From brian.goetz at oracle.com Mon Aug 2 15:45:03 2021 From: brian.goetz at oracle.com (Brian Goetz) Date: Mon, 2 Aug 2021 11:45:03 -0400 Subject: Fwd: Primitive objects vs memory model In-Reply-To: <6e25759d-1ca8-4f98-06fb-610132eaaf9b@gmail.com> References: <6e25759d-1ca8-4f98-06fb-610132eaaf9b@gmail.com> Message-ID: From the -comments list. -------- Forwarded Message -------- Subject: Primitive objects vs memory model Date: Mon, 2 Aug 2021 16:55:50 +0200 From: Raffaello Giulietti To: valhalla-spec-comments at openjdk.java.net Hello, I'm wondering if the relationship between primitive objects and the memory model (JLS 17 and j.l.i.VarHandle) has been discussed before. I couldn't find aything in JEP 401, nor is the mailing list archive, spanning more than 5 years of intense discussions, easily searchable for keywords. For example, does plain access mode guarantee bitwise atomicity (aka access atomicity) on JEP 401 Point instances (2 double fields for the coords), as if they were 32 bit values? Or perhaps only opaque mode can make such guarantees? Greetings Raffaello From dl at cs.oswego.edu Mon Aug 2 16:20:20 2021 From: dl at cs.oswego.edu (Doug Lea) Date: Mon, 2 Aug 2021 12:20:20 -0400 Subject: Fwd: Primitive objects vs memory model In-Reply-To: References: <6e25759d-1ca8-4f98-06fb-610132eaaf9b@gmail.com> Message-ID: Unless I've missed some follow-up decision, the intent has always been to only guarantee atomicity under opaque-or-stronger mode. Although perhaps something could be said about only extracting one primitive field from a compound value. On 8/2/21 11:45 AM, Brian Goetz wrote: > From the -comments list. > > > -------- Forwarded Message -------- > Subject: Primitive objects vs memory model > Date: Mon, 2 Aug 2021 16:55:50 +0200 > From: Raffaello Giulietti > To: valhalla-spec-comments at openjdk.java.net > > > > Hello, > > I'm wondering if the relationship between primitive objects and the > memory model (JLS 17 and j.l.i.VarHandle) has been discussed before. I > couldn't find aything in JEP 401, nor is the mailing list archive, > spanning more than 5 years of intense discussions, easily searchable > for keywords. > > For example, does plain access mode guarantee bitwise atomicity (aka > access atomicity) on JEP 401 Point instances (2 double fields for the > coords), as if they were 32 bit values? Or perhaps only opaque mode > can make such guarantees? > > > Greetings > Raffaello From daniel.smith at oracle.com Wed Aug 11 14:38:25 2021 From: daniel.smith at oracle.com (Dan Smith) Date: Wed, 11 Aug 2021 14:38:25 +0000 Subject: EG meeting *canceled*, 2021-08-11 Message-ID: <8660E196-FB4A-4FA6-B658-E11BB93C1118@oracle.com> I don't see anything new to discuss, so we'll skip today's meeting. Next meeting: August 25. We're having some good internal discussions about default values & null, and will send something out when that settles into something stable. From daniel.smith at oracle.com Tue Aug 24 23:15:43 2021 From: daniel.smith at oracle.com (Dan Smith) Date: Tue, 24 Aug 2021 23:15:43 +0000 Subject: EG meeting *canceled*, 2021-08-25 Message-ID: Eh, still quiet on here. Guess we'll skip one more meeting. I expect we'll have a useful status update by the next one. Carry on! (I owe Dan a reply on JVMS for JEP 401, which I will get to soon.)