Build subzero with LLVM autoconf build

Build subzero with LLVM autoconf build.

This adds the subzero repo to toolchain_build along with a
source rule to sync it. It updates LLVM revision to
get the Makefile changes which cause it to build.

No tests running yet, but llvm2ice is built with the rest of the LLVM tools.

R=jvoung@chromium.org,stichnot@chromium.org
NOTRY=true

Committed: http://src.chromium.org/viewvc/native_client?view=rev&revision=13569

[Jim Stichnoth, 2014-05-22 23:52:04]

lgtm

[Derek Schuff, 2014-05-23 00:26:03]

Not sure I'm going to have a chance to fix it today but the mac build fails
because its old gcc doesn't like the comma at the end of the enumerator lists

[Jim Stichnoth, 2014-05-23 14:07:03]

On 2014/05/23 00:26:03, Derek (OoO 5-22 to 6-12) wrote:
> Not sure I'm going to have a chance to fix it today but the mac build fails
> because its old gcc doesn't like the comma at the end of the enumerator lists

Hopefully a fix will soon be ready - https://codereview.chromium.org/296823013/
.

[Derek Schuff, 2014-06-13 21:35:57]

On 2014/05/23 14:07:03, stichnot wrote:
> On 2014/05/23 00:26:03, Derek (OoO 5-22 to 6-12) wrote:
> > Not sure I'm going to have a chance to fix it today but the mac build fails
> > because its old gcc doesn't like the comma at the end of the enumerator
lists
> 
> Hopefully a fix will soon be ready -
https://codereview.chromium.org/296823013/
> .

Looks like you landed this already, closing now.

[jvoung (off chromium), 2014-06-14 00:10:39]

Wait, I don't think this actually landed yet (w/ rebased COMPONENT_REVISIONS)? I
did patched it in, and used it to sync though.

[Derek Schuff, 2014-06-14 00:21:24]

On 2014/06/14 00:10:39, jvoung wrote:
> Wait, I don't think this actually landed yet (w/ rebased COMPONENT_REVISIONS)?
I
> did patched it in, and used it to sync though.

oops, you're right, my client was not in the state I expected. Reopened and
rebased.

[Derek Schuff, 2014-07-30 16:28:44]

rebased again. I have a CL out for the mac failure. Can one of you take a look
at the failure on Windows?

[Jim Stichnoth, 2014-07-30 16:47:50]

On 2014/07/30 16:28:44, Derek Schuff wrote:
> rebased again. I have a CL out for the mac failure. Can one of you take a look
> at the failure on Windows?

There are a few warnings about unused variables, all of which are "unused"
because of assert() calls.  I'll fix those to tidy things up.

Then there are link errors presumably because of ODR violations.  Although I
haven't yet recreated the errors locally, I have a patch that may fix it, maybe
I can just try it out in the meantime...

[Derek Schuff, 2014-07-30 17:23:41]

On 2014/07/30 16:47:50, stichnot wrote:
> On 2014/07/30 16:28:44, Derek Schuff wrote:
> > rebased again. I have a CL out for the mac failure. Can one of you take a
look
> > at the failure on Windows?
> 
> There are a few warnings about unused variables, all of which are "unused"
> because of assert() calls.  I'll fix those to tidy things up.
> 
> Then there are link errors presumably because of ODR violations.  Although I
> haven't yet recreated the errors locally, I have a patch that may fix it,
maybe
> I can just try it out in the meantime...

also this:
http://build.chromium.org/p/tryserver.nacl/builders/nacl-toolchain-mac-pnacl-...

[jvoung (off chromium), 2014-07-30 17:41:57]

On 2014/07/30 17:23:41, Derek Schuff wrote:
> On 2014/07/30 16:47:50, stichnot wrote:
> > On 2014/07/30 16:28:44, Derek Schuff wrote:
> > > rebased again. I have a CL out for the mac failure. Can one of you take a
> look
> > > at the failure on Windows?
> > 
> > There are a few warnings about unused variables, all of which are "unused"
> > because of assert() calls.  I'll fix those to tidy things up.
> > 
> > Then there are link errors presumably because of ODR violations.  Although I
> > haven't yet recreated the errors locally, I have a patch that may fix it,
> maybe
> > I can just try it out in the meantime...
> 
> also this:
>
http://build.chromium.org/p/tryserver.nacl/builders/nacl-toolchain-mac-pnacl-...

Might just be missing the dtor for InstX8632Lockable. I can try a speculative
fix, but my mac doesn't have gcc (only LLVM), so it doesn't repro.

[Derek Schuff, 2014-07-30 17:50:47]

On 2014/07/30 17:41:57, jvoung wrote:
> On 2014/07/30 17:23:41, Derek Schuff wrote:
> > On 2014/07/30 16:47:50, stichnot wrote:
> > > On 2014/07/30 16:28:44, Derek Schuff wrote:
> > > > rebased again. I have a CL out for the mac failure. Can one of you take
a
> > look
> > > > at the failure on Windows?
> > > 
> > > There are a few warnings about unused variables, all of which are "unused"
> > > because of assert() calls.  I'll fix those to tidy things up.
> > > 
> > > Then there are link errors presumably because of ODR violations.  Although
I
> > > haven't yet recreated the errors locally, I have a patch that may fix it,
> > maybe
> > > I can just try it out in the meantime...
> > 
> > also this:
> >
>
http://build.chromium.org/p/tryserver.nacl/builders/nacl-toolchain-mac-pnacl-...
> 
> Might just be missing the dtor for InstX8632Lockable. I can try a speculative
> fix, but my mac doesn't have gcc (only LLVM), so it doesn't repro.

Once I land the update to binutils 2.24 i plan on making the bots all build with
clang.

[Derek Schuff, 2014-07-31 00:14:47]

The CQ bit was checked by dschuff@chromium.org

[commit-bot: I haz the power, 2014-07-31 00:15:47]

CQ is trying da patch. Follow status at

https://nativeclient-status.appspot.com/cq/dschuff@chromium.org/296053008/140001

[commit-bot: I haz the power, 2014-07-31 15:23:25]

Change committed as 13569