Bug 219860 - CrashTracer: com.apple.WebKit.Networking at WebKit: WTF::Detail::CallableWrapper<WebKit::PrivateClickMeasurementManager::firePendingAttributionRequests()
Summary: CrashTracer: com.apple.WebKit.Networking at WebKit: WTF::Detail::CallableWrap...
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: WebKit Misc. (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Kate Cheney
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2020-12-14 10:22 PST by Kate Cheney
Modified: 2020-12-15 10:28 PST (History)
4 users (show)

See Also:


Attachments
Patch (3.96 KB, patch)
2020-12-14 10:55 PST, Kate Cheney
no flags Details | Formatted Diff | Diff
Patch (6.98 KB, patch)
2020-12-14 12:10 PST, Kate Cheney
no flags Details | Formatted Diff | Diff
Patch (7.43 KB, patch)
2020-12-15 08:54 PST, Kate Cheney
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Kate Cheney 2020-12-14 10:22:29 PST
We are hitting this crash after PCM code was migrated to use SQLite.
Comment 1 Kate Cheney 2020-12-14 10:55:52 PST
Created attachment 416172 [details]
Patch
Comment 2 Kate Cheney 2020-12-14 10:56:28 PST
<rdar://problem/72254207>
Comment 3 Alex Christensen 2020-12-14 11:53:23 PST
Comment on attachment 416172 [details]
Patch

From the description, I think this should just add this:
if (!m_networkSession)
    return;
Comment 4 Kate Cheney 2020-12-14 11:58:27 PST
(In reply to Alex Christensen from comment #3)
> Comment on attachment 416172 [details]
> Patch
> 
> From the description, I think this should just add this:
> if (!m_networkSession)
>     return;

I thought if (auto* resourceLoadStatistics = m_networkSession->resourceLoadStatistics()) would check m_networkSession but maybe I am wrong? If so I should check m_networkSession in other PrivateClickMeasurementManager functions to be consistent.
Comment 5 Alex Christensen 2020-12-14 11:59:19 PST
if (auto* resourceLoadStatistics = m_networkSession ? m_networkSession->resourceLoadStatistics() : nullptr) checks m_networkSession, but what you have doesn't.
Comment 6 Kate Cheney 2020-12-14 12:10:19 PST
Created attachment 416180 [details]
Patch
Comment 7 EWS 2020-12-14 12:59:14 PST
Committed r270802: <https://trac.webkit.org/changeset/270802>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 416180 [details].
Comment 8 Radar WebKit Bug Importer 2020-12-14 13:01:44 PST
<rdar://problem/72310410>
Comment 9 John Wilander 2020-12-14 17:34:23 PST
Comment on attachment 416180 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=416180&action=review

An easier way to fix this is to move the null check into PrivateClickMeasurementManager::featureEnabled().

> Source/WebKit/NetworkProcess/PrivateClickMeasurementManager.cpp:302
> +        return;

We need to also call the completion handler here.

> Source/WebKit/NetworkProcess/PrivateClickMeasurementManager.cpp:342
> +        return;

We need to also call the completion handler here.
Comment 10 Alex Christensen 2020-12-14 18:25:13 PST
Comment on attachment 416180 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=416180&action=review

>> Source/WebKit/NetworkProcess/PrivateClickMeasurementManager.cpp:302
>> +        return;
> 
> We need to also call the completion handler here.

yes we certainly do.  Also in the case before if resourceLoadStatistics is null.
Comment 11 Kate Cheney 2020-12-15 08:54:02 PST
Reopening to attach new patch.
Comment 12 Kate Cheney 2020-12-15 08:54:03 PST
Created attachment 416253 [details]
Patch
Comment 13 EWS 2020-12-15 10:28:21 PST
Committed r270852: <https://trac.webkit.org/changeset/270852>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 416253 [details].