[llvm] [Loads] Support dereferenceable assumption with variable size. (PR #128436)

[LLVM Continuous Integration via llvm-commits]

llvm-ci wrote:

LLVM Buildbot has detected a new failure on builder `llvm-clang-key-instructions` running on `sie-linux-worker5` while building `llvm` at step 6 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/208/builds/2759

<details>
<summary>Here is the relevant piece of the build log for the reference</summary>

```
Step 6 (test-build-unified-tree-check-all) failure: test (failure)
******************** TEST 'LLVM :: Transforms/LoopVectorize/vect.stats.ll' FAILED ********************
Exit Code: 1

Command Output (stderr):
--
/home/buildbot/buildbot-root/llvm-ki/build/bin/opt < /home/buildbot/buildbot-root/llvm-ki/llvm-project/llvm/test/Transforms/LoopVectorize/vect.stats.ll -passes=loop-vectorize -force-vector-interleave=4 -force-vector-width=4 -debug-only=loop-vectorize -enable-early-exit-vectorization --disable-output -stats -S 2>&1 | /home/buildbot/buildbot-root/llvm-ki/build/bin/FileCheck /home/buildbot/buildbot-root/llvm-ki/llvm-project/llvm/test/Transforms/LoopVectorize/vect.stats.ll # RUN: at line 1
+ /home/buildbot/buildbot-root/llvm-ki/build/bin/opt -passes=loop-vectorize -force-vector-interleave=4 -force-vector-width=4 -debug-only=loop-vectorize -enable-early-exit-vectorization --disable-output -stats -S
+ /home/buildbot/buildbot-root/llvm-ki/build/bin/FileCheck /home/buildbot/buildbot-root/llvm-ki/llvm-project/llvm/test/Transforms/LoopVectorize/vect.stats.ll
/home/buildbot/buildbot-root/llvm-ki/llvm-project/llvm/test/Transforms/LoopVectorize/vect.stats.ll:8:10: error: CHECK: expected string not found in input
; CHECK: 1 loop-vectorize - Number of early exit loops vectorized
         ^
<stdin>:260:62: note: scanning from here
3 loop-vectorize - Number of loops analyzed for vectorization
                                                             ^
<stdin>:261:1: note: possible intended match here
1 loop-vectorize - Number of loops vectorized
^

Input file: <stdin>
Check file: /home/buildbot/buildbot-root/llvm-ki/llvm-project/llvm/test/Transforms/LoopVectorize/vect.stats.ll

-dump-input=help explains the following input dump.

Input was:
<<<<<<
           1:  
           2: LV: Checking a loop in 'vectorized' from <stdin> 
           3: LV: Loop hints: force=? width=4 interleave=4 
           4: LV: Found a loop: for.body 
           5: LV: Found an induction variable. 
           6: LV: Found FP op with unsafe algebra. 
           7: LV: We can vectorize this loop! 
           8: LV: Loop does not require scalar epilogue 
           9: LV: Found trip count: 0 
          10: LV: The max safe fixed VF is: 67108864. 
          11: LV: The max safe scalable VF is: vscale x 0. 
          12: LV: Found uniform instruction: %cmp2 = icmp sgt i64 %indvars.iv.next, %size 
          13: LV: Found uniform instruction: %arrayidx = getelementptr inbounds float, ptr %a, i64 %indvars.iv2 
          14: LV: Found uniform instruction: %indvars.iv2 = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %for.body.preheader ] 
          15: LV: Found uniform instruction: %indvars.iv.next = add nuw nsw i64 %indvars.iv2, 1 
          16: LV: Found scalar instruction: %indvars.iv2 = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %for.body.preheader ] 
          17: LV: Found scalar instruction: %indvars.iv.next = add nuw nsw i64 %indvars.iv2, 1 
          18: LV: Found an estimated cost of 1 for VF 4 For instruction: %indvars.iv2 = phi i64 [ %indvars.iv.next, %for.body ], [ 0, %for.body.preheader ] 
          19: LV: Found an estimated cost of 0 for VF 4 For instruction: %arrayidx = getelementptr inbounds float, ptr %a, i64 %indvars.iv2 
          20: LV: Found an estimated cost of 1 for VF 4 For instruction: %0 = load float, ptr %arrayidx, align 4 
          21: LV: Found an estimated cost of 1 for VF 4 For instruction: %mul = fmul float %0, %0 
          22: LV: Found an estimated cost of 1 for VF 4 For instruction: store float %mul, ptr %arrayidx, align 4 
          23: LV: Found an estimated cost of 1 for VF 4 For instruction: %indvars.iv.next = add nuw nsw i64 %indvars.iv2, 1 
          24: LV: Found an estimated cost of 1 for VF 4 For instruction: %cmp2 = icmp sgt i64 %indvars.iv.next, %size 
          25: LV: Found an estimated cost of 1 for VF 4 For instruction: br i1 %cmp2, label %for.end.loopexit, label %for.body 
...

```

</details>

https://github.com/llvm/llvm-project/pull/128436