[ThinLTO] Don't convert functions to declarations if force-import-all is enabled
#134541
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
llvmbot
added
backend:AMDGPU
LTO
Contributor
Author
This stack of pull requests is managed by Graphite. Learn more about stacking. |
Member
|
@llvm/pr-subscribers-llvm-transforms Author: Shilei Tian (shiltian) ChangesOn one hand, we intend to With this PR, functions will no longer be converted to declarations when Full diff: https://github.com/llvm/llvm-project/pull/134541.diff 4 Files Affected:
diff --git a/llvm/lib/Transforms/IPO/FunctionImport.cpp b/llvm/lib/Transforms/IPO/FunctionImport.cpp
index f1dce5d7904f9..19a6bc55b0147 100644
--- a/llvm/lib/Transforms/IPO/FunctionImport.cpp
+++ b/llvm/lib/Transforms/IPO/FunctionImport.cpp
@@ -1701,13 +1701,16 @@ void llvm::thinLTOFinalizeInModule(Module &TheModule,
if (NewLinkage == GV.getLinkage())
return;
+ bool ForceImportFunction = isa<Function>(GV) && ForceImportAll;
+
// Check for a non-prevailing def that has interposable linkage
// (e.g. non-odr weak or linkonce). In that case we can't simply
// convert to available_externally, since it would lose the
// interposable property and possibly get inlined. Simply drop
// the definition in that case.
if (GlobalValue::isAvailableExternallyLinkage(NewLinkage) &&
- GlobalValue::isInterposableLinkage(GV.getLinkage())) {
+ GlobalValue::isInterposableLinkage(GV.getLinkage()) &&
+ !ForceImportFunction) {
if (!convertToDeclaration(GV))
// FIXME: Change this to collect replaced GVs and later erase
// them from the parent module once thinLTOResolvePrevailingGUID is
diff --git a/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll b/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll
new file mode 100644
index 0000000000000..6f93f880c3157
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll
@@ -0,0 +1,15 @@
+define void @f2(ptr %v) #0 {
+entry:
+ %v.addr = alloca ptr, align 8, addrspace(5)
+ store ptr %v, ptr addrspace(5) %v.addr, align 8
+ call void @f3(ptr %v)
+ ret void
+}
+
+define weak hidden void @f3(ptr %v) #0 {
+entry:
+ store i32 12345, ptr %v
+ ret void
+}
+
+attributes #0 = { noinline }
diff --git a/llvm/test/ThinLTO/AMDGPU/force-import-all.ll b/llvm/test/ThinLTO/AMDGPU/force-import-all.ll
new file mode 100644
index 0000000000000..f8beb0dae390f
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/force-import-all.ll
@@ -0,0 +1,26 @@
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %s -o %t1.bc
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %p/Inputs/noinline.ll -o %t2.bc
+; RUN: llvm-lto -thinlto-action=run -force-import-all %t1.bc %t2.bc -thinlto-save-temps=%t3.
+; RUN: llvm-dis %t3.0.3.imported.bc -o - | FileCheck %s --check-prefix=MOD1
+; RUN: llvm-dis %t3.1.3.imported.bc -o - | FileCheck %s --check-prefix=MOD2
+
+define i32 @f1(ptr %p) #0 {
+entry:
+ call void @f2(ptr %p)
+ ret i32 0
+}
+
+define weak hidden void @f3(ptr %v) #0 {
+entry:
+ store i32 12345, ptr %v
+ ret void
+}
+
+declare void @f2(ptr)
+
+attributes #0 = { noinline }
+
+; MOD1: define weak hidden void @f3
+; MOD1: define available_externally void @f2
+; MOD2: define void @f2
+; MOD2: define available_externally hidden void @f3
diff --git a/llvm/test/ThinLTO/AMDGPU/lit.local.cfg b/llvm/test/ThinLTO/AMDGPU/lit.local.cfg
new file mode 100644
index 0000000000000..7c492428aec76
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/lit.local.cfg
@@ -0,0 +1,2 @@
+if not "AMDGPU" in config.root.targets:
+ config.unsupported = True
|
Member
|
@llvm/pr-subscribers-backend-amdgpu Author: Shilei Tian (shiltian) ChangesOn one hand, we intend to With this PR, functions will no longer be converted to declarations when Full diff: https://github.com/llvm/llvm-project/pull/134541.diff 4 Files Affected:
diff --git a/llvm/lib/Transforms/IPO/FunctionImport.cpp b/llvm/lib/Transforms/IPO/FunctionImport.cpp
index f1dce5d7904f9..19a6bc55b0147 100644
--- a/llvm/lib/Transforms/IPO/FunctionImport.cpp
+++ b/llvm/lib/Transforms/IPO/FunctionImport.cpp
@@ -1701,13 +1701,16 @@ void llvm::thinLTOFinalizeInModule(Module &TheModule,
if (NewLinkage == GV.getLinkage())
return;
+ bool ForceImportFunction = isa<Function>(GV) && ForceImportAll;
+
// Check for a non-prevailing def that has interposable linkage
// (e.g. non-odr weak or linkonce). In that case we can't simply
// convert to available_externally, since it would lose the
// interposable property and possibly get inlined. Simply drop
// the definition in that case.
if (GlobalValue::isAvailableExternallyLinkage(NewLinkage) &&
- GlobalValue::isInterposableLinkage(GV.getLinkage())) {
+ GlobalValue::isInterposableLinkage(GV.getLinkage()) &&
+ !ForceImportFunction) {
if (!convertToDeclaration(GV))
// FIXME: Change this to collect replaced GVs and later erase
// them from the parent module once thinLTOResolvePrevailingGUID is
diff --git a/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll b/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll
new file mode 100644
index 0000000000000..6f93f880c3157
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll
@@ -0,0 +1,15 @@
+define void @f2(ptr %v) #0 {
+entry:
+ %v.addr = alloca ptr, align 8, addrspace(5)
+ store ptr %v, ptr addrspace(5) %v.addr, align 8
+ call void @f3(ptr %v)
+ ret void
+}
+
+define weak hidden void @f3(ptr %v) #0 {
+entry:
+ store i32 12345, ptr %v
+ ret void
+}
+
+attributes #0 = { noinline }
diff --git a/llvm/test/ThinLTO/AMDGPU/force-import-all.ll b/llvm/test/ThinLTO/AMDGPU/force-import-all.ll
new file mode 100644
index 0000000000000..f8beb0dae390f
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/force-import-all.ll
@@ -0,0 +1,26 @@
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %s -o %t1.bc
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %p/Inputs/noinline.ll -o %t2.bc
+; RUN: llvm-lto -thinlto-action=run -force-import-all %t1.bc %t2.bc -thinlto-save-temps=%t3.
+; RUN: llvm-dis %t3.0.3.imported.bc -o - | FileCheck %s --check-prefix=MOD1
+; RUN: llvm-dis %t3.1.3.imported.bc -o - | FileCheck %s --check-prefix=MOD2
+
+define i32 @f1(ptr %p) #0 {
+entry:
+ call void @f2(ptr %p)
+ ret i32 0
+}
+
+define weak hidden void @f3(ptr %v) #0 {
+entry:
+ store i32 12345, ptr %v
+ ret void
+}
+
+declare void @f2(ptr)
+
+attributes #0 = { noinline }
+
+; MOD1: define weak hidden void @f3
+; MOD1: define available_externally void @f2
+; MOD2: define void @f2
+; MOD2: define available_externally hidden void @f3
diff --git a/llvm/test/ThinLTO/AMDGPU/lit.local.cfg b/llvm/test/ThinLTO/AMDGPU/lit.local.cfg
new file mode 100644
index 0000000000000..7c492428aec76
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/lit.local.cfg
@@ -0,0 +1,2 @@
+if not "AMDGPU" in config.root.targets:
+ config.unsupported = True
|
Member
|
@llvm/pr-subscribers-lto Author: Shilei Tian (shiltian) ChangesOn one hand, we intend to With this PR, functions will no longer be converted to declarations when Full diff: https://github.com/llvm/llvm-project/pull/134541.diff 4 Files Affected:
diff --git a/llvm/lib/Transforms/IPO/FunctionImport.cpp b/llvm/lib/Transforms/IPO/FunctionImport.cpp
index f1dce5d7904f9..19a6bc55b0147 100644
--- a/llvm/lib/Transforms/IPO/FunctionImport.cpp
+++ b/llvm/lib/Transforms/IPO/FunctionImport.cpp
@@ -1701,13 +1701,16 @@ void llvm::thinLTOFinalizeInModule(Module &TheModule,
if (NewLinkage == GV.getLinkage())
return;
+ bool ForceImportFunction = isa<Function>(GV) && ForceImportAll;
+
// Check for a non-prevailing def that has interposable linkage
// (e.g. non-odr weak or linkonce). In that case we can't simply
// convert to available_externally, since it would lose the
// interposable property and possibly get inlined. Simply drop
// the definition in that case.
if (GlobalValue::isAvailableExternallyLinkage(NewLinkage) &&
- GlobalValue::isInterposableLinkage(GV.getLinkage())) {
+ GlobalValue::isInterposableLinkage(GV.getLinkage()) &&
+ !ForceImportFunction) {
if (!convertToDeclaration(GV))
// FIXME: Change this to collect replaced GVs and later erase
// them from the parent module once thinLTOResolvePrevailingGUID is
diff --git a/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll b/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll
new file mode 100644
index 0000000000000..6f93f880c3157
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/Inputs/noinline.ll
@@ -0,0 +1,15 @@
+define void @f2(ptr %v) #0 {
+entry:
+ %v.addr = alloca ptr, align 8, addrspace(5)
+ store ptr %v, ptr addrspace(5) %v.addr, align 8
+ call void @f3(ptr %v)
+ ret void
+}
+
+define weak hidden void @f3(ptr %v) #0 {
+entry:
+ store i32 12345, ptr %v
+ ret void
+}
+
+attributes #0 = { noinline }
diff --git a/llvm/test/ThinLTO/AMDGPU/force-import-all.ll b/llvm/test/ThinLTO/AMDGPU/force-import-all.ll
new file mode 100644
index 0000000000000..f8beb0dae390f
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/force-import-all.ll
@@ -0,0 +1,26 @@
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %s -o %t1.bc
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %p/Inputs/noinline.ll -o %t2.bc
+; RUN: llvm-lto -thinlto-action=run -force-import-all %t1.bc %t2.bc -thinlto-save-temps=%t3.
+; RUN: llvm-dis %t3.0.3.imported.bc -o - | FileCheck %s --check-prefix=MOD1
+; RUN: llvm-dis %t3.1.3.imported.bc -o - | FileCheck %s --check-prefix=MOD2
+
+define i32 @f1(ptr %p) #0 {
+entry:
+ call void @f2(ptr %p)
+ ret i32 0
+}
+
+define weak hidden void @f3(ptr %v) #0 {
+entry:
+ store i32 12345, ptr %v
+ ret void
+}
+
+declare void @f2(ptr)
+
+attributes #0 = { noinline }
+
+; MOD1: define weak hidden void @f3
+; MOD1: define available_externally void @f2
+; MOD2: define void @f2
+; MOD2: define available_externally hidden void @f3
diff --git a/llvm/test/ThinLTO/AMDGPU/lit.local.cfg b/llvm/test/ThinLTO/AMDGPU/lit.local.cfg
new file mode 100644
index 0000000000000..7c492428aec76
--- /dev/null
+++ b/llvm/test/ThinLTO/AMDGPU/lit.local.cfg
@@ -0,0 +1,2 @@
+if not "AMDGPU" in config.root.targets:
+ config.unsupported = True
|
shiltian
requested review from
MaskRay,
cachemeifyoucan,
teresajohnson and
yxsamliu
shiltian
force-pushed
the
users/shiltian/dont-convert-to-decl-if-force-import-all
branch
from
70db06f to
60dd2c2
Compare
shiltian
force-pushed
the
users/shiltian/dont-convert-to-decl-if-force-import-all
branch
2 times, most recently
from
fd9f457 to
35d8ee7
Compare
cachemeifyoucan
requested changes
Apr 8, 2025
| ; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %s -o %t.main.bc | ||
| ; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %p/Inputs/in-0.ll -o %t.in.0.bc | ||
| ; RUN: opt -mtriple=amdgcn-amd-amdhsa -module-summary %p/Inputs/in-1.ll -o %t.in.1.bc | ||
| ; RUN: llvm-lto -thinlto-action=run -force-import-all %t.main.bc %t.in.0.bc %t.in.1.bc --thinlto-save-temps=%t.2. |
Collaborator
There was a problem hiding this comment.
Don't you just need to run -thinlto-action=import instead of running the full pipeline and rely on save-temps?
Contributor
Author
There was a problem hiding this comment.
I wanted to do that as well, but it looks like thinLTOResolvePrevailingGUID doesn't run if we simply does import, as shown below:
; RUN: llvm-lto -thinlto-action=thinlink -force-import-all %t.main.bc %t.in.bc -o %t.index.bc
; RUN: llvm-lto -thinlto-action=import -force-import-all %t.main.bc %t.in.bc --thinlto-index=%t.index.bc
; RUN: llvm-dis %t.main.bc.thinlto.imported.bc -o - | FileCheck --check-prefix=MOD1 %s
; RUN: llvm-dis %t.in.bc.thinlto.imported.bc -o - | FileCheck --check-prefix=MOD2 %s
Did I miss anything here?
…l` is enabled On one hand, we intend to `force-import-all` functions when the option is enabled. On the other hand, we currently drop definitions of some functions and convert them to declarations, which contradicts this intent. With this PR, functions will no longer be converted to declarations when `force-import-all` is enabled.
shiltian
force-pushed
the
users/shiltian/dont-convert-to-decl-if-force-import-all
branch
from
35d8ee7 to
3f4e024
Compare
Contributor
Author
|
@cachemeifyoucan How about this version? |
shiltian
deleted the
users/shiltian/dont-convert-to-decl-if-force-import-all
branch
Collaborator
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/59/builds/15917 Here is the relevant piece of the build log for the reference |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
On one hand, we intend to force import all functions when the option is enabled.
On the other hand, we currently drop definitions of some functions and convert
them to declarations, which contradicts this intent.
With this PR, functions will no longer be converted to declarations when
force-import-allis enabled.