GN: add sanitize arg

GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack: Clang, GN, and Ninja
all track changes to xsan.blacklist without our help.

This has the incidental effect of upgrading the compiler used by *SAN
bots from Clang 3.8 to Clang 3.9.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002

Committed: https://skia.googlesource.com/skia/+/2b3c2a3ff9166e92ca1700e695c30ae729a3f10c

[mtklein_C, 2016-08-30 20:03:27]

Description was changed from

==========
GN: add sanitize arg

BUG=skia:
==========

to

==========
GN: add sanitize arg

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-08-30 20:06:16]

Description was changed from

==========
GN: add sanitize arg

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

MSAN will need more work (namely, compiling libc++ with MSAN) before it's
useful.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-08-30 20:48:03]

Description was changed from

==========
GN: add sanitize arg

MSAN will need more work (namely, compiling libc++ with MSAN) before it's
useful.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over TSAN and ASAN builds.
MSAN will need more work (namely, compiling libc++ with MSAN) before it's
useful.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-07 19:28:08]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over TSAN and ASAN builds.
MSAN will need more work (namely, compiling libc++ with MSAN) before it's
useful.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-07 20:53:05]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 13:58:23]

mtklein@chromium.org changed reviewers:
+ borenet@google.com, jcgregorio@google.com

[mtklein_C, 2016-09-08 13:58:25]

PS 29 has a bunch of passing trybot runs, and I'm running them again on the
rebase PS 30 just in case.

[mtklein_C, 2016-09-08 14:03:41]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:04:33]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:05:41]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:06:40]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:09:28]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

Unlike the old *SAN bots, this compiles in the compile step and tests/perfs
in the test/perf steps, which means we can get compile sharing now!  I've
seen it happen even.  The flip side of this is we need to do a little more
work to keep MSAN builds pointed at the right libc++.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots depend on the clang_linux package too.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:10:05]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now too.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots depend on the clang_linux package too.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots depend on the clang_linux package too.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:10:34]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots depend on the clang_linux package too.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:10:50]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack... Clang, GN, and Ninja
all seem to track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack: Clang, GN, and Ninja
all track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[mtklein_C, 2016-09-08 14:18:53]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack: Clang, GN, and Ninja
all track changes to xsan.blacklist without our help.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack: Clang, GN, and Ninja
all track changes to xsan.blacklist without our help.

This has the incidental effect of upgrading the compiler used by *SAN
bots from Clang 3.8 to Clang 3.9.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

[jcgregorio, 2016-09-08 14:46:54]

On 2016/09/08 at 13:58:25, mtklein wrote:
> PS 29 has a bunch of passing trybot runs, and I'm running them again on the
rebase PS 30 just in case.

LGTM for the GN changes. Deferring to Eric for the rest.

[borenet, 2016-09-08 14:47:33]

Looks mostly good but can't really speak to the GN config stuff.

https://codereview.chromium.org/2289343002/diff/570001/DEPS
File DEPS (right):

https://codereview.chromium.org/2289343002/diff/570001/DEPS#newcode56
DEPS:56: 'llvm': { }
Yay!

https://codereview.chromium.org/2289343002/diff/570001/infra/bots/recipe_modu...
File infra/bots/recipe_modules/flavor/gn_flavor.py (right):

https://codereview.chromium.org/2289343002/diff/570001/infra/bots/recipe_modu...
infra/bots/recipe_modules/flavor/gn_flavor.py:32: clang_linux =
str(self.m.vars.slave_dir.join('clang_linux'))
Why str() instead of using pathObj.join(...) ?

[mtklein, 2016-09-08 14:51:09]

mtklein@google.com changed reviewers:
+ mtklein@google.com

[mtklein, 2016-09-08 14:51:11]

https://codereview.chromium.org/2289343002/diff/570001/infra/bots/recipe_modu...
File infra/bots/recipe_modules/flavor/gn_flavor.py (right):

https://codereview.chromium.org/2289343002/diff/570001/infra/bots/recipe_modu...
infra/bots/recipe_modules/flavor/gn_flavor.py:32: clang_linux =
str(self.m.vars.slave_dir.join('clang_linux'))
On 2016/09/08 14:47:33, borenet wrote:
> Why str() instead of using pathObj.join(...) ?

Just seems easier to convert to string here once than to have to think about
whether things are Paths or strs every later time I need it as a str, e.g. '-L'
+ str(clang_linux.join('msan'))

[borenet, 2016-09-08 15:06:39]

Recipes LGTM

https://codereview.chromium.org/2289343002/diff/570001/infra/bots/recipe_modu...
File infra/bots/recipe_modules/flavor/gn_flavor.py (right):

https://codereview.chromium.org/2289343002/diff/570001/infra/bots/recipe_modu...
infra/bots/recipe_modules/flavor/gn_flavor.py:32: clang_linux =
str(self.m.vars.slave_dir.join('clang_linux'))
On 2016/09/08 14:51:10, mtklein wrote:
> On 2016/09/08 14:47:33, borenet wrote:
> > Why str() instead of using pathObj.join(...) ?
> 
> Just seems easier to convert to string here once than to have to think about
> whether things are Paths or strs every later time I need it as a str, e.g.
'-L'
> + str(clang_linux.join('msan'))

Ok, I'm not sure that str(pathObj) will resolve the path in the same way as
going through the recipe API, but if the trybots are successful then it must be
okay.

[mtklein_C, 2016-09-08 15:08:03]

> Ok, I'm not sure that str(pathObj) will resolve the path in the same way as
going through the recipe API, but if the trybots are successful then it must be
okay.

Yeah, it may help that these are all relative paths?

[mtklein_C, 2016-09-08 15:08:06]

The CQ bit was checked by mtklein@chromium.org

[commit-bot: I haz the power, 2016-09-08 15:08:09]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-08 15:39:36]

Description was changed from

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack: Clang, GN, and Ninja
all track changes to xsan.blacklist without our help.

This has the incidental effect of upgrading the compiler used by *SAN
bots from Clang 3.8 to Clang 3.9.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002
==========

to

==========
GN: add sanitize arg

Attempt to take over all *SAN builds.

MSAN has a lot of coordination required between gn/BUILD.gn and gn_flavor.py.
I'd like to follow up to move more of this into gn/BUILD.gn, to make it easier
to use locally.

The compile steps should be much faster now.  We no longer build CMake
and Clang for every run, instead using the clang_linux CIPD package.  This
removes the need for all the third_party/externals/llvm/... dependencies.

Similarly, since we're using the clang_linux package, we no longer depend
on Chrome's Clang, and thus no longer need to sync chromium on these bots.

Instead of packaging up MSAN libraries and llvm-symbolizer in the compile 
output, I have the test / perf bots also depend on the clang_linux package.
These do not vary from build to build.

No more need for the xsan.blacklist -include hack: Clang, GN, and Ninja
all track changes to xsan.blacklist without our help.

This has the incidental effect of upgrading the compiler used by *SAN
bots from Clang 3.8 to Clang 3.9.

BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2289343002

Committed:
https://skia.googlesource.com/skia/+/2b3c2a3ff9166e92ca1700e695c30ae729a3f10c
==========

[commit-bot: I haz the power, 2016-09-08 15:39:37]

Committed patchset #30 (id:570001) as
https://skia.googlesource.com/skia/+/2b3c2a3ff9166e92ca1700e695c30ae729a3f10c