Subject: Re: Suggested draft TS for TM-light
From: Michael L. Scott (mlscott_at_[hidden])
Date: 2020-01-10 12:58:17
I continue to have reservations about the limitation to constexpr functions, but Iâm ok with submitting, and would like to be an author. One question:
[ Note: The following holds for a data-race-free program: If the start of an atomic block T is sequenced before an evaluation A, A is sequenced before the end of T, and A strongly happens before some evaluation B, then the end of T strongly happens before B. If an evaluation C strongly happens before that evaluation A, then C strongly happens before the start of T. These properties in turn imply that in any simple interleaved (sequentially consistent) execution, the operations of each atomic block appear to be contiguous in the interleaving. -- end note ]
Iâm not sure I completely understand that final sentence. I would have guessed that we needed to say âfor any sequentially consistent execution _there exists_ a (potentially different) sequentially consistent execution in which the operations of each atomic block are contiguousâ. Is that existentiality covered by âappear to beâ? After all, if atomic blocks A and B touch only local variables, their operations can be arbitrarily interleaved in a sequentially consistent execution.
> On Jan 10, 2020, at 12:17 PM, Jens Maurer via SG5 <sg5_at_[hidden]> wrote:
> In the interest of not surprising anyone, I intend to submit
> this to the pre-meeting mailing.
> Anyone on the current authors list that doesn't expressly
> approve being there will be moved to an "Acknowledgements"
> section before submission.
> On 09/01/2020 22.06, Jens Maurer via SG5 wrote:
>> After today's teleconference, I couldn't resist, so
>> I created a draft TS for TM-light, mostly copying
>> from Victor's Google doc. See attached.
>> The new part is in 8.8, where we decree implementation-defined
>> behavior when "bad" things are evaluated inside an atomic block.
>> Possibly surprising changes vs. the status quo of the discussion:
>> - constexpr functions only
>> (as explained in the teleconference, the compiler might not retain
>> other functions from the same translation unit)
>> - throw-expressions are forbidden entirely
>> (Did we want to require support for throw/catch within the atomic
>> block? Seems more-than-minimal.)
>> - Coroutines are forbidden entirely.
>> (I haven't seen talk about them vs. TM. Also, they have funny
>> control flow.)
>> This would be in time to submit to the pre-Prague meeting,
>> for SG1 and EWG amusement there, if I get approval on this list
>> until Monday.
>> I've added all participants to the authors list, please complain
>> if that's wrong.
> SG5 mailing list
SG5 list run by herb.sutter at gmail.com
Older Archives on Google Groups