> -----Original Message----- > From: Shilimkar, Santosh > Sent: Wednesday, August 04, 2010 4:08 PM > To: Cousson, Benoit > Cc: G, Manjunath Kondaiah; Kevin Hilman; > linux-omap@xxxxxxxxxxxxxxx; Paul Walmsley; Tony Lindgren; > Sawant, Anand; Nayak, Rajendra; Basak, Partha; Varadarajan, Charulatha > Subject: RE: [PATCH 04/11] OMAP3: DMA: HWMOD: Add hwmod data > structures > > > > > -----Original Message----- > > From: Cousson, Benoit > > Sent: Wednesday, August 04, 2010 4:06 PM > > To: Shilimkar, Santosh > > Cc: G, Manjunath Kondaiah; Kevin Hilman; > linux-omap@xxxxxxxxxxxxxxx; Paul > > Walmsley; Tony Lindgren; Sawant, Anand; Nayak, Rajendra; > Basak, Partha; > > Varadarajan, Charulatha > > Subject: Re: [PATCH 04/11] OMAP3: DMA: HWMOD: Add hwmod > data structures > > > <snip> > > >> > > >> The point is to "enable" the device when at least one > channel is in use > > >> and "disable" it when every channel are released. We do > have only one > > >> DMA IP that managed several channels, since pm_runtime is doing > > >> reference counting, we can call get for each channel > request and put > > for > > >> each release. > > >> Did I miss something? > > > We are saying same thing more or less :) > > > > That was my impression as well, but I was not sure :-) > > > > > Since it's one IP, and say have one clock, then doing per channel > > runtime > > > get/put is redundant, No? > > > > If the DMA driver is already counting the number of channel in use, > > you're right we can avoid that. It is already the case today? > > > Yes. That's the case I was refrering > Agreed. I will incorporate this comment in next version. -Manjunath-- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html