RFR: 8262073: assert(allocates2(pc)) failed: not in CodeBuffer memory

Yangfei (Felix) felix.yang at huawei.com
Tue Feb 23 01:46:35 UTC 2021


Hi again,

> -----Original Message-----
> From: jdk8u-dev [mailto:jdk8u-dev-retn at openjdk.java.net] On Behalf Of
> Andrew Haley
> Sent: Sunday, February 21, 2021 2:08 AM
> To: jdk8u-dev at openjdk.java.net
> Subject: Re: RFR: 8262073: assert(allocates2(pc)) failed: not in CodeBuffer
> memory
> 
> On 20/02/2021 12:49, Yangfei (Felix) wrote:
> >   Performed full jtreg test on aarch64-linux-gnu.  OK?
> 
> OK, looks reasonable.

May I ask to make one more change here? 
In addition to rework vtable/itable stub size calculation, JDK-8207343 introduced this safety check [1] in aarch64 during itable stub creation. 
With this change, we won't use CodeCache for itable stub allocations if it is already full instead of failing the entire VM. 

For aarch64 8u, we already have the check for vtable stub creation, I think we should also add same check for itable stub case to be safe. 
New webrev adding this extra check: http://cr.openjdk.java.net/~fyang/8262073/webrev.01/ 
Still passed full jtreg test. 

[1] http://hg.openjdk.java.net/jdk/jdk/rev/54b344d9dd4e#l1.122

Thanks,
Felix


More information about the jdk8u-dev mailing list