We will re-spin this patch with the following changes: 1) Remove the extra context release introduced in 'scheduler_test_groups' 2) Remove the 'CU_ASSERT_FATAL(ret == 0)' change (or restore it back)
New API will not be handled as part of this patch. Any concerns? Thanks, Honnappa On 10 April 2017 at 02:20, Savolainen, Petri (Nokia - FI/Espoo) <[email protected]> wrote: > > > From: Kevin Wang [mailto:[email protected]] > Sent: Monday, April 10, 2017 5:47 AM > To: Bill Fischofer <[email protected]> > Cc: Ola Liljedahl <[email protected]>; Savolainen, Petri (Nokia - > FI/Espoo) <[email protected]>; Kevin Wang > <[email protected]>; [email protected] > Subject: Re: [lng-odp] [PATCH] validation: scheduler: Release context before > the end of the scheduler test > > So seems a new API odp_schedule_release_context() is required. So I'll > redirect the patch to api-next. Should it be an internal function? > > Your scheduler or this patch set does not require a new API function. So, > better keep it simple and focus on master branch only. Common release_ctx may > be added into the API later on. It has (may have) more overhead than correct > usage of atomic/ordered release calls. > > -Petri > > >
