Hi Shijith, > -----Original Message----- > From: Shijith Thotton <shijith.thot...@gmail.com> > Sent: Wednesday, March 10, 2021 3:15 AM > To: Carrillo, Erik G <erik.g.carri...@intel.com> > Cc: Shijith Thotton <sthot...@marvell.com>; Pavan Nikhilesh > <pbhagavat...@marvell.com>; Jerin Jacob <jer...@marvell.com>; > dev@dpdk.org > Subject: Re: [dpdk-dev] [PATCH 1/3] eventdev: introduce adapter flags for > periodic mode > > Hi Erik, > > On Tue, Mar 09, 2021 at 08:04:32PM +0000, Carrillo, Erik G wrote: > > > A timer adapter in periodic mode can be used to arm periodic timers. > > > This patch adds flags used to advertise capability and configure > > > timer adapter in periodic mode. Capability flag should be set for > > > adapters which support periodic mode. > > > > > > Below is a programming sequence on the usage: > > > /* check for periodic mode support by reading capability. */ > > > rte_event_timer_adapter_caps_get(...); > > > > > > /* create adapter in periodic mode by setting periodic flag > > > (RTE_EVENT_TIMER_ADAPTER_F_PERIODIC) and resolution. */ > > > rte_event_timer_adapter_create_ext(...); > > > > It looks like periodic support is an operating mode of the adapter > > itself, and that all timers created with a periodic adapter instance > > will be periodic timers. > > > > Is it possible to instead have "periodic/single-shot" be an attribute > > of an event timer itself, such that a single adapter instance could > > support either type of timer? > > > > With single type of timer per adapter, application can decide to create > multiple adapters of required type/mode and use as needed. > > For an adapter to support both type of timers, driver ops implementation has > to follow different paths based on timer type and new capability flag should > be introduced to expose this feature. Our HW only supports single type of > timer per adapter. > > Please let me know the approach you are aligned with.
Having a single type of timer per adapter is surprising initially, in my opinion, but I think it does make using periodic timers simple. Unless there are any other comments to the contrary, I think we can proceed with this approach. I'll respond in a separate post with comments on the patches. Thanks, Erik > > Thanks, > Shijith
