1752907 - Cut over Hardware Concurrency to finer-grained RFP Check

Status: RESOLVED FIXED

(Core :: DOM: Workers, enhancement)

Description

[Alisha Sharma]

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/97.0.4692.99 Safari/537.36

Comment 1

[Alisha Sharma]

Attachment: Bug 1752907: Cut over Hardware Concurrency to finer-grained RFP Check r?tjr

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/370f46a043e4
Cut over Hardware Concurrency to finer-grained RFP Check r=tjr

Comment 3

[Norisz Fay [:noriszfay]]

Backed out for causing mochitest failures on browser_navigator.js

Backout link

Push with failures

Failure log

Failure line(s): TEST-UNEXPECTED-FAIL | browser/components/resistfingerprinting/test/browser/browser_navigator.js | Checking spoofed navigator.hardwareConcurrency. - Got 6, expected 2

Comment 4

[BugBot [:suhaib / :marco/ :calixte]]

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:asharm28, could you have a look please?
If you still have some work to do, you can add an action "Plan Changes" in Phabricator.
For more information, please visit auto_nag documentation.

Comment 5

[Tom Ritter [:tjr]]

diff --git a/dom/workers/WorkerNavigator.cpp b/dom/workers/WorkerNavigator.cpp
--- a/dom/workers/WorkerNavigator.cpp
+++ b/dom/workers/WorkerNavigator.cpp
@@ -171,17 +171,21 @@ void WorkerNavigator::GetUserAgent(nsStr
 
   runnable->Dispatch(Canceling, aRv);
 }
 
 uint64_t WorkerNavigator::HardwareConcurrency() const {
   RuntimeService* rts = RuntimeService::GetService();
   MOZ_ASSERT(rts);
 
-  return rts->ClampedHardwareConcurrency();
+  WorkerPrivate* aWorkerPrivate = GetCurrentThreadWorkerPrivate();
+  bool rfp = aWorkerPrivate->ShouldResistFingerprinting();
+  fprintf(stderr, "ZTZT WorkerNavigator::HardwareConcurrency %s\n", rfp ? "yes" : "no");
+
+  return rts->ClampedHardwareConcurrency(rfp);
 }
 
 StorageManager* WorkerNavigator::Storage() {
   if (!mStorageManager) {
     WorkerPrivate* workerPrivate = GetCurrentThreadWorkerPrivate();
     MOZ_ASSERT(workerPrivate);
 
     RefPtr<nsIGlobalObject> global = workerPrivate->GlobalScope();
diff --git a/dom/workers/WorkerPrivate.cpp b/dom/workers/WorkerPrivate.cpp
--- a/dom/workers/WorkerPrivate.cpp
+++ b/dom/workers/WorkerPrivate.cpp
@@ -2644,16 +2644,17 @@ nsresult WorkerPrivate::GetLoadInfo(JSCo
     loadInfo.mHasStorageAccessPermissionGranted =
         aParent->HasStorageAccessPermissionGranted();
     loadInfo.mCookieJarSettings = aParent->CookieJarSettings();
     loadInfo.mOriginAttributes = aParent->GetOriginAttributes();
     loadInfo.mServiceWorkersTestingInWindow =
         aParent->ServiceWorkersTestingInWindow();
     loadInfo.mShouldResistFingerprinting =
         aParent->ShouldResistFingerprinting();
+    fprintf(stderr, "ZTZT %s\n", loadInfo.mShouldResistFingerprinting ? "yes" : "no");
     loadInfo.mIsThirdPartyContextToTopWindow =
         aParent->IsThirdPartyContextToTopWindow();
     loadInfo.mParentController = aParent->GlobalScope()->GetController();
     loadInfo.mWatchedByDevTools = aParent->IsWatchedByDevTools();
   } else {
     AssertIsOnMainThread();
 
     // Make sure that the IndexedDatabaseManager is set up
@@ -2785,16 +2786,17 @@ nsresult WorkerPrivate::GetLoadInfo(JSCo
       loadInfo.mFromWindow = true;
       loadInfo.mWindowID = globalWindow->WindowID();
       loadInfo.mStorageAccess = StorageAllowedForWindow(globalWindow);
       loadInfo.mUseRegularPrincipal = document->UseRegularPrincipal();
       loadInfo.mHasStorageAccessPermissionGranted =
           document->HasStorageAccessPermissionGranted();
       loadInfo.mShouldResistFingerprinting =
           nsContentUtils::ShouldResistFingerprinting(document);
+      fprintf(stderr, "ZTZT %s\n", loadInfo.mShouldResistFingerprinting ? "yes" : "no");
 
       // This is an hack to deny the storage-access-permission for workers of
       // sub-iframes.
       if (loadInfo.mHasStorageAccessPermissionGranted &&
           StorageAllowedForDocument(document) != StorageAccess::eAllow) {
         loadInfo.mHasStorageAccessPermissionGranted = false;
       }
       loadInfo.mIsThirdPartyContextToTopWindow =
diff --git a/dom/workers/WorkerPrivate.h b/dom/workers/WorkerPrivate.h
--- a/dom/workers/WorkerPrivate.h
+++ b/dom/workers/WorkerPrivate.h
@@ -847,16 +847,17 @@ class WorkerPrivate final
   }
 
   // Determine if the SW testing per-window flag is set by devtools
   bool ServiceWorkersTestingInWindow() const {
     return mLoadInfo.mServiceWorkersTestingInWindow;
   }
 
   bool ShouldResistFingerprinting() const {
+    fprintf(stderr, "ZTZT WP::ShouldResistFingerprinting() %s\n", mLoadInfo.mShouldResistFingerprinting ? "yes" : "no");
     return mLoadInfo.mShouldResistFingerprinting;
   }
 
   // Determin if the worker was created under a third-party context.
   bool IsThirdPartyContextToTopWindow() const {
     return mLoadInfo.mIsThirdPartyContextToTopWindow;
   }

Getting really weird result with the WorkerPrivate loadinfo data not being persisted; need to investigate....

Comment 6

[Tom Ritter [:tjr]]

Attachment: Bug 1752907: Correctly populate the RFP member on Workers r?asuth

Comment 7

[Tom Ritter [:tjr]]

Attachment: Bug 1752907: Add some helpful log messages to test r?asuth

Comment 8

[Tom Ritter [:tjr]]

Attachment: Bug 1752907: Add a new RFP exemptions subtest r?asuth

Because right now we only have hardware concurrency wired
up to the fine-grained RFP checks this is a copy of the
regular RFP test but with the hardware concurrency field
set to the non-spoofed value

Depends on D141772

Comment 9

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/94d4b31ef7d0
Correctly populate the RFP member on Workers r=asuth

Comment 10

[Atila Butkovits]

https://hg.mozilla.org/mozilla-central/rev/94d4b31ef7d0

Comment 11

[Tom Ritter [:tjr]]

Need to land a couple more on this bug.

Comment 13

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/e6c18f35ac5f
Cut over Hardware Concurrency to finer-grained RFP Check r=tjr
https://hg.mozilla.org/integration/autoland/rev/e50ad7304c18
Add some helpful log messages to test r=asuth
https://hg.mozilla.org/integration/autoland/rev/747092282e35
Add a new RFP exemptions subtest r=asuth

Comment 14

[Cristina Cozmuta (:CrissCozmuta)]

Backed out for causing failures on browser_navigator.js. CLOSED TREE

Backout link: https://hg.mozilla.org/integration/autoland/rev/f749cf663753da619cf06d946d51535c89dc67b5

Push with failures : https://treeherder.mozilla.org/jobs?repo=autoland&resultStatus=pending%2Crunning%2Csuccess%2Ctestfailed%2Cbusted%2Cexception%2Crunnable&searchStr=windows%2C7%2Cwebrender%2Copt%2Cmochitests%2Cwith%2Cfission%2Cenabled%2Ctest-windows7-32-qr%2Fopt-mochitest-browser-chrome-fis-e10s%2Cbc7&revision=747092282e35bcd902865963dd72b76550a2f066

Link to failure log: https://treeherder.mozilla.org/logviewer?job_id=372215458&repo=autoland&lineNumber=5581

Failure message:
TEST-UNEXPECTED-FAIL | browser/components/resistfingerprinting/test/browser/browser_navigator.js | Checking spoofed navigator.oscpu. - Got "Windows NT 6.1", expected "Windows NT 10.0; Win64; x64"

Comment 15

[Tom Ritter [:tjr]]

This failure is unrelated to this patch.

This test is supposed to be skipped when it's 32 bit Firefox running on 64 bit Windows. The skip-if says 'debug', the test that failed was 'opt' but I'm not sure why the 'debug' is in there, it seems like it should be both. Chris, does that seem right?

Comment 16

[Chris Peterson [:cpeterson]]

(In reply to Tom Ritter [:tjr] (ni for sec-approvals and such) from comment #15)

This failure is unrelated to this patch.

This test is supposed to be skipped when it's 32 bit Firefox running on 64 bit Windows. The skip-if says 'debug', the test that failed was 'opt' but I'm not sure why the 'debug' is in there, it seems like it should be both. Chris, does that seem right?

I agree:

• That test failure is not this patch's fault.
• We should probably skip this test for both debug and opt in 32-bit Firefox on 64-bit Windows.

I looking into fixing this test to pass in 32-bit Firefox on 64-bit Windows, but it wasn't a high priority and I forgot. I'll try to fix the test soon or just disable it for both debug and opt.

Comment 17

[Tom Ritter [:tjr]]

Attachment: Bug 1752907: Disable the navigator test on x86 Windows r?cpeterson

Comment 18

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/085cb7354660
Disable the navigator test on x86 Windows r=cpeterson

Comment 19

[Norisz Fay [:noriszfay]]

https://hg.mozilla.org/mozilla-central/rev/085cb7354660

Comment 20

[BugBot [:suhaib / :marco/ :calixte]]

Since the status are different for nightly and release, what's the status for beta?
For more information, please visit auto_nag documentation.

Comment 21

[Tom Ritter [:tjr]]

We haven't resolved this yet.

Comment 22

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/b43752720ca6
Cut over Hardware Concurrency to finer-grained RFP Check r=tjr
https://hg.mozilla.org/integration/autoland/rev/904ca697efbb
Add some helpful log messages to test r=asuth
https://hg.mozilla.org/integration/autoland/rev/d386072b56b6
Add a new RFP exemptions subtest r=asuth

Comment 23

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/mozilla-central/rev/b43752720ca6
https://hg.mozilla.org/mozilla-central/rev/904ca697efbb
https://hg.mozilla.org/mozilla-central/rev/d386072b56b6