RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-01 16:33:37 +01:00
Code Issues Projects Releases Wiki Activity
cf35a852f5
llvm-mirror/test/Assembler/AutoUpgradeIntrinsics.ll

56 lines
2.0 KiB
LLVM
Raw Normal View History

This is the patch to provide clean intrinsic function overloading support in LLVM. It cleans up the intrinsic definitions and generally smooths the process for more complicated intrinsic writing. It will be used by the upcoming atomic intrinsics as well as vector and float intrinsics in the future. This also changes the syntax for llvm.bswap, llvm.part.set, llvm.part.select, and llvm.ct* intrinsics. They are automatically upgraded by both the LLVM ASM reader and the bitcode reader. The test cases have been updated, with special tests added to ensure the automatic upgrading is supported. llvm-svn: 40807
2007-08-04 03:51:18 +02:00
; Tests to make sure intrinsics are automatically upgraded.
Replace the "movnt" intrinsics with a native store + nontemporal metadata bit. <rdar://problem/8460511> llvm-svn: 130791
2011-05-03 23:11:17 +02:00
; RUN: llvm-as < %s | llvm-dis | FileCheck %s
This is the patch to provide clean intrinsic function overloading support in LLVM. It cleans up the intrinsic definitions and generally smooths the process for more complicated intrinsic writing. It will be used by the upcoming atomic intrinsics as well as vector and float intrinsics in the future. This also changes the syntax for llvm.bswap, llvm.part.set, llvm.part.select, and llvm.ct* intrinsics. They are automatically upgraded by both the LLVM ASM reader and the bitcode reader. The test cases have been updated, with special tests added to ensure the automatic upgrading is supported. llvm-svn: 40807
2007-08-04 03:51:18 +02:00
Reapply r129401 with patch for clang. llvm-svn: 129419
2011-04-13 02:36:11 +02:00
declare <4 x float> @llvm.x86.sse.loadu.ps(i8*) nounwind readnone
declare <16 x i8> @llvm.x86.sse2.loadu.dq(i8*) nounwind readnone
declare <2 x double> @llvm.x86.sse2.loadu.pd(double*) nounwind readnone
define void @test_loadu(i8* %a, double* %b) {
%v0 = call <4 x float> @llvm.x86.sse.loadu.ps(i8* %a)
%v1 = call <16 x i8> @llvm.x86.sse2.loadu.dq(i8* %a)
%v2 = call <2 x double> @llvm.x86.sse2.loadu.pd(double* %b)
rip out a ton of intrinsic modernization logic from AutoUpgrade.cpp, which is for pre-2.9 bitcode files. We keep x86 unaligned loads, movnt, crc32, and the target indep prefetch change. As usual, updating the testsuite is a PITA. llvm-svn: 133337
2011-06-18 08:05:24 +02:00
; CHECK: load i128* {{.*}}, align 1
; CHECK: load i128* {{.*}}, align 1
; CHECK: load i128* {{.*}}, align 1
Reapply r129401 with patch for clang. llvm-svn: 129419
2011-04-13 02:36:11 +02:00
ret void
}
Replace the "movnt" intrinsics with a native store + nontemporal metadata bit. <rdar://problem/8460511> llvm-svn: 130791
2011-05-03 23:11:17 +02:00
declare void @llvm.x86.sse.movnt.ps(i8*, <4 x float>) nounwind readnone
declare void @llvm.x86.sse2.movnt.dq(i8*, <2 x double>) nounwind readnone
declare void @llvm.x86.sse2.movnt.pd(i8*, <2 x double>) nounwind readnone
declare void @llvm.x86.sse2.movnt.i(i8*, i32) nounwind readnone
define void @f(<4 x float> %A, i8* %B, <2 x double> %C, i32 %D) {
; CHECK: store{{.*}}nontemporal
call void @llvm.x86.sse.movnt.ps(i8* %B, <4 x float> %A)
; CHECK: store{{.*}}nontemporal
call void @llvm.x86.sse2.movnt.dq(i8* %B, <2 x double> %C)
; CHECK: store{{.*}}nontemporal
call void @llvm.x86.sse2.movnt.pd(i8* %B, <2 x double> %C)
; CHECK: store{{.*}}nontemporal
call void @llvm.x86.sse2.movnt.i(i8* %B, i32 %D)
ret void
}
Add one more argument to the prefetch intrinsic to indicate whether it's a data or instruction cache access. Update the targets to match it and also teach autoupgrade. llvm-svn: 132976
2011-06-14 06:58:37 +02:00
declare void @llvm.prefetch(i8*, i32, i32) nounwind
define void @p(i8* %ptr) {
; CHECK: llvm.prefetch(i8* %ptr, i32 0, i32 1, i32 1)
tail call void @llvm.prefetch(i8* %ptr, i32 0, i32 1)
ret void
}
Split the init.trampoline intrinsic, which currently combines GCC's init.trampoline and adjust.trampoline intrinsics, into two intrinsics like in GCC. While having one combined intrinsic is tempting, it is not natural because typically the trampoline initialization needs to be done in one function, and the result of adjust trampoline is needed in a different (nested) function. To get around this llvm-gcc hacks the nested function lowering code to insert an additional parent variable holding the adjust.trampoline result that can be accessed from the child function. Dragonegg doesn't have the luxury of tweaking GCC code, so it stored the result of adjust.trampoline in the memory GCC set aside for the trampoline itself (this is always available in the child function), and set up some new memory (using an alloca) to hold the trampoline. Unfortunately this breaks Go which allocates trampoline memory on the heap and wants to use it even after the parent has exited (!). Rather than doing even more hacks to get Go working, it seemed best to just use two intrinsics like in GCC. Patch mostly by Sanjoy Das. llvm-svn: 139140
2011-09-06 15:37:06 +02:00
declare i32 @nest_f(i8* nest, i32)
declare i8* @llvm.init.trampoline(i8*, i8*, i8*)
define void @test_trampolines() {
; CHECK: call void @llvm.init.trampoline(i8* null, i8* bitcast (i32 (i8*, i32)* @nest_f to i8*), i8* null)
; CHECK: call i8* @llvm.adjust.trampoline(i8* null)
call i8* @llvm.init.trampoline(i8* null,
i8* bitcast (i32 (i8*, i32)* @nest_f to i8*),
i8* null)
ret void
}
Reference in New Issue Copy Permalink