727864 - Enable DLL preloading when MozillaMaintenance service disabled prefetch

Status: VERIFIED FIXED

(Core :: Widget: Win32, defect)

Description

[Brian R. Bondy [:bbondy]]

We currently only do DLL preloading when we think that the Windows prefetch isn't used.  We should also do DLL preloading when the maintenance service has performed a successful clear prefetch operation.

Comment 1

[Brian R. Bondy [:bbondy]]

Attachment: Patch v1.

Comment 2

[Brian R. Bondy [:bbondy]]

Alfred Kayser 2012-02-16 23:59:42 PST

I would move the LOG statement of the failing WritePrefetchClearedReg function inside that function, just as the other LOG statement is also there.
+  if (retCode != ERROR_SUCCESS) {
+    LOG(("Could not open key. (%d)\n", retCode));
+  }
And, add a 'return false' here after the LOG.

Comment 3

[Brian R. Bondy [:bbondy]]

Attachment: Patch v2.

Implemented review comments and rebased to m-c tip.

Comment 4

[Jim Mathies [:jimm]]

Comment on attachment 598233 [details] [diff] [review]
Patch v2.

Review of attachment 598233 [details] [diff] [review]:
-----------------------------------------------------------------

::: browser/app/nsBrowserApp.cpp
@@ +251,5 @@
>    // in the program. Luckily 1 coincides with when prefetch is
>    // enabled. If Windows prefetch didn't happen we can do our own
>    // faster dll preloading.
> +  // The MozillaMaintenance service issues a command to disable the
> +  // pretch by replacing all found .pf files with 0 byte read only

nit - 'pretch'

::: toolkit/components/maintenanceservice/prefetch.cpp
@@ +54,5 @@
> +  LONG retCode = RegOpenKeyExW(HKEY_LOCAL_MACHINE, 
> +                               BASE_SERVICE_REG_KEY, 0,
> +                               KEY_WRITE | KEY_WOW64_64KEY, &baseKeyRaw);
> +  if (retCode != ERROR_SUCCESS) {
> +    LOG(("Could not open key. (%d)\n", retCode));

does this LOG macro do file name / line number logging? If not you might want to make that msg more specific.

Comment 5

[Brian R. Bondy [:bbondy]]

Attachment: Patch v3.

Implemented nits, carrying forward r+.

Comment 6

[Brian R. Bondy [:bbondy]]

Comment on attachment 599602 [details] [diff] [review]
Patch v3.

Actually I don't think this one needs a super-review so clearing that flag.

Comment 7

[Brian R. Bondy [:bbondy]]

Attachment: Patch v4

Rebased to tip

Comment 8

[Brian R. Bondy [:bbondy]]

http://hg.mozilla.org/mozilla-central/rev/21ae85ff4286

Comment 9

[Jason Smith [:jsmith]]

Is there anything to verify here from an end-user perspective?

Comment 10

[Brian R. Bondy [:bbondy]]

No this was verified by seeing telemetry data.