maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/anyext.ll

56 lines
1.5 KiB
LLVM
Raw Normal View History

[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown | FileCheck %s --check-prefix=X32
; RUN: llc < %s -mtriple=x86_64-unknown | FileCheck %s --check-prefix=X64
Don't use INSERT_SUBREG to model anyext operations on x86-64, as it leads to partial-register definitions. To help avoid redundant zero-extensions, also teach the h-register matching patterns that use movzbl to match anyext as well as zext. llvm-svn: 80099
2009-08-26 22:59:13 +08:00
As Dan pointed out, movzbl, movsbl, and friends are nicer than their alias (movzx/movsx) because they give more information. Revert that part of the patch. llvm-svn: 129498
2011-04-14 09:46:37 +08:00
; Use movzbl to avoid partial-register updates.
Don't use INSERT_SUBREG to model anyext operations on x86-64, as it leads to partial-register definitions. To help avoid redundant zero-extensions, also teach the h-register matching patterns that use movzbl to match anyext as well as zext. llvm-svn: 80099
2009-08-26 22:59:13 +08:00
define i32 @foo(i32 %p, i8 zeroext %x) nounwind {
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X32-LABEL: foo:
; X32: # BB#0:
; X32-NEXT: movzbl {{[0-9]+}}(%esp), %eax
VirtRegMap: Replace some identity copies with KILL instructions. An identity COPY like this: %AL = COPY %AL, %EAX<imp-def> has no semantic effect, but encodes liveness information: Further users of %EAX only depend on this instruction even though it does not define the full register. Replace the COPY with a KILL instruction in those cases to maintain this liveness information. (This reverts a small part of r238588 but this time adds a comment explaining why a KILL instruction is useful). llvm-svn: 274952
2016-07-09 08:19:07 +08:00
; X32-NEXT: # kill: %EAX<def> %EAX<kill> %AX<def>
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X32-NEXT: divb {{[0-9]+}}(%esp)
; X32-NEXT: movzbl %al, %eax
; X32-NEXT: andl $1, %eax
; X32-NEXT: retl
;
; X64-LABEL: foo:
; X64: # BB#0:
; X64-NEXT: movzbl %dil, %eax
VirtRegMap: Replace some identity copies with KILL instructions. An identity COPY like this: %AL = COPY %AL, %EAX<imp-def> has no semantic effect, but encodes liveness information: Further users of %EAX only depend on this instruction even though it does not define the full register. Replace the COPY with a KILL instruction in those cases to maintain this liveness information. (This reverts a small part of r238588 but this time adds a comment explaining why a KILL instruction is useful). llvm-svn: 274952
2016-07-09 08:19:07 +08:00
; X64-NEXT: # kill: %EAX<def> %EAX<kill> %AX<def>
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X64-NEXT: divb %sil
; X64-NEXT: movzbl %al, %eax
; X64-NEXT: andl $1, %eax
; X64-NEXT: retq
Don't use INSERT_SUBREG to model anyext operations on x86-64, as it leads to partial-register definitions. To help avoid redundant zero-extensions, also teach the h-register matching patterns that use movzbl to match anyext as well as zext. llvm-svn: 80099
2009-08-26 22:59:13 +08:00
%q = trunc i32 %p to i8
%r = udiv i8 %q, %x
%s = zext i8 %r to i32
%t = and i32 %s, 1
ret i32 %t
}
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
Don't use INSERT_SUBREG to model anyext operations on x86-64, as it leads to partial-register definitions. To help avoid redundant zero-extensions, also teach the h-register matching patterns that use movzbl to match anyext as well as zext. llvm-svn: 80099
2009-08-26 22:59:13 +08:00
define i32 @bar(i32 %p, i16 zeroext %x) nounwind {
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X32-LABEL: bar:
; X32: # BB#0:
[X86] Fix PR23155 by turning on X86FixupBWInsts by default. Differential Revision: http://reviews.llvm.org/D18866 llvm-svn: 265830
2016-04-09 02:58:29 +08:00
; X32-NEXT: movzwl {{[0-9]+}}(%esp), %eax
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X32-NEXT: xorl %edx, %edx
; X32-NEXT: divw {{[0-9]+}}(%esp)
VirtRegMap: Replace some identity copies with KILL instructions. An identity COPY like this: %AL = COPY %AL, %EAX<imp-def> has no semantic effect, but encodes liveness information: Further users of %EAX only depend on this instruction even though it does not define the full register. Replace the COPY with a KILL instruction in those cases to maintain this liveness information. (This reverts a small part of r238588 but this time adds a comment explaining why a KILL instruction is useful). llvm-svn: 274952
2016-07-09 08:19:07 +08:00
; X32-NEXT: # kill: %AX<def> %AX<kill> %EAX<def>
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X32-NEXT: andl $1, %eax
; X32-NEXT: retl
;
; X64-LABEL: bar:
; X64: # BB#0:
; X64-NEXT: xorl %edx, %edx
[X86] Teach X86FixupBWInsts to promote MOV8rr/MOV16rr to MOV32rr. This re-applies r268760, reverted in r268794. Fixes http://llvm.org/PR27670 The original imp-defs assertion was way overzealous: forward all implicit operands, except imp-defs of the new super-reg def (r268787 for GR64, but also possible for GR16->GR32), or imp-uses of the new super-reg use. While there, mark the source use as Undef, and add an imp-use of the old source reg: that should cover any case of dead super-regs. At the stage the pass runs, flags are unlikely to matter anyway; still, let's be as correct as possible. Also add MIR tests for the various interesting cases. Original commit message: Codesize is less (16) or equal (8), and we avoid partial dependencies. Differential Revision: http://reviews.llvm.org/D19999 llvm-svn: 268831
2016-05-07 09:11:17 +08:00
; X64-NEXT: movl %edi, %eax
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X64-NEXT: divw %si
VirtRegMap: Replace some identity copies with KILL instructions. An identity COPY like this: %AL = COPY %AL, %EAX<imp-def> has no semantic effect, but encodes liveness information: Further users of %EAX only depend on this instruction even though it does not define the full register. Replace the COPY with a KILL instruction in those cases to maintain this liveness information. (This reverts a small part of r238588 but this time adds a comment explaining why a KILL instruction is useful). llvm-svn: 274952
2016-07-09 08:19:07 +08:00
; X64-NEXT: # kill: %AX<def> %AX<kill> %EAX<def>
[X86] More thorough partial-register division checks For when grep counts are just not enough... llvm-svn: 260891
2016-02-15 22:09:35 +08:00
; X64-NEXT: andl $1, %eax
; X64-NEXT: retq
Don't use INSERT_SUBREG to model anyext operations on x86-64, as it leads to partial-register definitions. To help avoid redundant zero-extensions, also teach the h-register matching patterns that use movzbl to match anyext as well as zext. llvm-svn: 80099
2009-08-26 22:59:13 +08:00
%q = trunc i32 %p to i16
%r = udiv i16 %q, %x
%s = zext i16 %r to i32
%t = and i32 %s, 1
ret i32 %t
}