[clang] 1768898 - [MemProf] Control availability of hot/cold operator new from LTO link
Teresa Johnson via cfe-commits
cfe-commits at lists.llvm.org
Mon May 8 08:02:33 PDT 2023
Author: Teresa Johnson
Date: 2023-05-08T08:02:21-07:00
New Revision: 176889868024d98db032842bc47b416997d9e349
URL: https://github.com/llvm/llvm-project/commit/176889868024d98db032842bc47b416997d9e349
DIFF: https://github.com/llvm/llvm-project/commit/176889868024d98db032842bc47b416997d9e349.diff
LOG: [MemProf] Control availability of hot/cold operator new from LTO link
Adds an LTO option to indicate that whether we are linking with an
allocator that supports hot/cold operator new interfaces. If not,
at the start of the LTO backends any existing memprof hot/cold
attributes are removed from the IR, and we also remove memprof metadata
so that post-LTO inlining doesn't add any new attributes.
This is done via setting a new flag in the module summary index. It is
important to communicate via the index to the LTO backends so that
distributed ThinLTO handles this correctly, as they are invoked by
separate clang processes and the combined index is how we communicate
information from the LTO link. Specifically, for distributed ThinLTO the
LTO related processes look like:
```
# Thin link:
$ lld --thinlto-index-only obj1.o ... objN.o -llib ...
# ThinLTO backends:
$ clang -x ir obj1.o -fthinlto-index=obj1.o.thinlto.bc -c -O2
...
$ clang -x ir objN.o -fthinlto-index=objN.o.thinlto.bc -c -O2
```
It is during the thin link (lld --thinlto-index-only) that we have
visibility into linker dependences and want to be able to pass the new
option via -Wl,-supports-hot-cold-new. This will be recorded in the
summary indexes created for the distributed backend processes
(*.thinlto.bc) and queried from there, so that we don't need to know
during those individual clang backends what allocation library was
linked. Since in-process ThinLTO and regular LTO also use a combined
index, for consistency we query the flag out of the index in all LTO
backends.
Additionally, when the LTO option is disabled, exit early from the
MemProfContextDisambiguation handling performed during LTO, as this is
unnecessary.
Depends on D149117 and D149192.
Differential Revision: https://reviews.llvm.org/D149215
Added:
clang/test/CodeGen/thinlto-distributed-supports-hot-cold-new.ll
llvm/test/LTO/X86/memprof-supports-hot-cold-new.ll
llvm/test/ThinLTO/X86/memprof-supports-hot-cold-new.ll
Modified:
llvm/include/llvm/IR/ModuleSummaryIndex.h
llvm/include/llvm/LTO/LTO.h
llvm/lib/Bitcode/Reader/BitcodeReader.cpp
llvm/lib/IR/ModuleSummaryIndex.cpp
llvm/lib/LTO/LTO.cpp
llvm/lib/LTO/LTOBackend.cpp
llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp
llvm/test/ThinLTO/X86/memprof-basic.ll
llvm/test/ThinLTO/X86/memprof-duplicate-context-ids.ll
llvm/test/ThinLTO/X86/memprof-duplicate-context-ids2.ll
llvm/test/ThinLTO/X86/memprof-funcassigncloning.ll
llvm/test/ThinLTO/X86/memprof-indirectcall.ll
llvm/test/ThinLTO/X86/memprof-inlined.ll
llvm/test/ThinLTO/X86/memprof-inlined2.ll
llvm/test/Transforms/MemProfContextDisambiguation/basic.ll
llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids.ll
llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids2.ll
llvm/test/Transforms/MemProfContextDisambiguation/funcassigncloning.ll
llvm/test/Transforms/MemProfContextDisambiguation/indirectcall.ll
llvm/test/Transforms/MemProfContextDisambiguation/inlined.ll
llvm/test/Transforms/MemProfContextDisambiguation/inlined2.ll
Removed:
################################################################################
diff --git a/clang/test/CodeGen/thinlto-distributed-supports-hot-cold-new.ll b/clang/test/CodeGen/thinlto-distributed-supports-hot-cold-new.ll
new file mode 100644
index 000000000000..e213fbaf3fa1
--- /dev/null
+++ b/clang/test/CodeGen/thinlto-distributed-supports-hot-cold-new.ll
@@ -0,0 +1,70 @@
+; REQUIRES: x86-registered-target
+
+;; Test that passing -supports-hot-cold-new to the thin link prevents memprof
+;; metadata and attributes from being removed from the distributed ThinLTO
+;; backend, and vice versa without passing -supports-hot-cold-new.
+
+;; First check with -supports-hot-cold-new.
+; RUN: opt -thinlto-bc %s >%t.o
+; RUN: llvm-lto2 run %t.o -save-temps \
+; RUN: -supports-hot-cold-new \
+; RUN: -thinlto-distributed-indexes \
+; RUN: -r=%t.o,main,plx \
+; RUN: -r=%t.o,_Znam, \
+; RUN: -o %t.out
+
+;; Ensure that the index file reflects the -supports-hot-cold-new, as that is
+;; how the ThinLTO backend behavior is controlled.
+; RUN: llvm-dis %t.out.index.bc -o - | FileCheck %s --check-prefix=CHECK-INDEX-ON
+;; Flags are printed in decimal, but this corresponds to 0x161, and 0x100 is
+;; the value indicating -supports-hot-cold-new was enabled.
+; CHECK-INDEX-ON: flags: 353
+
+; RUN: %clang -target x86_64-unknown-linux-gnu -O2 -o %t1.o -x ir %t.o -c -fthinlto-index=%t.o.thinlto.bc -save-temps=obj
+
+; RUN: llvm-dis %t.s.0.preopt.bc -o - | FileCheck %s --check-prefix=CHECK-IR
+; CHECK-IR: !memprof {{.*}} !callsite
+; CHECK-IR: "memprof"="cold"
+
+;; Next check without -supports-hot-cold-new, we should not perform
+;; context disambiguation, and we should strip memprof metadata and
+;; attributes before optimization during the distributed backend.
+; RUN: llvm-lto2 run %t.o -save-temps \
+; RUN: -thinlto-distributed-indexes \
+; RUN: -r=%t.o,main,plx \
+; RUN: -r=%t.o,_Znam, \
+; RUN: -o %t.out
+
+;; Ensure that the index file reflects not having -supports-hot-cold-new.
+; RUN: llvm-dis %t.out.index.bc -o - | FileCheck %s --check-prefix=CHECK-INDEX-OFF
+;; Flags are printed in decimal, but this corresponds to 0x61, without 0x100 set.
+; CHECK-INDEX-OFF: flags: 97
+
+; RUN: %clang -target x86_64-unknown-linux-gnu -O2 -o %t1.o -x ir %t.o -c -fthinlto-index=%t.o.thinlto.bc -save-temps=obj
+
+; RUN: llvm-dis %t.s.0.preopt.bc -o - | FileCheck %s \
+; RUN: --implicit-check-not "!memprof" --implicit-check-not "!callsite" \
+; RUN: --implicit-check-not "memprof"="cold"
+
+source_filename = "thinlto-distributed-supports-hot-cold-new.ll"
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+define i32 @main() #0 {
+entry:
+ %call = call ptr @_Znam(i64 0), !memprof !0, !callsite !5
+ %call1 = call ptr @_Znam(i64 0) #1
+ ret i32 0
+}
+
+declare ptr @_Znam(i64)
+
+attributes #0 = { noinline optnone }
+attributes #1 = { "memprof"="cold" }
+
+!0 = !{!1, !3}
+!1 = !{!2, !"notcold"}
+!2 = !{i64 9086428284934609951, i64 -5964873800580613432, i64 2732490490862098848, i64 8632435727821051414}
+!3 = !{!4, !"cold"}
+!4 = !{i64 9086428284934609951, i64 -5964873800580613432, i64 2732490490862098848, i64 -3421689549917153178}
+!5 = !{i64 9086428284934609951}
diff --git a/llvm/include/llvm/IR/ModuleSummaryIndex.h b/llvm/include/llvm/IR/ModuleSummaryIndex.h
index 6fdc378ab5c3..b411148b7722 100644
--- a/llvm/include/llvm/IR/ModuleSummaryIndex.h
+++ b/llvm/include/llvm/IR/ModuleSummaryIndex.h
@@ -1305,6 +1305,9 @@ class ModuleSummaryIndex {
/// Indicates that summary-based synthetic entry count propagation has run
bool HasSyntheticEntryCounts = false;
+ /// Indicates that we linked with allocator supporting hot/cold new operators.
+ bool WithSupportsHotColdNew = false;
+
/// Indicates that distributed backend should skip compilation of the
/// module. Flag is suppose to be set by distributed ThinLTO indexing
/// when it detected that the module is not needed during the final
@@ -1513,6 +1516,9 @@ class ModuleSummaryIndex {
bool hasSyntheticEntryCounts() const { return HasSyntheticEntryCounts; }
void setHasSyntheticEntryCounts() { HasSyntheticEntryCounts = true; }
+ bool withSupportsHotColdNew() const { return WithSupportsHotColdNew; }
+ void setWithSupportsHotColdNew() { WithSupportsHotColdNew = true; }
+
bool skipModuleByDistributedBackend() const {
return SkipModuleByDistributedBackend;
}
diff --git a/llvm/include/llvm/LTO/LTO.h b/llvm/include/llvm/LTO/LTO.h
index dabc367c80c8..b7dfd630ad82 100644
--- a/llvm/include/llvm/LTO/LTO.h
+++ b/llvm/include/llvm/LTO/LTO.h
@@ -95,6 +95,11 @@ setupStatsFile(StringRef StatsFilename);
/// ordered indices to elements in the input array.
std::vector<int> generateModulesOrdering(ArrayRef<BitcodeModule *> R);
+/// Updates MemProf attributes (and metadata) based on whether the index
+/// has recorded that we are linking with allocation libraries containing
+/// the necessary APIs for downstream transformations.
+void updateMemProfAttributes(Module &Mod, const ModuleSummaryIndex &Index);
+
class LTO;
struct SymbolResolution;
class ThinBackendProc;
diff --git a/llvm/lib/Bitcode/Reader/BitcodeReader.cpp b/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
index 64616240d4a3..a1d704455892 100644
--- a/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
+++ b/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
@@ -8067,7 +8067,7 @@ static Expected<bool> getEnableSplitLTOUnitFlag(BitstreamCursor &Stream,
case bitc::FS_FLAGS: { // [flags]
uint64_t Flags = Record[0];
// Scan flags.
- assert(Flags <= 0xff && "Unexpected bits in flag");
+ assert(Flags <= 0x1ff && "Unexpected bits in flag");
return Flags & 0x8;
}
diff --git a/llvm/lib/IR/ModuleSummaryIndex.cpp b/llvm/lib/IR/ModuleSummaryIndex.cpp
index 99e5044bdb9b..7cba03e93e97 100644
--- a/llvm/lib/IR/ModuleSummaryIndex.cpp
+++ b/llvm/lib/IR/ModuleSummaryIndex.cpp
@@ -107,11 +107,13 @@ uint64_t ModuleSummaryIndex::getFlags() const {
Flags |= 0x40;
if (withWholeProgramVisibility())
Flags |= 0x80;
+ if (withSupportsHotColdNew())
+ Flags |= 0x100;
return Flags;
}
void ModuleSummaryIndex::setFlags(uint64_t Flags) {
- assert(Flags <= 0xff && "Unexpected bits in flag");
+ assert(Flags <= 0x1ff && "Unexpected bits in flag");
// 1 bit: WithGlobalValueDeadStripping flag.
// Set on combined index only.
if (Flags & 0x1)
@@ -145,6 +147,10 @@ void ModuleSummaryIndex::setFlags(uint64_t Flags) {
// Set on combined index only.
if (Flags & 0x80)
setWithWholeProgramVisibility();
+ // 1 bit: WithSupportsHotColdNew flag.
+ // Set on combined index only.
+ if (Flags & 0x100)
+ setWithSupportsHotColdNew();
}
// Collect for the given module the list of function it defines
diff --git a/llvm/lib/LTO/LTO.cpp b/llvm/lib/LTO/LTO.cpp
index fee09fcd76bd..de9b8f1da62b 100644
--- a/llvm/lib/LTO/LTO.cpp
+++ b/llvm/lib/LTO/LTO.cpp
@@ -76,6 +76,10 @@ cl::opt<bool> EnableLTOInternalization(
cl::desc("Enable global value internalization in LTO"));
}
+/// Indicate we are linking with an allocator that supports hot/cold operator
+/// new interfaces.
+extern cl::opt<bool> SupportsHotColdNew;
+
/// Enable MemProf context disambiguation for thin link.
extern cl::opt<bool> EnableMemProfContextDisambiguation;
@@ -1079,6 +1083,14 @@ Error LTO::run(AddStreamFn AddStream, FileCache Cache) {
return StatsFileOrErr.takeError();
std::unique_ptr<ToolOutputFile> StatsFile = std::move(StatsFileOrErr.get());
+ // TODO: Ideally this would be controlled automatically by detecting that we
+ // are linking with an allocator that supports these interfaces, rather than
+ // an internal option (which would still be needed for tests, however). For
+ // example, if the library exported a symbol like __malloc_hot_cold the linker
+ // could recognize that and set a flag in the lto::Config.
+ if (SupportsHotColdNew)
+ ThinLTO.CombinedIndex.setWithSupportsHotColdNew();
+
Error Result = runRegularLTO(AddStream);
if (!Result)
Result = runThinLTO(AddStream, Cache, GUIDPreservedSymbols);
@@ -1089,6 +1101,37 @@ Error LTO::run(AddStreamFn AddStream, FileCache Cache) {
return Result;
}
+void lto::updateMemProfAttributes(Module &Mod,
+ const ModuleSummaryIndex &Index) {
+ if (Index.withSupportsHotColdNew())
+ return;
+
+ // The profile matcher applies hotness attributes directly for allocations,
+ // and those will cause us to generate calls to the hot/cold interfaces
+ // unconditionally. If supports-hot-cold-new was not enabled in the LTO
+ // link then assume we don't want these calls (e.g. not linking with
+ // the appropriate library, or otherwise trying to disable this behavior).
+ for (auto &F : Mod) {
+ for (auto &BB : F) {
+ for (auto &I : BB) {
+ auto *CI = dyn_cast<CallBase>(&I);
+ if (!CI)
+ continue;
+ if (CI->hasFnAttr("memprof"))
+ CI->removeFnAttr("memprof");
+ // Strip off all memprof metadata as it is no longer needed.
+ // Importantly, this avoids the addition of new memprof attributes
+ // after inlining propagation.
+ // TODO: If we support additional types of MemProf metadata beyond hot
+ // and cold, we will need to update the metadata based on the allocator
+ // APIs supported instead of completely stripping all.
+ CI->setMetadata(LLVMContext::MD_memprof, nullptr);
+ CI->setMetadata(LLVMContext::MD_callsite, nullptr);
+ }
+ }
+ }
+}
+
Error LTO::runRegularLTO(AddStreamFn AddStream) {
// Setup optimization remarks.
auto DiagFileOrErr = lto::setupLLVMOptimizationRemarks(
@@ -1142,6 +1185,8 @@ Error LTO::runRegularLTO(AddStreamFn AddStream) {
}
}
+ updateMemProfAttributes(*RegularLTO.CombinedModule, ThinLTO.CombinedIndex);
+
// If allowed, upgrade public vcall visibility metadata to linkage unit
// visibility before whole program devirtualization in the optimizer.
updateVCallVisibilityInModule(*RegularLTO.CombinedModule,
diff --git a/llvm/lib/LTO/LTOBackend.cpp b/llvm/lib/LTO/LTOBackend.cpp
index d574ae5737d5..a18963fcaf85 100644
--- a/llvm/lib/LTO/LTOBackend.cpp
+++ b/llvm/lib/LTO/LTOBackend.cpp
@@ -565,6 +565,8 @@ Error lto::thinBackend(const Config &Conf, unsigned Task, AddStreamFn AddStream,
// the module, if applicable.
Mod.setPartialSampleProfileRatio(CombinedIndex);
+ updateMemProfAttributes(Mod, CombinedIndex);
+
updatePublicTypeTestCalls(Mod, CombinedIndex.withWholeProgramVisibility());
if (Conf.CodeGenOnly) {
diff --git a/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp b/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp
index e65b9dad1380..a3ca910a6b15 100644
--- a/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp
+++ b/llvm/lib/Transforms/IPO/MemProfContextDisambiguation.cpp
@@ -104,6 +104,12 @@ static cl::opt<std::string> MemProfImportSummary(
cl::desc("Import summary to use for testing the ThinLTO backend via opt"),
cl::Hidden);
+// Indicate we are linking with an allocator that supports hot/cold operator
+// new interfaces.
+cl::opt<bool> SupportsHotColdNew(
+ "supports-hot-cold-new", cl::init(false), cl::Hidden,
+ cl::desc("Linking with hot/cold operator new interfaces"));
+
/// CRTP base for graphs built from either IR or ThinLTO summary index.
///
/// The graph represents the call contexts in all memprof metadata on allocation
@@ -3190,6 +3196,17 @@ bool MemProfContextDisambiguation::processModule(
if (ImportSummary)
return applyImport(M);
+ // TODO: If/when other types of memprof cloning are enabled beyond just for
+ // hot and cold, we will need to change this to individually control the
+ // AllocationType passed to addStackNodesForMIB during CCG construction.
+ // Note that we specifically check this after applying imports above, so that
+ // the option isn't needed to be passed to distributed ThinLTO backend
+ // clang processes, which won't necessarily have visibility into the linker
+ // dependences. Instead the information is communicated from the LTO link to
+ // the backends via the combined summary index.
+ if (!SupportsHotColdNew)
+ return false;
+
ModuleCallsiteContextGraph CCG(M, OREGetter);
return CCG.process();
}
@@ -3241,6 +3258,14 @@ void MemProfContextDisambiguation::run(
ModuleSummaryIndex &Index,
function_ref<bool(GlobalValue::GUID, const GlobalValueSummary *)>
isPrevailing) {
+ // TODO: If/when other types of memprof cloning are enabled beyond just for
+ // hot and cold, we will need to change this to individually control the
+ // AllocationType passed to addStackNodesForMIB during CCG construction.
+ // The index was set from the option, so these should be in sync.
+ assert(Index.withSupportsHotColdNew() == SupportsHotColdNew);
+ if (!SupportsHotColdNew)
+ return;
+
IndexCallsiteContextGraph CCG(Index, isPrevailing);
CCG.process();
}
diff --git a/llvm/test/LTO/X86/memprof-supports-hot-cold-new.ll b/llvm/test/LTO/X86/memprof-supports-hot-cold-new.ll
new file mode 100644
index 000000000000..3ed68e8137a3
--- /dev/null
+++ b/llvm/test/LTO/X86/memprof-supports-hot-cold-new.ll
@@ -0,0 +1,65 @@
+;; Test that passing -supports-hot-cold-new to the LTO link allows context
+;; disambiguation to proceed, and also prevents memprof metadata and attributes
+;; from being removed from the LTO backend, and vice versa without passing
+;; -supports-hot-cold-new.
+
+;; Note that this tests regular LTO (with a summary) due to the module flag
+;; disabling ThinLTO.
+
+;; First check with -supports-hot-cold-new.
+; RUN: opt -module-summary %s >%t.o
+; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
+; RUN: -r=%t.o,main,plx \
+; RUN: -r=%t.o,_Znam, \
+; RUN: -memprof-dump-ccg \
+; RUN: -save-temps \
+; RUN: -o %t.out 2>&1 | FileCheck %s --check-prefix=DUMP
+; DUMP: Callsite Context Graph:
+
+; RUN: llvm-dis %t.out.0.0.preopt.bc -o - | FileCheck %s --check-prefix=IR
+; IR: !memprof {{.*}} !callsite
+; IR: "memprof"="cold"
+
+;; Next check without -supports-hot-cold-new, we should not perform
+;; context disambiguation, and we should strip memprof metadata and
+;; attributes before optimization.
+; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -r=%t.o,main,plx \
+; RUN: -r=%t.o,_Znam, \
+; RUN: -memprof-dump-ccg \
+; RUN: -save-temps \
+; RUN: -o %t.out 2>&1 | FileCheck %s --allow-empty \
+; RUN: --implicit-check-not "Callsite Context Graph:"
+
+; RUN: llvm-dis %t.out.0.0.preopt.bc -o - | FileCheck %s \
+; RUN: --implicit-check-not "!memprof" --implicit-check-not "!callsite" \
+; RUN: --implicit-check-not "memprof"="cold"
+
+source_filename = "memprof-supports-hot-cold-new.ll"
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+define i32 @main() #0 {
+entry:
+ %call = call ptr @_Znam(i64 0), !memprof !0, !callsite !5
+ %call2 = call ptr @_Znam(i64 0) #1
+ ret i32 0
+}
+
+declare ptr @_Znam(i64)
+
+attributes #0 = { noinline optnone }
+attributes #1 = { "memprof"="cold" }
+
+!llvm.module.flags = !{!6}
+
+!0 = !{!1, !3}
+!1 = !{!2, !"notcold"}
+!2 = !{i64 9086428284934609951, i64 -5964873800580613432, i64 2732490490862098848, i64 8632435727821051414}
+!3 = !{!4, !"cold"}
+!4 = !{i64 9086428284934609951, i64 -5964873800580613432, i64 2732490490862098848, i64 -3421689549917153178}
+!5 = !{i64 9086428284934609951}
+
+;; Force regular LTO even though we have a summary.
+!6 = !{i32 1, !"ThinLTO", i32 0}
diff --git a/llvm/test/ThinLTO/X86/memprof-basic.ll b/llvm/test/ThinLTO/X86/memprof-basic.ll
index ac3119fff2ae..0d466830ba57 100644
--- a/llvm/test/ThinLTO/X86/memprof-basic.ll
+++ b/llvm/test/ThinLTO/X86/memprof-basic.ll
@@ -36,6 +36,7 @@
; RUN: opt -thinlto-bc %s >%t.o
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
; RUN: -r=%t.o,sleep, \
@@ -55,6 +56,7 @@
;; Try again but with distributed ThinLTO
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -thinlto-distributed-indexes \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
diff --git a/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids.ll b/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids.ll
index 55aab1924108..5bc2f3005f46 100644
--- a/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids.ll
+++ b/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids.ll
@@ -58,6 +58,7 @@
; RUN: opt -thinlto-bc %s >%t.o
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
; RUN: -r=%t.o,sleep, \
@@ -78,6 +79,7 @@
;; Try again but with distributed ThinLTO
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -thinlto-distributed-indexes \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
diff --git a/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids2.ll b/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids2.ll
index b77df883d35b..13d2036822b2 100644
--- a/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids2.ll
+++ b/llvm/test/ThinLTO/X86/memprof-duplicate-context-ids2.ll
@@ -95,6 +95,7 @@
; RUN: opt -thinlto-bc %s >%t.o
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_Z1Db,plx \
; RUN: -r=%t.o,_Z1Cb,plx \
diff --git a/llvm/test/ThinLTO/X86/memprof-funcassigncloning.ll b/llvm/test/ThinLTO/X86/memprof-funcassigncloning.ll
index 25e974e98f45..9a72ae43b2f1 100644
--- a/llvm/test/ThinLTO/X86/memprof-funcassigncloning.ll
+++ b/llvm/test/ThinLTO/X86/memprof-funcassigncloning.ll
@@ -51,6 +51,7 @@
; RUN: opt -thinlto-bc %s >%t.o
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
; RUN: -r=%t.o,sleep, \
@@ -65,6 +66,7 @@
;; Try again but with distributed ThinLTO
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -thinlto-distributed-indexes \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
diff --git a/llvm/test/ThinLTO/X86/memprof-indirectcall.ll b/llvm/test/ThinLTO/X86/memprof-indirectcall.ll
index 8b9b64f0c192..76273959f4f4 100644
--- a/llvm/test/ThinLTO/X86/memprof-indirectcall.ll
+++ b/llvm/test/ThinLTO/X86/memprof-indirectcall.ll
@@ -56,6 +56,7 @@
; RUN: opt -thinlto-bc %s >%t.o
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,sleep, \
; RUN: -r=%t.o,_Znam, \
@@ -78,6 +79,7 @@
;; Try again but with distributed ThinLTO
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -thinlto-distributed-indexes \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
diff --git a/llvm/test/ThinLTO/X86/memprof-inlined.ll b/llvm/test/ThinLTO/X86/memprof-inlined.ll
index bc537a5f3fbb..feb9c9434422 100644
--- a/llvm/test/ThinLTO/X86/memprof-inlined.ll
+++ b/llvm/test/ThinLTO/X86/memprof-inlined.ll
@@ -46,6 +46,7 @@
; RUN: opt -thinlto-bc %s >%t.o
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
; RUN: -r=%t.o,sleep, \
@@ -67,6 +68,7 @@
;; Try again but with distributed ThinLTO
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -thinlto-distributed-indexes \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_ZdaPv, \
diff --git a/llvm/test/ThinLTO/X86/memprof-inlined2.ll b/llvm/test/ThinLTO/X86/memprof-inlined2.ll
index 02baf9fb4cd8..9f67849e1aa3 100644
--- a/llvm/test/ThinLTO/X86/memprof-inlined2.ll
+++ b/llvm/test/ThinLTO/X86/memprof-inlined2.ll
@@ -44,6 +44,7 @@
; RUN: opt -thinlto-bc %s >%t.o
; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
; RUN: -r=%t.o,main,plx \
; RUN: -r=%t.o,_Z3barv,plx \
; RUN: -r=%t.o,_Z3bazv,plx \
diff --git a/llvm/test/ThinLTO/X86/memprof-supports-hot-cold-new.ll b/llvm/test/ThinLTO/X86/memprof-supports-hot-cold-new.ll
new file mode 100644
index 000000000000..9e69377d1444
--- /dev/null
+++ b/llvm/test/ThinLTO/X86/memprof-supports-hot-cold-new.ll
@@ -0,0 +1,57 @@
+;; Test that passing -supports-hot-cold-new to the LTO link allows context
+;; disambiguation to proceed, and also prevents memprof metadata and attributes
+;; from being removed from the LTO backend, and vice versa without passing
+;; -supports-hot-cold-new.
+
+;; First check with -supports-hot-cold-new.
+; RUN: opt -thinlto-bc %s >%t.o
+; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -supports-hot-cold-new \
+; RUN: -r=%t.o,main,plx \
+; RUN: -r=%t.o,_Znam, \
+; RUN: -memprof-dump-ccg \
+; RUN: -save-temps \
+; RUN: -o %t.out 2>&1 | FileCheck %s --check-prefix=DUMP
+; DUMP: Callsite Context Graph:
+
+; RUN: llvm-dis %t.out.1.0.preopt.bc -o - | FileCheck %s --check-prefix=IR
+; IR: !memprof {{.*}} !callsite
+; IR: "memprof"="cold"
+
+;; Next check without -supports-hot-cold-new, we should not perform
+;; context disambiguation, and we should strip memprof metadata and
+;; attributes before optimization.
+; RUN: llvm-lto2 run %t.o -enable-memprof-context-disambiguation \
+; RUN: -r=%t.o,main,plx \
+; RUN: -r=%t.o,_Znam, \
+; RUN: -memprof-dump-ccg \
+; RUN: -save-temps \
+; RUN: -o %t.out 2>&1 | FileCheck %s --allow-empty \
+; RUN: --implicit-check-not "Callsite Context Graph:"
+
+; RUN: llvm-dis %t.out.1.0.preopt.bc -o - | FileCheck %s \
+; RUN: --implicit-check-not "!memprof" --implicit-check-not "!callsite" \
+; RUN: --implicit-check-not "memprof"="cold"
+
+source_filename = "memprof-supports-hot-cold-new.ll"
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+define i32 @main() #0 {
+entry:
+ %call = call ptr @_Znam(i64 0), !memprof !0, !callsite !5
+ %call2 = call ptr @_Znam(i64 0) #1
+ ret i32 0
+}
+
+declare ptr @_Znam(i64)
+
+attributes #0 = { noinline optnone }
+attributes #1 = { "memprof"="cold" }
+
+!0 = !{!1, !3}
+!1 = !{!2, !"notcold"}
+!2 = !{i64 9086428284934609951, i64 -5964873800580613432, i64 2732490490862098848, i64 8632435727821051414}
+!3 = !{!4, !"cold"}
+!4 = !{i64 9086428284934609951, i64 -5964873800580613432, i64 2732490490862098848, i64 -3421689549917153178}
+!5 = !{i64 9086428284934609951}
diff --git a/llvm/test/Transforms/MemProfContextDisambiguation/basic.ll b/llvm/test/Transforms/MemProfContextDisambiguation/basic.ll
index bd938754ce9d..483582c6ced9 100644
--- a/llvm/test/Transforms/MemProfContextDisambiguation/basic.ll
+++ b/llvm/test/Transforms/MemProfContextDisambiguation/basic.ll
@@ -34,7 +34,7 @@
;; -stats requires asserts
; REQUIRES: asserts
-; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: opt -passes=memprof-context-disambiguation -supports-hot-cold-new \
; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
; RUN: -memprof-export-to-dot -memprof-dot-file-path-prefix=%t. \
; RUN: -stats -pass-remarks=memprof-context-disambiguation \
@@ -45,6 +45,15 @@
;; We should have cloned bar, baz, and foo, for the cold memory allocation.
; RUN: cat %t.ccg.cloned.dot | FileCheck %s --check-prefix=DOTCLONED
+;; Check again without -supports-hot-cold-new and ensure all MIB are cold and
+;; that there is no cloning.
+; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
+; RUN: -memprof-export-to-dot -memprof-dot-file-path-prefix=%t. \
+; RUN: -stats -pass-remarks=memprof-context-disambiguation \
+; RUN: %s -S 2>&1 | FileCheck %s --implicit-check-not="Callsite Context Graph" \
+; RUN: --implicit-check-not="created clone"
+
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
diff --git a/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids.ll b/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids.ll
index 1f23ad3c6a51..53de404c8f28 100644
--- a/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids.ll
+++ b/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids.ll
@@ -56,7 +56,7 @@
;; -stats requires asserts
; REQUIRES: asserts
-; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: opt -passes=memprof-context-disambiguation -supports-hot-cold-new \
; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
; RUN: -memprof-export-to-dot -memprof-dot-file-path-prefix=%t. \
; RUN: -stats -pass-remarks=memprof-context-disambiguation \
diff --git a/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids2.ll b/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids2.ll
index d1659b524860..aa51006d263e 100644
--- a/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids2.ll
+++ b/llvm/test/Transforms/MemProfContextDisambiguation/duplicate-context-ids2.ll
@@ -93,7 +93,7 @@
;;
;; The IR was then reduced using llvm-reduce with the expected FileCheck input.
-; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: opt -passes=memprof-context-disambiguation -supports-hot-cold-new \
; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
; RUN: -memprof-export-to-dot -memprof-dot-file-path-prefix=%t. \
; RUN: %s -S 2>&1 | FileCheck %s --check-prefix=DUMP
diff --git a/llvm/test/Transforms/MemProfContextDisambiguation/funcassigncloning.ll b/llvm/test/Transforms/MemProfContextDisambiguation/funcassigncloning.ll
index b94e9b855b74..ac19eeca3b33 100644
--- a/llvm/test/Transforms/MemProfContextDisambiguation/funcassigncloning.ll
+++ b/llvm/test/Transforms/MemProfContextDisambiguation/funcassigncloning.ll
@@ -48,7 +48,7 @@
;; -stats requires asserts
; REQUIRES: asserts
-; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: opt -passes=memprof-context-disambiguation -supports-hot-cold-new \
; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
; RUN: -stats -pass-remarks=memprof-context-disambiguation \
; RUN: %s -S 2>&1 | FileCheck %s --check-prefix=DUMP --check-prefix=IR \
diff --git a/llvm/test/Transforms/MemProfContextDisambiguation/indirectcall.ll b/llvm/test/Transforms/MemProfContextDisambiguation/indirectcall.ll
index f3216aa13d88..a37aa4ccf1f4 100644
--- a/llvm/test/Transforms/MemProfContextDisambiguation/indirectcall.ll
+++ b/llvm/test/Transforms/MemProfContextDisambiguation/indirectcall.ll
@@ -54,7 +54,7 @@
;; -stats requires asserts
; REQUIRES: asserts
-; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: opt -passes=memprof-context-disambiguation -supports-hot-cold-new \
; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
; RUN: -memprof-export-to-dot -memprof-dot-file-path-prefix=%t. \
; RUN: -stats -pass-remarks=memprof-context-disambiguation \
diff --git a/llvm/test/Transforms/MemProfContextDisambiguation/inlined.ll b/llvm/test/Transforms/MemProfContextDisambiguation/inlined.ll
index f1b74f13fb14..06162430449a 100644
--- a/llvm/test/Transforms/MemProfContextDisambiguation/inlined.ll
+++ b/llvm/test/Transforms/MemProfContextDisambiguation/inlined.ll
@@ -44,7 +44,7 @@
;; -stats requires asserts
; REQUIRES: asserts
-; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: opt -passes=memprof-context-disambiguation -supports-hot-cold-new \
; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
; RUN: -memprof-export-to-dot -memprof-dot-file-path-prefix=%t. \
; RUN: -stats -pass-remarks=memprof-context-disambiguation \
diff --git a/llvm/test/Transforms/MemProfContextDisambiguation/inlined2.ll b/llvm/test/Transforms/MemProfContextDisambiguation/inlined2.ll
index 5a247f860799..cc97b5290e25 100644
--- a/llvm/test/Transforms/MemProfContextDisambiguation/inlined2.ll
+++ b/llvm/test/Transforms/MemProfContextDisambiguation/inlined2.ll
@@ -42,7 +42,7 @@
;;
;; The IR was then reduced using llvm-reduce with the expected FileCheck input.
-; RUN: opt -passes=memprof-context-disambiguation \
+; RUN: opt -passes=memprof-context-disambiguation -supports-hot-cold-new \
; RUN: -memprof-verify-ccg -memprof-verify-nodes -memprof-dump-ccg \
; RUN: %s -S 2>&1 | FileCheck %s --check-prefix=DUMP
More information about the cfe-commits
mailing list