Skip to content

Conversation

@wangzpgi
Copy link
Contributor

@wangzpgi wangzpgi commented May 6, 2025

The point of ignore_tkr(c) is to ignore both contiguous warnings and errors for arguments of all attribute types.

@wangzpgi wangzpgi requested review from clementval and klausler May 6, 2025 21:01
@wangzpgi wangzpgi self-assigned this May 6, 2025
@llvmbot llvmbot added flang Flang issues not falling into any other category flang:semantics labels May 6, 2025
@llvmbot
Copy link
Member

llvmbot commented May 6, 2025

@llvm/pr-subscribers-flang-semantics

Author: Zhen Wang (wangzpgi)

Changes

The point of ignore_tkr(c) is to ignore both contiguous warnings and errors for arguments of all attribute types.


Full diff: https://github.com/llvm/llvm-project/pull/138762.diff

2 Files Affected:

  • (modified) flang/lib/Semantics/check-call.cpp (+2-1)
  • (added) flang/test/Semantics/cuf20.cuf (+42)
diff --git a/flang/lib/Semantics/check-call.cpp b/flang/lib/Semantics/check-call.cpp
index dfaa0e028d698..58271d7ca2e87 100644
--- a/flang/lib/Semantics/check-call.cpp
+++ b/flang/lib/Semantics/check-call.cpp
@@ -1016,7 +1016,8 @@ static void CheckExplicitDataArg(const characteristics::DummyDataObject &dummy,
       }
     }
     if (dummyDataAttr == common::CUDADataAttr::Device &&
-        (dummyIsAssumedShape || dummyIsAssumedRank)) {
+        (dummyIsAssumedShape || dummyIsAssumedRank)
+        !dummy.ignoreTKR.test(common::IgnoreTKR::Contiguous)) {
       if (auto contig{evaluate::IsContiguous(actual, foldingContext,
               /*namedConstantSectionsAreContiguous=*/true,
               /*firstDimensionStride1=*/true)}) {
diff --git a/flang/test/Semantics/cuf20.cuf b/flang/test/Semantics/cuf20.cuf
new file mode 100644
index 0000000000000..222ff2a1b7c6d
--- /dev/null
+++ b/flang/test/Semantics/cuf20.cuf
@@ -0,0 +1,42 @@
+! RUN: %python %S/test_errors.py %s %flang_fc1
+
+! Test case 1: Device arrays with ignore_tkr(c)
+subroutine test_device_arrays()
+  interface bar
+    subroutine bar1(a)
+!dir$ ignore_tkr(c) a
+      real :: a(..)
+!@cuf attributes(device) :: a
+    end subroutine
+  end interface
+
+  integer :: n = 10, k = 2
+  real, device :: a(10), b(10), c(10)
+  
+  call bar(a(1:n))     ! Should not warn about contiguity
+  call bar(b(1:n:k))   ! Should not warn about contiguity
+  call bar(c(1:n:2))   ! Should not warn about contiguity
+end subroutine
+
+! Test case 2: Managed arrays with ignore_tkr(c)
+subroutine test_managed_arrays()
+  interface bar
+    subroutine bar1(a)
+!dir$ ignore_tkr(c) a
+      real :: a(..)
+!@cuf attributes(device) :: a
+    end subroutine
+  end interface
+
+  integer :: n = 10, k = 2
+  real, managed :: a(10), b(10), c(10)
+  
+  call bar(a(1:n))     ! Should not warn about contiguity
+  call bar(b(1:n:k))   ! Should not warn about contiguity
+  call bar(c(1:n:2))   ! Should not warn about contiguity
+end subroutine
+
+program main
+  call test_device_arrays()
+  call test_managed_arrays()
+end program 
\ No newline at end of file

@github-actions
Copy link

github-actions bot commented May 6, 2025

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

}
if (dummyDataAttr == common::CUDADataAttr::Device &&
(dummyIsAssumedShape || dummyIsAssumedRank)) {
(dummyIsAssumedShape || dummyIsAssumedRank)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How does this code even compile?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Missed && in the source.

@wangzpgi wangzpgi requested a review from klausler May 6, 2025 21:18
Copy link
Contributor

@clementval clementval left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

program main
call test_device_arrays()
call test_managed_arrays()
end program No newline at end of file
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

New line at end of line

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Strangely I don't see new line after end program in the source code.

@wangzpgi wangzpgi merged commit ce69a60 into llvm:main May 7, 2025
6 of 9 checks passed
@wangzpgi wangzpgi deleted the ignoretkrc branch May 7, 2025 16:23
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

flang:semantics flang Flang issues not falling into any other category

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants