1492766 - Fingerprinting protection for pointerEvent.pointerid

Status: RESOLVED FIXED

(Core :: DOM: Events, enhancement, P2)

Description

[Tim Huang[:timhuang]]

According to the spec[1], the pointerid is not related to the pointer type. However, in practice, the pointerid could reveal the real pointer type. We should also spoof the pointerid when fingerprinting resistance is on.

[1] https://www.w3.org/TR/pointerevents/#dom-pointerevent-pointerid

Comment 1

[Tim Huang[:timhuang]]

Attachment: Bug 1492766 - Part 1: Spoofing the pointer id of pointer events when resistfingerprinting is enabled

This patch makes it to spoof the pointer id into mouse pointer id
when resistfingerprinting is enabled. And we will only spoof for
content but not for chrome. Since we don't know the pointer id
beforehand, we have to cache the pointer id of the mouse interface when
it is activated. And use that pointer for the purpose of fingerprinting
resistance.

Comment 2

[Tim Huang[:timhuang]]

Attachment: Bug 1492766 - Part 2: Makes the set/releasePointerCapture working properly when fingerprinting resistance is enabled

When fingerprinting resistance is enabled, content should only view the
pointer capture events from the spoofed interface. In order to do so,
first, we need to restrict content to only set or release pointer capture
for only the spoofed pointer id. Second, we have to map other interfaces
into the spoofed one for pointer capture events.

Depends on D9531

Comment 3

[Tim Huang[:timhuang]]

Attachment: Bug 1492766 - Part 3: Update the test_pointer_event.html for test of pointer id spoofing

This test add a check for pointer id to make sure it always reply the
spoofed pointer id when resistfingerpritning is enabled. It also extend
the original test of pointer capture events into a test which tests
mouse and touch interfaces to make sure pointer capture events is
correctly filed even with the spoofed interface id. In addition, it also
adds test for that set/releasePointerCapture only accepts spoofed
pointer id when fingerprinting resistance is enabled.

Depends on D9532

Comment 4

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Sorry for the delay to review. I'll try to review them next week.

Comment 5

[Tim Huang[:timhuang]]

Try looks good.

https://treeherder.mozilla.org/#/jobs?repo=try&revision=82e59e5a06717e4b36903ad590c9e45606c13513

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a575162d9f5b
Part 1: Spoofing the pointer id of pointer events when resistfingerprinting is enabled r=masayuki,smaug,arthuredelstein
https://hg.mozilla.org/integration/autoland/rev/76017cad8320
Part 2: Makes the set/releasePointerCapture working properly when fingerprinting resistance is enabled r=masayuki,arthuredelstein,smaug
https://hg.mozilla.org/integration/autoland/rev/d48ee9ed6819
Part 3: Update the test_pointer_event.html for test of pointer id spoofing r=masayuki,arthuredelstein

Comment 7

[Arthur Iakab [arthur_iakab]]

https://hg.mozilla.org/mozilla-central/rev/a575162d9f5b
https://hg.mozilla.org/mozilla-central/rev/76017cad8320
https://hg.mozilla.org/mozilla-central/rev/d48ee9ed6819

Comment 8

[Arthur Iakab [arthur_iakab]]

https://hg.mozilla.org/mozilla-central/rev/a575162d9f5b
https://hg.mozilla.org/mozilla-central/rev/76017cad8320
https://hg.mozilla.org/mozilla-central/rev/d48ee9ed6819