I have the same problem. Here is a short demo to reproduce this.

Deleted: scroll-bug-62.html 512 bytes

scroll-bug-62.html 512 bytes View Download

Tested on latest Chrome Stable #62.0.3202.62, Canary #64.0.3247.0 on Windows 10 & Ubuntu 14.04 and able to reproduce the issue.

Using the per-revision bisect providing the bisect results,
Good build: 62.0.3178.0 (492239)
Bad build: 62.0.3179.0 (492477)

You are probably looking for a change made after 492372 (known good), but no later than 492373 (first known bad).

CHANGELOG URL:
The script might not always return single CL as suspect as some perf builds might get missing due to failure.

https://chromium.googlesource.com/chromium/src/+log/8ae3ea1c2efcf9d2786cb1af66315b08f0a95051..01cf6fd553bcc6d5a9ca6763eed42fbca5ecc494
https://chromium.googlesource.com/chromium/src/+/01cf6fd553bcc6d5a9ca6763eed42fbca5ecc494

From the CL above, assigning the issue to the owner concerned.

@szager:  Could you please look into the issue, pardon me if it has nothing to do with your changes and if possible please assign it to owner concerned.

Review-URL: https://chromium-review.googlesource.com/c/chromium/src/+/585892

Thanks!

Similar problem appears with Facebook-Plugin: Two scrollbars on the side. 

https://developers.facebook.com/docs/plugins/page-plugin

Deleted: 2017-10-23_https___developers.facebook.com_docs_plugins_page-plugin.png 127 KB

	2017-10-23_https___developers.facebook.com_docs_plugins_page-plugin.png 127 KB View Download

Punting to M-63 stable release.

@szager: Request you to please take a look into it as issue is tagged with a stable blocker and targeted to M63 which is set to be pushed to Beta Soon.

Thanks.!

[Bulk Edit]
URGENT - PTAL.
M63 Stable promotion is coming soon and your bug is labelled as Stable ReleaseBlock, pls make sure to land the fix and get it merged into the release branch ASAP. Thank you.

The root cause of this bug is improper rounding of border width throughout the layout code.

When borders are painted, their size is rounded to the nearest device pixel (with a minimum size of one device pixel).  However, the layout code uses the un-rounded CSS pixel size when computing box sizing.  This can cause a box to be sized too small to fit its contents.

I tried a couple of hacky fixes for this, by adding a fudge factor of up to one pixel when determing whether to add a scrollbar to an overflow:auto container.  Unfortunately, each of these attempts caused test failures.  I think it's too risky to merge a hacky fix that has user-visible effects; it will likely break as many things as it fixes.

Reproducing this bug requires the page to be zoomed, which means it's already somewhat narrow.  Extra scrollbars are ugly to behold, but they don't break the functionality of the page.  For these reasons, and because we don't have a good fix available, I propose we remove the stable-blocker label and shoot for the next release.  I am working on a more comprehensive fix, but we will want to give it time to bake in dev/beta.

I concur. Removing RB-S label.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/7f73e60a130464757b3773339655958391f6b40c

commit 7f73e60a130464757b3773339655958391f6b40c
Author: Stefan Zager <szager@chromium.org>
Date: Sat Nov 11 04:55:08 2017

Make PaintLayer::size_ a LayoutSize

Snapping PaintLayer size to the pixel grid too early causes a couple of problems:

- It can inadvertently trigger auto scrollbars to be created for a box with
non-pixel-aligned location and non-pixel-aligned content size.  This is the root cause
of the bug linked below.

- There are a few places in the code that call LayoutBox::SetLocation without updating
the box'es PaintLayer's size.  Since PaintLayer size relies on snapping to the box'es
location, this can cause the PaintLayer size to be off by one pixel.  That's what
happened in the new test baseline in this patch: the layer size in the old expectation
is actually off by one pixel.

BUG= 777095 

Cq-Include-Trybots: master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
Change-Id: I58362d64eff39c5499417f038f5e6ed5eed6f1e3
Reviewed-on: https://chromium-review.googlesource.com/752030
Reviewed-by: Tien-Ren Chen <trchen@chromium.org>
Reviewed-by: Steve Kobes <skobes@chromium.org>
Reviewed-by: Philip Rogers <pdr@chromium.org>
Commit-Queue: Stefan Zager <szager@chromium.org>
Cr-Commit-Position: refs/heads/master@{#515826}
[modify] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/LayoutTests/platform/mac/media/controls/controls-page-zoom-in-expected.txt
[modify] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/LayoutTests/platform/mac/virtual/new-remote-playback-pipeline/media/controls/controls-page-zoom-in-expected.txt
[add] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/LayoutTests/scrollbars/sub-pixel-overflow.html
[modify] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/Source/core/paint/PaintLayer.cpp
[modify] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/Source/core/paint/PaintLayer.h
[modify] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/Source/core/paint/PaintLayerClipper.cpp
[modify] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/Source/core/paint/PaintLayerScrollableArea.cpp
[modify] https://crrev.com/7f73e60a130464757b3773339655958391f6b40c/third_party/WebKit/Source/core/paint/ScrollableAreaPainter.cpp

szager@: Do you think this is safe to be merged to M63? Chrome Settings UI ( https://crbug.com/769199 ) as well as some other sites ( https://crbug.com/775986 ) are affected by it. 

No, I don't think this is safe to merge.  It needs to bake in dev and/or beta, to make sure it doesn't have unintended consequences or cause other sites to break.

 Issue 772380  has been merged into this issue.

 Issue 775986  has been merged into this issue.

 Issue 779556  has been merged into this issue.

 Issue 782721  has been merged into this issue.

 Issue 781654  has been merged into this issue.

 Issue 780035  has been merged into this issue.

 Issue 772360  has been merged into this issue.

 Issue 799412  has been merged into this issue.

 Issue 806359  has been merged into this issue.

Maybe rename this issue to avoid futur duplicate, it's not related to gmail.
repro here if it can help: https://codepen.io/frlinw/pen/qpeJJN

A release in Chrome 65 is possible?

The fix is in all M64 releases, and will be in M65 (and all future releases)

posting this here because I believe this issue is related/similar , please take a look https://bugs.chromium.org/p/chromium/issues/detail?id=807790 

another very similar bug, please take a look https://bugs.chromium.org/p/chromium/issues/detail?id=818873