RPCS3/llvm
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm.git synced 2025-02-26 21:56:11 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

[AVR] Fix the 'avr-tiny.ll' and 'avr25.ll' subtarget feature tests

Browse Source 
When these tests were originally written, the middle end would introduce
an unnecessary copy from r24:r23->GPR16->r24:r23, and these tests
mistakenly relied on it.

The most optimal codegen for the functions in the test cases before this patch
would be NOPs. This is because the first i16 argument always gets the same register
allocation as an i16 return value in the AVR calling convention.

These tests broke in r362963 when the codegen was improved and the
redundant copy was eliminated. After this, the test functions
were lowered to their optimal form - a 'ret' and nothing else.

This patch prepends an extra i16 operand to each of the test functions
so that a 16-bit copy must be inserted for the program to be correct.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363131 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Dylan McKay 2019-06-12 08:31:07 +00:00
parent 8dcec92474
commit 27a775c1fd
2 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
test/CodeGen/AVR/features/avr-tiny.ll
View File

@ -1,9 +1,9 @@
; RUN: llc -mattr=avrtiny -O0 < %s -march=avr | FileCheck %s
define i16 @reg_copy16(i16 %a) {
define i16 @reg_copy16(i16, i16 %a) {
; CHECK-LABEL: reg_copy16
; CHECK: mov r18, r24
; CHECK: mov r19, r25
; CHECK: mov r24, r22
; CHECK: mov r25, r23
ret i16 %a
}

4
test/CodeGen/AVR/features/avr25.ll
View File

@ -1,8 +1,8 @@
; RUN: llc -mattr=avr25 -O0 < %s -march=avr | FileCheck %s
; On most cores, the 16-bit 'MOVW' instruction can be used
define i16 @reg_copy16(i16 %a) {
define i16 @reg_copy16(i16, i16 %a) {
; CHECK-LABEL: reg_copy16
; CHECK: movw r18, r24
; CHECK: movw r24, r22
ret i16 %a
}