Here is a slightly different use case: https://jsfiddle.net/j5zksate/

In the OK browsers mentioned, the links become clickable once you scroll past the 'Overlay' section.

Chrome does not scroll at all, but the links do seem to become clickable if the scrollTop of the sections within the article are forced out of view.

It seems to be something to do with the overflow: hidden + overflow-y: auto, as shown here:

- https://jsfiddle.net/j5zksate/1/
- https://jsfiddle.net/j5zksate/2/

I can't scroll at all on Linux for either example, even uising the scrollbar. I'll bisect to see if this is recent.

Regression range is this:
https://chromium.googlesource.com/chromium/src/+log/04a322f25adf8b9cab272e985f44db9a01814450..34ca6f6990519850bd0831d0da74bf511d056290

Of which the one to blame seems to be this:
https://chromium.googlesource.com/chromium/src/+/ebcfc3309fba892b25d84f943d26bbdab6d100ee "currentlyscrollinglayer cleared in scrollAnimatedBegin."

Not clear why it breaks, but it does.

That's Chroime 55 when it broke, so this must really be a rare case.

FWIW in Windows the reported bug is reproducible here only 1) if site-per-process is disabled and I test the provided jsfiddle which use iframes 2) if I run the actual html/css in top document from a local file.

I don't think the regression is related to the suspected cl:
https://chromium.googlesource.com/chromium/src/+/ebcfc3309fba892b25d84f943d26bbdab6d100ee
Since it has three lines of code in layer_tree_host_impl.cc and the rest of the changes are comments or DCHECKS and that three lines of code are deleted in https://codereview.chromium.org/2486673008 later on.

schenny@ please feel free to assign it back to me if you still think https://chromium.googlesource.com/chromium/src/+/ebcfc3309fba892b25d84f943d26bbdab6d100ee is the suspected cl.

Able to reproduce this issue on reported version 67.0.3360.1, on latest canary 67.0.3371.0 using Windows , Linux and Mac.

This issue is seen from M60.As per comment#10 bisected, issue is broken in M-51.

Observations: Till 51.0.2686.0 able to scroll and click on links, but from 51.0.2687.0 unable to scroll.

Good Build: 51.0.2686.0
Bad Build: 51.0.2687.0

You are probably looking for a change made after 382335 (known good), but no later than 382346 (first known bad).
CHANGELOG URL:
  https://chromium.googlesource.com/chromium/src/+log/6bebadddb399ad1c1f91592efed410040617e708..5840ac85a7c912d9ddb764007b6a9ec65a26ad22

Suspecting https://codereview.chromium.org/1776843003 from changelog.

@ dtapuska: Please confirm the bug and help in re-assigning if it is not related to your change. 

As @dtapuska is OOO assigning to nzolghadr@ for further inputs on this.

Deleted: 818444 in M50.mp4 445 KB

	818444 in M50.mp4 445 KB View Download

Deleted: 818444 Actual.mp4 410 KB

	818444 Actual.mp4 410 KB View Download

dtapuska@ this is not our pointer events. Do you know who owns the CSS property pointerevents?

css pointerevents is essentially a hit-testing related feature so whoever owns hit-testing should look into this.

So I suspect input-dev may be the right owner. In particular since according to 
comment #11 the regression appears to be related to enabling wheel gesture events.

Historically, compositor hit-testing has not been taking pointer-events: none;
into account (sunxd@ is working to fix that). Perhaps by enabling wheel gesture events causes this case to now hit compositor which causes this issue.  

nzolghadr@ can you confrim that this is fixed if you disable threaded-scrolling?
If so then this is a compositor hit-testing issue and I think sunxd@ can be the right
owner.

This looks like the event is blocked by the parent. Currently cc does not have any concept of pointer-events: none, so cc's behaviour would allow any events targeted at pointer-events none elements. It seems to me this bug is still in blink side.

sunxd@, Did we even establish that an event is initiated but then blocked vs. the hit test never triggering any event?

I just trying to understand where to route this. Event propagation is one team, hit testing another.

If pointer-events ever affect this scroll I think it denied the event at this line: https://cs.chromium.org/chromium/src/third_party/blink/renderer/core/layout/layout_object.h?rcl=26e9592db563218aec7e8187e7a8c19a24349035&l=1738.

But I'm not 100% sure.

The NextAction date has arrived: 2019-01-07