+Ligi this may be the same issue as we've been seeing in other bugs, but it would be worth trying to repro on a Win 10 device.

I have USB port with  version 1.1in my Win 7 system, but do not have any problem in launching chrome version - 54.0.2840.59.

Lopping to Reilly , who is the the appropriate DEV for taking a look into this issue.

Tagging with RB Stable to get proper attention.

I'm the dev to look into the issue. If this issue is with vendor USB 3.0 drivers then it should be resolved on Win 10 but show up on random Win 7 machines depending on which vendor driver is installed. USB 3.0 support on Win 7 is an utter mess. :(

The shutdown issue is likely the same as  issue 652416  that jam@ reported last week. I'll take a look at his machine to see if I can find a workaround.

If we can't find a workaround we can prevent initialization of the USB subsystem in Chrome instances that don't need it for the chrome.usb API by turning off the WebUSB feature flag:

https://cs.chromium.org/chromium/src/chrome/browser/chrome_browser_main.cc?sq=package:chromium&rcl=1476702996&l=1534

Thanks for the update. We are planning a full Stable RC cut tomorrow Tuesday (10/18)@ 4.00 PM PST. 

Wondering whether turning off the WebUSB feature flag can be done from server side?Will there be any consequences if we do the same?

Thanks for investigation.

We're going to have M54 Stable Release on Wednesday (OCT 19).Please ensure to land the fix today before 8.00 PM to have enough baking time in trunk. If all looks good CL can be merged in time for tomorrow's stable cut at 4 PM (OCT 18).

I am in the process of disabling WebUSB on M-54 using Finch. In the meantime I can land a patch to disable WebUSB entirely or to disable only the part that is triggering this issue for so many users.

Since the code causing this issue has been part of Chrome for so long and is being exercised by any user with a Chrome App using the chrome.usb API I feel comfortable leaving WebUSB available for developers to play with given that users will rarely encounter a site that triggers this. It is the fact that this is being enabled on Chrome startup for everyone that has caused such a spike in reports.

Opinions?

Just using Finch sounds good to me.  It may cause some unintended side effects to disable WebUSB completely, so I'd rather just disable it by default for M-54.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/dfd32df76aea1a2d4f1aad69c74ded742a2ac69b

commit dfd32df76aea1a2d4f1aad69c74ded742a2ac69b
Author: reillyg <reillyg@chromium.org>
Date: Tue Oct 18 02:42:16 2016

Disable WebUSB detector logic on Windows.

Initializing the WebUSB device detector logic causes the USB subsystem
to attempt to fully enumerate the system's USB bus. This has been found
to cause jank and hangs on Chrome's FILE thread on Windows. Disabling
this component of WebUSB while the issue is investigated.

WebUSB will continue to function but this change will prevent the issue
from impacting users who do not visit sites using it.

BUG= 656702 
TEST=Verified USB subsystem is not initialized on startup on Windows by
checking chrome://device-log. Verified still initialized on Linux.

Review-Url: https://codereview.chromium.org/2422403003
Cr-Commit-Position: refs/heads/master@{#425867}

[modify] https://crrev.com/dfd32df76aea1a2d4f1aad69c74ded742a2ac69b/chrome/browser/usb/web_usb_detector.cc
[modify] https://crrev.com/dfd32df76aea1a2d4f1aad69c74ded742a2ac69b/chrome/browser/usb/web_usb_detector_unittest.cc

Thanks for the fix, unfortunately tonights canary- 56.0.2891.0 missed the above CL. Will trigger a new one for verifying the fix.

Marking as fixed to trigger processing of the merge requests. I will track fixing the underlying problem in  issue 652416  because that includes more technical detail.

Thanks for the quick turnaround. 

Tagging  issue 652416  with appropriate labels so that it will be in our triaging queue.

Merge approved for M54

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/2ac51a5dd890c0944f4d689792319426fb5f9850

commit 2ac51a5dd890c0944f4d689792319426fb5f9850
Author: Reilly Grant <reillyg@chromium.org>
Date: Tue Oct 18 22:17:45 2016

Disable WebUSB detector logic on Windows.

Initializing the WebUSB device detector logic causes the USB subsystem
to attempt to fully enumerate the system's USB bus. This has been found
to cause jank and hangs on Chrome's FILE thread on Windows. Disabling
this component of WebUSB while the issue is investigated.

WebUSB will continue to function but this change will prevent the issue
from impacting users who do not visit sites using it.

BUG= 656702 
TEST=Verified USB subsystem is not initialized on startup on Windows by
checking chrome://device-log. Verified still initialized on Linux.

Review-Url: https://codereview.chromium.org/2422403003
Cr-Commit-Position: refs/heads/master@{#425867}
(cherry picked from commit dfd32df76aea1a2d4f1aad69c74ded742a2ac69b)

Review URL: https://codereview.chromium.org/2425903003 .

Cr-Commit-Position: refs/branch-heads/2840@{#751}
Cr-Branched-From: 1ae106dbab4bddd85132d5b75c670794311f4c57-refs/heads/master@{#414607}

[modify] https://crrev.com/2ac51a5dd890c0944f4d689792319426fb5f9850/chrome/browser/usb/web_usb_detector.cc
[modify] https://crrev.com/2ac51a5dd890c0944f4d689792319426fb5f9850/chrome/browser/usb/web_usb_detector_unittest.cc

Looking to merge this to M-55 as well.

Approving merge to M55 branch 2883. Please merge ASAP. Thank you.

Tested the issue on windows 7 USB 3.0 with using chrome version 54.0.2840.68 with the below steps

1.Installed and Launch chrome
2.Go to URL chrome://device-log
3.No log entries
4.Exit and relaunch chrome again
5.Chrome opened without any any performance issue

reillyg@ Please find the attached screen cast and provide us the steps to verify the issue from test team end.

Note :: Also tested the issue on reported version 54.0.2840.59 observed USB entries under chrome://device-log for the first chrome launch.After restart not observed any logs.

Thanks,

Deleted: 54.0.2840.68 .mp4 1.5 MB

	54.0.2840.68 .mp4 1.5 MB View Download

Deleted: 54.0.2840.59 .mp4 2.9 MB

	54.0.2840.59 .mp4 2.9 MB View Download

Hi guys,
I'm continuing the investigations with users:
I suggested to test --disable-webusb-security in a shortcut (just because it's the only webusb thing I could find).

Obviously, we can't test Chrome's behavior when chrome://flags/#enable-webusb is turned off, because one can't reach flags when Chrome crashes at its start.

The main solution, to update USB 3.x drivers, remains the most effective, but not for all. Some admins complain on this issue and mention devices even without USB 3.x at all (ASUS H61MK). Asked them to update chipset drivers.
https://productforums.google.com/forum/?utm_medium=email&utm_source=footer#!topic/chrome-ru/MMnAsHRKAiw;context-place=topicsearchin/chrome-ru/800%7Csort:date

pisarevden@, to disable this feature on the command line add: --disable-features=WebUSB

kavvaru@, WebUSB was disabled via Finch for 54.0.2840.59 so after a restart (when Chrome gets a chance to update its experiment configuration) you won't see any log entries in chrome://device-log.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b3da8702f5e857a68eb081ea253120120cfccb66

commit b3da8702f5e857a68eb081ea253120120cfccb66
Author: Reilly Grant <reillyg@chromium.org>
Date: Wed Oct 19 16:52:16 2016

Disable WebUSB detector logic on Windows.

Initializing the WebUSB device detector logic causes the USB subsystem
to attempt to fully enumerate the system's USB bus. This has been found
to cause jank and hangs on Chrome's FILE thread on Windows. Disabling
this component of WebUSB while the issue is investigated.

WebUSB will continue to function but this change will prevent the issue
from impacting users who do not visit sites using it.

BUG= 656702 
TEST=Verified USB subsystem is not initialized on startup on Windows by
checking chrome://device-log. Verified still initialized on Linux.

Review-Url: https://codereview.chromium.org/2422403003
Cr-Commit-Position: refs/heads/master@{#425867}
(cherry picked from commit dfd32df76aea1a2d4f1aad69c74ded742a2ac69b)

Review URL: https://codereview.chromium.org/2440463002 .

Cr-Commit-Position: refs/branch-heads/2883@{#187}
Cr-Branched-From: 614d31daee2f61b0180df403a8ad43f20b9f6dd7-refs/heads/master@{#423768}

[modify] https://crrev.com/b3da8702f5e857a68eb081ea253120120cfccb66/chrome/browser/usb/web_usb_detector.cc
[modify] https://crrev.com/b3da8702f5e857a68eb081ea253120120cfccb66/chrome/browser/usb/web_usb_detector_unittest.cc

reillyg@ Thanks for the update.Adding TE-Verified labels for 54.0.2840.68 

Verified the issue on windows 7 USB 3.0 with using chrome version 55.0.2883.21.Not observed any log entries under chrome://device-log after chrome launch and restart.
Please find the attached screen shot for the same.
Adding Verified labels.

Deleted: 656702.png 72.2 KB

	656702.png 72.2 KB View Download

Some of users are enterprise admins and complaining that they have 800+ PCs affected. If sum couple of amounts they mentioned this will give us a value around 4000 PCs affected and counting.
Based on user feedback I can say the following:
1. Disable WebUSB detector logic on Windows.
did not help.
2. --disable-features=WebUSB
did not help.
3. Canary
has the same issue.
4. Some of affected PCs do not even have USB 3.x on board!

All of them are asking to fix this ASAP and to mention ETA, asking for more solutions, ready to test other workarounds.
Some of them are replacing chrome.exe with a M53 one and this helps for a time.

Please write questions for them and give me more possible solutions and ETA.

And yes, 5. crash reports about this issue are not saving

And obviously we have the same issue with Chrome for Work:
https://productforums.google.com/forum/?utm_source=digest&utm_medium=email#!topic/chrome/7_OqANDEnV0;context-place=forum/chrome

Found a thread which could help you with troubleshooting:
https://productforums.google.com/forum/?utm_source=digest&utm_medium=email#!topic/chrome/QgFp-GNoMsI;context-place=forum/chrome

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b3da8702f5e857a68eb081ea253120120cfccb66

commit b3da8702f5e857a68eb081ea253120120cfccb66
Author: Reilly Grant <reillyg@chromium.org>
Date: Wed Oct 19 16:52:16 2016

Disable WebUSB detector logic on Windows.

Initializing the WebUSB device detector logic causes the USB subsystem
to attempt to fully enumerate the system's USB bus. This has been found
to cause jank and hangs on Chrome's FILE thread on Windows. Disabling
this component of WebUSB while the issue is investigated.

WebUSB will continue to function but this change will prevent the issue
from impacting users who do not visit sites using it.

BUG= 656702 
TEST=Verified USB subsystem is not initialized on startup on Windows by
checking chrome://device-log. Verified still initialized on Linux.

Review-Url: https://codereview.chromium.org/2422403003
Cr-Commit-Position: refs/heads/master@{#425867}
(cherry picked from commit dfd32df76aea1a2d4f1aad69c74ded742a2ac69b)

Review URL: https://codereview.chromium.org/2440463002 .

Cr-Commit-Position: refs/branch-heads/2883@{#187}
Cr-Branched-From: 614d31daee2f61b0180df403a8ad43f20b9f6dd7-refs/heads/master@{#423768}

[modify] https://crrev.com/b3da8702f5e857a68eb081ea253120120cfccb66/chrome/browser/usb/web_usb_detector.cc
[modify] https://crrev.com/b3da8702f5e857a68eb081ea253120120cfccb66/chrome/browser/usb/web_usb_detector_unittest.cc

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/2ac51a5dd890c0944f4d689792319426fb5f9850

commit 2ac51a5dd890c0944f4d689792319426fb5f9850
Author: Reilly Grant <reillyg@chromium.org>
Date: Tue Oct 18 22:17:45 2016

Disable WebUSB detector logic on Windows.

Initializing the WebUSB device detector logic causes the USB subsystem
to attempt to fully enumerate the system's USB bus. This has been found
to cause jank and hangs on Chrome's FILE thread on Windows. Disabling
this component of WebUSB while the issue is investigated.

WebUSB will continue to function but this change will prevent the issue
from impacting users who do not visit sites using it.

BUG= 656702 
TEST=Verified USB subsystem is not initialized on startup on Windows by
checking chrome://device-log. Verified still initialized on Linux.

Review-Url: https://codereview.chromium.org/2422403003
Cr-Commit-Position: refs/heads/master@{#425867}
(cherry picked from commit dfd32df76aea1a2d4f1aad69c74ded742a2ac69b)

Review URL: https://codereview.chromium.org/2425903003 .

Cr-Commit-Position: refs/branch-heads/2840@{#751}
Cr-Branched-From: 1ae106dbab4bddd85132d5b75c670794311f4c57-refs/heads/master@{#414607}

[modify] https://crrev.com/2ac51a5dd890c0944f4d689792319426fb5f9850/chrome/browser/usb/web_usb_detector.cc
[modify] https://crrev.com/2ac51a5dd890c0944f4d689792319426fb5f9850/chrome/browser/usb/web_usb_detector_unittest.cc

There nothing that warrants Restrict-View-Google here.