[Flang][Driver] Emit module summary for Full LTO #164302

anchuraj · 2025-10-20T18:53:18Z

This PR enables module summary for Full LTO. Module summaries are enabled by default for Full LTO in clang, this change makes the flang behaviour consistent. Reference PR: https://reviews.llvm.org/D34156

llvmbot · 2025-10-20T18:53:52Z

@llvm/pr-subscribers-flang-driver

Author: Anchu Rajendran S (anchuraj)

Changes

This PR enables module summary for LTO. Information in module summary is used in performing LTO especially in Thin LTO.

Full diff: https://github.com/llvm/llvm-project/pull/164302.diff

3 Files Affected:

(modified) flang/lib/Frontend/FrontendActions.cpp (+10-6)
(modified) flang/test/CMakeLists.txt (+1)
(modified) flang/test/Driver/lto-bc.f90 (+25-11)

diff --git a/flang/lib/Frontend/FrontendActions.cpp b/flang/lib/Frontend/FrontendActions.cpp
index 0c630d2ba876d..0ff5ba17ca874 100644
--- a/flang/lib/Frontend/FrontendActions.cpp
+++ b/flang/lib/Frontend/FrontendActions.cpp
@@ -1019,11 +1019,15 @@ void CodeGenAction::runOptimizationPipeline(llvm::raw_pwrite_stream &os) {
 
   // Create the pass manager.
   llvm::ModulePassManager mpm;
+  // The module summary should be emitted by default for regular LTO
+  // except for ld64 targets.
+  bool emitSummary = (opts.PrepareForThinLTO || opts.PrepareForFullLTO) &&
+                     (triple.getVendor() != llvm::Triple::Apple);
+
+  if (emitSummary && !opts.PrepareForThinLTO)
+    llvmModule->addModuleFlag(llvm::Module::Error, "ThinLTO", uint32_t(0));
+
   if (opts.PrepareForFatLTO) {
-    // The module summary should be emitted by default for regular LTO
-    // except for ld64 targets.
-    bool emitSummary = opts.PrepareForThinLTO || opts.PrepareForFullLTO ||
-                       triple.getVendor() != llvm::Triple::Apple;
     mpm = pb.buildFatLTODefaultPipeline(level, opts.PrepareForThinLTO,
                                         emitSummary);
   } else if (opts.PrepareForFullLTO)
@@ -1034,9 +1038,9 @@ void CodeGenAction::runOptimizationPipeline(llvm::raw_pwrite_stream &os) {
     mpm = pb.buildPerModuleDefaultPipeline(level);
 
   if (action == BackendActionTy::Backend_EmitBC)
-    mpm.addPass(llvm::BitcodeWriterPass(os));
+    mpm.addPass(llvm::BitcodeWriterPass(os, false, emitSummary));
   else if (action == BackendActionTy::Backend_EmitLL)
-    mpm.addPass(llvm::PrintModulePass(os));
+    mpm.addPass(llvm::PrintModulePass(os, "", false, emitSummary));
 
   // FIXME: This should eventually be replaced by a first-class driver option.
   // This should be done for both flang and clang simultaneously.
diff --git a/flang/test/CMakeLists.txt b/flang/test/CMakeLists.txt
index da557f9ec3443..8c8e92faa787a 100644
--- a/flang/test/CMakeLists.txt
+++ b/flang/test/CMakeLists.txt
@@ -72,6 +72,7 @@ if (NOT FLANG_STANDALONE_BUILD)
     FileCheck
     count
     not
+    llvm-bcanalyzer
     llvm-dis
     llvm-objcopy
     llvm-objdump
diff --git a/flang/test/Driver/lto-bc.f90 b/flang/test/Driver/lto-bc.f90
index 5e34cdb87c5b1..fbf964b849eb7 100644
--- a/flang/test/Driver/lto-bc.f90
+++ b/flang/test/Driver/lto-bc.f90
@@ -1,21 +1,35 @@
 ! Test that the output is LLVM bitcode for LTO and not a native objectfile by
-! disassembling it to LLVM IR.
-! Right now there is nothing special about it and it is similar to non-lto IR,
-! more work is needed to add things like module summaries.
+! disassembling it to LLVM IR. Also tests module summaries are emitted for LTO
 
 ! RUN: %flang %s -c -o - | not llvm-dis -o %t
 ! RUN: %flang_fc1 %s -emit-llvm-bc -o - | llvm-dis -o - | FileCheck %s
-
-! RUN: %flang -flto %s -c -o - | llvm-dis -o - | FileCheck %s
-! RUN: %flang -flto=thin %s -c -o - | llvm-dis -o - | FileCheck %s
-
 ! CHECK: define void @_QQmain()
 ! CHECK-NEXT:  ret void
 ! CHECK-NEXT: }
+! CHECK-NOT: !{{.*}} = !{i32 1, !"ThinLTO", i32 0}
+! CHECK-NOT: ^{{.*}} = module:
+! CHECK-NOT: ^{{.*}} = gv: (name:
+! CHECK-NOT: ^{{.*}} = blockcount:
+
+! RUN: %flang -flto %s -c -o - | llvm-dis -o - | FileCheck %s --check-prefix=FULL
+! THIN: define void @_QQmain()
+! THIN-NEXT:  ret void
+! THIN-NEXT: }
+! THIN-NOT: !{{.*}} = !{i32 1, !"ThinLTO", i32 0}
+! THIN: ^{{.*}} = module:
+! THIN: ^{{.*}} = gv: (name:
+! THIN: ^{{.*}} = blockcount:
 
-! CHECK-NOT: ^0 = module:
-! CHECK-NOT: ^1 = gv: (name:
-! CHECK-NOT: ^2 = flags:
-! CHECK-NOT: ^3 = blockcount:
+! RUN: %flang -flto=thin %s -c -o - | llvm-dis -o - | FileCheck %s --check-prefix=THIN
+! FULL: define void @_QQmain()
+! FULL-NEXT:  ret void
+! FULL-NEXT: }
+! FULL: !{{.*}} = !{i32 1, !"ThinLTO", i32 0}
+! FULL: ^{{.*}} = module:
+! FULL: ^{{.*}} = gv: (name:
+! FULL: ^{{.*}} = blockcount:
 
+! RUN: %flang_fc1 -flto -emit-llvm-bc %s -o - | llvm-bcanalyzer -dump| FileCheck --check-prefix=MOD-SUMM %s
+! MOD-SUMM: FULL_LTO_GLOBALVAL_SUMMARY_BLOCK
+program main
 end program

flang/lib/Frontend/FrontendActions.cpp

flang/test/Driver/lto-bc.f90

tarunprabhu · 2025-10-21T00:59:41Z

flang/test/Driver/lto-bc.f90

+! THIN: define void @_QQmain()
+! THIN-NEXT:  ret void
+! THIN-NEXT: }
+! THIN-NOT: !{{.*}} = !{i32 1, !"ThinLTO", i32 0}


When setting -flto=full, metadata named ThinLTO is present in the module (line 27), but when -flto=thin, it is not? Does the 0 value for "ThinLTO" mean that the ThinLTO is off?

Also, it looks like the output in both the thin and full LTO cases are the same - or at least, what is being checked in this test is the same. What is the difference between the two schemes?

I apologize for my ignorance - as I said, I am not very familiar with LLVM's LTO implementation. Do you mind clarifying some of this? Thanks.

Thanks for the clarification (in the response to a separate comment)

To be clear, if testing the output of full vs. thin is too much, we need not do it - presumably that will have been tested elsewhere, perhaps in llvm/LTO. But if we can test something other than the presence/absence of "ThinLTO" metadata, we probably should.

Thank you! Working on these.

Hi, ThinLTO requires more work to enable module summaries. I was wrong to test only the emitted module summary. At least https://github.com/llvm/llvm-project/blob/main/clang/lib/CodeGen/BackendUtil.cpp#L1168C9-L1182C8 needs to be enabled. I have removed the tests for thinLTO. This PR enables only module summary for Full LTO by default (same as https://reviews.llvm.org/D34156). Thank you for pointing out this

In this, will -flto=thin and -flto=full produce the same output? Or will the output of -flto=thin be the same as the output if a -flto= option was not given at all?

Right now, flto = thin does not emit the summary, only flto = full emits. Thin LTO summary emission logic does not change - it is same as previous:

if (action == BackendActionTy::Backend_EmitBC) mpm.addPass(llvm::BitcodeWriterPass(os)); else if (action == BackendActionTy::Backend_EmitLL) mpm.addPass(llvm::PrintModulePass(os));

flang hello.f90 -c -emit-llvm -o -| llvm-dis -o - and flang -flto=thin hello.f90 -c -o -| llvm-dis -o - will produce the same output in ll files.

Sorry, I got a bit confused and asked a strange question. Let me try again.

In this test, when -flto=full, we are checking for the "ThinLTO = 0" metadata and the presence of "module:, gv:, blockcount:" (I assume that this is the summary, is that right?).

However, when flto=thin, we are only checking for the absence of the "ThinLTO" metadata. We are not checking for the absence of the summaries. Should we be doing that as well?

Yes, it is the module summary. Thanks a lot for reviewing this and helping me to improve my change. Since I am learning through this as well, there are misses which I try to correct. As this is a work in progress and as is a feature not yet implemented, I did not have preferences for checking or not checking this. I have added it back since it is more consistent with the existed test case.

github-actions · 2025-10-21T18:55:25Z

✅ With the latest revision this PR passed the C/C++ code formatter.

tarunprabhu · 2025-10-21T20:05:15Z

flang/lib/Frontend/FrontendActions.cpp

-                       triple.getVendor() != llvm::Triple::Apple;
+  // The module summary should be emitted by default for regular LTO
+  // except for ld64 targets.
+  bool emitSummary = (opts.PrepareForThinLTO || opts.PrepareForFullLTO) &&


Since ThinLTO support is not being enabled here, should emitSummary only be true if opt.PrepareForFullLTO is set? Should PrepareForThinLTO be ignored here?

Summary should be emitted for thinLTO. I will work on this as a follow upThe flag set logic is correct. However it is not used for thin LTO right now.

I was thinking that since emitSummary is ignored in the if (opts.PrepareForThinLTO) branch on line 1038, the code should look like this for now:

bool emitSummary = opts.PrepareForFullLTO && triple.getVendor() != llvm::Triple::Apple;

When support for summaries with ThinLTO is implemented, it can become:

bool emitSummary = (opts.PrepareForThinLTO || opts.PrepareForFullLTO) && triple.getVendor() != llvm::Triple::Apple;

However, since you intend to work on enabling summaries for thin LTO right away (if I understand you correctly), it should be ok to leave this code as it is. Otherwise, it can be a bit confusing.

What do you think?

Updated. Agree that it can be confusing.

tarunprabhu · 2025-10-21T20:42:57Z

flang/lib/Frontend/FrontendActions.cpp

+  if (action == BackendActionTy::Backend_EmitBC ||
+      action == BackendActionTy::Backend_EmitLL || opts.PrepareForFatLTO) {
+    if (opts.PrepareForThinLTO) {
+      // TODO: ThinLTO module summary support is yet to be enabled.


Does -flto=thin imply that module summaries are enabled? If so, it may be a good idea to issue a diagnostic saying that thin-lto support is incomplete. Do you intend to implement module summaries for thin-lto? If that is likely to land relatively soon, it may be ok not to emit a diagnostic.

The warning exists: https://github.com/llvm/llvm-project/blob/main/flang/test/Driver/lto-flags.f90#L33 I will work on this.

Thanks for working on this

tarunprabhu · 2025-10-21T20:46:48Z

flang/test/Driver/lto-bc.f90

+! THIN: define void @_QQmain()
+! THIN-NEXT:  ret void
+! THIN-NEXT: }
+! THIN-NOT: !{{.*}} = !{i32 1, !"ThinLTO", i32 0}


In this, will -flto=thin and -flto=full produce the same output? Or will the output of -flto=thin be the same as the output if a -flto= option was not given at all?

tarunprabhu

Thanks for addressing my comments. Unfortunately, I was not at all clear, and I have tried again. Thanks for your patience.

tarunprabhu · 2025-10-22T17:04:18Z

flang/lib/Frontend/FrontendActions.cpp

-                       triple.getVendor() != llvm::Triple::Apple;
+  // The module summary should be emitted by default for regular LTO
+  // except for ld64 targets.
+  bool emitSummary = (opts.PrepareForThinLTO || opts.PrepareForFullLTO) &&


I was thinking that since emitSummary is ignored in the if (opts.PrepareForThinLTO) branch on line 1038, the code should look like this for now:

bool emitSummary = opts.PrepareForFullLTO && triple.getVendor() != llvm::Triple::Apple;

When support for summaries with ThinLTO is implemented, it can become:

bool emitSummary = (opts.PrepareForThinLTO || opts.PrepareForFullLTO) && triple.getVendor() != llvm::Triple::Apple;

However, since you intend to work on enabling summaries for thin LTO right away (if I understand you correctly), it should be ok to leave this code as it is. Otherwise, it can be a bit confusing.

What do you think?

tarunprabhu · 2025-10-22T17:12:59Z

flang/test/Driver/lto-bc.f90

+! THIN: define void @_QQmain()
+! THIN-NEXT:  ret void
+! THIN-NEXT: }
+! THIN-NOT: !{{.*}} = !{i32 1, !"ThinLTO", i32 0}


Sorry, I got a bit confused and asked a strange question. Let me try again.

In this test, when -flto=full, we are checking for the "ThinLTO = 0" metadata and the presence of "module:, gv:, blockcount:" (I assume that this is the summary, is that right?).

However, when flto=thin, we are only checking for the absence of the "ThinLTO" metadata. We are not checking for the absence of the summaries. Should we be doing that as well?

tarunprabhu

Thanks for your patience and for the changes.

This PR enables module summary for Full LTO. Module summaries are enabled by default for Full LTO in clang, this change makes the flang behaviour consistent. Reference PR: https://reviews.llvm.org/D34156

llvmbot added flang:driver flang Flang issues not falling into any other category labels Oct 20, 2025

[Flang][Driver] Emit module summary for LTO

3309893

anchuraj force-pushed the enable-module-summary branch from 8d5366a to 3309893 Compare October 20, 2025 19:03

anchuraj requested review from abidh, dpalermo, mjklemm and tarunprabhu October 20, 2025 19:09

tarunprabhu reviewed Oct 21, 2025

View reviewed changes

R2: Addressing review comments

1ecbeb9

anchuraj force-pushed the enable-module-summary branch from 139bc9b to 7ed30f8 Compare October 21, 2025 18:53

anchuraj changed the title ~~[Flang][Driver] Emit module summary for LTO~~ [Flang][Driver] Emit module summary for Full LTO Oct 21, 2025

anchuraj requested a review from tarunprabhu October 21, 2025 19:00

Removing the changes from ThinLTO

51e02c1

anchuraj force-pushed the enable-module-summary branch from 7ed30f8 to 51e02c1 Compare October 21, 2025 19:01

tarunprabhu reviewed Oct 21, 2025

View reviewed changes

anchuraj requested a review from tarunprabhu October 22, 2025 17:12

tarunprabhu reviewed Oct 22, 2025

View reviewed changes

anchuraj added 2 commits October 22, 2025 12:20

R3: Addressing review comments

c050671

R4: Addressing review comments

ee40244

anchuraj force-pushed the enable-module-summary branch from a6dd62b to ee40244 Compare October 22, 2025 17:40

anchuraj requested a review from tarunprabhu October 22, 2025 17:53

tarunprabhu approved these changes Oct 22, 2025

View reviewed changes

anchuraj merged commit 1031f1b into llvm:main Oct 23, 2025
10 checks passed

[Flang][Driver] Emit module summary for Full LTO #164302

[Flang][Driver] Emit module summary for Full LTO #164302

Uh oh!

Conversation

anchuraj commented Oct 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

llvmbot commented Oct 20, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anchuraj Oct 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Oct 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anchuraj Oct 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anchuraj Oct 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tarunprabhu left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tarunprabhu left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

anchuraj commented Oct 20, 2025 •

edited

Loading

anchuraj Oct 22, 2025 •

edited

Loading

github-actions bot commented Oct 21, 2025 •

edited

Loading

anchuraj Oct 22, 2025 •

edited

Loading

anchuraj Oct 22, 2025 •

edited

Loading