-
Notifications
You must be signed in to change notification settings - Fork 116
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make tests compatible with upstream llvm #2865
Conversation
Allow the new disjoint flag on `or`s, which was recently added upstream.
Test summary for commit 133b7a5CTS tests (Failed: 0/138443)
Ubuntu navi3x, SrdcvkUbuntu navi2x, Srdcvk |
Test summary for commit 2099fa8CTS tests (Failed: 0/138378)
Ubuntu navi3x, SrdcvkUbuntu navi2x, Srdcvk |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
These changes look ok
Some of the tests checks are auto-generated
Maybe that's ok though - once someone re-generates the checks there's a strong possibility that the llvm versions will be up-to-date.
@@ -21,7 +21,7 @@ void main() | |||
// CHECK-NEXT: [[TMP0:%.*]] = call i64 @llvm.amdgcn.s.getpc() | |||
// CHECK-NEXT: [[TMP1:%.*]] = and i64 [[TMP0]], -4294967296 | |||
// CHECK-NEXT: [[TMP2:%.*]] = zext i32 [[USERDATA0]] to i64 | |||
// CHECK-NEXT: [[TMP3:%.*]] = or i64 [[TMP1]], [[TMP2]] | |||
// CHECK-NEXT: [[TMP3:%.*]] = or {{(disjoint )?}}i64 [[TMP1]], [[TMP2]] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This test has auto-generated checks - presumably the optional disjoint will get changed if that's run?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, once llvm is updated and someone regenerates the test, these will be made permanent :)
@@ -85,7 +85,7 @@ attribute[0].offset = 0 | |||
; SHADERTEST-NEXT: [[DOTI1:%.*]] = fptosi float [[TMP19]] to i32 | |||
; SHADERTEST-NEXT: [[TMP12:%.*]] = and i64 [[TMP11]], -4294967296 | |||
; SHADERTEST-NEXT: [[TMP13:%.*]] = zext i32 [[DESCTABLE0:%.*]] to i64 | |||
; SHADERTEST-NEXT: [[TMP14:%.*]] = or i64 [[TMP12]], [[TMP13]] | |||
; SHADERTEST-NEXT: [[TMP14:%.*]] = or {{(disjoint )?}}i64 [[TMP12]], [[TMP13]] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Also auto-generated
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok - let's go with this approach with the tests.
I suppose this might actually be better than updating and disabling the test - and then re-enabling at some point when all the changes are propagated.
Allow the new disjoint flag on
or
s, which was recently added upstream.