mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-22 18:54:02 +01:00
94988510e6
Adds ARMBankConflictHazardRecognizer. This hazard recognizer looks for a few situations where the same base pointer is used and then checks whether the offsets lead to a bank conflict. Two parameters are also added to permit overriding of the target assumptions: arm-data-bank-mask=<int> - Mask of bits which are to be checked for conflicts. If all these bits are equal in the offsets, there is a conflict. arm-assume-itcm-bankconflict=<bool> - Assume that there will be bank conflicts on any loads to a constant pool. This hazard recognizer is enabled for Cortex-M7, where the Technical Reference Manual states that there are two DTCM banks banked using bit 2 and one ITCM bank. Differential Revision: https://reviews.llvm.org/D93054
39 lines
1.4 KiB
LLVM
39 lines
1.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=thumbv7m-none-eabi -mcpu=cortex-m7 | FileCheck %s --check-prefix=CHECK
|
|
; RUN: llc < %s -mtriple=thumbv7m-none-eabi -mcpu=cortex-m7 -arm-data-bank-mask=-1 | FileCheck %s --check-prefix=NOBANK
|
|
|
|
; This tests the cortex-m7 bank conflict hazard recognizer.
|
|
; Normally both loads would be scheduled early (both in the first cycle) due to
|
|
; their latency. But will bank conflict to TCM so are scheduled in different
|
|
; cycles.
|
|
|
|
define i32 @test(i32* %x0, i32 %y, i32 %z) {
|
|
; CHECK-LABEL: test:
|
|
; CHECK: @ %bb.0: @ %entry
|
|
; CHECK-NEXT: ldr r3, [r0]
|
|
; CHECK-NEXT: subs r1, r3, r1
|
|
; CHECK-NEXT: ldr r0, [r0, #8]
|
|
; CHECK-NEXT: subs r1, r1, r2
|
|
; CHECK-NEXT: adds r1, #1
|
|
; CHECK-NEXT: muls r0, r1, r0
|
|
; CHECK-NEXT: bx lr
|
|
; NOBANK-LABEL: test:
|
|
; NOBANK: @ %bb.0: @ %entry
|
|
; NOBANK-NEXT: ldr r3, [r0]
|
|
; NOBANK-NEXT: ldr r0, [r0, #8]
|
|
; NOBANK-NEXT: subs r1, r3, r1
|
|
; NOBANK-NEXT: subs r1, r1, r2
|
|
; NOBANK-NEXT: adds r1, #1
|
|
; NOBANK-NEXT: muls r0, r1, r0
|
|
; NOBANK-NEXT: bx lr
|
|
entry:
|
|
%0 = load i32, i32* %x0, align 4
|
|
%mul3 = add nsw i32 %0, 1
|
|
%mul = sub nsw i32 %mul3, %y
|
|
%sub = sub nsw i32 %mul, %z
|
|
%arrayidx1 = getelementptr inbounds i32, i32* %x0, i32 2
|
|
%1 = load i32, i32* %arrayidx1, align 4
|
|
%mul2 = mul nsw i32 %sub, %1
|
|
ret i32 %mul2
|
|
}
|