Speculative fix to VS 2015 code-gen bug

Speculative workaround for VS 2015 code-gen bug

On PGO builds (but no others) a pair of functions in
process_power_collector.cc get compiled such that the callee returns a
floating-point value in xmm0 and the caller looks for it in st(0). The
details are in the bug, but if floating-point exceptions are enabled
(perhaps happening due to third-party injected software changing the FPU
state) then this can lead to crashes. Even without the crashes it is
clearly wrong.

Since the bug only reproes in PGO it is difficult to test a fix but
disabling optimizations around the two functions *should* do the trick.

BUG=640588
Committed: https://crrev.com/764c7362bc01b0f1765f2f9043ca89b71a8d0c76
Cr-Commit-Position: refs/heads/master@{#417994}

[brucedawson, 2016-09-10 00:04:36]

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

[brucedawson, 2016-09-10 00:05:20]

Description was changed from

==========
Speculative fix to VS 2015 code-gen bug

On PGO builds (but no others) a pair of functions in
process_power_collector.cc get compiled such that the callee returns a
floating-point value in xmm0 and the caller looks for it in st(0). The
details are in the bug, but if floating-point exceptions are enabled
(perhaps happening due to third-party injected software changing the FPU
state) then this can lead to crashes. Even without the crashes it is
clearly wrong.

Since the bug only reproes in PGO it is difficult to test a fix but
disabling optimizations around the two functions *should* do the trick.

BUG=640588
==========

to

==========
Speculative workaround for VS 2015 code-gen bug

On PGO builds (but no others) a pair of functions in
process_power_collector.cc get compiled such that the callee returns a
floating-point value in xmm0 and the caller looks for it in st(0). The
details are in the bug, but if floating-point exceptions are enabled
(perhaps happening due to third-party injected software changing the FPU
state) then this can lead to crashes. Even without the crashes it is
clearly wrong.

Since the bug only reproes in PGO it is difficult to test a fix but
disabling optimizations around the two functions *should* do the trick.

BUG=640588
==========

[commit-bot: I haz the power, 2016-09-10 00:05:30]

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

[brucedawson, 2016-09-10 00:08:12]

brucedawson@chromium.org changed reviewers:
+ sivachandra@chromium.org

[brucedawson, 2016-09-10 00:08:12]

Speculative fix for a compiler bug - PTAL?

I can't guarantee it fixes things, but it shouldn't make things worse, since
it's just turning off optimizations for two non-performance-critical functions.

[commit-bot: I haz the power, 2016-09-10 00:19:02]

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

[commit-bot: I haz the power, 2016-09-10 00:19:03]

Dry run: Try jobs failed on following builders:
  chromeos_amd64-generic_chromium_compile_only_ng on
master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromeos_amd64-...)

[brucedawson, 2016-09-10 00:33:00]

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

[commit-bot: I haz the power, 2016-09-10 00:33:28]

Dry run: 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, 2016-09-10 01:23:29]

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

[commit-bot: I haz the power, 2016-09-10 01:23:30]

Dry run: This issue passed the CQ dry run.

[brucedawson, 2016-09-12 17:57:58]

brucedawson@chromium.org changed reviewers:
+ derat@chromium.org, dhnishi@chromium.org

[brucedawson, 2016-09-12 17:57:59]

Adding other owners...

Does this look like a reasonable speculative fix for this code-gen bug? I'd like
to get this committed so I can see in tomorrow's canary if the bug is resolved.
Local testing is impractical.

[Daniel Nishi, 2016-09-12 18:14:07]

On 2016/09/12 17:57:59, brucedawson wrote:
> Adding other owners...
> 
> Does this look like a reasonable speculative fix for this code-gen bug? I'd
like
> to get this committed so I can see in tomorrow's canary if the bug is
resolved.
> Local testing is impractical.

Forewarning: I'm no longer on Chrome and haven't touched this code since I was
an intern.

This lgtm.

That said, doing some quick code searches, I'm not sure if these code path is
needed anymore. It was originally part of my intern project to revamp content
settings for web apps on desktop, but that effort got folded into tbuckley@'s
Settings redesign it appears. I haven't dug deeply into the mocks, but I'm not
sure if we're surfacing the info out of the collector anymore.

[brucedawson, 2016-09-12 18:21:00]

The CQ bit was checked by brucedawson@chromium.org

[commit-bot: I haz the power, 2016-09-12 18:21:40]

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

[brucedawson, 2016-09-12 18:23:12]

On 2016/09/12 18:14:07, Daniel Nishi wrote:
> On 2016/09/12 17:57:59, brucedawson wrote:
> > Adding other owners...
> > 
> > Does this look like a reasonable speculative fix for this code-gen bug? I'd
> like
> > to get this committed so I can see in tomorrow's canary if the bug is
> resolved.
> > Local testing is impractical.
> 
> Forewarning: I'm no longer on Chrome and haven't touched this code since I was
> an intern.
> 
> This lgtm.
> 
> That said, doing some quick code searches, I'm not sure if these code path is
> needed anymore. It was originally part of my intern project to revamp content
> settings for web apps on desktop, but that effort got folded into tbuckley@'s
> Settings redesign it appears. I haven't dug deeply into the mocks, but I'm not
> sure if we're surfacing the info out of the collector anymore.

Thanks Daniel. I'll land this change and see if it resolves the code-gen bug
because that's a valuable piece of information. The result from
UpdatePowerConsumption() is being thrown away (which may be part of what
confuses the compiler) but I couldn't tell if the code was having some greater
effect.

[commit-bot: I haz the power, 2016-09-12 19:10:40]

Description was changed from

==========
Speculative workaround for VS 2015 code-gen bug

On PGO builds (but no others) a pair of functions in
process_power_collector.cc get compiled such that the callee returns a
floating-point value in xmm0 and the caller looks for it in st(0). The
details are in the bug, but if floating-point exceptions are enabled
(perhaps happening due to third-party injected software changing the FPU
state) then this can lead to crashes. Even without the crashes it is
clearly wrong.

Since the bug only reproes in PGO it is difficult to test a fix but
disabling optimizations around the two functions *should* do the trick.

BUG=640588
==========

to

==========
Speculative workaround for VS 2015 code-gen bug

On PGO builds (but no others) a pair of functions in
process_power_collector.cc get compiled such that the callee returns a
floating-point value in xmm0 and the caller looks for it in st(0). The
details are in the bug, but if floating-point exceptions are enabled
(perhaps happening due to third-party injected software changing the FPU
state) then this can lead to crashes. Even without the crashes it is
clearly wrong.

Since the bug only reproes in PGO it is difficult to test a fix but
disabling optimizations around the two functions *should* do the trick.

BUG=640588
==========

[commit-bot: I haz the power, 2016-09-12 19:10:45]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-09-12 19:13:46]

Description was changed from

==========
Speculative workaround for VS 2015 code-gen bug

On PGO builds (but no others) a pair of functions in
process_power_collector.cc get compiled such that the callee returns a
floating-point value in xmm0 and the caller looks for it in st(0). The
details are in the bug, but if floating-point exceptions are enabled
(perhaps happening due to third-party injected software changing the FPU
state) then this can lead to crashes. Even without the crashes it is
clearly wrong.

Since the bug only reproes in PGO it is difficult to test a fix but
disabling optimizations around the two functions *should* do the trick.

BUG=640588
==========

to

==========
Speculative workaround for VS 2015 code-gen bug

On PGO builds (but no others) a pair of functions in
process_power_collector.cc get compiled such that the callee returns a
floating-point value in xmm0 and the caller looks for it in st(0). The
details are in the bug, but if floating-point exceptions are enabled
(perhaps happening due to third-party injected software changing the FPU
state) then this can lead to crashes. Even without the crashes it is
clearly wrong.

Since the bug only reproes in PGO it is difficult to test a fix but
disabling optimizations around the two functions *should* do the trick.

BUG=640588

Committed: https://crrev.com/764c7362bc01b0f1765f2f9043ca89b71a8d0c76
Cr-Commit-Position: refs/heads/master@{#417994}
==========

[commit-bot: I haz the power, 2016-09-12 19:13:47]

Patchset 2 (id:??) landed as
https://crrev.com/764c7362bc01b0f1765f2f9043ca89b71a8d0c76
Cr-Commit-Position: refs/heads/master@{#417994}