[TLI] Add support for lgamma libcall. #114536
Open
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
|
@llvm/pr-subscribers-llvm-analysis Author: Kenji Mouri / 毛利 研二 (MouriNaruto) ChangesThis patch adds basic support for Related issue: #113804 Note: I'm sorry that I made a similar PR that I had created (#114343) recently because one of my friends, @fawdlstty, hopes me to help him solve that issue in private. I think it's a good time for me to practice again. I hope I can make this PR merged successfully without reverting. I hope I can create PRs for different topics that I am interested in for the next time instead of solving friends' issues. I know this style PRs should be left to other people who want to participate in the LLVM for the first time. I hope @arsenm and @c8ef can help me review the code. Kenji Mouri Full diff: https://github.com/llvm/llvm-project/pull/114536.diff 6 Files Affected:
diff --git a/llvm/include/llvm/Analysis/TargetLibraryInfo.def b/llvm/include/llvm/Analysis/TargetLibraryInfo.def
index fd53a26ef8fc11..04d6aef56783d5 100644
--- a/llvm/include/llvm/Analysis/TargetLibraryInfo.def
+++ b/llvm/include/llvm/Analysis/TargetLibraryInfo.def
@@ -1726,6 +1726,21 @@ TLI_DEFINE_ENUM_INTERNAL(ldexpl)
TLI_DEFINE_STRING_INTERNAL("ldexpl")
TLI_DEFINE_SIG_INTERNAL(LDbl, LDbl, Int)
+/// double lgamma(double x);
+TLI_DEFINE_ENUM_INTERNAL(lgamma)
+TLI_DEFINE_STRING_INTERNAL("lgamma")
+TLI_DEFINE_SIG_INTERNAL(Dbl, Dbl)
+
+/// float lgammaf(float x);
+TLI_DEFINE_ENUM_INTERNAL(lgammaf)
+TLI_DEFINE_STRING_INTERNAL("lgammaf")
+TLI_DEFINE_SIG_INTERNAL(Flt, Flt)
+
+/// long double lgammal(long double x);
+TLI_DEFINE_ENUM_INTERNAL(lgammal)
+TLI_DEFINE_STRING_INTERNAL("lgammal")
+TLI_DEFINE_SIG_INTERNAL(LDbl, LDbl)
+
/// long long int llabs(long long int j);
TLI_DEFINE_ENUM_INTERNAL(llabs)
TLI_DEFINE_STRING_INTERNAL("llabs")
diff --git a/llvm/lib/Analysis/TargetLibraryInfo.cpp b/llvm/lib/Analysis/TargetLibraryInfo.cpp
index 7f0b98ab3c1514..a54ab8c9966c3c 100644
--- a/llvm/lib/Analysis/TargetLibraryInfo.cpp
+++ b/llvm/lib/Analysis/TargetLibraryInfo.cpp
@@ -301,6 +301,7 @@ static void initializeLibCalls(TargetLibraryInfoImpl &TLI, const Triple &T,
TLI.setUnavailable(LibFunc_floorf);
TLI.setUnavailable(LibFunc_fmodf);
TLI.setUnavailable(LibFunc_hypotf);
+ TLI.setUnavailable(LibFunc_lgammaf);
TLI.setUnavailable(LibFunc_log10f);
TLI.setUnavailable(LibFunc_logf);
TLI.setUnavailable(LibFunc_modff);
@@ -334,6 +335,7 @@ static void initializeLibCalls(TargetLibraryInfoImpl &TLI, const Triple &T,
TLI.setUnavailable(LibFunc_frexpl);
TLI.setUnavailable(LibFunc_hypotl);
TLI.setUnavailable(LibFunc_ldexpl);
+ TLI.setUnavailable(LibFunc_lgammal);
TLI.setUnavailable(LibFunc_log10l);
TLI.setUnavailable(LibFunc_logl);
TLI.setUnavailable(LibFunc_modfl);
diff --git a/llvm/lib/Transforms/Utils/BuildLibCalls.cpp b/llvm/lib/Transforms/Utils/BuildLibCalls.cpp
index e039457f313b29..35001809daeecb 100644
--- a/llvm/lib/Transforms/Utils/BuildLibCalls.cpp
+++ b/llvm/lib/Transforms/Utils/BuildLibCalls.cpp
@@ -1221,6 +1221,9 @@ bool llvm::inferNonMandatoryLibFuncAttrs(Function &F,
case LibFunc_isascii:
case LibFunc_isdigit:
case LibFunc_labs:
+ case LibFunc_lgamma:
+ case LibFunc_lgammaf:
+ case LibFunc_lgammal:
case LibFunc_llabs:
case LibFunc_log:
case LibFunc_log10:
diff --git a/llvm/test/Transforms/InferFunctionAttrs/annotate.ll b/llvm/test/Transforms/InferFunctionAttrs/annotate.ll
index 452d90aa98d88d..e10e60f3a86763 100644
--- a/llvm/test/Transforms/InferFunctionAttrs/annotate.ll
+++ b/llvm/test/Transforms/InferFunctionAttrs/annotate.ll
@@ -619,6 +619,15 @@ declare float @ldexpf(float, i32)
; CHECK: declare x86_fp80 @ldexpl(x86_fp80, i32) [[NOFREE_WILLRETURN]]
declare x86_fp80 @ldexpl(x86_fp80, i32)
+; CHECK: declare double @lgamma(double) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
+declare double @lgamma(double)
+
+; CHECK: declare float @lgammaf(float) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
+declare float @lgammaf(float)
+
+; CHECK: declare x86_fp80 @lgammal(x86_fp80) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
+declare x86_fp80 @lgammal(x86_fp80)
+
; CHECK: declare i64 @llabs(i64) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
declare i64 @llabs(i64)
diff --git a/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml b/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml
index 6702725e4fc8a4..83e35b50c12920 100644
--- a/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml
+++ b/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml
@@ -34,7 +34,7 @@
#
# CHECK: << Total TLI yes SDK no: 18
# CHECK: >> Total TLI no SDK yes: 0
-# CHECK: == Total TLI yes SDK yes: 271
+# CHECK: == Total TLI yes SDK yes: 274
#
# WRONG_DETAIL: << TLI yes SDK no : '_ZdaPv' aka operator delete[](void*)
# WRONG_DETAIL: >> TLI no SDK yes: '_ZdaPvj' aka operator delete[](void*, unsigned int)
@@ -48,14 +48,14 @@
# WRONG_DETAIL: << TLI yes SDK no : 'fminimum_numl'
# WRONG_SUMMARY: << Total TLI yes SDK no: 19{{$}}
# WRONG_SUMMARY: >> Total TLI no SDK yes: 1{{$}}
-# WRONG_SUMMARY: == Total TLI yes SDK yes: 270
+# WRONG_SUMMARY: == Total TLI yes SDK yes: 273
#
## The -COUNT suffix doesn't care if there are too many matches, so check
## the exact count first; the two directives should add up to that.
## Yes, this means additions to TLI will fail this test, but the argument
## to -COUNT can't be an expression.
-# AVAIL: TLI knows 522 symbols, 289 available
-# AVAIL-COUNT-289: {{^}} available
+# AVAIL: TLI knows 525 symbols, 292 available
+# AVAIL-COUNT-292: {{^}} available
# AVAIL-NOT: {{^}} available
# UNAVAIL-COUNT-233: not available
# UNAVAIL-NOT: not available
@@ -634,6 +634,18 @@ DynamicSymbols:
Type: STT_FUNC
Section: .text
Binding: STB_GLOBAL
+ - Name: lgamma
+ Type: STT_FUNC
+ Section: .text
+ Binding: STB_GLOBAL
+ - Name: lgammaf
+ Type: STT_FUNC
+ Section: .text
+ Binding: STB_GLOBAL
+ - Name: lgammal
+ Type: STT_FUNC
+ Section: .text
+ Binding: STB_GLOBAL
- Name: llabs
Type: STT_FUNC
Section: .text
diff --git a/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp b/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp
index 982d00c5d33593..34d8cae9ab3927 100644
--- a/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp
+++ b/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp
@@ -272,6 +272,9 @@ TEST_F(TargetLibraryInfoTest, ValidProto) {
"declare i32 @ilogb(double)\n"
"declare i32 @ilogbf(float)\n"
"declare i32 @ilogbl(x86_fp80)\n"
+ "declare double @lgamma(double)\n"
+ "declare float @lgammaf(float)\n"
+ "declare x86_fp80 @lgammal(x86_fp80)\n"
"declare double @logb(double)\n"
"declare float @logbf(float)\n"
"declare x86_fp80 @logbl(x86_fp80)\n"
|
|
@llvm/pr-subscribers-llvm-transforms Author: Kenji Mouri / 毛利 研二 (MouriNaruto) ChangesThis patch adds basic support for Related issue: #113804 Note: I'm sorry that I made a similar PR that I had created (#114343) recently because one of my friends, @fawdlstty, hopes me to help him solve that issue in private. I think it's a good time for me to practice again. I hope I can make this PR merged successfully without reverting. I hope I can create PRs for different topics that I am interested in for the next time instead of solving friends' issues. I know this style PRs should be left to other people who want to participate in the LLVM for the first time. I hope @arsenm and @c8ef can help me review the code. Kenji Mouri Full diff: https://github.com/llvm/llvm-project/pull/114536.diff 6 Files Affected:
diff --git a/llvm/include/llvm/Analysis/TargetLibraryInfo.def b/llvm/include/llvm/Analysis/TargetLibraryInfo.def
index fd53a26ef8fc11..04d6aef56783d5 100644
--- a/llvm/include/llvm/Analysis/TargetLibraryInfo.def
+++ b/llvm/include/llvm/Analysis/TargetLibraryInfo.def
@@ -1726,6 +1726,21 @@ TLI_DEFINE_ENUM_INTERNAL(ldexpl)
TLI_DEFINE_STRING_INTERNAL("ldexpl")
TLI_DEFINE_SIG_INTERNAL(LDbl, LDbl, Int)
+/// double lgamma(double x);
+TLI_DEFINE_ENUM_INTERNAL(lgamma)
+TLI_DEFINE_STRING_INTERNAL("lgamma")
+TLI_DEFINE_SIG_INTERNAL(Dbl, Dbl)
+
+/// float lgammaf(float x);
+TLI_DEFINE_ENUM_INTERNAL(lgammaf)
+TLI_DEFINE_STRING_INTERNAL("lgammaf")
+TLI_DEFINE_SIG_INTERNAL(Flt, Flt)
+
+/// long double lgammal(long double x);
+TLI_DEFINE_ENUM_INTERNAL(lgammal)
+TLI_DEFINE_STRING_INTERNAL("lgammal")
+TLI_DEFINE_SIG_INTERNAL(LDbl, LDbl)
+
/// long long int llabs(long long int j);
TLI_DEFINE_ENUM_INTERNAL(llabs)
TLI_DEFINE_STRING_INTERNAL("llabs")
diff --git a/llvm/lib/Analysis/TargetLibraryInfo.cpp b/llvm/lib/Analysis/TargetLibraryInfo.cpp
index 7f0b98ab3c1514..a54ab8c9966c3c 100644
--- a/llvm/lib/Analysis/TargetLibraryInfo.cpp
+++ b/llvm/lib/Analysis/TargetLibraryInfo.cpp
@@ -301,6 +301,7 @@ static void initializeLibCalls(TargetLibraryInfoImpl &TLI, const Triple &T,
TLI.setUnavailable(LibFunc_floorf);
TLI.setUnavailable(LibFunc_fmodf);
TLI.setUnavailable(LibFunc_hypotf);
+ TLI.setUnavailable(LibFunc_lgammaf);
TLI.setUnavailable(LibFunc_log10f);
TLI.setUnavailable(LibFunc_logf);
TLI.setUnavailable(LibFunc_modff);
@@ -334,6 +335,7 @@ static void initializeLibCalls(TargetLibraryInfoImpl &TLI, const Triple &T,
TLI.setUnavailable(LibFunc_frexpl);
TLI.setUnavailable(LibFunc_hypotl);
TLI.setUnavailable(LibFunc_ldexpl);
+ TLI.setUnavailable(LibFunc_lgammal);
TLI.setUnavailable(LibFunc_log10l);
TLI.setUnavailable(LibFunc_logl);
TLI.setUnavailable(LibFunc_modfl);
diff --git a/llvm/lib/Transforms/Utils/BuildLibCalls.cpp b/llvm/lib/Transforms/Utils/BuildLibCalls.cpp
index e039457f313b29..35001809daeecb 100644
--- a/llvm/lib/Transforms/Utils/BuildLibCalls.cpp
+++ b/llvm/lib/Transforms/Utils/BuildLibCalls.cpp
@@ -1221,6 +1221,9 @@ bool llvm::inferNonMandatoryLibFuncAttrs(Function &F,
case LibFunc_isascii:
case LibFunc_isdigit:
case LibFunc_labs:
+ case LibFunc_lgamma:
+ case LibFunc_lgammaf:
+ case LibFunc_lgammal:
case LibFunc_llabs:
case LibFunc_log:
case LibFunc_log10:
diff --git a/llvm/test/Transforms/InferFunctionAttrs/annotate.ll b/llvm/test/Transforms/InferFunctionAttrs/annotate.ll
index 452d90aa98d88d..e10e60f3a86763 100644
--- a/llvm/test/Transforms/InferFunctionAttrs/annotate.ll
+++ b/llvm/test/Transforms/InferFunctionAttrs/annotate.ll
@@ -619,6 +619,15 @@ declare float @ldexpf(float, i32)
; CHECK: declare x86_fp80 @ldexpl(x86_fp80, i32) [[NOFREE_WILLRETURN]]
declare x86_fp80 @ldexpl(x86_fp80, i32)
+; CHECK: declare double @lgamma(double) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
+declare double @lgamma(double)
+
+; CHECK: declare float @lgammaf(float) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
+declare float @lgammaf(float)
+
+; CHECK: declare x86_fp80 @lgammal(x86_fp80) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
+declare x86_fp80 @lgammal(x86_fp80)
+
; CHECK: declare i64 @llabs(i64) [[NOFREE_NOUNWIND_WILLRETURN_WRITEONLY]]
declare i64 @llabs(i64)
diff --git a/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml b/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml
index 6702725e4fc8a4..83e35b50c12920 100644
--- a/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml
+++ b/llvm/test/tools/llvm-tli-checker/ps4-tli-check.yaml
@@ -34,7 +34,7 @@
#
# CHECK: << Total TLI yes SDK no: 18
# CHECK: >> Total TLI no SDK yes: 0
-# CHECK: == Total TLI yes SDK yes: 271
+# CHECK: == Total TLI yes SDK yes: 274
#
# WRONG_DETAIL: << TLI yes SDK no : '_ZdaPv' aka operator delete[](void*)
# WRONG_DETAIL: >> TLI no SDK yes: '_ZdaPvj' aka operator delete[](void*, unsigned int)
@@ -48,14 +48,14 @@
# WRONG_DETAIL: << TLI yes SDK no : 'fminimum_numl'
# WRONG_SUMMARY: << Total TLI yes SDK no: 19{{$}}
# WRONG_SUMMARY: >> Total TLI no SDK yes: 1{{$}}
-# WRONG_SUMMARY: == Total TLI yes SDK yes: 270
+# WRONG_SUMMARY: == Total TLI yes SDK yes: 273
#
## The -COUNT suffix doesn't care if there are too many matches, so check
## the exact count first; the two directives should add up to that.
## Yes, this means additions to TLI will fail this test, but the argument
## to -COUNT can't be an expression.
-# AVAIL: TLI knows 522 symbols, 289 available
-# AVAIL-COUNT-289: {{^}} available
+# AVAIL: TLI knows 525 symbols, 292 available
+# AVAIL-COUNT-292: {{^}} available
# AVAIL-NOT: {{^}} available
# UNAVAIL-COUNT-233: not available
# UNAVAIL-NOT: not available
@@ -634,6 +634,18 @@ DynamicSymbols:
Type: STT_FUNC
Section: .text
Binding: STB_GLOBAL
+ - Name: lgamma
+ Type: STT_FUNC
+ Section: .text
+ Binding: STB_GLOBAL
+ - Name: lgammaf
+ Type: STT_FUNC
+ Section: .text
+ Binding: STB_GLOBAL
+ - Name: lgammal
+ Type: STT_FUNC
+ Section: .text
+ Binding: STB_GLOBAL
- Name: llabs
Type: STT_FUNC
Section: .text
diff --git a/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp b/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp
index 982d00c5d33593..34d8cae9ab3927 100644
--- a/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp
+++ b/llvm/unittests/Analysis/TargetLibraryInfoTest.cpp
@@ -272,6 +272,9 @@ TEST_F(TargetLibraryInfoTest, ValidProto) {
"declare i32 @ilogb(double)\n"
"declare i32 @ilogbf(float)\n"
"declare i32 @ilogbl(x86_fp80)\n"
+ "declare double @lgamma(double)\n"
+ "declare float @lgammaf(float)\n"
+ "declare x86_fp80 @lgammal(x86_fp80)\n"
"declare double @logb(double)\n"
"declare float @logbf(float)\n"
"declare x86_fp80 @logbl(x86_fp80)\n"
|
c8ef
approved these changes
Nov 1, 2024
In my experience, constant folding the gamma function is not easy because it requires invoking the corresponding libc function, which is implemented differently on various platforms. I have noticed different return values on Windows and Linux (x86), two of the most commonly used platforms. Writing a test case to satisfy the LLVM build bot may not be that trivial. Good luck! |
Thanks. I will try my best to implement this constant fold from the current LLVM codebase. I think I will learn a lot if I can make a proper PR for that. Kenji Mouri |
arsenm
reviewed
Nov 1, 2024
| @@ -334,6 +335,7 @@ static void initializeLibCalls(TargetLibraryInfoImpl &TLI, const Triple &T, | |||
| TLI.setUnavailable(LibFunc_frexpl); | |||
| TLI.setUnavailable(LibFunc_hypotl); | |||
| TLI.setUnavailable(LibFunc_ldexpl); | |||
| TLI.setUnavailable(LibFunc_lgammal); | |||
There was a problem hiding this comment.
As a pre-commit can you add new baseline TLI checker tests for the windows and linux cases? This is untested, and this is a problem in every one of these new libcall patches.
There was a problem hiding this comment.
Here is the PR: #114556
arsenm
approved these changes
Nov 5, 2024
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.
This patch adds basic support for
lgamma. Constant folding support will be submitted in a subsequent patch.Related issue: #113804
Note: I'm sorry that I made a similar PR that I had created (#114343) recently because one of my friends, @fawdlstty, hopes me to help him solve that issue in private. (It is important to finish promises made to friends.) I think it's a good time for me to practice again. I hope I can make this PR merged successfully without reverting.
I hope I can create PRs for different topics that I am interested in for the next time instead of solving friends' issues. I know this style PRs should be left to other people who want to participate in the LLVM for the first time.
I hope @arsenm and @c8ef can help me review the code.
Kenji Mouri