOW,HIGH} to tcg-internal.h

Move the error-generating fallback from tcg-op.c, and
replace "_link_error" with modern QEMU_ERROR markup.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/sparc: Avoid TCGV_{LOW,HIGH}

Use the official extend/extract functions instead of routines
that will shortly be internal to tcg.

Cc: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Set cflags_next_tb in cpu_common_initfn

While we initialize this value in cpu_common_reset, that
isn't called during startup, so set it as well in init.
This fixes -singlestep versus the very first TB.

Fixes: 04f5b647ed07 ("accel/tcg: Handle -singlestep in curr_cflags")
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Introduce paired register allocation

There are several instances where we need to be able to
allocate a pair of registers to related inputs/outputs.
Add 'p' and 'm' register constraints for this, in order to
be able to allocate the even/odd register first or second.

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

tcg: Massage process_op_defs()

In preparation of introducing paired registers,
massage a bit process_op_defs()'s switch case.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
[PMD: Split from bigger patch, 1/3]
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221219220925.79218-2-philmd@linaro.org>

tcg: Remove check_regs

We now check the consistency of reg_to_temp[] with each update,
so the utility of checking consistency at the end of each
opcode is minimal.  In addition, the form of this check is
quite expensive, consuming 10% of a checking-enabled build.

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

tcg: Centralize updates to reg_to_temp

Create two new functions, set_temp_val_{reg,nonreg}.
Assert that the reg_to_temp mapping is correct before
any changes are made.

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

tcg: Fix tcg_reg_alloc_dup*

The assignment to mem_coherent should be done with any
modification, not simply with a newly allocated register.

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

tci: MAX_OPC_PARAM_IARGS is no longer used

Unused since commit 7b7d8b2d9a ("tcg/tci: Use ffi for calls").

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Remove TCG_TARGET_STACK_GROWSUP

The hppa host code has been removed since 2013; this
should have been deleted at the same time.

Fixes: 802b5081233a ("tcg-hppa: Remove tcg backend")
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Tidy tcg_reg_alloc_op

Replace goto allocate_in_reg with a boolean.
Remove o_preferred_regs which isn't used, except to copy.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Use QEMU_IOTHREAD_LOCK_GUARD in io_readx/io_writex

Narrow the scope of the lock to the actual read/write,
moving the cpu_transation_failed call outside the lock.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

hw/ppc: Use QEMU_IOTHREAD_LOCK_GUARD in ppc_set_irq

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/riscv: Use QEMU_IOTHREAD_LOCK_GUARD in riscv_cpu_update_mip

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/ppc: Use QEMU_IOTHREAD_LOCK_GUARD in cpu_interrupt_exittb

In addition, use tcg_enabled instead of !kvm_enabled.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/ppc: Use QEMU_IOTHREAD_LOCK_GUARD in ppc_maybe_interrupt

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

hw/mips: Use QEMU_IOTHREAD_LOCK_GUARD in cpu_mips_irq_request

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

qemu/main-loop: Introduce QEMU_IOTHREAD_LOCK_GUARD

Create a wrapper for locking/unlocking the iothread lock.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Cleanup trailing whitespace

Remove whitespace at end of line, plus one place this also
highlights some missing braces.

Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/s390x: Fix coding style

We are going to modify this code, so fix its style first to avoid:

  ERROR: spaces required around that '*' (ctx:VxV)
  #281: FILE: tcg/s390x/tcg-target.c.inc:1224:
  +        uintptr_t mask = ~(0xffffull << i*16);
                                            ^

Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221130132654.76369-2-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

meson: Move CONFIG_TCG_INTERPRETER to config_host

Like CONFIG_TCG, the enabled method of execution is a host property
not a guest property.  This exposes the define to compile-once files.

Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: convert tcg/README to rst

Convert tcg/README to rst and move it to docs/devel as a new "TCG Intermediate
Representation" page. There are a few minor changes to improve the aesthetic
of the final output which are as follows:

  - Rename the title from "Tiny Code Generator - Fabrice Bellard" to "TCG
    Intermediate Representation"

  - Remove the section numbering

  - Add the missing parameters to the ssadd_vec operations in the "Host
    vector operations" section

  - Change the path to the Atomic Operations document to use a proper
    reference

  - Replace tcg/README in tcg.rst with a proper reference to the new document

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Message-Id: <20221130100434.64207-2-mark.cave-ayland@ilande.co.uk>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

.gitlab-ci.d/windows: Work-around timeout and OpenGL problems of the MSYS2 jobs

The windows jobs (especially the 32-bit job) recently started to
hit the timeout limit. Bump it a little bit to ease the situation
(80 minutes is quite long already - OTOH, these jobs do not have to
wait for a job from the container stage to finish, so this should
still be OK).

Additionally, some update on the container side recently enabled
OpenGL in these jobs - but the corresponding code fails to compile.
Thus disable OpenGL here for the time being until someone figured
out the proper fix in the shader code for this.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-id: 20230104123559.277586-1-thuth@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

Merge tag 'pull-testing-next-231222-1' of https://gitlab.com/stsquad/qemu into staging

testing updates:

  - fix minor shell-ism that can break check-tcg
  - turn off verbose logging on custom runners
  - make configure echo call in CI
  - fix unused variable in linux-test
  - add binary compiler docker image for hexagon
  - disable doc and gui builds for tci and disable-tcg builds

# gpg: Signature made Fri 23 Dec 2022 15:18:41 GMT
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [full]
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* tag 'pull-testing-next-231222-1' of https://gitlab.com/stsquad/qemu:
  gitlab-ci: Disable docs and GUIs for the build-tci and build-tcg-disabled jobs
  tests/docker: use prebuilt toolchain for debian-hexagon-cross
  tests/tcg: fix unused variable in linux-test
  configure: repeat ourselves for the benefit of CI
  gitlab: turn off verbose logging for make check on custom runners
  configure: Fix check-tcg not executing any tests

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

Merge tag 'pull-9p-20221223' of https://github.com/cschoenebeck/qemu into staging

9pfs: Windows host prep, cleanup

* Next preparatory patches for upcoming Windows host support.

* Cleanup patches.

# gpg: Signature made Fri 23 Dec 2022 11:04:26 GMT
# gpg:                using RSA key 96D8D110CF7AF8084F88590134C2B58765A47395
# gpg:                issuer "qemu_oss@crudebyte.com"
# gpg: Good signature from "Christian Schoenebeck <qemu_oss@crudebyte.com>" [unknown]
# 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: ECAB 1A45 4014 1413 BA38  4926 30DB 47C3 A012 D5F4
#      Subkey fingerprint: 96D8 D110 CF7A F808 4F88  5901 34C2 B587 65A4 7395

* tag 'pull-9p-20221223' of https://github.com/cschoenebeck/qemu:
  hw/9pfs: Replace the direct call to xxxat() APIs with a wrapper
  hw/9pfs: Drop unnecessary *xattr wrapper API declarations
  qemu/xattr.h: Exclude <sys/xattr.h> for Windows
  MAINTAINERS: Add 9p test client to section "virtio-9p"
  9pfs: Fix some return statements in the synth backend

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

gitlab-ci: Disable docs and GUIs for the build-tci and build-tcg-disabled jobs

These jobs use their own "script:" section and thus do not profit from
the global "--disable-docs" from the template. While we're at it, disable
also some GUI front ends here since we do not gain any additional test
coverage by compiling those here again.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20221208135945.99975-1-thuth@redhat.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221221090411.1995037-7-alex.bennee@linaro.org>

tests/docker: use prebuilt toolchain for debian-hexagon-cross

The current docker image for cross compiling hexagon guests
is manually built since it takes >2 hours to build from source.

This patch:
 1. Solves the above issue by using the prebuilt clang
    toolchain hosted on CodeLinaro [1] and maintained by QUIC [2].
 2. The dockerfile is also switched from multi-stage to single stage
    build to allow the CI docker engine to reuse the layer cache.
 3. Re-enables the hexagon-cross-container job to be always run in
    CI and makes it a non-optional dependency for the
    build-user-hexagon job.

The changes for 1 & 2 together bring down the build time to
~3 minutes in GitLab CI when cache is reused and ~9 minutes
when cache cannot be reused.

[1]: https://github.com/CodeLinaro/hexagon-builder
[2]: https://github.com/quic/toolchain_for_hexagon/releases/

Signed-off-by: Mukilan Thiyagarajan <quic_mthiyaga@quicinc.com>
[AJB: also tweak MAINTAINERS, remove QEMU_JOB_ONLY_FORKS and comment]
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221219144354.11659-1-quic_mthiyaga@quicinc.com>
Message-Id: <20221221090411.1995037-6-alex.bennee@linaro.org>

tests/tcg: fix unused variable in linux-test

The latest hexagon compiler picks up that we never consume wcount.
Given the name of the #define that rcount checks against is WCOUNT_MAX
I figured the check just got missed.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221221090411.1995037-5-alex.bennee@linaro.org>

configure: repeat ourselves for the benefit of CI

Our CI system echos the lines it executes but not the expansions. For
the sake of a line of extra verbosity during the configure phase lets
echo the invocation of script to stdout as well as the log when on CI.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221221090411.1995037-4-alex.bennee@linaro.org>

gitlab: turn off verbose logging for make check on custom runners

The verbosity adds a lot of unnecessary output to the CI logs which
end up getting truncated anyway. We can always extract information
from the meson test logs on a failure and for the custom runners its
generally easier to re-create failures anyway.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20221221090411.1995037-3-alex.bennee@linaro.org>

configure: Fix check-tcg not executing any tests

After configuring with --target-list=hexagon-linux-user
running `make check-tcg` just prints the following:

```
make: Nothing to be done for 'check-tcg'
```

In the probe_target_compiler function, the 'break'
command is used incorrectly. There are no lexically
enclosing loops associated with that break command which
is an unspecfied behaviour in the POSIX standard.

The dash shell implementation aborts the currently executing
loop, in this case, causing the rest of the logic for the loop
in line 2490 to be skipped, which means no Makefiles are
generated for the tcg target tests.

Fixes: c3b570b5a9a24d25 (configure: don't enable
cross compilers unless in target_list)

Signed-off-by: Mukilan Thiyagarajan <quic_mthiyaga@quicinc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Link: https://patchew.org/QEMU/20221207082309.9966-1-quic._5Fmthiyaga@quicinc.com/
Message-Id: <20221207082309.9966-1-quic_mthiyaga@quicinc.com>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221221090411.1995037-2-alex.bennee@linaro.org>

hw/9pfs: Replace the direct call to xxxat() APIs with a wrapper

xxxat() APIs are only available on POSIX platforms. For future
extension to Windows, let's replace the direct call to xxxat()
APIs with a wrapper.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Message-Id: <20221219102022.2167736-4-bin.meng@windriver.com>
Signed-off-by: Christian Schoenebeck <qemu_oss@crudebyte.com>

hw/9pfs: Drop unnecessary *xattr wrapper API declarations

These are not used anywhere in the source tree. Drop them.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Message-Id: <20221219102022.2167736-3-bin.meng@windriver.com>
Signed-off-by: Christian Schoenebeck <qemu_oss@crudebyte.com>

qemu/xattr.h: Exclude <sys/xattr.h> for Windows

Windows does not have <sys/xattr.h>.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Message-Id: <20221219102022.2167736-2-bin.meng@windriver.com>
Signed-off-by: Christian Schoenebeck <qemu_oss@crudebyte.com>

MAINTAINERS: Add 9p test client to section "virtio-9p"

The 9p test cases use a dedicated, lite-weight 9p client implementation
(using virtio transport) under tests/qtest/libqos/ to communicate with
QEMU's 9p server.

It's already there for a long time. Let's officially assign it to 9p
maintainers.

Signed-off-by: Christian Schoenebeck <qemu_oss@crudebyte.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>
Message-Id: <E1ozhlV-0007BU-0g@lizzy.crudebyte.com>

9pfs: Fix some return statements in the synth backend

The qemu_v9fs_synth_mkdir() and qemu_v9fs_synth_add_file() functions
currently return a positive errno value on failure. This causes
checkpatch.pl to spit several errors like the one below:

ERROR: return of an errno should typically be -ve (return -EAGAIN)
+        return EAGAIN;

Simply change the sign. This has no consequence since callers
assert() the returned value to be equal to 0.

Reported-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <166930551818.827792.10663674346122681963.stgit@bahia>
[C.S.: - Resolve conflict with 66997c42e02c. ]
Signed-off-by: Christian Schoenebeck <qemu_oss@crudebyte.com>

Merge tag 'pull-ppc-20221221' of https://gitlab.com/danielhb/qemu into staging

ppc patch queue for 2022-12-21:

This queue contains a MAINTAINERS update, the implementation of the Freescale eSDHC,
the introduction of the DEXCR/HDEXCR instructions and other assorted fixes (most of
them for the e500 board).

# gpg: Signature made Wed 21 Dec 2022 17:18:53 GMT
# gpg:                using EDDSA key 17EBFF9923D01800AF2838193CD9CA96DE033164
# gpg:                issuer "danielhb413@gmail.com"
# gpg: Good signature from "Daniel Henrique Barboza <danielhb413@gmail.com>" [unknown]
# 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: 17EB FF99 23D0 1800 AF28  3819 3CD9 CA96 DE03 3164

* tag 'pull-ppc-20221221' of https://gitlab.com/danielhb/qemu:
  target/ppc: Check DEXCR on hash{st, chk} instructions
  target/ppc: Implement the DEXCR and HDEXCR
  hw/ppc/e500: Move comment to more appropriate place
  hw/ppc/e500: Resolve variable shadowing
  hw/ppc/e500: Prefer local variable over qdev_get_machine()
  hw/ppc/virtex_ml507: Prefer local over global variable
  target/ppc/mmu_common: Fix table layout of "info tlb" HMP command
  target/ppc/mmu_common: Log which effective address had no TLB entry found
  hw/ppc/spapr: Reduce "vof.h" inclusion
  hw/ppc/vof: Do not include the full "cpu.h"
  target/ppc/kvm: Add missing "cpu.h" and "exec/hwaddr.h"
  hw/ppc/e500: Add Freescale eSDHC to e500plat
  hw/sd/sdhci: Support big endian SD host controller interfaces
  MAINTAINERS: downgrade PPC KVM/TCG CPUs and pSeries to 'Odd Fixes'

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

target/ppc: Check DEXCR on hash{st, chk} instructions

Adds checks to the hashst and hashchk instructions to only execute if
enabled by the relevant aspect in the DEXCR and HDEXCR.

This behaviour is guarded behind TARGET_PPC64 since Power10 is
currently the only implementation which has the DEXCR.

Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Nicholas Miehlbradt <nicholas@linux.ibm.com>
Message-Id: <20221220042330.2387944-3-nicholas@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc: Implement the DEXCR and HDEXCR

Define the DEXCR and HDEXCR as special purpose registers.

Each register occupies two SPR indicies, one which can be read in an
unprivileged state and one which can be modified in the appropriate
priviliged state, however both indicies refer to the same underlying
value.

Note that the ISA uses the abbreviation UDEXCR in two different
contexts: the userspace DEXCR, the SPR index which can be read from
userspace (implemented in this patch), and the ultravisor DEXCR, the
equivalent register for the ultravisor state (not implemented).

Signed-off-by: Nicholas Miehlbradt <nicholas@linux.ibm.com>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20221220042330.2387944-2-nicholas@linux.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/e500: Move comment to more appropriate place

The TLB entries are set up in mmubooke_create_initial_mapping(), not in
booke206_page_size_to_tlb().

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20221216145709.271940-7-shentey@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/e500: Resolve variable shadowing

Assign to the outer variable instead which even saves some code.

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20221216145709.271940-6-shentey@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/e500: Prefer local variable over qdev_get_machine()

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221216145709.271940-5-shentey@gmail.com>
[danielhb: remove linebreak in object_property_add_child()]
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/virtex_ml507: Prefer local over global variable

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Edgar E. Iglesias <edgar@zeroasic.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221216145709.271940-4-shentey@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc/mmu_common: Fix table layout of "info tlb" HMP command

Starting with the URWX columns the columns didn't line up.
Before:

  QEMU 7.2.50 monitor - type 'help' for more information
  (qemu) info tlb

  TLB0:
  Effective          Physical           Size TID   TS SRWX URWX WIMGE U0123
  0x0000000000a80000 0x000000000105d000   4K 117   0  SR--UR-- --M-- U----
  0x0000000000100000 0x000000000114e000   4K 117   0  SR--UR-- --M-- U----
  <snip

  TLB1:
  Effective          Physical           Size TID   TS SRWX URWX WIMGE U0123
  0x00000000c0000000 0x0000000000000000  16M 0     0  SR-XU--- --M-- U----
  0x00000000c1000000 0x0000000001000000  16M 0     0  SRW-U--- --M-- U----
  <snip>
  (qemu)

After:

  QEMU 7.2.50 monitor - type 'help' for more information
  (qemu) info tlb

  TLB0:
  Effective          Physical           Size TID   TS SRWX URWX WIMGE U0123
  0x00000000b7a00000 0x000000000fcf5000   4K 18    0  SR-- UR-- --M-- U----
  0x0000000000800000 0x000000000fd73000   4K 18    0  SR-- UR-X --M-- U----
  <snip>

  TLB1:
  Effective          Physical           Size TID   TS SRWX URWX WIMGE U0123
  0x00000000c0000000 0x0000000000000000  16M 0     0  SR-X U--- --M-- U----
  0x00000000c1000000 0x0000000001000000  16M 0     0  SRW- U--- --M-- U----
  <snip>
  (qemu)

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221216145709.271940-3-shentey@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc/mmu_common: Log which effective address had no TLB entry found

Let's not leave developers in the dark where this log message comes
from.

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221216145709.271940-2-shentey@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/spapr: Reduce "vof.h" inclusion

Currently objects including "hw/ppc/spapr.h" are forced to be
target specific due to the inclusion of "vof.h" in "spapr.h".

"spapr.h" only uses a Vof pointer, so doesn't require the structure
declaration. The only place where Vof structure is accessed is in
spapr.c, so include "vof.h" there, and forward declare the structure
in "spapr.h".

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20221213123550.39302-4-philmd@linaro.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/vof: Do not include the full "cpu.h"

"vof.h" doesn't need the full "cpu.h" to get the target_ulong
definition, including "exec/cpu-defs.h" is enough.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20221213123550.39302-3-philmd@linaro.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

target/ppc/kvm: Add missing "cpu.h" and "exec/hwaddr.h"

kvm_ppc.h is missing various declarations from "cpu.h":

  target/ppc/kvm_ppc.h:128:40: error: unknown type name 'CPUPPCState'; did you mean 'CPUState'?
  static inline int kvmppc_get_hypercall(CPUPPCState *env,
                                         ^~~~~~~~~~~
                                         CPUState
  include/qemu/typedefs.h:45:25: note: 'CPUState' declared here
  typedef struct CPUState CPUState;
                          ^
  target/ppc/kvm_ppc.h:134:40: error: unknown type name 'PowerPCCPU'
  static inline int kvmppc_set_interrupt(PowerPCCPU *cpu, int irq, int level)
                                         ^
  target/ppc/kvm_ppc.h:285:38: error: unknown type name 'hwaddr'
                                       hwaddr ptex, int n)
                                       ^
  target/ppc/kvm_ppc.h:220:15: error: unknown type name 'target_ulong'
  static inline target_ulong kvmppc_configure_v3_mmu(PowerPCCPU *cpu,
                ^
  target/ppc/kvm_ppc.h:286:38: error: unknown type name 'ppc_hash_pte64_t'
  static inline void kvmppc_read_hptes(ppc_hash_pte64_t *hptes,
                                       ^

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20221213123550.39302-2-philmd@linaro.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/ppc/e500: Add Freescale eSDHC to e500plat

Adds missing functionality to e500plat machine which increases the
chance of given "real" firmware images to access SD cards.

Signed-off-by: Bernhard Beschow <shentey@gmail.com>
Message-Id: <20221018210146.193159-8-shentey@gmail.com>
[PMD: Simplify using create_unimplemented_device("esdhc")]
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Bernhard Beschow <shentey@gmail.com>
Message-Id: <20221101222934.52444-4-philmd@linaro.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

hw/sd/sdhci: Support big endian SD host controller interfaces

Some SDHCI IP can be synthetized in various endianness:
https://github.com/u-boot/u-boot/blob/v2021.04/doc/README.fsl-esdhc

 - CONFIG_SYS_FSL_ESDHC_BE

   ESDHC IP is in big-endian mode. Accessing ESDHC registers can be
   determined by ESDHC IP's endian mode or processor's endian mode.

Our current implementation is little-endian. In order to support
big endianness:

- Rename current MemoryRegionOps as sdhci_mmio_le_ops ('le')
- Add an 'endianness' property to SDHCIState (default little endian)
- Set the 'io_ops' field in realize() after checking the property
- Add the sdhci_mmio_be_ops (big-endian) MemoryRegionOps.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Bernhard Beschow <shentey@gmail.com>
Reviewed-by: Bernhard Beschow <shentey@gmail.com>
Message-Id: <20221101222934.52444-3-philmd@linaro.org>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

MAINTAINERS: downgrade PPC KVM/TCG CPUs and pSeries to 'Odd Fixes'

The maintainer is no longer being paid to maintain these components. All
maintainership work is being done in his personal time since the middle
of the 7.2 development cycle.

Change the status of PPC KVM CPUs, PPC TCG CPUs and the pSeries machine
to 'Odd Fixes', reflecting that the maintainer no longer has exclusive
time to dedicate to them. It'll also (hopefully) keep expectations under
check when/if these components are used in a customer product.

Cc: Cédric Le Goater <clg@kaod.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20221117153218.182835-1-danielhb413@gmail.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>

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

virtio,pc,pci: features, cleanups, fixes

make TCO watchdog work by default
part of generic vdpa support
asid interrupt for vhost-vdpa
added flex bus port DVSEC for cxl

misc fixes, cleanups, documentation

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Wed 21 Dec 2022 12:32:36 GMT
# gpg:                using RSA key 5D09FD0871C8F85B94CA8A0D281F0DB8D28D5469
# gpg:                issuer "mst@redhat.com"
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [full]
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>" [full]
# 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: (41 commits)
  contrib/vhost-user-blk: Replace lseek64 with lseek
  libvhost-user: Switch to unsigned int for inuse field in struct VuVirtq
  hw/virtio: Extract QMP related code virtio-qmp.c
  hw/virtio: Extract config read/write accessors to virtio-config-io.c
  hw/virtio: Constify qmp_virtio_feature_map_t[]
  hw/virtio: Guard and restrict scope of qmp_virtio_feature_map_t[]
  hw/virtio: Rename virtio_ss[] -> specific_virtio_ss[]
  hw/virtio: Add missing "hw/core/cpu.h" include
  hw/cxl/device: Add Flex Bus Port DVSEC
  hw/acpi: Rename tco.c -> ich9_tco.c
  acpi/tests/avocado/bits: add mformat as one of the dependencies
  docs/acpi/bits: document BITS_DEBUG environment variable
  pci: drop redundant PCIDeviceClass::is_bridge field
  remove DEC 21154 PCI bridge
  vhost: fix vq dirty bitmap syncing when vIOMMU is enabled
  acpi/tests/avocado/bits: add SPDX license identifiers for bios bits tests
  include/hw: attempt to document VirtIO feature variables
  vhost-user: send set log base message only once
  vdpa: always start CVQ in SVQ mode if possible
  vdpa: add shadow_data to vhost_vdpa
  ...

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

Merge tag 'pull-tcg-20221220' of https://gitlab.com/rth7680/qemu into staging

Use interval trees for user-only vma mappings.
Assorted cleanups to page locking.

# gpg: Signature made Wed 21 Dec 2022 05:00:30 GMT
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* tag 'pull-tcg-20221220' of https://gitlab.com/rth7680/qemu:
  accel/tcg: Restrict page_collection structure to system TB maintainance
  accel/tcg: Factor tb_invalidate_phys_range_fast() out
  accel/tcg: Rename tb_invalidate_phys_page_fast{,__locked}()
  accel/tcg: Remove trace events from trace-root.h
  accel/tcg: Restrict cpu_io_recompile() to system emulation
  accel/tcg: Move remainder of page locking to tb-maint.c
  accel/tcg: Move PageDesc tree into tb-maint.c for system
  accel/tcg: Use interval tree for user-only page tracking
  accel/tcg: Move page_{get,set}_flags to user-exec.c
  accel/tcg: Drop PAGE_RESERVED for CONFIG_BSD
  accel/tcg: Use interval tree for TARGET_PAGE_DATA_SIZE
  accel/tcg: Use interval tree for TBs in user-only mode
  accel/tcg: Rename page_flush_tb
  util: Add interval-tree.c

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

block/io: Check for replay-enabled in bdrv_drain_all_begin()

In commit da0bd74434 we refactored bdrv_drain_all_begin() to pull out
the non-polling part into bdrv_drain_all_begin_nopoll().  This change
broke record-and-replay, because the "return early if replay enabled"
check is now in the sub-function bdrv_drain_all_begin_nopoll(), and
so it only causes us to return from that function, and not from the
calling bdrv_drain_all_begin().

Fix the regression by checking whether replay is enabled in both
functions.

The breakage and fix can be tested via 'make check-avocado': the
tests/avocado/reverse_debugging.py:ReverseDebugging_X86_64.test_x86_64_pc
tests/avocado/reverse_debugging.py:ReverseDebugging_AArch64.test_aarch64_virt
tests were both broken by this.

Fixes: da0bd744344adb1f285 ("block: Factor out bdrv_drain_all_begin_nopoll()")
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Tested-by: Fabiano Rosas <farosas@suse.de>
Message-id: 20221220174638.2156308-1-peter.maydell@linaro.org

contrib/vhost-user-blk: Replace lseek64 with lseek

64bit off_t is already in use since build uses _FILE_OFFSET_BITS=64
already. Using lseek/off_t also makes it work with latest musl without
using _LARGEFILE64_SOURCE macro. This macro is implied with _GNU_SOURCE
when using glibc but not with musl.

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
CC: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <20221218220740.315839-1-raj.khem@gmail.com>
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>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>

libvhost-user: Switch to unsigned int for inuse field in struct VuVirtq

It seems there is no need to keep the inuse field signed and end up with
compiler warnings for sign-compare.

  CC       libvhost-user.o
libvhost-user.c: In function ‘vu_queue_pop’:
libvhost-user.c:2763:19: error: comparison of integer expressions of different signedness: ‘int’ and ‘unsigned int’ [-Werror=sign-compare]
 2763 |     if (vq->inuse >= vq->vring.num) {
      |                   ^~
libvhost-user.c: In function ‘vu_queue_rewind’:
libvhost-user.c:2808:13: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]
 2808 |     if (num > vq->inuse) {
      |             ^

Instead of casting the comparision to unsigned int, just make the inuse
field unsigned int in the fist place.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Message-Id: <20221219175337.377435-8-marcel@holtmann.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/virtio: Extract QMP related code virtio-qmp.c

The monitor decoders are the only functions using the CONFIG_xxx
definitions declared in the target specific CONFIG_DEVICES header.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221213111707.34921-7-philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <a class="moz-txt-link-rfc2396E" href="mailto:philmd@linaro.org">&lt;philmd@linaro.org&gt;</a>

hw/virtio: Extract config read/write accessors to virtio-config-io.c

These config helpers use the target-dependent LD/ST API.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221213111707.34921-6-philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/virtio: Constify qmp_virtio_feature_map_t[]

These arrays are only accessed read-only, move them to .rodata.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221213111707.34921-5-philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Jonah Palmer<jonah.palmer@oracle.com>
Signed-off-by: Philippe Mathieu-Daudé <a class="moz-txt-link-rfc2396E" href="mailto:philmd@linaro.org">&lt;philmd@linaro.org&gt;</a>

hw/virtio: Guard and restrict scope of qmp_virtio_feature_map_t[]

Commit f3034ad71f ("qmp: decode feature & status bits in
virtio-status") did not guard all qmp_virtio_feature_map_t
arrays with the corresponding #ifdef'ry used in
qmp_decode_features(). Fix that and reduce the arrays scope
by declaring them static.

Suggested-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221213111707.34921-4-philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Jonah Palmer<jonah.palmer@oracle.com>
Suggested-by: Richard Henderson <a class="moz-txt-link-rfc2396E" href="mailto:richard.henderson@linaro.org">&lt;richard.henderson@linaro.org&gt;</a>
Signed-off-by: Philippe Mathieu-Daudé <a class="moz-txt-link-rfc2396E" href="mailto:philmd@linaro.org">&lt;philmd@linaro.org&gt;</a>

hw/virtio: Rename virtio_ss[] -> specific_virtio_ss[]

Since virtio_ss[] is added to specific_ss[], rename it as
specific_virtio_ss[] to make it clearer.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221213111707.34921-3-philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/virtio: Add missing "hw/core/cpu.h" include

virtio.c uses target_words_bigendian() which is declared in
"hw/core/cpu.h". Add the missing header to avoid when refactoring:

  hw/virtio/virtio.c:2451:9: error: implicit declaration of function 'target_words_bigendian' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
    if (target_words_bigendian()) {
        ^

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221213111707.34921-2-philmd@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/cxl/device: Add Flex Bus Port DVSEC

The Flex Bus Port DVSEC was missing on type 3 devices which was blocking
RAS checks.[1]

Add the Flex Bus Port DVSEC to type 3 devices as per CXL 3.0 8.2.1.3.

[1] https://lore.kernel.org/linux-cxl/167096738875.2861540.11815053323626849940.stgit@djiang5-desk3.ch.intel.com/

Cc: Dave Jiang <dave.jiang@intel.com>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Ben Widawsky <bwidawsk@kernel.org>
Cc: qemu-devel@nongnu.org
Cc: linux-cxl@vger.kernel.org
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Message-Id: <20221213-ira-flexbus-port-v2-1-eaa48d0e0700@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>

hw/acpi: Rename tco.c -> ich9_tco.c

tco.c contains the ICH9 implementation of its "total cost
of ownership". Rename it accordingly to emphasis this is
a part of the ICH9 model.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221212105115.2113-1-philmd@linaro.org>
Acked-by: Igor Mammedov <imammedo@redhat.com>

acpi/tests/avocado/bits: add mformat as one of the dependencies

mformat is needed by grub-mkrescue and hence, add this as one of the
dependencies to run bits tests. This avoids errors such as the following:

/var/tmp/acpi-bits-wju6tqoa.tmp/grub-inst-x86_64-efi/bin/grub-mkrescue: 360: mformat: not found

Signed-off-by: Ani Sinha <ani@anisinha.ca>
Message-Id: <20221203132407.34539-1-ani@anisinha.ca>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

docs/acpi/bits: document BITS_DEBUG environment variable

Debug specific actions can be enabled in bios bits acpi tests by passing
BITS_DEBUG in the environment variable while running the test. Document that.

CC: qemu-trivial@nongnu.org
Signed-off-by: Ani Sinha <ani@anisinha.ca>
Message-Id: <20221203132346.34479-1-ani@anisinha.ca>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Wilfred Mallawa <wilfred.mallawa@wdc.com>

pci: drop redundant PCIDeviceClass::is_bridge field

and use cast to TYPE_PCI_BRIDGE instead.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20221129101341.185621-3-imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>

remove DEC 21154 PCI bridge

Code has not been used practically since its inception (2004)
  f2aa58c6f4a20 UniNorth PCI bridge support
or maybe even earlier, but it was consuming contributors time
as QEMU was being rewritten.
Drop it for now. Whomever would like to actually
use the thing, can make sure it actually works/reintroduce
it back when there is a user.

PS:
I've stumbled upon this when replacing PCIDeviceClass::is_bridge
field with QOM cast to PCI_BRIDGE type. Unused DEC 21154
was the only one trying to use the field with plain PCIDevice.
It's not worth keeping the field around for the sake of the code
that was commented out 'forever'.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20221129101341.185621-2-imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost: fix vq dirty bitmap syncing when vIOMMU is enabled

When vIOMMU is enabled, the vq->used_phys is actually the IOVA not
GPA. So we need to translate it to GPA before the syncing otherwise we
may hit the following crash since IOVA could be out of the scope of
the GPA log size. This could be noted when using virtio-IOMMU with
vhost using 1G memory.

Fixes: c471ad0e9bd46 ("vhost_net: device IOTLB support")
Cc: qemu-stable@nongnu.org
Tested-by: Lei Yang <leiyang@redhat.com>
Reported-by: Yalan Zhang <yalzhang@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221216033552.77087-1-jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

acpi/tests/avocado/bits: add SPDX license identifiers for bios bits tests

Added the SPDX license identifiers for biosbits tests.
Also added a comment on each of the test scripts to indicate that they run
from within the biosbits environment and hence are not subjected to the regular
maintenance activities for QEMU and is excluded from the dependency management
challenges in the host testing environment.

Cc: Daniel P. Berrangé <berrange@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Maydell Peter <peter.maydell@linaro.org>
Cc: John Snow <jsnow@redhat.com>
Cc: Thomas Huth <thuth@redhat.com>
Cc: Alex Bennée <alex.bennee@linaro.org>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Michael Tsirkin <mst@redhat.com>
Cc: Thomas Huth <thuth@redhat.com>
Cc: qemu-trivial@nongnu.org
Signed-off-by: Ani Sinha <ani@anisinha.ca>
Message-Id: <20221125044138.962137-1-ani@anisinha.ca>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

include/hw: attempt to document VirtIO feature variables

We have a bunch of variables associated with the device and the vhost
backend which are used inconsistently throughout the code base. Lets
start trying to bring some order by agreeing what each variable is
for.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Cc: Stefano Garzarella <sgarzare@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Stefan Hajnoczi <stefanha@gmail.com>
Message-Id: <20221123152134.179929-2-alex.bennee@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost-user: send set log base message only once

Vhost message VHOST_USER_SET_LOG_BASE is device wide. So only
send it once with the first queue pair.

Signed-off-by: Yajun Wu <yajunw@nvidia.com>
Acked-by: Parav Pandit <parav@nvidia.com>
Message-Id: <20221122051447.248462-1-yajunw@nvidia.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: always start CVQ in SVQ mode if possible

Isolate control virtqueue in its own group, allowing to intercept control
commands but letting dataplane run totally passthrough to the guest.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20221215113144.322011-13-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>

vdpa: add shadow_data to vhost_vdpa

The memory listener that thells the device how to convert GPA to qemu's
va is registered against CVQ vhost_vdpa. memory listener translations
are always ASID 0, CVQ ones are ASID 1 if supported.

Let's tell the listener if it needs to register them on iova tree or
not.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-12-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: store x-svq parameter in VhostVDPAState

CVQ can be shadowed two ways:
- Device has x-svq=on parameter (current way)
- The device can isolate CVQ in its own vq group

QEMU needs to check for the second condition dynamically, because CVQ
index is not known before the driver ack the features. Since this is
dynamic, the CVQ isolation could vary with different conditions, making
it possible to go from "not isolated group" to "isolated".

Saving the cmdline parameter in an extra field so we never disable CVQ
SVQ in case the device was started with x-svq cmdline.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-11-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: add asid parameter to vhost_vdpa_dma_map/unmap

So the caller can choose which ASID is destined.

No need to update the batch functions as they will always be called from
memory listener updates at the moment. Memory listener updates will
always update ASID 0, as it's the passthrough ASID.

All vhost devices's ASID are 0 at this moment.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-10-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: allocate SVQ array unconditionally

SVQ may run or not in a device depending on runtime conditions (for
example, if the device can move CVQ to its own group or not).

Allocate the SVQ array unconditionally at startup, since its hard to
move this allocation elsewhere.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-9-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: move SVQ vring features check to net/

The next patches will start control SVQ if possible. However, we don't
know if that will be possible at qemu boot anymore.

Since the moved checks will be already evaluated at net/ to know if it
is ok to shadow CVQ, move them.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-8-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: request iova_range only once

Currently iova range is requested once per queue pair in the case of
net. Reduce the number of ioctls asking it once at initialization and
reusing that value for each vhost_vdpa.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20221215113144.322011-7-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasonwang@redhat.com>

vdpa: add vhost_vdpa_net_valid_svq_features

It will be reused at vdpa device start so let's extract in its own
function.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-6-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost: move iova_tree set to vhost_svq_start

Since we don't know if we will use SVQ at qemu initialization, let's
allocate iova_tree only if needed. To do so, accept it at SVQ start, not
at initialization.

This will avoid to create it if the device does not support SVQ.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-5-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost: allocate SVQ device file descriptors at device start

The next patches will start control SVQ if possible. However, we don't
know if that will be possible at qemu boot anymore.

Delay device file descriptors until we know it at device start. This
will avoid to create them if the device does not support SVQ.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-4-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost: set SVQ device call handler at SVQ start

By the end of this series CVQ is shadowed as long as the features
support it.

Since we don't know at the beginning of qemu running if this is
supported, move the event notifier handler setting to the start of the
SVQ, instead of the start of qemu run. This will avoid to create them if
the device does not support SVQ.

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-3-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: use v->shadow_vqs_enabled in vhost_vdpa_svqs_start & stop

This function used to trust in v->shadow_vqs != NULL to know if it must
start svq or not.

This is not going to be valid anymore, as qemu is going to allocate svq
array unconditionally (but it will only start them conditionally).

Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20221215113144.322011-2-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa-dev: mark the device as unmigratable

The generic vDPA device doesn't support migration currently, so
mark it as unmigratable temporarily.

Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Longpeng <longpeng2@huawei.com>
Message-Id: <20221215134944.2809-5-longpeng2@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: add vdpa-dev-pci support

Supports vdpa-dev-pci, we can use the device as follow:

-device vhost-vdpa-device-pci,vhostdev=/dev/vhost-vdpa-X

Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Longpeng <longpeng2@huawei.com>
Message-Id: <20221215134944.2809-4-longpeng2@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: add vdpa-dev support

Supports vdpa-dev, we can use the deivce directly:

-M microvm -m 512m -smp 2 -kernel ... -initrd ... -device \
vhost-vdpa-device,vhostdev=/dev/vhost-vdpa-x

Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Longpeng <longpeng2@huawei.com>
Message-Id: <20221215134944.2809-3-longpeng2@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio: get class_id and pci device id by the virtio id

Add helpers to get the "Transitional PCI Device ID" and "class_id"
of the device specified by the "Virtio Device ID".

These helpers will be used to build the generic vDPA device later.

Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Longpeng <longpeng2@huawei.com>
Message-Id: <20221215134944.2809-2-longpeng2@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

ich9: honour 'enable_tco' property

An 'ICH9-LPC.enable_tco' property has been exposed for a
very long time, but attempts to set it have never been
honoured.

Originally, any user provided 'enable_tco' value was force
replaced by a value passed from the machine type setup
code that was determine by machine type compat properties.

  commit d6b304ba924b95d12edfddaac99777b577301309
  Author: Eduardo Habkost <ehabkost@redhat.com>
  Date:   Sat Jan 23 14:02:10 2016 -0200

    machine: Remove no_tco field

    The field is always set to zero, so it is not necessary anymore.

After legacy Q35 machine types were deleted in:

  commit 86165b499edf8b03bb2d0e926d116c2f12a95bfe
  Author: Eduardo Habkost <ehabkost@redhat.com>
  Date:   Sat Jan 23 14:02:09 2016 -0200

    q35: Remove old machine versions

the machine type code ended up just unconditionally passing
'true', all the time, so this was further simplified in

  commit d6b304ba924b95d12edfddaac99777b577301309
  Author: Eduardo Habkost <ehabkost@redhat.com>
  Date:   Sat Jan 23 14:02:10 2016 -0200

    machine: Remove no_tco field

    The field is always set to zero, so it is not necessary anymore.

  commit 18d6abae3ea092950629e5d26aff1dcfc9a2d78e
  Author: Eduardo Habkost <ehabkost@redhat.com>
  Date:   Sat Jan 23 14:02:11 2016 -0200

    ich9: Remove enable_tco arguments from init functions

    The enable_tco arguments are always true, so they are not needed
    anymore.

Leaving the ich9_pm_init to just force set 'enable_tco' to true.
This still overrides any user specified property. The initialization
of property defaults should be done when properties are first
registered, rather than during object construction.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20221216125749.596075-6-berrange@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/isa: enable TCO watchdog reboot pin strap by default

The TCO watchdog implementation default behaviour from POV of the
guest OS relies on the initial values for two I/O ports:

  * TCO1_CNT == 0x0

    Since bit 11 (TCO Timer Halt) is clear, the watchdog state
    is considered to be initially running

  * GCS == 0x20

    Since bit 5 (No Reboot) is set, the watchdog will not trigger
    when the timer expires

This is a safe default, because the No Reboot bit will prevent the
watchdog from triggering if the guest OS is unaware of its existance,
or is slow in configuring it. When a Linux guest initializes the TCO
watchdog, it will attempt to clear the "No Reboot" flag, and read the
value back. If the clear was honoured, the driver will treat this as
an indicator that the watchdog is functional and create the guest
watchdog device.

QEMU implements a second "no reboot" flag, however, via pin straps
which overrides the behaviour of the guest controlled "no reboot"
flag:

  commit 5add35bec1e249bb5345a47008c8f298d4760be4
  Author: Paulo Alcantara <pcacjr@gmail.com>
  Date:   Sun Jun 28 14:58:58 2015 -0300

    ich9: implement strap SPKR pin logic

This second 'noreboot' pin was defaulted to high, which also inhibits
triggering of the requested watchdog actions, unless QEMU is launched
with the magic flag "-global ICH9-LPC.noreboot=false".

This is a bad default as we are exposing a watchdog to every guest OS
using the q35 machine type, but preventing it from actually doing what
it is designed to do. What is worse is that the guest OS and its apps
have no way to know that the watchdog is never going to fire, due to
this second 'noreboot' pin.

If a guest OS had no watchdog device at all, then apps whose operation
and/or data integrity relies on a watchdog can refuse to launch, and
alert the administrator of the problematic deployment. With Q35 machines
unconditionally exposing a watchdog though, apps will think their
deployment is correct but in fact have no protection at all.

This patch flips the default of the second 'no reboot' flag, so that
configured watchdog actions will be honoured out of the box for the
7.2 Q35 machine type onwards, if the guest enables use of the watchdog.

See also related bug reports

  https://bugzilla.redhat.com/show_bug.cgi?id=2080207
  https://bugzilla.redhat.com/show_bug.cgi?id=2136889
  https://bugzilla.redhat.com/show_bug.cgi?id=2137346

Reviewed-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20221216125749.596075-5-berrange@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

pc: clean up compat machines

We can move setting default_cpu_version into the base machine options,
and we need to unset alias and is_default only once.

Suggested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20221212152145.124317-3-cohuck@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: Add compat machines for 8.0

Add 8.0 machine types for arm/i440fx/m68k/q35/s390x/spapr.

Reviewed-by: Cédric Le Goater <clg@kaod.org> [ppc]
Reviewed-by: Thomas Huth <thuth@redhat.com> [s390x]
Reviewed-by: Greg Kurz <groug@kaod.org> [ppc]
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20221212152145.124317-2-cohuck@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>

hw/watchdog: add trace events for watchdog action handling

The tracepoints aid in debugging the triggering of watchdog devices.

Reviewed-by: Richard W.M. Jones <rjones@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20221216125749.596075-4-berrange@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/isa: add trace events for ICH9 LPC chip config access

These tracepoints aid in understanding and debugging the guest drivers
for the TCO watchdog.

Reviewed-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20221216125749.596075-3-berrange@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

hw/acpi: add trace events for TCO watchdog register access

These tracepoints aid in understanding and debugging the guest drivers
for the TCO watchdog.

Reviewed-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20221216125749.596075-2-berrange@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

accel/tcg: Restrict page_collection structure to system TB maintainance

Only the system emulation part of TB maintainance uses the
page_collection structure. Restrict its declaration (and the
functions requiring it) to tb-maint.c.

Convert the 'len' argument of tb_invalidate_phys_page_fast__locked()
from signed to unsigned.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-6-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Factor tb_invalidate_phys_range_fast() out

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-5-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Rename tb_invalidate_phys_page_fast{,__locked}()

Emphasize this function is called with pages locked.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-4-philmd@linaro.org>
[rth: Use "__locked" suffix, to match other instances.]
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Remove trace events from trace-root.h

Commit d9bb58e510 ("tcg: move tcg related files into accel/tcg/
subdirectory") introduced accel/tcg/trace-events, so we don't
need to use the root trace-events anymore.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-3-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Restrict cpu_io_recompile() to system emulation

Missed in commit 6526919224 ("accel/tcg: Restrict cpu_io_recompile()
from other accelerators").

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20221209093649.43738-2-philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>