[jdk11u-dev] RFR: 8292177: InitialSecurityProperty JFR event

[Andrew John Hughes]

On Sat, 29 Jun 2024 16:21:19 GMT, Andrew John Hughes <andrew at openjdk.org> wrote:

> This backport introduces the JFR security event, `InitialSecurityProperty`, and along with it, the caching of the initial security properties which is also used by [JDK-8281658](https://bugs.openjdk.org/browse/JDK-8281658) to show the security properties with `-XshowSettings`.
> 
> The change was backported to the Oracle fork of 11u in 11.0.20.
> 
> The following changes were necessary for the backport:
> 
> * `JavaSecurityPropertiesAccess` is moved to `jdk.internal.misc` where `SharedSecrets` lives in 11u
> * `ensureClassInitialized` in `SharedSecrets` is called from the `unsafe` instance in 11u, as with other `get*Access()` methods in that class
> * The patch to `module-info.java` is not needed as `jdk.jfr` already has access to `jdk.internal.misc` in 11u. The 17u addition is to `jdk.internal.access`.
> * There are context differences in `JDKEvents.java` due to events introduced in later JDK versions.
> * The `EventNames.java` test includes a huge unrelated change to reorganise the order of the variable modifiers. This was applied manually to the names in 11u, which differ slightly from those in 17u.
> 
> All `jdk.jfr.event` tests passed, including the new one, with the exception of `TestNative`, but that seems to be a setup issue rather than a regression caused by this patch (`Error. Use -nativepath to specify the location of native code`)

Not sure why the Mac port won't build (no error message), but I can't see a Java-only change causing this.

-------------

PR Comment: https://git.openjdk.org/jdk11u-dev/pull/2827#issuecomment-2202757433