Add Chrome OS's minigbm library to third_party

Add Chrome OS's minigbm library to third_party

This library is used by Chrome on Chrome OS to allocate accelerated
graphics buffers. The version avaiable on Ubuntu is too old, and as a
result we can't run unit tests for the gbm platform code on developer
workstations or on the waterfall.

This adds the library to third_party and fixes up the build files to use
either the OS copy (for Chrome OS release builds) or gardened copy (for
testing on the waterfall) depending on the value of "use_system_minigbm".

BUG=530712
TEST=
  GYP_DEFINES="use_ozone=1 chromeos=1 ozone_platform_gbm=1 use_system_minigbm=0" gclient runhooks
  ninja -C out/Release ozone_unittests
  ./out/Release/ozone_unittests

  gn gen out_gn_ozone --args='target_os="chromeos" use_ozone=true ozone_platform_gbm=true use_system_minigbm=false'
  ninja -C out_gn_ozone ozone_unittests
  ./out_gn_ozone/Release/ozone_unittests

Committed: https://crrev.com/c231bffd37754a1ab09fbbb4a1260438be638df4
Cr-Commit-Position: refs/heads/master@{#348758}

[spang, 2015-09-11 20:14:24]

spang@chromium.org changed reviewers:
+ brettw@chromium.org

[brettw, 2015-09-11 20:56:36]

You should send a note to chrome-eng-review as described on
http://www.chromium.org/developers/adding-3rd-party-libraries Since this is
small and Google code, this should be straightforward.

I'm confused about the use_system_minigbm value. It defaults to true with the
comment seeming to indicate that it will fail on workstations if this isn't set
to false. I think Developers should expect to check out a build on their
workstation and have it work. Can you clarify this?

[spang, 2015-09-11 21:03:26]

On 2015/09/11 20:56:36, brettw wrote:
> You should send a note to chrome-eng-review as described on
> http://www.chromium.org/developers/adding-3rd-party-libraries Since this is
> small and Google code, this should be straightforward.

Ok.

> 
> I'm confused about the use_system_minigbm value. It defaults to true with the
> comment seeming to indicate that it will fail on workstations if this isn't
set
> to false. I think Developers should expect to check out a build on their
> workstation and have it work. Can you clarify this?

You're absolutely correct it needs to be set to false to get the desired benefit
on the waterfall & on eng workstations. However in order to do that we need to
turn it "on" explicitly in the CrOS build, but I haven't done that yet.

I'm happy to make the CrOS change first, then we can set it to false right away.

[brettw, 2015-09-11 21:08:13]

I don't entirely follow, but I just want all developers to always have working
builds without setting any flags.

[spang, 2015-09-11 21:35:23]

On 2015/09/11 21:08:13, brettw wrote:
> I don't entirely follow, but I just want all developers to always have working
> builds without setting any flags.

I'm saying I don't want to change the way CrOS release builds work. They use the
system version.

I'll check a change into CrOS to set GYP_DEFINES="use_system_minigbm=1", and
after that the default can be "0" so it works out of the box on Ubuntu.

Developer won't have to specify use_system_minigbm on their own - definitely
agree that would suck. The CrOS SDK will do it for them when doing a build that
targets CrOS hardware.

[spang, 2015-09-14 17:00:34]

On 2015/09/11 21:35:23, spang wrote:
> On 2015/09/11 21:08:13, brettw wrote:
> > I don't entirely follow, but I just want all developers to always have
working
> > builds without setting any flags.
> 
> I'm saying I don't want to change the way CrOS release builds work. They use
the
> system version.
> 
> I'll check a change into CrOS to set GYP_DEFINES="use_system_minigbm=1", and
> after that the default can be "0" so it works out of the box on Ubuntu.
> 
> Developer won't have to specify use_system_minigbm on their own - definitely
> agree that would suck. The CrOS SDK will do it for them when doing a build
that
> targets CrOS hardware.

I've updated the default to not use the system version.

I'll land https://chromium-review.googlesource.com/#/c/299690/ first so it won't
affect CrOS release builds.

[brettw, 2015-09-14 17:13:15]

LGTM if you land it in a way that nobody breaks :)

[spang, 2015-09-14 21:32:27]

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

[commit-bot: I haz the power, 2015-09-14 21:36:05]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1341473003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1341473003/20001

[spang, 2015-09-14 22:25:10]

The CQ bit was unchecked by spang@chromium.org

[spang, 2015-09-14 22:25:10]

The CQ bit was checked by spang@chromium.org

[commit-bot: I haz the power, 2015-09-14 22:26:57]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1341473003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1341473003/20001

[commit-bot: I haz the power, 2015-09-14 23:17:05]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-09-14 23:17:59]

Patchset 2 (id:??) landed as
https://crrev.com/c231bffd37754a1ab09fbbb4a1260438be638df4
Cr-Commit-Position: refs/heads/master@{#348758}

[commit-bot: I haz the power, 2015-09-23 12:38:44]

Patchset 2 (id:??) landed as
https://crrev.com/c231bffd37754a1ab09fbbb4a1260438be638df4
Cr-Commit-Position: refs/heads/master@{#348758}