Bug 232970 - We should not kill all WebContent processes whenever the WebAuthn process crashes
Summary: We should not kill all WebContent processes whenever the WebAuthn process cra...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebKit2 (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Chris Dumez
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2021-11-10 16:15 PST by Chris Dumez
Modified: 2021-11-11 10:45 PST (History)
8 users (show)

See Also:


Attachments
Patch (4.93 KB, patch)
2021-11-10 16:19 PST, Chris Dumez
no flags Details | Formatted Diff | Diff
Patch (7.44 KB, patch)
2021-11-10 17:50 PST, Chris Dumez
ews-feeder: commit-queue-
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Chris Dumez 2021-11-10 16:15:17 PST
We should not kill all WebContent processes whenever the WebAuthn process crashes. This is overly aggressive. We should instead do like for the network process and have the WebProcess re-initiate the connection to the WebAuthn process when it's gone.
Comment 1 Chris Dumez 2021-11-10 16:15:40 PST
<rdar://83941760>
Comment 2 Chris Dumez 2021-11-10 16:19:20 PST
Created attachment 443876 [details]
Patch
Comment 3 Geoffrey Garen 2021-11-10 16:24:05 PST
Comment on attachment 443876 [details]
Patch

r=me
Comment 4 Geoffrey Garen 2021-11-10 16:24:40 PST
Does this API test (or another) demonstrate that use of WebAuthn does succeed after restarting the WebAuthn process?
Comment 5 Chris Dumez 2021-11-10 16:28:59 PST
(In reply to Geoffrey Garen from comment #4)
> Does this API test (or another) demonstrate that use of WebAuthn does
> succeed after restarting the WebAuthn process?

Hmm, this particular test only calls `internals.setMockWebAuthenticationConfiguration({ })` which causes the WebAuthn process to get re-launched but it indeed doesn't attempt to use WebAuthn afterwards.

This seems to be the only test that involves killing the WebAuthn process...

I know very little about WebAuthn :o) but I'll look into extending the test to actually use WebAuthn after the crash.
Comment 6 Chris Dumez 2021-11-10 17:50:10 PST
Created attachment 443887 [details]
Patch
Comment 7 EWS 2021-11-10 21:03:22 PST
Committed r285619 (244120@main): <https://commits.webkit.org/244120@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 443887 [details].