audio/pw: Report more accurate error when connecting to PipeWire fails

According to its man page [1], pw_context_connect() sets errno on
failure:

  Returns a Core on success or NULL with errno set on error.

It may be handy to see errno when figuring out why PipeWire
failed to connect. That leaves us with just one possible path to
reach 'fail_error' label which is then moved to that path and
also its error message is adjusted slightly.

1: https://docs.pipewire.org/group__pw__core.html#ga5994e3a54e4ec718094ca02a1234815b

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-ID: <3a78811ad5b0e87816b7616ab21d2eeef00b9c52.1726647033.git.mprivozn@redhat.com>

tests: add basic -display dbus Map.Unix test

Only check we eventually get a shared memory scanout.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-20-marcandre.lureau@redhat.com>

ui: refactor using a common qemu_pixman_shareable

Use a common shareable type for win32 & unix, and helper functions.
This simplify the code as it avoids a lot of #ifdef'ery.

Note: if it helps review, commits could be reordered to introduce the
common type before introducing shareable memory for unix.

Suggested-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-19-marcandre.lureau@redhat.com>

virtio-gpu: allocate shareable 2d resources on !win32

Similar to what was done in commit 9462ff46 ("virtio-gpu/win32: allocate
shareable 2d resources/images") for win32, allocate resource memory with
memfd, so the associated display surface memory can be shared with a
different process.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-18-marcandre.lureau@redhat.com>

ui/dbus: implement Unix.Map

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-17-marcandre.lureau@redhat.com>

ui/dbus: add Listener.Unix.Map interface XML

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-16-marcandre.lureau@redhat.com>

ui/dbus: make Listener.Win32.Map win32-specific

There are no types specific to Windows, so the code compiles on other
platforms, but its useless on !Windows.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-15-marcandre.lureau@redhat.com>

meson: find_program('gdbus-codegen') directly

gio.pc variable is a bit bogus in context of cross-compilation, since it
contains an absolute path, relative to the sysroot directory. On Fedora, it ends
up as:
/usr/x86_64-w64-mingw32/sys-root/mingw/lib/pkgconfig/usr/bin/gdbus-codegen
path which does not exist because it is not shipped by Fedora mingw
packages.

Instead, we can rely on meson find_program() behaviour to do a better
job based on its search order and capabilities.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-14-marcandre.lureau@redhat.com>

ui/surface: allocate shared memory on !win32

Use qemu_memfd_alloc() to allocate the display surface memory, which
will fallback on tmpfile/mmap() on systems without memfd, and allow to
share the display with other processes.

This is similar to how display memory is allocated on win32 since commit
09b4c198 ("console/win32: allocate shareable display surface").

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-13-marcandre.lureau@redhat.com>

ui/dbus: add trace for can_share_map

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-12-marcandre.lureau@redhat.com>

ui/dbus: do not limit to one listener per connection / bus name

This is an arbitrary limitation that doesn't concern QEMU directly and
may make some use cases unnecessarily more complicated.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-11-marcandre.lureau@redhat.com>

ui/pixman: generalize shared_image_destroy

Learn to free memfd-allocated shared memory.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-10-marcandre.lureau@redhat.com>

util/memfd: report potential errors on free

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-9-marcandre.lureau@redhat.com>

ui/dbus: discard pending CursorDefine on new one

Similar to scanout updates, let's discard pending cursor changes.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-8-marcandre.lureau@redhat.com>

ui/dbus: discard display messages on disable

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-7-marcandre.lureau@redhat.com>

ui/dbus: fix filtering all update messages

Filtering pending messages when a new scanout is given shouldn't discard
pending cursor changes, for example.

Since filtering happens in a different thread, use atomic set/get.

Fixes: fa88b85dea ("ui/dbus: filter out pending messages when scanout")
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-6-marcandre.lureau@redhat.com>

ui/win32: fix potential use-after-free with dbus shared memory

DisplaySurface may be free before the pixman image is freed, since the
image is refcounted and used by different objects, including pending
dbus messages.

Furthermore, setting the destroy function in
create_displaysurface_from() isn't appropriate, as it may not be used,
and may be overriden as in ramfb.

Set the destroy function when the shared handle is set, use the HANDLE
directly for destroy data, using a single common helper
qemu_pixman_win32_image_destroy().

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-5-marcandre.lureau@redhat.com>

ui/dbus: fix leak on message filtering

A filter function that wants to drop a message should return NULL, in
which case it must also unref the message itself.

Fixes: fa88b85de ("ui/dbus: filter out pending messages when scanout")
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-4-marcandre.lureau@redhat.com>

hw/audio/hda: fix memory leak on audio setup

When SET_STREAM_FORMAT is called, we should clear the existing setup.

Factor out common function to close a stream.

Direct leak of 144 byte(s) in 3 object(s) allocated from:
    #0 0x7f91d38f7350 in calloc (/lib64/libasan.so.8+0xf7350) (BuildId: a4ad7eb954b390cf00f07fa10952988a41d9fc7a)
    #1 0x7f91d2ab7871 in g_malloc0 (/lib64/libglib-2.0.so.0+0x64871) (BuildId: 36b60dbd02e796145a982d0151ce37202ec05649)
    #2 0x562fa2f447ee in timer_new_full /home/elmarco/src/qemu/include/qemu/timer.h:538
    #3 0x562fa2f4486f in timer_new /home/elmarco/src/qemu/include/qemu/timer.h:559
    #4 0x562fa2f448a9 in timer_new_ns /home/elmarco/src/qemu/include/qemu/timer.h:577
    #5 0x562fa2f47955 in hda_audio_setup ../hw/audio/hda-codec.c:490
    #6 0x562fa2f4897e in hda_audio_command ../hw/audio/hda-codec.c:605

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-3-marcandre.lureau@redhat.com>

hw/audio/hda: free timer on exit

Fixes: 280c1e1cd ("audio/hda: create millisecond timers that handle IO")
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Akihiko Odaki <akihiko.odaki@daynix.com>
Message-ID: <20241008125028.1177932-2-marcandre.lureau@redhat.com>

Merge tag 'pull-tcg-20241013' of https://gitlab.com/rth7680/qemu into staging

linux-user/i386: Emulate orig_ax
linux-user/vm86: Fix compilation with Clang
tcg: remove singlestep_enabled from DisasContextBase
accel/tcg: Add TCGCPUOps.tlb_fill_align
target/hppa: Handle alignment faults in hppa_get_physical_address
target/arm: Fix alignment fault priority in get_phys_addr_lpae

# -----BEGIN PGP SIGNATURE-----
#
# iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmcMRU4dHHJpY2hhcmQu
# aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV9dSQf+MUJq//oig+bDeUlQ
# v3uBMFVi1DBYI1Y/xVODADpn8Ltv5s9v7N+/phi+St2W65OzGNYviHvq/abeyhdo
# M40LGtOvjO6Mns+Z9NKTobtT8n4ap4JJyoFjuXFTHkMMDiQ/v7FkEJJoS3W2bemi
# zmKYF/vWe3bwI+d3+dyaUjA92gSs+Hlj8uEVBlzn3ubA19ZdvtyfKURPQynrkwlo
# dFtAOFRFBU6vrlJSBElxUfYO4jC4Cng19EOrWvIsuKAkACuhiHgah10i3WKw8Asz
# 1iRUYXe0EOlX2RYNTD+Oj5j0cViRylirgPtIhEIPBuDP7m1Jy1JO4dVARUJBBU71
# Zd4Uuw==
# =EX+a
# -----END PGP SIGNATURE-----
# gpg: Signature made Sun 13 Oct 2024 23:10:22 BST
# 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-20241013' of https://gitlab.com/rth7680/qemu: (27 commits)
  target/arm: Fix alignment fault priority in get_phys_addr_lpae
  target/arm: Implement TCGCPUOps.tlb_fill_align
  target/arm: Move device detection earlier in get_phys_addr_lpae
  target/arm: Pass MemOp to get_phys_addr_lpae
  target/arm: Pass MemOp through get_phys_addr_twostage
  target/arm: Pass MemOp to get_phys_addr_nogpc
  target/arm: Pass MemOp to get_phys_addr_gpc
  target/arm: Pass MemOp to get_phys_addr_with_space_nogpc
  target/arm: Pass MemOp to get_phys_addr
  target/hppa: Implement TCGCPUOps.tlb_fill_align
  target/hppa: Handle alignment faults in hppa_get_physical_address
  target/hppa: Fix priority of T, D, and B page faults
  target/hppa: Perform access rights before protection id check
  target/hppa: Add MemOp argument to hppa_get_physical_address
  accel/tcg: Use the alignment test in tlb_fill_align
  accel/tcg: Add TCGCPUOps.tlb_fill_align
  include/exec/memop: Introduce memop_atomicity_bits
  include/exec/memop: Rename get_alignment_bits
  include/exec/memop: Move get_alignment_bits from tcg.h
  accel/tcg: Assert noreturn from write-only page for atomics
  ...

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

target/arm: Fix alignment fault priority in get_phys_addr_lpae

Now that we have the MemOp for the access, we can order
the alignment fault caused by memory type before the
permission fault for the page.

For subsequent page hits, permission and stage 2 checks
are known to pass, and so the TLB_CHECK_ALIGNED fault
raised in generic code is not mis-ordered.

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

target/arm: Implement TCGCPUOps.tlb_fill_align

Fill in the tlb_fill_align hook.  Handle alignment not due to
memory type, since that's no longer handled by generic code.
Pass memop to get_phys_addr.

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

target/arm: Move device detection earlier in get_phys_addr_lpae

Determine cache attributes, and thence Device vs Normal memory,
earlier in the function.  We have an existing regime_is_stage2
if block into which this can be slotted.

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

target/arm: Pass MemOp to get_phys_addr_lpae

Pass the value through from get_phys_addr_nogpc.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/arm: Pass MemOp through get_phys_addr_twostage

Pass memop through get_phys_addr_twostage with its
recursion with get_phys_addr_nogpc.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/arm: Pass MemOp to get_phys_addr_nogpc

Zero is the safe do-nothing value for callers to use.
Pass the value through from get_phys_addr_gpc and
get_phys_addr_with_space_nogpc.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/arm: Pass MemOp to get_phys_addr_gpc

Zero is the safe do-nothing value for callers to use.
Pass the value through from get_phys_addr.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/arm: Pass MemOp to get_phys_addr_with_space_nogpc

Zero is the safe do-nothing value for callers to use.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/arm: Pass MemOp to get_phys_addr

Zero is the safe do-nothing value for callers to use.

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

target/hppa: Implement TCGCPUOps.tlb_fill_align

Convert hppa_cpu_tlb_fill to hppa_cpu_tlb_fill_align so that we
can recognize alignment exceptions in the correct priority order.

Resolves: https://bugzilla.kernel.org/show_bug.cgi?id=219339
Tested-by: Helge Deller <deller@gmx.de>
Reviewed-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/hppa: Handle alignment faults in hppa_get_physical_address

In Chapter 5, Interruptions, the group 3 exceptions lists
"Unaligned data reference trap" has higher priority than
"Data memory break trap".

Reviewed-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/hppa: Fix priority of T, D, and B page faults

Drop the 'else' so that ret is overridden with the
highest priority fault.

Fixes: d8bc1381250 ("target/hppa: Implement PSW_X")
Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/hppa: Perform access rights before protection id check

In Chapter 5, Interruptions, the group 3 exceptions lists
"Data memory access rights trap" in priority order ahead of
"Data memory protection ID trap".

Swap these checks in hppa_get_physical_address.

Reviewed-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/hppa: Add MemOp argument to hppa_get_physical_address

Just add the argument, unused at this point.
Zero is the safe do-nothing value for all callers.

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

accel/tcg: Use the alignment test in tlb_fill_align

When we have a tlb miss, defer the alignment check to
the new tlb_fill_align hook.  Move the existing alignment
check so that we only perform it with a tlb hit.

Reviewed-by: Helge Deller <deller@gmx.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Add TCGCPUOps.tlb_fill_align

Add a new callback to handle softmmu paging.  Return the page
details directly, instead of passing them indirectly to
tlb_set_page.  Handle alignment simultaneously with paging so
that faults are handled with target-specific priority.

Route all calls of the two hooks through a tlb_fill_align
function local to cputlb.c.

As yet no targets implement the new hook.
As yet cputlb.c does not use the new alignment check.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

include/exec/memop: Introduce memop_atomicity_bits

Split out of mmu_lookup.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

include/exec/memop: Rename get_alignment_bits

Rename to use "memop_" prefix, like other functions
that operate on MemOp.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

include/exec/memop: Move get_alignment_bits from tcg.h

This function is specific to MemOp, not TCG in general.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

accel/tcg: Assert noreturn from write-only page for atomics

There should be no "just in case"; the page is already
in the tlb, and known to be not readable.

Reviewed-by: Helge Deller <deller@gmx.de>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/vm86: Fix compilation with Clang

Since commit 95b9c27c81 ("linux-user: Remove unused handle_vm86_fault")
a bunch of other "static inline" function are now unused, too. Clang
warns about such unused "static inline" functions in .c files, so the
build currently breaks when compiling with "--enable-werror". Remove
the unused functions to get it going again.

Fixes: 95b9c27c81 ("linux-user: Remove unused handle_vm86_fault")
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Dr. David Alan Gilbert <dave@treblig.org>
Message-ID: <20241011161845.417342-1-thuth@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tests/tcg: Run test-proc-mappings.py on i386

Now that orig_ax is exposed and GDB is happy, don't skip
test-proc-mappings.py on i386. In fact, it's broken only on
m68k now, so skip only this architecture.

Acked-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <20240912093012.402366-6-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/i386/gdbstub: Expose orig_ax

Copy XML files describing orig_ax from GDB and glue them with
CPUX86State.orig_ax.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <20240912093012.402366-5-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

target/i386/gdbstub: Factor out gdb_get_reg() and gdb_write_reg()

i386 gdbstub handles both i386 and x86_64. Factor out two functions
for reading and writing registers without knowing their bitness.

While at it, simplify the TARGET_LONG_BITS == 32 case.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <20240912093012.402366-4-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

linux-user/i386: Emulate orig_ax

The kernel uses orig_rax/orig_eax to store the syscall number before
a syscall. One can see this value in core dumps and ptrace.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <20240912093012.402366-3-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

include/exec: Introduce env_cpu_const()

It's the same as env_cpu(), but for const objects.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Message-ID: <20240912093012.402366-2-iii@linux.ibm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: remove singlestep_enabled from DisasContextBase

It is used in a couple of places only, both within the same target.
Those can use the cflags just as well, so remove the separate field.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010083641.1785069-1-pbonzini@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

Merge tag 'for-upstream' of https://gitlab.com/bonzini/qemu into staging

* first commit for Rust support
* add CI job using Fedora + Rust nightly
* fix detection of ATOMIC128 on x86_64
* fix compilation with Sphinx 8.1.0

# -----BEGIN PGP SIGNATURE-----
#
# iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmcJEKUUHHBib256aW5p
# QHJlZGhhdC5jb20ACgkQv/vSX3jHroOSZQf+LlvZm9npHR6lZ9DEruhu/uf0c1gO
# 9+dBJiKQ1OWopSQOqEgOsLL0J123Ls4V8O3tzZwIDuuRofCB2+wKswad6CHoydJx
# 4p9rRXv6MLlnTqqGxemm/dPZqJ7+6L0poHoDKW+s7AgfVDshhj1RSbQfs8Ujh41F
# f1sdi3DzopVWtK4CE+8/UeLy5Cxlixke9SKhYQrFHrdsANARP81gxQjczKApMc1z
# v9qkrLtkM06VUyuvbPps7CHSHDpzx9mXcmkkPgLqLX9MfbCztzi44aVSaS9HYk5G
# y54dSKdY7VJEuGhG916G+GMDJyow4nhT9Gk6tWtk63TQN5nExVsoZMOmdw==
# =PFGL
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 11 Oct 2024 12:48:53 BST
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* tag 'for-upstream' of https://gitlab.com/bonzini/qemu:
  docs: use consistent markup for footnotes
  docs: avoid footnotes consisting of just URLs
  docs: fix invalid footnote syntax
  gitlab-ci: add Rust-enabled CI job
  dockerfiles: add a Dockerfile using a nightly Rust toolchain
  meson: ensure -mcx16 is passed when detecting ATOMIC128
  meson: define qemu_isa_flags
  meson: fix machine option for x86_version
  rust: add PL011 device model
  rust: add utility procedural macro crate
  scripts/archive-source: find directory name for subprojects
  rust: add crate to expose bindings and interfaces
  meson.build: add HAVE_GLIB_WITH_ALIGNED_ALLOC flag
  .gitattributes: add Rust diff and merge attributes
  rust: add bindgen step as a meson dependency
  configure, meson: detect Rust toolchain
  build-sys: Add rust feature option
  Require meson version 1.5.0

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

docs: use consistent markup for footnotes

Unfortunately, the definition of the footnote syntax requires
the author to use the awkward escaped space "\ " in the really common
case of "footnote marker at end of word or sentence"; and in fact the rST
documentation's examples of footnote syntax contain only artificial
examples that do *not* use the syntax.  This resulted in ugly rendering
of footnotes throughout QEMU's documentation.  Ensure the space is escaped
whenever the footnote must attach to the preceding word, and also use
a named reference for clarity.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

docs: avoid footnotes consisting of just URLs

Replace the footnotes with inline links whenever the footnote text
consists of nothing but the URL.  While at it, make the link texts
consistent in the surrounding areas, for example avoiding usage of
"here" for the link's text.

In the case of acpi-bits.rst this fixes a build failure with Sphinx
8.1.0, because the FOSDEM link was duplicated in the paragraph and the
new version is a lot stricter about unreferenced footnotes.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

docs: fix invalid footnote syntax

All footnotes must come after a separator in reStructuredText.  Fix the
two files in which this does not happen.

This mistake causes the link to be rendered literally:

   ...from the venv itself[#distlib]_. If no...

and is caught by Sphinx 8.1.0 as an unreferenced footnote.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

gitlab-ci: add Rust-enabled CI job

Add a job using --enable-rust, to ensure that the toolchain is installed
correctly by the Dockerfile and that QEMU builds with Rust enabled on
at least one platform.

Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Acked-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

dockerfiles: add a Dockerfile using a nightly Rust toolchain

This will be useful in order to test QEMU's Rust code with nightly
clippy.  It can also be used to check that the code builds, until the
minimum supported Rust version is lowered enough to allow enabling Rust
in other Dockerfiles too.

Use a separate container, instead of the Fedora one, to avoid that
CI breaks for everyone if for some reason the rustup build turns out
to be shaky.

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Acked-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: ensure -mcx16 is passed when detecting ATOMIC128

Moving -mcx16 out of CPU_CFLAGS caused the detection of ATOMIC128 to
fail, because flags have to be specified by hand in cc.compiles and
cc.links invocations (why oh why??).

Ensure that these tests enable all the instruction set extensions that
will be used to build the emulators.

Fixes: c2bf2ccb266 ("configure: move -mcx16 flag out of CPU_CFLAGS", 2024-05-24)
Reported-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: define qemu_isa_flags

Create a separate variable for compiler flags that enable
specific instruction set extensions, so that they can be used with
cc.compiles/cc.links.

Note that -mfpmath=sse is a code generation option but it does not
enable new instructions, therefore I did not make it part of
qemu_isa_flags.

Suggested-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson: fix machine option for x86_version

s/mbmi1/mbmi/

When configuring with -Dx86_version >= 3, meson step works, but
compilation fails because option -mbmi1 is unknown.

Signed-off-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Link: https://lore.kernel.org/r/20241004223715.1275428-1-pierrick.bouvier@linaro.org
Cc: qemu-stable@nongnu.org
Fixes: ef7d1adfa85 ("meson: allow configuring the x86-64 baseline", 2024-06-28)
Revieved-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

rust: add PL011 device model

This commit adds a re-implementation of hw/char/pl011.c in Rust.

How to build:

1. Configure a QEMU build with:
   --enable-system --target-list=aarch64-softmmu --enable-rust
2. Launching a VM with qemu-system-aarch64 should use the Rust version
   of the pl011 device

Co-authored-by: Junjie Mao <junjie.mao@intel.com>
Co-authored-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Link: https://lore.kernel.org/r/6ec1d4fb8db2a1d7ba94c73e65d9770371b7857d.1727961605.git.manos.pitsidianakis@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

rust: add utility procedural macro crate

This commit adds a helper crate library, qemu-api-macros for derive (and
other procedural) macros to be used along qemu-api.

It needs to be a separate library because in Rust, procedural macros, or
macros that can generate arbitrary code, need to be special separate
compilation units.

Only one macro is introduced in this patch, #[derive(Object)]. It
generates a constructor to register a QOM TypeInfo on init and it must
be used on types that implement qemu_api::definitions::ObjectImpl trait.

Reviewed-by: Junjie Mao <junjie.mao@hotmail.com>
Signed-off-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Link: https://lore.kernel.org/r/dd645642406a6dc2060c6f3f17db2bc77ed67b59.1727961605.git.manos.pitsidianakis@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

scripts/archive-source: find directory name for subprojects

Rust subprojects have the semantic version (followed by -rs) in the subproject
name, but the full version (without -rs) is used by crates.io for the root
directory of the tarball.  Teach scripts/archive-source.sh to look for the
root directory name in wrap files.

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

rust: add crate to expose bindings and interfaces

Add rust/qemu-api, which exposes rust-bindgen generated FFI bindings and
provides some declaration macros for symbols visible to the rest of
QEMU.

Co-authored-by: Junjie Mao <junjie.mao@intel.com>
Co-authored-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Link: https://lore.kernel.org/r/0fb23fbe211761b263aacec03deaf85c0cc39995.1727961605.git.manos.pitsidianakis@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

meson.build: add HAVE_GLIB_WITH_ALIGNED_ALLOC flag

Rust crates, introduced from the next commit onwards, can optionally use
the glib allocator API and need to know whether g_aligned_alloc etc are
available.

This commit adds a define in config_host_data that depends on glib
version >= 2.72.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Link: https://lore.kernel.org/r/23f7b0cc9801d315f5d7835e30d775e133ec2fb9.1727961605.git.manos.pitsidianakis@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

.gitattributes: add Rust diff and merge attributes

Set rust source code to diff=rust (built-in with new git versions)
and merge=binary for Cargo.lock files (they should not be merged but
auto-generated by cargo)

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Signed-off-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Link: https://lore.kernel.org/r/278e3878b40ecc7c424dec1ed978eedf21469f52.1727961605.git.manos.pitsidianakis@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

rust: add bindgen step as a meson dependency

Add bindings_rs target for generating rust bindings to target-independent
qemu C APIs.

The bindings need be created before any rust crate that uses them is
compiled.

The bindings.rs file will end up in BUILDDIR/bindings.rs and have the
same name as a target:

  ninja bindings.rs

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Link: https://lore.kernel.org/r/1be89a27719049b7203eaf2eca8bbb75b33f18d4.1727961605.git.manos.pitsidianakis@linaro.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

configure, meson: detect Rust toolchain

Include the correct path and arguments to rustc in the native
and cross files (native compilation is needed for procedural
macros).

Based on the host architecture and OS, the compiler and optionally the argument
to --cpu, the Rust target triple can be detected automatically for either a
native or a cross compiler.

In general, it is only a matter of translating the architecture and OS, and
adding a machine to form the triple, but there are some special cases (e.g.
detecting soft vs. hard floating point on ARM) and some inconsistencies.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Link: https://lore.kernel.org/r/207d2640b32d511e9c27478ce3192f5bb0bf3169.1727961605.git.manos.pitsidianakis@linaro.org
[Leave disabled by default until CI covers the Rust code on supported
 distros. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Merge tag 'crypto-fixes-pull-request' of https://gitlab.com/berrange/qemu into staging

Introduce new cryptography hashing APIs

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEE2vOm/bJrYpEtDo4/vobrtBUQT98FAmcH/iIACgkQvobrtBUQ
# T9+Yjg/+NReYV5BDjOLk6vfgTsK6Ku0/hdis2cf9OS8Ud1VXzKaxfhwkchtw9QVI
# kuAthesQNocEPfQfl2K4+f4oaKfysO7awDwYto/JhY/m1iCZ8iqofZWehOITszvM
# EvWlNBr83NtpGFIwQWIxFEVZo42gaUnA69iAjBo7YQnE5xufJuPIbgMjB/O4/zar
# Xlo15A69TP9dBJTvIDdrhkt3Quiysa7a68BW+piAAKvplOjOfugCEo3ebLwlZYOh
# dK0Cg9v24+BMAqQ7kDMroS4uHC+OEs2AOvfYh01QqWxNkk7RsPjb9VAA60Ng89eC
# 6BU4jw17zUAqL67of+M1cTTX4UPGBWGIUXt8CtO1DpByxiGXXfEkBrBmIyDJvxn9
# EzB4WpAXpVo2AG6vYpYSBGyxycWQs33ljfBb/qR6xu5PnA+Jc/jfJkVv5iYP96wW
# F6pJm6FoK69aTJU7K4kAJPjD2fZum+iHVWc283NIkq9HQJLz2EYE0LIfOOY5feJK
# S0tjEE5ZLqKG5JAdpsaCe5V/vExc512/D56Xb5fY4mC2DPb/b6fM66Oc5M7DTuK1
# LxCgnEuqm1Lo3CMR0k4W8Xezs7hWp+u3tr+i705l5qFxklYkmFeVAzTWdQ56JOGk
# Z1XKUbcPUnweormPMxMQXyxXpey4DBwUGbjC98iqE8tjUg6NA3o=
# =yVgk
# -----END PGP SIGNATURE-----
# gpg: Signature made Thu 10 Oct 2024 17:17:38 BST
# gpg:                using RSA key DAF3A6FDB26B62912D0E8E3FBE86EBB415104FDF
# gpg: Good signature from "Daniel P. Berrange <dan@berrange.com>" [full]
# gpg:                 aka "Daniel P. Berrange <berrange@redhat.com>" [full]
# Primary key fingerprint: DAF3 A6FD B26B 6291 2D0E  8E3F BE86 EBB4 1510 4FDF

* tag 'crypto-fixes-pull-request' of https://gitlab.com/berrange/qemu:
  tests/unit: Add a assert for test_io_channel_unix_listen_cleanup
  crypto: drop obsolete back compat logic for old nettle
  crypto/hashpriv: Remove old hash API function
  crypto/hash-afalg: Remove old hash API functions
  crypto/hash-nettle: Remove old hash API functions
  crypto/hash-gnutls: Remove old hash API functions
  crypto/hash-gcrypt: Remove old hash API functions
  crypto/hash-glib: Remove old hash API functions
  tests/unit/test-crypto-hash: accumulative hashing
  crypto/hash: Implement and use new hash API
  crypto/hash-afalg: Implement new hash API
  util/iov: Introduce iov_send_recv_with_flags()
  crypto/hash-nettle: Implement new hash API
  crypto/hash-gnutls: Implement new hash API
  crypto/hash-gcrypt: Implement new hash API
  crypto/hash-glib: Implement new hash API
  crypto: accumulative hashing API

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

tests/unit: Add a assert for test_io_channel_unix_listen_cleanup

Calling bind without checking return value. Add a assert for it.

Signed-off-by: Kunwu <chentao@kylinos.cn>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: drop obsolete back compat logic for old nettle

The nettle 2.x series declared all the hash functions with 'int' for
the data size. Since we dropped support for anything older than 3.4
we can assume nettle is using 'size_t' and thus avoid the back compat
looping logic.

Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hashpriv: Remove old hash API function

Remove old hash_bytesv function, as it was replaced by the 4
new functions.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-afalg: Remove old hash API functions

Removes the old hash API functions in the afalg driver,
and modifies the hmac function to use the new helper functions.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[ clg: - Checkpatch fixes ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-nettle: Remove old hash API functions

Removes old hash implementation in the nettle hash driver.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[ clg: - Fixed spelling in commit log ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-gnutls: Remove old hash API functions

Removes old hash implementation in the gnutls hash driver.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[ clg: - Fixed spelling in commit log ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-gcrypt: Remove old hash API functions

Removes old hash implementation in the gcrypt hash driver.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[ clg: - Fixed spelling in commit log ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-glib: Remove old hash API functions

Removes old hash implement-ion in the GLib hash driver.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
[ clg: - Fixed spelling in commit log ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

tests/unit/test-crypto-hash: accumulative hashing

Added an accumulative hashing test. Checks for functionality of
the new hash create, update, finalize and free functions.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Improved test_hash_accumulate() with g_autofree variables ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash: Implement and use new hash API

Changes the public hash API implementation to support accumulative hashing.

Implementations for the public functions are added to call the new
driver functions that implement context creation, updating,
finalization, and destruction.

Additionally changes the "shortcut" functions to use these 4 new core
functions.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Reworked qcrypto_hash_bytesv() error handling
       - Used hash->driver int qcrypto_hash_new(), qcrypto_hash_free()
         qcrypto_hash_updatev()
       - Introduced qcrypto_hash_supports() check in
         qcrypto_hash_new()
       - Introduced g_autofree variables in qcrypto_hash_finalize_digest()
         and qcrypto_hash_finalize_base64()
       - Re-arrranged code in qcrypto_hash_digestv() and
         qcrypto_hash_digest()
       - Checkpatch fixes ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-afalg: Implement new hash API

Updates the afalg hash driver to support the new accumulative
hashing changes as part of the patch series.

Implements opening/closing of contexts, updating hash data
and finalizing the hash digest.

In order to support the update function, a flag needs to be passed
to the kernel via the socket send call (MSG_MORE) to notify it that more
data is to be expected to calculate the hash correctly.
As a result, a new function was added to the iov helper utils to allow
passing a flag to the socket send call.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Handled qcrypto_afalg_hash_ctx_new() errors in
         qcrypto_afalg_hash_new()
       - Freed alg_name in qcrypto_afalg_hash_new()
       - Reworked qcrypto_afalg_recv_from_kernel()
       - Split iov changes from original patch ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

util/iov: Introduce iov_send_recv_with_flags()

In order to support a new update function, a flag needs to be passed
to the kernel via the socket send call (MSG_MORE) to notify it that
more data is to be expected to calculate the hash correctly.

Add a new iov helper for this purpose.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Split iov changes from original patch
       - Checkpatch fixes ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-nettle: Implement new hash API

Implements the new hashing API in the nettle hash driver.
Supports creating/destroying a context, updating the context
with input data and obtaining an output hash.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Dropped qcrypto_hash_supports() in qcrypto_nettle_hash_new() ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-gnutls: Implement new hash API

Implements the new hashing API in the gnutls hash driver.
Supports creating/destroying a context, updating the context
with input data and obtaining an output hash.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Dropped qcrypto_hash_supports() in qcrypto_gnutls_hash_new()
       - Reworked qcrypto_gnutls_hash_finalize()
       - Handled gnutls_hash_init() errors in qcrypto_gnutls_hash_new()
       - Replaced gnutls_hash_deinit() by gnutls_hash_output() in
         qcrypto_gnutls_hash_finalize()
       - Freed resources with gnutls_hash_deinit() in
         qcrypto_gnutls_hash_free() ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-gcrypt: Implement new hash API

Implements the new hashing API in the gcrypt hash driver.
Supports creating/destroying a context, updating the context
with input data and obtaining an output hash.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Dropped qcrypto_hash_supports() in qcrypto_gcrypt_hash_new()
       - Reworked qcrypto_gcrypt_hash_finalize()
       - Handled gcry_md_open() errors in qcrypto_gcrypt_hash_new()
       - Checkpatch fixes ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto/hash-glib: Implement new hash API

Implements the new hashing API in the GLib hash driver.
Supports creating/destroying a context, updating the context
with input data and obtaining an output hash.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Dropped qcrypto_hash_supports() in qcrypto_glib_hash_new()
       - Removed superfluous cast (GChecksum *) in qcrypto_glib_hash_free()
       - Reworked qcrypto_glib_hash_finalize() ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

crypto: accumulative hashing API

Changes the hash API to support accumulative hashing.
Hash objects are created with "qcrypto_hash_new",
updated with data with "qcrypto_hash_update", and
the hash obtained with "qcrypto_hash_finalize".

These changes bring the hashing API more in line with the
hmac API.

Signed-off-by: Alejandro Zeise <alejandro.zeise@seagate.com>
[ clg: - Changed documentation "non-zero on error" -> "-1 on error" ]
Signed-off-by: Cédric Le Goater <clg@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>

Merge tag 'migration-20241009-pull-request' of https://gitlab.com/peterx/qemu into staging

Migration pull request

- Ani's patch to complete the memory API on coalesced IO / eventfd notifies
- Fabiano's Coverity fix on using pstrcpy() over strncpy()
- Dave's series on removing/deprecating zero-blocks and uffd cleanups
- Juraj's one more fix on multifd/cancel test where it can fail when
  cancellation happens too slow on src
- Dave's one more remove deadcode patch in iova-tree.c
- Yuan's build fix for multifd qpl compressor

# -----BEGIN PGP SIGNATURE-----
#
# iIgEABYKADAWIQS5GE3CDMRX2s990ak7X8zN86vXBgUCZwZ6CBIccGV0ZXJ4QHJl
# ZGhhdC5jb20ACgkQO1/MzfOr1wa3ZwD9HiAN9m7WOfZxXKOVIIwhOjUNTw0FiFeO
# HMxp8A2jeYsBAK+d5lYGX1V2FtQ152YiOJQzRW31MkdAOishJzcHCXgO
# =gBW0
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 09 Oct 2024 13:41:44 BST
# gpg:                using EDDSA key B9184DC20CC457DACF7DD1A93B5FCCCDF3ABD706
# gpg:                issuer "peterx@redhat.com"
# gpg: Good signature from "Peter Xu <xzpeter@gmail.com>" [marginal]
# gpg:                 aka "Peter Xu <peterx@redhat.com>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: B918 4DC2 0CC4 57DA CF7D  D1A9 3B5F CCCD F3AB D706

* tag 'migration-20241009-pull-request' of https://gitlab.com/peterx/qemu:
  migration/multifd: fix build error when qpl compression is enabled
  util/iova-tree: Remove deadcode
  tests/migration-test: Wait for cancellation sooner in multifd cancel
  util/userfaultfd: Remove unused uffd_poll_events
  migration/postcopy: Use uffd helpers
  util/userfaultfd: Return -errno on error
  migration: Remove unused socket_send_channel_create_sync
  migration: Deprecate zero-blocks capability
  migration: Remove unused migrate_zero_blocks
  migration: Remove migrate_cap_set
  migration/multifd: Ensure packet->ramblock is null-terminated
  memory: notify hypervisor of all eventfds during listener (de)registration

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

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

chardev: introduce 'reconnect-ms' and deprecate 'reconnect'
chardev: add path option for pty backend

# -----BEGIN PGP SIGNATURE-----
#
# iQJQBAABCAA6FiEEh6m9kz+HxgbSdvYt2ujhCXWWnOUFAmcGQTYcHG1hcmNhbmRy
# ZS5sdXJlYXVAcmVkaGF0LmNvbQAKCRDa6OEJdZac5crYEACdw2EjdHm+OSkC2cUt
# p9x0spZ8xkUJpU7f7Xc+9fIYHTqqmtgKhbC2/iSaeeL8aSN5LugOAP7u0ya9gXo2
# M1pXPx4WYwy7yftmanwVSvD94GPiCT5EN2W3zFMzvkey9b2AdayGyCXaZY2ago4+
# QJq0pPwWu5VFR0b3ocXIJM1WiEzbR6wr+R0xuPGMLAp0uOnNIyE8AD9MhHJ/BROB
# dDCD7xhG6MB1CnRG6+saGV/Aon9mml5i5MmiwmQ6JjC0zZ0w4arkWfjkW5r+yZOc
# BrpC6P+MdwX7t7W//4bszlW0Lv+qw9Q9FBavtLrOSTCy/h+cq5XgNjZnf0j0edNp
# /EGwQ6D7lqbHp0fGX1O+vnWe0IcyGMKWTU05UGpb3TBDzfzlmjNinQxU6LZPfizh
# evRkKELRsy+WTOyJzyeKemw/SdoP8o7RmDF9x3uCP6dF8Q8/UD3CgnwVa8h9L+Jz
# 591Ek6srXQhEHkUJdpf48LHvCho4eWkpIy2M7/iurXFZBhpz77A27DCnfBedYg/O
# f67MreVDz3C/RSlsnGj7/c0eKnZHPuNhI4AxCAH01bCy1PEv65LrWUdeAtJRm84X
# upbn9alqP4LCviUBIwjDqSDhb48mYbgmP8sC3oXeyZObSGQUz94oXh1QQapUHSzz
# BCImJN+feMFgLLuEXreJP1N/kQ==
# =musy
# -----END PGP SIGNATURE-----
# gpg: Signature made Wed 09 Oct 2024 09:39:18 BST
# gpg:                using RSA key 87A9BD933F87C606D276F62DDAE8E10975969CE5
# gpg:                issuer "marcandre.lureau@redhat.com"
# gpg: Good signature from "Marc-André Lureau <marcandre.lureau@redhat.com>" [full]
# gpg:                 aka "Marc-André Lureau <marcandre.lureau@gmail.com>" [full]
# Primary key fingerprint: 87A9 BD93 3F87 C606 D276  F62D DAE8 E109 7596 9CE5

* tag 'chr-pull-request' of https://gitlab.com/marcandre.lureau/qemu:
  chardev: add path option for pty backend
  chardev: introduce 'reconnect-ms' and deprecate 'reconnect'

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

Merge tag 'pull-tcg-20241008' of https://gitlab.com/rth7680/qemu into staging

linux-user: Fix parse_elf_properties GNU0_MAGIC check
linux-user: Various improvements to strace
linux-user: Add openat2 support
linux-user/flatload: Take mmap_lock in load_flt_binary()
accel/tcg: Make page_set_flags() documentation public
tcg/ppc: Use TCG_REG_TMP2 for scratch tcg_out_qemu_st
tcg/ppc: Use TCG_REG_TMP2 for scratch index in prepare_host_addr
target/m68k: Always return a temporary from gen_lea_mode

# -----BEGIN PGP SIGNATURE-----
#
# iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmcFeLQdHHJpY2hhcmQu
# aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV+8Wgf9FJb/71zPAg7idUo8
# k+P5hOmE0PzPKdkoCh5WP0+TF4m/gMc3ABMFO2ABx4EcAkWxUmNP1Po5Lb0/A9Gq
# LuSDRJVk3jSeGebmMsIH0SVzKY3ewQv7ZGB6A2uFdXtkwDUCxgHUQcfQY9yvjFut
# eU11zWkPbmbI4bwz9Y+Tyv9LOXRk/Br5B/uQEzQ9/Go/0OeJP0iTXre/veGufqND
# wK35AoDCVNomjw6h2Fg0t3kCt/mtR8W8D16+4QhsH+iE9IQMmhj4itz8Uq2ikh29
# lQIUT8acHC8fNUQJXw0nPXbNsimp/GyRGJiYCH7g7V67o/fjDUSbYMxk2TMZPuJV
# dGy3KQ==
# =UeX8
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 08 Oct 2024 19:23:48 BST
# 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-20241008' of https://gitlab.com/rth7680/qemu:
  accel/tcg: Make page_set_flags() documentation public
  target/m68k: Always return a temporary from gen_lea_mode
  tcg/ppc: Use TCG_REG_TMP2 for scratch index in prepare_host_addr
  tcg/ppc: Use TCG_REG_TMP2 for scratch tcg_out_qemu_st
  linux-user: Add strace for recvfrom()
  linux-user: Add strace for sendto()
  linux-user: Factor print_buf_len() out
  linux-user: Display sockaddr buffer as pointer
  linux-user: Correct print_sockaddr() format
  linux-user: Trace wait4()'s and waitpid()'s wstatus
  linux-user: add strace support for openat2
  linux-user: add openat2 support in linux-user
  linux-user: Fix parse_elf_properties GNU0_MAGIC check
  linux-user/flatload: Take mmap_lock in load_flt_binary()

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

tests/qtest: Bump qmp-cmd-test timeout to 120s

The qmp-cmd-test test takes typically about 15s on my local machine.
On the k8s runners it takes usually 20s but sometimes about 60s,
because the k8s runners have wildly variable execution time.  If
they're running slow, we hit the default timeout.  Bump the
qmp-cmd-test timeout to 120s to avoid this.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-id: 20241008141337.2790423-1-peter.maydell@linaro.org

migration/multifd: fix build error when qpl compression is enabled

The page_size member has been removed from the MultiFDSendParams
and MultiFDRecvParams. The function multifd_ram_page_size is used to
provide the page size in the multifd compressor.

Fixes: 90fa121c6c ("migration/multifd: Inline page_size and page_count")
Signed-off-by: Yuan Liu <yuan1.liu@intel.com>
Link: https://lore.kernel.org/r/20241008104527.3516755-1-yuan1.liu@intel.com
Signed-off-by: Peter Xu <peterx@redhat.com>

chardev: add path option for pty backend

Add path option to the pty char backend which will create a symbolic
link to the given path that points to the allocated PTY.

This avoids having to make QMP or HMP monitor queries to find out what
the new PTY device path is.

Based on patch from Paulo Neves:

https://patchew.org/QEMU/1548509635-15776-1-git-send-email-ptsneves@gmail.com/

Tested with the following invocations that the link is created and
removed when qemu stops:

  qemu-system-x86_64 -nodefaults -mon chardev=compat_monitor \
  -chardev pty,path=test,id=compat_monitor0

  qemu-system-x86_64 -nodefaults -monitor pty:test

  # check QMP invocation with path set
  qemu-system-x86_64 -nodefaults -qmp tcp:localhost:4444,server=on,wait=off
  nc localhost 4444
  > {"execute": "qmp_capabilities"}
  > {"execute": "chardev-add", "arguments": {"id": "bar", "backend": {
      "type": "pty", "data": {"path": "test" }}}}

  # check QMP invocation with path not set
  qemu-system-x86_64 -nodefaults -qmp tcp:localhost:4444,server=on,wait=off
  nc localhost 4444
  > {"execute": "qmp_capabilities"}
  > {"execute": "chardev-add", "arguments": {"id": "bar", "backend": {
      "type": "pty", "data": {}}}}

Also tested that when a link path is not passed invocations still work, e.g.:

  qemu-system-x86_64 -monitor pty

Co-authored-by: Paulo Neves <ptsneves@gmail.com>
Signed-off-by: Paulo Neves <ptsneves@gmail.com>
[OP: rebase and address original patch review comments]
Signed-off-by: Octavian Purdila <tavip@google.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-ID: <20240806010735.2450555-1-tavip@google.com>

chardev: introduce 'reconnect-ms' and deprecate 'reconnect'

The 'reconnect' option only allows to specify the time in seconds,
which is way too long for certain workflows.

We have a lightweight disk backend server, which takes about 20ms to
live update, but due to this limitation in QEMU, previously the guest
disk controller would hang for one second because it would take this
long for QEMU to reinitialize the socket connection.

Introduce a new option called 'reconnect-ms', which is the same as
'reconnect', except the value is treated as milliseconds. These are
mutually exclusive and specifying both results in an error.

'reconnect' is also deprecated by this commit to make it possible to
remove it in the future as to not keep two options that control the
same thing.

Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
Acked-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
Acked-by: Markus Armbruster <armbru@redhat.com>
Tested-by: Markus Armbruster <armbru@redhat.com>
Message-ID: <20240913094604.269135-1-d-tatianin@yandex-team.ru>

util/iova-tree: Remove deadcode

iova_tree_find_address, and iova_tree_foreach have never been
used since the code was originally added by:
  eecf5eedbd ("util: implement simple iova tree")

Remove them.

Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Link: https://lore.kernel.org/r/20240918142515.153074-1-dave@treblig.org
Signed-off-by: Peter Xu <peterx@redhat.com>

tests/migration-test: Wait for cancellation sooner in multifd cancel

The source QEMU might not finish the cancellation of the migration
before we start setting up the next attempt. During the setup, the
test_migrate_start() function and others might need to interact with the
source in a way that is not possible unless the migration is fully
canceled. For example, setting capabilities when the migration is still
running leads to an error.

By moving the wait before the setup, we ensure this does not happen.

Cc: Peter Xu <peterx@redhat.com>
Signed-off-by: Juraj Marcin <jmarcin@redhat.com>
Link: https://lore.kernel.org/r/20240920161319.2337625-1-jmarcin@redhat.com
Signed-off-by: Peter Xu <peterx@redhat.com>

util/userfaultfd: Remove unused uffd_poll_events

uffd_poll_events has been unused since it was added; it's also
just a wrapper around a plain old poll call, so doesn't add anything.

Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Link: https://lore.kernel.org/r/20240919134626.166183-8-dave@treblig.org
Signed-off-by: Peter Xu <peterx@redhat.com>

migration/postcopy: Use uffd helpers

Use the uffd_copy_page, uffd_zero_page and uffd_wakeup helpers
rather than calling ioctl ourselves.

They return -errno on error, and print an error_report themselves.
I think this actually makes postcopy_place_page actually more
consistent in it's callers.

Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Link: https://lore.kernel.org/r/20240919134626.166183-7-dave@treblig.org
[peterx: fix i386 build]
Signed-off-by: Peter Xu <peterx@redhat.com>

util/userfaultfd: Return -errno on error

Convert (the currently unused) uffd_wakeup, uffd_copy_page and
uffd_zero_page to return -errno on error rather than -1.

That will make it easier to reuse in postcopy.

Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Link: https://lore.kernel.org/r/20240919134626.166183-6-dave@treblig.org
Signed-off-by: Peter Xu <peterx@redhat.com>

migration: Remove unused socket_send_channel_create_sync

socket_send_channel_create_sync only use was removed by
  d0edb8a173 ("migration: Create the postcopy preempt channel asynchronously")

Remove it.

Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Link: https://lore.kernel.org/r/20240919134626.166183-5-dave@treblig.org
Signed-off-by: Peter Xu <peterx@redhat.com>

migration: Deprecate zero-blocks capability

The zero-blocks capability was meant to be used along with the block
migration, which has been removed already in commit eef0bae3a7
("migration: Remove block migration").

Setting zero-blocks is currently a noop, but the outright removal of
the capability would cause and error in case some users are still
setting it. Put the capability through the deprecation process.

Signed-off-by: Fabiano Rosas <farosas@suse.de>
Reviewed-by: Peter Xu <peterx@redhat.com>
Link: https://lore.kernel.org/r/20240919134626.166183-4-dave@treblig.org
Signed-off-by: Peter Xu <peterx@redhat.com>

migration: Remove unused migrate_zero_blocks

migrate_zero_blocks is unused since
  eef0bae3a7 ("migration: Remove block migration")

Remove it.

Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Link: https://lore.kernel.org/r/20240919134626.166183-3-dave@treblig.org
Signed-off-by: Peter Xu <peterx@redhat.com>

migration: Remove migrate_cap_set

migrate_cap_set has been unused since
  18d154f575 ("migration: Remove 'blk/-b' option from migrate commands")

Remove it.

Signed-off-by: Dr. David Alan Gilbert <dave@treblig.org>
Reviewed-by: Fabiano Rosas <farosas@suse.de>
Link: https://lore.kernel.org/r/20240919134626.166183-2-dave@treblig.org
Signed-off-by: Peter Xu <peterx@redhat.com>

migration/multifd: Ensure packet->ramblock is null-terminated

Coverity points out that the current usage of strncpy to write the
ramblock name allows the field to not have an ending '\0' in case
idstr is already not null-terminated (e.g. if it's larger than 256
bytes).

This is currently harmless because the packet->ramblock field is never
touched again on the source side. The destination side reads only up
to the field's size from the stream and forces the last byte to be 0.

We're still open to a programming error in the future in case this
field is ever passed into a function that expects a null-terminated
string.

Change from strncpy to QEMU's pstrcpy, which puts a '\0' at the end of
the string and doesn't fill the extra space with zeros.

(there's no spillage between iterations of fill_packet because after
commit 87bb9e953e ("migration/multifd: Isolate ram pages packet data")
the packet is always zeroed before filling)

Resolves: Coverity CID 1560071
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Link: https://lore.kernel.org/r/20240919150611.17074-1-farosas@suse.de
Signed-off-by: Peter Xu <peterx@redhat.com>