Can confirm this bug, any attempt to open Options will crash Chrome on Windows XP x64

I have Windows XP 64-Bit and this works fine for me...

I'm on XPx64 as well.  it crashes on "Options", "Clear browsing Data", "Import
bookmarks and settings", and "Report bug or broken website".  basically anything that
pops-up a dialog box.  "About" is okay.

During install, I had Firefox up, and it asked me to close Firefox so it can import
stuff.  I said Skip.  The first Chrome browser that came up after install, I couldn't
go anywhere (ie. www.google.com).  After I restart the Chrome browser, it works.

oem23.inf:ATI.Mfg.NTamd64:ati2mtag_M54:8.501.0.0:pci\ven_1002&dev_7145

ATI Radeon X1400 Insprion 9400 Vista x64

Works fine here.

This happens to me too for XP64sp2.  I have full DEP enabled (i.e., not just for
Windows services) if that matters.

XP x64

Just tried changing pretty much every option, no issues here.

If it helps, here's the undecoded backtrace I get from a debugger:

>	00000000()	
 	03812544()	
 	chrome.dll!02b81adc() 	
 	[Frames below may be incorrect and/or missing, no symbols loaded for chrome.dll]	
 	chrome.dll!02843b06() 	
 	chrome.dll!02b5edad() 	
 	chrome.dll!02b612ed() 	
 	chrome.dll!0280f99e() 	
 	user32.dll!_InternalCallWinProc@20()  + 0x28 bytes	

Maybe it's trying to invoke a null function pointer?

3 of the 4 xp64 machines I have access to are crashing like this

I'm a Windows XP x64 user and it crashes just like everyone else has said.
Its when I click on something on the interface to 'submit' or do something. Eg. Add 
bookmark.

The import feature crashes on click as well as the report bug.
I was searching where I could report it, as the report bug button didn't work.

Also would like to confirm this crash in XP x64

also confirmed (to the point i started a dupe bug report by accident)

Yup, totally crashes for me as well, on the above mentioned screens/submenus, every 
single time.

I also can confirm this issue. I also have the problem of using the star/bookmark 
function.

I also tried to reinstall and see if it would help the issues. Unfortunately, the 
problems still exists.

I can confirm this bug too. Crashes when immediately when opening "Options".

I confirm this very annoying bug too! It crashes immediately when clicking "Options".

Latest build from http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/ seems
to work fine.

I confirm this. I got Win Pro 64bit fully patched, Asus A8N-E mobo, nvidia video 
card, Logitech software for keyboard/mouse, Avast AV.

Yes! Chrome is crashing in XP-64bit version when you click options.

confirmed, and build 1675 works fine. (havent tried earlier builds)

Confirmed, Running MS XP-64bit Fully patched. 

Crashes when:
- I attempt to Open options
- Import from firefox 

the latest build from http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/ 
fixed the problem for me too.

I can confirm this bug too, in windows XP x64

definitely grab one of the newer builds. I no longer have this problem.

i can confirm the same errors discribed here and the newest build 1694 these issues
apear to be resolved.

this bug is weired. chrome works like it should on my home win64 instalation. with my 
wifes acount on the same computer it crashes,...just like here in office on win64

I have this same crash for "Options". Windows XP Pro x64. I used the bug report with 
the google-desktop analyzer text dump in the submission.

I can confirm it too... When you click on OPTIONS it sometimes pop-up the windows and 
in like 100 ms or something like that close the google chrome and popup another 
window saying that google chrome crashed restart now~~~ bla bla bla!

I have two account on my home 64-bit XP, it crashes on one account and does not 
crash on the other.

i have the same bug. downloaded last build. if some ppl have a bug on one account and 
not on another, should it mean its related to some configuration?

i have also cfosspeed running on my pc, i just mention that coz i've had some bugs 
with other programs, that were confirmed to be related to cfoss.

(No comment was entered for this change.)

I can verify that this problem occurs on my Windows XP Professional x64 SP2 system
with the latest Chrome release (build 1978, 0.2.149.29); however, the problem is not
present in Chromium build 1824 (0.2.152.0).

Confirmed

Hi folks, I am trying to reproduce the crash with the latest public build on a 64-bit 
WinXP machine. I can't get it to crash on that machine. Any specific repro steps 
would be helpful.

The reproduction steps are:
1. Select "Options" from the menu.

That's it.  Either it'll work for you or it'll crash.  Seems like it's dependent on
something in the environment, but there's not enough information to tell exactly what
that is.

What is there to do here?  It seems like it's already been fixed in the nightly
chromium builds.

I confirm the bug too on 0.2.149.29 (1798) ...

But on version 0.2.152.0 (Developer Build 1974) it works..

version 0.2.149.29 crashes on my AMD 64 bit and XP pro 64 when i attempt to open 
options

Version 0.3.154.0 of Chrome still crashes on my AMD x64 box, but the nightly build of 
Chromium doesn't.

0.2.149.30 crashes for me opening the Options dialog or clicking on the star next to 
the address bar to create a favorite. Running Athlon64 XP Pro x64 SP2.

This is the same as <b/1363680>

Nick tried 149.30 on a XP64 VM and had no luck reproducing it. I wonder if there is 
something else on the systems that trigger the bug?

Based on the comments above, it looks like this is fixed in the Dev channel.

Still crashing XP x64 here with 0.3.154.6, both opening options and trying to star an 
address. Not fixed.

I never tested Chromium, or any dev-channel releases. However, I just installed 0.3.154.9 and the crash is fixed.

Weird.  0.3.154.9 still crashes for me even though Chromium works fine.

0.3.154.9 crashes for me as well.

For those still crashing... What were your install or upgrade steps?

I went to http://www.google.com/chrome/index.html , downloaded the Google Chrome
installer, and ran it.  I didn't explicitly uninstall any earlier versions.

Then I ran Google Chrome and tried to access the Options dialog, and it crashed.

And I have verified that the About dialog reports that it's 0.3.154.9.

Same with me. Weirdest thing is when I installed the new version that came out last
week it worked ok, I was able to import FF favorites and so. And then the next day
when I tried opening the options menu... bam crash ! Has been doing it ever since.

Also on XP x64 with an Athlon X2, 2 Gb RAM.

Reopening in light of recent reports.

With some more research reveals that this is a bad virtual function call. The top of 
the stack lists SkCanvas::internalRestore which is incorrect, this function is never 
called from View::GetInsets. The rest of the stack, which looks valid, is:

ChromeViews::View::GetInsets()  Line 512
ChromeViews::Label::GetPreferredSize(WTL::CSize * out=)  Line 66
ChromeViews::GridLayout::GetPreferredSize(ChromeViews::View * host=0x03d64ca0, 
WTL::CSize * out=0x002de050)  Line 795
ChromeViews::View::GetPreferredSize(WTL::CSize * out=0x002de050)  Line 200
ChromeViews::ColumnSet::CalculateSize()  Line 582
ChromeViews::GridLayout::SizeRowsAndColumns(bool layout=false, int width=0, int 
height=0, WTL::CSize * pref=0x002de180)  Line 818
ChromeViews::GridLayout::GetPreferredSize(ChromeViews::View * host=0x03d64f70, 
WTL::CSize * out=0x002de180)  Line 795
ChromeViews::View::GetPreferredSize(WTL::CSize * out=0x002de180)  Line 200
ChromeViews::ColumnSet::CalculateSize()  Line 582
ChromeViews::GridLayout::SizeRowsAndColumns(bool layout=true, int width=498, int 
height=460, WTL::CSize * pref=0x002de200)  Line 818
ChromeViews::GridLayout::Layout(ChromeViews::View * host=0x03d5d410)  Line 771
ChromeViews::View::Layout()  Line 248
GeneralPageView::Layout()  Line 735
ChromeViews::View::Layout()  Line 255 + 0x7 bytes
ChromeViews::TabbedPane::DoSelectTabAt(int index=0)  Line 214
ChromeViews::TabbedPane::SelectTabAt(int index=0)  Line 132
OptionsWindowView::ShowOptionsPage(OptionsPage page=OPTIONS_PAGE_GENERAL, 
OptionsGroup highlight_group=OPTIONS_GROUP_NONE)  Line 124
ShowOptionsWindow(OptionsPage page=OPTIONS_PAGE_DEFAULT, OptionsGroup 
highlight_group=OPTIONS_GROUP_NONE, Profile * profile=0x0230f8e0)  Line 228

In r5328 I checked in a patch to change Label::GetPreferredSize which seems to be 
involved in this. The existing use of a 0x0 canvas was a little suspicious. We can go 
faster by just using the desktop DC to measure text in this way, and that also avoids 
the 0x0 bitmap.

r5328 was merged to the release branch @ r5399. It will be in 154.22 and later 
releases.

Unless we get better data, we'll not be doing more work here for 1.0.

I have version 0.3.154.9 and I and getting the crash as well, I know you guys are 
working on it. It there any other way to set some of the options without going to the 
options screen that causes the crash for me? I want to be able to set chrome to be my 
default browser, anyway of doing this without the options page?

One way to do that is to reinstall, because setting Chrome as the default browser is 
one of the options in the wizard before Chrome is launched after the installation.

Or just delete the "First Run" from the Chrome Application folder and then start 
Chrome.

It should be identical to running the first time after an install.

I am still experiencing this, so it doesn't seem fixed in 0.4.154.22.  At least it
properly restores tabs after the crash now.  (It wasn't restoring tabs after these
crashes before, for me.)

Also crashing on attempt to add bookmark (clicking the star) or clicking the "Report
bug or broken website..." item, as other people have been mentioning.  (Manage
Bookmarks and About both seem to work okay, though.)

aysz88: could you follow the steps at http://google.com/support/chrome/bin/answer.py?
answer=107788 and attach the crash report to this bug?

Thanks!

I haven't been able to get it to work, it only writes a checkpoint file with the text:
GBP1
20081120
1

(I'm not sure if this is related, but I noticed that the crash counts in Local State
-> "user experience metrics" -> "stability" were all zero.)

I tried enabling debug logging on the command line, but it doesn't seem to give much
either:
*startup*
[4656:16136:242281531:INFO:install_util.cc(64)] No existing Chrome install found.
[4656:16136:242281609:INFO:bookmark_model.cc(126)] Loading bookmarks
[4656:16136:242281656:INFO:bookmark_model.cc(454)] Loaded bookmarks, file_exists=1
from_history=0
[4656:16136:242281687:INFO:history.cc(648)] History backend finished loading
*click wrench -> Options*
[4656:16136:242284515:INFO:metrics_log.cc(106)] Metrics: Hash numeric
[ShowOptions]=[7316488890418352152]
*crash*

I might break out a copy of Visual Studio and try to dig up something that way....

aysz88: If you can use Visual Studio, it would be awesome! You can find me on IRC 
(brettx) if you have questions.

0.4.154.23 still shows the bug, absolutely no progress here...

And guess what... also with 0.4.154.25 !

Hmm... that is unfortunate.

There are a few other ways to generate a dump file. It is best to turn _off_ "send crashes and 
statistics to Google" and restart Chrome for these three approaches to work...

1) Make Chrome crash again and see if you get a file (with a timestamp that corresponds to the 
crash) under C:\Documents and Settings\All Users\Application Data\Microsoft\Dr Watson\user.dmp. 
The location might be different, you can verify by going to Start \ Run, and typing in 
drwtsn32.exe. The location can be found in the "Crash Dump" text field. If this doesn't work, you 
can try either of these methods:

2) If you attach to the process that crashed with Visual Studio (assuming you have that 
installed) and select a menu item that appears under Debug called "Save dump as". You can then 
send us this dump file.

3) If you don't have Visual Studio, you can download userdump.exe from Microsoft 
(http://www.microsoft.com/downloads/details.aspx?FamilyID=E089CA41-6A87-40C8-BF69-
28AC08570B7E&displaylang=en&displaylang=en) and extract it (by default it goes to something 
like C:\kktools\userdump8.1). Then make Chrome crash, but don't dismiss the crash dialog. Instead 
run...
  C:\kktools\userdump8.1\x86\userdump.exe chrome.exe
... which will create dump files for all running Chrome processes (including the one that 
crashed).

Feel free to send these files to one of the Chromium guys (whose emails appear in the comments 
for this bug).

I followed the directions in step 3 of #65.  I had a bunch of tabs open I didn't want 
to lose, so I opened a new window, observed the new pid, and then crashed the app by 
clicking on the bookmark star.  I also get crashes on "report bug" and "under the 
hood" in options.

I uploaded the dump of the new window here: http://www.zshare.net/download/5204982145cc2283/

OK, so :
1. No dump file there
2. No VS
3. Much worse : I installed that userdump.exe thing, and now everytime I try to lauch
chrome it CRASHES THE WHOLE SYSTEM ! Yes, BSOD ! Which NEVER happens under XP X64 !

Now I can't even lauch Chrome anymore !

Ok, so I uninstalled this userdump s**t and now Chrome is lauching again. But the
newer 0.4.154.29 version still shows the bug...

javlopez, thanks for your diligence. Unfortunately the crash you uploaded isn't very 
interesting. Here's the steps you need to do:

. Get Chrome to crash so that you see the Dr Watson dialog appear.
. Run userdump.exe with chrome as an option.

This should generate a bunch of a dump files. If you could zip them up and upload 
them that would be great!

Thanks!

julx64, the download page for user dump is confusing. When you run the downloaded 
file it creates a setup.exe, there is no need to run this. Instead follow the steps I 
outline in 69 and upload the dump files some where.

Thanks!

sky, the userdump executable generates over 1GB of data (due the number of tabs I 
have open) so I don't really want to figure out how to send them all. 

I did manage to get a user.dmp from watson today.  Is this useful?

user.dmp 75.0 KB   Download

I also got this.

drwtsn32.log 255 KB   Download

I don't think we need any more dumps. The stacks are always similar (computing table 
sizes for the dialog or popup). It's not clear from the stack what the problem is.

I suspect this is happening because of some other software or some configuration on 
the system we don't like, since we run great on many XP64 systems.

We really just need somebody that is experencing this error to get the source and do 
some quick debugging. If you are such a person, I can help walk you through things.

Someone asked about an antivirus, well I have Avast, just in case this may help.

javlopez,

Frustratingly the dump isn't very helpful:( Any chance you could turning on page heap 
and try again? You can download page heap here: http://www.microsoft.com/whdc/devtools/debugging/default.mspx .

To turn on page heap for Chrome do:

Gflags.exe –p /enable chrome.exe /full

Then try and crash it again. When done be sure and turn off page heap, otherwise 
Chrome will be quite slow:

Gflags.exe –p /disable chrome.exe /full

Thanks again for your help!

Also I'd be interested if temporarily disabling your antivirus affects this problem.

@brettw, I'm a java developer without Visual Studio.  Does that make me such a 
person?  I dl'ed the debugging tools for x64 and can open the process in windbg, but 
not really sure where to go from there.  Also I stopped all Symantec services but 
it's not affecting crashes.

@sky, no problem enabling page heap and crashing the app, but what output are you 
looking for?  The dump in watson?  I can only get it to appear every once in a while 
(even though it crashes every time).

javlopez, yes, I'm looking for a dump just like you uploaded in comment 71.

76 : No, no change when I disable Avast.

@sky, here is the latest dump.  Let me know if there's anything I can do with windbg.

user.dmp 70.6 KB   Download

drwtsn32.log 245 KB   Download

javlopez: From your setup, the easiest first thing to try is to get a full dump from 
you. The mini dumps only have stack information, so it's harder to tell what's going 
on.

If you can run Chrome in windbg with a clean profile (so you don't get any history in 
the dump file, and also keep it smaller) like so:
  windbg chrome.exe --user-data-dir=c:\temp_profile
Then do the thing that crashes. In windbg, do:
  .dump /mf c:\full_dump.dmp
it will create a "full" minidump. Then zip that up and email it to me (might be too 
big to post, it should be ~20MB).

If we can't get any further, I can get you a debug build to try the same thing in.

brettw: sent you the zip.  Here are my steps:

* open windbg from cmd
* press f5 until the chrome window appears
* click the bookmark star
* windbg hits a break
* submit the dump command

Let me know if I need to do anything differently.

lame, google bounced my email even though gmail let me send it.

http://rapidshare.com/files/169741831/full_dump.7z.html

Can't be anti-virus apps as I'm running ClamAV which has no background processing on
Windows (yet). I don't see much cause from other apps unless Chrome depends on system
services to be running that I don't have (I've disabled most that have no purpose for
me). And another potential cause is my Athlon 64 X2 CPU, but the only real issue with
that I know of is the RDTSC sync problem and should not cause any problems with
dialog sizing unless it's trying to do something unnecessarily fancy.

I also have an Athlon x2, if that can be of any help...

hi ...i use google chrome on Windows Vista Home Premeuim..Suddenly Google Chrome 
Crashed. This did not happened any rime with IE..Any solutions to this problem

javlopez: thanks for the great dump. I spent some time looking at it, and I have a 
speculative fix I'd like you to try. Can you email me: brettw@chromium.org and I'll 
reply with the address of the .zip file and instructions? It's big and I don't want 
to kill my little server with a bunch of people downloading it.

javlopez: I don't think we need that testing anymore. Scott fount documentation that 
my hunch was right. We're getting bitten by a bug in Windows documented on the bottom 
of this page:
  http://msdn.microsoft.com/en-us/library/bb775839(VS.85).aspx
Some more description is here:
  http://kbalertz.com/947841/computer-running-version-Windows-Server-version-Windows-
Visual-Studio-crashes-dialog.aspx

I'll make this change.

brettw: awesome, nice work!

Hopefully fixed on trunk in r6420. I'm leaving open so some users can confirm this.

If somebody experiencing this error can check a trunk build since then such as:
  http://build.chromium.org/buildbot/snapshots/chromium-rel-xp/6449/
that would be great.

Note that this is a continuous build, it won't autoupdate, and I know for a fact 
isn't as stable as we want, so you probably don't want to use it as your main browser 
beyond testing.

I can open the options window with that build (for the first time ever with Chrome),
so I guess whatever you did fixed the problem.

Thanks, I'll call that done.

Congratulations.

I'm eager to test it in the next autoupdate release. When will it be ?

 Issue 3547  has been merged into this issue.

This fix is now in the dev channel release 0.4.154.33.

Yep, works like a charm now, thanks !

Verified in the official build Chrome: V - 154.36

Very useful files search engine. http://indexoffiles.com is a search engine designed
to search files in various file sharing and uploading sites

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=60650

------------------------------------------------------------------------
r60650 | timurrrr@chromium.org | Mon Sep 27 08:32:36 PDT 2010

Changed paths:
 M http://src.chromium.org/viewvc/chrome/trunk/src/tools/valgrind/drmemory/suppressions.txt?r1=60650&r2=60649&pathrev=60650

Revert some Dr. Memory suppressions since they re-appeared as  issue 55  (was:  issue 49 )
TBR=glider
Review URL: http://codereview.chromium.org/3509002
------------------------------------------------------------------------

I did a right click on the Google chrome shortcut, went to properties, then to Shortcut. I then copied the target. 
Mine was "C:\Documents and Settings\Akshay\Local Settings\Application Data\Google\Chrome\Application\chrome.exe"

Then from the COMPATIBILITY Option> learn more about PROGRAM COMPATIBILITY OPTION.

It opens a Windows HELP & SUPPORT Page.
There is an option PROGRAM COMPATIBILITY WIZARD, Start that.
Follow the steps.
Paste the TARGET copied from GOOGLE chrome Shortcut.

Finally it will run the browser. Mine did.

Try it.

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=86790

------------------------------------------------------------------------
r86790 | nduca@chromium.org | Thu May 26 00:33:16 PDT 2011

Changed paths:
 M http://src.chromium.org/viewvc/chrome/trunk/src/DEPS?r1=86790&r2=86789&pathrev=86790

Roll ANGLE to r654

------------------------------------------------------------------------
r654 | nduca@chromium.org | 2011-05-24 18:55:06 -0700 (Tue, 24 May 2011) | 4 lines

Use D3DSWAPEFFECT_FLIPEX on Windows 7.

Reduces frame rate jitters on Windows 7 reported in crbug.com/82925
Review URL: http://codereview.appspot.com/4517096
------------------------------------------------------------------------
r653 | daniel@transgaming.com | 2011-05-17 11:34:24 -0700 (Tue, 17 May 2011) | 3 lines

Add missing newlines to source files

Issue=154
------------------------------------------------------------------------
r652 | daniel@transgaming.com | 2011-05-17 11:34:18 -0700 (Tue, 17 May 2011) | 4 lines

Improve robustness of Context::finish and flush

Trac #16690
Signed-off-by: Nicolas Capens
------------------------------------------------------------------------
r651 | daniel@transgaming.com | 2011-05-13 09:05:22 -0700 (Fri, 13 May 2011) | 7 lines

Fix the order of the error() parameters in Surface.cpp and advertise EGL_ANGLE_d3d_share_handle_client_buffer extension

Trac #16711

Also rename the extension file to match the extension name

Signed-off-by: Nicolas Capens
------------------------------------------------------------------------
r650 | jbauman@chromium.org | 2011-05-12 16:04:07 -0700 (Thu, 12 May 2011) | 6 lines

Add support for eglCreatePbufferFromClientBuffer, taking as input a D3D share handle corresponding to the enum EGL_D3D_TEXTURE_2D_SHARE_HANDLE_ANGLE. No spec as of yet, but it does the obvious thing.

BUG=129
TEST=

Review URL: http://codereview.appspot.com/4325048
------------------------------------------------------------------------
r649 | jbauman@chromium.org | 2011-05-12 13:53:34 -0700 (Thu, 12 May 2011) | 8 lines

Use SSE2 to swizzle RGBA to BGRA

Using SSE2 can drastically reduce the amount of time it takes to do glTexImage2D. I've also added a plain-C path that's much faster than the one that was there before.

BUG=151
TEST=

Review URL: http://codereview.appspot.com/4465052
------------------------------------------------------------------------
r648 | apatrick@chromium.org | 2011-05-12 11:15:03 -0700 (Thu, 12 May 2011) | 38 lines

Cache result of D3DPERF_GetStatus.

I profiled the WebGL acquarium with AMD CodeAnalyst and it turns out this function got a lot of hits. Top 10 before change:

CS:EIP     	Symbol + Offset                                                              	64-bit 	Timer samples
0x62a87260 	CBatchFilterI::ProcessBatch                                                  	       	13.41
0x62a171f1 	D3DPERF_GetStatus                                                            	       	6.04
0x629ce831 	CD3DBase::SetVertexShaderConstantF_FP                                        	       	5.12
0x62a88bea 	CBatchFilterI::GetBatchBufferPointer<_LH_SETPIXELSHADERCONSTIMM_TOKEN_SMALL> 	       	4.61
0x6298060b 	UpdateViewportCache                                                          	       	4.2
0x6298da3a 	CD3DBase::UpdateTextures                                                     	       	3.58
0x6298db6b 	CD3DDDIDX10::SetTexture                                                      	       	2.76
0x6298df1d 	CD3DDDIDX10::InsertStreamSource                                              	       	2.46
0x629d1c1a 	CD3DBase::SetPixelShaderConstantF_FP                                         	       	2.25
0x6297efc4 	CD3DHal::SetSamplerState_FP                                                  	       	2.05

10 functions, 186 instructions, Total: 454 samples, 46.47% of shown samples, 2.51% of total session samples

And after:

CS:EIP     	Symbol + Offset                                                              	64-bit 	Timer samples
0x69317260 	CBatchFilterI::ProcessBatch                                                  	       	13.87
0x69318bea 	CBatchFilterI::GetBatchBufferPointer<_LH_SETPIXELSHADERCONSTIMM_TOKEN_SMALL> 	       	5.84
0x6921060b 	UpdateViewportCache                                                          	       	5.29
0x6925e831 	CD3DBase::SetVertexShaderConstantF_FP                                        	       	4.93
0x6921da3a 	CD3DBase::UpdateTextures                                                     	       	4.38
0x6921e034 	CD3DBase::SetStreamSource_FP                                                 	       	3.65
0x69261c1a 	CD3DBase::SetPixelShaderConstantF_FP                                         	       	3.65
0x69227651 	CD3DBase::DrawIndexedPrimitive                                               	       	2.74
0x6920efc4 	CD3DHal::SetSamplerState_FP                                                  	       	2.37
0x6925e9f7 	CD3DBase::SetVertexShaderConstantIntF                                        	       	2.37

10 functions, 152 instructions, Total: 269 samples, 49.09% of shown samples, 0.80% of total session samples

UpdateViewportCache looks like it might be low hanging fruit as well.
Review URL: http://codereview.appspot.com/4535049
------------------------------------------------------------------------

Review URL: http://codereview.chromium.org/7077003
------------------------------------------------------------------------

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=86926

------------------------------------------------------------------------
r86926 | nduca@chromium.org | Thu May 26 16:22:38 PDT 2011

Changed paths:
 M http://src.chromium.org/viewvc/chrome/trunk/src/DEPS?r1=86926&r2=86925&pathrev=86926

Revert 86790 - Roll ANGLE to r654

------------------------------------------------------------------------
r654 | nduca@chromium.org | 2011-05-24 18:55:06 -0700 (Tue, 24 May 2011) | 4 lines

Use D3DSWAPEFFECT_FLIPEX on Windows 7.

Reduces frame rate jitters on Windows 7 reported in crbug.com/82925
Review URL: http://codereview.appspot.com/4517096
------------------------------------------------------------------------
r653 | daniel@transgaming.com | 2011-05-17 11:34:24 -0700 (Tue, 17 May 2011) | 3 lines

Add missing newlines to source files

Issue=154
------------------------------------------------------------------------
r652 | daniel@transgaming.com | 2011-05-17 11:34:18 -0700 (Tue, 17 May 2011) | 4 lines

Improve robustness of Context::finish and flush

Trac #16690
Signed-off-by: Nicolas Capens
------------------------------------------------------------------------
r651 | daniel@transgaming.com | 2011-05-13 09:05:22 -0700 (Fri, 13 May 2011) | 7 lines

Fix the order of the error() parameters in Surface.cpp and advertise EGL_ANGLE_d3d_share_handle_client_buffer extension

Trac #16711

Also rename the extension file to match the extension name

Signed-off-by: Nicolas Capens
------------------------------------------------------------------------
r650 | jbauman@chromium.org | 2011-05-12 16:04:07 -0700 (Thu, 12 May 2011) | 6 lines

Add support for eglCreatePbufferFromClientBuffer, taking as input a D3D share handle corresponding to the enum EGL_D3D_TEXTURE_2D_SHARE_HANDLE_ANGLE. No spec as of yet, but it does the obvious thing.

BUG=129
TEST=

Review URL: http://codereview.appspot.com/4325048
------------------------------------------------------------------------
r649 | jbauman@chromium.org | 2011-05-12 13:53:34 -0700 (Thu, 12 May 2011) | 8 lines

Use SSE2 to swizzle RGBA to BGRA

Using SSE2 can drastically reduce the amount of time it takes to do glTexImage2D. I've also added a plain-C path that's much faster than the one that was there before.

BUG=151
TEST=

Review URL: http://codereview.appspot.com/4465052
------------------------------------------------------------------------
r648 | apatrick@chromium.org | 2011-05-12 11:15:03 -0700 (Thu, 12 May 2011) | 38 lines

Cache result of D3DPERF_GetStatus.

I profiled the WebGL acquarium with AMD CodeAnalyst and it turns out this function got a lot of hits. Top 10 before change:

CS:EIP     	Symbol + Offset                                                              	64-bit 	Timer samples
0x62a87260 	CBatchFilterI::ProcessBatch                                                  	       	13.41
0x62a171f1 	D3DPERF_GetStatus                                                            	       	6.04
0x629ce831 	CD3DBase::SetVertexShaderConstantF_FP                                        	       	5.12
0x62a88bea 	CBatchFilterI::GetBatchBufferPointer<_LH_SETPIXELSHADERCONSTIMM_TOKEN_SMALL> 	       	4.61
0x6298060b 	UpdateViewportCache                                                          	       	4.2
0x6298da3a 	CD3DBase::UpdateTextures                                                     	       	3.58
0x6298db6b 	CD3DDDIDX10::SetTexture                                                      	       	2.76
0x6298df1d 	CD3DDDIDX10::InsertStreamSource                                              	       	2.46
0x629d1c1a 	CD3DBase::SetPixelShaderConstantF_FP                                         	       	2.25
0x6297efc4 	CD3DHal::SetSamplerState_FP                                                  	       	2.05

10 functions, 186 instructions, Total: 454 samples, 46.47% of shown samples, 2.51% of total session samples

And after:

CS:EIP     	Symbol + Offset                                                              	64-bit 	Timer samples
0x69317260 	CBatchFilterI::ProcessBatch                                                  	       	13.87
0x69318bea 	CBatchFilterI::GetBatchBufferPointer<_LH_SETPIXELSHADERCONSTIMM_TOKEN_SMALL> 	       	5.84
0x6921060b 	UpdateViewportCache                                                          	       	5.29
0x6925e831 	CD3DBase::SetVertexShaderConstantF_FP                                        	       	4.93
0x6921da3a 	CD3DBase::UpdateTextures                                                     	       	4.38
0x6921e034 	CD3DBase::SetStreamSource_FP                                                 	       	3.65
0x69261c1a 	CD3DBase::SetPixelShaderConstantF_FP                                         	       	3.65
0x69227651 	CD3DBase::DrawIndexedPrimitive                                               	       	2.74
0x6920efc4 	CD3DHal::SetSamplerState_FP                                                  	       	2.37
0x6925e9f7 	CD3DBase::SetVertexShaderConstantIntF                                        	       	2.37

10 functions, 152 instructions, Total: 269 samples, 49.09% of shown samples, 0.80% of total session samples

UpdateViewportCache looks like it might be low hanging fruit as well.
Review URL: http://codereview.appspot.com/4535049
------------------------------------------------------------------------

Review URL: http://codereview.chromium.org/7077003

TBR=nduca@chromium.org
Review URL: http://codereview.chromium.org/7077022
------------------------------------------------------------------------