tcg/ppc: Drop fallback constant loading in tcg_out_cmp

Use U and C constraints for brcond2 and setcond2, so that tcg_out_cmp2
automatically passes in-range constants to tcg_out_cmp.

Tested-by: Nicholas Piggin <npiggin@gmail.com>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_movcond_{i32,i64}

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

tcg: Convert movcond to TCGOutOpMovcond

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_brcond_{i32,i64}

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

tcg: Convert brcond to TCGOutOpBrcond

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_{neg}setcond_{i32,i64}`

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Convert setcond, negsetcond to TCGOutOpSetcond

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Remove TCG_TARGET_HAS_negsetcond_{i32,i64}

All targets now provide negsetcond, so remove the conditional.

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

tcg/tci: Support negsetcond

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

tcg/mips: Support negsetcond

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/loongarch64: Support negsetcond

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_mulu2_{i32,i64}

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

tcg: Convert mulu2 to TCGOutOpMul2

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_muls2_{i32,i64}

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

tcg: Convert muls2 to TCGOutOpMul2

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_ctpop_{i32,i64}

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Convert ctpop to TCGOutOpUnary

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_ctz_{i32,i64}

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

tcg: Convert ctz to TCGOutOpBinary

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_clz_{i32,i64}

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

tcg: Convert clz to TCGOutOpBinary

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_rot{l,r}_{i32,i64}

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Convert rotl, rotr to TCGOutOpBinary

For aarch64, arm, loongarch64, mips, we can drop rotl.
For ppc, s390x we can drop rotr.
Only x86, riscv, tci have both rotl and rotr.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Do not require both rotr and rotl from the backend

Many host architectures do not implement both rotate right
and rotate left and require the compiler to negate the
shift count to rotate the opposite direction.  We have been
requiring the backend to perform this transformation.
Do this during opcode expansion so that the next patch
can drop support where possible in the backend.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_sar_{i32,i64}

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

tcg: Convert sar to TCGOutOpBinary

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_shr_{i32,i64}

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

tcg: Convert shr to TCGOutOpBinary

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_shl_{i32,i64}

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

tcg: Convert shl to TCGOutOpBinary

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

tcg: Merge INDEX_op_remu_{i32,i64}

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

tcg: Convert remu to TCGOutOpBinary

For TCI, we're losing type information in the interpreter.
Introduce a tci-specific opcode to handle the difference.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_rem_{i32,i64}

Rename to INDEX_op_rems to emphasize signed inputs,
and mirroring INDEX_op_remu_*.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Convert rem to TCGOutOpBinary

For TCI, we're losing type information in the interpreter.
Introduce a tci-specific opcode to handle the difference.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_divu2_{i32,i64}

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

tcg: Convert divu2 to TCGOutOpDivRem

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_div2_{i32,i64}

Rename to INDEX_op_divs2 to emphasize signed inputs,
and mirroring INDEX_op_divu2_*.  Document the opcode.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Convert div2 to TCGOutOpDivRem

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

tcg: Merge INDEX_op_divu_{i32,i64}

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

tcg: Convert divu to TCGOutOpBinary

For TCI, we're losing type information in the interpreter.
Introduce a tci-specific opcode to handle the difference.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_div_{i32,i64}

Rename to INDEX_op_divs to emphasize signed inputs,
and mirroring INDEX_op_divu_*.

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

tcg: Convert div to TCGOutOpBinary

For TCI, we're losing type information in the interpreter.
Introduce a tci-specific opcode to handle the difference.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg: Merge INDEX_op_mulsh_{i32,i64}

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

tcg: Convert mulsh to TCGOutOpBinary

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

tcg: Merge INDEX_op_muluh_{i32,i64}

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

tcg: Convert muluh to TCGOutOpBinary

Remove unreachable mul[su]h_i32 leftovers from commit aeb6326ec5e
("tcg/riscv: Require TCG_TARGET_REG_BITS == 64").

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

tcg: Merge INDEX_op_mul_{i32,i64}

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

tcg: Convert mul to TCGOutOpBinary

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

tcg: Merge INDEX_op_not_{i32,i64}

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

tcg: Convert not to TCGOutOpUnary

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

tcg: Merge INDEX_op_neg_{i32,i64}

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

tcg: Convert neg to TCGOutOpUnary

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

tcg: Merge INDEX_op_sub_{i32,i64}

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

tcg: Convert sub to TCGOutOpSubtract

Create a special subclass for sub, because two backends can
support "subtract from immediate".  Drop all backend support
for an immediate as the second operand, as we transform sub
to add during optimize.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/arm: Fix constraints for sub

In 7536b82d288 we lost the rI constraint that allowed the use of
RSB to perform reg = imm - reg.  At the same time, drop support
for reg = reg - imm, which is now transformed generically to
addition, and need not be handled by the backend.

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

tcg: Merge INDEX_op_nor_{i32,i64}

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

tcg: Convert nor to TCGOutOpBinary

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

tcg/loongarch64: Do not accept constant argument to nor

The instruction set does not implement nor with immediate.
There is no reason to pretend that we do.

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

tcg: Merge INDEX_op_nand_{i32,i64}

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

tcg: Convert nand to TCGOutOpBinary

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

tcg: Merge INDEX_op_eqv_{i32,i64}

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

tcg: Convert eqv to TCGOutOpBinary

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

tcg/optimize: Fold eqv with immediate to xor

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

tcg: Merge INDEX_op_xor_{i32,i64}

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

tcg: Convert xor to TCGOutOpBinary

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

tcg: Merge INDEX_op_orc_{i32,i64}

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

tcg: Convert orc to TCGOutOpBinary

At the same time, drop all backend support for immediate
operands, as we now transform orc to or during optimize.

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

tcg/optimize: Fold orc with immediate to or

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

tcg: Merge INDEX_op_or_{i32,i64}

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

tcg: Convert or to TCGOutOpBinary

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

tcg: Merge INDEX_op_andc_{i32,i64}

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

tcg: Convert andc to TCGOutOpBinary

At the same time, drop all backend support for immediate
operands, as we now transform andc to and during optimize.

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

tcg/optimize: Emit add r,r,-1 in fold_setcond_tst_pow2

We canonicalize subtract with constant to add with constant.
Fix this missed instance.

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

tcg/optimize: Fold andc with immediate to and

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

tcg: Merge INDEX_op_and_{i32,i64}

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

tcg: Convert and to TCGOutOpBinary

Drop all backend support for an immediate as the first operand.
This should never happen in any case, as we swap commutative
operands to place immediates as the second operand.

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

tcg: Merge INDEX_op_add_{i32,i64}

Rely on TCGOP_TYPE instead of opcodes specific to each type.

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

tcg: Convert add to TCGOutOpBinary

Drop all backend support for an immediate as the first operand.
This should never happen in any case, as we swap commutative
operands to place immediates as the second operand.

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

tcg: Merge INDEX_op_mov_{i32,i64}

Begin to rely on TCGOp.type to discriminate operations,
rather than two different opcodes.  Convert mov first.
Introduce TCG_OPF_INT in order to keep opcode dumps the same.

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

tcg: Remove INDEX_op_ext{8,16,32}*

Use the fully general extract opcodes instead.

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

tcg: Use extract2 for cross-word 64-bit extract on 32-bit host

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

tcg: Add all_outop[]

Add infrastructure for more consolidated output of opcodes.
The base structure allows for constraints to be either static
or dynamic, and for the existence of those constraints to
replace TCG_TARGET_HAS_* and the bulk of tcg_op_supported.

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

tcg: Add TCGType to tcg_op_insert_{after,before}

We cannot rely on the value copied from TCGOP_TYPE(op), because
the relevant op could be typeless, such as INDEX_op_call.

Fixes: fb744ece3a78 ("tcg: Copy TCGOP_TYPE in tcg_op_insert_{after,before}")
Suggested-by: Nicholas Piggin <npiggin@gmail.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>

tcg/optimize: Introduce opt_insert_{before,after}

Consolidate the places we call tcg_op_insert_{before,after}
within the optimization pass.

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

tcg/loongarch64: Improve constraints for TCG_CT_CONST_VCMP

Use the TCGCond given to tcg_target_const_match to exactly match
the supported constant.  Adjust the code generation to assume this
has been done -- recall that encode_*_insn contain assertions that
the constants are valid.

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

tcg/loongarch64: Fix vec_val computation in tcg_target_const_match

Only use vece for a vector constant.  This avoids an assertion
failure in sextract64 when vece contains garbage.

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

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

Block layer patches

- Discard alignment fixes
- Remove unused callback .bdrv_aio_pdiscard()
- qemu-img bench: Input validation fix

# -----BEGIN PGP SIGNATURE-----
#
# iQJFBAABCAAvFiEE3D3rFZqa+V09dFb+fwmycsiPL9YFAmgLy7QRHGt3b2xmQHJl
# ZGhhdC5jb20ACgkQfwmycsiPL9YB4A//Zsbb+tVsyLBKeffwPpHF/cAzHVH7Q2dV
# GC2JJvfrwq0gykfjj+u4akVQnPh49QiQM623PX7O15IikwdLy45ddQcYL1qflYCs
# ZGmBOuz/deI74qjl67bZVqIm8WeRhwHkdutfXOL7GRe2IHbceLbwwGUcbCgOVavt
# LHu3E2MIbvkLJoHEgg8UbJhZZY9DTLGDaMt00Yhy3UvNHU8UDeIr8o4dxMVv3gOf
# +8kIjGQkYNqpWp7aCxy8vofdSFjbBp4lSCK4G83xikUw49qkwWcgZ6jyTzXALg0G
# V+nMjH+DnfIRqhi1skFTHQNmFc6upxr7FIOgC+G5amkKLHCPnX9j5/2pBwrk63R7
# kXqzIPfRmfOTnJX+m7a9K/pE6RU9aPfr8mQdokEcQtlJkEjc6QN9HKfy/CLnJ5Id
# Le8jQODSZ1zRsP6Z8jyG4unj0AuOucUoXjAKQ5EWK5RoRoLMirxqDEDd9tBjcPYB
# JQmB/j7aTrF3aDWBs5ragCQYdcoXJbAbqLAwhaofyVRmVyjYJmWEIkPGGo946GPd
# /BFgaUaea4qW5+iIpWFTD9TCQEY/A7RRpT4teu7anZ/hDzLUyXLJU28xYC6LxiDZ
# Yoy5M/U6MLvgkBVTNuss4T3CIutBrUI7a/DLuGB+cSM6KkigQvNwLuBqPzTDfEQP
# sQJOP4UsX6k=
# =8amc
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 25 Apr 2025 13:51:48 EDT
# gpg:                using RSA key DC3DEB159A9AF95D3D7456FE7F09B272C88F2FD6
# gpg:                issuer "kwolf@redhat.com"
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* tag 'for-upstream' of https://repo.or.cz/qemu/kevin:
  qemu-img: improve queue depth validation in img_bench
  block: Remove unused callback function *bdrv_aio_pdiscard
  block/io: skip head/tail requests on EINVAL
  file-posix: probe discard alignment on Linux block devices

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Merge tag 'single-binary-20250425' of https://github.com/philmd/qemu into staging

Various patches loosely related to single binary work:

- Replace cpu_list() definition by CPUClass::list_cpus() callback
- Remove few MO_TE definitions on Hexagon / X86 targets
- Remove target_ulong uses in  ARMMMUFaultInfo and ARM CPUWatchpoint
- Remove DEVICE_HOST_ENDIAN definition
- Evaluate TARGET_BIG_ENDIAN at compile time and use target_needs_bswap() more
- Rename target_words_bigendian() as target_big_endian()
- Convert target_name() and target_cpu_type() to TargetInfo API
- Constify QOM TypeInfo class_data/interfaces fields
- Get default_cpu_type calling machine_class_default_cpu_type()
- Correct various uses of GLibCompareDataFunc prototype
- Simplify ARM/Aarch64 gdb_get_core_xml_file() handling a bit
- Move device tree files in their own pc-bios/dtb/ subdir
- Correctly check strchrnul() symbol availability on macOS SDK
- Move target-agnostic methods out of cpu-target.c and accel-target.c
- Unmap canceled USB XHCI packet
- Use deposit/extract API in designware model
- Fix MIPS16e translation
- Few missing header fixes

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEE+qvnXhKRciHc/Wuy4+MsLN6twN4FAmgLqb8ACgkQ4+MsLN6t
# wN6nCQ//cmv1M+NsndhO5TAK8T1eUSXKlTZh932uro6ZgxKwN4p+j1Qo7bq3O9gu
# qUMHNbcfQl8sHSytiXBoxCjLMCXC3u38iyz75WGXuPay06rs4wqmahqxL4tyno3l
# 1RviFts9xlLn+tJqqrAR6+pRdALld0TY+yXUjXgr4aK5pIRpLz9U/sIEoh7qbA5U
# x0MTaceDG3A91OYo0TgrNbcMe1b9GqQZ+a4tbaP+oE37wbiKdyQ68LjrEbV08Y1O
# qrFF4oxquV31QJcUiuII1W7hC6psGrMsUA1f1qDu7QvmybAZWNZNsR9T66X9jH5J
# wXMShJmmXwxugohmuPPFnDshzJy90aFL6Jy2shrfqcG2v0W66ARY1ZnbJLCcfczt
# 073bnE2dnOVhd/ny37RrIJNJLLmYM0yFDeKuYtNNAzpK9fpA7Q2PI8QiqNacQ3Pa
# TdEYrGlMk7OeNck8xJmJMY5rATthi1D4dIBv3rjQbUolQvPJe2Y9or0R2WL1jK5v
# hhr6DY01iSPES3CravmUs/aB1HRMPi/nX45OmFR6frAB7xqWMreh81heBVuoTTK8
# PuXtRQgRMRKwDeTxlc6p+zba4mIEYG8rqJtPFRgViNCJ1KsgSIowup3BNU05YuFn
# NoPoRayMDVMgejVgJin3Mg2DCYvt/+MBmO4IoggWlFsXj59uUgA=
# =DXnZ
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 25 Apr 2025 11:26:55 EDT
# gpg:                using RSA key FAABE75E12917221DCFD6BB2E3E32C2CDEADC0DE
# gpg: Good signature from "Philippe Mathieu-Daudé (F4BUG) <f4bug@amsat.org>" [full]
# Primary key fingerprint: FAAB E75E 1291 7221 DCFD  6BB2 E3E3 2C2C DEAD C0DE

* tag 'single-binary-20250425' of https://github.com/philmd/qemu: (58 commits)
  qemu: Convert target_name() to TargetInfo API
  accel: Move target-agnostic code from accel-target.c -> accel-common.c
  accel: Make AccelCPUClass structure target-agnostic
  accel: Include missing 'qemu/accel.h' header in accel-internal.h
  accel: Implement accel_init_ops_interfaces() for both system/user mode
  cpus: Move target-agnostic methods out of cpu-target.c
  cpus: Replace CPU_RESOLVING_TYPE -> target_cpu_type()
  qemu: Introduce target_cpu_type()
  qapi: Rename TargetInfo structure as QemuTargetInfo
  hw/microblaze: Evaluate TARGET_BIG_ENDIAN at compile time
  hw/mips: Evaluate TARGET_BIG_ENDIAN at compile time
  target/xtensa: Evaluate TARGET_BIG_ENDIAN at compile time
  target/mips: Check CPU endianness at runtime using env_is_bigendian()
  accel/kvm: Use target_needs_bswap()
  linux-user/elfload: Use target_needs_bswap()
  target/hexagon: Include missing 'accel/tcg/getpc.h'
  accel/tcg: Correct list of included headers in tcg-stub.c
  system/kvm: make functions accessible from common code
  meson: Use osdep_prefix for strchrnul()
  meson: Share common C source prefixes
  ...

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

Merge tag 'pull-vfio-20250425' of https://github.com/legoater/qemu into staging

vfio queue:

* Updated IGD passthrough documentation
* Fixed L2 crash on pseries machines
* Reorganized code and renamed services
* Moved HostIOMMUDevice realize after device attachement to help
  adding support for nested IOMMU
* Fixed CPR registration with IOMMUFD backend
* Refactored vfio-pci code to prepare ground for vfio-user

# -----BEGIN PGP SIGNATURE-----
#
# iQIzBAABCAAdFiEEoPZlSPBIlev+awtgUaNDx8/77KEFAmgLS9IACgkQUaNDx8/7
# 7KG4pxAAt5M4hDy6J1itsHhb50ORqEejdLwxKgJbSUPhRznzJBLqDoKFjtf0iiu3
# sSqGzcBAOPIrUmLiS7F31v7KxG7WkKGeWoRa3xaltnQ/6Z9v5FgWPvDKfUsI5C52
# eGg0gmXQeFQeuN5N+pJ/oDhyvzH9UKvesEYvbZrSRhau+7k06zFfrT2vGMjM4+0h
# A6Rd56dh99jT2kxbULr9eHuuYZxwse/wv0XWLponutBWeroqUzeEwwjYctVRMKR4
# HMVzrz+05EoIlFpY1EELE3SIZ9IaYViI6K3CHV2wE1TqrUWvipvY1oh2eJalFG78
# 5pyUY1Kb/iFR5AXPowzJkPc6il/1duXEghxYhPs6qxIZus3wQag1UwHoT8bw2dzV
# gLV6yk4w277VzzrYSTI7kgrCM0FNs/rdXmDAk8F0C63W9BV6lFt8t7i3DA27kEGe
# 58Ee71OPKCswzuqyu+t7zJKXEGvcR7/g2sXHkiHWKcrCfVSWG4Va/hnXpMGxM7Iw
# 35wzToYHYBxtdn5W92g9DtFgYAlQKivriylXSsJzhz9fjnQAOmLGKacjsJTCEpBR
# CJrIEsz1Zg+jmeBK8AO2OJUetW4wLtKsHW9uNRJVgjQ+eY8FKnzA6lNwGrEwrCNB
# rql77fw08jNtMNbU7M3G7a8OIYJoSVyH6wwar0qZA8IZjLkHY00=
# =eSsm
# -----END PGP SIGNATURE-----
# gpg: Signature made Fri 25 Apr 2025 04:46:10 EDT
# gpg:                using RSA key A0F66548F04895EBFE6B0B6051A343C7CFFBECA1
# gpg: Good signature from "Cédric Le Goater <clg@redhat.com>" [full]
# gpg:                 aka "Cédric Le Goater <clg@kaod.org>" [full]
# Primary key fingerprint: A0F6 6548 F048 95EB FE6B  0B60 51A3 43C7 CFFB ECA1

* tag 'pull-vfio-20250425' of https://github.com/legoater/qemu: (50 commits)
  vfio: refactor out vfio_pci_config_setup()
  vfio: refactor out vfio_interrupt_setup()
  vfio: Register/unregister container for CPR only once for each container
  vfio: Remove hiod_typename property
  vfio: Cleanup host IOMMU device creation
  vfio/container: Move realize() after attachment
  vfio/iommufd: Move realize() after attachment
  vfio/iommufd: Make a separate call to get IOMMU capabilities
  MAINTAINERS: Add a maintainer for util/vfio-helpers.c
  vfio: Rename VFIOContainer related services
  vfio: Rename VFIODevice related services
  vfio: Rename vfio-common.h to vfio-device.h
  vfio: Introduce vfio_listener_un/register() routines
  vfio: Rename RAM discard related services
  vfio: Introduce new files for VFIO MemoryListener
  vfio: Rename vfio_get_dirty_bitmap()
  vfio: Rename vfio_devices_all_device_dirty_tracking()
  vfio: Rename vfio_devices_all_dirty_tracking_started()
  vfio: Make vfio_container_query_dirty_bitmap() static
  vfio: Make vfio_devices_query_dirty_bitmap() static
  ...

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

qemu-img: improve queue depth validation in img_bench

This error was discovered by fuzzing qemu-img.

Currently, running `qemu-img bench -d 0` in img_bench is allowed,
which is a pointless operation and causes qemu-img to hang.

Signed-off-by: Denis Rastyogin <gerben@altlinux.org>
Message-ID: <20250327162423.25154-5-gerben@altlinux.org>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>

qemu: Convert target_name() to TargetInfo API

Have target_name() be a target-agnostic method, dispatching
to a per-target TargetInfo singleton structure.
By default a stub singleton is used. No logical change
expected.

Inspired-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20250424222112.36194-3-philmd@linaro.org>

accel: Move target-agnostic code from accel-target.c -> accel-common.c

Various methods of accel-target.c don't use any target-specific
knowledge at all and can be built once in the target-agnostic
accel-common.c file.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Message-Id: <20250417165430.58213-8-philmd@linaro.org>

accel: Make AccelCPUClass structure target-agnostic

Move the target-agnostic parts of "accel/accel-cpu-target.h"
to "accel/accel-cpu.h".

Doing so we need to include missing "hw/core/cpu.h" header
in "accel/accel-cpu.h" otherwise we get:

  include/accel/accel-cpu-target.h:39:28: error: unknown type name 'CPUClass'
     39 |     void (*cpu_class_init)(CPUClass *cc);
        |                            ^

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20250417165430.58213-7-philmd@linaro.org>

accel: Include missing 'qemu/accel.h' header in accel-internal.h

The "qemu/accel.h" header is implicitly pulled in. Include
it explicitly in order to avoid when refactoring unrelated
headers:

  accel/accel-internal.h:13:32: error: unknown type name 'AccelClass'
     13 | void accel_init_ops_interfaces(AccelClass *ac);
        |                                ^

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20250417165430.58213-6-philmd@linaro.org>

accel: Implement accel_init_ops_interfaces() for both system/user mode

We want to build more common code, moving objects from meson's
specific_ss[] set to common_ss[]. Since the CONFIG_USER_ONLY
definitions isn't applied on the common_ss[] set, it is simpler
to add an empty accel_init_ops_interfaces() stub on user emulation,
removing any CONFIG_USER_ONLY use in accel-target.c.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20250417165430.58213-5-philmd@linaro.org>

cpus: Move target-agnostic methods out of cpu-target.c

Various methods of cpu-target.c don't use any target-specific
knowledge at all and can be built once in the target-agnostic
cpu-common.c file.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20250417165430.58213-4-philmd@linaro.org>

cpus: Replace CPU_RESOLVING_TYPE -> target_cpu_type()

Replace the target-specific CPU_RESOLVING_TYPE definition
by a call to the target-agnostic target_cpu_type() runtime
helper.

Since the big "cpu.h" is not required anymore in tcg-all.c,
remove it, using the tinier "cpu-param.h" header.

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Message-Id: <20250417165430.58213-3-philmd@linaro.org>

qemu: Introduce target_cpu_type()

Introduce the target_cpu_type() helper to access the
CPU_RESOLVING_TYPE target-specific definition from
target-agnostic code.

Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Message-Id: <20250417165430.58213-2-philmd@linaro.org>

qapi: Rename TargetInfo structure as QemuTargetInfo

The QAPI-generated 'TargetInfo' structure name is only used
in a single file. We want to heavily use another structure
similarly named. Rename the QAPI one, since structure names
are not part of the public API.

Suggested-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20250422145502.70770-2-philmd@linaro.org>

hw/microblaze: Evaluate TARGET_BIG_ENDIAN at compile time

Rather than evaluating TARGET_BIG_ENDIAN at preprocessing
time via #ifdef'ry, do it in C at compile time

Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20250417131004.47205-8-philmd@linaro.org>