[PATCH] D123928: [NVPTX][tests] Do not run the tests which are not supported by nvptx

[Artem Belevich via Phabricator via llvm-commits]

tra added inline comments.


================
Comment at: llvm/test/CodeGen/Generic/2007-12-17-InvokeAsm.ll:4
+; NVPTX does not support memory address operands
+; UNSUPPORTED: nvptx
+
----------------
This should probably be "REQUIRE: x86" instead, as it uses inline assembly with x86-specific constraints.


================
Comment at: llvm/test/CodeGen/Generic/2010-ZeroSizedArg.ll:4
 
+; UNSUPPORTED: nvptx
+
----------------
What exactly is the problem with this case -- the datatype with 0-element fields or vararg function printf?
If that's the latter, I think  `printf` could be removed as it's unused by the test.


================
Comment at: llvm/test/CodeGen/Generic/APIntLoadStore.ll:3
+
+; NVPTX failed to emit extern global var of type size > 64 (i65 and so on)
+; UxNSUPPORTED: nvptx
----------------
This kind of tests should probably be cloned into NVPTX-specific variants with appropriate REQUIRES tag. We do want to make sure that we do not regress on the types NVPTX does support.


================
Comment at: llvm/test/CodeGen/Generic/empty-load-store.ll:4
 
+; UNSUPPORTED: nvptx
+
----------------
What happens in this case? AFAICT, llc appears to work with this IR: https://godbolt.org/z/n1eK193br



================
Comment at: llvm/test/CodeGen/Generic/pr12507.ll:7
 
 define void @foo(i160 %x) {
 entry:
----------------
This test crashes LLVM, probably due to the use of i160. https://godbolt.org/z/W6s4n4vM7
We need at least a comment about that.


================
Comment at: llvm/test/CodeGen/Generic/pr24662.ll:4
 
+; UNSUPPORTED: nvptx
+
----------------
Another crasher: https://godbolt.org/z/cYPebKasG, probably due to `i670010`.  Please add a comment.




================
Comment at: llvm/test/CodeGen/Generic/stacksave-restore.ll:3
 
+; UNSUPPORTED: nvptx
+
----------------
Crasher: https://godbolt.org/z/67583fq9q

This time due to use of `@llvm.stacksave` and `@llvm.stackrestore`.


================
Comment at: llvm/test/CodeGen/Generic/zero-sized-array.ll:4
 
+; UNSUPPORTED: nvptx
+
----------------
This one compiles to PTX: https://godbolt.org/z/c3sE14j46
What's the reason for disabling it? Something to do with `[0 x i8]` ?

I suspect ptxas would be unhappy about `.param .align 1 .b8 f4_param_0[0]`.


================
Comment at: llvm/test/CodeGen/MLRegalloc/default-eviction-advisor.ll:11-15
 define void @f2(i64 %lhs, i64 %rhs, i64* %addr) {
   %sum = add i64 %lhs, %rhs
   store i64 %sum, i64* %addr
   ret void
 }
----------------
Again, it would be good to explain what's going on here. Considering that we don't see the message regardless of what `-regalloc-enable-advisor` is set to, I guess it's just not enalbed for NVPTX.


================
Comment at: llvm/test/Feature/optnone-llc.ll:9
 ; REQUIRES: asserts, default_triple
+; UNSUPPORTED: nvptx
 
----------------
This test makes implicit assumptions about machine-specific pass pipeline. It might be better to either trim the list of checks to the passes that are universal for all back-ends or restrict the test only to the back-end which does match these assumptions.

Excluding NVPTX is OK, but my guess is that this test may currently be failing on other 'exotic' back-ends.


================
Comment at: llvm/test/MC/AsmParser/include.ll:2
 ; RUN: llc -I %p/Inputs -filetype asm -o - %s | FileCheck %s
-; UNSUPPORTED: -zos
+; UNSUPPORTED: -zos,nvptx
 ; REQUIRES: default_triple
----------------
My guess is that the test is not universally applicable, as not all back-ends may support `.include` or `.macro` even if they do have support for parsing assembly.
OK to mark NVPTX unsupported, but the test may be problematic for other back-ends, too.


================
Comment at: llvm/test/MC/AsmParser/macro-same-context.ll:3
 ;; thus a definition is available to the whole file. PR36110
-; UNSUPPORTED: -zos
+; UNSUPPORTED: -zos,nvptx
 ; RUN: not llc < %s 2>&1 | FileCheck %s
----------------
Ditto.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D123928/new/

https://reviews.llvm.org/D123928