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>
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>
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>
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>
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>
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>
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>
Use the fully general extract opcodes instead.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
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>
We cannot rely on the value copied from TCGOP_TYPE(op), because
the relevant op could be typeless, such as INDEX_op_call.
Fixes: fb744ece3a ("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>
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>
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>
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>
In effect, hoist the check for mttcg from tcg_n_regions()
to tcg_init_machine().
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>
We already have two subdirectories for which we need
to build files twice, for user vs system modes.
Move this handling to the top level.
This cannot be combined with user_ss or system_ss,
because the formulation has not been extended to support
configuration symbols.
Reviewed-by: Pierrick Bouvier <pierrick.bouvier@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Be sure to allocate the temp frame if it wasn't.
In the resolved issues, incomplete dead code elimination left a load
at the top of an unreachable loop. We simply need to allocate the
stack slot to avoid crashing.
Fixes: c896fe29d6 ("TCG code generator")
Reported-by: Michael Tokarev <mjt@tls.msk.ru>
Reported-by: Helge Konetzka <hk@zapateado.de>
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2891
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2899
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20250401144332.41615-1-philmd@linaro.org>
These defines never should have been added as they were
never used. Only 32-bit hosts may have these opcodes and
they have them unconditionally.
Fixes: 6cb14e4de2 ("tcg/loongarch64: Add the tcg-target.h file")
Fixes: fb1f70f368 ("tcg/riscv: Add the tcg-target.h file")
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Eliminate code repetition by using the appropriate helpers.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Replace target-specific 'Z' with generic 'z'.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Replace target-specific 'Z' with generic 'z'.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Replace target-specific 'Z' with generic 'z'.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Replace target-specific 'Z' with generic 'z'.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
For loongarch, mips, riscv and sparc, a zero register is
available all the time. For aarch64, register index 31
depends on context: sometimes it is the stack pointer,
and sometimes it is the zero register.
Introduce a new general-purpose constraint which maps 0
to TCG_REG_ZERO, if defined. This differs from existing
constant constraints in that const_arg[*] is recorded as
false, indicating that the value is in a register.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Since we no longer support 64-bit guests on 32-bit hosts,
we can use a 32-bit type on a 32-bit host. This shrinks
the size of the structure to 16 bytes on a 32-bit host.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There is now always only one guest address register.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The guest address will now always fit in one register.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The guest address will now always fit in one register.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The guest address will now always fit in one register.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The guest address will now always be TCG_TYPE_I32.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Since 64-on-32 is now unsupported, guest addresses always
fit in one host register. Drop the replication of opcodes.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
These should have been removed with the rest. There are
a couple of hosts which can emit guest_base into the
constant pool: aarch64, mips64, ppc64, riscv64.
Fixes: a417ef8350 ("tcg: Remove TCG_TARGET_NEED_LDST_LABELS and TCG_TARGET_NEED_POOL_LABELS")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Generalize the existing optimization of "TSTNE x,sign" and "TSTNE x,-1".
This can be useful for example in the i386 frontend, which will generate
tests of zero-extended registers against 0xffffffff.
Ironically, on x86 hosts this is a very slight pessimization in the very
case it's meant to optimize because
brcond_i64 cc_dst,$0xffffffff,tsteq,$L1
(test %ebx, %ebx) is 1 byte smaller than
brcond_i64 cc_dst,$0x0,eq,$L1
(test %rbx, %rbx). However, in general it is an improvement, especially
if it avoids placing a large immediate in the constant pool.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20250102181601.1421059-3-richard.henderson@linaro.org>
Make deposit "unconditional" in the sense that the opcode is
always present. Rely instead on TCG_TARGET_deposit_valid,
now always defined.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Make extract and sextract "unconditional" in the sense
that the opcodes are always present. Rely instead on
TCG_TARGET_HAS_{s}extract_valid, now always defined.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We already have these assertions during opcode creation.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
