[PATCH][LoopStrengthReduce] Don't trim a formula that uses a subset of required registers

Chandler Carruth chandlerc at google.com
Mon Apr 28 16:10:36 PDT 2014


On Mon, Apr 28, 2014 at 3:55 PM, Chandler Carruth <chandlerc at google.com>wrote:

>
> On Fri, Apr 25, 2014 at 1:39 PM, Andrew Trick <atrick at apple.com> wrote:
>
>> Adam,
>>
>> This fix looks great to me! Please go ahead and commit.
>>
>
> Andy, how does this look good? ;]
>
> --- /dev/null
> +++ b/test/Transforms/LoopStrengthReduce/ARM64/req-regs.c
> @@ -0,0 +1,36 @@
> +// RUN: clang %s -O3 -target arm64-apple-ios -o - -S -mllvm
> -debug-only=loop-reduce 2>&1| FileCheck %s
> +// REQUIRES: asserts
> +
> +// LSR used to fail here due to a bug in the ReqRegs test.  To complicate
> +// things, this could only be reproduced with clang because the uses would
> +// come out in different order when invoked through llc.
> +
> +// CHECK: The chosen solution requires
> +// CHECK-NOT: No Satisfactory Solution
>
> This test runs clang, from the users PATH, with -O3, in an LLVM test. This
> doesn't work *at all*.
>
> That really doesn't work. Mind reverting until you have a reduced test
> case? This is blocking us, so I'd like to get it fixed ASAP.
>

FYI, caught Andy on IRC and he said it would be fine to just revert for
now. Sorry for the trouble. Happy to follow up here as needed.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20140428/3e9c8a21/attachment.html>


More information about the llvm-commits mailing list