Rollback Scavenge implemetation and rely on existing functionality to free...

Rollback Scavenge implemetation and rely on existing functionality to free

This is a landing of a patch provided by antonm.  See:
http://codereview.chromium.org/235022

Also included change to browser_about_handler.cc to fix build, and I set
TCMALLOC_RELEASE_RATE to 1.0 on line 40 of page_heap.cc (I think this
was an inadvertent rollback element).

r=antonm

[jar (doing other things), 2009-10-01 00:28:14]

I landed this, but gcl did not properly close it.  The message read:

Committed revision 27692.
Loaded authentication cookies from c:\Users\jar\.codereview_upload_cookies
Error accessing url /257009/close

I'll manually close this CL shortly.

[Anton Muhin, 2009-10-01 12:58:25]

Thanks a lot, Jim!

yours,
anton.

On Thu, Oct 1, 2009 at 4:28 AM,  <jar@chromium.org> wrote:
> I landed this, but gcl did not properly close it. =A0The message read:
>
> Committed revision 27692.
> Loaded authentication cookies from c:\Users\jar\.codereview_upload_cookie=
s
> Error accessing url /257009/close
>
> I'll manually close this CL shortly.
>
> http://codereview.chromium.org/257009
>

[Anton Muhin, 2009-10-01 13:05:52]

Jim,

setting rate to 0 was intentional---that disable IncrementalScavenge.

yours,
anton.

On Thu, Oct 1, 2009 at 1:41 AM,  <jar@chromium.org> wrote:
> Reviewers: antonm, Mike Belshe,
>
> Description:
> Rollback Scavenge implemetation and rely on existing functionality to fre=
e
>
> This is a landing of a patch provided by antonm. =A0See:
> http://codereview.chromium.org/235022
>
> Also included change to browser_about_handler.cc to fix build, and I set
> TCMALLOC_RELEASE_RATE to 1.0 on line 40 of page_heap.cc (I think this
> was an inadvertent rollback element).
>
> r=3Dantonm
>
> Please review this at http://codereview.chromium.org/257009
>
> SVN Base: svn://chrome-svn/chrome/trunk/src/
>
> Affected files:
> =A0M =A0 =A0 chrome/browser/browser_about_handler.cc
> =A0M =A0 =A0 chrome/renderer/render_thread.cc
> =A0M =A0 =A0 third_party/tcmalloc/config_linux.h
> =A0M =A0 =A0 third_party/tcmalloc/config_win.h
> =A0D =A0 =A0 third_party/tcmalloc/google/malloc_extension.h
> =A0D =A0 =A0 third_party/tcmalloc/malloc_extension.cc
> =A0M =A0 =A0 third_party/tcmalloc/page_heap.h
> =A0M =A0 =A0 third_party/tcmalloc/page_heap.cc
> =A0M =A0 =A0 third_party/tcmalloc/tcmalloc.cc
> =A0M =A0 =A0 third_party/tcmalloc/tcmalloc.gyp
>
>
>

[jar (doing other things), 2009-10-01 18:20:41]

The critical point was that without incremental scavenging, we'd currently
have no way to release memory in the browser process.  Having this
incremental scavenge gave us at least a chance to hammer down our browser
usage over time.
Was there some other way that the browser would have moved off its
high-water mark without this setting?

Jim

On Thu, Oct 1, 2009 at 6:05 AM, Anton Muhin <antonm@google.com> wrote:

> Jim,
>
> setting rate to 0 was intentional---that disable IncrementalScavenge.
>
> yours,
> anton.
>
> On Thu, Oct 1, 2009 at 1:41 AM,  <jar@chromium.org> wrote:
> > Reviewers: antonm, Mike Belshe,
> >
> > Description:
> > Rollback Scavenge implemetation and rely on existing functionality to
> free
> >
> > This is a landing of a patch provided by antonm.  See:
> > http://codereview.chromium.org/235022
> >
> > Also included change to browser_about_handler.cc to fix build, and I set
> > TCMALLOC_RELEASE_RATE to 1.0 on line 40 of page_heap.cc (I think this
> > was an inadvertent rollback element).
> >
> > r=antonm
> >
> > Please review this at http://codereview.chromium.org/257009
> >
> > SVN Base: svn://chrome-svn/chrome/trunk/src/
> >
> > Affected files:
> >  M     chrome/browser/browser_about_handler.cc
> >  M     chrome/renderer/render_thread.cc
> >  M     third_party/tcmalloc/config_linux.h
> >  M     third_party/tcmalloc/config_win.h
> >  D     third_party/tcmalloc/google/malloc_extension.h
> >  D     third_party/tcmalloc/malloc_extension.cc
> >  M     third_party/tcmalloc/page_heap.h
> >  M     third_party/tcmalloc/page_heap.cc
> >  M     third_party/tcmalloc/tcmalloc.cc
> >  M     third_party/tcmalloc/tcmalloc.gyp
> >
> >
> >
>

[Anton Muhin, 2009-10-01 18:24:19]

Not in the active tab (currently), but in the hidden yes.  Was
planning to invoke release for active tab as well, but now it's
probably unnecessary.

yours,
anton.

On Thu, Oct 1, 2009 at 10:20 PM, Jim Roskind <jar@chromium.org> wrote:
> The critical point was that without incremental scavenging, we'd currentl=
y
> have no way to release memory in the browser process. =A0Having this
> incremental scavenge gave us at least a chance to hammer down our browser
> usage over time.
> Was there some other way that the browser would have moved off its
> high-water mark without this setting?
>
> Jim
>
> On Thu, Oct 1, 2009 at 6:05 AM, Anton Muhin <antonm@google.com> wrote:
>>
>> Jim,
>>
>> setting rate to 0 was intentional---that disable IncrementalScavenge.
>>
>> yours,
>> anton.
>>
>> On Thu, Oct 1, 2009 at 1:41 AM, =A0<jar@chromium.org> wrote:
>> > Reviewers: antonm, Mike Belshe,
>> >
>> > Description:
>> > Rollback Scavenge implemetation and rely on existing functionality to
>> > free
>> >
>> > This is a landing of a patch provided by antonm. =A0See:
>> > http://codereview.chromium.org/235022
>> >
>> > Also included change to browser_about_handler.cc to fix build, and I s=
et
>> > TCMALLOC_RELEASE_RATE to 1.0 on line 40 of page_heap.cc (I think this
>> > was an inadvertent rollback element).
>> >
>> > r=3Dantonm
>> >
>> > Please review this at http://codereview.chromium.org/257009
>> >
>> > SVN Base: svn://chrome-svn/chrome/trunk/src/
>> >
>> > Affected files:
>> > =A0M =A0 =A0 chrome/browser/browser_about_handler.cc
>> > =A0M =A0 =A0 chrome/renderer/render_thread.cc
>> > =A0M =A0 =A0 third_party/tcmalloc/config_linux.h
>> > =A0M =A0 =A0 third_party/tcmalloc/config_win.h
>> > =A0D =A0 =A0 third_party/tcmalloc/google/malloc_extension.h
>> > =A0D =A0 =A0 third_party/tcmalloc/malloc_extension.cc
>> > =A0M =A0 =A0 third_party/tcmalloc/page_heap.h
>> > =A0M =A0 =A0 third_party/tcmalloc/page_heap.cc
>> > =A0M =A0 =A0 third_party/tcmalloc/tcmalloc.cc
>> > =A0M =A0 =A0 third_party/tcmalloc/tcmalloc.gyp
>> >
>> >
>> >
>
>