Issue 195732 in chromium: SetPrinterInfo eating up CPU

[chro...@googlecode.com]

Status: Unconfirmed
Owner: ----
Labels: Pri-2 Via-Wizard Type-Bug-Regression OS-Windows

New issue 195732 by earthso...@gmail.com: SetPrinterInfo eating up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

UserAgent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.33 (KHTML,
like Gecko) Chrome/27.0.1440.0 Safari/537.33

Steps to reproduce the problem:
1. Load a large page (such as
http://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?path=Win/)

What is the expected behavior?
The page renders after loading within a reasonable amount of time.

What went wrong?
The parent process of chrome starts eating up a good chunk of all 4
processors (~25% in total available CPU) and chromium is sluggish while
this is happening. It can last anywhere from 45 seconds to 2 minutes.

Using Process Hacker, I can see that this thread is the culprit:

Thread ID, CPU, Cycles Delta, Start Address, Priority:
2788, 24.61, 2,094,845,780, chrome.exe!SetPrinterInfo+0x48961, Normal

Inspecting it shows this:
0, ntoskrnl.exe!_misaligned_access+0x17e7
1, ntoskrnl.exe!KeAcquireSpinLockAtDpcLevel+0x93d
2, ntoskrnl.exe!KeWaitForSingleObject+0x19f
3, ntoskrnl.exe!_misaligned_access+0xba4
4, ntoskrnl.exe!_misaligned_access+0x1821
5, ntoskrnl.exe!_misaligned_access+0x1a97
6, chrome.dll!ovly_debug_event+0x111a124
7, 0x18e6b417780810
8, 0x2e105803eb8fd5
9, 0x3ebaf960018e6cc
10, 0x400000030
11, 0x1556700000000013
12, 0x3ebb6920018e6d8
13, 0x18e6e400000026
14, 0x2603ebb76c
15, 0x3ebb7de0018e6f4
16, 0x26
17, 0x2f868c00018e71c
18, 0x109af00000000026
19, 0x155670000c7dd2c8
20, 0x13
21, 0x1850797a00000002
22, 0x2f8978f0018e86c
23, 0x27f000200000e7f
24, 0x15567000109af000
25, 0xde0f6d80e6c2a3c
26, 0xcaea3000000000d
27, 0xe6c2a3c00000000
28, 0xde087e00e6c2a44
29, 0x200e6c2a20
30, chrome.dll!ovly_debug_event+0x111a124
31, chrome.dll!ovly_debug_event+0x1115925
32, chrome.dll!ovly_debug_event+0x11178e6
33, chrome.dll!ovly_debug_event+0x1117fe2
34, chrome.dll!ovly_debug_event+0x11180bc
35, chrome.dll!ovly_debug_event+0x111812e
36, chrome.dll!ovly_debug_event+0x1e3210
37, chrome.dll!ovly_debug_event+0x1e60df
38, chrome.dll!ovly_debug_event+0x1e64e4
39, chrome.dll!ovly_debug_event+0x1e6896
40, chrome.dll!ovly_debug_event+0x1d6ff1
41, chrome.dll!ovly_debug_event+0x1b5b85
42, chrome.dll!ovly_debug_event+0x1d73df
43, chrome.dll!ovly_debug_event+0xed2b44
44, chrome.dll!ovly_debug_event+0xed35c3
45, chrome.dll!ovly_debug_event+0xed36f6
46, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49b0f3
47, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49c23c
48, chrome.dll!ovly_debug_event+0xe7c80e
49, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49b10b
50, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49c23c
51, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49e953
52, chrome.dll!ovly_debug_event+0xe983df
53, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49b10b
54, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49c23c
55, chrome.dll!ovly_debug_event+0xdeba2a
56, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49b10b
57, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49c23c
58, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49e953
59, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49b10b
60, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49c23c
61, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49e953
62, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49b10b
63, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x49c23c
64, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x4b9d13
65, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x4bc25b
66, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x4bc3c9
67, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x4bb227
68, chrome.dll!ovly_debug_event+0x1dbadb
69, chrome.dll!ovly_debug_event+0x1dbb40
70, user32.dll!gapfnScSendMessage+0x332
71, user32.dll!GetThreadDesktop+0xd7
72, user32.dll!GetClientRect+0xc5
73, user32.dll!CallWindowProcW+0x1b
74, chrome.dll!ovly_debug_event+0x17749f9
75, chrome.dll!ovly_debug_event+0x1774a3d
76, user32.dll!gapfnScSendMessage+0x332
77, user32.dll!GetDC+0x52
78, user32.dll!GetThreadDesktop+0x185
79, user32.dll!GetThreadDesktop+0x1e1
80, ntdll.dll!KiUserCallbackDispatcher+0x2e
81, user32.dll!DispatchMessageW+0xf
82, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x4a00eb
83, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x5000fc
84, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x501298
85, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x4ff6b0
86, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x4e6d8e
87, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x50c6d3
88, chrome.dll!ovly_debug_event+0xf6f2eb
89, chrome.dll!ovly_debug_event+0xf6fdbe
90, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x3df73a
91, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x3df92d
92, chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x3dc7b0
93, chrome.dll!ChromeMain+0x20
94, chrome.exe!SetPrinterInfo+0x39c9
95, chrome.exe!SetPrinterInfo+0x3a98
96, chrome.exe!SetPrinterInfo+0x3b16
97, chrome.exe!SetPrinterInfo+0x4890e
98, kernel32.dll!BaseThreadInitThunk+0x12
99, ntdll.dll!RtlInitializeExceptionChain+0x63
100, ntdll.dll!RtlInitializeExceptionChain+0x36

Did this work before? Yes I've noticed this for some time, but
unfortunately I cannot give an accurate estimate of how long it's been. At
least weeks, if not months.

Chrome version: 27.0.1440.0 build 188004 Channel: n/a
OS Version: 6.1 (Windows 7, Windows Server 2008 R2)

I also notice something like this when working inside a Joomla admin
interface for a while. Eventually, the process for the tab eats up a lot of
memory and each time a new page is loaded it goes through this. I usually
get around it while editing inside Joomla by just closing the tab & logging
back in which clears out whichever memory leak(s) have taken place over
time. However, when loading large data (like the snapshots folder URL),
this isn't an option.

This is a 64-bit Windows 7 machine with 8 GB of RAM with a Xeon E5606 CPU
and chromium 27.0.1440.0 (build 188004).

--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings

[chro...@googlecode.com]

Updates:
Cc: rponn...@chromium.org
Labels: Needs-Feedback

Comment #1 on issue 195732 by rponn...@chromium.org: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

Tested this issue Windows7 & 27.0.1440.0 (Official Build 187987) . Unable
to repro this issue. While opening the mentioned URL, page opens up within
30 sec & not observed any high memory usage .

@earthsound: Kindly re-test and let us know the behavior. If issue still
exists, kindly provide screen cast will help to triage further.

[chro...@googlecode.com]

Comment #2 on issue 195732 by eric.zei...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

I can repro this on Win7 with (27.0.1453.12 dev-m). Doesn't use lots of
memory, but the thread with SetPrinterInfo does sit at 6% CPU on my machine
until chrome is restarted.

[chro...@googlecode.com]

Comment #3 on issue 195732 by lucifer6...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

I have a similar issue, there is a
thread "chrome.exe!SetPrinterInfo+0x5950a" using up 25% of the CPU.

After checking various extensions, I have narrowed it down. The extension
works fine on the site it was designed for but on some other URLs causes
this error. There is an "Uncaught TypeError" in the Javascript but the
extension has been unchanged for months.

It is only since the latest update that Chrome has been suffering this
problem. Could it be a problem in the Javascript parser or error handler?

Currently running Win7 64bit with Version 26.0.1410.43 m

[chro...@googlecode.com]

Comment #4 on issue 195732 by halil.oz...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

I can confirm this v. 26.0.1410.43 m on Win7 64bit in both home and work
machine. It doesn't eat a lot (~ 1-5% CPU) but it's nearly always there.
Inspection with Process Explorer reveals:

Thread:
chrome.exe!SetPrinterInfo+0x5950a

Stack:
KERNEL32.dll!GetConsoleMode+0xb4
chrome.dll!_ovly_debug_event+0x18b06f
chrome.dll!_ovly_debug_event+0x14d5c
chrome.dll!_ovly_debug_event+0x14cc7
chrome.dll!_ovly_debug_event+0x4d7341
chrome.dll!_ovly_debug_event+0x4d727a
chrome.dll!_ovly_debug_event+0x635ac
chrome.dll!ChromeMain+0xf22b
chrome.dll!ChromeMain+0xf1b2
chrome.dll!ChromeMain+0x32c
chrome.dll!ChromeMain+0x1e
chrome.exe+0x26b83
chrome.exe!SetPrinterInfo+0x76a
chrome.exe!SetPrinterInfo+0x7d5
chrome.exe!SetPrinterInfo+0x594b7
KERNEL32.dll!BaseThreadInitThunk+0x12
ntdll.dll!RtlInitializeExceptionChain+0x63
ntdll.dll!RtlInitializeExceptionChain+0x36

Also From time to time I'm having serious issues with ovly_debug_event. It
sometimes (as a separate thread) climbs up to ~50% CPU. I don't know the
relation of it to SetPrinterInfo (other than it shows up in the above stack
too) but after investigating a little bit more I'm thinking of opening a
separate issue about it.

[chro...@googlecode.com]

Comment #5 on issue 195732 by halil.oz...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

Note that I'm experiencing the issues I've mentioned regardless of the
pages I'm opening. It happens every day. Somedays I have to kill
ovly_debug_event thread in the second issue (and occasionally
SetPrinterInfo in the first issue), sometimes causing Chrome to crash,
sometimes not.

[chro...@googlecode.com]

Comment #6 on issue 195732 by m4des...@gmail.com: SetPrinterInfo eating up
CPU
http://code.google.com/p/chromium/issues/detail?id=195732

I just started noticing this issue, it's eating about 10% of CPU
continuously! The page responsible for this is:

http://blog.pandurangpatil.com/2013/01/intercept-gwt-rpc-request-and-response.html

[chro...@googlecode.com]

Comment #11 on issue 195732 by brendan....@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

I too am having this issue with version 27.0.1453.116 m

[chro...@googlecode.com]

Comment #12 on issue 195732 by arevin...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

Please fix it. I am used to keep many tabs open while working when it
hangs, it takes quite a while to restore normal operation.

[chro...@googlecode.com]

Comment #13 on issue 195732 by arevin...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

Additional data:
Screenshot: http://gyazo.com/505aab332d8b44b72920b12ddcdcfc51

Copy of the stacktrace for this CPU-eating Thread:

ntoskrnl.exe!KeWaitForMultipleObjects+0xc0a
wow64cpu.dll!TurboDispatchJumpAddressEnd+0x6c0
wow64cpu.dll!TurboDispatchJumpAddressEnd+0x1fb
wow64.dll!Wow64SystemServiceEx+0x1ce
wow64.dll!Wow64LdrpInitialize+0x429
ntdll.dll!RtlIsDosDeviceName_U+0x24c87
ntdll.dll!LdrInitializeThunk+0xe
chrome.dll+0x3004
chrome.dll!_ovly_debug_event+0x1f694
chrome.dll!ChromeMain+0x15317
chrome.dll!_ovly_debug_event+0x1c03b5
chrome.dll!_ovly_debug_event+0x6ed3b
chrome.dll!_ovly_debug_event+0x1c0286

No page gets to load. If I try to load a page, the tab only says 'Loading..'

[chro...@googlecode.com]

Comment #14 on issue 195732 by zhixiang...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

Same problem here. It makes me crazy because it hangs every day! Even in
the incognito mode and disabled all extensions/plugins.
The page chrome://tracing/ fail to open when it hangs.

Here is my thread info:
ntoskrnl.exe!memset+0x82f

wow64cpu.dll!TurboDispatchJumpAddressEnd+0x6c0
wow64cpu.dll!TurboDispatchJumpAddressEnd+0x1fb
wow64.dll!Wow64SystemServiceEx+0x1ce
wow64.dll!Wow64LdrpInitialize+0x429
ntdll.dll!RtlIsDosDeviceName_U+0x24c87
ntdll.dll!LdrInitializeThunk+0xe

chrome.dll+0x6da1
chrome.dll!_ovly_debug_event+0x1e1ea0
chrome.dll!_ovly_debug_event+0x1e1d7d
chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x3aafa8
chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x3ab0bc
chrome.dll!_ovly_debug_event+0x180cac
chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x353b99
chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x353d45
chrome.dll!RelaunchChromeBrowserWithNewCommandLineIfNeeded+0x353e09
chrome.dll!_ovly_debug_event+0x1c0435
chrome.dll!_ovly_debug_event+0x14446
chrome.dll!_ovly_debug_event+0x13e49
chrome.dll!_ovly_debug_event+0x18d26
chrome.dll!_ovly_debug_event+0x18c34
chrome.dll!_ovly_debug_event+0x13a05
chrome.dll!_ovly_debug_event+0x1395d
chrome.dll!_ovly_debug_event+0x18be4
chrome.dll!_ovly_debug_event+0x16639f
chrome.dll!_ovly_debug_event+0x1182a5
chrome.dll!_ovly_debug_event+0x18a60
chrome.dll!_ovly_debug_event+0x1897d
ntdll.dll!RtlInitializeExceptionChain+0x63
ntdll.dll!RtlInitializeExceptionChain+0x36

[chro...@googlecode.com]

Comment #15 on issue 195732 by arevin...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

Who is in charge here? Is anyone even listening?!

[chro...@googlecode.com]

Comment #17 on issue 195732 by halil.oz...@gmail.com: SetPrinterInfo eating
up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

@ligimole it's not always apparent, as it can be quite sporadic (at least
form me). I'll try to send feedback on new versions.

Anyway, multiply this for 10-20 open tabs and you suddenly have a sizable
CPU usage (10-20%) for nothing (ie I don't even have a printer installed).

[chro...@googlecode.com]

Updates:
Labels: Cr-Internals-Printing

Comment #19 on issue 195732 by to...@chromium.org: SetPrinterInfo eating up
CPU
http://code.google.com/p/chromium/issues/detail?id=195732

(No comment was entered for this change.)

[chro...@googlecode.com]

Comment #21 on issue 195732 by ema....@gmail.com: SetPrinterInfo eating up
CPU
http://code.google.com/p/chromium/issues/detail?id=195732

This happens for me as well.

The most noticing instance is on Facebook. Looking at some long pages of
pictures, those pages which load in segments using xmlhttp, the more I load
the more it gets tough and I detect noticeable rendering slowdowns as well
as SetPrinterInfo at page load it's 40-50%, during idle it's 4-5%, and
several ovly_debug_event.

Besides this issue, but maybe it's related, I always have the main
Chrome.exe and the tabs exe's Read and Write delta at 500B to 2KB each
second without even accessing the browser. Just letting it sit in idle and
I get so much IO workload.

[chro...@googlecode.com]

Updates:
Labels: -Cr-Internals-Printing

Comment #23 on issue 195732 by vitalyb...@chromium.org: SetPrinterInfo

eating up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

SetPrinterInfo is just last public symbol in chrome.exe binary so tools
display location as reference to that point.

Can you create crash dump when you see this next time?

[chro...@googlecode.com]

Comment #24 on issue 195732 by vitalyb...@chromium.org: SetPrinterInfo

eating up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

for #22 it's hangs somewhere at
(1ecc0) chrome!base::internal::JSONParser::StringBuilder::Convert+0x46
(1ed10) chrome!base::internal::JSONParser::StringBuilder::AsString

[chro...@googlecode.com]

Comment #29 on issue 195732 by vitalyb...@chromium.org: SetPrinterInfo

eating up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

No problem. Thanks for update.

[chro...@googlecode.com]

Comment #30 on issue 195732 by ema....@gmail.com: SetPrinterInfo eating up
CPU
http://code.google.com/p/chromium/issues/detail?id=195732

True, but I don't use that extension... it might not be the extension
itself but the API/dev-kit it uses or some internal function which, from
what it looks like, is used by multiple extensions.

So this is when I see lots of activity, mouse stuttering, freezing, partial
rendering if I scroll down (the rest of the page gets painted over
checkers), high CPU usage, but it's not the instance where it stays and
takes up constantly 5%, more or less. For this, I would have to browse for
some time to trigger it. Next time, if I spot it.



Attachments:
chrome (facebook).dmp 2.7 MB
chrome (main).dmp 6.0 MB

[chro...@googlecode.com]

Comment #31 on issue 195732 by ema....@gmail.com: SetPrinterInfo eating up
CPU
http://code.google.com/p/chromium/issues/detail?id=195732

Did the issue drop in interest?

[chro...@googlecode.com]

Updates:
Labels: Restrict-AddIssueComment-EditIssue

Comment #32 on issue 195732 by vitalyb...@chromium.org: SetPrinterInfo

eating up CPU
http://code.google.com/p/chromium/issues/detail?id=195732

This issue was closed. Please file new issues with dumps if you believe
that something wrong with your Chrome instance. I understand that it could
be hard, but without dump that shows such behavior we can't fix this.

I see nothing bad in these dumps.