[PATCH] D28797: [LangRef] Make @llvm.sqrt(x) return undef, rather than have UB, for negative x.

[Justin Lebar via Phabricator via llvm-commits]

jlebar created this revision.
Herald added a subscriber: wdng.

Some frontends emit a speculate-and-select idiom for sqrt, wherein they compute
sqrt(x), check if x is negative, and select NaN if it is:

  %cmp = fcmp olt double %a, -0.000000e+00
  %sqrt = call double @llvm.sqrt.f64(double %a)
  %ret = select i1 %cmp, double 0x7FF8000000000000, double %sqrt

This is technically UB as the LangRef is written today if %a is ever less than
-0.  But emitting code that's compliant with the current definition of sqrt
would require a branch, which would then prevent us from matching this idiom in
SelectionDAG (which we do today -- ISD::FSQRT has defined behavior on negative
inputs), because SelectionDAG looks at one BB at a time.

Nothing in LLVM takes advantage of this undefined behavior, as far as we can
tell, and the fact that llvm.sqrt has UB dates from its initial addition to the
LangRef.


https://reviews.llvm.org/D28797

Files:
  llvm/docs/LangRef.rst


Index: llvm/docs/LangRef.rst
===================================================================
--- llvm/docs/LangRef.rst
+++ llvm/docs/LangRef.rst
@@ -10061,7 +10061,7 @@
 
 The '``llvm.sqrt``' intrinsics return the sqrt of the specified operand,
 returning the same value as the libm '``sqrt``' functions would. Unlike
-``sqrt`` in libm, however, ``llvm.sqrt`` has undefined behavior for
+``sqrt`` in libm, however, ``llvm.sqrt`` returns undefined for
 negative numbers other than -0.0 (which allows for better optimization,
 because there is no need to worry about errno being set).
 ``llvm.sqrt(-0.0)`` is defined to return -0.0 like IEEE sqrt.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D28797.84640.patch
Type: text/x-patch
Size: 650 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170117/1dd8670f/attachment.bin>