RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-26 04:32:44 +01:00
Code Issues Projects Releases Wiki Activity
2fbf734dc0
llvm-mirror/test/CodeGen/ARM/fast-isel.ll
Paul Robinson 8d4bf186a4 [FastISel] Flush local value map on every instruction 
Local values are constants or addresses that can't be folded into
the instruction that uses them. FastISel materializes these in a
"local value" area that always dominates the current insertion
point, to try to avoid materializing these values more than once
(per block).

https://reviews.llvm.org/D43093 added code to sink these local
value instructions to their first use, which has two beneficial
effects. One, it is likely to avoid some unnecessary spills and
reloads; two, it allows us to attach the debug location of the
user to the local value instruction. The latter effect can
improve the debugging experience for debuggers with a "set next
statement" feature, such as the Visual Studio debugger and PS4
debugger, because instructions to set up constants for a given
statement will be associated with the appropriate source line.

There are also some constants (primarily addresses) that could be
produced by no-op casts or GEP instructions; the main difference
from "local value" instructions is that these are values from
separate IR instructions, and therefore could have multiple users
across multiple basic blocks. D43093 avoided sinking these, even
though they were emitted to the same "local value" area as the
other instructions. The patch comment for D43093 states:

  Local values may also be used by no-op casts, which adds the
  register to the RegFixups table. Without reversing the RegFixups
  map direction, we don't have enough information to sink these
  instructions.

This patch undoes most of D43093, and instead flushes the local
value map after(*) every IR instruction, using that instruction's
debug location. This avoids sometimes incorrect locations used
previously, and emits instructions in a more natural order.

In addition, constants materialized due to PHI instructions are
not assigned a debug location immediately; instead, when the
local value map is flushed, if the first local value instruction
has no debug location, it is given the same location as the
first non-local-value-map instruction.  This prevents PHIs
from introducing unattributed instructions, which would either
be implicitly attributed to the location for the preceding IR
instruction, or given line 0 if they are at the beginning of
a machine basic block.  Neither of those consequences is good
for debugging.

This does mean materialized values are not re-used across IR
instruction boundaries; however, only about 5% of those values
were reused in an experimental self-build of clang.

(*) Actually, just prior to the next instruction. It seems like
it would be cleaner the other way, but I was having trouble
getting that to work.

This reapplies commits cf1c774d and dc35368c, and adds the
modification to PHI handling, which should avoid problems
with debugging under gdb.

Differential Revision: https://reviews.llvm.org/D91734
2021-01-11 08:32:36 -08:00

200 lines
4.7 KiB
LLVM
Raw Blame History

; RUN: llc < %s -O0 -fast-isel-abort=1 -relocation-model=dynamic-no-pic -mtriple=armv7-apple-ios -verify-machineinstrs | FileCheck %s --check-prefix=ARM --check-prefix=ARM-MACHO
; RUN: llc < %s -O0 -fast-isel-abort=1 -relocation-model=dynamic-no-pic -mtriple=armv7-linux-gnueabi -verify-machineinstrs | FileCheck %s --check-prefix=ARM --check-prefix=ARM-ELF
; RUN: llc < %s -O0 -fast-isel-abort=1 -relocation-model=dynamic-no-pic -mtriple=thumbv7-apple-ios -verify-machineinstrs | FileCheck %s --check-prefix=THUMB
; Very basic fast-isel functionality.
define i32 @test0(i32 %a, i32 %b) nounwind {
entry:
%a.addr = alloca i32, align 4
%b.addr = alloca i32, align 4
store i32 %a, i32* %a.addr
store i32 %b, i32* %b.addr
%tmp = load i32, i32* %a.addr
%tmp1 = load i32, i32* %b.addr
%add = add nsw i32 %tmp, %tmp1
ret i32 %add
}
; Check truncate to bool
define void @test1(i32 %tmp) nounwind {
entry:
%tobool = trunc i32 %tmp to i1
br i1 %tobool, label %if.then, label %if.end
if.then: ; preds = %entry
call void @test1(i32 0)
br label %if.end
if.end: ; preds = %if.then, %entry
ret void
; ARM-LABEL: test1:
; ARM: tst r0, #1
; THUMB-LABEL: test1:
; THUMB: tst.w r0, #1
}
; Check some simple operations with immediates
define void @test2(i32 %tmp, i32* %ptr) nounwind {
; THUMB-LABEL: test2:
; ARM-LABEL: test2:
b1:
%a = add i32 %tmp, 4096
store i32 %a, i32* %ptr
br label %b2
; THUMB: add.w {{.*}} #4096
; ARM: add {{.*}} #4096
b2:
%b = add i32 %tmp, 4095
store i32 %b, i32* %ptr
br label %b3
; THUMB: addw {{.*}} #4095
; ARM: movw {{.*}} #4095
; ARM: add
b3:
%c = or i32 %tmp, 4
store i32 %c, i32* %ptr
ret void
; THUMB: orr {{.*}} #4
; ARM: orr {{.*}} #4
}
define void @test3(i32 %tmp, i32* %ptr1, i16* %ptr2, i8* %ptr3) nounwind {
; THUMB-LABEL: test3:
; ARM-LABEL: test3:
bb1:
%a1 = trunc i32 %tmp to i16
%a2 = trunc i16 %a1 to i8
%a3 = trunc i8 %a2 to i1
%a4 = zext i1 %a3 to i8
store i8 %a4, i8* %ptr3
%a5 = zext i8 %a4 to i16
store i16 %a5, i16* %ptr2
%a6 = zext i16 %a5 to i32
store i32 %a6, i32* %ptr1
br label %bb2
; THUMB: and
; THUMB: strb
; THUMB: and{{.*}}, #255
; THUMB: strh
; THUMB: uxth
; ARM: and
; ARM: strb
; ARM: and{{.*}}, #255
; ARM: strh
; ARM: uxth
bb2:
%b1 = trunc i32 %tmp to i16
%b2 = trunc i16 %b1 to i8
store i8 %b2, i8* %ptr3
%b3 = sext i8 %b2 to i16
store i16 %b3, i16* %ptr2
%b4 = sext i16 %b3 to i32
store i32 %b4, i32* %ptr1
br label %bb3
; THUMB: strb
; THUMB: sxtb
; THUMB: strh
; THUMB: sxth
; ARM: strb
; ARM: sxtb
; ARM: strh
; ARM: sxth
bb3:
%c1 = load i8, i8* %ptr3
%c2 = load i16, i16* %ptr2
%c3 = load i32, i32* %ptr1
%c4 = zext i8 %c1 to i32
%c5 = sext i16 %c2 to i32
%c6 = add i32 %c4, %c5
%c7 = sub i32 %c3, %c6
store i32 %c7, i32* %ptr1
ret void
; THUMB: ldrb
; THUMB: ldrh
; THUMB: and{{.*}}, #255
; THUMB: sxth
; THUMB: add
; THUMB: sub
; ARM: ldrb
; ARM: ldrh
; ARM: and{{.*}}, #255
; ARM: sxth
; ARM: add
; ARM: sub
}
; Check loads/stores with globals
@test4g = external global i32
define void @test4() {
%a = load i32, i32* @test4g
%b = add i32 %a, 1
store i32 %b, i32* @test4g
ret void
; Note that relocations are either movw/movt or constant pool
; loads. Different platforms will select different approaches.
; THUMB: {{(movw r0, :lower16:L_test4g\$non_lazy_ptr)|(ldr.n r0, .LCPI)}}
; THUMB: {{(movt r0, :upper16:L_test4g\$non_lazy_ptr)?}}
; THUMB: ldr [[REG:r[0-9]+]], [r0]
; THUMB: ldr [[REG1:r[0-9]+]], {{\[}}[[REG]]]
; THUMB: adds [[REG1]], #1
; THUMB: {{(movw r1, :lower16:L_test4g\$non_lazy_ptr)|(ldr.n r0, .LCPI)}}
; THUMB: {{(movt r1, :upper16:L_test4g\$non_lazy_ptr)?}}
; THUMB: ldr [[REG2:r[0-9]+]], [r1]
; THUMB: str [[REG1]], {{\[}}[[REG2]]]
; ARM-MACHO: {{(movw r0, :lower16:L_test4g\$non_lazy_ptr)|(ldr r0, .LCPI)}}
; ARM-MACHO: {{(movt r0, :upper16:L_test4g\$non_lazy_ptr)?}}
; ARM-MACHO: ldr [[REG:r[0-9]+]], [r0]
; ARM-ELF: movw [[REG:r[0-9]+]], :lower16:test4g
; ARM-ELF: movt [[REG]], :upper16:test4g
; ARM: ldr [[REG1:r[0-9]+]], {{\[}}[[REG]]]
; ARM: add [[REG2:r[0-9]+]], [[REG1]], #1
; ARM-MACHO: {{(movw r1, :lower16:L_test4g\$non_lazy_ptr)|(ldr r0, .LCPI)}}
; ARM-MACHO: {{(movt r1, :upper16:L_test4g\$non_lazy_ptr)?}}
; ARM-MACHO: ldr [[REG3:r[0-9]+]], [r1]
; ARM-ELF: movw [[REG3:r[0-9]+]], :lower16:test4g
; ARM-ELF: movt [[REG3]], :upper16:test4g
; ARM: str [[REG2]], {{\[}}[[REG3]]]
}
; ARM: @urem_fold
; THUMB: @urem_fold
; ARM: and r0, r0, #31
; THUMB: and r0, r0, #31
define i32 @urem_fold(i32 %a) nounwind {
%rem = urem i32 %a, 32
ret i32 %rem
}
define i32 @trap_intrinsic() noreturn nounwind {
entry:
; ARM: @trap_intrinsic
; THUMB: @trap_intrinsic
; ARM: trap
; THUMB: trap
tail call void @llvm.trap( )
unreachable
}
declare void @llvm.trap() nounwind
Reference in New Issue View Git Blame Copy Permalink