1979727 - Testcase generating eTLD + N URL and fetching it hangs the Parent-Process for 6minutes+. Chrome is instant (i think it errors out)

Status: VERIFIED FIXED

(Core :: Security: PSM, defect, P1)

Description

[Mayank Bansal]

Attachment: etld + N.html

Open testcase
Enter 500000 (or use a saner smaller number)
Click on generate. -- > This is instant, so thats good
Now click on Fetch button

Firefox (partial): https://share.firefox.dev/4faULbO (6minutes + )
Chrome: instant (or instant error-out)

Comment 1

[Mayank Bansal]

Here is a saner profile with N= 150000 : https://share.firefox.dev/4l1a08Q

Comment 2

[Sunil Mayya]

It looks like the source of the jank is this piece of code:
nsSiteSecurityService.cpp

Not sure if this parsing logic needs to be adjusted for this specific test case (an extremely long URL with "." delimiters).

I’ll defer to the security team to determine if any changes are necessary.

Comment 3

[Sunil Mayya]

Attachment: Screenshot 2025-08-04 at 17.41.28.png

Jank in nsSiteSecurityService::IsSecureHost()

Comment 4

[Dana Keeler (she/her) [:keeler]]

Well, you're never going to be able to resolve a hostname like that, so I wonder if it makes sense to add a limit to the number of subdomain components when constructing a URI (this may be what makes Chrome so fast here). Kershaw, what do you think?

Comment 5

[Kershaw Chang [:kershaw]]

(In reply to Dana Keeler (she/her) [:keeler] from comment #4)

Well, you're never going to be able to resolve a hostname like that, so I wonder if it makes sense to add a limit to the number of subdomain components when constructing a URI (this may be what makes Chrome so fast here). Kershaw, what do you think?

So, we have a limit here to check if a host name is valid.
I think we just need to reuse that check in nsSiteSecurityService::IsSecureHost().

Comment 6

[Dana Keeler (she/her) [:keeler]]

Sounds good!
Nikolas, would you be interested in taking this on?

Comment 7

[Nikolas Wipper [:nwipper]]

This does happen on my machine + I also get a consistent 7 seconds of JS jank regardless of N: https://share.firefox.dev/3JpXN02. Not sure how that is related.

Comment 8

[Nikolas Wipper [:nwipper]]

Attachment: Bug 1979727 - Return early for invalid host names when checking if they are secure. r=keeler

Comment 9

[Dana Keeler (she/her) [:keeler]]

(In reply to Nikolas Wipper [:nwipper] from comment #7)

This does happen on my machine + I also get a consistent 7 seconds of JS jank regardless of N: https://share.firefox.dev/3JpXN02. Not sure how that is related.

Might be these regexes: https://searchfox.org/mozilla-central/rev/ca035d76ad92757859375c2dc807c7a61055823b/devtools/client/netmonitor/src/utils/request-utils.js#354-356 - maybe file a bug to see what we can do there?

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/c7313046808a
https://hg.mozilla.org/integration/autoland/rev/670499a08a7c
Return early for invalid host names when checking if they are secure. r=keeler

Comment 11

[BugBot [:suhaib / :marco]]

The severity field is not set for this bug.
:keeler, could you have a look please?

For more information, please visit BugBot documentation.

Comment 12

[Serban Stanca [:SerbanS]]

https://hg.mozilla.org/mozilla-central/rev/670499a08a7c

Comment 13

[Mayank Bansal]

This is fixed on the latest Nightly - The browser seems to do nothing when i press the "fetch" button.
Thanks!