> ________________________________________ > From: ext Y, Kishore [kishore.y@xxxxxx] > Sent: Wednesday, December 23, 2009 12:22 > To: Valkeinen Tomi (Nokia-D/Helsinki) > Cc: Kevin Hilman; linux-omap@xxxxxxxxxxxxxxx > Subject: RE: [PATCH 0/3] OMAP: DSS: Low power refresh rate support for dss > > > -----Original Message----- > > From: Tomi Valkeinen [mailto:tomi.valkeinen@xxxxxxxxx] > > Sent: Tuesday, December 22, 2009 4:44 PM > > To: Y, Kishore > > Cc: Kevin Hilman; linux-omap@xxxxxxxxxxxxxxx > > Subject: Re: [PATCH 0/3] OMAP: DSS: Low power refresh rate support for dss > > > > Hi, > > > > On Tue, 2009-12-22 at 10:37 +0100, ext Y, Kishore wrote: > > > These patches add Low Power Refresh functionality for DSS. > > > Short info on patches > > > > > > DSS PM_WKEN to refill DMA > > > - DMA request to refill the FIFO will wake up the CORE domain > > > DSS2 LPR enable and disable function definition > > > - LPR enable and disable functions implementing FIFO merge > > > DSS2 sysfs enteries to export LPR enable to user > > > - sysfs entries for above methods > > > > I don't like this approach very much. Fifo merge should be automatic, > > not user configurable. It should be integrated into the apply/cache > > mechanism in manager.c. > > > > I actually implemented fifo merge at some point during DSS2 development, > > but I had to remove it when changing the driver/device model. I haven't > > had time to implement it again, as it not very trivial. The old version > > can be found in N900's kernel, and earlier snapshots of DSS2. > > Can you point me to the git for N900 code or ur old dss code? I couldn't find it. I don't know if there's a git tree, but the kernel source can be downloaded from, for example, http://repository.maemo.org/pool/fremantle/free/k/kernel/kernel-source_2.6.28-20094102.3+0m5_all.deb Use "dpkg -x file.deb ." to extract. > Please suggest: > How do we decide when to enable or disable lpr if we include it in manager.c? > Should we be doing it based on video pipeline usage...? If by LPR you mean fifo-merge, I think it's simple: fifo-merge can be used when only one overlay is in use. And as I don't know any drawbacks with fifo-merge, fifo-merge should be used whenever there's only one overlay in use. The complication comes from configuring fifo-merge. If I remember right, fifo-merge bit is not shadowed, but overlay disable/enable is, and thus you need to be careful that the overlays really are disabled when enabling fifo-merge etc. Tomi Ps. Removing Kevin from cc, I don't think this concerns him.-- 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
