[Openmp-commits] [PATCH] D105121: [OpenMP] Avoid checking parent reference count in targetDataBegin

[Narayanaswamy, Ravi via Openmp-commits]

>  Are you concerned about a misbehavior here?
I am just answering the question George asked about the expected behavior.

-----Original Message-----
From: Joel E. Denny via Phabricator <reviews at reviews.llvm.org> 
Sent: Friday, July 09, 2021 1:16 PM
To: jdenny.ornl at gmail.com; Rokos, Georgios <georgios.rokos at intel.com>; Narayanaswamy, Ravi <ravi.narayanaswamy at intel.com>; jdoerfert at anl.gov; jhuber6 at vols.utk.edu
Cc: xw111luoye at gmail.com; sourabhsingh.tomar at amd.com; sara.royuela at bsc.es; Saiyedul.Islam at amd.com; ron.lieberman at amd.com; pdhaliwal at amd.com; lin32 at llnl.gov; liao6 at llnl.gov; kuhnel at google.com; jatin.bhateja at gmail.com; greg63706 at gmail.com; deepak.eachempati <deepak.eachempati at hpe.com>; bhuvanendra.kumarn at amd.com; abidmuslim at gmail.com; a.bataev at hotmail.com; tianshilei1992 at gmail.com; clementval at gmail.com; protze.joachim at gmail.com; georgakoudis1 at llnl.gov; yaxun.liu at amd.com; zhang.guansong at gmail.com; stefomeister at gmail.com; openmp-commits at lists.llvm.org
Subject: [PATCH] D105121: [OpenMP] Avoid checking parent reference count in targetDataBegin

jdenny added a comment.

In D105121#2867779 <https://reviews.llvm.org/D105121#2867779>, @RaviNarayanaswamy wrote:

> #pragma omp target map(s1.p[0:10])
>
> Memory and data copy will happened for p[0:10] and the device copy of s1 will be updated with the device pointer to p

Are you concerned about a misbehavior here?

Do you want me to wait to push until some of the subscribers you've added have a chance to review?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D105121/new/

https://reviews.llvm.org/D105121