Update third_party/binutils to 2.26, add ICF fix

Update third_party/binutils to 2.26, add ICF fix

The new binutils 2.26 release incorporates two out of the three
fixes that we were applying to the 2.25 version.  There is also
an upstream fix, that's not in 2.26, for a bug that made the
linker crash when using --icf for x86-32, which we now apply
locally so it should become safe to use gold with --icf for x86.

BUG=610908
R=thakis@chromium.org

Committed: https://crrev.com/2a1fcea64a6130271e68cddb71fb71f71080c3e2
Cr-Commit-Position: refs/heads/master@{#392828}

[Roland McGrath, 2016-05-05 21:21:37]

I'm not clear on how I should "check that I am happy with the binaries" I built
locally.  I think I'd need to copy them into the Linux_x64 directory and then do
a non-goma build to be testing them.  Is that right?

[Nico, 2016-05-05 21:31:43]

That sounds right to me.

You'll also have to push the new binutils to goma before landing.

I've never built new binutils myself -- I'm cc'ing pcc who did this not too long
ago. He probably knows how exactly this should be done.

https://codereview.chromium.org/1950343004/diff/1/third_party/binutils/README...
File third_party/binutils/README.chromium (right):

https://codereview.chromium.org/1950343004/diff/1/third_party/binutils/README...
third_party/binutils/README.chromium:13: Ubuntu Trusty.
I wonder if this will be a problem -- does this change required glibc /
libstdc++ dependencies? Will this binary still be able to run on trusty bots?
(You'll know this better than me.)

[Roland McGrath, 2016-05-05 21:46:37]

I'll attempt that testing regime, but wait for pcc's advice.

https://codereview.chromium.org/1950343004/diff/1/third_party/binutils/README...
File third_party/binutils/README.chromium (right):

https://codereview.chromium.org/1950343004/diff/1/third_party/binutils/README...
third_party/binutils/README.chromium:13: Ubuntu Trusty.
On 2016/05/05 21:31:43, Nico wrote:
> I wonder if this will be a problem -- does this change required glibc /
> libstdc++ dependencies? Will this binary still be able to run on trusty bots?
> (You'll know this better than me.)

It would if the builds weren't in chroots that are still Precise, as the next
paragraph still correctly says.  I just changed this one not to be a lie
(probably was a lie before, if the last update was on someone's workstation :-),
not because the fact is actually relevant to anything--that's why we're using
chroots.

[pcc1, 2016-05-05 21:48:53]

pcc@chromium.org changed reviewers:
+ pcc@chromium.org

[pcc1, 2016-05-05 21:48:53]

> I've never built new binutils myself -- I'm cc'ing pcc who did this not too
long
> ago. He probably knows how exactly this should be done.

There's an upload.sh script in the same directory that you'd need to run. That
script also mentions the names of people you'd need to contact to update goma.

When I first ran it, I didn't have the right permissions -- please contact me
offline if that's an issue.

https://codereview.chromium.org/1950343004/diff/1/third_party/binutils/README...
File third_party/binutils/README.chromium (right):

https://codereview.chromium.org/1950343004/diff/1/third_party/binutils/README...
third_party/binutils/README.chromium:13: Ubuntu Trusty.
On 2016/05/05 21:31:43, Nico wrote:
> I wonder if this will be a problem -- does this change required glibc /
> libstdc++ dependencies? Will this binary still be able to run on trusty bots?
> (You'll know this better than me.)

I wouldn't have thought it would. The actual build takes place inside a precise
chroot (see the setup done by build-all.sh). It doesn't look like this patch
changes how the chroot is set up, so I think this comment can be left unchanged.

[Roland McGrath, 2016-05-06 17:32:35]

The CQ bit was checked by mcgrathr@chromium.org to run a CQ dry run

[Roland McGrath, 2016-05-06 17:32:49]

Verified ready in GOMA.  PTAL.

[commit-bot: I haz the power, 2016-05-06 17:32:59]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1950343004/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1950343004/40001

[commit-bot: I haz the power, 2016-05-06 18:51:39]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-05-06 18:51:40]

Dry run: This issue passed the CQ dry run.

[Nico, 2016-05-11 01:04:29]

lgtm

[Nico, 2016-05-11 01:04:38]

Description was changed from

==========
Update third_party/binutils to 2.26, add ICF fix

The new binutils 2.26 release incorporates two out of the three
fixes that we were applying to the 2.25 version.  There is also
an upstream fix, that's not in 2.26, for a bug that made the
linker crash when using --icf for x86-32, which we now apply
locally so it should become safe to use gold with --icf for x86.

BUG= none
R=thakis@chromium.org
==========

to

==========
Update third_party/binutils to 2.26, add ICF fix

The new binutils 2.26 release incorporates two out of the three
fixes that we were applying to the 2.25 version.  There is also
an upstream fix, that's not in 2.26, for a bug that made the
linker crash when using --icf for x86-32, which we now apply
locally so it should become safe to use gold with --icf for x86.

BUG=610908
R=thakis@chromium.org
==========

[Dirk Pranke, 2016-05-11 01:08:58]

The CQ bit was checked by dpranke@chromium.org

[commit-bot: I haz the power, 2016-05-11 01:09:31]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1950343004/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1950343004/40001

[commit-bot: I haz the power, 2016-05-11 02:22:12]

Description was changed from

==========
Update third_party/binutils to 2.26, add ICF fix

The new binutils 2.26 release incorporates two out of the three
fixes that we were applying to the 2.25 version.  There is also
an upstream fix, that's not in 2.26, for a bug that made the
linker crash when using --icf for x86-32, which we now apply
locally so it should become safe to use gold with --icf for x86.

BUG=610908
R=thakis@chromium.org
==========

to

==========
Update third_party/binutils to 2.26, add ICF fix

The new binutils 2.26 release incorporates two out of the three
fixes that we were applying to the 2.25 version.  There is also
an upstream fix, that's not in 2.26, for a bug that made the
linker crash when using --icf for x86-32, which we now apply
locally so it should become safe to use gold with --icf for x86.

BUG=610908
R=thakis@chromium.org
==========

[commit-bot: I haz the power, 2016-05-11 02:22:13]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-05-11 02:23:25]

Description was changed from

==========
Update third_party/binutils to 2.26, add ICF fix

The new binutils 2.26 release incorporates two out of the three
fixes that we were applying to the 2.25 version.  There is also
an upstream fix, that's not in 2.26, for a bug that made the
linker crash when using --icf for x86-32, which we now apply
locally so it should become safe to use gold with --icf for x86.

BUG=610908
R=thakis@chromium.org
==========

to

==========
Update third_party/binutils to 2.26, add ICF fix

The new binutils 2.26 release incorporates two out of the three
fixes that we were applying to the 2.25 version.  There is also
an upstream fix, that's not in 2.26, for a bug that made the
linker crash when using --icf for x86-32, which we now apply
locally so it should become safe to use gold with --icf for x86.

BUG=610908
R=thakis@chromium.org

Committed: https://crrev.com/2a1fcea64a6130271e68cddb71fb71f71080c3e2
Cr-Commit-Position: refs/heads/master@{#392828}
==========

[commit-bot: I haz the power, 2016-05-11 02:23:26]

Patchset 3 (id:??) landed as
https://crrev.com/2a1fcea64a6130271e68cddb71fb71f71080c3e2
Cr-Commit-Position: refs/heads/master@{#392828}