We wanted the VPSRAQ variant with the scalar vector shift operand,
not the variant with an immediate operand.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1022
Fixes: 47b331b2a8 ("tcg/i386: Implement avx512 scalar shift")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The general ternary logic operation can implement BITSEL.
Funnel the 4-operand operation into three variants of the
3-operand instruction, depending on input operand overlap.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512VL has a general ternary logic operation, VPTERNLOGQ,
which can implement NOT, ORC, NAND, NOR, EQV.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512DQ has VPMULLQ.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512VL has VPABSQ, VPMAXSQ, VPMAXUQ, VPMINSQ, VPMINUQ.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Expand 32-bit and 64-bit scalar rotate with VPRO[LR]V;
expand 16-bit scalar rotate with VPSHLDV.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There is no such instruction on x86, so we should
not be pretending it has arguments.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
While there are no specific 16-bit rotate instructions, there
are double-word shifts, which can perform the same operation.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will use VPSHLD, VPSHLDV and VPSHRDV for 16-bit rotates.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512VL has VPROLVD and VPRORVQ.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512VL has VPROLD and VPROLQ, layered onto the same
opcode as PSHIFTD, but requires EVEX encoding and W1.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512 has VPSRAQ with immediate operand, in the same form as
with AVX, but requires EVEX encoding and W1.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512VL has VPSRAQ.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
AVX512VL has VPSRAVQ, and
AVX512BW has VPSLLVW, VPSRAVW, VPSRLVW.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The condition for UMIN/UMAX availability is about to change;
use the canonical version.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The evex encoding is added here, for use in a subsequent patch.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There are some operation sizes in some subsets of AVX512 that
are missing from previous iterations of AVX. Detect them.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
__get_cpuid_max returns an unsigned value.
For consistency, store the result in an unsigned variable.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Richard Henderson <rth@twiddle.net>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Renaming defines for quad in their various forms so that their signedness is
now explicit.
Done using git grep as suggested by Philippe, with a bit of hand edition to
keep assignments aligned.
Signed-off-by: Frédéric Pétrot <frederic.petrot@univ-grenoble-alpes.fr>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20220106210108.138226-2-frederic.petrot@univ-grenoble-alpes.fr
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
We're about to move this out of tcg.h, so rename it
as we did when moving MemOp.
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We have lacked expressive support for memory sizes larger
than 64-bits for a while. Fixing that requires adjustment
to several points where we used this for array indexing,
and two places that develop -Wswitch warnings after the change.
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We need to be able to represent VEX.W on a 32-bit host, where REX.W
will always be zero. Fixes the encoding for VPSLLVQ and VPSRLVQ.
Fixes: a2ce146a06 ("tcg/i386: Support vector variable shift opcodes")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/385
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Retain the current rorw bswap16 expansion for the zero-in/zero-out case.
Otherwise, perform a wider bswap plus a right-shift or extend.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Change the return value to bool, because that's what is should
have been from the start. Pass the ct mask instead of the whole
TCGArgConstraint, as that's the only part that's relevant.
Change the value argument to int64_t. We will need the extra
width for 32-bit hosts wanting to match vector constants.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
There are two different versions of prototype for tcg_out_op and
tcg_out_vec_op functions:
1) using const TCGArg *args and const int *const_args arguments
2) using const TCGArg args[TCG_MAX_OP_ARGS] and const int
const_args[TCG_MAX_OP_ARGS] aguments.
This duality causes warnings on GCC 11 and prevents build using
--enable-werror. As second version provides more information,
unify functions prototypes to this variant.
Signed-off-by: Miroslav Rezanina <mrezanin@redhat.com>
Message-Id: <20210312121418.139093-1-mrezanin@redhat.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This exports the constraint sets from tcg_target_op_def to
a place we will be able to manipulate more in future.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This eliminates the target-specific function target_parse_constraint
and folds it into the single caller, process_op_defs. Since this is
done directly into the switch statement, duplicates are compilation
errors rather than silently ignored at runtime.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Create symbolic constants for all low-byte-addressable
and second-byte-addressable registers. Create a symbol
for the registers that need reserving for softmmu.
There is no functional change for 's', as this letter is
only used for i386. The BYTEL name is correct for the
action we wish from the constraint.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Rather than check the type when filling in the constraint,
check it when matching the constant. This removes the only
use of the type argument to target_parse_constraint.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
These are now completely covered by mov from a
TYPE_CONST temporary.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
While we don't store more than tcg_target_long in TCGTemp,
we shouldn't be limited to that for code generation. We will
be able to use this for INDEX_op_dup2_vec with 2 constants.
Also pass along the minimal vece that may be said to apply
to the constant. This allows some simplification in the
various backends.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Having dupi pass though movi is confusing and arguably wrong.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Now that all native tcg hosts support splitwx,
make this pointer const.
Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Now that all native tcg hosts support splitwx,
make this pointer const.
Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We must change all targets at once, since all must match
the declaration in tcg.c.
Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Simplify the arguments to always use s->code_ptr instead of
take it as an argument. That makes it easy to ensure that
the value_ptr is always the rx version.
Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We must change all targets at once, since all must match
the declaration in tcg.c.
Reviewed-by: Joelle van Dyne <j@getutm.app>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This value is constant across all thread-local copies of TCGContext,
so we might as well move it out of thread-local storage.
Reviewed-by: Joelle van Dyne <j@getutm.app>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Enable this on i386 to restrict the set of input registers
for an 8-bit store, as required by the architecture. This
removes the last use of scratch registers for user-only mode.
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Always true when movbe is available, otherwise leave
this to generic code.
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The previous change wrongly stated that 32-bit avx2 should have
used VPBROADCASTW. But that's a 16-bit broadcast and we want a
32-bit broadcast.
Fixes: 7b60ef3264
Cc: qemu-stable@nongnu.org
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This wasn't actually used for anything, really. All variable
operands must accept registers, and which are indicated by the
set in TCGArgConstraint.regs.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The union is unused; let "regs" appear in the main structure
without the "u.regs" wrapping.
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
With Makefiles that have automatically generated dependencies, you
generated includes are set as dependencies of the Makefile, so that they
are built before everything else and they are available when first
building the .c files.
Alternatively you can use a fine-grained dependency, e.g.
target/arm/translate.o: target/arm/decode-neon-shared.inc.c
With Meson you have only one choice and it is a third option, namely
"build at the beginning of the corresponding target"; the way you
express it is to list the includes in the sources of that target.
The problem is that Meson decides if something is a source vs. a
generated include by looking at the extension: '.c', '.cc', '.m', '.C'
are sources, while everything else is considered an include---including
'.inc.c'.
Use '.c.inc' to avoid this, as it is consistent with our other convention
of using '.rst.inc' for included reStructuredText files. The editorconfig
file is adjusted.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
