[llvm] r228168 - Fixes a bug in vector load legalization that confused bits and bytes.

[Kuperstein, Michael M]

I'd say yes.

I don't think there's a risk - the previous version of this would either crash or produce completely incorrect code in the cases where the BitOffset >= WideBits evaluated to true, and nothing else is affected.

Michael

-----Original Message-----
From: hwennborg at google.com [mailto:hwennborg at google.com] On Behalf Of Hans Wennborg
Sent: Wednesday, February 04, 2015 23:05
To: Kuperstein, Michael M
Cc: llvm-commits
Subject: Re: [llvm] r228168 - Fixes a bug in vector load legalization that confused bits and bytes.

On Wed, Feb 4, 2015 at 10:54 AM, Michael Kuperstein <michael.m.kuperstein at intel.com> wrote:
> Author: mkuper
> Date: Wed Feb  4 12:54:01 2015
> New Revision: 228168
>
> URL: http://llvm.org/viewvc/llvm-project?rev=228168&view=rev
> Log:
> Fixes a bug in vector load legalization that confused bits and bytes.
>
> Differential Revision: http://reviews.llvm.org/D7400

PR20417 reproduces on the 3.6 branch. Should we merge this fix?
---------------------------------------------------------------------
Intel Israel (74) Limited

This e-mail and any attachments may contain confidential material for
the sole use of the intended recipient(s). Any review or distribution
by others is strictly prohibited. If you are not the intended
recipient, please contact the sender and delete all copies.