[llvm-commits] Patch to safely restore optimization to testcase from PR3225 on x86

[Dan Gohman]

On Dec 18, 2008, at 11:42 PM, Eli Friedman wrote:

> On Thu, Dec 18, 2008 at 5:31 PM, Dan Gohman <gohman at apple.com> wrote:
>>> +  if (KnownZero == VTBits) {
>>
>> If the shift count is known to be all-zero, shouldn't it
>> have been simplified already by pre-legalize DAGCombine,
>> if not InstCombine?
>
> The masked bits check is only checking the top bit, i.e. whether the
> shift amount is < 32 or >= 32.

Thanks, I misread that.  The patch looks ok to me, assuming you can
add a testcase to go along with it :-).  Thanks!

Dan