Re: [rvm-research] Why both insertion and deletion barrier in sync1/2 phases of on-the-fly Garbage collector are needed?

"Hosking, Antony L" <[email protected]> Wed, 6 Jul 2016 21:24:32 +0000
Newsgroups gmane.comp.java.jikes.rvm.devel
Message-ID <[email protected]>
Hi Lokesh,

SATB needs only deletion barrier so long as collector transitions are stop-the-world, but when on-the-fly the insertion barrier may be necessary to get across the mutator handshakes to initiate collection.  Our book has an example.  DLG, which since expressed in logic is a GC specification rather than an implementation, optionally disables the insertion barrier when not handshaking.

Cheers,

Tony

On 7 Jul 2016, at 6:03 AM, Lokesh Gidra <[email protected]<mailto:[email protected]>> wrote:

Hi,

Given that the write barrier and object allocation are supposed to be atomic wrt. handshakes, why does an SATB based on-the-fly implementation (like the famous DLG from POPl'94 paper) need to use both, insertion and deletion barriers during sync1 and sync2 phases? Won't just using deletion barrier suffice?


Kind Regards,
Lokesh
------------------------------------------------------------------------------
Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San
Francisco, CA to explore cutting-edge tech and listen to tech luminaries
present their vision of the future. This family event has something for
everyone, including kids. Get more information and register today.
http://sdm.link/attshape_______________________________________________
Jikesrvm-researchers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jikesrvm-researchers

------------------------------------------------------------------------------
Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San
Francisco, CA to explore cutting-edge tech and listen to tech luminaries
present their vision of the future. This family event has something for
everyone, including kids. Get more information and register today.
http://sdm.link/attshape

_______________________________________________
Jikesrvm-researchers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jikesrvm-researchers