RFR: Aarch64 cleanup
Aleksey Shipilev
shade at redhat.com
Wed Jul 5 11:04:32 UTC 2017
On 07/05/2017 12:57 PM, Roman Kennke wrote:
> I'm scrutinizing our JDK9 diff vs. our baseline. Here's some cleanups
> for trivial stuff that I found in aarch64:
>
> http://cr.openjdk.java.net/~rkennke/cleanup-aarch64/webrev.01/
> <http://cr.openjdk.java.net/%7Erkennke/cleanup-aarch64/webrev.01/>
>
> one is a trivial whitespace change, the other removes code that is dead
> because surrounded by #ifdef 0, and the 3rd is removing two
> read-barriers in monitorenter/monitorexit that follow a store-check,
> which is bogus. There's no read-barriers required there because we
> always ensure stuff on the stack is in to-space. That's what the
> store-check is checking. We have no read-barriers in the same place in
> x86 either.
>
> (Note: I found no bad stuff vs upstream in x86)
>
> Ok?
Yup.
-Aleksey
More information about the shenandoah-dev
mailing list