[PATCH] D155307: [InstCombine] Allow KnownBits to be propagated

[Nikita Popov via Phabricator via llvm-commits]

nikic added inline comments.


================
Comment at: llvm/test/Transforms/InstCombine/2023-07-13-arm-infiniteloop.ll:62
+  ret void
+}
+
----------------
nikic wrote:
> pmatos wrote:
> > nikic wrote:
> > > pmatos wrote:
> > > > nikic wrote:
> > > > > Please minimize the test -- I don't see how this second function could possibly be related for example.
> > > > I did run llvm-reduce... I wonder if llvm-reduce is not removing  functions. I will take a look.
> > > llvm-reduce should be removing functions...
> > hummm, you're right it does. Which is why in this case the test cannot be reduced further and there's no infinite loop if I manually remove the function bpf_prog_calc_tag. However, I cannot see straightforwardly why that is.
> I don't get an infinite loop with your current test either -- I think your current test might be against something like `-O2` rather than `-passes=instcombine`?
> 
> Here's a test that works against just instcombine: https://gist.github.com/nikic/a2c0b936686940ae31ae2b23ec9efbcd (Not reduced)
I wasn't able to reduce it substantially beyond that, but here's a cleaned up test case:
```
declare i1 @llvm.is.constant.i32(i32)

define i32 @test(i32 %arg) {
entry:
  %zext = zext i32 %arg to i64
  %lshr1 = lshr i64 %zext, 32
  %trunc = trunc i64 %lshr1 to i32
  %cond = call i1 @llvm.is.constant.i32(i32 %trunc)
  br i1 %cond, label %else, label %if

if:
  br label %exit

else:
  %fshl = call i32 @llvm.fshl.i32(i32 %trunc, i32 0, i32 0)
  br label %exit

exit:
  %phi = phi i32 [ %fshl, %else ], [ 0, %if ]
  %lshr2 = lshr i32 %phi, 1
  ret i32 %lshr2
}

declare i32 @llvm.fshl.i32(i32, i32, i32)
```


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D155307/new/

https://reviews.llvm.org/D155307