Re: [chromium-dev] Re: Chrome.exe still running after exiting

[Peter Kasting]

This kind of mail is not appropriate for chromium-dev, which is a technical development list.  Report bugs in the bug tracker at http://crbug.com/ .

PK

[hoang nguyen]

I think i know this problem.

because we use chrome make default browser, so when we start window it will be run. To me, i use ie is default browser. it decreas CPU usage <20%.
On Thursday, February 9, 2012 10:14:56 PM UTC+7, pdwiki wrote:

Hi,

One problem I am facing is that when the user exit's chrome, a single
chrome.exe process is still running. Initially 2 chrome.exe's were
running (i.e. when the browser was in use) but selecting exit from the
menu doesn't kill both the processes. It seems
BrowserList::AttemptUserExit(); in browser.cc is responsible for the
exit. Any idea why a single chrome.exe is still running? This is on a
Windows 7 PC.

Thanks.

[Greg Thompson]

On Tue, Jan 15, 2013 at 10:25 PM, hoang nguyen <hoangng...@gmail.com> wrote:

I think i know this problem.

because we use chrome make default browser, so when we start window it will be run.

This is not correct: making Chrome the default browser does not cause it to run when Windows starts, or when a user logs on. There are features in Chrome that cause it to run in background mode at user logon, but these are orthogonal to Chrome being the default browser.

 

To me, i use ie is default browser. it decreas CPU usage <20%.

On Thursday, February 9, 2012 10:14:56 PM UTC+7, pdwiki wrote:

Hi,

One problem I am facing is that when the user exit's chrome, a single
chrome.exe process is still running. Initially 2 chrome.exe's were
running (i.e. when the browser was in use) but selecting exit from the
menu doesn't kill both the processes. It seems
BrowserList::AttemptUserExit(); in browser.cc is responsible for the
exit. Any idea why a single chrome.exe is still running? This is on a
Windows 7 PC.

Thanks.

--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev

[Darius J]

Has there been any repairs/fixes for this? Im having and entire community of people with this same issue. 50+ and still i cannot get to the bottom of this problem.

[Greg Thompson]

On Mon, Oct 7, 2013 at 9:58 AM, Darius J <dov...@gmail.com> wrote:

Has there been any repairs/fixes for this? Im having and entire community of people with this same issue. 50+ and still i cannot get to the bottom of this problem.

Have you checked to see if any of your users have background apps or extensions installed? Please see the top of this thread for more info.

 

On Thursday, February 9, 2012 10:14:56 AM UTC-5, pdwiki wrote:

Hi,

One problem I am facing is that when the user exit's chrome, a single
chrome.exe process is still running. Initially 2 chrome.exe's were
running (i.e. when the browser was in use) but selecting exit from the
menu doesn't kill both the processes. It seems
BrowserList::AttemptUserExit(); in browser.cc is responsible for the
exit. Any idea why a single chrome.exe is still running? This is on a
Windows 7 PC.

Thanks.

--

[Darius J]

Absolutely, I'm the "comp tech" @ a school district and finally found this thread. I've carefully read through the whole thing and applied all troubleshooting/repairs&workaround listed here.

Some more information about the problem here: https://productforums.google.com/forum/#!msg/chrome/OqgzZrKOzZA/3mZ6vZp0OHYJ  

Please help im going nuts right now, and as far as i know this has been happening on win 7 64. With all extensions off and running in background ticked off and cloud print ticked off...?

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Drew Wilson]

Sorry for the naive question, but how are you exiting Chrome? By closing the window, or by choosing "Exit" from the menu in the browser window?

[Darius J]

"In my line of work there are no naive questions" but to answer your question: I have done both. in which case the same problem exists.

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Scott Hess]

There are bugs like http://crbug.com/10658 which imply that certain system utilities can cause this, too.  That bug is so old, it's likely that it's been repurposed a couple times, but it's possible that there are other hints of similar things in the bug database.

-scott

[Drew Wilson]

Wow. Exiting from the in-browser menu is a pretty hard exit and bypasses all background settings, so this points to something else (not background apps) keeping the browser process alive.

Anyone have any ideas about how to figure out what this is?

[Darius J]

From reading all of your posts it seemed as if it's only tied to a 64 bit 7 pro system(as an issue that is), i'm saying that because i don't have any problems with computers running 32 bit 7 pro. 

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.

[Zachary Turner]

What's the full command line of the chrome.exe that's still running?  (In task manager or process explorer, enable the command line column).

[Darius J]

C:\Users\PHS\AppData\Local\Google\Chrome\Application

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Darius J]

C:\Users\PHS\AppData\Local\Google\Chrome\Application\chrome.exe

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[PhistucK]

The full command line usually contain flags (--type=...). If it does not, it means it is the browser. Oh oh. :(

☆PhistucK

[Darius J]

here is a pic of running tasks

[Darius J]

[PhistucK]

You have a lot of Chrome processes running. I thought you meant that only one process is left active for some reason.

Does the Applications tab list Google Chrome? Can you right click and Bring to Front? Does it show the browser?

☆PhistucK

--

[Darius J]

When closed yes then only one is running, otherwise when i open chrome after process is closed it opens multiple tabs.

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Sigurður Ásgeirsson]

It would be interesting to get a crash dump of the hung Chrome browser process. If you right-click it in task manager and click "Create Dump file" a crash dump will be captured. Could you do that and somehow make the (zipped) .dmp file available?

[Darius J]

https://docs.google.com/file/d/0B7WaT4CaMUPpamdVaENwZHAxVGc/edit?usp=sharing​

 chromeDump.rar

​

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Darius J]

Let me know if you can access it, if not ill try sending another way. Thank You!

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Sigurður Ásgeirsson]

There's a hung thread in there that's preventing shutdown:

The main thread is stuck on this:

chrome_58d00000!content::GamepadProvider::~GamepadProvider

and the thread this is trying to join in stuck in direct input:

0:023:x86> k

ChildEBP RetAddr  

0aa0eef8 769add54 ntdll_77250000!NtReadFile+0x15

0aa0ef5c 767c3ec7 KERNELBASE!ReadFile+0x118

0aa0efa4 6b506508 kernel32!ReadFileImplementation+0xf0

0aa0f230 6b50247c dinput8!CHid_Init+0x2bb

0aa0f248 6b4f48cf dinput8!CHid_New+0x36

0aa0f270 6b4fcef2 dinput8!CDIDev_Initialize+0x8f

0aa0f290 6b4fcf32 dinput8!CDIObj_CreateDeviceHelper+0x45

0aa0f2ac 6b50040e dinput8!CDIObj_CreateDeviceW+0x2c

0aa0f2d0 6b4fd06d dinput8!CDIDEnum_Next+0x6b

0aa0f744 5a4a2c09 dinput8!CDIObj_EnumDevicesW+0xe1

0aa0f798 5a4a2d60 chrome_58d00000!content::GamepadPlatformDataFetcherWin::EnumerateDevices+0xfd [c:\b\build\slave\win\build\src\content\browser\gamepad\gamepad_platform_data_fetcher_win.cc @ 298]

0aa0f7bc 5a47df8d chrome_58d00000!content::GamepadPlatformDataFetcherWin::GetGamepadData+0x7a [c:\b\build\slave\win\build\src\content\browser\gamepad\gamepad_platform_data_fetcher_win.cc @ 338]

0aa0f7dc 58d2a9a1 chrome_58d00000!content::GamepadProvider::DoPoll+0x48 [c:\b\build\slave\win\build\src\content\browser\gamepad\gamepad_provider.cc @ 174]

0aa0f848 58d2ac49 chrome_58d00000!base::MessageLoop::RunTask+0x221 [c:\b\build\slave\win\build\src\base\message_loop\message_loop.cc @ 484]

0aa0f8b4 58d2b463 chrome_58d00000!base::MessageLoop::DoDelayedWork+0xcb [c:\b\build\slave\win\build\src\base\message_loop\message_loop.cc @ 646]

0aa0f93c 58d2b404 chrome_58d00000!base::MessagePumpForIO::DoRunLoop+0x54 [c:\b\build\slave\win\build\src\base\message_loop\message_pump_win.cc @ 538]

0aa0f95c 58d2a329 chrome_58d00000!base::MessagePumpWin::Run+0x3e [c:\b\build\slave\win\build\src\base\message_loop\message_pump_win.h @ 48]

0aa0f980 58d2a280 chrome_58d00000!base::MessageLoop::RunInternal+0x73 [c:\b\build\slave\win\build\src\base\message_loop\message_loop.cc @ 437]

0aa0f994 58d2b3b4 chrome_58d00000!base::RunLoop::Run+0x59 [c:\b\build\slave\win\build\src\base\run_loop.cc @ 46]

0aa0f9bc 58d2b278 chrome_58d00000!base::Thread::Run+0x34 [c:\b\build\slave\win\build\src\base\threading\thread.cc @ 158]

0aa0faac 58d2b0db chrome_58d00000!base::Thread::ThreadMain+0xdc [c:\b\build\slave\win\build\src\base\threading\thread.cc @ 207]

0aa0fac0 767c336a chrome_58d00000!base::`anonymous namespace'::ThreadFunc+0x58 [c:\b\build\slave\win\build\src\base\threading\platform_thread_win.cc @ 78]

0aa0facc 77289f72 kernel32!BaseThreadInitThunk+0xe

0aa0fb0c 77289f45 ntdll_77250000!__RtlUserThreadStart+0x70

0aa0fb24 00000000 ntdll_77250000!_RtlUserThreadStart+0x1b

I don't know anything about direct input, but this shouldn't be happening - I'll try and find someone who know of this code ...

Perhaps you and your users are seeing this because the machines you manage all have some sort of oddball gampad/input device? 

Siggi

[Darius J]

interesting....? we all are running latitude E5430 and E5420's with 64 bit win 7. All have capabilities of connecting to a docking station. I do have mine in a docking station right now but the remaining teachers don't. At least not yet, and have the same issue. Right now that's the only type of input device I can think of. 

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Justin TerAvest]

I'm pretty sure Flash once caused the gamepad polling thread to be started, regardless of whether a gamepad was present or not:

https://code.google.com/p/chromium/issues/detail?id=277691

I don't know what version of Chrome this should be fixed in, though.

[Darius J]

Interesting you should write that, i disabled plugins one at a time and flash was the one causing the issue, but since some websites  are flash/html(as well as other code types) and others mostly flash dominant chrome leaves a link to re-enable the plugin which then still causes the same issue to re-occur. I wasn't sure how to progress from here though(which then your response came up). Whats interestingly enough though is your post at the same time of me playing with the plugins, perfect timing!

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Darius J]

here is the (chrome://profiler)

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Darius J]

I see it running, but i'm not sure how to move forward on this. Is there something i can uninstall, or just wait for a newer version of chrome/flash?

Best Regards,

Darius Jankowicz

Computer Technician

Phone: 8622230525

Email: djankowicz@passaic-city.k12.nj.us

Passaic City Public Schools

101 Passaic Ave., Passaic, NJ 07055

www.passaic-city.k12.nj.us

[Sigurður Ásgeirsson]

FYI: I filed this as http://crbug.com/305267 - anyone here know who'd be good to take this bug on?

Note that this happens in 30.0.1599.69 which is our current stable - and we'll be blind to this from crashes :(.

[Junk Spamming]

I was having this issue too, however I resolved it by turning off Print Sharing in Advanced Settings.  I figured the process was enabled after exiting since Google still needed to run to print remotely.  Disabled the printer option and the process went away.

[Darius J]

It's more than just that.

On Thursday, October 17, 2013, Junk Spamming wrote:

I was having this issue too, however I resolved it by turning off Print Sharing in Advanced Settings.  I figured the process was enabled after exiting since Google still needed to run to print remotely.  Disabled the printer option and the process went away.

--

--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev

To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.

--

[Rick Westbrock]

I was having the same problem of multiple (usually about ten) chrome.exe processes still running after closing the app and making sure it is gone from the Applications tab of Task Manager but I'm running Win7 Home Premium 64-bit. I started disabling extensions one by one to narrow it down and it once I disabled Fitbolt for Google Chrome the problem went away. I don't know if that extension uses Flash but I suspect it may since it has things like a countdown clock. At any rate I am installing the desktop client for Fitbolt instead so that I can remove the Chrome extension. (The web site is fitbolt.com if anyone is inclined to research it.)

[AR modder]

On Thursday, February 9, 2012 10:14:56 AM UTC-5, pdwiki wrote:

Hi,

One problem I am facing is that when the user exit's chrome, a single
chrome.exe process is still running. Initially 2 chrome.exe's were
running (i.e. when the browser was in use) but selecting exit from the
menu doesn't kill both the processes. It seems
BrowserList::AttemptUserExit(); in browser.cc is responsible for the
exit. Any idea why a single chrome.exe is still running? This is on a
Windows 7 PC.

Thanks.

 Yeah this has happened to me too on and off and if the other Chrome.exe is running it wont allow me to open chrome again unless I kill the process or restart my comp, it is actually more then just an inconvenience it is also doubling the CPU load seems like another issue google needs to iron out.

[Ryan Sleevi]

-bcc:chromium-dev

> --
> --
> Chromium Developers mailing list: chromi...@chromium.org
> View archives, change email options, or unsubscribe:
> http://groups.google.com/a/chromium.org/group/chromium-dev

Resurrecting a two year old thread on a development list is not the best way to file a bug or help resolve the issue.

Please see http://crbug.com/new to file a bug, which can help the right people follow-up.

Cheers.

[Glen Lawrence]

This worked for me.  In my case it was the McAfee Site Adviser Extension.  I have 2 other extensions that work just fine.  It was the McAfee extension that was the culprit.  Thanks for your post!!

On Wednesday, October 3, 2012 1:05:08 PM UTC-4, Cau Trindade wrote:

I had the same problem and solve it !!

Go to TOOLS > EXTENSIONS  menu. Uncheck / deselect ALL extensions you have installed (don't delete them, just deactivated all.

Quit Chrome and check if all tasks/processes were ended. If so, it means one or more of the installed extensions are causing this problem.

Now, activate one by one quiting Chrome and checking the processess for all of each of extensions activation. You will find which one is causing the problem.

In my case the problem "DvdVideoSoft Free Youtube Download 1.0.0.0" extension. 

[bailey.r...@gmail.com]

Hi. I'm having the same problem. Disabled ALL extensions and it still hasn't resolved the issue. Any other things I could try?

Thanks.

[Benjamin Carl]

I am having this same problem where Chrome continues to run after closing, even after disabling all extensions and plugins.  I already unchecked the option to allow background processes to run.  I've tried closing Chrome both by clicking the X and by choosing Exit from the menu.  

When I have Chrome open with just a single tab, I run the Chrome Task Manager, it only shows 3 processes: the browser, the tab that I have open, and the GPU process.  But when I open Windows Task Manager, it shows many more processes, over a dozen.  When I close Chrome, a few of the processes go away, but at least 8 or 9 stay running.  When I try to kill the processes, they come right back.  

This is causing my computer to run hard and churn all day long, the fan runs constantly.  I know that it's Chrome doing it, because the problem goes away as soon as I uninstall Chrome.

I have attached a screenshot of my Windows task manager while Chrome is closed.  I have added the columns for path and command line so you can see them

[PhistucK]

You seem to be having malware.

https://www.google.com/search?q=gcc+controller.exe

Note the --user-data-dir that points to ...Temp/GC/SOME-GUID. Unless you created this one yourself, this one is run by malware.

I figured it due to the GCC/Controller.exe that is also running.

☆PhistucK

--

--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev

To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.

[Dave Gurney]

Hi,

I think I have a worse problem.  Chrome.exe is still running (anywhere from 4 to 9 processes in task manager) after uninstalling Google Chrome and rebooting.
The chrome.exe processes in task manager even come back after manually ending the processes.

I hope someone has a fix for this problem.  It will probably fix your problem at the same time.
Dave

PS:  I have 5 chrome.exe  processes in task manager now.

On Thursday, February 9, 2012 10:14:56 AM UTC-5, pdwiki wrote:

[PhistucK]

Post the command lines of the processes​, it may be malware as well.

☆PhistucK

--

--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev

[Cindy Davis]

I have to close Chrome in my task manager, because if I don't it will never open up after I exit. This is really getting bothersome in some way I need a fix.

[PhistucK]

Post the command lines of the zombie chrome.exe processes. You can use Process Explorer or the Windows 7 or later task manager (add the command line column).

☆PhistucK

--

--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev

[Michael Landry]

FYI this is happening sporadically on my Windows 8.1 machine. Same experience as Cindy Davis above... one chrome.exe process remains running after exit, and you cannot open another Chrome window until you kill the orphan process. Commandline is simply "chrome.exe", no commandline parameters. 

[Sigurður Ásgeirsson]

If you right-click the process in task manager and select "Create a dump file" and upload it somewhere. From the dump file, we ought to be able to figure out what's up...

[Cynthia Meyerhoefer]

I have this problem in particular with any site that uses Flash Player.  I just started looking at  at my chrome task manager; at one point yesterday and it was over 500,000 (I also see those sorts of numbers on Windows Task Manager for a Chrome Process, had my suspicions that is was Flash Player, and now have confirmation).

[KonnaBeans]

I was able to solve this exact problem with Settings > System > uncheck 
"Continue running background apps when Google Chrome is closed."

[Sarkie]

I was having this issue.

I disabled the Mozilla Flash, same issue, disabled the Google one, fixed the issue, re-enabled the Google one and the issue is gone, now I can't get it back. 

I did take a dump of it lingering around though.

https://drive.google.com/file/d/0BxvwWozaN8ehNHBQNTlaOW5WY0k/edit?usp=sharing

The command line is straight to the chrome.exe itself.

[Sang Lee]

This is an old post but I reply this for anybody having the same problem with me.

I use Windows 8.1 and Lenovo Laptop.

In my case,"LocationTaskManager" service from Lenovo was using Chrome

even after I closed chrome or when I never started chrome.

LocationTaskManager service is installed with "Lenovo Settings" app(from Windows Store)

for its "Location Awareness" function.

Lenovo's Location Awareness function is something like this:

when I move to somewhere(so I "change my location'), some environments such as WiFi, ethernet,.. also changes.

Location Awareness function makes "profiles" for these different environments, 

and automatically update the settings for that environment using these profiles whenever I move to somewhere.

For this functionality to be possible it needs to catch my "location" or "location change".

It guess that Lenovo's Location Awareness uses some function of Chrome's to know

my laptop's current location.

After I disabled "LocationTaskManager" service and reboot, this problem was solved.

----

In addition, it seems that Lenovo's LocationTaskManager occupies Chrome's Temp folder,

so I couldn't save any settings of Chrome(theme...etc.).

after disabling the service, this problem is also solved.

---

Anybody having the same problem with me and using Lenovo's software

please try to disable Lenovo's LocationTaskManager service or

uninstalling Lonovo Settings app.

hope this can be a help.

[Zachary Turner]

chromium-dev -> bcc

This is not a chrome support mailing list, it is for developers making changes to the chrome codebase.  If this does not describe you, chances are you are in the wrong place.  In the interest of helping this thread die permanently, however, please check the value of the setting "Continue running background apps when Google Chrome is closed" in chrome://settings.  If it is enabled, then chrome will sometimes continue to run after exiting.  This is not a bug.

If chrome continues to run after exiting even when this setting is unchecked, then please do not respond to this thread, but instead file a bug at http://crbug.com, or alternatively, you can ask in a designated chrome support forum.

--

[dave white]

alt - e > settings > show advanced settings 

under SYSTEM make sure the box beside "Continue running background apps when Google Chrome is closed" is unticked

that helped me.

[eos]

I have a similar problem.  Did you have any luck getting rid of the location task manger or the offending process? I cant find any way to get rid of it, so Chome forgets everything (settings, extensions, bookmarks etc) every time I start chrome.

[Reece]

I've found a fix that at least worked for me but it does involve uninstalling chrome but with 3rd party uninstaller that removes the registry keys and files associated with chrome.

http://www.revouninstaller.com/revo_uninstaller_free_download.html

Uninstall chrome with revo uninstaller and simply reinstall.

Good luck !

[Dustin Johnson]

I had this exact same issue.  How I fixed it was going to Settings (alt+e), then scroll to the far bottom (under advanced settings) and unchecking the box that says "Continue running background apps when Google Chrome is closed".  Fixed the issue for me.  

[Mike Tsao]

That'll fix the issue for sure, but be sure you understand what you're doing. Chrome Apps such as Hangouts and Photos normally run in the background (e.g., you'll miss messages if your chat client isn't running!). All that functionality stops if you exit the Chrome browser when the option is disabled.

--

[Sean K]

I have the similar issue -- where there are 3 or 5 chrome.exe running after I closed all visible chrome windows on my 64 bit windows 8.1 laptop.

I used to have the stand-alone hangout app installed -- but I removed it.

So, why are there so many chrome processes still running after I exited the chrome browser windows?

[PhistucK]

Sounds like malware, really. Perhaps try https://www.google.com/chrome/srt/.

☆PhistucK

On Sun, Jun 28, 2015 at 12:04 AM, Ed Wilson <ed.l...@gmail.com> wrote:

I tried unchecking the box that allows chrome to run in the background. but it still keeps a process running after i close chrome. and re opens after it is terminated.

It has been taking control of my camera and microphone, and requires them to be running at all times.

Pretty much every time i try to use my camera it tells me it is in use by another program. I.e. Chrome.

Know of any reasons why chrome needs to have my camera and mic on 24/7?

On Thursday, February 9, 2012 at 7:14:56 AM UTC-8, pdwiki wrote:

Hi,

One problem I am facing is that when the user exit's chrome, a single
chrome.exe process is still running. Initially 2 chrome.exe's were
running (i.e. when the browser was in use) but selecting exit from the
menu doesn't kill both the processes. It seems
BrowserList::AttemptUserExit(); in browser.cc is responsible for the
exit. Any idea why a single chrome.exe is still running? This is on a
Windows 7 PC.

Thanks.

--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev

To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev...@chromium.org.

[Philippe Verdy]

You should use the Process Viewer in Chrome to look at the command lines that allows laucnhing a Chrome instance. You may then find that there are parameters connecting Chrome to an external application lauching it (beware notably if this application uses "nacl32.exe" meaning that it is capable of running arbitrary native code without being contrained by the sandbox (the sandbox only isolates Chrome itself and its internal APIs from what the native code will run in the background with a direct access to you host).

Use your antivirus or if it does not detect it, use MalwareBytes AntiMalware to locate that malicious tool. If you have a difficulty to stop this malware, may be you can force it to crash by unplugging your webcam or micro when it is active : try with the device manager, and disable these devices. However many malwares are also automatically restarted after a crash (don't forget to inspect your task scheduler which monitors crashing apps in order to relaunch them: you may need to remove them.

If you remove this task and immediately reappears, it is because the malware uses multiple processes and you can't kill them all at the same time: all those processes are monitoring each other som that when you kill one of them, the others will relaunch the missing one immediately. The only way to avoid that is to boot in safe mode (you may enable networking, this will allow your antivirus or antimalware to use online updates or more advanced detection.cleanup tools specific to what it discovers on your system, to make sure they clean every entry point allowing the malware to restart: many malwares have self-defenses now and even if the really dangerous part is in a single process, there are lots of very small "companion" tools whose only role is to reinstall the malware or download new variants and entry points, and unlocking some system protections to allow the full malware to run again with full privileges).

When you are in safe mode, don't forget aftert performing a cleanup to reboot again in safe mode and restart the antimalware scanning: frequently it will find other entry points which were previously inactive but only present to be activated when another entry point was non functional to hide them.

Then use SFC /SCANNOW on windows to make sure your system files are correct. If this still does not work, look for "CheckSUR" on the internet to inspect the content of the registry of installed windows components, containing the list of installable or installed packages and their digital signatures: some malwares are dropping entries there to make sure that SFC /SCANNOW will not work correctly: you may have such problems if the "optional features" in the control panel do not list any package but shows an blank window: you need a safe and complete copy of this database of packages and signatures.

If CheckSur detects some altered packages, you'll have to drop them if you can't get a safe copy from another installed PC. When all this is done, SFC /SCANNOW will work and will be able to inspect also the critical system drivers (including thoise that are runnin in safe mode, notably the SCSI disk drivers and fielsystem drivers, or power management drivers).

If you're still locked, the simplest is to boot your PC from another safe image (a CDROM or USB key in order to run the recovery tools from there, or to boot another OS such as Linux, or to mount and create a separate minimal OS installation that you'll run in dual boot mode). Another solution is to boot the Windows installer to perform an upgrade or install another separate copy, and use that alternate OS to inspect the filesystems used by the first OS, but first make sure that you don't boot anything from your origiunal system disk: boot from another device using the BIOS function key (F8 key is for the multiboot in Windows, F12 is frequently used for selecting the first boot device, or some BIOSes propose it by pressing DEL to enter their menu in which you can select the device to boot first.

But for your case, I really think that what you have is only an external software relauching chrome and using it for rendering HTML contents, or inspecting your private data store: booting in safe mode with networking, you should uninstall Chrome and reinstall it completely (but don't connect immediately to your online Google account, as it will start reinstalling your preferences, including reactivating some dangerous plugins/extensions.

First perform your cleanup by running Chrome without connecting it to your Google account.

(in the interim, you could still use the internet using another browser, including IE, Firefox, or Opera, or a more basic browser: there are some tiny webbrowsers that do not require any installation and can run from a single executable, creating only temporary caches, and with minimal or no support of multimedia extensions such as video, advanced javascripts) An alternative is to find help on the internet from another connected device (use your smartphone or tablet, or the old PC that you gave to your children or your connected smart TV, disable the public Wifi or make sure it has a WPA protection, or change its setting so that other devices on your home network will be isolated and not polluted by other infected devices in your private LAN, unplug the unnecessary Ethernet cables; rebooting a router is also a good way to forcfe your connectin to get a new IP and avoid immediate reinfection from the internet by remote hosts that know your current public IP)

If you have doubts about the configuration of your home router, use its builtin "factory reset" to force it back to its initial safe configuration (this should drop forwarding rules, and connected devices will ahve to be reassociated individually. Make sure you get a clean device fully updated with its system updates before trying to cleanup another one). Use this safe configuration without tuning it for specific softwares. Some routers also offer their own "safe mode" in which you don't need to reset the standard configuration, but forcing it to use the safe one as long you your don't reboot it normally (this way your previous configuration will be kept and reused later, notably post forwarding rules; some others just propose to cleanup the port forwarding rules, and dissociate all known local devices on your LAN). Finally don't forget to perform router system updates (if they are proposed in its configuration menus)

[PhistucK]

On Mon, Jun 29, 2015 at 1:30 AM, Philippe Verdy <ver...@gmail.com> wrote:

You should use the Process Viewer in Chrome to look at the command lines that allows laucnhing a Chrome instance. You may then find that there are parameters connecting Chrome to an external application lauching it (beware notably if this application uses "nacl32.exe" meaning that it is capable of running arbitrary native code without being contrained by the sandbox (the sandbox only isolates Chrome itself and its internal APIs from what the native code will run in the background with a direct access to you host).

Note that, as far as I know, Native Client is, in fact, sandboxed​ and does not allow unsandboxed code.

PPAPI, on the other hand, does allow unsandboxed code, but only in its non Native Client mode (and those modules tend to be verified by hard coded mechanisms within the browser).

☆PhistucK

[Victor Khimenko]

On Mon, Jun 29, 2015 at 8:31 AM, PhistucK <phis...@gmail.com> wrote:

On Mon, Jun 29, 2015 at 1:30 AM, Philippe Verdy <ver...@gmail.com> wrote:

You should use the Process Viewer in Chrome to look at the command lines that allows laucnhing a Chrome instance. You may then find that there are parameters connecting Chrome to an external application lauching it (beware notably if this application uses "nacl32.exe" meaning that it is capable of running arbitrary native code without being contrained by the sandbox (the sandbox only isolates Chrome itself and its internal APIs from what the native code will run in the background with a direct access to you host).

Note that, as far as I know, Native Client is, in fact, sandboxed​ and does not allow unsandboxed code.

Indeed. In fact it's harder to bypass sandbox of NaCl that ty bypass sandbox of renderer. NaCl is often used to exploit errors in renderer, but I've never seen an exploit which circumvents NaCl protection directly (although I've seen three (or was it four?) errors which made it possible to create proof-of-concept exploit, but only one was actually a mistake in NaCl code, most were CPU-induced problems). 

 

PPAPI, on the other hand, does allow unsandboxed code, but only in its non Native Client mode (and those modules tend to be verified by hard coded mechanisms within the browser).

The only way to load third-part PPAPI module is via the command line switch, and if you could alter Chrome's command line then you probably have a way to do many other interesting things.

[Philippe Verdy]

I've recently been infected by an adware installed without my consent, via NaCl, using NaCl also to run and spying all my internet navigation, using also about 1 mbps of download bandwodth and nearly 1mbps of upload (certainly runing an open relay), and that constantly used 100% of one CPU core. I used lots of cleaners and antivirus tools they did not detect anything.

Finally I saw that it had corrupted the Windows Component Store to remove the digital signature of some components, notably for the boot recovery and a version of the C runtime library used in the kernel. I had also installed a new hidden user account, stored in an encrypted volume mounted on an encypted partition that it had allocated at end of the disk (but which was not even visible in diskpart: I had to unmount the disk to inspect it on another PC to see that hidden partition and the fact that it was using another partition table displaying a disk with a reduced size not matching the hardware size.

Deleting the partition was not enough, but the most active part of the adware was constantly running in NaCl, but its traffic on the Internet was in fact driven to an external process using a debufgging interface to connect to the NaCl hosted process, and then to relay the internet trafic without using Chrome's API. Effectively there was an hidden VM running in parallel to Windows and a hidden hypervisor started from boot (this hypervisor was using a Linux-based core). There was a strange behavior of the CPU which had some functionality bits altered (only emulated as if they were working: the noexec bit for example was emulated but not functional at all, and the memory map contained a PCI allocation for a large block mapped to an unknown device in Windows but left without driver, this was apparently the way for the adware to communicate secretly with the hypervisor using some memory mapped protocol in that block configured in the PCI map).

If I detected it, that was because of my (8-core/16-threads) CPU constantly using 12.5-13%% total time  where it should have been nearly 0%(that NaCl process used two cores runing 2 threads with extreme activity, probably trying to crack passwords or to steal my resources for computing bitcoins given to someone else, because I noted that there was recurrent installation of a Python engine regularly reinstalled in temporary folders where it could not be deleted even by the administrator, and it was not the Python engine used by Google Drive; sometimes it was running some java applet and there was a JVM installed as well in a temporary folder).

Also because that CPU was also using a lot of critical sections blocking other CPUs too often (the UI was extremely slow to respond, videos had frequent pauses, played sounds had too many audible "cracks" caused by timeouts, that NaCl was contantly taking the first priority on CPU time, GPU time, or in disk I/O, runing above threads with real-time constraints for kernel drivers; even a debugger could not stop that NaCl process), even after stopping Chrome and disabling all background processes. And it was also constantly using a significant part of the GPU. The system was also extremely hot and thermal regulation had difficutlies to maintain temperatures just below 100°C, when usually my PC runs applications around 60°C, and idle time is at about 45°C. The 100°C were also reached directly within the BIOS menu were regulation was completely inactive.

It was really a surprise to see it because it was running on a fresh new installation of Windows, apparently it could be installed along with the OS itself, and the first software that I installed immediately was Avast (even before waiting for the full installation of system updates in the background), and then I installed Chrome and rebooted. This NaCl process was then running (including when runing Windows in Safe Mode, where Windows apparently allows arbitrary SCSI drivers to run in that mode, even if they are not signed and it is enough to lauche the code that will then run arbitrary processes or infect Windows including in the middle of its installation: it had corrupted the Component store, CheckSur detected that there were missing components but no source for them and no signatures for all executable files, and SFC /SCANNOW did not detect any problem, the MBR and UEFI partitions seemed cleaned as well, inspecting the disks from another PC was not sufficient, there was also some code stored in the non volatile flash memory of PCI devices: it was also necessary to purge the PCI configuration of these devices, something imposible to do from the BIOS menu that was already virtualized by the spyware)

[PhistucK]

Interesting (I only read the first paragraph, sorry). Was it the actual Google Native Client (it located alongside chrome.exe, in the official path of Google Chrome)? Or was it some executable disguised as "nacl32.exe"?

If the latter, then mind your accusations and false facts (that Native Client is unsandboxed). If the former, it would have been great to send this to Google, but I guess you cleaned it already.

☆PhistucK

[Torne (Richard Coles)]

I have no idea what is actually the case here but it is at least theoretically possible that a piece of malware could itself use nacl as a component :) If I was going to run a botnet I wouldn't want to let the people I rented it to run arbitrary code on it, I would want to sandbox it ;)

[Victor Khimenko]

On Tue, Jun 30, 2015 at 12:34 AM, Torne (Richard Coles) <to...@chromium.org> wrote:

I have no idea what is actually the case here but it is at least theoretically possible that a piece of malware could itself use nacl as a component :) If I was going to run a botnet I wouldn't want to let the people I rented it to run arbitrary code on it, I would want to sandbox it ;)

At that point NaCl sandbox was effectively disabled:

On Mon, Jun 29, 2015 at 11:31 PM, Philippe Verdy <ver...@wanadoo.fr> wrote:

There was a strange behavior of the CPU which had some functionality bits altered (only emulated as if they were working: the noexec bit for example was emulated but not functional at all, and the memory map contained a PCI allocation for a large block mapped to an unknown device in Windows but left without driver, this was apparently the way for the adware to communicate secretly with the hypervisor using some memory mapped protocol in that block configured in the PCI map).

NaCl process has two sandboxes, but first relies on the noexec bit (on Windows x64; it uses segment on 32 bit Windows) and second one relies on the ability to restrict access to Windows kernel API. From the description it looks like hypervisor installed hooks to disable/bypass both sandboxes.

I wouldn't say that this example shows that you could use NaCl to execute arbitrary code: of course if you installed malicious hypervisor you could disable any and all sanboxes and could inject malware into any piece of software. Malware authors just decided to use NaCl to run their code for some reason (not really sure why).

[pdwiki]

[Julian Pastarmov]

Hi,
Could it be you are running some extensions in background mode? Those
can keep a chrome instance running it the system tray AFAIK.
-Julian

[Mike Frysinger]

alt+e (pull up wrench menu) -> view background processes

alt+e (pull up wrench menu) -> settings -> under the hood -> background apps
-mike

[Albert Bodenhamer]

Have you enabled the cloud print service?  That sets up a persistent service process for handling print jobs.

Can you get the command line of the last running process?

--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
   http://groups.google.com/a/chromium.org/group/chromium-dev

--
Albert Bodenhamer | Software Engineer | abodenha@chromium.org 

[Finnur Thorarinsson]

I think we may have introduced a bad regression somewhere in our code that causes shutdown not to occur properly.

I don't know much about the problem at this point, but I've now seen various reports on Win, Linux and Mac from a number of people.

Some symptoms:

Canary suggests an update being available, yet won't restart when you ask it to.

Chrome still running in the background after shutting down.

Closing Chrome while debugging in Visual Studio doesn't end the debugging session.

Cannot build Chrome because .exe or .dll is in use.

If anyone has any info/bug number or is looking into it, please update the thread so we can keep track.

[Eric Roman]

This was happening on Raman's computer last night too, and causing problems.

For starters, Chrome was rarely actually exitting after closing the last window, which might be "working as intended" based on having the "Continue running background apps when Google Chrome is closed" option on. Whatever the case, it  was very confusing and annoying from a user perspective...

The bigger problem is the browser would occasionally get into a wedged state where after re-launching no page would actually load!

It is unclear what the cause for this was since we lost the ability to reproduce after mucking around.

It looked like what was happening is the browser was spawning off renderer processes as expected, however they would immediately commit seppuku. The end result was a blank page that would spin indefinitely at "Loading..." (why no awe snap?)

Doesn't seem to be happening for him anymore after disabling the "Continue running background apps when Google Chrome is closed".

Raman is seeing if he can reproduce the issues again, but it seems related to background apps.

[Raman Tenneti]

Today, wasn't able to reproduce the wedged state Eric and I were seeing last night (my profile has been upgraded to M17 from M16).

Details

----------

Wasn't able to disable the following extension/renderer processes.  There were two renderer processes (one is an internal process) running in my windows task manager (after closing all windows). If we disable "Let Google chrome run in the background", all processes of chrome exit when we close the last window.

- Oflline Google mail

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=renderer --lang=en-US --force-fieldtest=ConnCountImpact/conn_count_6/ConnnectBackupJobs/ConnectBackupJobsEnabled/DnsImpact/default_enabled_prefetch/IdleSktToImpact/idle_timeout_10/Instant/AllControlA/Prerender/ContentPrefetchPrerender2/PrerenderFromOmnibox/OmniboxPrerenderEnabled/ProxyConnectionImpact/proxy_connections_32/SpdyCwnd/cwndMin16/SpdyImpact/npn_with_spdy/WarmSocketImpact/warm_socket/ --extension-process --renderer-print-preview --channel=6980.051D9380.1447726051 /prefetch:3

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=renderer --lang=en-US --force-fieldtest=ConnCountImpact/conn_count_6/ConnnectBackupJobs/ConnectBackupJobsEnabled/DnsImpact/default_enabled_prefetch/GlobalSdch/global_enable_sdch/IdleSktToImpact/idle_timeout_10/Instant/AllControlA/Prerender/ContentPrefetchPrerender2/PrerenderFromOmnibox/OmniboxPrerenderEnabled/ProxyConnectionImpact/proxy_connections_32/SpdyCwnd/cwndMin16/SpdyImpact/npn_with_spdy/WarmSocketImpact/warm_socket/ --extension-process --renderer-print-preview --channel=6980.0542C000.133991452 /prefetch:3

After reading email and open/close windows and then close all browser windows, the following processes were running (was able to exit the browser by right clicking on chrome icon in System Tray and then clicking on Exit). 

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=renderer --lang=en-US --force-fieldtest=ConnCountImpact/conn_count_6/ConnnectBackupJobs/ConnectBackupJobsEnabled/DnsImpact/default_enabled_prefetch/IdleSktToImpact/idle_timeout_10/Instant/AllControlA/Prerender/ContentPrefetchPrerender2/PrerenderFromOmnibox/OmniboxPrerenderEnabled/ProxyConnectionImpact/proxy_connections_32/SpdyCwnd/cwnd10/SpdyImpact/npn_with_spdy/WarmSocketImpact/warmest_socket/ --extension-process --renderer-print-preview --channel=13160.050BE1C0.836011504 /prefetch:3

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=renderer --lang=en-US --force-fieldtest=ConnCountImpact/conn_count_6/ConnnectBackupJobs/ConnectBackupJobsEnabled/DnsImpact/default_enabled_prefetch/GlobalSdch/global_enable_sdch/IdleSktToImpact/idle_timeout_10/Instant/AllControlA/Prerender/ContentPrefetchPrerender2/PrerenderFromOmnibox/OmniboxPrerenderEnabled/ProxyConnectionImpact/proxy_connections_32/SpdyCwnd/cwnd10/SpdyImpact/npn_with_spdy/WarmSocketImpact/warmest_socket/ --extension-process --renderer-print-preview --channel=13160.050BE540.1809318103 /prefetch:3

C:\Windows\system32\rundll32.exe "C:\Users\rtenneti\DOWNLO~1\CHROME~1\CHROME~1\gcswf32.dll",BrokerMain browser=chrome

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=plugin --plugin-path="C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\gcswf32.dll" --lang=en-US --channel=13160.0B6CAF00.1848515607 --flash-broker=8688 /prefetch:4

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=plugin --plugin-path="C:\Users\rtenneti\AppData\Roaming\Mozilla\plugins\npgoogletalk.dll" --lang=en-US --channel=13160.0B6CA550.1577640654 /prefetch:4

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=service --lang=en-US

"C:\Users\rtenneti\Downloads\chrome-win32\chrome-win32\chrome.exe" --type=gpu-process --channel=13160.0DE14500.1909462736 /prefetch:12

Last night, Eric and I were trying to upgrade from M16 to M18 Beta to test 18.0.1025.11. We could have been running into profile related issues. But today, wasn't able to reproduce it with either M17 (17.0.963.46) or M18 (18.0.1025.11).

thanks,

raman

[Drew Wilson]

FWIW, all background apps should exit when you force a restart via the wrench menu (either "upgrade chrome now" or "exit" menu items) because that should generate a NOTIFICATION_APP_TERMINATING, which BackgroundModeManager turns into a call to BrowserList::EndKeepAlive().

-atw

[pdwiki]

Sorry for the late reply guys.

@Julian - No extensions are running.

@Mike - I unchecked the 'Continue running background apps when Google
Chrome is closed' and it still does not end the final chrome.exe
process. If I view the background processes, I see the Browser process
still running at the end.

@Albert - No, I don't have cloud print service enabled.

@Finnur - Yes, closing Chrome while debugging in Visual Studio does
end the debugger.

If this is fixed in the latest revision, does anyone know exactly
which portion of the code is in use? I am running 15.0.859.0
(Developer Build 97580 Windows) and don't really have the option of
updating the files to the latest version as it would mean a lot of
changes for me.

Thanks,
Prabal.

[Patrick Dubroy]

I opened a bug for the issue of not being able to restart when an update is available: http://crbug.com/113964. It may or may not be the same issue.

[Finnur Thorarinsson]

I am able to simulate Chrome not shutting down (under Visual Studio) by introducing a NULL pointer exception during shutdown. For example, if I add this to the destructor of BrowserActionsContainer (and start and close Chrome)...

   int* i = NULL; *i = 42;

When I do this I see the structured exception handler in wrapped_window_proc.h (way down at the bottom of the stack) kick in. This sequence of events bypasses a whole slew of shutdown code, including the BrowserView destructor, which is responsible for setting browser_ to NULL (which in turn triggers the module_ref_count_ to go to 0, etc). 

This would all be fine if the process terminated as a result of this, yet for some reason it does not. 

Now, one might think that's because our structured exception handler (CrashForException in breakpad_win.cc) is eating the exception or hanging while trying to communicate with BreakPad or something, but commenting it out gives the same result. That makes sense because our handler actually returns EXCEPTION_CONTINUE_SEARCH (because g_breakpad is NULL while debugging in Visual Studio) so the next exception handler should be consulted. And things start to get hard to follow after that.

The end result, though, is incomplete shutdown and Chrome continues to run.

(This all may be a red herring but I thought I should post my partial findings in case someone has any great ideas -- and besides, I needed something to do while rebuilding Chrome from scratch). :)

-F

[Ricardo Vargas]

On Mon, Feb 13, 2012 at 7:50 AM, Finnur Thorarinsson <fin...@chromium.org> wrote:

I am able to simulate Chrome not shutting down (under Visual Studio) by introducing a NULL pointer exception during shutdown. For example, if I add this to the destructor of BrowserActionsContainer (and start and close Chrome)...

   int* i = NULL; *i = 42;

When I do this I see the structured exception handler in wrapped_window_proc.h (way down at the bottom of the stack) kick in. This sequence of events bypasses a whole slew of shutdown code, including the BrowserView destructor, which is responsible for setting browser_ to NULL (which in turn triggers the module_ref_count_ to go to 0, etc). 

This would all be fine if the process terminated as a result of this, yet for some reason it does not. 

Now, one might think that's because our structured exception handler (CrashForException in breakpad_win.cc) is eating the exception or hanging while trying to communicate with BreakPad or something, but commenting it out gives the same result. That makes sense because our handler actually returns EXCEPTION_CONTINUE_SEARCH (because g_breakpad is NULL while debugging in Visual Studio) so the next exception handler should be consulted. And things start to get hard to follow after that.

The end result, though, is incomplete shutdown and Chrome continues to run.

(This all may be a red herring but I thought I should post my partial findings in case someone has any great ideas -- and besides, I needed something to do while rebuilding Chrome from scratch). :)

Red herring. The whole point of that exception handler is to crash the process (if BreakPad is there), 'cause the standard OS behavior is to eat the exception and keep the process alive. So this may affect local debugging, but not a properly installed Chrome.

[Finnur Thorarinsson]

> Red herring. So this may affect local debugging, but not a properly installed Chrome.

Yeah, I worried that was the case. 

OK, then this exercise is meaningless and I've reached a dead-end on this issue. :( 

I was going to file a new bug for the behavior Brian Rakowski ran into, but when searching through the database I found various other such bugs. I marked two Mac specific bugs as a duplicate of this bug:

http://code.google.com/p/chromium/issues/detail?id=62583 ...since it has requests for users to provide more data and users have obliged. The bug is not owned though. Maybe we could get someone with Mac experience to take a look?

[Nico Weber]

On Tue, Feb 14, 2012 at 2:08 AM, Finnur Thorarinsson
<fin...@chromium.org> wrote:
>
>> Red herring. So this may affect local debugging, but not a properly
>> installed Chrome.
>
> Yeah, I worried that was the case.
>
> OK, then this exercise is meaningless and I've reached a dead-end on this
> issue. :(
>
> I was going to file a new bug for the behavior Brian Rakowski ran into, but
> when searching through the database I found various other such bugs. I
> marked two Mac specific bugs as a duplicate of this bug:
> http://code.google.com/p/chromium/issues/detail?id=62583 ...since it has
> requests for users to provide more data and users have obliged. The bug is
> not owned though. Maybe we could get someone with Mac experience to take a
> look?

The bug at the top of this thread was on windows 7. Bug 62583, we had
several bugs that could've caused this on mac back then (since fixed).
It's probably better to file a new bug and close the old one.

Nico

[Finnur Thorarinsson]

Correct. However, since no one has a consistent repro case on Windows and the Mac one has additional data provided I thought it might possibly be of more use. I don't see the point of me adding another bug for Windows if I don't have a repro case.

[Finnur Thorarinsson]

> 'cause the standard OS behavior is to eat the exception and keep the process alive. 

For those (perhaps few?) scratching their heads (like I was), this needs a bit of explanation, which Rick provided offline.

Rick is talking about the standard exception handling of 64-bit Windows specifically, which differs from the world of 32-bit Windows. In 32-bit land, we are used to the process dying on unhandled exceptions, whereas in 64-bit land you'll see one of three different behaviors, only one of which terminates the process after an unhandled exception. :/ 

For details, see the Remarks section here:

http://msdn.microsoft.com/en-us/library/windows/desktop/ms633573(v=vs.85).aspx 

Hmm... From reading the document it looks like in Windows 7 SP1 they are going back to terminating the process (32-bit, 64-bit or  WOW64). Interesting.

On Mon, Feb 13, 2012 at 19:46, Ricardo Vargas <rva...@chromium.org> wrote:

[Finnur Thorarinsson]

It looks like you meant 'Reply all' instead of 'Reply'.

An extra process being created as a result of adding SingleSplitView doesn't seem normal to me, but I don't know that class well so I don't know what it does. From glancing at the implementation I didn't see anything related to Chrome process creation. I guess a more pertinent question is what you are showing inside the split view as that may be creating an extra process.

>  I could not debug to the point where the additional threads are starting...

How come? Because of the multiprocess architecture of Chrome?

There are some tips that might assist with debugging listed here, in case you haven't seen them:

http://www.chromium.org/Home

On Thu, Apr 26, 2012 at 10:10, pdwiki <praba...@gmail.com> wrote:

Well, I am back at this issue after a long time. I think I have
finally found the source of this issue...or at least a lead to the
source. I added an extra SingleSplitView to the browser_view.cc and
made the necessary adjustments in the browser_view_layout.cc file and
the number of chrome.exe's in the task manager shot up by 1 (3 in
total). My curiosity peaked and I added another SingleSplitView and
the number of chrome.exe's became 4. The chrome.exe's are not getting
added at the place where the single split view's are being added to
the layout (at that point only 1 chrome.exe is running) but at some
later stage the chrome.exe's running is always proportional to the
number of split view's running + 1 (for the main chrome.exe).
Oh...initially there are 2 chrome.exe's running because of the
devtools split view. I could not debug to the point where the
additional threads are starting...but I believe this is what is
happening. So if someone with a little more insight on this portion
could help me out with solving this issue then it will be very helpful
for me.

Thanks.

On Feb 14, 10:55 pm, Finnur Thorarinsson <fin...@chromium.org> wrote:
> Correct. However, since no one has a consistent repro case on Windows and
> the Mac one has additional data provided I thought it might possibly be of
> more use. I don't see the point of me adding another bug for Windows if I
> don't have a repro case.
>
>
>
>
>
>
>

> On Tue, Feb 14, 2012 at 16:25, Nico Weber <tha...@chromium.org> wrote:
> > On Tue, Feb 14, 2012 at 2:08 AM, Finnur Thorarinsson

> > <fin...@chromium.org> wrote:
>
> > >> Red herring. So this may affect local debugging, but not a properly
> > >> installed Chrome.
>
> > > Yeah, I worried that was the case.
>
> > > OK, then this exercise is meaningless and I've reached a dead-end on this
> > > issue. :(
>
> > > I was going to file a new bug for the behavior Brian Rakowski ran into,
> > but
> > > when searching through the database I found various other such bugs. I
> > > marked two Mac specific bugs as a duplicate of this bug:

> > >http://code.google.com/p/chromium/issues/detail?id=62583...since it has

> > > requests for users to provide more data and users have obliged. The bug
> > is
> > > not owned though. Maybe we could get someone with Mac experience to take
> > a
> > > look?
>
> > The bug at the top of this thread was on windows 7. Bug 62583, we had
> > several bugs that could've caused this on mac back then (since fixed).
> > It's probably better to file a new bug and close the old one.
>
> > Nico
>

> > > On Mon, Feb 13, 2012 at 19:46, Ricardo Vargas <rvar...@chromium.org>

> > >>>> On Sat, Feb 11, 2012 at 2:20 PM, pdwiki <prabal....@gmail.com> wrote:
>
> > >>>>> Sorry for the late reply guys.
>
> > >>>>> @Julian - No extensions are running.
>
> > >>>>> @Mike - I unchecked the 'Continue running background apps when Google
> > >>>>> Chrome is closed' and it still does not end the final chrome.exe
> > >>>>> process. If I view the background processes, I see the Browser
> > process
> > >>>>> still running at the end.
>
> > >>>>> @Albert - No, I don't have cloud print service enabled.
>
> > >>>>> @Finnur - Yes, closing Chrome while debugging in Visual Studio does
> > >>>>> end the debugger.
>
> > >>>>> If this is fixed in the latest revision, does anyone know exactly
> > >>>>> which portion of the code is in use? I am running 15.0.859.0
> > >>>>> (Developer Build 97580 Windows) and don't really have the option of
> > >>>>> updating the files to the latest version as it would mean a lot of
> > >>>>> changes for me.
>
> > >>>>> Thanks,
> > >>>>> Prabal.
>
> > >>>>> --

> > >>>>> Chromium Developers mailing list: chromium-...@chromium.org

> > >>>>> View archives, change email options, or unsubscribe:
> > >>>>>    http://groups.google.com/a/chromium.org/group/chromium-dev
>
> > >>>> --

> > >>>> Chromium Developers mailing list: chromium-...@chromium.org

> > >>>> View archives, change email options, or unsubscribe:
> > >>>>http://groups.google.com/a/chromium.org/group/chromium-dev
>
> > >>> --

> > >>> Chromium Developers mailing list: chromium-...@chromium.org

> > >>> View archives, change email options, or unsubscribe:
> > >>>http://groups.google.com/a/chromium.org/group/chromium-dev
>
> > > --

> > > Chromium Developers mailing list: chromium-...@chromium.org

[Prabal]

Yeah, I meant to hit Reply All :)

By 'I could not debug to the point where the additional threads are starting' I actually wanted to mean that I could not link up the relation between the split views and the additional threads. I must be missing something (still relatively new at this). These additional threads are the ones which do not exit upon exiting chrome (right now I have 3 chrome.exe's running after exiting). Still trying to figure out the relation...just wanted to get the word out.

[Finnur Thorarinsson]

I'm wondering if you can maybe upload the changes you made, over to Rietveld, so we can try to repro?

http://www.chromium.org/developers/contributing-code#TOC-Create-a-change-list-CL-

[AFjetpilot]

The Google Chrome doesn't close when you exit, and it causes big problems.

When you start Chrome again, because you don't know it's running in the backround, you get nothing but trouble.

Pages don't load, pages lock up, you get the warning that the page crashed, the Ah Snap bad news warning.

Chrome doesn't work with Verizon.com Central.  Google Chrome has been broken for ovver a month.

It just doen't work anymore, and nobody fixes it.  I don't like IE, but I will switch back to it, at least it works.

Too bad, Google Chrome is a wonderful program, but it's broke and they can't seem to fix it.

[Ed Wilson]

I tried unchecking the box that allows chrome to run in the background. but it still keeps a process running after i close chrome. and re opens after it is terminated.

It has been taking control of my camera and microphone, and requires them to be running at all times.

Pretty much every time i try to use my camera it tells me it is in use by another program. I.e. Chrome.

Know of any reasons why chrome needs to have my camera and mic on 24/7?

[Bill Mayhew]

Why does this problem (or, of course, one that "looks just like it", i.e. exhibits the same failure mode)  STILL exist, years later?

Using chrome updated within the last week - it STILL refuses to exit when I close all windows.

Platform: Win 7 SP1

(looked at recent posts and saw no solutions posted, perhaps in another thread somewhere)

[PhistucK]

It depends on your settings (cloud print, background applications, perhaps also push notifications? Not sure).

☆PhistucK

--
--
Chromium Developers mailing list: chromi...@chromium.org

View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev

---
You received this message because you are subscribed to the Google Groups "Chromium-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev+unsubscribe@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/def4f4a7-b1c1-4622-88b0-222909b6e10c%40chromium.org.

[Philippe Verdy]

Look into the system tray, you've got an Google Chrome icon that allows you to enable or disable this background process. If you disable it, you'll get a warning saying which services will stop working. That icon also includes in its menu a "Quit" option that you can use at any time to stop all background services... Use it, you won't have any chrome.Exe process running.

So clearly this is working as expected and I don't see any bug, and this is effective since years, even if you did not notice it, probably you never look at what is in your system tray, you may have various other tools running there that you did not notice and that you may need to tune: most of these notification icons include an option to exit their process.

Note however that Google Drive has a separate icon for its background service and to quit it, you need to open the window to locate the menu in the top corner (same thing for Dropbox or MS OneDrive): some of these tools run their own local instance of Chrome for displaying their interface. Same thing also for Steam (that also embeds its own separate Chrome instance). Some of these tools are using a Python instance instead of a browser (installed in the user's temporary folder and deleted when you exit these tools): Python is much lighter than a browser (in storage space), but it is often slower than modern Javascript engines in browsers and may sometime use more memory for the VM they are running, because the code is compiled on demand and often interpreted, but less efficiently than Javascript engines of modern browsers. Some newer tools use a VM to run Lua scripts (which is even lighter than Python) or PHP scripts. As the performance cost is huge when these VM are starting up, generally the VM is initialized and will run in the background, to be instantly ready to reply to user interaction.

A Chrome instance is also sometimes used to render the interface of some antivirus/security tools, or for some media players or media providers (e.g. Spotify). These specific instances of Chrome-based browsers are installed separately, with their own local version and their own settings of the specific plugins they need (these installations do not share any common process, unlike Adove Flash or IE and Edge which are intrusive and allow unsecure interprocess communications.

In Windows 8/8.1/10, you've got many "UWX" processes running in the background and preinitialized within a .Net VM: they take lot of virtual memory, but will stay idely for long and will swap to disk (in a new "swapfile" distinct from the paging file used for generic processes: they can be shutdown completely and restarted using their separate application-specific data store for the current user). This is the case for all apps that have interactive tiles in the start menu or in the logon screen (you may configure these apps to use static tiles instead if you want to save many background processes)

So what Chrome is doing is not exclusive, many other programs are doing it already, including the OS itself for many of its "services".

--
--
Chromium Developers mailing list: chromi...@chromium.org
View archives, change email options, or unsubscribe:
http://groups.google.com/a/chromium.org/group/chromium-dev
---

You received this message because you are subscribed to a topic in the Google Groups "Chromium-dev" group.
To unsubscribe from this topic, visit https://groups.google.com/a/chromium.org/d/topic/chromium-dev/0rKJn87f8_0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to chromium-dev+unsubscribe@chromium.org.

[Bill Mayhew]

Pure vanilla.  No Google apps running other than the browser.

[Bill Mayhew]

The Chrome icon in my system tray has NO SUCH MENU OPTIONS... its only local functions are New window, new incognito window, unpin, close.  (As mentioned, my platform is Windows 7.)

[PhistucK]

You probably mean the task bar and not the system tray (the small icons next to the time).

☆PhistucK

You received this message because you are subscribed to the Google Groups "Chromium-dev" group.

To unsubscribe from this group and stop receiving emails from it, send an email to chromium-dev+unsubscribe@chromium.org.

To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/chromium-dev/CAOziKqoS9groWC%2BuYE594V2x7j%3DFQ%2BuZQq0_W3YoqZk5fnzSCw%40mail.gmail.com.

[Bill Mayhew]

Good point; on my notebook, the task bar and system tray may as well be one "thing." 

Backing up then - I don't recall ever seeing a Google icon in the system tray.

[PhistucK]

Have you verified that this chrome.exe is Chrome and not some rogue Chromium (the open source version of Chrome that is sometimes used by malware) or other rogue software?

chrome.exe is usually located at -

C:\Users\USERNAME\AppData\Local\google\chrome\application\chrome.exe (where USERNAME is your user name)

Or

C:\Program Files (x86)\Google\Chrome\Application\chrome.exe

☆PhistucK

[Bill Mayhew]

Yes, it's version 58.0.3029.81 in 

Program Files (x86)/Google/Chrome/Application

[PhistucK]

Does it happen a lot? If so -

You can search crbug.com for an existing issue and star it. If you cannot find one, file a new issue using the "New issue" link on the same page.

Please, do not add a "+1" or "Me too" or "Confirmed" (or similar) comment. It just wastes the time of Chrome engineers and sends unnecessary e-mails to all of the people who starred the issue.

You can reply with a link to the found or created issue and might get triaged (and fixed) faster.

Thank you.

☆PhistucK

[Philippe Verdy]

No I REALLY mean the small icons in the system tray, NOT icons in the task bar !!!

There's a Google icon there, if you've got background processes enabled.

[PhistucK]

Yep, I know, I was correcting Bill. :)

☆PhistucK

[James G Needham]

Terminate Google Chrome files when exiting:
Go to Settings...
Page down to Advanced Settings...
Page down to System...
Toggle "Continue running background apps when Google Chrome is closed" to "Off" (Gray shading).

[ghislain mega lutumba]

for more information add to the location in the Chrome iOS project go to www.kivumakers.com