Commit Graph

10 Commits

Author SHA1 Message Date
Julian Seward
6f8a54b6a5 Adds support for legalizing CLZ, CTZ and POPCOUNT on baseline x86_64 targets.
Changes:

* Adds a new generic instruction, SELECTIF, that does value selection (a la
  conditional move) similarly to existing SELECT, except that it is
  controlled by condition code input and flags-register inputs.

* Adds a new Intel x86_64 variant, 'baseline', that supports SSE2 and
  nothing else.

* Adds new Intel x86_64 instructions BSR and BSF.

* Implements generic CLZ, CTZ and POPCOUNT on x86_64 'baseline' targets
  using the new BSR, BSF and SELECTIF instructions.

* Implements SELECTIF on x86_64 targets using conditional-moves.

* new test filetests/isa/intel/baseline_clz_ctz_popcount.cton
  (for legalization)

* new test filetests/isa/intel/baseline_clz_ctz_popcount_encoding.cton
  (for encoding)

* Allow lib/cretonne/meta/gen_legalizer.py to generate non-snake-caseified
  Rust without rustc complaining.

Fixes #238.
2018-02-06 09:43:00 -08:00
Tyler McMullen
1a11c351b5 Add tests and documentation for x86_(push|pop). Fix up encoding issues revealed by tests. 2017-12-05 11:49:12 -08:00
Tyler McMullen
4eb9a54096 Convert x86_(push|pop) operations to be explicitly limited to 32-bit and 64-bit values. 2017-12-05 11:49:12 -08:00
Tyler McMullen
b8275f5713 Add (some) encodings for x86_push/pop instructions. Simple uses actually pass the legalizer now. 2017-12-05 11:49:12 -08:00
Tyler McMullen
8ed37e352e Add x86_push and x86_pop instructions. 2017-12-05 11:49:12 -08:00
Jakob Stoklund Olesen
1fe7890700 Add x86_fmin and x86_fmax instructions.
These Intel-specific instructions represent the semantics of the minss /
maxss Intel instructions which behave more like a C ternary operator
than the WebAssembly fmin and fmax instructions.

They will be used as building blocks for implementing the WebAssembly
semantics.
2017-09-27 09:17:09 -07:00
Jakob Stoklund Olesen
ac69f3bfdf Add an Intel-specific x86_cvtt2si instruction.
This is used to represent the non-trapping semantics of the cvttss2si and
cvttsd2si instructions (and their vectorized counterparts).

The overflow behavior of this instruction is specific to the Intel ISAs.

There is no float-to-i64 instruction on the 32-bit Intel ISA.
2017-09-26 15:44:41 -07:00
Jakob Stoklund Olesen
54534e2147 Add Intel legalization for division and multiplication.
These operations need custom legalization in order to use Intel's div
and idiv instructions.
2017-07-28 16:41:59 -07:00
Dimo
b448574a49 Assert all InstructionGroups are closed in TargetIsa.__init__(); Close x86 group 2017-07-24 14:08:44 -07:00
Jakob Stoklund Olesen
c4db4c124b Begin an Intel-specific instruction group.
Add instructions representing Intel's division instructions which use a
numerator that is twice as wide as the denominator and produce both the
quotient and remainder.

Add encodings for the x86_[su]divmodx instructions.
2017-07-18 11:20:00 -07:00