I mark this patch as "plan changed" since the assert message indicates that something wrong in IRBuilder or Codegen, however, main branch also have the same issue so I think this patch does not trigger the issue.

Use irbuilder for masked and master construct: https://reviews.llvm.org/D100874.

Initially we did not have an OMPIRBuilder object unconditionally, now we have. Let's move over everything that is ready. So master and critical should be good to go as well I suppose.

In D100514#2693601, @jdoerfert wrote:

Also, don't forget to mark it as done in https://clang.llvm.org/docs/OpenMPSupport.html :)

In D100514#2693600, @jdoerfert wrote:

Any reason we should not unconditionally use the OMPIRBuilder impl? (btw, many thanks for providing one!)
We have an OMPIRBuilder always around in clang's codegen, so there is little reason not to use it if it is feature complete.

Update test and fix code format

Fix lit-test failure for dispatch_ast_print and rebase

Fix/Add test for "can not have more than one filter clause" issue and rebase.

Rebase and fix based on comments.

Fix comments

Rebase

Update codegen test

Add some codegen tests

LGTM.

I think some existing Clang OpenMP codegen tests will break if they are testing against __kmpc_push_target_tripcount.

@alokmishra.besu, I'm trying to play/learn with your patch so I downloaded the diff and apply it to the master branch, however, I'm getting error messages everywhere if llvm::omp::OMPD_metadirective appears:

Rebase

Fix outdated comment in test

LGTM.

ping

Below test asserts inside generateInfoForComponentList:

Assertion failed: (!IsPointer && "Unexpected base element with the pointer type."), function generateInfoForComponentList, file /Users/cchen/workspace/llvm-project/clang/lib/CodeGen/CGOpenMPRuntime.cpp, line 7754.
PLEASE submit a bug report to https://bugs.llvm.org/ and include the crash backtrace, preprocessed source, and associated run script.

Just tried this patch and found that below test abort inside generateInfoCapture.

#include <stdio.h>
#define LEN 100
int buf[LEN];

Add more tests and store location for present modifier

Update based on comment

I now only update target_defaultmap_messages test since I'm not sure what I've done is correct for the new version. Will add all other defaultmap_messages tests once corrected the stuff I added in target_defaultmap_messages.

Rebase and fix coding style

I think I'll just land it since it's an nfc.

Refactor target_data_update

In D82245#2404499, @tianshilei1992 wrote:

Would you please take a thorough check with your patch? I can still find many code that still don't conform with the standard. Thanks. The test case is fine as we don't have unified standard.

Update style

Apply LLVM coding style

In D82245#2386544, @tianshilei1992 wrote:

In D82245#2383337, @cchen wrote:

Fix coding style

I'm not sure whether this is still WIP because I didn't see any change in the touched code.

Fix coding style

Fix build error

Rebase

In D82245#2377081, @grokos wrote:

The link in the description for the clang patch is outdated (that patch has been abandoned), can you replace it with https://reviews.llvm.org/D84192? Thanks!

Rebase and fix based on feedback

Refactor

Fix coding style

Fix based on feedback

In D90704#2372303, @ABataev wrote:

It would be good if you could identify the object which leads to a crash, I mean a target region, variable, etc.

In D90704#2371938, @jdoerfert wrote:

I do not understand the commit message. Can you try to make it clearer?

Separate test to a independent file

In D74144#2307494, @ABataev wrote:

In D74144#2307454, @cchen wrote:

@ABataev, the below test is extracted from Sollve test suite and Clang now emit:

test.c:17:35: error: subscripted value is not an array or pointer
    #pragma omp target update to( (([N][N])foo)[1:M] )
                                  ^~~~~~~~~~~~~
test.c:17:5: error: expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'
    #pragma omp target update to( (([N][N])foo)[1:M] )

This error message came from the ActOnOMPArraySectionExpr which is called inside ParsePostfixExpressionSuffix. The issue is that the base expression in ActOnOMPArraySectionExpr looks like:

ParenExpr 0x122859be0 '<OpenMP array shaping type>' lvalue
`-OMPArrayShapingExpr 0x122859b98 '<OpenMP array shaping type>' lvalue
  |-IntegerLiteral 0x122859b38 'int' 5
  |-IntegerLiteral 0x122859b58 'int' 5
  `-DeclRefExpr 0x122859b78 'int *' lvalue Var 0x1228599d0 'foo' 'int *'

which is not a base that we would expect in an array section expr. I've tried relaxing the base type check in ActOnOMPArraySectionExpr but not sure it's the way to go. (or should I just extract the DeclReExpr from ArrayShapingExpr before calling ActOnOMPArraySectionExpr?)

#define N 5
#define M 3

int main(void) {
    int tmp[N][N];
    for(int i=0; i<N; i++)
        for(int j=0; j<N; j++)
            tmp[i][j] = N*i + j;

    int *foo = &tmp[0][0];

    // This compiles just fine
    //#pragma omp target update to( ([N][N])foo )

    // This is rejected by the compiler
    #pragma omp target update to( (([N][N])foo)[1:M] )
}

I don't think it is allowed by the standard.

According to the standard, The shape-operator can appear only in clauses where it is explicitly allowed.
In this case, array shaping is used as a base expression of array section (or subscript) expression, which does not meet the standard. Tje array sjaping operation is not used in clause, instead it is used as a base subexpression of another expression.

@ABataev, the below test is extracted from Sollve test suite and Clang now emit:

test.c:17:35: error: subscripted value is not an array or pointer
    #pragma omp target update to( (([N][N])foo)[1:M] )
                                  ^~~~~~~~~~~~~
test.c:17:5: error: expected at least one 'to' clause or 'from' clause specified to '#pragma omp target update'
    #pragma omp target update to( (([N][N])foo)[1:M] )

This error message came from the ActOnOMPArraySectionExpr which is called inside ParsePostfixExpressionSuffix. The issue is that the base expression in ActOnOMPArraySectionExpr looks like:

ParenExpr 0x122859be0 '<OpenMP array shaping type>' lvalue
`-OMPArrayShapingExpr 0x122859b98 '<OpenMP array shaping type>' lvalue
  |-IntegerLiteral 0x122859b38 'int' 5
  |-IntegerLiteral 0x122859b58 'int' 5
  `-DeclRefExpr 0x122859b78 'int *' lvalue Var 0x1228599d0 'foo' 'int *'

which is not a base that we would expect in an array section expr. I've tried relaxing the base type check in ActOnOMPArraySectionExpr but not sure it's the way to go. (or should I just extract the DeclReExpr from ArrayShapingExpr before calling ActOnOMPArraySectionExpr?)

LGTM

In PR45212 comment 5 (https://bugs.llvm.org/show_bug.cgi?id=45212#c5), there is a reduced test case that failed after adding this patch. The assertion is from OMPChildren::getInnermostCapturedStmt.

Fix commit message

ping

Ping, compiler support is in this patch: https://reviews.llvm.org/D84192.

@ABataev, do I need to wait for the runtime patch to commit this one? If so, do you have some recommend reviewers for me to add to that patch? I have pinged several times for that patch but haven't got many reviews for it. Thanks!