summaryrefslogtreecommitdiff
path: root/test
diff options
context:
space:
mode:
authorChandler Carruth <chandlerc@gmail.com>2014-02-06 04:37:03 +0000
committerChandler Carruth <chandlerc@gmail.com>2014-02-06 04:37:03 +0000
commit57732bff1e6166817b7408d2099a5adb666747f0 (patch)
tree41d594b3daa9e9f9be9cde5f099f8c7cb70f594f /test
parent9b71cd85ffa6a81a9b7959ca6e2185fcce03735a (diff)
downloadllvm-57732bff1e6166817b7408d2099a5adb666747f0.tar.gz
llvm-57732bff1e6166817b7408d2099a5adb666747f0.tar.bz2
llvm-57732bff1e6166817b7408d2099a5adb666747f0.tar.xz
[PM] Add a new "lazy" call graph analysis pass for the new pass manager.
The primary motivation for this pass is to separate the call graph analysis used by the new pass manager's CGSCC pass management from the existing call graph analysis pass. That analysis pass is (somewhat unfortunately) over-constrained by the existing CallGraphSCCPassManager requirements. Those requirements make it *really* hard to cleanly layer the needed functionality for the new pass manager on top of the existing analysis. However, there are also a bunch of things that the pass manager would specifically benefit from doing differently from the existing call graph analysis, and this new implementation tries to address several of them: - Be lazy about scanning function definitions. The existing pass eagerly scans the entire module to build the initial graph. This new pass is significantly more lazy, and I plan to push this even further to maximize locality during CGSCC walks. - Don't use a single synthetic node to partition functions with an indirect call from functions whose address is taken. This node creates a huge choke-point which would preclude good parallelization across the fanout of the SCC graph when we got to the point of looking at such changes to LLVM. - Use a memory dense and lightweight representation of the call graph rather than value handles and tracking call instructions. This will require explicit update calls instead of some updates working transparently, but should end up being significantly more efficient. The explicit update calls ended up being needed in many cases for the existing call graph so we don't really lose anything. - Doesn't explicitly model SCCs and thus doesn't provide an "identity" for an SCC which is stable across updates. This is essential for the new pass manager to work correctly. - Only form the graph necessary for traversing all of the functions in an SCC friendly order. This is a much simpler graph structure and should be more memory dense. It does limit the ways in which it is appropriate to use this analysis. I wish I had a better name than "call graph". I've commented extensively this aspect. This is still very much a WIP, in fact it is really just the initial bits. But it is about the fourth version of the initial bits that I've implemented with each of the others running into really frustrating problms. This looks like it will actually work and I'd like to split the actual complexity across commits for the sake of my reviewers. =] The rest of the implementation along with lots of wiring will follow somewhat more rapidly now that there is a good path forward. Naturally, this doesn't impact any of the existing optimizer. This code is specific to the new pass manager. A bunch of thanks are deserved for the various folks that have helped with the design of this, especially Nick Lewycky who actually sat with me to go through the fundamentals of the final version here. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@200903 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test')
-rw-r--r--test/Analysis/LazyCallGraph/basic.ll126
1 files changed, 126 insertions, 0 deletions
diff --git a/test/Analysis/LazyCallGraph/basic.ll b/test/Analysis/LazyCallGraph/basic.ll
new file mode 100644
index 0000000000..ebadb75154
--- /dev/null
+++ b/test/Analysis/LazyCallGraph/basic.ll
@@ -0,0 +1,126 @@
+; RUN: opt -disable-output -passes=print-cg %s 2>&1 | FileCheck %s
+;
+; Basic validation of the call graph analysis used in the new pass manager.
+
+define void @f() {
+; CHECK-LABEL: Call edges in function: f
+; CHECK-NOT: ->
+
+entry:
+ ret void
+}
+
+; A bunch more functions just to make it easier to test several call edges at once.
+define void @f1() {
+ ret void
+}
+define void @f2() {
+ ret void
+}
+define void @f3() {
+ ret void
+}
+define void @f4() {
+ ret void
+}
+define void @f5() {
+ ret void
+}
+define void @f6() {
+ ret void
+}
+define void @f7() {
+ ret void
+}
+define void @f8() {
+ ret void
+}
+define void @f9() {
+ ret void
+}
+define void @f10() {
+ ret void
+}
+define void @f11() {
+ ret void
+}
+define void @f12() {
+ ret void
+}
+
+declare i32 @__gxx_personality_v0(...)
+
+define void @test0() {
+; CHECK-LABEL: Call edges in function: test0
+; CHECK-NEXT: -> f
+; CHECK-NOT: ->
+
+entry:
+ call void @f()
+ call void @f()
+ call void @f()
+ call void @f()
+ ret void
+}
+
+define void ()* @test1(void ()** %x) {
+; CHECK-LABEL: Call edges in function: test1
+; CHECK-NEXT: -> f12
+; CHECK-NEXT: -> f11
+; CHECK-NEXT: -> f10
+; CHECK-NEXT: -> f7
+; CHECK-NEXT: -> f9
+; CHECK-NEXT: -> f8
+; CHECK-NEXT: -> f6
+; CHECK-NEXT: -> f5
+; CHECK-NEXT: -> f4
+; CHECK-NEXT: -> f3
+; CHECK-NEXT: -> f2
+; CHECK-NEXT: -> f1
+; CHECK-NOT: ->
+
+entry:
+ br label %next
+
+dead:
+ br label %next
+
+next:
+ phi void ()* [ @f1, %entry ], [ @f2, %dead ]
+ select i1 true, void ()* @f3, void ()* @f4
+ store void ()* @f5, void ()** %x
+ call void @f6()
+ call void (void ()*, void ()*)* bitcast (void ()* @f7 to void (void ()*, void ()*)*)(void ()* @f8, void ()* @f9)
+ invoke void @f10() to label %exit unwind label %unwind
+
+exit:
+ ret void ()* @f11
+
+unwind:
+ %res = landingpad { i8*, i32 } personality i32 (...)* @__gxx_personality_v0
+ cleanup
+ resume { i8*, i32 } { i8* bitcast (void ()* @f12 to i8*), i32 42 }
+}
+
+@g = global void ()* @f1
+@g1 = global [4 x void ()*] [void ()* @f2, void ()* @f3, void ()* @f4, void ()* @f5]
+@g2 = global {i8, void ()*, i8} {i8 1, void ()* @f6, i8 2}
+@h = constant void ()* @f7
+
+define void @test2() {
+; CHECK-LABEL: Call edges in function: test2
+; CHECK-NEXT: -> f7
+; CHECK-NEXT: -> f6
+; CHECK-NEXT: -> f5
+; CHECK-NEXT: -> f4
+; CHECK-NEXT: -> f3
+; CHECK-NEXT: -> f2
+; CHECK-NEXT: -> f1
+; CHECK-NOT: ->
+
+ load i8** bitcast (void ()** @g to i8**)
+ load i8** bitcast (void ()** getelementptr ([4 x void ()*]* @g1, i32 0, i32 2) to i8**)
+ load i8** bitcast (void ()** getelementptr ({i8, void ()*, i8}* @g2, i32 0, i32 1) to i8**)
+ load i8** bitcast (void ()** @h to i8**)
+ ret void
+}