<div dir="ltr">Here's an updated version, jvoung pointed out that the change for ARM and Thumb long in fast-isel-intrinsic.ll wasn't quite right. The other files are unchanged.</div><div class="gmail_extra"><br><br>
<div class="gmail_quote">On Mon, May 13, 2013 at 3:34 PM, JF Bastien <span dir="ltr"><<a href="mailto:jfb@google.com" target="_blank">jfb@google.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div dir="ltr">Right, the reserved register vary with other triples, so my change is setting up the tests so that they'll work on other triples too (specifically, armv7-linux-gnueabi). I only added regex to registers when allocation differs with other triples.</div>
<div class="HOEnZb"><div class="h5">
<div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, May 13, 2013 at 3:29 PM, Eric Christopher <span dir="ltr"><<a href="mailto:echristo@gmail.com" target="_blank">echristo@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Actually, wait, reserved register differences? Shouldn't be different<br>
with the same triple, etc or is this just setting up for adding more<br>
triples? :)<br>
<span><font color="#888888"><br>
-eric<br>
</font></span><div><div><br>
On Mon, May 13, 2013 at 3:26 PM, Eric Christopher <<a href="mailto:echristo@gmail.com" target="_blank">echristo@gmail.com</a>> wrote:<br>
> LGTM.<br>
><br>
> -eric<br>
><br>
> On Mon, May 13, 2013 at 3:23 PM, JF Bastien <<a href="mailto:jfb@google.com" target="_blank">jfb@google.com</a>> wrote:<br>
>> ARM FastISel is currently only enabled for iOS non-Thumb1, and I'm working<br>
>> on enabling it for other targets. As a first step I've fixed some of the<br>
>> tests.<br>
>><br>
>> Changes to ARM FastISel tests:<br>
>> - Different triples don't generate the same relocations (especially<br>
>> movw/movt versus constant pool loads). Use a regex to allow either.<br>
>> - Mangling is different. Use a regex to allow either.<br>
>> - The reserved registers are sometimes different, so registers get allocated<br>
>> in a different order. Capture the names only where this occurs.<br>
>> - Add -verify-machineinstrs to some tests where it works. It doesn't work<br>
>> everywhere it should yet.<br>
>> - Add -fast-isel-abort to many tests that didn't have it before.<br>
>> - Split out the VarArg test from fast-isel-call.ll into its own test. This<br>
>> simplifies test setup because of --check-prefix.<br>
>><br>
>> I'm mostly done with code changes to enable ARM FastISel and will upload<br>
>> separate patches for this. I still have an issue outstanding with lencod,<br>
>> which I'm lead to understand might be tricky.<br>
</div></div></blockquote></div><br></div>
</div></div></blockquote></div><br></div>